Human Error Rate Research Articles

Comparison of task complexity measures for emergency operating procedures: Convergent validity and predictive validity

Human performance while executing operating procedures is critically important for the safety of complex industrial systems. To predict and model human performance, several complexity measures have been developed. This study aims to compare the convergent validity and predictive validity of three existing complexity measures, step complexity (SC), task size, and task complexity (TC), using operator performance data collected from an emergency operating procedure (EOP) experiment. This comparative study shows that these measures have a high convergent validity with each other, most likely because all of them involve the size dimension of complexity. These measures and their sub-measures also have a high predictive validity for operation time and a moderate-to-high predictive validity for error rate, except the step logic complexity (SLC) measure, a component of the SC measure. SLC appears not to contribute to the predictive validity in the experimental EOPs. The use of visual, auditory, cognitive, and psychomotor (VACP) rating scales in the TC measure seems to be significantly beneficial for explaining the human error rate; however, these rating scales appear not to adequately reflect the complexity differences among the meta-operations in EOPs.

Cross-disciplinary method for predicting and reducing human error probabilities in manual assembly operations

Currently, methods for the evaluation of human reliability are mainly applied in safety critical industries (aerospace engineering, nuclear engineering). The Expert System for Task Taxonomy (ESAT) was developed to evaluate the error rate of control and surveillance activities in safety critical areas. In contrast, currently industrial assembly processes are mainly developed concerning an optimal use of the time factor. In this article, the suitability of ESAT for the evaluation of human reliability in manual assembly operations will be examined. By the analysis of an exemplarily considered assembly process it is shown whether and how its work content can be depicted by the ESAT method. In this context, it is also analysed how a modification of the marginal conditions – e.g. the level of specification – affects the calculation results. In the first analysis six manual assembly operations from the automotive industry were analysed and appropriate ESAT-data for its sequences of action were calculated. As a result, the new method allows quantifying potential human error rates before the start of production and comparing planning alternatives under time and cost aspects early. Consequently, the evaluation of human mistakes during the planning process allows developing not only time-optimised but also quality-optimised manual assembly operations.

An empirical study on the basic human error probabilities for NPP advanced main control room operation using soft control

By adopting new human–system 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 specific screen, then choose the controller, and finally 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 insufficient 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.

Application of Integrated Human Error Identification Techniques for Chemicals Cylinder Change Task

This paper outlines the human error identification techniques currently existent to assess and analyze the risk and human error. Even if many formal error identification techniques have existed for years, none could be validated for use to every different domain. This study considers lots of possible error modes that include external error modes, internal error modes, performance shaping factors and psychological error mechanisms, and integrates a number of execution procedures of human error identification techniques, including task analysis, criticality analysis and reliability analysis, to assess the latent human error and hazard risk in a case plant. It is concluded that the integrated method is a valuable aid to develop much safer operational processes and can be used to predict human error rates of critical tasks in the plant.

Quality risk analysis in a cGMP environment: multiple models for comprehensive failure mode identification during the computer system lifecycle

Pharmaceutical quality systems use various inputs to ensure product quality and prevent failures that might have patient consequences. These inputs are generally data from failures that have already occurred, for example process deviations or customer complaints. Risk analysis techniques are well-established in certain other industries and have become of interest to pharmaceutical manufacturers because they allow potential quality failures to be predicted and mitigating action taken in advance of their occurring. Failure mode and effects analysis (FMEA) is one such technique, and in this study it was applied to implement a computerized manufacturing execution system in a pharmaceutical manufacturing environment. After introduction, the system was monitored to detect failures that did occur and these were analyzed to determine why the risk analysis method failed to predict them. Application of FMEA in other industries has identified weaknesses in predicting certain error types, specifically its dependence on other techniques to model risk situations and its poor analysis of non-hardware risks, such as human error, and this was confirmed in this study. Hierarchical holographic modeling (HHM), a technique for identifying risk scenarios in wide-scope analyses, was applied subsequently and identified additional potential failure modes. The technique for human error rate prediction (THERP) has previously been used for the quantitative analysis of human error risk and the event tree from this technique was adapted and identified further human error scenarios. These were input to the FMEA for prioritization and mitigation, thereby strengthening the risk analysis in terms of failure modes considered.

