Human Error Probability Estimation Research Articles

Dynamic human reliability analysis method for nuclear power plant main control room based on SPAR-H and SD

Human reliability analysis (HRA) plays a vital role in probabilistic risk assessment (PRA) by quantifying the reliability and safety of complex technical systems through the estimation of human error probabilities (HEPs). Standardized plant analysis risk human (SPAR-H) reliability analysis is widely used for HRA; it adjusts the nominal HEP by assigning different multipliers to the performance shaping factors (PSFs). However, SPAR-H cannot capture the time-varying nature of HEPs; therefore, it cannot realistically respond to the accumulation and fluctuation of human risk factors across the mission. Hence, this study proposed a systematic method of dynamic HRA. It includes decision-making trial and evaluation laboratory-interpretative structural modeling (DEMATEL-ISM) to qualitatively and quantitatively analyze the hierarchical causality among the PSFs and their influence on each other based on their relative relationship in SPAR-H and construct hierarchical causality diagrams among the PSFs. Additionally, it uses a system dynamics method in conjunction with the analysis of time-dependent PSFs based on the hierarchical causality diagrams of PSFs to respond to the time-varying nature of HEPs. The simulation and rationale check results of the cases show that the proposed dynamic HRA method can realize a more realistic and dynamic estimation of HEPs.

Accounting for dependencies among performance shaping factors in SPAR-H using a regularized autoencoder and WINGS-AISM

The standardized plant analysis risk human reliability analysis (SPAR-H) method is widely used for human reliability analysis to adjust the nominal human error probability (HEP) by assigning different multipliers to the performance shaping factors (PSFs). Nevertheless, SPAR-H suffers from assuming PSFs to be independent without considering any overlaps and dependencies. Therefore, this study introduces a new systematic method to analyze the relationships among the PSFs in SPAR-H qualitatively and quantitatively to obtain more reasonable HEP estimation results. The proposed method comprises three primary aspects: 1) a regularized autoencoder for the denoising and feature extraction of expert evaluation results, 2) the weighted influence non-linear gauge system-based adversarial interpretive structure modeling (WINGS-AISM) method to analyze the relationships among the PSFs and construct their causal hierarchy, and 3) a new relative weighting system to modify the PSF multipliers based on this hierarchy. The results of experiments comparing the proposed method with conventional methods highlight that our method effectively reduces the double counting of overlapping PSFs in SPAR-H, providing more reasonable and accurate HEP estimation results.

Task reliability index for operator performance and failure probability assessment in control room simulators

Main control room simulators are an important information source for improving the empirical basis of human reliability analysis (HRA). Both quantitative and qualitative observations provide valuable insight on the operators’ ability to successfully perform their tasks. However, relevant portions of this evidence are not used in typical approaches for estimation of human error probabilities (HEP), which are based on failure counts. With small sample sizes and the operator performance levels representative for nuclear power plants, very little information enters the probability estimates. This paper presents a novel approach to formalize the collection and use of reliability-relevant evidence from simulators. The methodology treats task reliability with the plant-centred orientation characteristic of the systems or Probabilistic Safety Assessment (PSA) perspective as well as a human-centred view of task performance based in human factors engineering. These are combined to form a Task reliability index (TRI) able to distinguish finer performance variations and consider indications of potential failures. Bayesian updating of HEP distributions is suggested by translating the TRI data into pseudo-failures and trials, preserving the convenient coupling of Beta-Binomial distributions in the Bayesian analysis. The approach is demonstrated using data from the International HRA Empirical Study and is shown to produce results that are comparable to those obtained in the empirical study, which were based on comprehensive expert analyses.

