On the Subject of Solar Vehicles and the Benets of the Technology

John Connors
Jack Baskin School of Engineering University of California, Santa Cruz Santa Cruz, California, 95064

Abstract The face of the automotive industry is being reshaped by concerns over oil supplies, international policy and fuel costs. A wide variety of hybrid technologies are now available including discussion of hydrogen possibilities. The solar powered car, one of the oldest alternative energy vehicles, has many applications to the emerging electric vehicle market. The development of a telemetry system for a solar powered race car aids in a better understanding of the energy usage of a vehicle and the aspects applicable to electric vehicles as a whole. This paper surveys the history and future of solar and electric vehicles and provides an overview of a typical solar car.

I. INTRODUCTION Automotive consumers are growing increasingly more interested in highly fuel efcient vehicles, environmentalists are concerned over vehicle emissions, and politicians are ghting to ease foreign relations and maintain international oil trade. The effects of the combustion engine automobile are diverse and widespread. Research in ethanol, hydrogen, biodiesel and electric vehicles aims to improve energy efciencies while decreasing pollution and dependency on foreign oil. In the early 1900s, the majority of pleasure vehicles sold in the United States were electrically powered. Since the acceptance of the internal combustion engine, electric vehicles became an insignicant portion of the automobile market, but have been gaining market share in recent years. In the 1990s, California passed the Zero Emissions Vehicle Mandate and forced automotive companies to produce electric vehicles. The law was later changed and automakers gained increased exibility in the requirement to produce electric cars [9]. As a result, the electric vehicles in todays market are primarily produced by smaller companies but are slowly becoming more common in the market. The development of electric vehicles as practical, marketable automobiles has been a continual, though slow process. Numerous technologies have developed that continue to increase range and performance while decreasing expense to consumers. For three decades, considerable academic efforts have been applied to the development of solar powered cars. This research has produced many advances, though the technologies have not yet been applied to the more general electric vehicle. Many of the improvements could benet electric cars and further their efciencies in a demanding market. This paper is covers three topics: a history and overview of solar vehicles, the design and function of a telemetry

Fig. 1. Started in 1990, UC Berkeleys California Dreamin was one of the earlier solar vehicles built.

system and a discussion of the solar car technologies that are applicable to the broader electric vehicle market. II. SOLAR VEHICLES A. History of Solar Vehicles The rst combination of photovoltaic devices and electric vehicles happened in the late 1970s. Pressured by the oil crisis, engineers and environmentalists began looking for alternative energy sources and eventually turned to solar. To generate more publicity and research interest in solar powered transportation, Hans Tholstrup organized a 1,865 mi (3,000 km) race across the Australian outback in 1987. Called the World Solar Challenge (WSC), competitors were invited from industry research groups and top universities around the globe. General Motors (GM) won the event by a large margin, achieving speeds over 40 mph with their Sunraycer vehicle [7]. In response to their victory, GM teamed with the US Department of Energy (DOE) to hold the GM Sunrayce in 1990 [8]. Approximately the same length as the WSC, Sunrayce is considered to be a more difcult race due to more diverse terrain and climates as well as more challenging road surfaces and trafc congestion. Further WSC events have been held ever three years along the original route from Darwin to Adelaide, Australia. The Sunrayce, recently renamed the

American Solar Challenge in 2001, then the North American Solar Challenge in 2005, is held every two years across different routes. In 2005, the race set a new record for the longest solar vehicle race, covering 2460 mi (3960 km) from Austin, Texas, USA to Calgary, Alberta, Canada. Despite initially being dominated and funded by General Motors, the design and construction of solar vehicles has produced its own unique development process. Originally motivated by research, the building of solar vehicles is now referred to as brain sport, developing dozens of new vehicles each year for the sole purpose of competition, not production. In contrast to research developing new technologies, solar vehicle competition is an environment in which technology has been built for research. Manufacturers actively produce power trackers, motors and even tires specically for use with solar vehicles. The various competitions were designed to promote exposure and interest in the vehicles and demonstrate a proof of concept for electric and solar vehicles. Today, the vehicles remain as a collegiate sport but continue to develop and improve valuable technologies. Due to the unique nature of the solar community and events, these technologies remain an untapped resource. Signicant improvements and understanding of electric vehicles has been developed that can be applied to a broader range of automobiles to provide more efcient and cleaner alternatives over combustion engine vehicles. B. Solar Vehicle Design Because solar cars are not built in production quantities, each vehicle is unique, though many common characteristics are shared as some components have few variations. All solar vehicles must be light and efcient in order to be practical given the low power output of a solar array. For the most part, each vehicle is composed of the same basic elements. Most modern solar vehicles use either a tubular space frame chassis or a composite monocoque shell (Figure 2). Space frame designs tend to be easier and less expensive to build than a monocoque chassis. The monocoque structure is lighter and simpler than a space frame, and allows for vehicles to easily be split into compartments to separate different mechanical and electrical subsystems. The composite designs are more difcult to analyze and integrate with suspension and steering components without compromising the structural integrity of the shell. The outer body of a solar vehicle is crucial for minimizing the aerodynamic drag induced by air ow at higher speeds. The drag forces are characterized by the shape of the vehicle as well as the vehicles frontal area, the projection of the vehicle onto a plane perpendicular to the ow of air. Solar cars therefore have very low proles and tend to be completely enclosed. To maintain proper air ow, the only moving parts outside the vehicle body are the portions of the wheels contacting the road. All suspension, brakes and inner structural components are enclosed by smooth surfaces to avoid altering the air stream. Small openings are occasionally used to help ventilate the driver and battery pack.

