Reliability Modelling And Efficient Evaluation Of Multi-performance Multi-state Power Generation System

With the rapid development of society and the expansion of power demand,the power generation system is becoming increasingly complex.The analysis of the risk and reliability for power generation system(PGS)is key to the system safety and operation.In PGS,the factors like unit failures will affect the system reliability.In order to carefully model the system reliability,the multi-state system(MSS)model is applied.The MSS model is the extension of binary state system,where the component only admits binary states: "normal operation" or "complete failure",while MSS can have multiple states.In order to model the characteristics of generators more accurately,the number of unit states gradually increases;meanwhile,the number of generators in the PGS also increases,thus the system scale is becoming larger and larger.How to efficiently evaluate the reliability of large-scale PGS becomes a big challenge.On the other hand,more energy forms get involved in PGS.With the rapid development of integrated energy systems,more multi-energy generators,like combined heat and power(CHP)unit,are participating in the PGS.Unlike traditional single-performance generators,these generators output multiple forms of energy such as electric and heat,and traditional MSS model cannot analyze the multiple performance system reliability.Therefore,it is necessary to analyze and evaluate the reliability of a multi-performance power generation system.This paper focuses on the reliability analysis of PGS,studies the efficient reliability evaluation of large-scale PGS,and the reliability model for multi-performance PGS.The main content is as follows:(1)The multi-state power generation system reliability model is proposed.The reliability problem for industrial PGS is clearly illustrated,and the universal generating function method is applied to evaluate the system reliability.The system structure tree method is proposed to represent the PGS structure.(2)A new approximation method,termed as continuization-discretization approximation(CDA)method is proposed to evaluate the reliability of large-scale PGS.First,based on the system structure tree method,a series-parallel PGS is divided into the series subsystems and the parallel subsystems.Second,based on the central limit theory,the continuization process is applied to approximate the reliability curve,and based on random sampling,the discretization process is employed to simplify the calculation.Different acceleration strategies can effectively improve the calculation efficiency of the algorithm.Finally,adjust the algorithm parameters to balance the calculation efficiency and accuracy of the approximation method.The proposed CDA method show its high advantage in efficiency in large-scale PGS reliability evaluation with satisfactory accuracy.(3)Reliability modeling and analysis method for multi-performance PGS are proposed.First,the key role of performance measure vector is represented and 4 ordering methods for performance measure vector are proposed.Second,the fundamental properties of system reliability such as separability,monotonicity,and relevance are developed in the multi-performance multi-state system.Finally,the multi-performance universal generation function method is introduced to evaluate the system reliability,and the reliability evaluation for the performance-coupling system is analyzed.