Professional Documents
Culture Documents
Master Thesis
Author(s):
Yu, Qianchen
Publication Date:
2018-04
Permanent Link:
https://doi.org/10.3929/ethz-b-000257364
Rights / License:
In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For more
information please consult the Terms of use.
ETH Library
Qianchen Yu
Master’s Thesis
Supervision
M.Sc. Arne Meeuw
Prof. Dr. Elgar Fleisch
Dr. L. Andrew Bollinger
April 2018
Abstract
To mitigate climate change and increase energy efficiency, researchers have studied a wide range
of solutions, for instance, smart grid, district heating, renewables, and storage. Dropping feed-
in tariffs encourage end-users to take a more active role. Renewable energy integration, system
operation, transparent billing, and data integrity, are common challenges for grid operators when
adopting the aforementioned solutions. A holistic view of those technologies in operation (i.e., on
district or community level) still has many aspects to be investigated and optimized.
One kind of Distributed Leger Technologies (DLTs), blockchain, has been flourishing for the past
decade. It has particular advantages in securing immutability and guaranteeing transparency in
transactions. Its decentralized nature and automatic computing feature bring the potential of
adding value in the energy sector, especially regarding energy dispatching, billing, as well as energy
financing.
The first three chapters of this thesis briefly introduce the basics of blockchain, the concept of dis-
trict multi-energy systems, and summarize the latest academic and industrial projects of blockchain
technology in the energy industry. Moreover, this thesis proposes a novel process to coordinate, al-
locate and settle energy transactions in a district multi-carrier energy system. The process operates
in a decentralized way, fully on-chain.
The design leaves producers the freedom to choose their preferred pricing strategy for profit maxi-
mization. The price-availability-based allocation system guarantees consumers the lowest possible
cost. In addition, an incentive mechanism for maximum local distribution of energy was introduced.
Each device reacts for the common good of the community and follows the principle of relieving
the stress of operational limits on the lower grid level.
This design was implemented using the Ethereum blockchain and was tested with real consumption
and production profiles of electricity and heating systems. Three pricing strategies, namely zero-
intelligence real-time-pricing, inversed-production peak pricing, and game-theory based pricing,
were compared and evaluated. An analysis of the energy consumption and operation cost of such
a system was performed in the end.
In comparison to other pricing strategies, the game-theory based pricing strategy yields the best
local energy dispatch in the district energy system. More than 80 % of the energy demand from
households are able to be supplied by community green energy resources. The consuming house-
holds also benefit economically from the dynamic market competition and are able to get the lowest
prices. The battery in the system effectively reduces the capacity requirement of the grid.
i
Contents
1 Introduction 1
1.1 Distributed Ledger Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Blockchain Techonologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Technical Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2.2 Adding a Validated Transaction on the Chain . . . . . . . . . . . . . . . . . 6
1.3 Machine-to-Machine (M2M) Friendly Blockchain . . . . . . . . . . . . . . . . . . . 7
1.4 Comparison of Some Featured Blockchain Platforms . . . . . . . . . . . . . . . . . 8
1.4.1 Ethereum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5 Prospect of Blockchain Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4 Project Setup 27
4.1 Design of the Energy System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1.1 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.1.2 Sizing and Data Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.2 Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.3 Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.4 Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
4.4.1 Zero-Intelligence Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.4.2 Inversed-Production Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.4.3 Game-Theory Based Pricing . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
4.5 ÐApp Solution Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
iii
5.2.2 Scenario 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5.2.3 Scenario 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.2.4 Scenario 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.2.5 Scenario 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.2.6 Scenario 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.3 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
5.4 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.4.1 Time and Gas Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
5.4.2 Analysis and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
A Appendix 87
A.1 Source Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
A.2 Data Retrieve Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
A.3 ÐApp implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
A.3.1 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
A.3.2 Action Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Bibliography 93
iv
Chapter 1
Introduction
With a significant rise in the price of Bitcoin since 2016, cryptocurrency and its underlying technol-
ogy - blockchain - has created headlines. Despite many research streams that cautiously evaluate
the real potential of blockchain [1–3], industries, especially startups, contribute dynamically for the
evolution of the technology. Plenty of blockchain-based concepts and applications have emerged
across a wide span of sectors, including finance [4, 5], digital identity [6, 7], agriculture [8], health
care [9], smart city [10], real estate [11], entertainment [12], intellectual property [13], precious
metals [14], and diamonds [15].
The regulation of cryptocurrency and blockchain technology have also been in the spotlight of
many countries. A recent assessment [16] from the Congress’ Joint Economic Committee of the
United States includes an entire section on cryptocurrencies and blockchain, which is interpreted
as the latest sign that the U.S. government is taking the subject of blockchain and cryptocurrencies
seriously.
In this chapter, I introduce the general features of blockchain technology providing some basic
concepts and technical terms. They aim to pave a fundamental layer for readers who are not
familiar with this technology to quickly grasp the essentials for further reading. As the chapter is
going to cover only a small piece of the iceberg, those who are interested in particular areas are
encouraged to dig through the recommended references provided at the end of each section.
1
2 1.1. Distributed Ledger Technology
Blockchain The first type of DLT - blockchain has many well-known applications, such as Bit-
coin [21] and Ethereum [22]. Technical details related to this data structure will be discussed in
detail in section 1.2.
Figure 1.1: The Hashgraph data structure developed over time by including events in exchange
of validating two other events. Each event holds a collision-resistant hash function, where its
transactions and parent nodes are encoded. A validated event means the hash matches parent and
valid signature. An event included in the Hashgraph leads to two valid parent nodes, therefore four
grandparent nodes, resulting in a valid record of a complete history of its ancestors with edges
linking between these ancestors. Nodes hold an identical copy of the hashgraph via "gossiping".
Each node constantly talks with another randomly selected node, to share the events they validated
in topological order. The other node includes the unknown ones into its local ledger [23].
Directed Acyclic Graph (DAG) Two typical DLT applications using a DAG structure are
IOTA (see section 1.3) and Hashgraph.
Figure1.1 shows the data structure managed by one node in Hashgraph. Vertices in Hashgraph
are called “events”. An event consists of a set of transactions, the event’s parents, a timestamp,
and a signature from the node that created the event [24]. Nodes continuously communicate with
each other by randomly selecting a peer and sending events that have not been validated yet.
The timestamp indirectly determines the sequence in the network of events. By validating the two
selected events, a new event is included in the network and thus updated among all nodes. Unlike
the blockchain, the DAG applied in permissioned systems accelerates the transaction speed. In
Harshgraph, more than 250,000 transactions per second pre-sharding1 can be reached in a fair
(mathematically proven fairness with consensus timestamping) and secure (Asynchronous Byzan-
tine Fault Tolerant) manner [23].
Tempo Tempo ledger is the core technology for Redix, a DLT platform that is IoT friendly
planning to be launched in Q3 2018. An instance of Tempo is called a Universe, divided into
shards. Any event within a Universe, such as a message or transaction, is represented by an object
called an Atom, Payload Atom and Transfer Atom respectively.
1 Sharding is to divide the network into several smaller component networks (called shards) that are capable of
Figure 1.2: Map of a Temporal Proof process that verifies state changes and reaches consensus
across shards. It starts with the validation of the Atom (αx ) on a node (N1 ) where it contains
the shard of the destinations. If the validation succeeds, node N1 passes the Atom on to the next
node. This process repeats until the required number of nodes validates the Atom or any validation
process fails [25].
Figure 1.2 shows the validation process of an Atom. All Atoms contain at least one end-point
destination, represented by an end-point address. End-point addresses are derived from an identity,
such as a user’s public key. End-point addresses are used to route events through the network.
Before an event can be presented to the entire network for global acceptance, an initial validation
of the event is performed by a subset of nodes which, if successful, results in: A Temporal Proof
being constructed and associated with the Atom, a network-wide broadcast of the Atom, and its
Temporal Proof [25].
Among all the data structures for DLT, blockchain is only one of the possible, yet the most
famous subcategory. It is a distributed, transactional database, creating a shared ledger system
maintaining the integrity of data [26]. Globally distributed nodes are linked by a peer-to-peer (P2P)
communication network with its own layer of protocol messages for node communication and peer
discovery [27].
When a new data entry is agreed upon, in accordance with a consensus protocol, it is packed
together with other agreed entries taking place at the same time. This package of transactions,
as well as the hash of the latest block, is then sealed into a new "block". This block is appended
to the longest chain of blocks that have reached consensus previously. This structure links all the
valid records in sequential order. Section 1.2.2 demonstrates the full process.
The signed, chained data records provide security, anonymity, and data integrity [28], without a
third party’s verification of the authenticity of the transaction [29]. With this primary advantage,
many industries envision the blockchain technology to be the backbone of the disruptive revolution
in the future, regardless of its current technical challenges and limitations. Section 1.5 gives some
examples of these challenges and limitations.
Due to the novelty of the underlying technologies, there is no well-known definition of DLT and
blockchain. Researchers and practitioners may confuse these two terms [30] despite the quality of
the work they deliver in the field of blockchain. Some studies [17] define the common interchangeable
usage between the term "blockchain technology" and "distributed ledger technology".
4 1.2. Blockchain Techonologies
Glaser et al. [31] gave a proposal for the taxonomy of the decentralized consensus systems. How-
ever, due to rapid development, their work might not be comprehensive. The definition of certain
terminologies might not be adaptable for emerging projects, but the bone of this proposal does
give an excellent and valuable overview of the blockchain-related ecosystem.
Before demonstrating the workflow of blockchain in section 1.2.2, some critical technical termi-
nologies are introduced in this section for better understanding.
Permissioned and Permissionless DLT Distributed ledger systems can either be permis-
sioned or permissionless. In a permissioned DLT system, all the members are pre-selected by either
the owner or an administrator of the ledger, who owns higher rights than other participants. The
transactions in a permissioned system are less costly in comparison to a permissionless system
because users of the permissioned system do not necessarily need to pay for the computing power
when during validating transactions. However, the anonymity of data on such a chain is less. It
holds the risk of ex-post modification by the admin account. In contrary to a permissioned DLT
system, a permissionless chain is open to any user. Each user is identical to another. Table 1.1
shows a comparison between these two chains.
Public and Private Keys Each user/agent can hold account(s) on the blockchain. On a per-
missionless blockchain, there is no limit of creating account(s); While permission is required if a
new user/agent wants to participate in a permissioned blockchain, the creation of the account is
subject to the application-specific conditions. Upon the creation of an account, a pair of public
and private keys is generated. Keys are large numbers and represented in hexadecimal or base 64
Chapter 1. Introduction 5
form. The public key serves as a publishable identifier in the network. Every account holder, or
even an outsider (for public blockchain), can see the public key and eventually track the activities
related to the public key. The private key is, on the other hand, kept by the owner of the account.
This asymmetric cryptographic system performs two activities: authentication and encryption.
In authentication, user A uses its private key as a digital signature to approve transactions on the
chain. Other users can validate the transaction by checking with A’s public key.
In encryption, user A uses the public key of user B to encrypt a message. User B can later decrypt
the message with his private key.
Smart Contract Smart contracts are scripts running on the blockchain that can process inputs
automatically and provide results according to the predefined logical calculation.
Hash A hash function used in cryptography is a function that maps data of arbitrary size to
data of fixed size. An ideal cryptographic hash function can easily produce a hash for any input,
but it is difficult to use the hash to derive the input. Additionally, any changes in the original data
would result in extensive and seemingly uncorrelated changes to the hash [32]. The hash function
serves for guaranteeing the immutability of the blockchain technology.
Consensus Algorithm To make all nodes in a decentralized system reach a common global view
on the information spread on the network, a set of embedded rules/algorithms agreed by all nodes
on the chain is pivotal. The consensus mechanism is the key to avoid information asymmetries.
A few of the commonly used consensus mechanisms have been listed here. Tschorsch et al. [33]
conducted a comprehensive review on the Proof-of-X algorithm.
• Proof-of-Authority (PoA)
Nodes with particular keys hold the status of validators. Only validators are allowed to add
blocks on the chain. They stake their reputation and status to secure the blockchain.
• Proof-of-Work (PoW)
The verification of transactions in the Proof-of-Work (PoW) system is conducted by solving
a complex cryptographic computation by the nodes in the network. The work is the compu-
tation time that relates to the transactions contained in the block, which is difficult to fake,
thus, considered as a proof of validity [31].
Bitcoin and Ethereum use the SHA-256 algorithm [34] and Litecoin uses Scrypt [35].
• Proof-of-Stake (PoS)
The miners who carry out PoW computations can be viewed as entities who vote on blocks of
transactions that the users recently broadcasted on the network, so that the decision-making
power of each miner is in proportion to the amount of computational power that it has. While
in the Proof-of-Stake (PoS) system, the validation of transactions and mining of new blocks
depend on the stake, i.e., the amount of cryptocurrency of the account. Miners need to put
up some cryptocurrency as collateral. The larger stake the account has, the higher chance it
has to mine a block successfully. The rationale behind the Proof-of-Stake algorithm is that
entities who hold stake in the system are well-suited to maintain its security, since their stake
will diminish in value when the security of the system erodes [36]. Although this consensus
mechanism can significantly reduce the energy consumption related to the mining, depending
on the actual algorithm a chain applies, there is a risk of encountering the "Nothing-at-risk"
6 1.2. Blockchain Techonologies
problem. There are some PoS types of alternatives being introduced against the aforesaid
problem [36].
Figure 1.3: A diagram of the data structure of Bitcoin’s transactions [21]. A digital asset is passed
from the owner 0 on to the owner 3 via a series of transactions. The grey box in the middle
represents the transaction from owner 1 to owner 2. To authenticate this transaction, a signature
from owner 1 is needed. This signature is made when the public key of the destination account
(owner 2) and the previous transaction (the grey box on the left) are hashed and signed by the
current owner (owner 1) with its private key. This transaction is verified by comparing the public
key of the owner 1 with its signature and the signed message. If successful, the owner of the asset
becomes owner 2.
Figure 1.3 gives an example of the registration of transactions on a blockchain. A digital asset is
passed from its old owner 0 on to the owner 3 via a series of transactions. The grey box in the
middle represents the transaction from owner 1 to owner 2. To authenticate this transaction, a
signature from owner 1 is needed. This signature is made when the public key of the destination
account (owner 2) and the previous transaction (the grey box on the left) are hashed and signed
by the current owner (owner 1) with its private key. This transaction is verified by comparing the
public key of the owner 1 with its signature and the signed message. If successful, the owner of the
asset becomes owner 2. Such data structure works for all transactions on the blockchain.
After the validation of a group of transactions effectuated during an agreed-upon time interval,
these transactions are collected and sealed into a ready-to-be-chained block at a certain timestamp.
This action is undertaken by alleged miners. They are chosen according to the consensus mechanism
of the blockchain.
Chapter 1. Introduction 7
After packaging a new block, miners broadcast it onto the network. The nodes then add the block
onto the longest chain, once the hash of the previous block and valid transactions are verified to
have existed.
Apart from the pure digitalized application, to leverage blockchain in real-world use cases, the
combination of blockchain and physical objects is necessary. The linking Internet of Things (IoT)
devices with blockchain is vital. Christidis et al. [39] investigated the potential and constraints of
such a development. A blockchain network assures a unique, transparent, and immutable shared
database to verify the integrity and validity of applications, as well as to provide access for other
service providers for revenue generation. It also ensures the anonymity of IoT devices [40].
Recent research and projects have revealed the progress on blockchain hardware integration.
EthEmbedded allows Ethereum blockchain to run on small single-board computers, such as Rasp-
berry Pi and BeagleBone [41]. If these mini-computers are still too power-hungry for some mer-
chandises in certain supply chain applications, Moore’s law may help to solve this issue. IBM
announced the smallest computer in the world in March 2018 [42]. With the computing power
similar to an x86 chip, a shape comparable to a grain of salt and the cost of merely 0.1 $, this
blockchain-enabled innovation may lift the industrial scale blockchain-IoT solutions to the next
level.
However, the data collected from the IoT devices are extremely large. By using the data in real-
time, can one benefit their full value? Moreover, due to the mining algorithm of the blockchain,
the computational overhead induced increases both energy consumption and processing delay.
Therefore, a suitable blockchain is crucial to guarantee a cost-effective operation of IoT on a
blockchain.
In this section, I briefly introduce two blockchain projects that are designed for Machine-to-Machine
communication.
IOTA
IOTA is a platform for IoT data sharing in exchange for payment within its community. Each user
needs to find a neighbor to be able to participate in the platform. The data collected from IoT
devices are communicated in the form of transaction in the network.
The validation process of IOTA is similar to that of Ethereum, as it solves puzzles similar to a
hash puzzle. The transaction storage mechanism used by IOTA is called tangle, a directed acyclic
graph (DAG). It uses Markov Chain Monte Carlo (MCMC) algorithms to constitute the tangle
graph [43]. The tangle incentivizes the maintenance of its consensus mechanism among each and
every user itself rather than having a clear role of a miner. Whenever a transaction takes place, the
sender must validate two previous transactions. The validation work is imposed on all the active
participants on the chain, which, in turn, reduces transaction time and eradicates the fee.