An Integrated CREAM for Human Reliability Analysis Based on Consensus Reaching Process under Probabilistic Linguistic Environment

Human reliability analysis (HRA) is widely used to evaluate the impact of human errors on various complex human–machine systems for enhancing their safety and reliability. Nevertheless, it is hard to estimate the human error probability (HEP) in reality due to the uncertainty of state assessment information and the complex relations among common performance conditions (CPCs). In this paper, we aim to present a new integrated cognitive reliability and error analysis method (CREAM) to solve the HRA problems under probabilistic linguistic environment. First, the probabilistic linguistic term sets (PLTSs) are utilized to handle the uncertain task state assessments provided by experts. Second, the minimum conflict consensus model (MCCM) is employed to deal with conflict task state assessment information to assist experts reach consensus. Third, the entropy weighting method is used to determine the relative objective weights of CPCs. Additionally, the CPC effect indexes are introduced to assess the overall effect of CPCs on performance reliability and obtain the HEP estimation. Finally, the reliability of the proposed CREAM is demonstrated via a healthcare practical case. The result shows that the new integrated CREAM can not only effectively represent experts’ uncertain task state assessments but also determine more reliable HEP estimation in HRA.

Quantitative assessment of human error of emergency behavior for hazardous chemical spills in chemical parks

In-depth research on human error is of great significance to improve the human reliability in the emergency rescue as well as to avoid the occurrence of secondary accidents. Therefore, the standardized process and key emergency behaviors in the emergency rescue process of hazardous chemical leakage accidents are first analyzed. Based on the Cognitive Reliability and Error Analysis Method, the human error mode was sorted out, and the human error probability estimation model was established. Combining the improved Analytic Hierarchy Process and entropy weight method, a model for estimating the severity of human error was established, realizing the quantitative analysis of the human reliability of the emergency behavior. The study showed that the 21 emergency behaviors were clustered into 3 categories, with clustering 1 including 2 emergency behaviors, which have the highest level of severity, clustering 2 including 11 emergency behaviors, and clustering 3 including 8 emergency behaviors. Finally, countermeasures were proposed in four areas, including increasing the level of Common Performance Conditions and the ability to transmit information on accidents, etc. It indicates that the method can widely guide the improvement of emergency response capability in chemical parks.

Path to modeling dynamic performance shaping factors in nuclear power plants operation – A review

AbstractThe evolution of the mechanism of human behavior formation analysis has significantly influenced the development of human reliability analysis (HRA), which aims to calculate human error probability (HEP) with performance shaping factors (PSFs). This paper reviews the typical HRA methods in different generations, the role of PSFs, and their interrelation‐ships in human risk modeling, with the background of nuclear power plants (NPPs). In a retrospective of typical HRA methods, PSF plays a fundamental role in assessing human performance during task operation. However, the subjectivity in defining and evaluating PSFs often leads to a partial representation of human behavior characteristics and human risk evolution, resulting in the neglect of PSF inter‐relationships and conservative HEP estimation. Recent studies have emphasized employing simulation platforms to simulate the task process and obtain data relevant to PSFs that can enable the exploration of the mutual effects to support the calculation of HEP more accurately. Compared to certain previous methods involving over‐simplification and inappropriate assumptions resulting in inaccurate results, current HRA methods are prone to the construction of HEP models based on objective data acquisition and dynamic calculations with process models. This shift enables a better illustration of the intricate relationships among PSFs. Reflecting on the current trend of HRA methodology, this paper proposes a possible PSF quantification based on physiological measurement providing accessible and objective data. It improves the shortcomings in data scarcity and time‐invariance of HEP calculation, thus more accurately and realistically responds to the accumulation and fluctuation of human risks throughout a task.

