You are on page 1of 9

Nuclear Engineering and Design 257 (2013) 7987

Contents lists available at SciVerse ScienceDirect

Nuclear Engineering and Design


journal homepage: www.elsevier.com/locate/nucengdes

An empirical study on the basic human error probabilities for NPP advanced main control room operation using soft control
Inseok Jang a , Ar Ryum Kim a , Mohamed Ali Salem Al Harbi b , Seung Jun Lee c , Hyun Gook Kang a , Poong Hyun Seong a,
a b c

Department of Nuclear and Quantum Engineering, Korea Advanced Institute of Science and Technology, 373-1, Guseong-dong, Yuseong-gu, Daejeon 305-701, Republic of Korea Department of Nuclear Engineering, Khalifa University of Science, Technology and Research, P.O. Box 127788, Abu Dhabi, United Arab Emirates Integrated Safety Assessment Division, Korea Atomic Energy Research Institute, 150-1, Dukjin-dong, Yuseong-gu, Daejeon 305-353, Republic of Korea

h i g h l i g h t s
The operation environment of MCRs in NPPs has changed by adopting new HSIs. The operation action in NPP Advanced MCRs is performed by soft control. Different basic human error probabilities (BHEPs) should be considered. BHEPs in a soft control operation environment are investigated empirically. This work will be helpful to verify if soft control has positive or negative effects.

a r t i c l e

i n f o

a b s t r a c t
By adopting new humansystem interfaces that are based on computer-based technologies, the operation environment of main control rooms (MCRs) in nuclear power plants (NPPs) has changed. The MCRs that include these digital and computer technologies, such as large display panels, computerized procedures, soft controls, and so on, are called Advanced MCRs. Among the many features in Advanced MCRs, soft controls are an important feature because the operation action in NPP Advanced MCRs is performed by soft control. Using soft controls such as mouse control, touch screens, and so on, operators can select a specic screen, then choose the controller, and nally manipulate the devices. However, because of the different interfaces between soft control and hardwired conventional type control, different basic human error probabilities (BHEPs) should be considered in the Human Reliability Analysis (HRA) for advanced MCRs. Although there are many HRA methods to assess human reliabilities, such as Technique for Human Error Rate Prediction (THERP), Accident Sequence Evaluation Program (ASEP), Human Error Assessment and Reduction Technique (HEART), Human Event Repository and Analysis (HERA), Nuclear Computerized Library for Assessing Reactor Reliability (NUCLARR), Cognitive Reliability and Error Analysis Method (CREAM), and so on, these methods have been applied to conventional MCRs, and they do not consider the new features of advance MCRs such as soft controls. As a result, there is an insufcient database for assessing human reliabilities in advanced MCRs. In this paper, BHEPs in a soft control operation environment are investigated empirically for BHEPs to apply advanced MCRs. A soft control operation environment is constructed by using a compact nuclear simulator (CNS), which is a mockup for advanced MCRs. Before the experiments, all tasks that should be performed by subjects are analyzed using one of the task analysis methods, Systematic Human Error Reduction and Prediction Approach (SHERPA). Human errors are then checked to analyze BHEPs, human error mode, and the cause of human error when using soft control. 2013 Elsevier B.V. All rights reserved.

Article history: Received 2 November 2012 Received in revised form 22 January 2013 Accepted 23 January 2013

1. Introduction The assessment of what can go wrong with large scale systems such as nuclear power plants is of considerable current interest, given the past decades record of accidents attributable to human error. Such assessments are formal and technically complex evaluations of the potential risks of systems, and are called probabilistic safety assessments (PSAs). A PRA today consider not just hardware

Corresponding author. Tel.: +82 42 350 3820; fax: +82 42 350 3810. E-mail addresses: nuclear82@kaist.ac.kr (I. Jang), arryum@kaist.ac.kr (A.R. Kim), 100035556@kustar.ac.ae (M.A.S.A. Harbi), sjlee@kaeri.re.kr (S.J. Lee), hyungook@kaist.ac.kr (H.G. Kang), phseong@kaist.ac.kr (P.H. Seong). 0029-5493/$ see front matter 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nucengdes.2013.01.003