The advantage of IOTA is that there is no fee involved in the transaction. Nevertheless, it is
still questionable about whether their consensus will work. The prerequisite of preventing double-
spending in IOTA is letting all nodes update with all the transactions. Besides, the requirement
for hardware to have an additional chip to run the chain will induce an increase in production
cost. What’s more, IOTA does not natively support smart contracts. The IOTA Foundation has
publicly stated that smart contracts will not be a feature of the IOTA core but provided as a layer
on top [44].
8 1.4. Comparison of Some Featured Blockchain Platforms
Streamr
Also working in the IoT space, Streamr [45] shares similar vision and mission with IOTA. It enables
live storing and sharing of IoT data streams (through the API of hardware). Unlike IOTA, Streamr
chooses to build on top of an existing chain, rather than building an IoT specific blockchain.
Streamr has three major components: Streamr Network, Streamr Engine, and Streamr Marketplace.
The Streamr Network is the infrastructure for data communication between nodes; Streamr Engine
allows data processing, visualization, and interaction with its blockchain; Streamr Marketplace is
a platform for data trading. This three-layered design helps to minimize the operation cost and
assures the transaction time by differentiating between on-chain and off-chain communication.
Streamr is blockchain agnostic. The Streamr Network runs off-chain. Currently, only the Streamr
Engine, which is in charge of user authentication and payment, runs on Ethereum.
Another solution to the M2M data communication is via the distributed data storage. To prevent
the high cost related to the large volume of data flooding into the blockchain, IoT devices can
store the data on a distributed data storage system, for instance, IPFS [46] and Swarm [47], and
get the returned key, which is the hash of data registry. It can later store the unique hash key in
the smart contract for sale. Once the financial settlement is done, the new owner of the data set
may retrieve the data from the same distributed data storage system with the hash key.
1.4.1 Ethereum
Table 1.2: Comparison of blockchain technologies: Ethereum, Hyperledger Fabric and Corda [48]
is to move into a proof-of-stake system, Casper [50]. This fork is expected to be fully implemented
in both main network and testnets in 2018.
Another challenge in Ethereum is its scalability. Factors such as the block size limit, transaction
processing capacity, limit Ethereum to be scalable. The network can currently process ~15 trans-
actions per second. Several projects tackling the scalability issue are planned to be implemented
in 2018:
• Sharding will alleviate the workload of each node by running the blockchain only partially [51].
• Raiden network will enable off-chain processing [52];
• Plasma allows upgradable smart contracts [53].
These planned projects bring more value to Ethereum blockchain. Right after the release of the
Casper update on the testnet, the price of Ether rose to 1000 $.
For our project, the ideal blockchain platform is Ethereum, because it is a proven working blockchain
network, with a well-designed ecosystem and smart contract editing feature, albeit the transaction
cost and time may leave only a little margin for operation.
Figure 1.4: There are concerns on trust associated with various levels in the Blockchain Engineering
Framework. The critical issue happens between the behavioral layer and the agent layer, when
the communication crosses the "trust frontier". As many use-cases of blockchain technology need
actual interactions of agents in the real-world (in the behavioral layer), behaviors done outside the
closed blockchain ecosystem needs to translate into the agent layer. However, the trust provided by
blockchain technology can only be guaranteed when the application operates as a closed ecosystem
within its boundaries, i.e. inside the "trust frontier" [55].
virtual money investment - then what is the most cost-efficient way for
tracking and verification? [16]
Cybersecurity The security to protect the digital assets against theft and fraud. The
DAO hack happened in 2016 drained more than 3 million USD worth of
Ether away from the crowdfunded account by an anonymous hacker [56].
How shall the loss of value be recognized legally? How to enhance the
cybersecurity for blockchain-based application in general?
Privacy In a permissionless blockchain, although only the public key of each
account is published at the moment of a transaction to promote trans-
parency while maintaining privacy to a certain degree, the risk of leak-
ing privacy still exists. By analyzing all the transactions interacted with
certain accounts, a behavioral pattern may be extracted [57].
Legal enforceability Legal enforceability of smart contracts is limited. Regardless the com-
pliant execution of an audited smart contract, a transacting entity may
dispute the outcome of a smart contract operation [39].
Societal issues With the automation of smart contracts and transparency in data
transaction, blockchain technology can substitute trusted third par-
ties, for instance, platform providers. However, trust problems are re-
Chapter 1. Introduction 11
A multi-energy system (MES) combines multiple energy carriers, whereby electricity, heating, cool,
and so on optimally interact with each other at various levels (for instance, within a district, or
a city, or at a regional level). An MES with distributed generation of multiple energy vectors,
described as distributed multi-generation (DMG), is a key option to decarbonize the energy sector
while improving the technical, economic, and environmental performance [59].
To reduce CO2 emissions, Europe envisions to have at least 27 % of the power produced by
renewable energy by 2030 [60]. Besides carbon-free power generation, renewable energy sources are
also sulfur-free alternatives for conventional fuel supplies, which is a key component in abating air
pollution for countries that still heavily rely on fossil fuels, i.e., China and India [61].
Not only for the benefit of the environment are renewable energy sources the critical piece for the
future, but it is also a good choice for investment against the rise in interest rates. The world
economy has raised the likelihood of tightening monetary policies. The long-term interest rates
have increased from 2 % to 2.5 %. In this direction, the high-CAPEX and low-OPEX technologies
would be more competitive [62]. As there is no fuel required to generate electricity with renewable
energies, such as with solar energy and wind power, the marginal cost of renewable energies is
almost zero. The nature that against the uncertainty in conventional market risk incentivizes the
adaptation of the renewables.
To leverage renewable energies to its full extent, an integration with battery system can improve
its flexibility and increase the reliability of energy supply. As the cost of battery has decreased
over the years, a 20–30 % yearly cost reduction could bring conventional lithium-ion batteries at a
commercial or utility scale to the point of mass adoption potential before 2020 [63]. When the cost
of batteries reaches around $100 {kWh, the mass adoption of batteries will redefine energy supply
with innovative integrated solutions and novel business models.
Distributed Energy Resources (DERs) include small-scale generators (e.g., rooftop solar panels),
small-scale energy storage systems, plug-in electric vehicles (EVs), and household appliances with
energy storage capabilities and flexible power requirements that permit them to function as virtual
batteries [64]. There are primarily renewable energy sources (RES) in the DERs.
13
14 2.2. From Centralized to Distributed Energy Generation
With the growing penetration of distributed renewables in the market, the traditional electrical
grid structure is exposed to the potential risk of broad area failure. A small grid failure may lead
to cascading outages and entice large-scale blackout [65]. The need for flexible power and ancillary
service provision has increased [66].
One solution is to segment the wide-area synchronized power system into small or medium sized
cells at the distribution network level and control the subsystems (or "cells") asynchronously [67].
The size of these subsystems may vary from a single household to a city or region.
A district-level multi-energy system is an ideal representative of such subsystem that sits in the
distribution network. The electrical network of such a system remains at a low-voltage (LV) level,
with the connection to the grid via controllable transformer(s), giving a single connection to the
medium-voltage network. The complexity of operation is reduced for such a system in comparison
to a system that involves management of a higher voltage network. Within the same setting, a
district heating network and other bi-directional DERs, like electric vehicles, batteries and hydrogen
storages, could also be added up on top to form a district multi-energy system.
The flourishment of distributed renewable energy generation has challenged the operation of mod-
ern electric power systems. China will reach a turning point where it will build more distribution
grid-connected solar projects than large transmission grid-connected projects. China will also dou-
ble the volume of behind-the-meter solar projects built to 14GW in 2018, according to a study
of Bloomberg New Energy Finance [62]. Although a distributed generation connected to lower
voltage grids is geographically closer to the end-users, which relieves line stress in the grid when
production runs smoothly, the line stress suddenly increases when the distributed energy sources
cannot provide necessary energy, as it occurs when PV cells do not receive sunlight for example.
The grid still needs to have lines of substantial capacity to provide back up when the storages are
depleted. The high variability and volatility of its energy production make this type of system’s
energy scheduling more complicated both for the long-term (over seasons or years) and for the
short-term (over 15 min in the intra-day energy dispatch) [68]. The demand for flexible power
and ancillary service is therefore increasing. One of the solutions is a Transactive Energy System
(TES).
A TES is defined as a set of economic and controllable mechanisms that permit supply and demand
of power to be balanced over time across an entire electrical infrastructure, where these mechanisms
are designed to enhance value for the transacting parties consistent with overall system reliability.
An electrical infrastructure consists of electrical devices connected to a physical grid. The electrical
devices are capable of supplying power (generation) and/or of absorbing power (consuming/storing
power), and the grid permits the delivery of power from supplier devices to absorption devices [66].
To date, researchers have primarily focused on the implementation of TES designs for distribution
systems, to facilitate the reliable and efficient management of prosumer-owned distributed energy
resources (DERs) [64]. Here the term Prosumers refers to the decision-making entities that have
the possibility of either providing energy or consuming energy, or even doing both at the same
time.
Since there are mainly prosumers in the aforesaid district-level multi-energy system, the TES model
is suitable for the electric network in such a system.
Challenges
As mentioned in the previous sections, the stability of energy supply and flexibility of the grid are
issues caused by the rising share of RES. These problems can be solved from either demand side
or supply side, or from both sides.
Chapter 2. District Multi-Energy System 15
There are several general methods that guarantee stability and flexibility of the distribution level
grid [69, 70]:
• Supply-side solutions
– Flexible generation
Generation plants with rapid ramp-up rate can adjust its generation schedule based on
the current load profile, for instance, combined-heat-and-power (CHP) plant and gas
plant. They can reduce their production when RES has a surplus or rise the production
in case of deficiency in RES.
– Intelligent scheduling
The production of RES has a heavy dependence on the environment. By integrating
sensors and advanced data analytics of artificial intelligence (AI), suppliers can gain
insights in consumption prediction, with which the production can be accommodated.
– Improving network utilization
By adding power flow control devices and voltage control devices for instance Phase
Shifting Transformers, FACTS devices, HVDC lines).
• Demand-side solutions
– Demand Response
Demand response (DR) is a broad term that encompasses all the options for end-users
to actively operate their electrical loads in order to support the operation of the power
system. It covers a range of time-scales and services, from hourly load shifting to long-
term capacity provision, and further for end-use efficiency improvements [71].
– Dispatchable load
Dispatchable loads are devices of which the consumption can be switched to timing
where the energy system has more capacity. A typical example of dispatchable loads
is electric vehicles. These loads are, in general, managed by the network’s stability
managing party, per user’s preference [72].
• Energy storage
Storage devices can store directly in a battery or convert it to another form of energy, such
as hydrogen or flywheels for further energy conservation during the peak generation period.
This energy storage can later be converted back or directly retrieved according to the need
on the demand side and the network condition [73].
By 2016, around half of the total energy consumption in EU was for space heating, domestic hot
water, and process heat [74]. Compared to the centralized heating, district heating improves energy
efficiency. The short transmission distance of district heating curtails the loss in the heating system.
The shared infrastructure reduces its deployment cost in comparison with the fully decentralized
heating system. The optimal point that balances the energy efficiency and financial stress provided
by district heating makes countries expand its application. District energy networks also promote
the deployment of renewables. Especially, heat pumps and electric boilers can shape the peak of
electricity production and provide flexibility for variable renewable electricity [75–77].
An estimation of IRENA shows the demand for district heating in Switzerland will increase by
21PJ in 2030 compared to 2015, which amounts to a share of 7.9 % as a fraction of total national
heating and cooling demand comparing to that of 6.0 % in 2015 [78]. Admitting the portion of
renewables in district heating remains small - only 7 % in 2015 globally and 28 % in the European
Union. The high cost of electricity and gas in Switzerland by 2030 will make the exploitation of
renewable resources highly competitive [79]. Swiss cantons introduced incentive policies in favor of
the penetration of renewable district heating [80].
16 2.3. Efficient Operation for DER
Figure 2.1: Prediction of the levelised cost of district heating technologies in Switzerland by 2030. It
indicates an advantage of cost-saving in deploying heat pumps in Swiss district heating system [78].
The main options of renewables for district heating are biomass, solar thermal, geothermal, and
heat pumps [79]. The choice of the technology depends on the availability of the local resource
and the cost. District heating powered by biomass and geothermal has a dominant presence in
Switzerland [78]. However, IRENA’s REmap programme [79] predicts that new large-scale electric
heat pumps will lead the market, counting for 4.7 PJ/year as final district energy demand. The
levelised cost of heating shown in figure 2.1 also indicates an advantage of cost in deploying heat
pumps in Swiss district heating system.
2.2.3 Storage
In the district multi-energy system, the imbalance between the load and supply creates stability
problems as well as raises the marginal cost of generation for both electricity network and heating
network.
On the distribution level, storage can help shave the peak load, leading to a more stable electricity
supply [81, 82]. In association with power electronics, controllable energy storage system holds
various advantages on both technical and financial aspects. Mohd et al. [83] gave a comprehensive
overview on this point. For district heating system, there is some inherent thermal energy storage
carried in the circulating water, with which the gap between the heat load and the production can
be smoothened to some extent due to its lantency [84].
Depending on the technology, its predictability and responsibility vary. Some storage technologies
need few seconds or up to minute to respond to the control signal. The more the storage is pre-
dictable, the more suitable it is for urgent demand, such as frequency control. In general, there
are wide options of technology for a specific requirement, ranging from the regenerative fuel cell
to compressed air energy storage (CAES), battery and thermal [85].
To stimulate effective interactions with network users, price signals can play an important role [86].
The design for the policy may be based on various principles, including efficiency, equity, simplicity,
Chapter 2. District Multi-Energy System 17
Figure 2.2: Techno-economic alignment of decentralization in electricity markets [85].In our project,
we are going to use the direct load control method driven by nodal price signals, to reduce the
impact on system operation and to maximize the utility function for each prosumer in the district
multi-energy system.
consistency, transparency, stability, and additivity [85]. Among the incentive schema, there are two
general guiding methods: price-based method and interruptible demand response [87].
As shown in the figure 2.2, Eid et al. [85] summarized the appropriate incentives concepts or control
methods for different use cases. In our project, we are going to use the direct load control method
driven by nodal price signals, to minimize the operational impact technically as well as maximizing
the utility function for each prosumer in the district multi-energy system.
Many dynamic pricing strategies, meaning a time-varying energy rate, are available and have been
investigated by researchers. Table 2.1 shows some of the possible ones. Two pricing strategies
presented in chapter 4, Zero-Intelligence Pricing (cf. to section 4.4.1) and Inversed-Production
Pricing (cf. to section 4.4.2), are derived from the Real-Time-Pricing (RTP) and Critical Peak
Pricing (CPP) models respectively.
18 2.3. Efficient Operation for DER
Table 2.1: Possible dynamic pricing options for DER management [85].In the following chapter, two
basic dynamic pricing options, Real-Time-Pricing (RTP) and Critical Peak Pricing (CPP), together
with a Maximum-Community-Sharing pricing option based on game theory, are presented.
Chapter 3
This chapter provides an overview over current blockchain-in-energy projects taking place in
academia and industry. We will attempt to compare these projects, analyze the trend of the de-
velopment and summarize the challenges.
The intersection between blockchain and distributed energy system is the inherently distributed
nature, as well as the need for keeping a record of the energy and financial transactions. As the
features of both systems tend to go into the same direction, many studies and applications are
carried out in this field.
These projects can be grouped into two categories: One is about energy dispatching and billing,
the other one is about tokenization and derivatives.
In the first category of energy dispatching and billing, studies focus on the energy transactions
taking place on two levels: Firstly the "TSO or DSO centric" top-down applications, and secondly
the "community-level" bottom-up applications.
TSO and DSO stand for the Transmission System Operator and the Distribution System Operator
respectively. They are jointly in charge of maintaining the quality and the security of electricity
supply [88]. The "TSO or DSO centric" top-down applications aim at implementing the blockchain
technology in the existing power market to improve processes, find a new business models and new
revenue streams for traditional players in the energy industry. Projects under this scope usually
focus on business-to-business (B2B), targeting the energy wholesale market. In the electricity
wholesale market, a blockchain-based trading system can reduce the overhead cost and increase
transparency of billing [89, 90]. PONTON leads a project called Enerchain [91], in which many
major European utility companies are involved in the peer-to-peer wholesale trading and grid
management process optimization. BP, Eni Trading & shipping, and other partners launched a
pilot project of a full-cycle gas trading platform [92].
19
20 3.1. Current Blockchain Applications in the Energy Sector
The "community-wise" bottom-up applications are in general projects that examine energy trading
on the distribution level. They encompass architecture design, incentive mechanisms, allocation
strategies, control systems, etc.