Interactive requirements prioritization using a genetic algorithm

ContextThe order in which requirements are implemented affects the delivery of value to the end-user, but it also depends on technical constraints and resource availability. The outcome of requirements prioritization is a total ordering of requirements that best accommodates the various kinds of constraints and priorities. During requirements prioritization, some decisions on the relative importance of requirements or the feasibility of a given implementation order must necessarily resort to a human (e.g., the requirements analyst), possessing the involved knowledge. ObjectiveIn this paper, we propose an Interactive Genetic Algorithm (IGA) that includes incremental knowledge acquisition and combines it with the existing constraints, such as dependencies and priorities. We also assess the performance of the proposed algorithm. MethodThe validation of IGA was conducted on a real case study, by comparing the proposed algorithm with the state of the art, interactive prioritization technique Incomplete Analytic Hierarchy Process (IAHP). ResultsThe proposed method outperforms IAHP in terms of effectiveness, efficiency and robustness to decision maker errors. ConclusionIGA produces a good approximation of the reference requirements ranking, requiring an acceptable manual effort and tolerating a reasonable human error rate.

Human reliability analysis of the Tehran research reactor using the SPAR-H method

The purpose of this paper is to cover human reliability analysis of the Tehran research reactor using an appropriate method for the representation of human failure probabilities. In the present work, the technique for human error rate prediction and standardized plant analysis risk-human reliability methods have been utilized to quantify different categories of human errors, applied extensively to nuclear power plants. Human reliability analysis is, indeed, an integral and significant part of probabilistic safety analysis studies, without it probabilistic safety analysis would not be a systematic and complete representation of actual plant risks. In addition, possible human errors in research reactors constitute a significant part of the associated risk of such installations and including them in a probabilistic safety analysis for such facilities is a complicated issue. Standardized plant analysis risk-human can be used to address these concerns; it is a well-documented and systematic human reliability analysis system with tables for human performance choices prepared in consultation with experts in the domain. In this method, performance shaping factors are selected via tables, human action dependencies are accounted for, and the method is well designed for the intended use. In this study, in consultations with reactor operators, human errors are identified and adequate performance shaping factors are assigned to produce proper human failure probabilities. Our importance analysis has revealed that human action contained in the possibility of an external object falling on the reactor core are the most significant human errors concerning the Tehran research reactor to be considered in reactor emergency operating procedures and operator training programs aimed at improving reactor safety.

A Sampling Inspection Plan with Human Error: Considering the Relationship between Visual Inspection Time and Human Error Rate

Objective: The aim of this study is to design a sampling inspection plan with human error which is changing according to inspection time. Background: Typical sampling inspection plans have been established typically based on an assumption of the perfect inspection without human error. However, most of all inspection tasks include human errors in the process of inspection. Therefore, a sampling inspection plan should be designed with consideration of imperfect inspection. Method: A model for single sampling inspection plans were proposed for the cases that visual inspection error rate is changing according to inspection time. Additionally, a sampling inspection plan for an optimal inspection time was proposed. In order to show an applied example of the proposed model, an experiment for visual inspection task was performed and the inspection error rates were measured according to the inspection time. Results: Inspection error rates changed according to inspection time. The inspection error rate could be reflected on the single sampling inspection plans for attribute. In particular, inspection error rate in an optimal inspection time may be used for a reasonable single sampling plan in a practical view. Conclusion: Human error rate in inspection tasks should be reflected on typical single sampling inspection plans. A sampling inspection plan with consideration of human error requires more sampling number than a typical sampling plan with perfect inspection. Application: The result of this research may help to determine more practical sampling inspection plan rather than typical one.

Human error rates for speaker recognition

It is commonly assumed that speaker identification by human listeners is an innate skill under certain conditions. As such, human listening tests have served as the benchmark for automatic recognition systems. In recent evaluations comparing human and machine performance on a speaker comparison task, error rates of naïve human listeners far exceed those of machines [special session on Human Assisted Speaker Recognition, IEEE ICASSP, Prague, 2011]. In this presentation, we quantify the performance of naïve listeners in a variety of challenging channel conditions and we compare these results against automatic systems and trained human listeners. The results of these experiments impact the admissibility of both forensic voice analysis and courtroom testimony by human listeners.