80

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987

failures and environmental events that can impact upon risk but also human error contributions (Kirwan, 1992). The importance of human reliability related problems in securing the safety of complex process systems has been clearly demonstrated in recent decades. Unfortunately, although many people have devoted efforts to clarifying why human performance deviates from a certain expected level, there are several difculties in scrutinizing human reliability related problems. One critical difculty is that the amount of available knowledge from operating experiences is extremely small because of the infrequency of real accidents (Park and Jung, 2005). Similarly, in a report on one of the renowned HRA methods, Technique for Human Error Rate Prediction (THERP), it is pointed out that The paucity of actual data on human performance continues to be a major problem for estimating HEPs and performance times in nuclear power plant (NPP) task (Swain and Guttmann, 1983). Due to the lack of real accident data in NPPs, data such as the results of experiments and elds studies, experiments using articial tasks, and simulation data have been used as a database for Human Error Probability (HEP) estimation. However, another critical difculty is that most current HRA databases deal with operation in conventional type of MCRs. With the adoption of new humansystem interfaces that are based on computer-based technologies, the operation environment of MCRs in NPPs has changed. The MCRs including these digital and computer technologies, such as large display panels, computerized procedures, soft controls, and so on, are called advanced MCRs (Stubler et al., 2000). Because of the different interfaces, different basic human error probabilities (BHEPs) should be considered in human reliability analyses (HRAs) for advanced MCRs. Although there are many HRA methods to assess human reliabilities such as Technique for Human Error Rate Prediction (THERP), Accident Sequence Evaluation Program (ASEP), Human Error Assessment and Reduction Technique (HEART), Human Event Repository and Analysis (HERA), Nuclear Computerized Library for Assessing Reactor Reliability (NUCLARR), and Cognitive Reliability and Error Analysis Method (CREAM) (Swain and Guttmann, 1983; Swain, 1987; Hallbert et al., 2006, 2007; Wendy and David, 1992; Hollnagel, 1998), these methods have been applied to conventional MCRs, and they do not consider the new features of advance MCRs such as soft controls. This study carries out an empirical analysis of human error considering soft controls under an advanced MCR mockup called compact nuclear simulator (CNS). The aim of this work is not only to compile a database using the simulator for advanced MCRs but also to compare BHEPs with those of a conventional MCR database. Moreover, human error modes and causes of human error when using soft control are also identied.

General characteristics of soft controls are as follows: multiple locations for access, serial access, present and available, physical decoupling of input and display interfaces, interface management control, multiple modes, software-dened functions, and interface system (Stubler et al., 2000; Lee et al., 2011). Based on these characteristics and functions of soft control, human error may be reduced or these changes may conversely increase human error due to interface management complexity. 2.2. Task analysis for soft control A task analysis helps the analyst to understand and represent human and system performance in a particular task and scenario. A task analysis involves identifying tasks, collecting task data, analyzing the data so that tasks are understood, and producing a documented representation of the analyzed tasks. Typical task analysis methods are used for understanding the required humanmachine and humanhuman interactions and breaking down tasks or scenarios into component task steps or physical operations. A task analysis can be dened as the study of what the operator (or a team of operators) is required to do (i.e., the operators actions and cognitive processes) in order to achieve system goals (Jang et al., 2012.). In new operation environment, advanced MCR, the operation actions of operators are divided into primary tasks (e.g., providing control inputs to plant systems) or secondary tasks (e.g., manipulating the user interface to access information or controls or to change control modes). Interface management tasks are referred to as secondary tasks because they are concerned with controlling the interface rather than the plant. Operators should perform secondary tasks to nd appropriate screens or devices by screen navigations and screen selections before they perform the primary task to control a device. Human errors of primary tasks may result in the execution of inappropriate control actions. Human errors involving secondary tasks are likely to cause delays in accessing controls and displays, to disorient the operator within the display system, or to select wrong controls and displays. While conventional MCRs do not have secondary tasks, the secondary tasks of soft control take a relatively large portion. Therefore, not only human error for primary tasks but also that for secondary tasks should be analyzed in soft controls (Lee et al., 2011). During the HRA process, a task analysis should be implemented in advance. In this study, a task analysis of soft control is performed based on the Emergency Operating Procedure (EOP) considering the features of soft control such as navigation tasks, interface management tasks, and so on. There are two kinds of tasks in EOP when using soft control, control of non-safety related functions and control of safety related functions. All tasks including primary and secondary tasks are analyzed using a task analysis method, Systematic Human Error Reduction and Prediction Approach (SHERPA) (Lee et al., 2011; Embry, 1986). As an example, Fig. 1 shows a task analysis using SHERPA. The goal of the task is to reset the safety injection and auxiliary feedwater actuation signal. In order to achieve the goal, the operator selects Reactivity system screen from the operator console (secondary task) and resets the safety injection signal (primary task). For reset of the safety injection signal, there are other subtasks: Press bypass button from the operator console (secondary task), Press the acknowledge button (secondary task), and nally Press bypass button using the input device for the safety component (primary task). Another subtask, Reset the auxiliary feedwater actuation signal, performed to reset the safety injection signal, is then analyzed. The unit tasks can be rearranged as shown in Fig. 2. Each unit task is included in one of four steps: operation selection, screen selection, control device selection, and operation execution. While