Fig. 2. Numerous tradeoffs exist between tubular space frames and composite monocoque chassis. Designs from UC Berkeley and Kansas State.

In addition to providing aerodynamic advantages, the outer shell also functions as the primary support for the solar array. Though any array size would constitute a solar vehicle, competition cars are usually limited in the total square footage of array space. Numerous materials have been found to be photovoltaic, though the most common cells are made of single junction silicon (Si) or multijunction gallium-arsenide (GaAs). Silicon cells represent the majority of the solar market and, as a result, remain the most inexpensive type of solar cell available. The most efcient Si cells only reach energy efciencies of 20% while GaAs cells have been made with efciencies as high as 40%. GaAs cells are manufactured for space grade applications and remain very expensive for the higher efciency units. Performance drop offs in low light conditions and a narrow wave length spectrum limit the applications of GaAs cells while Si cells tend to perform well in these situations. Solar vehicles remain one of the few electric vehicles to utilize brushless DC motors. Though many early vehicles used lower efciency brushed motors and chain drives, modern vehicles generally use wheel mounted brushless motors that are capable of maintaining highway speeds and achieving electrical efciencies of 96%-98%. Hub mounted motors eliminate the need for a transmission or other gear reductions. Some motors allow the user to vary the gap between the rotor and stator, changing the relationship between speed and torque. These adjustments allow the vehicle to maintain optimal performance through a variety of driving scenarios.

Fig. 3. Manufacturing and assembly irregularities add to the unique shape of a photovoltaic IV curve and variations in the peak power point.

Fig. 4. Maximum Power Point Tracker used to maximize the performance of photovoltaic device and act as a DC-DC converter.

used as a reserve to compensate for poor weather conditions or harsh driving environments. The power needed for propulsion varies with speed, grade, wind and road surface. Likewise, the available solar power varies with weather, incident light angle and temperature. Efcient management of a battery pack or other storage system can minimize the performance effects caused by these variations. Generally, vehicles track the state of charge of a battery system to plan power usage over various terrains and weather conditions. Battery storage density varies as a function of discharge current, but an accurate state of charge estimate can be calculated through the use of Peukerts equations. The greatest advance in solar vehicles over the last few decades is due to improvements in battery technology. Early vehicles used lead-acid batteries, a chemistry still found in most combustion engine vehicles. Though this type of battery is inexpensive and easy to manage, lead-acid cells can represent nearly half the weight of a typical solar vehicle. Improvements in technology led to the use of nickel-metal hydride (NiMH) and nickel-cadmium (NiCad) batteries which have better power to weight ratios than lead-acid batteries [4]. Both nickel based cells are more costly and require more maintenance than their lead-acid counterparts. Today, vehicles must use lithium-ion (Li-ion) or lithium polymer (Li-poly) batteries in order to be successful in competition [6]. Lithium based cells have approximately six times more power per weight over lead-acid batteries. This improvement has resulting in cars that weigh as little as 400 lbs and achieve speeds over 60 mph. Similar to the introduction of NiMH and NiCad, lithium based batteries offer great weight improvements but increase cost and complexity. Lithium technology requires active monitoring of cell voltage, current and temperature to avoid chemical instabilities that degrades the life of the cell and has been known to lead to chemical res. III. TELEMETRY Recent automotive developments have led to vehicles that are considerably more practical than those of the 20th century. In order to continue to develop these vehicles, designers must understand their limitations. From Sun to road, an understanding of all aspects of the car are crucial to the development of more efcient vehicles. To gain a better understanding of the issues involved, we have designed a telemetry system to track the power usage of a modern solar car. A. The Design The vehicle is naturally segmented into subsystems: the solar array, battery pack, motor and peripheral devices, such as displays and a rear view camera. The telemetry system is designed to sample the power usage and other statistics of each subsystem and relay that information via a wireless link to another machine which logs and analyzes the data. The motor controller internally monitors the motor behavior through temperature and hall effect sensors. The controller maintains a series of registers with the most recently logged data which are available through a provided RS-232 interface.

