Studies On Dynamic Evolution And Reliability Forecasting Of Operators’ Behaviors In Digital Main Control Room Of Nuclear Power Plants

Digitalization is a trend in large-scale complex systems such as nuclear power plants(NPPs). It changes the way digital NPPs main control room(MCR) operators interact with systems because new technologies have been introduced. These new technologies changed the way of information display, procedures, technique system, and human-system interface(HSI), which are affect human reliability. All these changes result in operators’ behaviors mode also been changed. While digital technology greatly enhancing system functionality and automation level, it also bring some new risks to MCR operators. Whether the reliability of operators’ behaviors can meet safety and economic requirements has become one of the urgent problems that NPPs must solve. How to systematically do the human reliability analysis(HRA) for operators’ behaviors in digital MCR has become the hot and difficult issue no matter in engineering or academic. Considering the importance of this issue and lack of relevant research, this dissertation first investigated operators’ four behaviors: monitoring and detection(MD), situation assessment(SA), response planning(RP), and response implementation(RI), then explored the characteristic of operators’ behaviors in digital MCR context, finally reveal the logical dynamic evolution mechanism of operators’ behaviors and constructed reliability forecasting method for operators’ behaviors. All of these works are based on a large number of experimental studies. The following are the main research works:(1) Identified the human factor characteristic of digital MCR. Compared with traditional analog MCR, the digital MCR has some significant advantages in many aspects. But from the view of contribution to human errors, digital control technology has changed the cognitive context and situational environment, which maybe affect operators’ behaviors reliability. We investigated Lingdong NPP by observing the full size simulator and the MCR, questionnaires to operators, interviewing with operators, observing the training process when they requalification, and discussing with trainee. On the basis of these investigations we summarized the human factor characteristic of digital MCR. We also identified the typical performance shaping factors(PSFs) in digital control system(DCS) from technique system, human-machine interface, procedures, tasks, and crew communication and cooperation.(2) Study on dynamic evolution of operators’ behaviors in digital NPPs MCR. To date, it is still lack of HRA studies on human performance of digital MCR operators. Moreover most of current studies are focus on one of operators’ behaviors, cannot take account the operators’ ’ behaviors as a whole part, and the analysis methods are static. Considering the human-machine system of NPPs is a logical evolution system, it is as the same with the dynamic process of boolean network(BN), so we incorporates BN analysis into the field of HRA to help constructing the analysis model which used to describe the dynamic evolution process of operators’ behaviors. By applying semi-tensor product of matrix(STP), the dynamics evolution of operators’ behavior can be expressed in an algebraic form. Utilizing this algebraic representation, a BN analysis model is proposed, on which we based a qualitative analysis. An illustrative example is given to show how to construct the BN model via experimental data.(3) Reliability forecasting for operators’ behaviors in digital NPPs MCR. While adopting DCS in NPPs, new requirements have been proposed to operators. Faced with digital technical changes, MCR operators need to continue improving their behaviors reliability level. In addition to evaluate operators’ behaviors reliability, managers and shift supervisors also want to forecast their reliability level. By literature review and feasibility analysis, we determined to use “time series analysis plus dynamic network model(DNM)” to do the reliability forecasting for operators’ behaviors. First we identified the influence factors associated with operators’ behaviors reliability, and then we developed the reliability model, finally we proposed a reliability forecasting model by integrating time series forecasting method with DNM. We demonstrated how to predict operators’ SA reliability as an example, and the prediction results are consistent with measurement results.Results of this study expected to improve the existing HRA system; it also can provide a scientific basis for improving the integrity and accuracy of the existing analytical techniques. It can help to provide guidelines for error analysis and error prevention of operators’ behaviors, and to contribute to operators’ training and evaluation, HSI design, human failure event protection. We think it can improve the NPPs safety level. The research results have significant theoretical value and practical significance, and can be extend to other large-scale complex systems which adopting DCS.