Concerning the architecture design, Lombardi et al. [93] introduced a blockchain-based three-
layered infrastructure architecture for the transactive grid. The three layers are the Policy Man-
agement Layer (PML), the Energy Trading Layer (ETL) and the Security Enhancement Layer
(SEL). The ETL and SEL were implemented as smart contracts. This architecture allows one to
carry out energy auctions within the grid, isolate prosumers’ hosting devices with known vulner-
abilities and update the latter as soon as security patches are applied. However, no testing of the
architecture design was conducted.
For the energy allocation strategy, Marshall et al. [94] developed a fractional accounting structure
and a quota-based accounting structure for P2P electricity allocation rules, as basic structures
for tariff and incentive designs. Sikorski et al. [95] tested a bilateral settlement. They built a
proof-of-concept (PoC) for real-time machine-to-machine energy trading, where a bilateral finan-
cial agreement between two electricity producers and one electricity consumer over blockchain is
presented. Lamparter et al. [96] proposed an agent-based double-auction mechanism. Mengelkamp
et al. [97] implemented a closed discrete double auction market with single market clearing price.
Yan et al. [98] summarized several patterns in the design of distributed energy trading mechanisms,
especially on the principle of energy allocation, including:
Some studies research the day-ahead scheduling with real-time control. Munsing et al. [99] im-
plemented a distributed optimization and control system to schedule a mix of batteries, shapable
loads and deferrable loads for the day-ahead market, using the Alternating Direction Method of
Multipliers (ADMM), on the Ethereum blockchain. Giotitsas et al. [100] introduced and evaluated
a P2P electricity sharing network that consists of loads, PVs, and batteries. It is operated by Power
Management Unit (PMU) with day-ahead scheduling. The system enables scalable deployment of
distributed generation to provide less expensive electricity in off-grid areas, compared to Solar
Home Systems (SHSs).
Linking to the predominant application of blockchain technology [101], cryptocurrency, many busi-
nesses tend to monetize the positive externalities of renewable energy via tokens. The idea of these
tokens is first to quantify the renewable energy generation and exchange the metered value with
a certain amount of tokens. These tokens can, therefore, be considered as a proof of the energy
generation or renewable energy capacity [102]. With this proof, users can either trade the virtual
energy on the market [103], or trade the carbon credits given by certifying the amount of CO 2
reduced thanks to the usage of renewable energy [104].
Use case Country Start Year Participating Entities Usage of blockchain Size Highlight
Digital,Grid Platform Corporation,
University of Tokyo,
Neighbourhood
Tateyama Kagaku Group Inc. ,
Measurement and execution result recording; (5 household prosumers,
PROTON - Kansai Electric Power Co. Treat power clump of constant magnitude as commodity,
[67] Japan 2018 Price matching; 10 household consumers,
Urawa Misono Ltd, Tokyo Electric Power Company Holdings,Inc. , P2P control with Digital Grid Router (DGR).
Receive virtual currency; and PV producers
Hitachi EI system,
in a large shopping center AEON.)
Tessera Technology Inc.
NTT Data Inc.
District first plan to demonstrate peer-to-peer energy transactions
BlockCypher Settle energy transactions
unknown [106] U.S.A 2018 (two test homes in NREL’s involving distributed, energy resources (DER)
National Renewable Energy Laboratory (NREL) across multiple blockchains
Energy Systems Integration Facility) over the dash cryptocurrency network
Curtin University,
Murdoch University,
Curtin Institute of Computation,
LandCorp,
unknown CSIRO/Data61, Blockchain-powered distributed City ? Field test on how cities can effectively use blockchain
[107] Australia 2017
(Power Ledger) CISCO, energy and water system (City of Fremantel) technologies to moderate energy and water usage.
Power Ledger,
Australian Energy Market Operator (AEMO),
Western Power,
CRC for Low Carbon Living.
Innogy, City
Conjoule [108] Germany 2015 P2P energy trading among rooftop PVs Purchase energy on local market
Tokyo Electric Power Company (TEPCO) (Essen Kettwig and Mülheim)
Hardware "agent"
Grid+ [109] U.S.A 2017 ConsenSys Direct access to wholesale energy markets.
energy prepaid with either Ether or fiat currencies
Distributed marketplace for OTC trading of
Ponton,
Trading physical power and gas products, wholesale energy products (B2B);
Enerchain [91] Europe 2017 Enel, European power market
spot and forward for any European delivery zone. TSO, DSO and aggregators centric grid management;
E.On
Blockchian agnostic
District Testing a range of blockchain use-cases in anticipation of
unknown [110] Austria 2018 Wien Energie
(Wien Viertel Zwei) rolling out fully-fledged services.
Brooklyn P2P energy and gas trading platform: Aiming at grid-level service provision
[111] U.S.A 2016 LO3 Energy District
Microgrid transactiveGrid (DER aggregation, balancing, wholesale trading)
Electron, National Grid, Siemens, Baringa,
Meter Registration platform; A shared market place for energy assets
Statkraft, Shell, Northern Powergrid,
Energy eBay [112] UK 2017 Flexibility trading platform; to respond to pricing signals;
Power Networks, Flexitricity, EDF Energy,
Community Energy Projects; Bilaterial trading first test in the middle of 2018;
Open Energi and Kiwi Power
Successful in all 8 test scenarios;
BTL, BP, Eni Trading & shipping, Wien Energie
The pilot proved that Interbit blockchain
Eni Trading & Shipping,
European gas trading on can streamline many trading and back office processes
OneOffice [92] Europe 2018 Total, Gazprom Marketing & Trading Limited,
BTL’s Interbit platform; across the energy trade life cycle, such as,
Mercuria,
confirmations, actualizations, invoice generation, settlement,
Vattenfall, Petroineos, and Freepoint
audit, reporting and regulatory compliance.
P2P Energy trading;
City
Power Ledger [113] Australia 2016 Power Ledger wholesale market settlement;
(Perth)
Electric Vehicles and carbon trading
P2P energy trading with preference of
Piclo [114] UK 2014 Open Utility, Good Energy
renewable energy provides and end users
Grid Singularity,
Grid Singularity [115] International 2016
Energy Web Foundation
Daijie [116] UK Transaction recording IoT-enabled P2P energy exchange;
AI powered high frequency energy trading;
Drift [117] U.S.A 2014 Drift Energy purchase
with DLT.
The first retail electricity marketplace in Southeast-Asia
Electrify [118] Southeast Asia 2017 Electrify.Asia Energy purchase from retailers or P2P. addressing the need for
transparency and security in the consumption of energy.
Table 3.2: Summary of some energy related tokens
WPR token Renewable energy producers can pre-sell the production in the global market
WePower [121] 2017
and energy token in exchange of capital.
Renewable energy finance.
MPQ token MPQ token allows purchasers votes on the decision of project deployment;
ImpactPPA [122] 2017
and GEN Credit GEN Creadit is the energy currency to purchase energy created
by the generation systems.
TGR Veridium Labs [123] 2017 Carbon credit trading
Climatecoin Climatecoin Foundation [104] 2017 Carbon credit trading
Crowdfunding PV down to the level of an individual solar cell and leasing
SUNEX the Sun Exchange [124] 2014
them to schools and businesses in Africa.
A peer to peer investing application which unlocks billions of dollars
MyBit Mybit [125] 2016
in IoT revenue.
Purchasing existing sustainable energy systems and smart devices
GREEN Greeneum [126] 2015 on the open marketplace.
Purchasing Greeneum Carbon Certificates .
23
24 3.2. Unresearched Points
Apart from these two main trends of applications, there are also projects testing the potential of
blockchain applications in other circumstances. For instance, in the Burgundy region of France, a
blockchain-enabled infrastructure is connected to water meters. By analyzing the metering data,
the system automatically triggers a call to a repair team in the event of a leak [105].
A summary of the industry paradigms is in table 3.1. This table is extended from [127–129].
Unfortunately, some projects have not sufficiently revealed their details to the public until the
completion of this thesis. Some fields are left blank for this reason. Some examples of token projects
are briefly summarized in table 3.2.
Project Setup
As described in the previous chapter, our project analyzes the near real-time multi-carrier energy
dispatching and settling system. The dispatching logic is implemented in smart contracts. Energy
transactions and financial settlements are both recorded on a blockchain.
The multi-carrier energy system that we investigate is derived from the eHub demonstrator of
EMPA in a simplified form. The system consists of prosumers and two networks: an electricity
network and a heating network. The electricity network comprises electricity storages and the
heating network consists of heating storages, heat pump and heat exchangers.
eHub is an energy research and technology transfer platform, aiming at optimized energy manage-
ment at the district level and evaluating its influence on the overall energy system [132]. Together
with other Research and Technology Transfer Platforms (RTTPs) set up by Empa, such as Next
Evolution in Sustainable Building Technologies (NEST ) [133] and move [134], it simulates real-
world conditions to test specific projects.
There are several units in the NEST that represent, e.g. offices, apartments, a wellness center.
A combination of these units is comparable to a small district community. Therefore, this work
requires a layout based on the actual setup of eHub for a district multi-energy system and uses
real data collected from sensors installed in units, in order to represent actual human behaviors.
Nevertheless, there is a disparity between the layout of eHub and the intention of this project, so
is between the configuration of this project and that of eHub.
In order to illustrate and evaluate the working pattern of each type of device in a rather simple
way, the project is not a duplicate of the complex NEST system, but rather emphasizes only on the
essential representative equipments of the complex NEST system. Hence, certain units in NEST
are chosen and scaled for our project accordingly. For instance, the size of the water tanks is scaled
down and the capacity of PV is scaled up. The characteristics of the devices are specified in the
following sections.
4.1.1 Architecture
As shown in figure 4.1 and figure 4.2, there are three households, three photovoltaic (PV) panels,
one battery, three heat pumps and four hot water tanks in the two networks of our project. Each
device included in the system - represented as a node i - can be characterized by its energy load.
During one day in the near real-time energy trading system, a node can consume or supply energy
27
28 4.1. Design of the Energy System
Figure 4.1: Layout of the electricity network in our district multi-energy system. It comprises of
three households, three photovoltaic (PV) panels, and one battery.
(or remain inactive) at a timeslot t. The energy load of such a device at time t is denoted as li ptq.
Table 4.1 shows the abstraction of different devices.
When a device consumes energy, meaning it receives electricity or heat from other devices, its
energy load is denoted as li ptq ¡ 0. Such a device is considered as Consumer C. Examples for kind
of devices are households and offices. When it produces energy, meaning there is electricity or heat
generated by the device, flowing to other devices, the energy load of such a device is li ptq 0.
Such device is classified as Producer P, for instance, PV panels. If a device can store energy, there
are three possible behaviors at that time t. If the device charges at time t, it acts as a Consumer,
thus li ptq ¡ 0. If it discharges energy, it works as a Producer, thus li ptq 0. li ptq 0, if the device
is not active. Such a device is called Storage S, by which load can be shifted.
It worth noticing that not all devices
that have the aforementioned three possible behaviors are
considered as storage, i.e. S pC P q. The grid G, can feed energy into the system in case
of production deficiency in the system and absorb energy in case of surplus, while its energy is
transient. It requires a sequence of control to maintain voltage stability, otherwise the energy fed
into the grid would dissipate into useless heat or cause frequency drops when too much energy
is extracted. Similar to the grid, heat pumps consume electricity (consumer) and generate heat
(producer), depending on the device or network they address. Although, the grid and heat pumps
can behave as consumers or producers, due to their transient and interface-like properties, they
are not considered as Storage.
House The project simulates a small community powered by as much renewable energy as pos-
sible. We have used three residential and commercial buildings, all generally named as houses
Hi , i P r0, 2s to represent different energy consumption patterns.
Each house consumes electricity and heat from the respective network. The electricity load is
measured in real time with a smart meter. However, it is more complicated to track the heat
Chapter 4. Project Setup 29
Figure 4.2: Layout of the heat network in our district multi-energy system. It comprises of three
households, three heat pumps, and four hot water tanks, among which one is blacklisted by House 0 .
consumption in real time because there is no direct measurement for the supplied heat. Instead,
heat consumption can be calculated with measurements, such as temperature.
In residential and commercial building, there are two types of heat supply, i.e. Domestic Hot Water
(DHW) at medium temperature and Space Heating (SH) at low to medium temperature [135]. Fis-
cher et al. [136] introduced a stochastic bottom-up model for space heating and domestic hot water
load profiles, in which both physical (e.g. ambient temperature 4.2b and season fluctuation 4.2c)
and behavioral 4.2d aspects are considered. This model is used to describe the heat consumption
of households in our system.
Equation set 4.1 shows the model for calculating SH energy consumption. Equation set 4.2 shows
the model for DHW. The total demand for domestic heating is finally calculated.
Cm building mass
∆Tm temperature change
∆t time step
30 4.1. Design of the Energy System
Q9 dhw ptq fseason ptq Q9 losses ptq 9 ptq cw Tw,h ptq Tw,c,0
m rWs (4.2a)
Tω,c pnday q Tamb 3K nday nday,offset rKs
2π
(4.2b)
365
Tω,c,nominal Tω,c pnday q
fseason pnday q 1 (4.2c)
∆Tω,0
Q9 g,int ptq Pel ptq npers ptq 65.0 rWs (4.2d)
Where
PV Among all the possible renewable electricity technologies, PV is the most viable and cost-
effective for distributed households. Also, as there are PV panels installed on the NEST demon-
strator, we can leverage the real-time data input. PV panels are the only energy producer in this
system. PV systems can be acquired by single households or jointly by multiple members of the
system. There is no locational restriction or amount limitation of PVs. Via a bilateral agreement or
a leasing contract, the owner of a PV system can place it on the rooftop of another person’s house.
In all the cases, energy generated by a PV system should be first supplied to the house at which
the panel is located to prevent unnecessary energy losses. If the household owns the PV system,
the two are combined and considered as Prosumer. However, in our specific market design, there
is no necessity to treat both as a pair. Among the three PV systems P Vi , i P r0, 2s, H0 and H1
decide to share the resources from their rooftops, P V1 and P V2 , having the same access priority
in terms of energy usage.
Battery The system has a battery B0 shared among all members of the community. The battery
can be connected to PV systems as a solution for system flexibility, where it shaves the production
peak and smoothens the consumption peak. At each time-slot t, the battery B0 can decide whether
to actively acquire a certain amount of energy or not, to set a certain price for selling stored energy.
However, in case of unbalanced demand and supply in the electricity network, the battery must
react and is obliged to balance the system. In this case, the amount of energy stored in B 0 follows
Chapter 4. Project Setup 31
pmaxq
0 ¤ VB0 ptq ¤ VB0 . On the contrary to its energetic obligations, the battery can independently
set the price for its balancing service.
The ownership may vary depending on the analyzed case. For instance, a battery can be the
property of a particular household, serving for the optimization of the household’s energy cost, if
the investment is 100% private. Another case could be that the battery is owned and operated by
a Distribution System Operators (DSO). In the LEAFS project [137], the DSO is entitled to the
operation of storage units for grid stability purposes [138].
In our particular design, the DSO cannot operate the battery B0 . However, it is possible to intro-
duce storages to the system for other purposes, especially for the stability of grid. It needs to be
noticed that any solution introduced by the DSO, shall be differentiated from the properties of the
community. Grid stability follows different economic logics than the behavior of prosumers that
try to maximize their profits. DSO is allowed to participate in the local market as an entity and
charge users for the provided grid service.
Grid As a backup for the district energy system, the grid can provide energy in case of surplus
or deficiency by virtue of controlling the system and balancing with other users in the same
network. The grid can set the energy price and feed-in tariff, based on grid’s condition at each time
slot. Due to grid security and congestion issues, there is a capacity limit for the grid, denoted as
pminq ¤ l ptq ¤ lpmaxq .
lg g g
The connection of each device or household to the grid is not mandatory. In a rural area where
the grid is less accessible, there is no other choice than operating the energy system as an island
system. Our application is designed with the capability of customizing the connectivity to the grid
for each device. It is set by default to be connected to the grid. To reflect the layout of eHub, all
the devices are connected to the grid.
Water Tank Water tanks (W Ti , i P r0, 3s) provide flexibility for heating supply. As mentioned
above, DHW and SH are at different temperature level. To efficiently conserve hot water at two
temperature levels, water tanks need to be temperature specific. There is at least one water tank
dedicated to each domestic heating supply, as shown in figure 4.2. In line with the layout of
eHub [139], the medium temperature water is stored at TM 38 C while the high temperature
water storage operates at TH 65 C.
Each water tank is linked to a heat pump as the only source of thermal energy supply. The reason
will be detailed in the paragraph concerning the heat pump.
Each household’s water consumption is directly requested from the linked water tank with the
volume of the hot water at a particular temperature T at time t. Similar to storage capacity of the
pmaxq
battery, water tanks operate in the volumetric range of 0 ¤ VW Ti ptq ¤ VW Ti , i P r0, 3s. A water
tank sends water to the household once the action is approved by the smart contract. It declares
the demand for hot water supply to heat pumps (in L), according to its information on the current
profile and the usage pattern.
Heat Pump A heat pump is an effective means for a real-time setup compared to other thermal
generators. It works as an interface between electricity network and the heating network, which
uses purchased electricity to heat up the ground water to a desired temperature (cf. paragraph on
water tank).