Further optimizations allow the motor poles to be reversed to produce power on deceleration, limiting the power loss of the system. Managing the power from a solar array is an important aspect in the operation of a solar vehicle. The unique shape of a photovoltaic IV curve (Figure 3) results in variations in the available power as the load varies. To maximize the power produced from an array, designers use peak power trackers (PPTs) (Figure 4) to guarantee the cells remain at their peak energy output. Because the shape and magnitude of the curve vary with temperature, light intensity and angle of incidence, a precalculation of the peak power point is not practical and almost certain to be incorrect. PPTs employ various methods to track the maximum power point as light levels and other conditions vary over time. More advanced trackers are also able to overcome local maximums (Figure 3) caused by imperfections in array manufacturing and assembly. Power trackers generally also include a method for transforming the array power to a voltage consistent with the vehicle bus voltage and maintain electrical efciencies as high as 97%-99%. Though not essential, most vehicles utilize a storage element, typically in the form of a battery pack. The storage is

The primary data of interest is the DC voltage and current, motor speed, torque, temperature and the presence of any internal fault conditions. The use of a lithium-ion battery system requires signicant infrastructure to guarantee the cells do not violate various current, voltage, and temperature constraints. As such, data is gathered from groups of cells and compared to the operating restrictions imposed by the manufacturer. The telemetry system can access this data through another RS-232 link and monitor not only the entire battery system but statistics on subcomponents of the batter pack to track any inconsistencies. An additional device is used to monitor the state of charge of the battery system. An e-meter independently measures the overall battery pack voltage and DC current in order to estimate the state of charge by integration with Peukerts Law. As the battery system is vital to the success of a solar vehicle, the e-meter allows for redundant sampling of the battery statistics to protect against failures. This device also has an RS-232 interface and makes the battery pack singular in having the only systems with a continual feed of data without the need for external polling. Power generated from the solar array is transformed to meet the battery voltage by the PPTs. As mentioned previously, these devices also track the maximum power point of the solar array to compensate for numerous dynamic performance factors. Each PPT provides access to tracking data as well as performance information that can be accessed through a CAN bus. The CAN protocol was developed by Bosch for use in automotive sensor networks and works well in this environment. In this vehicle, the array power is spread over ve separate PPTs and allows variations in array performance to be monitored. The energy efcient nature of a solar powered vehicle does not allow of the use of onboard data logging or statistical computation. For this design we implemented a wireless serial modem to relay the telemetry data to another vehicle for analysis. The commercial modem utilizes yet another RS-232 port for communication. With all of these systems in place, a telemetry system was designed to centralize all of this data in one location. The data is packetized and sent over the wireless link to be analyzed. Including the serial modem, this system runs half a dozen communication interfaces at any given time. In order to service all of these links, and to guarantee availability of the processing unit, an FPGA platform was chosen over a traditional microcontroller. The Xilinx XC2S50 offers adequate logic and memory area and the exibility of FPGA designs. Using the Pegasus development board from Digilent Inc., an expansion board was fabricated to interface to four RS232 ports, a CAN interface, and numerous ADC inputs for onboard sensors. Each port uses a line level translation device to convert the digital signal of the FPGA to match the required voltage and coupling of each communication protocol. Interfacing with all the independent subsystems in the vehicle, this board (Figure 6) acts a hub for all communications and a large register le for the telemetric data.

Fig. 5. An FPGA development and custom breakout boards are used as a telemetry system in a solar powered vehicle.

Fig. 6. Data is gathered through a variety of communication interface, packetized and transmitted over a wireless modem.