Machine Translation Errors

Errors in machine translations of English-Iraqi Arabic dialogues were analyzed using the methods developed for the Human Translation Error Rate measure (HTER). Human annotations were used to refine the Translation Error Rate (TER) annotations. The analyses were performed on approximately 100 translations into each language from four translation systems. Results include high frequencies of pronoun errors and errors involving the copula in translations to English. High frequencies of errors in subject/person inflection and closed-word classes characterized translations to Iraqi Arabic. There were similar frequencies of word order errors in both translation directions and low frequencies of polarity errors. The problems associated with many errors can be predicted from structural differences between the two languages. Also problematic is the need to insert lexemes not present in the source or vice versa. Some problems associated with deictic elements like pronouns will require knowledge of the discourse context to resolve.

ICONE19-43321 Development of Integrated Parameter Database for Risk Assessment at the Rokkasho Reprocessing Plant

A study to develop a parameter database for Probabilistic Safety Assessment (PSA) for the application of risk information on plant operation and maintenance activity is important because the transparency, consistency, and traceability of parameters are needed to explanation adequacy of the evaluation to third parties. Application of risk information for the plant operation and maintenance activity, equipment reliability data, human error rate, and 5 factors of "five-factor formula" for estimation of the amount of radioactive material discharge (source term) are key inputs. As a part of the infrastructure development for the risk information application, we developed the integrated parameter database, "R-POD" (Rokkasho reprocessing Plant Omnibus parameter Database) on the trial basis for the PSA of the Rokkasho Reprocessing Plant. This database consists primarily of the following 3 parts, 1) an equipment reliability database, 2) a five-factor formula database, and 3) a human reliability database. The underpinning for explaining the validity of the risk assessment can be improved by developing this database. Furthermore, this database is an important tool for the application of risk information, because it provides updated data by incorporating the accumulated operation experiences of the Rokkasho reprocessing plant.

A Safety Approach to Predict Human Error in Critical Flight Tasks

Based on accident analysis in recent years, aircrew error has become an increasingly important theme in the aviation industry, due to its crucial contribution to air incidents. The aircrew's task, to provide safe flight under icing condition, has been analyzed in this research. The Human Error Probabilities of each action within the task are calculated in the light of the HEART taxonomy, followed by the prediction of total task failure rate. A tentative validation has been performed according to operational data of the US civil air transport system. The results show that the HEART technique can be used to predict human error rates of critical tasks in the aviation domain.

Human reliability: Benchmark and prediction

A straightforward, general, and simple methodology is given for evaluating and predicting the human error probability (HEP) in transients for accident risk prediction and reduction purposes. The result is validated against all the available data, producing a completely independent assessment of the uncertainty in HEP for safety and risk analysis, and a new validated prediction method. Previous published work established the technical basis for the existence of learning curves for predicting human performance, system outcomes, and accident rates, which was derived and validated against extensive system outcomes and human learning trial data. This latest second-generation minimum error rate equation (MERE)/universal learning curve (ULC) prediction is based on the well-known and proven learning hypothesis. It is therefore applicable to transient and accident analysis, providing the probability of failure or success for individuals as well as for systems. In this paper, new simulator data and three major existing human reliability analysis (HRA) methods used in safety and risk analysis (e.g. technique of human error rate prediction (THERP), human error and assessment technique (HEART), and human cognitive reliability (HCR)) are benchmarked against the new prediction derived directly from the general ULC. Since these existing methods utilize empirical factors, expert judgement and task analysis, the paper demonstrates a totally new objective approach to benchmark human reliability analysis.