In order to simplify the model, the heat pump is the only device to supply the heating network.
Noted that it is also possible to supply hot water with water at lower-temperature hot water and
do the reverse cycle with heat exchanger. However, for the simplicity of the network, we consider
the sole option of heating from ambient water resource and stored in the water tank.
32 4.1. Design of the Energy System
Depending on the output temperature, heat pumps used in our system are classified into two types:
high-temperature (HT) heat pump and medium-temperature (MT) heat pump. They can also be
written in short as "HHP" and "MHP" respectively.
The performance of a heat pump is characterized by the Coefficient of Performance (COP), where
COP QWH , indicating the amount of heat QH that is possible to be produced per amount of
electricity W .
V9 ρw cw Tw,h Tw,c,0
W
QH
(4.3)
COP COP
Where
V9 volume of water
cw specific heating capacity of water
ρw density of water
Tw,h hot water temperature
Tw,c,0 cold water temperature
To simulate a live energy dispatching and trading, the load profile is simulated iwht 15 min intervals,
thus 96 sets of data are needed for an entire day. The daily electricity and hot water consumption
for all the devices are real data retrieved from the Energy Hub (eHub) demonstrator on an averaged
typical winter day, constructed from data from December 21 and 22, 2017. We only take a typical
winter day here because of the state of the NEST project: Only by winter 2017, the buildings and
sensors were ready to retrieve useful data. Table A.1 summarizes the parameters used in the SQL
query of Empa eHub database.
Household The appliance usage and thus the consumption pattern for each household is inher-
ently different. On the electricity side, the consumption profile of H0 is derived from the averaged
profile of three "Offices" at NEST on December 21, 2017. The consumption profile of H1 is ab-
stracted from the averaged "apartments" of NEST on the same date. The consumption profile of
the third household H2 is derived from the average "apartments" profile on December 22, 2017.
The original data represent the power consumption for each minute. Therefore, data are integrated
into 15 min slot, as shown in figure 4.3a.
The usage profile of hot water is retrieved from the hourly load profiles of the work from Fang et
al. [140]. The consumption in each hour is divided evenly over the 15 min intervals for individual
trading, as shown in figure 4.3b.
PV The generation profiles of PV panels are taken from NEST’s rooftop PV installations. The
detailed query entries are6 listed in the table A.1. The generation profile is constructed by using
the accumulated energy in each 15-minute interval throughout the day as shown in figure 4.4.
Battery The battery shared within the community has a capacity of 200 kWh with an initial state
of charge of 100 kWh. The unit price for an electricity producer in case of excess energy absorption
is very low, at 3 cent CHF/kWh, while the sales price for electricity is 40 cent CHF/kWh.
Grid The feed-in and supply power limit are both 100 kW, therefore 100 kW ¤ lg ptq ¤ 100 kW.
Feed-in and supply are both priced at 40 cCHF/kWh for simplicity reasons.
Chapter 4. Project Setup 33
House0 15 House0
House1 House1
30 House2 House2
Consumption (kW)
Consumption (L)
10
20
5
10
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) Daily electricity consumption per household (b) Daily hot water consumption per household
Figure 4.3: Electricity and water consumption profile per household on a winter day.
25
PV0
PV1
20 PV2
Production (kW)
15
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Water Tank Here we collected data from the eHub database with parameters listed in table A.1.
Then, the volumetric fluctuation pattern over a day is retrieved. The hot water tank used in NEST
serves for a whole neighborhood-level building, hence its size is too large comparing to the size of
our system layout. Therefore, the consumption pattern from the orignial data (cf. to figure 4.5a)
is scaled down from a volume of around 1000 L to around 100 L - 120 L (cf. to figure 4.5b). The
difference of the consecutive time slot is the active water demand requested from the heating
network (cf. to figure 4.6) eWTi VWTi pt ∆tq VWTi ptq
Heat Pump According to the technical specification of heat pumps in the eHub project, the
medium-temperature (MT) heat pump has a COP of 4.5 and the high-temperature (HT) heat
pump is has a COP of 3. Since the scripting language Solidity of Ethereum cannot properly handle
floating numbers, the calculation is simplified by rounding parameters are rounded to integers.
Upon the calculation using equation 4.3, assuming the ambient temperature of remains around
3 C in December in Zurich, equation set 4.4 is the electricity-heat conversion function for heat
pumps. Equation 4.4a means that to supply 11 litres of water at medium temperature 38 C with
a MT heat pump, it consumes 0.1 kWh. Equation 4.4b shows that 0.1 kWh of electricity can heat
34 4.2. Policy
1,000
100
800
80
Volume (L)
Volume (L)
600
60
400
40
200 20
Water Tank0 Water Tank1 Water Tank0 Water Tank1
Water Tank2 Water Tank2
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
Figure 4.5: Original and processed volume profile per water tank of eHub on a winter day
40
Consumption (L)
30
20
10
Water Tank0
Water Tank1
0 Water Tank2
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 4.6: Hourly hot water consumption profile per water tank of the examined system on a
winter day
WMT Vwater
11
[L]
0.1 kWh (4.4a)
WHT Vwater
5
[L]
0.1 kWh (4.4b)
4.2 Policy
With regard to forging solidarity with the community, increaseing the penetration of renewable
energies, and minimizing cost of energy, here we present a policy that incorporate metrics for
availability, connectivity and priority of devices in the system. We assume that with appropriate
hardware (including communication devices, controlling syte, etc.) the system can safely operate
and devices are able to be switched on and off accordingly.
The principle of trading in this system is to empower the community and to incentivize the usage
of renewable energies while minimizing the cost for each participant.
Chapter 4. Project Setup 35
In general, the connectivity between devices limits the choices of trading among all devices. Let δi,j
denotes the connection between node i and node j. If they are linked, energy can be transferred via
the distribution grid with δi,j 1, otherwise δi,j 0. If a producer and a consumer are connected,
there is the possibility to transfer energy between them. This condition is therefore expressed as
δi,j 1, i P P, j P C
Here I assume that households have the willingness to purchase as much renewable energies as
possible compared to conventional energy or mixed energy supplied by the grid, as long as the
energy transfer is possible over the grid (connectivity check) and the price is reasonable. All the
energy generated or stored in the district multi-energy system is considered as community energy.
When households query the available energy supply, the community energy owns a higher priority
over the energy from the grid. Within the community, households tend to select the supplier with
the lower price, in order to reduce their cost. Moreover, cost and green energy are not the only
criteria. Households are not merely price-takers, but they balance against other players in the
community by a reputation-preference system. This means that households have the right to put
malicious or unavailable devices into a blacklist. In this case, even if two participants are physically
connected, their bids are not placed on the same table.
PV systems have the incentive the sell as much generated energy to the community as possible.
PV systems need to set a unit sale’s price pP Vi ptq for the produced energy at a given time interval
t. Since they are the price setter and the transaction settlement partially relies on the price, they
need to use an appropriate pricing strategy to manage the destination of their generated energy.
When there is a surplus in generation, there is no other way for PV systems than to sell to batteries
at a close-to-zero price or feed back into grid. However, since the community-owned battery has a
higher priority, the PV systems have to sell the energy to the battery, if battery is still capable to
accept, no matter the attractive feed-in price provided by the grid.
The general rule for the battery is to purchase energy at low prices and sell it at high prices.
Therefore, the battery should have a good overview and prediction on the total energy generation
and consumption in the entire system. The battery has the right to decide whether it actively
declares the need for charging. If bids, the declared amount of energy is secured to be supplied, at
a price defined by the supplier(s). However, if the battery chooses not to bid, there is still possibility
that it receives the some energy. This takes place when there is surplus of energy generation in
the system. The excess energy first goes into the battery, at a dumping price that is significantly
lower than its selling price. The point of active bidding is to assure that the volume of the battery
remains at a certain level for further trading. In case, the battery foresees a decrease in renewable
energy generation in the system, the connected consumers will start to ask for energy supply from
the battery, δB0 ,ci 1 and lci ¡ 0. Battery is therefore certain to sell energy at its preferred prices,
to make profits. If the battery does not actively bid in the first place, it has lower priority for
energy supply. In this case, ther is a risk of not being able to sufficiently charge in preparation for
further energy trading.
The reason for a non-uniform (i.e. heterogeneous) within the system is due to asymmetric connec-
tivity and availability of devices. The resources are not evenly distributed within the system. If all
nodes were in a fully meshed network, the unit price would converge to a uniform market price.
Furthermore, the price of renewable energies is mainly lower than that of conventional or mixed
energy provided by the grid, as it internalizes the positive social, environmental and economic
externalities, such as no SOx or NOx emissions, climate change mitigation, positive contribution
to the community.
The following section 4.3 delineates operational steps for the introduced policy and section 4.4
details the pricing strategies of the system used for further testing.
36 4.3. Mechanism
4.3 Mechanism
A 15-minute trading interval consists of 5 steps. Each step has an allocated period of time, which
is globally coordinated by one smart contract. Figure 4.7 illustrates the sequence of actions in one
round. Activities take place at the same time in both networks.
Figure 4.7: Timeline of the 5-step trading mechanism. Left-hand side is for electricity network;
Right-hand side is for heating network.
During the first step, all the devices register basic information for the upcoming 15 min. All the
devices are aware of the expected load except for transient devices, i.e li ptq, i P Dz tC X P u Y S is
set, where D is the set of all the nodes.
Figure 4.8: Consumers rank offers from suppliers based on updated unit price and the status of
update. The community green energy supplier with the updated lowest price ranks on the top of
the list. The grid is added at the bottom of the sorted pricing list.
In the second step, all the consumers send out price queries to renewable energy producers or to
storage units of the community. After all the queries have been responded to, consumers sort all
the possible offers according to the updated unit price, as shown in figure 4.8.
Chapter 4. Project Setup 37
A response of a pricing query comprises of three pieces of information: the address of the producer
i P P, the unit price of its generated energy for the upcoming trading round pi ptq, and the status
of update. If the device succeeds in replying the pricing query in time, the status of update is
calculated as true, otherwise, the status of update is false. The received responses are sorted by
the following guidelines:
• Sort from the lowest price to the highest price.
• Rank down the unresponsive devices.
The lower the price is, the higher its rank is. However, if the unit price is not updated for the
upcoming time slot, there is a chance that further communication with this particular device may
fail. Therefore, devices with outdated information are placed at the bottom of the list. The option
of buying energy from the grid is attached at the end of the list, due to the lower priority it has
comparing to devices in the community.
At the same time, the heat pump receives the demand of hot water from its connected water
tank. At this step, it converts the requested volume into needed electricity with equation 4.4 and
broadcasts the demand to connected devices.
Meanwhile another action from water tanks also take place. To calculate the new price for the
upcoming trading period, each water tank needs to recalculate the averaged unit price for the
stored hot water. Since the hot water supplied by the heat pumps is using electricity at a different
unit price, the price for re-selling this hot water should also be changed over time according to the
equation 4.5.
° n1 ° n1
pn1 max V n1 Hi Vout,i , 0 i Vin,i pni 1 fprofit
pn V ptq
(4.5)
pn designates the averaged unit price of a water tank during round n, which takes place during time
t. pn1 denotes the averaged unit price of the water tank during round n 1, which is equivalent
to time t ∆t. fprofit is the factor for gaining profit. At fprofit 1, the newly calculated price
equals the cost of hot water generation, at which the water tank breaks even. If fprofit ¡ 1, the
water tank makes profit.
Figure 4.9: Producers sort the list of consumers based on their price-oriented preference. A con-
sumer who puts the producer on the top of its sorted pricing list (small number in Rank), is ranked
up on the list of the producer. Total is the second criterium to consider. A smaller Total value
represents fewer other energy producers that a consumer is connected to, therefore, a higher possi-
bility that this producer supplies its generated energy. Consumption is the last critierum to define
the position of a consumer on the producer’s list. Consumers with large Consumption is ranked
down.
Moving on to the thrid step, it is the producers’ turn to ask for the sorted result from their
connected consumers and make their own list of ranks based on the preference of consumers, as
shown in figure 4.9.
In the returned information, there are four elements: Consumption, Rank, Total, and Updated.
Consumption contains the amount of energy a consumer demands. Rank means the place that this
38 4.4. Pricing
producer holds among all the possible energy suppliers in the particular energy consumer’s list.
Total tells the number connected producers that a consumer has. Updated indicates whether this
piece of information is up-to-date. The guideline for producers to sort their connected consumers
follows:
• Outdated information is not taken into account for the sorting.
• Sort all consumers based on its Rank. The smallest Rank is placed on the hightest in the list.
• For two items with the same Rank, the smaller Total connection ranks up.
• The device with larger consumption is ranked down.
The smaller the number of Rank is, the sooner an energy transaction between the two parties
should be taking place because the producer provides the best price for this particular consumer.
If multiple consumers have allocated the same Rank number, the second sorting criterium to look at
is the total amount of connected producers this particular consumer has. The smaller this number
is, the more scarce energy is for this user or, in other words, there is not much community power
accessibility for this household. Therefore, users with a smaller Total number are given priority. If
both of these measures are identical, the less the Consumption a consumer has, the higher its rank
is. This is to promote the sharing behavior with the highest number of community members and
the idea of reducing unnecessary energy consumption for conservation.
In the heating network, the water tanks asks houses for the hot water consumption at this step.
Until now, electricity grid and heating grid are working in parallel and independently. Starting
from step 4, information between two networks starts flowing.
At step 4, nodes with negative energy load initiate energy transactions based on the sorted rank
list. If one heat pump receives energy during this stage, it immediately initiates another transaction
in the heating network to water tank, as it consumes the electricity to produce heat. Similarly, in
the heating grid, each water tank transacts hot water to the corresponding households.
After this step, there may still be residues of generation or lack of supply. This is the final clearing
phase where producers and consumers act at the same time with storage or grid for the surplus or
deficiency.
It is possible for devices in the community to put some of their connected devices on the "blacklist".
If a producer is on the blacklist of a consumer, this consumer does not send pricing query to the
producer in the second step. If a consumer is blacklisted by a producer, the producer does not send
the ranking query to the producer in the fourth step. This mechanism allows punishment against
deceit or unaffordable energy price.
A sequence diagram that illustrates the activities in the electricity network is in Appendix A.3.2.
The concrete list of action for each device is listed in the table A.2.
4.4 Pricing
From a system-level perspective, a trading mechanism built on a pure financial strategy cannot
necessarily perform better technically [141]. This means that an optimization based on price or
cost does not lead to an optimization for the system’s physical operation. In some cases, techni-
cal difficulties (e.g. congestion, instability) may occur and induce an operation default/security
concern/damage on the infrastructure, etc. Those increase the maintenance cost and dispatching
effort required by the grid operator, who will in return charge the energy users with higher bills.
Therefore, a good design of pricing strategy that internalizes all the potential externalities, such as
operation constrains, environmental impacts (carbon taxes) and human factors (encourage certain
connection within the community) would be beneficial.
Chapter 4. Project Setup 39
Pricing strategies are adopted by the devices in the system with information provided by agents.
Agents act as the owner of the device, to perform price setting or consumption setting. Here we
Scenario 4 Scenario 6
Scenario 5
Figure 4.10: Relationship between pricing scenarios. Scenarios with the same type of shading have
identical connectivity among devices in the system. Scenarios with the same color (in the same
column) share the same pricing principle. In terms of pricing strategy, scenario 1 (in light blue)
follows the zero-intelligence pricing strategy. Scenario 2, 4, and 5 (in dark blue) adopt inversed-
production pricing strategy. Scenario 3 and 6 (in orange) use game-theory based pricing strategy.
In terms of the connection of devices, scenario 1, 2, 3, and 6 remain the proposed layout shown in
figure 4.1 and figure 4.2. In scenario 5, the battery is connected to all devices, while there is no
battery in the scenario 4.
adapt several pricing scenarios in three categories as shown in figure 4.10. First, zero-intelligence
pricing (light-blue area in figure 4.10, containing scenario 1), a basic pricing strategy where agents
have little information on both the performance of their own and others’ devices. This makes
the pricing strategy totally random and independent of any system impulses. Secondly, inversed-
production pricing (dark-blue area in figure 4.10, containing scenarios 4 and 5), where agents need
to predict the production of their own devices for the next 15 min interval based on historical
data and set up the price according to a general demand-supply relation. Each agent performs
independently, relying only on the historical information of its own device. Thirdly, a collective
non-cooperative pricing (orange area in figure 4.10, containing scenarios 3 and 6), where each energy
producing agent can adjust the selling price based on the behavior of other energy producers who
hold a similar role, to maximize their utility function. Agents firstly share necessary information
among each other and make their own decisions in combination with the historical performance of
their own device.
Figures 4.11, figure 4.12, and figure 4.13 illustrate the price profile of PV systems under those three
types of pricing strategy.
In this thesis, the trading policy was based on the pricing signal. Each device can decide on its
own pricing mechanism. To briefly sum up the difference between each scenario, comparing to the
basic layout shown in figure 4.1 and figure 4.2:
• Scenario 1
Basic layout, zero-intelligence random and independent pricing.