2. Soft control 2.1. Denition and general characteristics of soft control In NUREG-CR/6635, soft controls are dened as devices having connections with control and display system that are mediated by software rather than physical connections (Stubler et al., 2000). This denition directly reects the characteristics of advanced MCRs, including that the operator does not need to provide control input through hard-wired, spatially dedicated control devices that have xed functions. Because of this characteristic, the function of soft control may be variable and context dependent rather than statically dened. Also, devices may be located virtually rather than spatially dedicated. That is, personnel may be able to access a particular soft control from multiple places within a display soft control from multiple places within a display system (Stubler et al., 2000; Lee et al., 2011).

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987

81

Fig. 1. Task analysis using SHERPA.

some tasks are similar to those of conventional controls, there are unique features of soft control such as the screen selection step, and some steps, such as control device selection and operation execution steps, have different unit tasks from those of conventional controls. The operation process of soft controls can vary according to the interface features of advanced MCRs (Lee et al., 2011). Once the task analysis is completed, all of the unit tasks are categorized as diagnosis tasks and execution tasks. Based on a review of rst and second generation HRA methods, there are two general task categories: execution and diagnosis. Examples of execution tasks include operating equipment, performing line-ups, starting pumps, conducting calibration or testing, and other activities performed during the course of following plant procedures or work orders. Diagnosis tasks consist of reliance on knowledge and experience to understand existing conditions, planning and prioritizing activities, and determining appropriate courses of action (Gertman et al., 2005). BHEPs for the classied two tasks are estimated by an empirical simulation study in this work.

3. Experiment in simulation environment 3.1. Compact nuclear simulator (CNS) As the name indicates, this simulator is compact and is not a full scope simulator. The reference plant of this simulator is Kori 3 Nuclear Power Unit in Korea, which is a Westinghouse 3 Loop PWR plant. As shown in Fig. 3, the interface of CNS is fully digitalized to make the experimental environment similar to an advanced MCR. The thermal hydraulic part is from SMABRE code, which was developed in Finland, and one group diffusion equation is used for ux calculations. This is linked to the PWR code model and special routines for the steam ow, the turbine, the condenser, and the condensate and feed-water systems. The chemical and volume control system and the protection system are also described. In addition, there are some peripheral parts such as the electrical system, the containment, and the pressure relief tank. Fig. 4 shows 7 subsystems (Reactor Coolant System, Residual Heat Removal System, Main Steam/Turbine System, Feedwater

Fig. 2. Sequence analysis of a soft control task.