Handling dependencies among performance shaping factors in SPAR-H through DEMATEL method

The Standardized Plant Analysis Risk-Human Reliability Analysis (SPAR-H) method is a widely used method in human reliability analysis (HRA). Performance shaping factors (PSFs) refer to the factors that may influence human performance and are used to adjust nominal human error probabilities (HEPs) in SPAR-H. However, the PSFs are assumed to be independent, which is unrealistic and can lead to unreasonable estimation of HEPs. In this paper, a new method is proposed to handle the dependencies among PSFs in SPAR-H to obtain more reasonable results. Firstly, the dependencies among PSFs are analyzed by using decision-making trial and evaluation laboratory (DEMATEL) method. Then, PSFs are assigned different weights according to their dependent relationships. Finally, multipliers of PSFs are modified based on the relative weights of PSFs. A case study is illustrated that the proposed method is effective in handling the dependent PSFs in SPAR-H, where the duplicate calculations of the dependent part can be reduced. The proposed method can deal with a more general situation that PSFs are dependent, and can provide more reasonable results.

Human error contribution to accidents in the manufacturing sector: A structured approach to evaluate the interdependence among performance shaping factors

Since the 1970s, Human Reliability Analysis (HRA) methods have received a great interest for the quantification of the Human Error Probability (HEP) in Nuclear Power Plants (NPPs). To this purpose, the second-generation HRA methods consider contextual and cognitive factors - named Performance Shaping Factors (PSFs) - that may influence the workers’ performance during tasks execution. Despite the recent extension of HRA methods to different fields, only few studies refer to the manufacturing sector. In addition, the majority of contributions assume the independence among PSFs, which may result in an over or under estimation of HEP. Therefore, the present paper focuses on the manufacturing sector to propose a Fuzzy DEMATEL (FDEMATEL) based method to support the risk analyst in the quantification of PSF interrelationships and importance, when computing HEP.As a result, the most influential human factors on which taking priority actions to improve the overall human reliability may be identified accurately. Based on a selected list of PSFs, the methodological approach is implemented in an Italian textile company, where experience and training factors are demonstrated to be the most central ones to increase the human reliability when performing the weaving process tasks. The designed approach is well structured and effortless as well as it allows at considering the uncertainty and vagueness of input data and a group decision context.

A novel integrated methodology for human reliability assessment in hydrogen fuelling stations

Hydrogen fuel cell vehicles have proliferated in recent years, leading to a subsequent increase in hydrogen fuelling stations (HFSs). Operation and maintenance activities of an HFS prone to undesired events caused by human errors. Estimating and evaluating the risk associated with operators' performance can help to enhance the safety of such operations. The human error assessment and reduction technique (HEART) methodology has been widely used to assess the reliability of human performance. However, this technique is not entirely reliable since it depends on experts' experience and judgment, and the model's final outcome is subject to uncertainty. Besides, the HEART method ignores the dependencies of all the tasks required to assess human error probability (HEP). This paper presents a new integrated methodology for quantifying the impacts of different conditions on the reliability of operators' performances in HFSs. The proposed framework is constructed based on the Bayesian network (BN) for complimenting the application of HEART to deal with the lack of data and experts' judgment. Hence, it reduces the inaccuracy of HEP assessed by HEART by plotting it to a BN and considering the dependencies of all involved tasks. Subsequently, the best-worst method (BWM) is utilised to deal with the subjectivity of expert judgment. The integrated BN-BWM-HEART technique enhances the safety and reliability of the current operation and maintenance practices of an HFS. The results indicate that the integrated BN-BWM-HEART technique has the potential capacity to reduce the uncertainty associated with HEP estimation. So, the outcome of the developed methodology enhances the safety and reliability of operations and maintenance activities in an HFS.

A Bayesian two-stage approach to integrate simulator data and expert judgment in human error probability estimation

With the ongoing efforts to collect new data for Human Reliability Analysis (HRA) (in particular, from nuclear power plant control room simulators), it becomes important that the coming data will be processed traceably, addressing its underlying variability, eventually in combination with expert judgment. In this direction, this work presents a two-stage Bayesian model to integrate expert-elicited probability estimates and empirical evidence from simulator data in the quantification of HEP values and of the associated variability distributions. The general aim is to provide a data aggregation framework able to mathematically combine diverse information sources throughout the HEP estimation process, in a systematic and reproducible way, contributing to strengthening the empirical basis of future HRA methods. The Bayesian model can be used to produce reference values and bounds for HRA methods as well as to improve the quality of plant-specific HEP estimates for use in Probabilistic Safety Assessment applications. The model is first verified with artificial data and then applied to quantify the HEP of human failure events from literature. Model sensitivity to biases in expert estimates is also investigated.