• Scenario 2
Basic layout, prices are set up based on an inversed relation with production.
• Scenario 3
Basic layout, unit prices are adjusted by 3 cent CHF/kWh per time step based on the actions
taken by other similar devices at previous steps.
40 4.4. Pricing
• Scenario 4
Based on the basic layout, however, the battery is isolated from all other devices. Prices are
set up based on an inversed relation with production, identical to that of scenario 2.
• Scenario 5
based on the basic layout, however, the battery is connected all the producers and consumers.
• Scenario 6
Basic layout shown in figure 4.1 and figure 4.2, unit prices are reduced by 20 % of the current
price at each time step based on the actions taken by other similar devices at the previous
steps.
Price (CHF/kWh)
Price (CHF/kWh)
Production (kW)
Production (kW)
Production (kW)
14
30 10 30 30
20
12
20 20 20
5 15
10
Production Production Production
10 10 10
Price Price Price
0 10 8
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h)
Price (CHF/kWh)
Price (CHF/kWh)
30
Production (kW)
Production (kW)
Production (kW)
15
20 10 20 20
20 10
10 5 10 10
Production Production 10 Production 5
Price Price Price
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Price (CHF/kWh)
Price (CHF/kWh)
Production (kW)
Production (kW)
Production (kW)
20 15 20 20 20 20
10
10 10 10 10 10
Production 5 Production Production
Price Price Price
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 4.11: Production and Figure 4.12: Production and Figure 4.13: Production and
price profile of PVs in the 1st price profile of PVs in the 2nd price profile of PVs in the 3rd
category, on a winter day (Dec category, on a winter day (Dec category, on a winter day (Dec
21, 2017) 21, 2017) 21, 2017)
Chapter 4. Project Setup 41
The zero-intelligence pricing strategy is one kind of RTP strategy, mentioned in section 2.3. Pricing
signals for PVs are generated randomly with multiple probability distributions for the initial test.
This simulates the case where users of the community do not have information on the pricing
strategy of the counterparts and do not have the intention to compete with other producers in the
system. Their pricing strategies are completely independent and random. The Pv 0 used Gaussian
distribution with a mean of 20 cent CHF/kWh, with a standard deviation of 5. The price of PV 1
is generated with a uniform distribution between 5 cent CHF/kWh and 10 cent CHF/kWh. PV 2
applies an exponential distribution with a λ of 0.5.
Under this category, each device has only limited information on the behavior of any participant
in the system. For the price setter, i.e. PV, no information is available on either its competitors
(PV or battery) or its clients (house and heat pump), nor on its’ own production profile. This
corresponds to the situation that the system might encounter in its early stages. Since there is no
information on the market price, a device can only set up a pricing strategy based on its leveraged
cost with a the desired margin plus a risk premium. Here we assume that the leveraged costs are
the same for all the PVs; the preferred margin varies according to the owner, but remains in the
range of (10 to 15 %); the risk premium is set based on the market risk assessed by the owner
of the device. Producers arrive at different estimates because of the difference in the connection
layout of each device. This risk part fluctuates as the other factors, such as weather and consumers’
demand, vary along one single day. All of these factors result in choosing different risk Probability
Distribution Functions (PDFs). The choice of the exact PDF also depends on the risk adversity
of the user. Therefore, we use three different PDFs here. The figure below illustrates the pricing
strategy adopted in this scenario.
The inversed-production pricing strategy belongs to the CPP strategy mentioned in section 2.3. In
this pricing category, PV systems set up their price by applying desired rates to their basic price of
production. The basic price includes the market risk premium as well as a fluctuating part which
is in reversed correlation its volume of the generated energy. It is based on the assumption that
the higher the current production volume one PV system has, the higher the production of PVs
in the same community, since they are exposed to similar condition. According to the demand-
supply theory, assuming that the demand from consumers remains relatively stable (due to a lack
of information), the equilibrium price decreases as the quantity of supply increases.
It’s worth noticing that it is not necessary for all the producers to follow this relation. For those
who serve as the unique energy provider to certain consumers, their monopoly status allows them
to retain a high price even when the market price drops. In our model, for simplicity, we adopt the
same pattern for all the producers regardless their connectivity configuration.
$
pP V0 ptq
&10 l 70
P V0 ptq if lP V0 ptq 0,
(4.6)
%10, otherwise
$
pP V1 ptq
&11 l 80
P V1t p q2 if lP V1 ptq 0,
(4.7)
%11, otherwise
$
pP V2 ptq
&6l P V2
100
p q2
t if lP V2 ptq 0,
(4.8)
%6, otherwise
42 4.5. ÐApp Solution Architecture
Strategies in this category have an agile approach to modify their offered price based on the planned
price, as planned in the Inversed-production pricing in section 4.4.2. This pricing strategy keeps
track of the portion of energy sold out to the community at each timestep, which is defined as the
community factor fC in the equation 4.9c. If a comparison of the community factor of a producer
with that of other similar players in the market reveals that it sells less energy, the unit selling
price that producer offers will be reduced for the next time interval according to equation 4.9d.
The increment in price, padjustment , is adjusted by the internal policy of the producer. For example,
in the scenario 3 under this category, only 3 cent CHF/kWh is reduced when the condition meets;
While in scenario 6, the new price is reduced by 20 % comparted to the original one.
p4q
∆communityP Vi ptq lP Vi ptq lP Vi ptq
p1q (4.9a)
p5q p4q
∆gridP Vi ptq lP Vi ptq lP Vi ptq (4.9b)
∆communityP Vi ptq
fC,P Vi ptq
∆communityP Vi ptq ∆gridP Vi ptq
(4.9c)
padjusted
P Vi pt ∆tq porginal
P Vi pt ∆tq padjustment
( (4.9d)
if fC,P Vi ptq fC,P Vj ptq, where j P r0, 2s zi
Where
Figure 4.14: Schema of the smart contract structure of the proposed solution
The role of Configuration.sol is an entity who does the due-diligence work with respect to the
physical setup in the real world. Lines and tubes are indispensable for energy carriers flowing from
the location of its generation to the destination. As energy cannot be transferred or transmitted
wirelessly, the presence of such physical channels is the pre-requisite for P2P energy transaction.
Configuration.sol ensures this pre-requisite is met.
From a business perspective, the owner of the Configuration.sol can have an infrastructure rental
business that is setting up an agreement with the prosumers for their usage of the wires and
tunnels; They also turn to the gird or other out-of-community energy suppliers, to whom they can
offer a physical connection upon agreement; Then their third role comes naturally, the middle-man
who links the grid and hors-community energy suppliers to individual households. The linkage is
optional, requested by the node. The payment for establishing the connection could also be settled
on the chain according to the agreement among all affected users within this distributed prosumer
community. This part is not discussed in detail as it is not the focus of this thesis.
The financial settlement is recorded with a digital value in the contract associated with the device.
Given the inflation risk in the exchange of crypto-currency and fiat money, transactions are not
settled directly with Ether.
Chapter 5
45
46 5.2. Result
5.2 Result
5.2.1 Scenario 1
60 Sold to community 50
Production
103
Dumped to grid
and/or battery Price
40 15
Production (kWh)
Price (CHF/kWh)
Production (kW)
40
30
10
20
20
5
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) PV0 (a) PV0
Price (CHF/kWh)
Production (kW)
20 20 10
10 10 5
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) PV1 (b) PV1
30
Sold to community
30
Production
103
To grid Price 20
Production (kWh)
Price (CHF/kWh)
Production (kW)
20 20 15
10
10 10
5
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.1 Electricity production of PVs Figure 5.2 PV production and price under
under the 1st pricing scenario. the 1st pricing scenario.
48 5.2. Result
Residue 12
Payment
30
8
20
6
10 4
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) House0 (a) House0
Residue Payment
40 Accumulated Payment (CHF)
Provided by prosumers
Purchased from grid 15
Consumption (kWh)
30
10
20
10 5
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) House1 (b) House1
50
Residue 12 Payment
Accumulated Payment (CHF)
Provided by prosumers
40 Purchased from grid 10
Consumption (kWh)
30 8
6
20
4
10
2
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.3 Electricity settlements for Figure 5.4 Accumulated payment for
houses under the 1st pricing scenario. houses under the 1st pricing scenario.
Chapter 5. Testing and Results 49
110
8 Demand
5
90 4
0
5 80 2
10 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) Water tank0 (a) House0
20 10
80
60
0 5
40
20 20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) Water tank1 (b) House1
150
Volume Demand
DHW Consumption (L/15 min)
30 Outflow Supply
In- & Outflow (L/15 min)
10
5
0 50
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.5 Water flow powered by Figure 5.6 Domestic hot water
different source of energy for water tanks consumption and supply for houses under
under the 1st pricing scenario. the 1st pricing scenario.
50 5.2. Result
100
Volume
80
Volume (kWh)
60
40
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.7 State of charge of battery0 shared by House0 and House1 under the 1st pricing
scenario.
100 Injection
Maximum capacity
Absorption
50
Capacity (kW)
50
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.8 Electricity transfer from the district energy system to the grid under the 1 st pricing
scenario.
Chapter 5. Testing and Results 51
5.2.2 Scenario 2
In the second scenario, price setters, meaning PVs, have the knowledge of their historical perfor-
mance. With this data, PVs can predict their energy production profile and adjust their pricing
strategy based on the demand-supply relation, in order to divert the production surplus and prevent
overheat damage on itself.
This is a strategy that fits into the demand response scheme. Demand response describes the
changes in electric usage by end-use customers from their normal consumption patterns in response
to changes in the price of electricity over time. It can also refer to mechanisms such as incentive
payments designed to induce lower electricity use at times of high wholesale market prices or when
system reliability is jeopardized [142].
Price (CHF/kWh)
Production (kW)
40 20
30
20
20 10
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) PV0 (a) PV0
30
Sold to community
Production
40103
Dumped to grid 30
and/or battery Price
Production (kWh)
Price (CHF/kWh)
30
Production (kW)
20 20
20
10 10
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) PV1 (b) PV1
3
3010
30 Sold to community 30
Production
Dumped to grid
Price
and/or battery
Production (kWh)
Price (CHF/kWh)
Production (kW)
20 20 20
10 10 10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.9 Electricity production for Figure 5.10 PV production and price for PVs
PVs under the 2nd pricing scenario. under the 2nd pricing scenario.
Chapter 5. Testing and Results 53
Residue 20 Payment
15
20
10
10
5
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) House0 (a) House0
Residue Payment
40 Accumulated Payment (CHF)
Provided by prosumers
Purchased from grid 30
Consumption (kWh)
30
20
20
10 10
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) House1 (b) House1
50
Residue Payment
Accumulated Payment (CHF)
25
Provided by prosumers
40 Purchased from grid
Consumption (kWh)
20
30
15
20 10
10 5
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.11 Electricity settlements for Figure 5.12 Accumulated payment for
houses under the 2nd pricing scenario. houses under the 2nd pricing scenario.
54 5.2. Result
110
8 Demand
5
90
4
0
5 80
2
10
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) Water tank0 (a) House0
20
10
80
60
0 5
40
20 20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) Water tank1 (b) House1
150
Volume Demand
DHW Consumption (L/15 min)
30 Outflow Supply
In- & Outflow (L/15 min)
10
5
0 50
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.13 Water flow powered by Figure 5.14 Domestic hot water
different source of energy for water tanks consumption and supply for houses under
under the 2nd pricing scenario. the 2nd pricing scenario.
Chapter 5. Testing and Results 55
100
Volume
80
Volume (kWh)
60
40
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.15 State of charge of battery0 shared by House0 and House1 under the 2nd pricing
scenario.
100 Injection
Maximum Capacity
Absorption
50
Capacity (kW)
50
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.16 Electricity transfer from the district energy system to the grid under the 2 nd pricing
scenario.
56 5.2. Result
5.2.3 Scenario 3
In scenario 3, an agent-based nodal pricing approach is implemented according to the game theory,
where the price of PV is set based on not only the current production but also the volume of
electricity that has been successfully sold out to the community.
60 Sold to community 50
Production
103
Dumped to grid 20
and/or battery Price
40
Production (kWh)
Price (CHF/kWh)
Production (kW)
15
40
30
10
20
20
10 5
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) PV0 (a) PV0
30
Sold to community
Production
103
Dumped to grid 30
and/or battery Price 25
Production (kWh)
Price (CHF/kWh)
Production (kW)
20
20 20
15
10 10
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) PV1 (b) PV1
3
3010
30 Sold to community 30
Production
Dumped to grid
Price
and/or battery
Production (kWh)
Price (CHF/kWh)
Production (kW)
20 20 20
10 10 10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.17 Electricity production for Figure 5.18 PV production and price for PVs
PVs under the 3rd pricing scenario. under the 3rd pricing scenario.
58 5.2. Result
4 Payment
Residue
20
2
10
1
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) House0 (a) House0
Residue Payment
40 Accumulated Payment (CHF)
Provided by prosumers
Purchased from grid 6
Consumption (kWh)
30
4
20
10 2
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Time (h)
(b) House1 (b) House1
50
5 Payment
Residue
Accumulated Payment (CHF)
Provided by prosumers
40 Purchased from grid 4
Consumption (kWh)
30 3
20 2
10 1
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.19 Electricity settlements for Figure 5.20 Accumulated payment for
houses under the 3rd pricing scenario. houses under the 3rd pricing scenario.
Chapter 5. Testing and Results 59
95
Volume 8 Demand
0 4
85
5 2
80
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
Volume
DHW Consumption (L/15 min) 15 Demand
40 Outflow 120 Supply
In- & Outflow (L/15 min)
60
5
0
40
20 20 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
150
Demand
DHW Consumption (L/15 min)
Volume
30 Supply
Outflow
In- & Outflow (L/15 min)
10
5
0 50
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.21 Water flow powered by Figure 5.22 Domestic hot water
different source of energy for water tanks consumption and supply for houses under
under the 3rd pricing scenario. the 3rd pricing scenario.
60 5.2. Result
100
Volume
80
Volume (kWh)
60
40
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.23 State of charge of battery0 shared by House0 and House1 under the 3rd pricing
scenario.
40 Injection
Maximum Capacity
Absorption
20
Capacity (kW)
20
40
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.24 Electricity transfer from the district energy system to the grid under the 3 rd pricing
scenario.
Chapter 5. Testing and Results 61
5.2.4 Scenario 4
Scenario 4 is based on the scenario 2. The only difference between the two is the layout: In scenario
4, there is no battery in the electricity network. The only flexibility provider is the grid.
50 30103
Sold to community
Production
60 Price
Dumped to grid
and/or battery 40
Price (CHF/kWh)
Production (kW)
Production (kWh)
20
40 30
20
10
20
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Time (h) Time (h)
Sold to community
Production
40103
30 30
Dumped to grid
Price
and/or battery
Production (kWh)
Price (CHF/kWh)
30
Production (kW)
20 20
20
10 10
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) PV1 (b) PV1
3
3010
30 Sold to community 30
Dumped to grid Production
and/or battery Price
Production (kWh)
Price (CHF/kWh)
Production (kW)
20
20 20
10
10 10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Time (h)
(c) PV2
(c) PV2
Residue 20 Payment
15
20
10
10
5
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) House0 (a) House0
Residue Payment
40 Accumulated Payment (CHF) 40
Provided by prosumers
Purchased from grid
Consumption (kWh)
30 30
20 20
10 10
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) House1 (b) House1
50
20 Payment
Residue
Accumulated Payment (CHF)
Provided by prosumers
40 Purchased from grid
Consumption (kWh)
15
30
10
20
5
10
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.27 Electricity settlements for Figure 5.28 Accumulated payment for
houses under the 4th pricing scenario. houses under the 4th pricing scenario.
64 5.2. Result
Volume 8 Demand
5
90 4
0
5 80
2
10 70
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) Water tank0 (a) House0
Volume Demand
40 120 DHW Consumption (L/15 min) 15
Outflow
In- & Outflow (L/15 min)
Supply
Level in Water Tank (L)
20 10
80
60
0 5
40
20 20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Time (h)
(b) Water tank1
(b) House1
150
Volume Demand
DHW Consumption (L/15 min)
30 Outflow Supply
In- & Outflow (L/15 min)
10
0 50 5
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.29 Water flow powered by Figure 5.30 Domestic hot water
different source of energy for water tanks consumption and supply for houses under
under the 4th pricing scenario. the 4th pricing scenario.
Chapter 5. Testing and Results 65
120
Volume
110
Volume (kWh)
100
90
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.31 State of charge of battery0 shared by House0 and House1 under the 4th pricing
scenario.
100 Injection
Maximum Capacity
Absorption
50
Capacity (kW)
50
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.32 Electricity transfer from the district energy system to the grid under the 4 th pricing
scenario.
66 5.2. Result
5.2.5 Scenario 5
Similar to the scenario 4, scenario 5 is also based on the second scenario, with all the devices
connected to the battery. The battery is a fully community owned good, shared by all the prosumers.