Using various software to edit and synthesize designs, submodules were created within the XC2S50 to control the various interfaces. Each module is dedicated to monitoring a separate component, polling for information and lling FIFO queues with the most current data samples. In this way, the data in the register le is guaranteed to be correlated within sub second accuracy at any given time. A separate module, triggered off the e-meter data, collects each components data into a queue driving the wireless modem. A subset of this data is processed in the vehicle and available on a custom display for analysis by the driver. The driver is able to make adjustments to driving behavior by monitoring the vehicles speed and power consumption while high-level strategy decisions can be made by team members able to access the full log of telemetry data. Once completed, the rmware designs are stored in nonvolatile ash memory. When the vehicle is powered up, the design is loaded into the XC2S50 and the system begins tracking immediately. Once the sensor data is acquired and correlated, it is transmitted in packets to a remote machine. A data rate of one sample per second was chosen to balance the need for information and the power required by the wireless modem communication device. This data rate is capable of demonstrating the trends found in most driving circumstances including acceleration and heavy braking. As each packet is transmitted, it is logged and stored on the

remote computer. This data can be graphed and analyzed for trends and oddities in behavior. Array faults, battery damage, and even tire wear can be detected by losses in efciency. This data may also be run through simulators to determine the best use for the power available. Based on historical weather patterns and terrain data, alternative routes can be found and strategies can be planned to manage power reserves for current weather systems. The data gathered by systems such as this are useful to better understand the behavior of photovoltaic systems and electric vehicles. Improvements in efciencies and uses of photovoltaic power can be made to better convert the Suns power into a useful resource. This type of system has contributed to many of the technology advances inherent in modern solar cars. IV. APPLICATION TO ELECTRIC VEHICLES Electric vehicles can become more practical and usable by applying many of the technologies rened within the solar vehicle community. To maximize performance and range, designers must maximize the efciency of the electrical components while minimizing the power requirements for motion. The power needed to maintain a constant velocity V must overcome the aerodynamic drag, road grade and rolling resistance of the vehicle [2] and is written as P (V ) = Paero + Pgrade + Proll where Paero = 0.5air Cd Af V 3 Pgrade = mgzV Proll = mgV Cr for air , the density of the air, Cd , the coefcient of aerodynamic drag of the vehicle, Af , the frontal area of the vehicle, V , the velocity, m, the mass of the vehicle, g, the acceleration due to gravity, z, the road grade and Cr , the vehicles coefcient of rolling resistance. In any given environment, air , g and z are constants not within the control of the vehicle. The greatest power loses are due to aerodynamic drag, Paero , being a function of V 3 . As discussed above, solar vehicles tend to have very low frontal area. Most solar cars accommodate only one person, and those that carry a passenger usually do so in a tandem conguration. The drivers lay at a low seating angle, minimizing their height within the vehicle and generally have very little room around them. These optimizations lower the frontal area of the vehicle a great deal but are generally not practical for a marketable vehicle. In addition, limiting suspension travel, battery pack height and the vertical size of other components can decrease the required height of the vehicle. For control and stability issues, decreasing the vehicle width is not always advisable. Further suspension and braking renements, in conjunction with decreased tire width, limits the frontal area of the wheel wells (Figure 7). The vehicles coefcient of drag, Cd , is determined by the vehicles shape and surface. Any disruptions in the ow of air

Fig. 7. Proper airow plays a key roll in minimizing the aerodynamic drag of a solar vehicle.

Fig. 8. Vehicles like the Pontiac Solstice follow a traditional layout of an engine bay, passenger compartment and fuel/trunk storage area.

around the vehicle require greater power to overcome. Figure 7 shows that the given shape not only maintains a smooth ow of air, but also successfully rejoins the airow again on the trailing edge of the vehicle. Traditional vehicle shapes have been dened by the need for an engine bay, passenger compartment and trunk/fuel storage space (Figure 8). Electric vehicle motors tend to require far less space than a combustion engine, generally do not involve a complicated transmission and require no exhaust or emissions equipment. As a result, the traditional vehicle layout is not necessarily applicable or ideal. Motor space can be moved toward the rear of the vehicle, eliminating the need for a drive train. Battery packs can be broken up and spread around, between or under the passengers. More aerodynamic shapes can be used as larger leading and trailing spaces are no longer needed. Furthermore, the vehicle undercarriage can be covered to provide a smoother surface and less disruption to the airow between the vehicle and road surface. The total mass, m, of the vehicle is important to both Proll and Pgrade as well as decreasing the power needed for acceleration. Therefore, minimizing vehicle weight is crucial and demonstrates why battery improvements have been benecial for electric vehicles. The use of aluminum chassis and composite shells also contribute to a lower vehicle weight. All vehicle systems are optimized to minimize the cars mass. Because the vehicle weight is far less than a traditional combustion engine vehicle, suspension and braking components are not required to be as strong, further decreasing the vehicles overall weight. Increased weight savings generally results in