Estimation of human error probabilities in marine safety services: The case of lifeboat and davits inspection

Shipboard Operation Human Reliability Analysis (SOHRA) method is recognized as a practical tool to predict human error probability (HEP) of operators engaging marine operations. Identifying the generic task types (GTTs) and marine specific error producing conditions (m-EPCs), the tool successfully derives HEP value distribution of critical operations in maritime environment. However, the real-time applications of SOHRA to prioritize and implement the suitable recovery actions have still open for development due to the limited time and expertise at pre-operation stage. This paper adapts SOHRA into lifeboat and davits inspection process as a critical marine safety service. Considering the tasks conducted by marine safety service engineers in routine and remote support modes, the GTTs and m-EPCs are assigned to estimate HEP values. In this context, a remote assistance system, is alternatively extended to involve standardization, camera tracking, and advisory support to enhance human reliability through inspection stages. The findings spotlight the deviation in HEP between routine ( 8.26E + 00) and remote support (4.04E-01) modes. A set of recovery actions (i.e. instructional materials) to remedy the HEP values in routine mode are suggested while it is not required at the remote mode assistance. The application illustrates that remote supporting to the marine service might reasonably reduce service engineers’ error rates. Consequently, the study is expected to enhance SOHRA applications in inspection period, particularly added value to marine service engineers in duty. The further studies on the proposed remote assistance concept as new generation solution will contribute to the service quality of marine safety companies.

New success likelihood index model for large group human reliability analysis considering noncooperative behaviors and social network

Human reliability analysis (HRA) is a systematic method to qualitatively and quantitatively assess and analyze potential human errors for enhancing the safety and reliability of complex socio-technical systems. As a representative HRA approach, the success likelihood index method (SLIM) has been widely applied to quantify the effects of performance shaping factors on human performance and derive the human error probabilities of operational tasks. The influence quantification of performance shaping factors mainly relies on experts’ subjective judgments, where conflict opinions inevitably exist because of their different knowledge and experiences. Besides, the current SLIM studies often involved a small group of experts and assumed them to be independent, which is insufficient to tackle increasingly complicated HRA problems. In this article, a large group SLIM model is developed to calculate the human error probabilities of operational tasks considering experts’ noncooperative behaviors and social relations. Specifically, a large group of experts are included to provide task state assessments under complex uncertain linguistic environment, and clustered into subgroups based on a social network analysis. Then a consensus reaching process is performed to tackle the noncooperative behaviors of experts to obtain a consensus human error probability estimation. Finally, a case study of cargo loading process is provided to illustrate the practicality and effectiveness of the proposed large group SLIM model.

Empirical estimation of human error probabilities based on the complexity of proceduralized tasks in an analog environment

The contribution of degraded human performance (e.g., human errors) is significant for the safety of diverse social-technical systems. Therefore, it is crucial to understand when and why the performance of human operators could be degraded. In this study, the occurrence probability of human errors was empirically estimated based on the complexity of proceduralized tasks. To this end, Logistic regression analysis was conducted to correlate TACOM (Task Complexity) scores with human errors collected from the full-scope training simulator of nuclear power plants equipped with analog devices (analog environment). As a result, it was observed that the occurrence probability of both errors of commission and errors of omission can be soundly estimated by TACOM scores. Since the effect of diverse performance influencing factors on the occurrence probabilities of human errors could be soundly distinguished by TACOM scores, it is also expected that TACOM scores can be used as a tool to explain when and why the performance of human operators starts to be degraded.

