Analysis Of Human Error And Projection Technologies For Soft Control At NPPs

With the more digital I&C technologies being applied in nuclear power plants, operation by computer-based soft control has been becoming an inreversible trend in the main control room. Replacing traditional operations by hardware device on hard panels will benefit the nuclear operator plant from more convenient environment for operation,but may have a negative effect on human error reduction because the man-machine interaction and human error modes have been changed significantly. Effectively identifying risk is the premise of controlling risk, reducing risk and ensuring the high level of safety, reliability and economy of nuclear power plant operation. For ensuring nuclear safety, it is necessary to analyze the potential human error modes and estimate their influence on nuclear power plant operation by considering the new features of advanced main control rooms.This thesis presented a research on human error analysis and estimation technology for soft control in advanced main control room. The major works of this thesis include:(1) Taking the normal cold shutdown condition of Chemical and Volume Control System(CVCS) at Pressurized Water Reactor (PWR) as an example, the missions of the operating procedures for restarting CVCS and the potential human errors were analyzed and identified.A platform for human error analysis of soft control (PHEASC) was designed and developed.The human error analysis model for each sub-task was built by CAE technology which is a claim, argument and evidence based methodology for safety justification.(2) Based on the human error analysis model of CVCS operating procedures, the formula for estimating human error probability by soft control in the main control room was presented.A case study was taken for verifying the performance of PHEASC in quantification of human errors of operators.(3) Uncertainty analysis was conducted for analyzing the effects of using different statistical data on human error probability estimation. In this thesis,probability importance was utilized for analyzing the sensitivity of model parameters. Some general suggestions were given for preventing some human error modes which are of high importance to human error probability reduction.(4) Criticality importance analysis was taken for further consider the sensitivity of model parameters and easiness of improving. This thesis proved that the corrective failure probability of primary task is of the most signification efficiency for reducing human error probability of the given operating procedures. Several study cases were selected for validating the influence of improving man-machine interface and conducting operation supervisory into main control room on human error reduction.(5) A general formula for estimating the human error probability was presented by considering the misleading issue of operation supervisory. The thesis discussed the effectiveness of applying intelligent operator support system for human error reduction of operators with different kill levels. A method of determining the misleading probability of intelligent operator support systems was also given.The methodologies and PHEASC for human error modeling and analysis platform will not only assist the technicians in design and improvement of human machine interface for advanced main control rooms at nuclear power plants, but also provide a theoretical basis for evaluating the criterion and applicability of intelligent operator support system and important inputs for probabilistic safety assessment of nuclear power plant.