Human Error Analysis And Its Prevention Technology In Advanced Main Control Room Of NPPs

The statistical results from recently accidents in nuclear power plants (NPPs) shows that human error has become the primary attribution of system risk. The purpose of human error analysis and human reliability analysis (HRA) is to identify, predict and prevent the possible human errors during operation and reduce risks caused by human factors. With the applications of computer technologies in instrumentation and control (I&C) system of NPPs,the development of the main control room turns towards more digital and advanced. On the one hand, the automation level, comprehensive control performance and reliability of digital systems are significantly improved. On the other hand, from the perspective of human factors,the context of operators has been changed because of the new characteristics of human machine interfaces (HMIs), and it results the change of the cognitive processes and operation manners of operators. The traditional human error analysis methods and data base may not be completely suitable for use. Additionally, the prevention technologies of human errors are necessary to reduce the risk attributed to human factors.The dissertation is based on a research projects hosted by Harbin Engineering University and China Nuclear Power Design Company,entitled “research of the intelligent main control room". Considering the human factor issue in design of main control room and human machine interface, the article presents the research of human error identification, human reliability evaluation and prevention technologies of human errors. The major works of the thesis are conducted as follows:1) Taking the soft control tasks defined by line up procedures of chemical and volume control system (RCV) of CPR-1000 as an example, hierarchical task analysis is performed considering design of HMIs and characteristics of soft control, all the human error modes especially slips happened in the performance of procedures are identified. Considered the task dependence and human error recovery, evaluation of human error probabilities (HEP) is performed based on statistical data. It provides evidence for identifying the weaknesses of tasks performance.2) Sensitivity analysis is conducted to evaluate the effects of parameters in quantitative human reliability model to the execution HEP of the whole soft control task, in particular nominal HEP of each soft control human error modes identified and empirical value of human error recovery failure probability. With the probability importance degree and crucial importance degree analysis, human error modes which can dramatically influence the task performance are identified and modes can be easily improved from engineering point of view are also explicit. The effects of several performance shaping factors (PSFs) for error recovery on execution HEP are discussed, it emphasizes that the design of human machine interface and procedure performance supervisory are important for preventing human errors.3) The dissertation presents a human error preventing technology based on multilevel flow model (MFM) which is a representative qualitative modeling framework. According to the available functions and achieved task goals in different duration in the process of performing line up procedures of RCV, each of MFM models representing a task phase or a specific system state is constructed and demonstrated. With the MFM models,it is expected that the keyhole effect would be solved. On the basis of MFM cause-consequence reasoning theory, it presents two kinds of procedure performance supervisory strategy that are goal achievement based and alarm analysis based,respectively. More effective and operation relevant information is provided for crews, the situation awareness is expected to be solved.In the end of the thesis, it quantitatively discusses the range of use of the computerized operator support systems (COSSs), which are applied to human error preventing and procedure performance monitoring,such as MFM. It is expected that the results can provide a technical measure for the design and final application of COSSs.To perform qualitative and quantitative analysis of human errors and to develop error prevention technologies as presented in the article, it is expected that the results can provide evidence for crew training, design of HMIs and operation management, and the risks attributed to human factor can be further reduced.