Based Model Of Human Reliability Of Nuclear Power Plant Control Room HMI Digital Optimization

The human-machine interface of main control room is the main place of the monitoring and controlling system safe running process for operators, main manipulating process, cognitive process, decision-making process and safe controlling center of whole system for operators. The experience shows that human-machine interface design brings influence to information obtained and judgment.Currently, digital human-machine interface is a new field. It is different from traditional human-machine interface. These differences include displaying information, controlling, operation, interfacing with and managing user interface, alarm system and regulations, which will undoubtedly bring new challenge to operators. Aiming at the situation, taking nuclear power plant for the reference background in the paper, from the aspect decreasing human factors events, the author optimizes the human-machine interface to decrease the burden and challenge for operators.In order to resolve the optimizing problem of the digital human-machine interface based on human factors reliability, the author in the paper does some researches as follows:(1) Analyzes the differences about traditional human-machine interface and digital human-machine interface in detail;(2) Analyzes the design factors about traditional human-machine interface and digital human-machine interface respectively;(3)Proposes qualitative optimizing methods for many factors of digital human-machine interface;(4) In the monitor process after event, the layout of among the position of function block for digital human-machine interface is very important for human factors reliability, namely, if the layout of among function blocks in the same display is better the errors of monitor process can be greatly decreased. In order to resolve the situation, the author proposes Bayesian networks digital human-machine interface monitor unit layout optimized model based on improved genetic algorithm. The author poses a complete optimized process iterative figure in the process, utilizes improved hybrid genetic algorithm to the combination of layout, makes use of Bayesian method as adapting function, takes human factors reliability for base. For the calculation of human factors reliability the author uses dynamic simulative function that includes the effective factors about time parameter and human being factors. The process of calculation in the model contains both continuity and discreteness, which is in keeping with realistic cases. Experiment proves that the improved hybrid generic algorithm proposed has good stability and sensitivity and the proposed optimized model has good precise and stability;(5) In monitor process after event, the quantity of parameters and alarm etc affect on the monitor process except the layout among function block of display.aiming at the problem, The author proposes Markov monitor transferred digital human-machine interface optimized model based on fuzzy immune segment evolutionary algorithm in order to provide an advanced optimized method for the quantity of digital human-machine interface parameter. The model mainly optimizes the quantity of parameters of display so that affect on the information reliability obtained by operators and monitor transferring of operators. In order to resolve the situation, the author firstly establishes a complete optimized process figure then researches and defines each parts of the process in detail. The author uses dynamic fuzzy method to come into fuzzy segment and poses improved stepwise method that can greatly improve searching performance in the searching factors and immune segment method in evolutionary process that can greatly improve the performance of evolutionary. The author proposes a dynamic model for designed quantity of optimized factors, obtains two balanced parameter values by many experiments, raises monitor transferred Markov model with condition for human factors reliability analysis of optimized factors. The proposed model has a good accuracy and stability by a lot of experiments. Finally, experiment analysis indicates that the error probability function of parameter quantity affinity of human-machine interface is very sensitive and has very good stability, that optimized factors searching performance algorithm of improved stepwise method is better than general algorithm, and that Markov monitor transferred human factors reliability optimized model is very stable performance and has very good sensitivity.(6) For digital human-machine interface in nuclear power plant, after events are appeared, only by the way that the operators need correctly finish every operation regulation step by step in limited time the whole task can be finished. The operators need switch and transfer among six different displays in the process. If the operation regulation automatic layout in the display is adapt the moving time and researching time can be decrease for operators such that decrease the human factors events. So, the author proposes neural networks human reliability optimized model of a algorithm based on digital human-machine interface regulations automatic layout shortest moving path under the condition of event. For the proposed model the author firstly illustrates qualitative analysis of event in detail then establishes an optimized process figure and finally analyzes a few parts for optimized process in detail. Firstly, the author poses a nearest path algorithm of dynamic sign, proves the proposed algorithm and analyzes the process of algorithm in detail; Secondly, utilizes neural networks method to build optimized model that designs some simulative dynamic functions and includes the obtained optimized factor of regulations layout, namely t; Finally, clearly shows the relative figure of error probability and the good performance of the model in the simulative process.(7) The author applies the established digital human-machine inferface optimized model in the paper to error safety inject events and obtain corresponding experiment results that include three parts as follows:①the most optimized layout for main interface in the event of error safety inject;②the most optimized quantity about parameters, alarm, character of a line for main interface in event of error safety inject;③the best result of executing regulations automatic layout in event of error safety inject.The research results of applicaton prove the effectiveness of the builded model.(8) The research emphasis exists in establishing three different quantitative optimized models for different factors of digital human-machine interface after events. The differences for three models are that digital human-machine interface monitor layout optimized model based on improved genetic algorithm mainly resolves how to layout for each function blocks of the same display, digital human-machine interface Markov monitor transferred human factors reliability optimized model based on fuzzy immune segment evolutional algorithm mainly resolves the optimized problems of the number of parameters and that neural networks human reliability optimized model of algorithm based on digital human-machine interface regulations layout shortest moving path mainly resolves how to make the regulations automatic layout in one of the six displays at certain time. On the other hand, there are common characters among them, firstly, in the optimized process they all take human factors reliability as basis. The common points embody that for optimized model of function blocks layout, if human factor reliability is more high the layout combination more proper, for the optimized model of parameter design, if human factor reliability and affinity are more suitable the number of parameters are more good that for the regulation layout, if regulation automatic layout is more good the human factor reliability is more high; Secondly, the three models all resolve the different factor optimized problems of digital human-machine interface from different aspects after event.