Analysis of dependencies among performance shaping factors in human reliability analysis based on a system dynamics approach

Performance shaping factors (PSFs) describe the influence of given contexts on human performance and are used to quantify human error probabilities (HEPs). HEPs are calculated within human reliability analysis, which is a part of probabilistic risk assessment. One challenge lies in the fact that the dependencies among PSFs are neglected, which leads to the poor estimation of HEPs. The objective of this research is to apply a system dynamics approach to the modeling and analysis of the dependencies of PSFs within the standardized plant analysis of risk–human reliability analysis method. Dependencies are built on the causalities of PSFs and are determined by the mutual information theory and the analytic hierarchy process.The method is demonstrated on three human failure events. The first one is the failure of initiating reactor coolant system depressurization. The second one is the failure of throttling the high-pressure injection to reduce reactor coolant system pressure. The third one is the failure of diagnosing the steam generator tube rupture. The sensitive analysis results show that our model is reasonable and robust when dependencies are taken into account.

A hybrid evaluation method for human error probability by using extended DEMATEL with Z-numbers: A case of cargo loading operation

Human error probability (HEP) evaluation and prediction is one of the most significant tasks to enhance human reliability in marine industry. Among various kinds of HEP evaluation techniques, the Human Error Assessment and Reduction Technique (HEART) technique is regarded as an effective and empirical tool that has been widely adopted in various fields. However, current HEART techniques are insufficient to address HEP evaluation problem in which the self-assurance of expert's judgment and inter-dependencies between Error-producing conditions (EPCs) are considered. Therefore, the purpose of this paper is to develop a hybrid HEART framework (H-HEART-F) to address this problem by integrating Z-numbers and the decision-making trial and evaluation laboratory (DEMATEL) method. First, the Z-numbers are introduced to model the uncertainty and self-assurance of evaluation information from various experts. Then, the Z-number power weighted average (Z-PWA) operator is proposed to aggregate the individual evaluation information into a group direct influenced matrix. Next, an extended DEMATEL method based on possibility degree measure is constructed to determine the proportion of effect of each EPC by considering the self-assurance of expert's judgment and inter-dependencies between EPCs. Finally, the HEP estimation for the cargo loading operation in oil or chemical tanker ship is presented to demonstrate the availability and feasibility of the H-HEART-F. After that, the sensitivity analysis and comparison study are conducted to further illustrate the reasonableness and effectiveness of the proposed method.

A modified human reliability analysis method for the estimation of human error probability in the offloading operations at oil terminals

AbstractHuman error is a significant factor contributing to accidents during oil offloading at oil terminals. Fuzzy cognitive reliability and error analysis method (CREAM) is one of the applicable methods to estimate human error probability (HEP). However, the currently used fuzzy CREAM models have issues which weaken the quality of HEP results. Both the max‐min fuzzy CREAM and the weighted fuzzy CREAM neglect the compatibility degree of each activated fuzzy if‐then rule; besides, the max‐min fuzzy CREAM focuses only on the maximum and minimum input data and ignores their weights. Therefore, this paper presents a modified fuzzy CREAM to address all these issues to strengthen the quality of HEP estimation. The proposed approach employs fuzzy degree of each common performance condition (CPC), the weight of each CPC, and the compatibility degree of each activated fuzzy if‐then rule together to calculate HEP. The proposed approach is applied for HEP estimation in an oil offloading operation at the selected oil terminal. The final HEP value is estimated at 8.929 × 10−3. It is concluded that the proposed fuzzy CREAM model can provide more reliable results.