50 30103
Sold to community
Production
60 Price
Dumped to grid
and/or battery 40
Price (CHF/kWh)
Production (kW)
Production (kWh)
20
40 30
20
10
20
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Time (h) Time (h)
Sold to community
Production
40103
30 30
Dumped to grid
Price
and/or battery
Production (kWh)
Price (CHF/kWh)
30
Production (kW)
20 20
20
10 10
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) PV1 (b) PV1
3
3010
30 Sold to community 30
Dumped to grid Production
and/or battery Price
Production (kWh)
Price (CHF/kWh)
Production (kW)
20
20 20
10
10 10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Time (h)
(c) PV2
(c) PV2
Residue 20 Payment
15
20
10
10
5
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) House0 (a) House0
Residue Payment
40 Accumulated Payment (CHF) 40
Provided by prosumers
Purchased from grid
Consumption (kWh)
30 30
20 20
10 10
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) House1 (b) House1
50
20 Payment
Residue
Accumulated Payment (CHF)
Provided by prosumers
40 Purchased from grid
Consumption (kWh)
15
30
10
20
5
10
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.35 Electricity settlements for Figure 5.36 Accumulated payment for
houses under the 5th pricing scenario. houses under the 5th pricing scenario.
Chapter 5. Testing and Results 69
120
8 Demand
5 80 2
10 70
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) Water tank0 (a) House0
20
10
80
60
0 5
40
20 20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) Water tank1 (b) House1
150
Volume Demand
DHW Consumption (L/15 min)
30 Outflow Supply
In- & Outflow (L/15 min)
10
5
0 50
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.37 Water flow powered by Figure 5.38 Domestic hot water
different source of energy for water tanks consumption and supply for houses under
under the 5th pricing scenario. the 5th pricing scenario.
70 5.2. Result
Volume
Volume (kWh) 80
60
40
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.39 State of charge of battery0 shared by House0 and House1 under the 5th pricing
scenario.
100 Injection
Maximum Capacity
Absorption
50
Capacity (kW)
50
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.40 Electricity transfer from the district energy system to the grid under the 5 th pricing
scenario.
Chapter 5. Testing and Results 71
5.2.6 Scenario 6
Scenario 6 shares a similar idea as the scenario 3, as it is also based on the game theory. The
adaptation behavior in the case unfavorable conditions is what differentiates the two scenarios.
Energy providers tend to take bold actions when they are not competitive against others. Therefore
the adaptation mechanism in scenario 6 is quite extreme.
50
Production
103
60 Sold to community
Dumped to grid Price
40 15
and/or battery
Price (CHF/kWh)
Production (kW)
Production (kWh)
40 30
10
20
20 5
10
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 Time (h) Time (h)
30
Sold to community
Production
103
Dumped to grid 30 20
Price
and/or battery
Production (kWh)
Price (CHF/kWh)
Production (kW)
15
20 20
10
10 10
5
0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) PV1 (b) PV1
3
3010
30 Sold to community 30
Dumped to grid Production
and/or battery Price
Production (kWh)
Price (CHF/kWh)
Production (kW)
20
20 20
10
10 10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Time (h)
(c) PV2
(c) PV2
Residue Payment
20
3
2
10
1
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(a) House0 (a) House0
Residue
8
40 Accumulated Payment (CHF) Payment
Provided by prosumers
Purchased from grid
Consumption (kWh)
6
30
20 4
10 2
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Time (h)
(b) House1 (b) House1
50
Residue Payment
Accumulated Payment (CHF)
Provided by prosumers 5
40 Purchased from grid
Consumption (kWh)
4
30
3
20
2
10
1
0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.43 Electricity settlements for Figure 5.44 Accumulated payment for
houses under the 6th pricing scenario. houses under the 6th pricing scenario.
74 5.2. Result
100
8 Demand
90
0 4
85
5 2
80
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
60
0 5
40
20 20 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h)
(b) Water tank1 (b) House1
150
Demand
DHW Consumption (L/15 min)
Volume
30 Outflow Supply
In- & Outflow (L/15 min)
10
5
0 50
10
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 5.45 Water flow powered by Figure 5.46 Domestic hot water
different source of energy for water tanks consumption and supply for houses under
under the 6th pricing scenario. the 6th pricing scenario.
Chapter 5. Testing and Results 75
100
Volume
80
Volume (kWh)
60
40
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.47 State of charge of battery0 shared by House0 and House1 under the 6th pricing
scenario.
100 Injection
Maximum Capacity
Absorption
50
Capacity (kW)
50
100
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h)
Figure 5.48 Electricity transfer from the district energy system to the grid under the 6 th pricing
scenario.
76 5.3. Comparison
5.3 Comparison
Table 5.1: Comparison of the destination of the electricity generated by each PV across scenarios
Nr. 1 2 3 4 5 6
50 50 50 50 50 50
Sold to community Sold to community Sold to community Sold to community Sold to community Sold to community
Dumped to grid Dumped to grid Dumped to grid Dumped to grid Dumped to grid Dumped to grid
40 and/or battery 40 and/or battery 40 and/or battery 40 and/or battery 40 and/or battery 40 and/or battery
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
30 30 30 30 30 30
PV0 20
10
20
10
20
10
20
10
20
10
20
10
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h) Time (h) Time (h) Time (h)
Sold to community Sold to community Sold to community Sold to community Sold to community Sold to community
30 30 30 30 30 30
Dumped to grid Dumped to grid Dumped to grid Dumped to grid Dumped to grid Dumped to grid
and/or battery and/or battery and/or battery and/or battery and/or battery and/or battery
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
20 20 20 20 20 20
PV1 10 10 10 10 10 10
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h) Time (h) Time (h) Time (h)
30 Sold to community
30 Sold to community
30 Sold to community
30 Sold to community
30 Sold to community
30 Sold to community
Dumped to grid Dumped to grid Dumped to grid Dumped to grid Dumped to grid Dumped to grid
and/or battery and/or battery and/or battery and/or battery and/or battery and/or battery
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
Production (kWh)
20 20 20 20 20 20
PV2 10 10 10 10 10 10
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h) Time (h) Time (h) Time (h)
Table 5.2: Percentage of the electricity generated by PV that is actively sold to the community
under each scenario
Nr. 1 2 3 4 5 6
PV0 79.9 % 92.4 % 91.2 % 92.4 % 92.4 % 89.9 %
PV1 100 % 100 % 100 % 100 % 100 % 96.0 %
PV2 100 % 100 % 99.7 % 100 % 100 % 89.6 %
Table 5.3: Percentage of the time where PVs sell electricity only to the community (Step 4 only)
under each scenario
Nr. Destination 1 2 3 4 5 6
Community only 41.7 % 47.2 % 80.6 % 47.2 % 47.2 % 77.8 %
PV0 Grid only 0.0 % 0.0 % 0.0 % 0.0 % 0.0 % 0.0 %
Both 58.3 % 52.8 % 19.4 % 52.8 % 52.8 % 22.2 %
Community only 100.0 % 100.0 % 100.0 % 100.0 % 100.0 % 94.1 %
PV1 Grid only 0.0 % 0.0 % 0.0 % 0.0 % 0.0 % 0.0 %
Both 0.0 % 0.0 % 0.0 % 0.0 % 0.0 % 5.9 %
Community only 100.0 % 100.0 % 97.1 % 100.0 % 100.0 % 82.9 %
PV2 Grid only 0.0 % 0.0 % 0.0 % 0.0 % 0.0 % 0.0 %
Both 0.0 % 0.0 % 2.9 % 0.0 % 0.0 % 17.1 %
Table 5.1 summarizes the destination to which PVs sell their generation across all the testing
scenarios, over one operation day. Table 5.2 shows the exact percentage of PVs’ contribution to the
community. Table 5.3 shows that the community takes indisputable priority in receiving renewable
energy supply. It can be found that in the game-theory based scenarios, electricity produced by
PV is better dispatched in the community. The time where PVs, especially PV 0 , have to dump
excess energy onto the grid, is decreased by half. This change in percentage indicates that the
game-theory based scenario is more efficient in scheduling.
Chapter 5. Testing and Results 77
Table 5.4: Comparison of the source of the electricity purchased by each house across scenarios
Nr. 1 2 3 4 5 6
Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers
30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
20 20 20 20 20 20
House0 10 10 10 10 10 10
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h) Time (h) Time (h) Time (h)
Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers
30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
20 20 20 20 20 20
House1 10 10 10 10 10 10
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h) Time (h) Time (h) Time (h)
Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers Provided by prosumers
30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid 30 Purchased from grid
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
Consumption (kWh)
20 20 20 20 20 20
House2 10 10 10 10 10 10
0 0 0 0 0 0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Time (h) Time (h) Time (h) Time (h) Time (h) Time (h)
Table 5.5: Percentage of the electricity volume purchased by house that directly comes from PV
under each scenario
Nr. 1 2 3 4 5 6
House0 73.1 % 72.8 % 100 % 68.3 % 71.6 % 98.4 %
House1 38.7 % 38.7 % 90.1 % 13.2 % 20.6 % 88.6 %
House2 17.3 % 18.2 % 96.1 % 43.4 % 48.6 % 98.6 %
Table 5.4 summarizes the source from which houses buy their electricity supply across all the
testing scenarios, over one operation day. Table 5.5 shows the exact percentage of houses’ direct
purchase from PVs in the community.
5.4 Evaluation
Here we track the usage of gas for each round of transaction. As introduced in Chapter 1, gas
measures the amount of work needed for executing a transaction and gas price is the unit price
that the transaction initiator would like to pay to the miner for its validation work. The higher
the gas price, the more expensive the transaction costs, but it also leads to faster validation of
the transaction. It is essential to find a balanced point between the cost and the speed for this
project, given the volume of transactions and the time constraint in this quasi-real-time trading
policy. Due to the inherent distributed nature of renewable energy systems, novel communication
standards (e.g., IEC 61850) have emerged to protect, control, and operate the network to fulfill the
stringent reliability, delay, and latency demands [143]. As suggested in INTEGRIS project [144], to
enable medium reliable information exchange and management between devices in the smart grid,
the transfer time shall be less than 5 minutes. Therefore, here we consider 5 min as the maximum
settlement time for each transaction, as there is on average one transaction per step in our system.
Here we apply the standard Gas Price (2 gwei) given by ETH Gas Station to get an estimation of
the transaction cost based on the assumed average transaction time.
The results given in table 5.7 and table 5.8 show the estimated average cost for household, PV,
battery, heat pump, and grid. Figure 5.49 shows the average gas consumption and the testing time
of each device per round. Each device needs to spend around 50 $ in the transaction for a full
78 5.4. Evaluation
Table 5.6: Percentage of the time where houses purchase electricity directly from PV under each
scenario
Nr. Source 1 2 3 4 5 6
Community only 46.9 % 45.8 % 100.0 % 29.2 % 41.6 % 95.8 %
House0 Grid only 46.9 % 46.9 % 0.0 % 63.5 % 52.1 % 3.2 %
Both 6.3 % 7.3 % 0.0 % 7.3 % 6.3 % 1.0 %
Community only 1.0 % 1.0 % 82.3 % 0.0 % 12.5 % 82.3 %
House1 Grid only 67.7 % 67.7 % 14.6 % 85.4 % 68.7 % 15.6 %
Both 31.3 % 31.3 % 3.1 % 14.6 % 18.8 % 2.1 %
Community only 17.7 % 15.6 % 97.9 % 13.6 % 28.2 % 95.8 %
House2 Grid only 72.9 % 65.6 % 0.0 % 70.8 % 58.3 % 4.2 %
Both 9.4 % 18.8 % 2.1 % 15.6 % 13.5 % 0.0 %
Table 5.7: Estimated transaction cost per round of the proposed model
day’s operation. To run a full-day 15-min interval testing for each pricing scenario, it takes around
25 to 30 minutes to complete it on a 4-core 3 GHz Intel® Xeon® CPU, when a query for account
balance to EVM (Ethereum Virtual Machine) is undertaken at each step. However, it takes only
around 12 minutes to complete a full-cycle testing, if queries to EVM are not attended. For such
a simple system layout, the testing time is long.
The household consumption profile of NEST cannot reflect the 100% real case, because there are,
in general, only people during the day instead of having domestic activities. Therefore, energy
consumption drops after 18h, while in the real case, this period is another peak in the daily
consumption profile.
In general, the price influence depends very much on the energy supply distribution. Game-theory
based strategy provides a feedback loop for devices to make better decisions with limited informa-
tion. The agility it provides is essential in the intra-day, high-frequency trading.
In all the scenarios, the battery is discharged at the end of the day. It is not a sustainable approach
because the situation changes as there is no initial supply from the battery in the beginning of
the day, where no renewable resource in the system can generate energy. The community system,
as it is now producing too little energy to maintain a strategy in which grid-power is only used
for backup. If a real system were to operate using this strategy, the PVs would need to be sized
more adequately. Another solution is to include other forms of green energy into the system. Some
clean energies, such as wind and biomass, do not operate depending on the radiation from nature.
They can produce energy during at night when PVs are not capable of doing so, or operate for the
Chapter 5. Testing and Results 79
Table 5.8: Estimated transaction cost per day of the proposed model
105 104
8 2.5
2
6
Time (ms)
1.5
Gas
4
House 1
PV
Battery 2
0.5
Heat Pump
Time
0 0
1 16 31 46 61 76 91
Round
base-load generation to supply the demand of the community and battery replenishment. Ideally,
additional rules would need to be introduced which lead to sustained average levels of both the
battery and the heat storage.
By choosing a competitive pricing strategy, households pay significantly less than in a determined
pricing case. The allocation of renewable energy is more in favor of the community compared to
that in the scenario without any feedback loop. The risk of having a monopolistic price setter exists
when there is only one available energy provider in the community. According to our current policy,
the community green energy supplier holds priority than other suppliers. During the time when
there is only one community green energy provider, it appears on the top list of all its connected
devices. Therefore, as long as there is demand from the connected devices of this particular provider,
its production in that time interval is ensured to be sold out. This leaves this particular energy
provider freedom to set up a price where no competitor is in the market, which in contrast, is a
risk for consumers. This risk is controlled when there are more energy providers of the same form,
meaning competitors, in the market. The better the connectivity of the consumers, the more robust
the system supply is, therefore, the less risky this issue is. The "blacklist" mechanism indirectly
eliminates the long-term malicious behaviors.
Battery plays an inevitable role in providing stability in the system shaving the peak. Comparing
the amount of energy that PV1 bought from the community in scenario 2, 4, and 5, the additional
link to the battery allows this house to purchase 12.5 % more of energy merely from the community.
For House0 and House2 , which have relatively good connectivity in the district energy system,
without the battery (in scenario 4), their demand to purchase energy from the grid have increased
by 5 - 15 %.
Chapter 6
Empowering Local Community Through the designed architecture and market policy, each
consumer can purchase all the necessary energy resources with access to a maximum amount of
local renewable energy resources and a minimized cost. The concept of community is therefore
strengthened towards a common goal of reducing pollution, by sharing resources. The resource of
energy is traced. The positive environmental externalities of renewable energy are incorporated.
Dynamic Tariff Model Our mechanism allows a self-determined flexible tariff setting model.
Testing results show that with a behavior and market price based game theory pricing strategy,
producers can effectively allocate their resources. Consumers can also minimize the cost of energy
consumption.
Transactive Control Since we can track the transactions settled on the chain, control signals
can be caught and communicated via API for transactive device control. This leads to a fully-
automated energy-trading platform. Both heat and electricity can be supplied without brokers or
energy companies.
Extension for Payment System The design of this project allows extensions to accept pay-
ments on cryptocurrencies, tokens, and legal currencies. All of these kinds of payments can be used
in combination with a deposition system. One of the best applications is to use the transaction
ledger as a Proof of Origin, i.e. confirmation of renewable energy usage. With association to the
blockchain-based carbon market, the use of renewable energy price can be further rewarded with
carbon credits or the energy can be traded alongside green certificates in the Cap-and-Trade mar-
ket.
81
82 6.1. Feature of the Project
Green certificate: On 3 February 2017, China’s National Development and Reform Commission
(NDRC) announced a pilot program for the issuance of green certificates to solar and wind power
producers, each certificate representing 1 MWh of electricity output. Power producers are allowed
to sell the certificates to buyers such as coal-fired power producers, grid operators, and corpo-
rates to evidence their use of clean energy. The price for the certificates is determined by bilateral
negotiation or by competitive bidding. The price is capped at the level of the subsidy currently re-
ceived by power producers. This program was voluntary in 2017 and NDRC may make the scheme
mandatory from 2018 [145].
Implementation on the Ethereum Blockchain This design has both pros and cons. The
trading platform is build based on Ethereum. Therefore the cost of transactions is related to the
exchange rate for ETH to CHF or other legal tenders. In contrast, the value of energy does not
necessarily relate to the value of ETH, it can be monetized with other tokens or fiat money. (as it
is written in the form of digits on the ledger) However, both contain the risk of fluctuation. The
inflation of Ethereum on the public change has a significant impact on the operation cost of the
trading platform; the value of the energy may fluctuate depending on the community. It can boost
the solidarity of the community, but on the flip side, also creates space for market speculation
when one entity owns a significant percentage of devices in the market and communicate off-chain
for a pricing strategy that maximizes their utility.