82

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987

3.2. Experiment In order to measure human error rate in an emergency situation, experiments with 21 students majoring in nuclear engineering majors are performed under a Steam Generator Tube Rupture (SGTR) accident scenario. The number of human errors is checked in a prepared checklist created from the task analysis and BHEPs are calculated based on the number of errors divided by the number of opportunities. 3.2.1. Procedure The specic experimental procedure of the simulation is based on the SGTR scenario. The CNS provides an accident that requires cognitive action in order to be solved. The subjects then try to nd the causes of the accident by responding symptoms of the accident and they know that the accident is SGTR and nally attempt to stabilize the plant. During the accident, the subjects respond to the alarm signals, plant parameters, and so on. In order to familiarize the subjects with the CNS and how to control the devices, training of each subject is performed. Training procedures are divided into two steps. First, the subjects are trained academically by being educated about Emergency Operating Procedures (EOPs) such as EOP development after TMI accidents, general structure of EOP, event and symptom-based approach, use of EOP, and so on. Second, practical training is performed by using compact nuclear simulator (CNS) used in this experiment. After both academic and practice

Fig. 3. Compact nuclear simulator.

System, Electrical System, Chemical and Volume Control System, and Condenser System) of the CNS. Other smaller subsystems, such as the SG Level control system, are not shown separately. Using this simulator, 79 malfunctions, which are briey summarized in Table 1, can be provided.

Fig. 4. Seven subsystems of the CNS.

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987 Table 2 Human error modes. Error mode E1 (operation omission) Cases

83

S/G pressure high/low indication failure RHR HX discharge valve position failure MSIV fails to shut

Loss of pressurizer proportional heater

Operation blocked

Failure in FW controller (loss of steam ow signal)

Condenser tube leak

E2 (wrong object) Steam line rupture (isolable) E3 (wrong operation) E4 (mode confusion) E5 (inadequate operation) E6 (delayed operation)

-Omission of a step in a procedure -Omission of an instruction in a step -Nonrecognition of unexpected action -Right operation on wrong object -Wrong operation on wrong object -Wrong operation on right object -Operation in wrong direction -Right operation on wrong mode -Wrong operation on wrong mode -Operation mistimed -Operation incomplete -Too late operation

Rod bank uncontrolled in PRZ level high/low indication failure

S/G level high/low indication failure

Pump breaker trip

Rod bank uncontrolled out PRZ pressure high/low indication failure RCP rotor seizure

Loss of air ejector

MSIV closure

Make-up water blocked

FW line break inside containment

Steam pressure indication failure high/low Minimal rod speed

Close all FW valves

Loss of off-site power

Position indication failure PRZ PORV stuck open

Steam dump valve stuck open

Reactor coolant pump trip

Undemanded turbine control valve movement

Bypass of feedwater preheaters Nuclear instrumentation system (NIS) failure

Regen. Heat exchanger leak

trainings are performed, the subjects are tested whether each subject is qualied or not by performing pilot tests. The SHERPA analysis is carried out to determine the tasks required and how operators perform, as shown in Fig 1. To measure the number of human errors for required tasks during the SGTR scenario, the task analysis using SHERPA that yields the optimal solution to the scenario should be completed rst. The checklist in this paper is based on the EOP-3, which is the response procedure for a SGTR. As explained in section 2.2, tasks required for the SGTR are classied into two groups, diagnosis tasks and execution tasks, so that estimation of BHEPs can be calculated with classication of different human error modes. The checklist for the scenario is presented in Fig. 5. From the results of the SHERPA analysis, there are 135 tasks (diagnosis tasks: 30, execution tasks: 105) in total that the subjects need to complete in the checklist. Moreover, it should be considered that each execution task could have different possible error modes. The human errors made by the subjects are then checked on the checklist. The cause of human error is also checked for a human error mode analysis and a further supplementary data analysis. In addition, to verify whether the subjects indeed made errors, the screens that subjects controlled are recorded using a screen saver program. 4. Results 4.1. Human error mode identication Many papers and reports have developed human error mode classication techniques. Although these techniques have used different taxonomies for human error mode, their denition is similar. In this paper, the human error modes were dened by considering the error modes of SHERPA and other human error analysis methods such as THERP. Referred human error modes were from E1 to E6 (operation omission, wrong object, wrong operation, mode confusion, inadequate operation, and delayed operation respectively), as shown in Table 2 (Lee et al., 2011; Embry, 1986; Swain and Guttmann, 1983). Referred human error modes were from E1 to E6 (operation omission, wrong object, wrong operation, mode confusion, inadequate operation, and delayed operation respectively), as shown in Table 2. Each human error mode was classied as follows: (Lee et al., 2011). Operation omission: an operator has a potential to omit a necessary operation when he/she selects an operation to be performed in operating procedures. Also, after selecting the right operation and selecting the right control device, the operation could not be performed due to an inappropriate manipulation of soft controls. Wrong object: when an operator selects the target device, he/she could select a wrong device. In this case, if the operator recognizes the wrong selection, then the operation could be recovered.