Using Expert Models in Human Reliability Analysis-A Dependence Assessment Method Based on Fuzzy Logic

In human reliability analysis (HRA), dependence analysis refers to assessing the influence of the failure of the operators to perform one task on the failure probabilities of subsequent tasks. A commonly used approach is the technique for human error rate prediction (THERP). The assessment of the dependence level in THERP is a highly subjective judgment based on general rules for the influence of five main factors. A frequently used alternative method extends the THERP model with decision trees. Such trees should increase the repeatability of the assessments but they simplify the relationships among the factors and the dependence level. Moreover, the basis for these simplifications and the resulting tree is difficult to trace. The aim of this work is a method for dependence assessment in HRA that captures the rules used by experts to assess dependence levels and incorporates this knowledge into an algorithm and software tool to be used by HRA analysts. A fuzzy expert system (FES) underlies the method. The method and the associated expert elicitation process are demonstrated with a working model. The expert rules are elicited systematically and converted into a traceable, explicit, and computable model. Anchor situations are provided as guidance for the HRA analyst's judgment of the input factors. The expert model and the FES-based dependence assessment method make the expert rules accessible to the analyst in a usable and repeatable way, with an explicit and traceable basis.

Simple summation rule for optimal fixation selection in visual search

When searching for a known target in a natural texture, practiced humans achieve near-optimal performance compared to a Bayesian ideal searcher constrained with the human map of target detectability across the visual field [Najemnik, J., & Geisler, W. S. (2005). Optimal eye movement strategies in visual search. Nature, 434, 387–391]. To do so, humans must be good at choosing where to fixate during the search [Najemnik, J., & Geisler, W.S. (2008). Eye movement statistics in humans are consistent with an optimal strategy. Journal of Vision, 8(3), 1–14. 4]; however, it seems unlikely that a biological nervous system would implement the computations for the Bayesian ideal fixation selection because of their complexity. Here we derive and test a simple heuristic for optimal fixation selection that appears to be a much better candidate for implementation within a biological nervous system. Specifically, we show that the near-optimal fixation location is the maximum of the current posterior probability distribution for target location after the distribution is filtered by (convolved with) the square of the retinotopic target detectability map. We term the model that uses this strategy the entropy limit minimization (ELM) searcher. We show that when constrained with human-like retinotopic map of target detectability and human search error rates, the ELM searcher performs as well as the Bayesian ideal searcher, and produces fixation statistics similar to human.

Super-human multi-talker speech recognition: A graphical modeling approach

We present a system that can separate and recognize the simultaneous speech of two people recorded in a single channel. Applied to the monaural speech separation and recognition challenge, the system out-performed all other participants – including human listeners – with an overall recognition error rate of 21.6%, compared to the human error rate of 22.3%. The system consists of a speaker recognizer, a model-based speech separation module, and a speech recognizer. For the separation models we explored a range of speech models that incorporate different levels of constraints on temporal dynamics to help infer the source speech signals. The system achieves its best performance when the model of temporal dynamics closely captures the grammatical constraints of the task. For inference, we compare a 2-D Viterbi algorithm and two loopy belief-propagation algorithms. We show how belief-propagation reduces the complexity of temporal inference from exponential to linear in the number of sources and the size of the language model. The best belief-propagation method results in nearly the same recognition error rate as exact inference.

Speech fragment decoding techniques for simultaneous speaker identification and speech recognition

This paper addresses the problem of recognising speech in the presence of a competing speaker. We review a speech fragment decoding technique that treats segregation and recognition as coupled problems. Data-driven techniques are used to segment a spectro-temporal representation into a set of fragments, such that each fragment is dominated by one or other of the speech sources. A speech fragment decoder is used which employs missing data techniques and clean speech models to simultaneously search for the set of fragments and the word sequence that best matches the target speaker model. The paper investigates the performance of the system on a recognition task employing artificially mixed target and masker speech utterances. The fragment decoder produces significantly lower error rates than a conventional recogniser, and mimics the pattern of human performance that is produced by the interplay between energetic and informational masking. However, at around 0 dB the performance is generally quite poor. An analysis of the errors shows that a large number of target/masker confusions are being made. The paper presents a novel fragment-based speaker identification approach that allows the target speaker to be reliably identified across a wide range of SNRs. This component is combined with the recognition system to produce significant improvements. When the target and masker utterance have the same gender, the recognition system has a performance at 0 dB equal to that of humans; in other conditions the error rate is roughly twice the human error rate.