Real Peer-to-Peer Structure This system prevents the single-point failure because the com-
munication is P2P without any inter-mediator or aggregator. When the system operates within
its technical constraints, if one device is disconnected/defective, the entire trading process is not
disturbed. Those devices that are not physically connected to the failure point would have zero
impact; while those connected devices start to look for the next available device to complete the
process until the defective device reconnects to the network.
Also, in a distributed contracts structure managed by contract factories, it eases the difficulty of
maintaining and upgrading.
This trading logic has good scalability. In this trading mechanism, the complexity of the execution
of each node increases with the number of devices connected to it, as they are the only interaction
the node has. In a fully decentralized energy market setup, the magnitude of connections that one
device can have remains comparable regardless the increase in the size of the community. Therefore,
this trading solution can easily scale up.
Anonymous and Transparent Ledger Although the accounts are publicly accessible to all
the users, data privacy is well protected in this system, thanks to the previously mentioned original
P2P structure. Data privacy and anonymity can be better guaranteed than that in the traditional
central database-driven business model because only certain users, i.e., connected devices, can
access to the non-public production/consumption and price information when situated in the right
time-slot.
Environmental Benefits The design principle of this project internalizes many environmental
externalities of renewable energies, which is reflected in the unit price. Results show that it im-
proves the penetration and usage of renewable energies in the community. With a game theory
based pricing strategy, more than 80 % of households’ electricity demand can be covered by the
green energy produced in the community. This autonomous system is a good solution in the local
renewable energy market.
Simple Interaction with Upper-Lever Grid The balancing mechanism (concretely speaking,
step 5 in the trading mechanism) makes a physical balancing group with a single connection point
Chapter 6. Conclusion and Outlook 83
with the upper-level grid. This structure is a complete substitute of current virtual balancing group
on the TSO-level power market. The traditional concept of balancing group in the power market
is completely disconnected with the physical operation. This near real-time trading system can
re-integrate the physical and financial aspects together.
In terms of technical operation, we should consider network constraints and line losses in the
transmission and distribution. Possible aspects to look into are: Incorporating energy conversion
losses for heat pumps; Heating losses in the tube, when sending hot water from a heat pump to a
water tank and on to a household. Detailing the model for the heat pump to make sure that the
change in performance when the temperature changes is included. Also, heat pumps should always
operate in the optimal performance range.
Optimization work, such as in [99], is one of the most aspects to consider in further work. Re-
searchers need to thoroughly evaluate elements of the design and evaluate whether it should be
put on-chain or off-chain, to increase the efficacy of data and to reduce computing cost, if ap-
plicable. For instance, the communication between peers taking place before the settlement of
transactions does not need to be performed using smart contracts. Applications can use the Whis-
per communication protocol instead, to speed up the communication.
In the mechanism design aspect, we are going to extend the agent implementation with more
sophisticated methods for strategy design.
In terms of business strategy,it would also be interesting to investigate possible business models
and technical operation codes that incorporate local ancillary services, which provide flexibility and
enhance the system’s stability. Various business models can be tested when applying blockchain
together with the ancillary service providers. For instance, ancillary service providers could receive
a premium along with paying deposit on a smart contract on the blockchain by claiming the
capability of offering flexibility if needed. According to the amount of the deposit, provides rank
along a list of queries. When the flexibility is needed, the smart contract goes through this list
and secure transactions with the higher ranks. Once the flexibility is provided, the smart contract
can initiate transactions to pay back the deposit as well as the payment for the service. If the
provider is not able to provide what he has claimed, the deposit then goes into the pool for
flexibility remuneration. Another example could be, to leverage the public data on the chain, to
provide maintenance services for the energy system. A company can subscribe and listen to the
transaction and predict a failure in, for example, a heat pump due to a lower transaction volume
and therefore provide fast maintenance service.
As for the usability aspect, we should improve the usability for non-tech-savvy. Although the
technology is based on the blockchain, users should only interact with a website or an App (where
84 6.2. Future Work
no good solution has been introduced) or directly via touchscreen on the devices where they could
configure the system, introduce and modify the selling price of the PV they own.
In the social aspect,we could examine the approaches to incentivize ethical behaviors among the
community members. One example is to introduce a reputation evaluation system as an extension
of the non-cooperative game theory by using the public data User behaviors can be further analyzed
to extract behavioral patterns. This information would be reported publicly for community usage
instead of by one individual or private entity for purely commercial or political purpose. The rating
could be implemented in the current system by letting it affect the decision making of the agents
themselves.
On the hardware side, there are also challenges. To record the real-time energy production and
consumption, smart meters are required. They hold a pivotal role as the primary data entry point.
However, electricity smart meters are not widely adopted by all households, especially by those built
before the 21st century and that haven’t been retrofitted. Moreover, the smart meters or sensors
for water tanks and heating facilities are not yet as well developed as smart meters for electrical
appliances. Also, to fully benefit from the smart grid setup, controllable devices are preferred yet
uncommon in an average household. The lack of adequate, accessible equipment creates a gap to
the adoption of the smart grid in the first place; not to mention that blockchain-enabled smart
meters or sensors are not yet on the market. Therefore, the concept has only been implemented in
small-scale pilot projects. Some further work may include the synchronized control of devices.
In a cyber security perspective,the security of the contract needs to be improved with caution.
As mentioned in the chapter 1, hackers may take significant advantages of a small bug in the system.
Since this work is only a proof-of-concept for academic use, the security of the data structure is
not the primary focus in the scope of this project. Since there is no direct crypto-currency or token
transaction in the system, the threat is less critical. However, for the data security reasons, this is
worth refining before future field tests.
As to the regulatory aspect, with the emerging interest for regulation for cryptocurrencies,
the legal compliance issue extends to all the blockchain applications that involve data and value
transfer. However, legal institutions have not established regulations yet. Aitzhan and Zhum-
abekuly [146] found that unlike conventional companies, cryptocurrency-centered companies in-
vesting in agriculture or green energy do not need to be compliant with securities law in Europe.
What is the role and responsibility of participants? In case of default, who shall take the respon-
sibility in various situations? Will utility or grid operator own control of certain devices in the
system? Shall the earning from such trading be reported and taxed? Many questions are still not
clear.
And lastly, let’s look beyond energy trading, since blockchain already has some proven
application in the financial industry. There is a high potential for the application of DLT in energy
projects’ financing and management, not only in the crowd funding of small-scale renewable energy
projects but also in the authenticating project execution. Such application relieves the workload of
due-diligence and secures the arrival of the claimed financial support from investors, thus increasing
the transparency for both sides.
Chapter 6. Conclusion and Outlook 85
6.3 Outlook
Although the blockchain technology is highly praised at the moment, the emergence of quantum
computers may threaten the viability of its current implementations. The consensus mechanism,
which allows for the verification of each transaction, requires codified communication between the
sender and the receiver of the transaction. The decryption of this codified communication, thought
to be impossible due to the computation time required would all of a sudden be possible using
quantum computers and yield much more than just information about the single transaction. A
hacker attack by a quantum computer would also reveal a blockchain account’s personal identifier
and therefor allow the hacker to take over the entire account. Having this new vulnerability in
mind it becomes obvious that the blockchain technology faces new challenges. There is a large
need for new innovations if the blockchain technology shall continue to be as disruptive over the
next decades as it has been in the last few years.
I am very excited to see what future brings. I hope you are too.
Appendix A
Appendix
Table A.1: Parameter used in the SQL query for consumption/production data from eHub database
A.3.1 Architecture
Here below is the detailed ÐApp smart contract architecture for the electricity network.
87
88 A.3. ÐApp implementation
Action List
[1] M. Pisa and M. Juden, “Blockchain and Economic Development: Hype vs. Reality,” p. 49,
Jul. 2017.
[2] S. McLean and S. Deane-Johns, “Demystifying Blockchain and Distributed Ledger
Technology – Hype or Hero?” Computer Law Review International, vol. 17, no. 4, Jan.
2016. [Online]. Available: http://www.degruyter.com/view/j/cri.2016.17.issue-4/cri-2016-
0402/cri-2016-0402.xml
[3] B. Notheisen, F. Hawlitschek, and C. Weinhardt, “Breaking down the blockchain hype –
towards a blockchain market engineering approach,” Guimarães, Portugal, Jun. 2017, pp.
1062–1080.
[4] Y. Guo and C. Liang, “Blockchain application and outlook in the banking industry,”
Financial Innovation, vol. 2, no. 1, p. 24, Dec. 2016. [Online]. Available: https:
//link.springer.com/article/10.1186/s40854-016-0034-9
[5] “Neufund.” [Online]. Available: https://neufund.org/
[6] “Sovrin - identity for all.” [Online]. Available: https://sovrin.org/
[7] “BanQu, the FinTech company focusing on economic identities and extreme poverty,” May
2013. [Online]. Available: http://www.banquapp.com/
[8] Y.-P. Lin, J. R. Petway, J. Anthony, H. Mukhtar, S.-W. Liao, C.-F. Chou, and Y.-F. Ho,
“Blockchain: The Evolutionary Next Step for ICT E-Agriculture,” Environments, vol. 4,
no. 3, p. 50, Jul. 2017. [Online]. Available: http://www.mdpi.com/2076-3298/4/3/50
[9] P. Zhang, D. C. Schmidt, J. White, and G. Lenz, “Blockchain Technology Use Cases in
Healthcare,” p. 32, 2018.
[10] J. Sun, J. Yan, and K. Z. K. Zhang, “Blockchain-based sharing services: What blockchain
technology can contribute to smart cities,” Financial Innovation, vol. 2, p. 26, Dec. 2016.
[Online]. Available: https://doi.org/10.1186/s40854-016-0040-y
[11] S. Kejriwal and S. Mahajan, “Blockchain in commercial real estate,” The
Deloitte Center for Financial Services, Tech. Rep., 2017. [Online]. Avail-
able: https://www2.deloitte.com/content/dam/Deloitte/us/Documents/financial-services/
us-fsi-rec-blockchain-in-commercial-real-estate.pdf
[12] “Ujo Music.” [Online]. Available: https://ujomusic.com
[13] “Bernstein - Blockchain for intellectual property.” [Online]. Available: https://www.
bernstein.io/
[14] “GOLDMINT | Digital Gold.” [Online]. Available: https://goldmint.io/
[15] “Everledger | A Digital Global Ledger.” [Online]. Available: https://www.everledger.io/
93
94 Bibliography
[33] F. Tschorsch and B. Scheuermann, “Bitcoin and Beyond: A Technical Survey on Decentralized
Digital Currencies,” IEEE Communications Surveys Tutorials, vol. 18, no. 3, pp. 2084–2123,
2016.
[34] Q. H. Dang, “Secure Hash Standard,” National Institute of Standards and Technology, Tech.
Rep. NIST FIPS 180-4, Jul. 2015. [Online]. Available: https://nvlpubs.nist.gov/nistpubs/
FIPS/NIST.FIPS.180-4.pdf
[35] C. Percival and S. Josefsson, “The scrypt Password-Based Key Derivation Function,” 2016.
[Online]. Available: http://www.rfc-editor.org/info/rfc7914
[36] I. Bentov, A. Gabizon, and A. Mizrahi, “Cryptocurrencies without Proof of Work,”
arXiv:1406.5694 [cs], Jun. 2014, arXiv: 1406.5694. [Online]. Available: http://arxiv.org/
abs/1406.5694
[37] M. Castro and B. Liskov, “Practical Byzantine fault tolerance and proactive recovery,” ACM
Transactions on Computer Systems (TOCS), vol. 20, no. 4, pp. 398–461, 2002.
[38] M. Vukolić, “The Quest for Scalable Blockchain Fabric: Proof-of-Work vs. BFT Replication,”
in Open Problems in Network Security, ser. Lecture Notes in Computer Science. Springer,
Cham, Oct. 2015, pp. 112–125. [Online]. Available: https://link.springer.com/chapter/10.
1007/978-3-319-39028-4_9
[39] K. Christidis and M. Devetsikiotis, “Blockchains and Smart Contracts for the Internet of
Things,” IEEE Access, vol. 4, pp. 2292–2303, 2016.
[40] F. Restuccia, S. D’Oro, and T. Melodia, “Securing the Internet of Things: New Perspectives
and Research Challenges,” arXiv:1803.05022 [cs], Mar. 2018, arXiv: 1803.05022. [Online].
Available: http://arxiv.org/abs/1803.05022
[41] “Eth(Embedded) – Ethereum Clients on Embedded Devices.” [Online]. Available:
http://ethembedded.com/
[42] “IBM’s latest computer is a blockchain-ready CPU smaller than a grain of salt,”
Mar. 2018. [Online]. Available: https://www.digitaltrends.com/computing/ibm-blockchain-
computer-salt/
[43] S. Popov, “The Tangle - IOTA white paper,” IOTA, Tech. Rep. [Online]. Available:
https://iota.org/IOTA_Whitepaper.pdf
[44] R. Rottmann, “About Smart Contracts in IOTA,” Mar. 2018. [Online]. Available:
https://medium.com/@ralf/about-smart-contracts-in-iota-626d2bd3619e
[45] “Streamr.” [Online]. Available: https://www.streamr.com/
[46] P. Labs, “IPFS is the Distributed Web.” [Online]. Available: https://ipfs.io/
[47] “Welcome to the swarm documentation! — swarm 0.2rc5 documentation.” [Online].
Available: https://swarm-guide.readthedocs.io/en/latest/
[48] M. Valenta and P. Sandner, “Comparison of Ethereum, Hyperledger Fabric and Corda,” p. 8.
[49] G. Chapron, “The environment needs cryptogovernance,” Nature News, vol. 545, no. 7655, p.
403, May 2017. [Online]. Available: http://www.nature.com/news/the-environment-needs-
cryptogovernance-1.22023
[50] “Proof of Stake FAQ · ethereum/wiki Wiki.” [Online]. Available: https://github.com/
ethereum/wiki/wiki/Proof-of-Stake-FAQ
[51] “Sharding FAQ · ethereum/wiki Wiki.” [Online]. Available: https://github.com/ethereum/
wiki/wiki/Sharding-FAQ
[52] “Raiden 101.” [Online]. Available: https://raiden.network/101.html
96 Bibliography
[53] J. Poon and V. Buterin, “Plasma: Scalable Autonomous Smart Contracts,” p. 47.
[54] R. Beck, C. Becker, J. Lindman, and M. Rossi, “Opportunities and Risks of Blockchain
Technologies,” p. 44.
[55] F. Hawlitschek, B. Notheisen, and T. Teubner, “The limits of trust-free systems: A
literature review on blockchain technology and trust in the sharing economy,” Electronic
Commerce Research and Applications. [Online]. Available: https://www.sciencedirect.com/
science/article/pii/S1567422318300292
[56] “Understanding The DAO Attack,” Jun. 2016. [Online]. Available: https://www.coindesk.
com/understanding-dao-hack-journalists/
[57] E. T. f. t. arXiv, “If you’ve used Bitcoin to buy anonymously, the transaction may not
have been private at all.” [Online]. Available: https://www.technologyreview.com/s/608716/
bitcoin-transactions-arent-as-anonymous-as-everyone-hoped/
[58] T. Wagenknecht, T. Teubner, and C. Weinhardt, “Peer Ratings and Assessment Quality in
Crowd-Based Innovation Processes,” Guimarães,Portugal, 2017.
[59] P. Mancarella, “MES (multi-energy systems): An overview of concepts and evaluation
models,” Energy, vol. 65, pp. 1–17, Feb. 2014. [Online]. Available: http://www.sciencedirect.
com/science/article/pii/S0360544213008931
[60] Anonymous, “2030 climate & energy framework,” Nov. 2016. [Online]. Available:
https://ec.europa.eu/clima/policies/strategies/2030_en
[61] J. C. Boudri, L. Hordijk, C. Kroeze, M. Amann, J. Cofala, I. Bertok, L. Junfeng,
D. Lin, Z. Shuang, H. Runquing, T. S. Panwar, S. Gupta, D. Singh, A. Kumar,
M. C. Vipradas, P. Dadhich, N. S. Prasad, and L. Srivastava, “The potential
contribution of renewable energy in air pollution abatement in China and India,”
Energy Policy, vol. 30, no. 5, pp. 409–424, Apr. 2002. [Online]. Available: http:
//www.sciencedirect.com/science/article/pii/S0301421501001070
[62] “The Force Is With Clean Energy: 10 Predictions for 2018,” Jan. 2018. [Online]. Available:
https://about.bnef.com/blog/clean-energy-10-predictions-2018/
[63] V. Bermudez, “Electricity storage supporting PV competitiveness in a reliable and
sustainable electric network,” Journal of Renewable and Sustainable Energy, vol. 9, no. 1, p.