Control rod speed malfunction Non-isolable FW line leak

Charging line rupture (isolable)

Steam ow indication failure high/low Failure of auto turbine runback

RCP shaft break

Open all FW valves

Steam dump valves open undemanded

Minimal rod speed

PRZ SV stuck open

RCP seal injection valve fails to close

Letdown heat exchanger leak

Turbine control valve stuck

FW pump trip

Steam dump valve closed undemanded Turbine trip

Stuck control rod in CBC Rupture of steam generator tubes

Drop of single control rod in CBB Leakage of primary coolant into containment Failure in PRZ pressure controller

Condensate pump breaker trip

Table 1 79 predetermined malfunctions in CNS.

Drop of all control rods in CBA Anticipated transient without scram (failure to scram) PRZ spray valve open, fails to close or jammed shut Loss of S/G level signal

Bypass valve position failure

Main steam line break, nonisolable

Stuck S/G safety valve

Loss of condenser circulating water pump Loss of auxiliary feedwater

Auxiliary FW ow uncontrollable

Boron concentration high/low RHR heat exchanger leak

MSR (reheater) steam supply control valve closed HP feedwater heater high - high level Auxiliary FW pump trip

VCT level control failure

RCP seal return ow high

Loss of RHR pump

Service water pump breaker trip

Reactor coolant pump stop transients RCP seal return ow low and high temperature Loss of charging pump

Loss of feedwater heaters

Overlap failure

Axial tilt

Load rejection

84

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987

EOP
E-3 E-3

Operators action/task
Check RCP whether or not it should be stopped Inspect ruptured S/G

Detailed operators action/ task


Charging Pump : at least 1 of them is operating RCS pressure : lower than 97 kg/cm2 Is there unexpected increase of narrowrange level of certain S/G Regulate controller set point ofruptured S/G PORV to 79.1 kg/cm2 Check whether ruptured S/G PORV is closed or not Close steam distribution valve which supply flow to turbine driven Check blow-down isolation valve of ruptured S/G whether or not it should be closed Close MSIV and bypass valve of ruptured S/G HV-108 HV-208 HV-308 HV-313 HV-314 HV-315 HV-304 HV-108 HV-208 HV-308

Error check

Cause

E-3

Isolate let-down flow from ruptured S/G

. . . .

. . . .
Reset SI Signal

. . . .

. . . .
SelectReactivitysystem from the operator console PressBypassfrom the operator console Press the Acknowledge button PressBypassbutton using the input device for the safety component PressResetsafeguardfrom the operator console Press the Acknowledge button PressResetbutton using the input device for the safety component

. . . .

E-3

Reset SI and AUX FW Actuation Signal

Reset AUX FW Actuation Signal

. . . .

. . . .

. . . .
Fig. 5. The error checklist from task analysis.

. . . .

. . . .

However, if he/she does not recognize that the selection is wrong and continues executing the operation on the wrong object, then a wrong operation will be performed. Wrong operation: even though an operator performed appropriate navigations of screens and right selection of the target device, an operator can perform a wrong operation such as pressing OPEN button instead of CLOSE button. Mode confusion: if a control window includes multimode, an operator could perform an operation on the wrong mode. An operator can make a mistake such as increasing the level of pressurizer instead of its pressure in case that the level and pressure of the pressurizer are controlled in the same control panel with mode switch. Inadequate operation: when an operator executes the right operation after appropriate navigations of screens and right selection of the target device, the operation could be executed insufciently, too early or too long/short. All operations that are performed incompletely Delayed operation: due to the wrong selections of screens or devices and recovery of them, an operation could not be performed at the right time. Additional time for reselection of screens or devices could be one of the reasons for such delayed operation.

Fig. 6. Number of human errors.