A Bayesian model to treat within-category and crew-to-crew variability in simulator data for Human Reliability Analysis

The models adopted in Human Reliability Analysis (HRA) characterize personnel tasks and performance conditions via categories of task and influencing factors (e.g. task types and Performance Shaping Factors, PSF). These categories cover the variability of the operational tasks and conditions affecting performance, and of the associated Human Error Probability (HEP). However, variability exists as well within such categories, for example because of the different scenarios and plants in which data is collected, as well as of the operating crew differences (within-category and crew-to-crew variability). This paper presents a Bayesian model to mathematically aggregate simulator data to estimate failure probabilities, explicitly accounting for the specific tasks, scenarios, plants and crew behavior variability, within a given “constellation” (i.e. combination) of task and factor categories. The general aim of the proposed work is to provide future HRA with reference data with stronger empirical basis for failure probability values, both for their nominal values as well as for their variability and uncertainty. Numerical applications with both artificially-generated data and real simulator data are provided to demonstrate the effects of modelling variability in HEP estimates, to avoid potential overconfidence and biases. The applicability of the proposed model to ongoing simulator data collection programs is also investigated.

Analysis and Estimation of Human Errors From Major Accident Investigation Reports

Abstract Risk analyses require proper consideration and quantification of the interaction between humans, organization, and technology in high-hazard industries. Quantitative human reliability analysis approaches require the estimation of human error probabilities (HEPs), often obtained from human performance data on different tasks in specific contexts (also known as performance shaping factors (PSFs)). Data on human errors are often collected from simulated scenarios, near-misses report systems, and experts with operational knowledge. However, these techniques usually miss the realistic context where human errors occur. The present research proposes a realistic and innovative approach for estimating HEPs using data from major accident investigation reports. The approach is based on Bayesian Networks used to model the relationship between performance shaping factors and human errors. The proposed methodology allows minimizing the expert judgment of HEPs, by using a strategy that is able to accommodate the possibility of having no information to represent some conditional dependencies within some variables. Therefore, the approach increases the transparency about the uncertainties of the human error probability estimations. The approach also allows identifying the most influential performance shaping factors, supporting assessors to recommend improvements or extra controls in risk assessments. Formal verification and validation processes are also presented.

Realistic estimation of human error probability through Monte Carlo thermal-hydraulic simulation

Human reliability analysis (HRA) plays an important role in probabilistic safety assessment because human error probability (HEP) strongly influences not only core damage frequency in the nuclear field, for example, but also risk insight across many industries. Conventional HRA methods though use thermal-hydraulic analysis with best-estimate or conservative conditions to estimate the available time for a given operator action, thus significantly overestimating HEP. In response, this paper suggests a new method to more precisely quantify HEP using realistic thermal-hydraulic analyses with exhaustive conditions. The developed method was applied to the rapid cooldown operation in a small loss-of-coolant accident in the APR1400 nuclear power plant. As a result, it was found that conventional HRA methods currently overestimate the risk-significance of this rapid cooldown operation by a factor of six. It is expected that the new HRA method can be applied to not only the nuclear field but also other complex systems to accurately estimate system failure probabilities for specific scenarios by means of more realistic HEP.

BN-SLIM: A Bayesian Network methodology for human reliability assessment based on Success Likelihood Index Method (SLIM)

Success Likelihood Index Model (SLIM) is one of the widely-used deterministic techniques in human reliability assessment especially when data is insufficient. However, this method suffers from epistemic uncertainty as it extremely relies on expert judgment for determining the model parameters such as the rates and weights of the performance shaping factors (PSFs). Besides, given an operation consisting of several tasks, SLIM calculates the human error probability (HEP) by ignoring possible dependencies among the tasks.The present study is aimed at using Bayesian Network (BN) for improving the performance of SLIM in handling uncertainty arising from experts opinion and lack of data. To this end, SLIM is combined with BN to form the so-called BN-SLIM technique. We demonstrate how BN-SLIM can consider uncertainty associated with the rates of PSFs by using probability distributions. BN-SLIM is also able to provide a better estimation of human error probability by considering conditional dependencies resulting from common PSFs. The probability updating feature of BN-SLIM can be used to identify the PSFs contributing the most to human failure event. The outperformance of BN-SLIM over SLIM is demonstrated via an illustrative example.