012301, Jan. 2017. [Online]. Available: https://aip.scitation.org/doi/full/10.1063/1.4974851
[64] H. T. Nguyen, S. Battula, R. R. Takkala, Z. Wang, and L. S. Tesfatsion, “Transactive Energy
Design for Integrated Transmission and Distribution Systems,” Iowa State University,
Department of Economics, Tech. Rep. 201802280800001000, Feb. 2018. [Online]. Available:
https://ideas.repec.org/p/isu/genstf/201802280800001000.html
[65] R. Abe, H. Taoka, and D. McQuilkin, “Digital Grid: Communicative Electrical Grids of the
Future,” IEEE Transactions on Smart Grid, vol. 2, no. 2, pp. 399–410, Jun. 2011.
[66] L. Tesfatsion, “Chapter 4 - Electric power markets in transition: Agent-based modeling tools
for transactive energy support,” B. H. o. C. Economics, Ed. Elsevier. [Online]. Available:
https://www.sciencedirect.com/science/article/pii/S1574002118300042
[67] C. Stöcker, “Software Defined Digital Grid on a P2p Network — On Systems of Autonomous
Energy Cells,” Mar. 2018. [Online]. Available: https://medium.com/cstoecker/software-
defined-digital-grid-on-a-p2p-network-cdd17c9017e4
[68] U. Ponsukcharoen, “IMPACT OF DISTRIBUTED PV GENERATION ON GRID OPERA-
TION,” p. 172.
[69] P. Denholm and R. M. Margolis, “Evaluating the limits of solar photovoltaics (PV)
in electric power systems utilizing energy storage and other enabling technologies,”
Bibliography 97
Energy Policy, vol. 35, no. 9, pp. 4424–4433, Sep. 2007. [Online]. Available: http:
//www.sciencedirect.com/science/article/pii/S030142150700095X
[70] E. Martinot, “Grid Integration of Renewable Energy: Flexibility, Innovation, and
Experience,” Annual Review of Environment and Resources, vol. 41, no. 1, pp. 223–
251, Nov. 2016. [Online]. Available: http://www.annualreviews.org/doi/10.1146/annurev-
environ-110615-085725
[71] N. O’Connell, E. Hale, I. Doebber, J. Jorgenson, and Technical University of Denmark,
Lyngby, Denmark, “On the Inclusion of Energy-Shifting Demand Response in Production
Cost Models: Methodology and a Case Study,” Tech. Rep. NREL/TP–6A20-64465, 1215318,
Jul. 2015. [Online]. Available: http://www.osti.gov/servlets/purl/1215318/
[72] A. Goyal and K. S. Swarup, “Automatic Demand Response for load Scheduling and Man-
agement based on Customer Participation,” p. 6.
[73] D. G. Papaefthymiou, K. Grave, and K. Dragoon, “Flexibility options in electricity sys-
tems,” ECOFYS, Tech. Rep. POWDE14426, Jan. 2014.
[74] A. Gierek, “EU Strategy on Heating and Cooling,” Strasbourg, Tech. Rep. A8-0232/2016,
2016. [Online]. Available: http://www.europarl.europa.eu/sides/getDoc.do?pubRef=-//EP/
/TEXT+REPORT+A8-2016-0232+0+DOC+XML+V0//EN
[75] B. P. Bhattarai, B. Bak-Jensen, J. R. Pillai, and M. Maier, “Demand flexibility from resi-
dential heat pump,” in 2014 IEEE PES General Meeting | Conference Exposition, Jul. 2014,
pp. 1–5.
[76] D. Vanhoudt, D. Geysen, B. Claessens, F. Leemans, L. Jespers, and J. Van Bael, “An
actively controlled residential heat pump: Potential on peak shaving and maximization of
self-consumption of renewable energy,” Renewable Energy, vol. 63, no. Supplement C, pp.
531–543, Mar. 2014. [Online]. Available: http://www.sciencedirect.com/science/article/pii/
S0960148113005521
[77] J. Hong, N. J. Kelly, I. Richardson, and M. Thomson, “Assessing heat pumps
as flexible load,” Proceedings of the Institution of Mechanical Engineers, Part A: Journal
of Power and Energy, vol. 227, no. 1, pp. 30–42, Feb. 2013. [Online]. Available:
https://doi.org/10.1177/0957650912454830
[78] IRENA, “Renewable Energy in District Heating and Cooling, A sector roadmap for
REmap,” International Renewable Energy Agency, Abu Dhabi, Tech. Rep., Mar. 2017.
[Online]. Available: www.irena.org/remap
[79] IEA and T. University, “District Energy Systems in China,” International Energy Agency,
Tech. Rep., 2017. [Online]. Available: https://www.iea.org/publications/freepublications/
publication/district-energy-systems-in-china.html
[80] “Energiestrategie 2050.” [Online]. Available: http://www.bfe.admin.ch/energiestrategie2050/
index.html
[81] W. J. A. v. Leeuwen, M. Bongaerts, G. M. A. Vanalme, B. Asare-Bediako, and W. L. Kling,
“Load Shifting by Heat Pumps using Thermal Storage,” in Universities’ Power Engineering
Conference (UPEC), Proceedings of 2011 46th International, Sep. 2011, pp. 1–6.
[82] M. Battaglia, R. Haberl, E. Bamberger, and M. Haller, “Increased self-consumption
and grid flexibility of PV and heat pump systems with thermal and electrical
storage,” Energy Procedia, vol. 135, pp. 358–366, Oct. 2017. [Online]. Available:
http://linkinghub.elsevier.com/retrieve/pii/S1876610217346398
[83] A. Mohd, E. Ortjohann, A. Schmelter, N. Hamsic, and D. Morton, “Challenges in integrating
distributed Energy storage systems into future smart grid,” in 2008 IEEE International
Symposium on Industrial Electronics, Jun. 2008, pp. 1627–1632.
98 Bibliography
[84] C. Eid, P. Codani, Y. Perez, J. Reneses, and R. Hakvoort, “Managing electric flexibility
from Distributed Energy Resources: A review of incentives for market design,” Renewable
and Sustainable Energy Reviews, vol. 64, pp. 237–247, Oct. 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S1364032116302222
[85] ——, “Managing electric flexibility from Distributed Energy Resources: A review of
incentives for market design,” Renewable and Sustainable Energy Reviews, vol. 64, pp.
237–247, Oct. 2016. [Online]. Available: http://www.sciencedirect.com/science/article/pii/
S1364032116302222
[86] I. J. Pérez-Arriaga and A. Bharatkumar, “A framework for redesigning distribution
network use of system charges under high penetration of distributed energy resources:
new principles for new problems,” Working Paper, 2014. [Online]. Available: http:
//cadmus.eui.eu//handle/1814/39570
[87] M. Muratori, B.-A. Schuelke-Leech, and G. Rizzoni, “Role of residential demand response
in modern electricity markets,” Renewable and Sustainable Energy Reviews, vol. 33, pp.
546–553, May 2014. [Online]. Available: http://www.sciencedirect.com/science/article/pii/
S1364032114001488
[88] “General Guidelines for Reinforcing the cooperation between TSOs And DSOs,” Brussels,
Tech. Rep., Nov. 2015. [Online]. Available: https://www.entsoe.eu/publications/position-
papers/position-papers-archive/Pages/Position%20Papers/general-guidelines-for-
reinforcing-the-cooperation-between-tsos-dsos.aspx
[89] N. Schwieters, J. van Hoof, D. Etheridge, and A. von Perfall, “Blockchain -
an opportunity for energy producers and consumers,” PwC, Tech. Rep. [Online].
Available: https://www.pwc.com/gx/en/industries/assets/pwc-blockchain-opportunity-for-
energy-producers-and-consumers.pdf
[90] M. Merz, “Potential of the Blockchain Technology in Energy Trading,” in Blockchain
technology Introduction for business and IT managers. Berlin, Boston: De Gruyter
Oldenbourg, 2016. [Online]. Available: https://www.ponton.de/downloads/mm/Potential-
of-the-Blockchain-Technology-in-Energy-Trading_Merz_2016.en.pdf
[91] “EnerChain.” [Online]. Available: https://enerchain.ponton.de/
[92] “BP, Eni deepen blockchain trading in European gas,” Reuters, Jun. 2017.
[Online]. Available: https://www.reuters.com/article/us-bp-eni-blockchain/bp-eni-deepen-
blockchain-trading-in-european-gas-idUSKBN18W1N2
[93] F. Lombardi, L. Aniello, S. De Angelis, A. Margheri, and V. Sassone, “A blockchain-based
infrastructure for reliable and cost-effective IoT-aided smart grids,” 2018.
[94] L. Marshall, A. Bruce, and I. MacGill, “Allocation Rules and Meter Timing Issues in Local
Energy or ‘Peer to Peer’ Networks,” p. 13.
[95] J. J. Sikorski, J. Haughton, and M. Kraft, “Blockchain technology in the chemical industry:
Machine-to-machine electricity market,” Applied Energy, vol. 195, pp. 234–246, Jun. 2017.
[Online]. Available: http://linkinghub.elsevier.com/retrieve/pii/S0306261917302672
[96] S. Lamparter, S. Becher, and J.-G. Fischer, “An agent-based market platform for smart
grids,” in Proceedings of the 9th international conference on autonomous agents and multia-
gent systems: industry track. International Foundation for Autonomous Agents and Multi-
agent Systems, 2010, pp. 1689–1696.
[97] E. Mengelkamp, B. Notheisen, C. Beer, D. Dauer, and C. Weinhardt, “A blockchain-based
smart grid: towards sustainable local energy markets,” Computer Science - Research and
Development, pp. 1–8, Aug. 2017. [Online]. Available: https://link.springer.com/article/10.
1007/s00450-017-0360-9
Bibliography 99
[98] H. YAN, B.-B. HUANG, and B.-W. HONG, “Distributed Energy Transaction Pattern and
Block Chain Based Architecture Design,” DEStech Transactions on Environment, Energy
and Earth Sciences, no. epee, 2017.
[99] E. Munsing, J. Mather, and S. Moura, “Blockchains for Decentralized Optimization of Energy
Resources in Microgrid Networks,” 2017.
[100] W. Inam, D. Strawser, K. K. Afridi, R. J. Ram, and D. J. Perreault, “Architecture and
system analysis of microgrids with peer-to-peer electricity sharing to create a marketplace
which enables energy access,” in 2015 9th International Conference on Power Electronics
and ECCE Asia (ICPE-ECCE Asia), Jun. 2015, pp. 464–469.
[101] G. Salviotti, L. M. De Rossi, and N. Abbatemarco, “A structured framework to assess the
business application landscape of blockchain technologies.” in Proceedings of the 51st Hawaii
International Conference on System Sciences, 2018.
[102] “SolarCoin | SolarCoin.” [Online]. Available: https://solarcoin.org/
[103] “Jouliette at De Ceuvel.” [Online]. Available: https://spectral.energy/about/news/jouliette-
at-deceuvel/
[104] “Climatecoin | The first carbon zero cryptocurrency in the world.” [Online]. Available:
https://climatecoin.io/
[105] “Blockchain energy – ENGIE.” [Online]. Available: https://www.engie.com/en/news/
blockchain-energy/
[106] “US Government Lab Looks to Blockchain for P2p Energy,” Jan. 2018. [On-
line]. Available: https://www.coindesk.com/us-government-lab-blockcypher-tap-blockchain-
for-p2p-energy-transactions/
[107] P. Ledger, “Australian Government Awards Grant to $8 Million Project In
The City Of Fremantle Using The Power. . . ,” Nov. 2017. [Online]. Avail-
able: https://medium.com/power-ledger/australian-government-awards-grant-to-8-million-
project-in-the-city-of-fremantle-using-the-power-2dbadfad50ae
[108] “Conjoule | Our Shared Energy Future.” [Online]. Available: http://conjoule.de/en/home/
[109] “Grid+.” [Online]. Available: https://gridplus.io/
[110] “Austrian Energy Group Eyes Blockchain With Vienna Test,” Feb. 2018. [Online]. Available:
https://www.coindesk.com/austrian-energy-group-eyes-blockchain-vienna-test/
[111] E. Mengelkamp, J. Gärttner, K. Rock, S. Kessler, L. Orsini, and C. Weinhardt, “Designing
microgrid energy markets: A case study: The Brooklyn Microgrid,” Applied Energy, vol.
210, pp. 870–880, Jan. 2018. [Online]. Available: http://www.sciencedirect.com/science/
article/pii/S030626191730805X
[112] “Blockchain Consortium Aims to Create an ’Energy eBay’ | Greentech Media.” [Online].
Available: https://www.greentechmedia.com/articles/read/blockchain-consortium-aims-to-
create-an-energy-ebay
[113] “Power Ledger receives part of $8 million government grant
for Fremantle blockchain energy project,” Nov. 2017. [Online].
Available: https://www.smartcompany.com.au/startupsmart/news-analysis/power-ledger-
receives-part-8-million-government-grant-fremantle-blockchain-energy-project/
[114] “Piclo – Local renewable energy for businesses".” [Online]. Available: https://piclo.uk/
[115] “gridsingularity.com.” [Online]. Available: http://gridsingularity.com/#/
[116] “DAJIE!” [Online]. Available: https://www.dajie.eu/#home
100 Bibliography
and Buildings, vol. 124, no. Supplement C, pp. 120–128, Jul. 2016. [Online]. Available:
http://www.sciencedirect.com/science/article/pii/S0378778816303358
[137] www.ait.ac.at, “Leafs - AIT Austrian Institute Of Technology.” [Online]. Available:
https://www.ait.ac.at/themen/smart-grids/projects/leafs/
[138] D.-I. D. F. Strempfl, “Perspectives for Blockchain and Digitalisation from an DSO’s point-of-
view,” in Heading Towards Sustainable Energy Systems: Evolution or Revolution?, 15th IAEE
European Conference, Sept 3-6, 2017. International Association for Energy Economics, 2017.
[139] M. HOHMANN, C. WAIBEL, R. EVINS, and J. CARMELIET, “Multi-objective optimiza-
tion of the design and operation of an energy hub for the Empa campus,” in Proceedings of
International Conference CISBAT 2015 Future Buildings and Districts Sustainability from
Nano to Urban Scale. LESO-PB, EPFL, 2015, pp. 591–596.
[140] Fang Aimin, Hou Yangguang, and Wang Fang, “Feasibility of Domestic Hot Water
Regulation for Power Grid Peak and Valley Balance: Hotel-Building Case Study,” Journal
of Energy Engineering, vol. 143, no. 5, p. 05017002, Oct. 2017. [Online]. Available:
https://ascelibrary.org/doi/full/10.1061/%28ASCE%29EY.1943-7897.0000441
[141] A. Aryandoust, “Modelling and Optimization Based Control for Demand Response in Active
Distribution Networks,” Ph.D. dissertation, RWTH Aachen University, Zurich, Jan. 2018.
[142] J. Aghaei and M.-I. Alizadeh, “Demand response in smart electricity grids equipped with
renewable energy sources: A review,” Renewable and Sustainable Energy Reviews, vol. 18,
pp. 64–72, Feb. 2013. [Online]. Available: http://www.sciencedirect.com/science/article/pii/
S1364032112005205
[143] A. Zaballos, J. Navarro, and R. Martín De Pozuelo, “A Custom Approach for a Flexible,
Real-Time and Reliable Software Defined Utility,” Sensors, vol. 18, no. 3, p. 718, Feb. 2018.
[Online]. Available: http://www.mdpi.com/1424-8220/18/3/718
[144] J. M. Selga, A. Zaballos, and J. Navarro, “Solutions to the Computer Networking Challenges
of the Distribution Smart Grid,” IEEE Communications Letters, vol. 17, no. 3, pp. 588–591,
Mar. 2013.
[145] “Moody’s: China’s green certificates will lower tariffs for wind and solar power producers,
but support development of renewable energy sector in the long term,” Feb. 2017. [Online].
Available: https://www.moodys.com/research/Moodys-Chinas-green-certificates-will-lower-
tariffs-for-wind-and--PR_361889
[146] N. Z. Aitzhan and D. Svetinovic, “Security and Privacy in Decentralized Energy Trading
through Multi-signatures, Blockchain and Anonymous Messaging Streams,” IEEE Transac-
tions on Dependable and Secure Computing, vol. PP, no. 99, pp. 1–1, 2016.
Declaration of originality
The signed declaration of originality is a component of every semester paper, Bachelor’s thesis,
Master’s thesis and any other degree paper undertaken during the course of studies, including the
respective electronic versions.
Lecturers may also require a declaration of originality for other written papers compiled for their
courses.
__________________________________________________________________________
I hereby confirm that I am the sole author of the written work here enclosed and that I have compiled it
in my own words. Parts excepted are corrections of form and content by the supervisor .
Title of work:
Supervision:
M.Sc. Arne Meeuw
Prof. Dr. Elgar Fleisch
Dr. L. Andrew Bollinger
Student:
Name: Qianchen Yu
E-mail: qyu@student.ethz.ch
Legi-Nr.: 15-927-437