However, in this study, PSFs were not investigated, because BHEPs according to human error modes should be determined in advance in the HRA process and then PSFs should be applied to the BHEPs for the nal modied HEP. Human errors that occurred were classied using the human error modes dened in Section 4.1. Fig. 7 shows the number of human errors according to human error modes for the execution error, and the diagnosis error mode is added independently. Also, BHEPs for each error mode were calculated, as shown in Fig. 8.

If human error modes excluding the dened human error modes were found in the experiment results, the human error modes were modied by adding new modes. 4.2. Basic human error probabilities 4.2.1. Statistical results According to the experiment procedure, the number of errors made by subjects was recorded in the prepared checklists. Fig. 6 shows the total number of errors regarding diagnosis and execution by each subject. While there were several subjects who did not make any errors, one of the subjects made 10 errors. This difference might be caused by personal Performance Shaping Factors (PSFs).

Fig. 7. Number of human errors according to error modes.

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987 Table 3 Number of human errors and probabilities of human errors according to error modes. Errors/opportunities E1 (operation omission) E2 (wrong object) E3 (wrong operation) E4 (mode confusion) E5 (inadequate operation) E6 (delayed operation) Diagnosis error 5/1281 11/756 5/441 1/42 12/504 71/504 9/360 Probabilities of human errors 0.0039 0.01455 0.01134 0.02381 0.02381 0.01389 0.0143 95% condence limits 0.00050.0073 0.00600.0231 0.00150.0213 0.01050.0371 0.00370.0241 0.00500.0236

85

Error factors 3.87 1.96 3.82 1.88 2.56 2.17

The number of human errors/opportunities, probabilities of human errors, 95% condence limits, and error factors according to human error modes are tabulated in Table 3. In case of mode confusion (E4), statistical analysis was not carried out because a small number of opportunities do not guarantee statistical results such as condence limits, and error factors calculation. Among the BHEPs in execution error mode, inadequate operation (E5) and mode confusion showed the highest error probability but because the number of tasks that related to mode confusion (E4) was relatively smaller than other tasks it was hard to conclude that probability of E5 and E4 were the same due to different sample and calculated error factors. Several BHEPs were compared with established HEP data in the form of THERP tables. In the case of omission error in THERP Table 20-7, shown in Table 4, HEP is 0.01 with an error factor of 3 while the probability of operation omission in this study is 0.0039 with a similar error factor of 3.87. This comparison implies that the soft control environment reduces human error regarding operation omission. In the case of selection errors in THERP Table 20-9 shown in Table 5, HEP is either 0.001 or 0.003 with an error factor of 3 while the probability of a wrong object in this study is 0.01455 with an error factor of 1.96. After the Error Factor (EF) in the THERP table is considered, comparing calculated error factor in this study, HEPs in the results of this study is similar to HEP from THERP table. In the case of failure of administrative control in THERP Table 20-6, shown in Table 6, HEP is 0.005 with an error factor of 10 while the probability of wrong operation in this study is 0.01134 with an error factor of 3.82. Because two values of uncertainty bounds (error factors) are considerably different, it was difcult to compare the result in this study with THERP table when soft control was used for probability of wrong operation. 4.2.2. Human errors observed in this experiment After the calculation of BHEPs, the performance checklists and the recorded operators screens were analyzed again to verify the detailed reasons for human errors. Examples of human errors in this study are given below. As an example of the operation omission (E1), there was one step in the SGTR procedure where subjects have to line up the

following valves: LV-161 open, LV-615 close, LV-459 open, and HV-1, 2, 3 open. However, several subjects failed to line up one of the valves in the procedures. Regarding the human error mode of wrong object (E2), one subject operated the wrong controller. The subject should have performed the following step in the procedure: Turn on pressurizer heater to saturate cooling water at the pressure of ruptured S/G. In order to perform this step, the subject had to select the BACKUP HEATER on the operators console, as shown in Fig. 9a. However, the subject selected the wrong controller, as shown in Fig. 9b. In the case of the human error mode of wrong operation (E3), the subject stopped all Reactor Coolant Pumps (RCPs), as shown in Fig. 10, although RCP 1 and 3 should have been stopped as explained in the instructions. Also, there was a case

Fig. 8. Basic human error probabilities (BHEPs) according to error modes.