Creation and evaluation of a multi-sensory virtual assembly environment

The multi-modal information presentation, integrated into the virtual environment (VE), has potential for stimulating different senses, improving the user’s impression of immersion, and increasing the amount of information that is accepted and processed by the user’s perception system. The increase of the useful feedback information may reduce the user’s cognitive load, thus enhancing the user’s efficiency and performance while interacting with VEs. This paper presents our creation of a multi-sensory virtual assembly environment (VAE) and the evaluation of the effects of multi-sensory feedback on the usability. The VAE brings together complex technologies such as constraint-based assembly simulation, optical motion tracking technology, and real-time 3D sound generation technology around a virtual reality workbench and a common software platform. The usability evaluation is in terms of its three attributes: efficiency of use, user satisfaction, and reliability. These are addressed by using task completion times (TCTs), questionnaires, and human performance error rates (HPERs), respectively. Two assembly tasks have been used to perform the experiments, using sixteen participants. The outcomes showed that the multi-sensory feedback could improve the usability. They also indicated that the integrated feedback offered better usability than either feedback used in isolation. Most participants preferred the integrated feedback to either feedback (visual or auditory) or no feedback. The participants’ comments demonstrated that nonrealistic or inappropriate feedback had negative effects on the usability, and easily made them feel frustrated. The possible reasons behind the outcomes are also analysed by using a unifying human computer interaction framework. The implications, concluded from the outcomes of this work, can serve as useful guidelines for improving VE system design and implementation.

A prototype safety system for new on-the-job training (OJT) in an anthropocentric cell manufacturing system

The anthropocentric cell manufacturing system (ACMS) discussed consists of a skill-based manufacturing scheduling system, a new on-the-job training (New OJT) system and a safety system. The New OJT was developed using e-learning and automated guided vehicle (AGV)-arm robot as a teacher for CNC machine training. Since the machines and the AGV-arm robot are automatic devices, controlled by computer programs, the dangers might arise during training. Therefore, the safety system is required to ensure that a learner will be safe. We propose a prototype of safety system with designed machine-protection in order to protect the learner from dangerous behaviours during a level 1 training of a New OJT in ACMS. The machine-protection focused on protection from danger using image processing with a background subtraction technique and comparison of objects’ position. We also evaluated safety assessments of the learners, consisting of human error probabilities (HEP) and a human error rate (HER). The prototype of this developed safety system showed that the machine-protection certainly raises a warning alarm when the learner rises into dangerous zones during the training. The evaluated safety assessments showed that our designed New OJT training level 1 course performed more efficient safety performance for self-training of the learner.

The Impact of Severe Accident Management Guidance on the Frequency of Containment Failure Modes of a Boiling Water Reactor

Probabilistic safety assessment (PSA) employs a systematic approach to estimate the risk associated with the operation of nuclear power plants (NPPs). Severe accident management guidance (SAMG), which delineates the mitigation actions of core meltdown accidents of NPPs, is developed to support operators and staffs in the technical support centers during the emergency responses of core melt accidents. Proper execution of SAMG could lower the failure probability of containment and reduce the amount of radionuclides released to the environment during the accident. It can be expected that the implementation of SAMG will reduce the risk of NPPs. However, SAMG is not available when most of the conventional level-2 PSA analyses are performed. In the present study, the mitigation actions of SAMG are incorporated into the level-2 PSA model of the ChinShan Nuclear Power Station of the Taiwan Power Company. The NPP analyzed employs a General Electric-designed boiling water reactor-4 with Mark I containment.The effectiveness of the mitigation actions specified in SAMG to terminate the progression of the accident is verified and validated using the MAAP4 code. The containment system event trees and containment phenomenological event trees of the level-2 PSA model are modified to incorporate the new mitigation actions specified in SAMG. The Human Cognitive Reliability (HCR) and Technique for Human Error Rate Prediction (THERP) models are used to quantify the human error probability (HEP) of all the actions in the level-2 PSA model. The MAAP4 code is used to perform thermohydraulic calculations to determine the demand time required in the HEP analysis.The results show that the total frequency of accident progression beyond vessel failure is reduced by 41% and the change in the probability of containment staying intact is not very significant because of the implementation of SAMG. After SAMG implementation, the frequency of containment early failure is reduced by 69.9%. The frequency of suppression pool venting is increased by 77.9%. The changes in the frequency of other containment failure modes are relatively insignificant. The most important human action is specified in Guideline RC/F of Severe Accident Guideline-1, i.e., In-Vessel Injection to Arrest Core Damage.