decreased luxury items and though practical in increasing efciencies, the process may decrease marketability. Modern solar powered cars weigh as little as 400-500 lbs but offer the minimum in driver comforts. Additional power loses are accumulated by a vehicles rolling resistance, or the internal friction of the wheels and drive train. Many commercial electric vehicles require a short drive train and gear reduction while some even use multiple gear ratios and require a transmission. Solar vehicles tend to use hub motors capable of higher rotational speeds and requiring no drive train or gearing. Electric vehicles could benet from the use of one or multiple hub motors to limit the drag effects of a drive train. Multiple hub motors do have the side effect of adding additional overall and un-sprung vehicle weight and require control strategies to coordinate their behavior. Improvements can also be implemented on non-driven wheels by limiting the friction of wheel hubs and axles. Solar cars often use xed axles and high precision bearings to minimize the rotational drag of the free wheels. Brakes discs and drums are aligned precisely to avoid brake rub and energy loss. For all wheels, the tire surface is an additional source of rolling resistance. Specially manufactured tires by Bridgestone and Michelin specically for solar vehicles offer slick surfaces as opposed to the knobby tread of conventional tires. Tires are also inated to higher pressures to avoid tire deection and contact area. Unfortunately, these thinner, slicker tires are more prone to puncture and wear and result a shorter life span than traditional tires. V. CONCLUSIONS A vehicle with a footprint of 8m2 can generate 16002400 watts or 2-3 horsepower given current photovoltaic efciencies. Assuming 100% efcient cells, the vehicle would still produce only 11 hp at midday. For this reason, it is unlikely to see solar vehicles produced in large quantities. Even so, in late 2006, Venturi announced the Astrolab solar vehicle, the rst production solar car. The drawbacks of a production vehicle are evident in the projected range of 68 miles. Though the majority of the population will never own or drive a solar vehicle, photovoltaic technologies are beginning to be applied to combustion engine vehicles. Work has been done to use solar power to offset the electrical need of conventional vehicles and reduce the engine load. One application involves the use of solar power to active cool the passenger compartment while the vehicle is parked [5]. The use of air conditioning to cool the vehicle introduces additional load to the engine and results in increased emissions. Additional technologies from solar vehicles can benet the electric cars being developed today. The use of electric vehicles is becoming increasingly more common and will continue to do so over the next few decades. The environmental, economic, and political concerns over combustion vehicles will contribute to an increase in the use of electric vehicles

and drive further advances in battery technologies vehicle efciencies. VI. ACKNOWLEDGEMENTS We wish to thank CalSol, the UC Berkeley Solar Vehicle Team for the opportunity to build this project and deploy it in their vehicle. We enjoyed the chance to participate in the solar vehicle community. We would also like to express our appreciation to Hybrid Technologies, Digilent Inc., Intel Corp. and Synplicity Inc. for their contributions to the project. R EFERENCES
[1] J. Connors, Telemetric development in solar vehicles with synplify pro software, The Syndicated, vol. 5, no. 1, pp. 1,35, 2005. [2] M. W. Daniels and P. R. Kumar, The optimal use of the solar powered automobile, Control Systems Magazine, IEEE, vol. 19, no. 3, pp. 1222, 1999. [3] M. Granovskii, I. Dincer, and M. A. Rosen, Environmental and economic aspects of hydrogen production and utilization in fuel cell vehicles, Journal of Power Sources, vol. 157, pp. 411421, 2006. [4] H. Hoshino, H. Uchida, H. Kimura, K. Takamoto, K. Hiroka, and Y. Matsumae, Preparation of a nickel-metal hydride (ni-mh) rechargable battery and its application to a solar vehicle, International Journal of Hydrogen Energy, vol. 26, pp. 873877, 2001. [5] K. D. Huang, S.-C. Tzeng, W.-P. Ma, and M.-F. Wu, Intelligent solarpowered automobile-ventilation system, Applied Energy, vol. 80, pp. 141154, 2005. [6] B. Kennedy, D. Patterson, and S. Camilleri, Use of lithium-ion batteries in electric vehicles, Journal of Power Sources, vol. 90, pp. 156162, 2000. [7] R. J. King, Photovoltaic applications for electric vehicles, Conference Record of the Twenty First IEEE Photovoltaic Specialists Conference, vol. 2, pp. 2125, 1990. [8] , Recent solar car technology developments including australian world solar challenge results, Conference Record of the Twenty Second IEEE Photovoltaic Specialists Conference, vol. 1, pp. 629634, 1991. [9] S. Shaheen, Californias zero-emission vehicle mandate, Institute of Transportation Studies, vol. UCD-ITS-RP-04-14, 2004. [10] H. Shimizu, J. Harada, C. Bland, K. Kawakami, and L. Chan, Advanced concepts in electric vehicle design, IEEE Transactions on Industrial Electronics, vol. 44, no. 1, pp. 1418, 1997.