Fig. 9. (a) Screenshot regarding right operation on right object. (b) Screenshot regarding right operation on wrong object.

86

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987

Table 4 THERP table 20-7: estimated probabilities of errors of omission per item of instruction when use of written procedures is specied. Omission of item: When procedures with checkoff provisions are correctly used When procedures without checkoff provisions are used, or when checkoff provisions are incorrectly used When written procedures are available and should be used but are not used Short list <10 items Long list >10 items Short list <10 items Long list >10 items HEP 0.001 0.003 0.003 0.01 0.05 Error factor 3 3 3 3 5

Table 5 THERP table 20-9: estimated probabilities of errors in selecting annunciated displays for quantitative or qualitative readings. Selection of wrong display When it is dissimilar to adjacent displays From similar-appearing displays when they are on a panel with clearly drawn mimic lines that include the displays From similar-appearing displays that are part of well-delineated functional group on a panel From an array of similar-appearing displays identied by labels only HEP Negligible 0.0005 0.001 0.003 EF 10 3 3

Table 6 THERP table 20-6: estimated HEP related to failure of administrative control. Task Carry out a plant policy of scheduled tasks such as periodic tests or maintenance performed weekly, monthly, of at longer intervals Initiated a scheduled shiftly checking or inspection function Use written operation procedures under normal operation procedure Use written operation procedures under abnormal operation procedure Use a valve change or restoration list Use written test or calibration procedures Use written maintenance procedures Use checklist properly HEP 0.01 0.001 0.01 0.005 0.01 0.5 0.3 0.5 EF 5 3 3 10 3 5 5 5

where the subject operated a controller in automatic mode when the controller should have been operated in manual mode; this is mode confusion (E4). As a human error mode of inadequate operation, a subject performed a task incompletely. There was a task where the subject had to reset the Safety Injection (SI) and actuated auxiliary feedwater signal in the procedure. For the completion of this task, the subject should follow these steps: 1. Select Reactivity Control System from the operator console, 2. Press bypass from the operator console, 3. Press acknowledge button, 4. Press bypass button using the input device for the safety component, 5. Press reset safeguard from the operator console, 6. Press acknowledge button, 7. Press reset button using the input device for the safety component. In this task, several subjects did not perform steps 24, which resulted in incomplete operation. Fig. 11 shows

complete and incomplete operation procedures, respectively. Next, several diagnosis errors were found. 5. Discussion The results from this research show BHEPs according to dened human error modes when soft control is used in advanced MCRs. In order to investigate BHEPs, a mockup facility, CNS, was set up to simulate an advanced MCR with a focus on soft control. Using CNS, 21 subjects participated in an experiment dealing with a SGTR. Recently, although numerous studies have proven the effects of PSFs on nal HEPs, this research did not consider either external (e.g. darkness, high temperature, excessive humidity, and high work requirement) or internal PSFs (e.g. high stress, excessive fatigue, deciencies in knowledge, skills and experience, etc.). In order to overcome this problem, the instructor tried to ensure a consistent and unbiased environment during the experiments and educated the subjects before the experiments in an effort to minimize the effects of the external and internal PSFs. By maintaining consistent conditions, BHEPs not considering PSFs were extracted. However, as shown in Fig. 6, it was thought that internal PSFs may have affected subjects 18 and 20. These two subjects caused the BHEP to be higher. Because the database of BHEPs are values where PSFs are not applied, PSFs in advanced MCRs should be studied and nal HEPs based on BHEPs (the results of this study) and PSFs (the results of further studies) should be calculated together in future work. Another aspect that should be discussed is the number of subjects and the number of opportunities for E4 in this experiment. This study focused on the effects of soft control on human performance as a starting point and pilot test. More experiments by the authors and other researchers should be performed with a large number of subjects, various scenarios that could extract or induce various error modes (especially E4 in this study), and even normal conditions so that BHEPs based on experimental study will be more concrete and reliable.

Fig. 10. Screenshot regarding wrong operation.

I. Jang et al. / Nuclear Engineering and Design 257 (2013) 7987

87

were then checked on error check lists, classifying human error modes during the accident scenario. Using the results of the error checklists, several statistical and graphical analyses were implemented, such as the number of human errors according to subjects, the number of human errors according to human error modes, and BHEP according to human error modes. Moreover, BHEPs using soft control were compared with various THERP tables to investigate the level of human error reduction when using soft control. These comparisons implied that the soft control environment reduces human error related to operation omission, but there was no significant effect on error regarding wrong operation and wrong object. Also, human errors observed in this experiment were investigated, and this might provide insight to modify soft control design or to prepare efcient operator training methods dealing with soft control. This empirical study will be helpful to verify whether soft control has positive or negative effects on human performance and will be able to provide modied BHEPs for advanced MCRs to various HRA methods if more data are collected continuously. Acknowledgement This research was supported by a Nuclear Research & Development Program of the National Research Foundation (NRF) grant funded by the Korean government (MEST) (grant code: 2012011506). References
Embry, D.E., 1986. SHERPA: a systematic human error reduction and prediction approach. In: International Topical Meeting on Advances in Human Factors in Nuclear Power Systems, Knoxville, TN. Gertman, D., Blackman, H., Marble, J., Byers, J., Smith, C., 2005. The SPAR-H Human Reliability Analysis Method. Idaho National Laboratory, NUREG/CR6883, USNRC. Hallbert, B., Boring, R., Gertman, D., Dudenhoeffer, D., Whaley, A., Marble, J., Joe, J., Lois, E., 2006. Human Event Repository and Analysis (HERA) System, Overview, vol. 1. Idaho National Laboratory, NUREG/CR-6903, USNRC. Hallbert, B., Whaley, A., Boring, R., McCabe, p., Chang, Y., 2007. Human Event Repository and Analysis (HERA): the HERA Coding Manual and Quality Assurance, vol. 2. Idaho National Laboratory, NUREG/CR-6903, USNRC. Hollnagel, E., 1998. Cognitive Reliability and Error Analysis Method CREAM. Elsevier Science, Oxford, UK. Jang, I., Park, J., Seong, P., 2012. An empirical study on the relationships between functional performance measure and task performance measure in NPP MCR. Ann. Nucl. Energy 42, 96103. Kirwan, B., 1992. Human error identication in human reliability assessment. Part 1: Overview of approaches. Appl. Ergon. 23, 299318. Lee, S.J., Kim, J., Jang, S.C., 2011. Human error mode identication for NPP main control room operation using soft controls. J. Nucl. Sci. Technol. 48, 902910. Park, J., Jung, W., 2005. A database for human performance under simulated emergencies of nuclear power plants. Nucl. Eng. Technol. 37, 491502. Stubler, W.F., OHara, J.M., Kramer, J., 2000. Soft Control: Technical Basis and Human Factors Review Guidance. Sandia National Laboratories, NUREG/CR6635, USNRC. Swain, A.D., Guttmann, H.E., 1983. Handbook of Human-Reliability Analysis with Emphasis on Nuclear Power Plant Application. Sandia National Laboratories, NUREG/CR-1278, USNRC. Swain, A.D., 1987. Accident Sequence Evaluation Program Human Reliability Analysis Procedure. Sandia National Laboratories, NUREG/CR-4772, USNRC. Wendy, J.R., David, I.G., 1992. NUCLARR: a workstation software package to support risk assessment. Reliab. Eng. Syst. Saf. 37, 173178.

Fig. 11. Screenshot regarding inadequate operation.

6. Conclusion This paper investigated basic human error probabilities (BHEPs) according to human error modes in a soft control operation environment empirically, because soft control is one of the most important features in advanced MCRs and there is no BHEP database related to the use of soft control in the HRA method. As a mockup facility, a compact nuclear simulator (CNS) that includes features of soft control and advanced MCRs such as large display panels and computer technologies was used to simulate advanced MCRs. In order to extract the BHEPs, a task analysis for soft control, which yielded an error checklist, was completed. From the soft control task analysis it was revealed that secondary tasks that are related to manipulating the user interface to access information or controls or to change control modes are new kinds of tasks that do not exist in conventional MCRs. The errors made by 21 subjects

You might also like