基于缺陷预测的设备状态维修决策与优化方法

发布时间：2018-05-30 10:50

本文选题：状态维修 + 故障模式　；参考：《江苏科技大学》2017年硕士论文

【摘要】：状态维修(CBM)是一种执行方式基于设备运行状态监测,利用监测的状态参数和各种分析手段判断设备状态,安排适当维修时间和维修方式的维修。本文围绕状态维修的重要组成部分--决策及决策过程进行深入的研究,研究了基于设备重要度和故障模式特征的维修方式选择方法、设备缺陷早期识别模型和缺陷状态预知模型,并在此基础上建立维修决策优化模型;通过对这些模型进行计算机仿真和案例研究,验证了决策过程相关模型的可行性和有效性。首先,从设备重要度和故障模式特征两个角度研究了设备维修方式的选择方法。研究了设备的重要度排序方法,分析了层次分析法以及熵值法的特点和适用条件,提出了两者相结合的组合权重分析方法,利用蒙特卡洛模拟确保了方法的客观性;研究了设备故障模式特征的识别方法,提出了一种改进的布谷鸟搜索与最小二乘支持向量机相融合的算法,对设备故障模式进行准确识别;最后以船舶动力设备为实例,验证了基于重要度排序及故障模式特征相结合的设备维修方式确定方法。其次,分析了设备运行阶段状态的特点,研究了设备缺陷状态早期识别方法。从设备状态劣化的角度考虑,对隐半马尔可夫模型(HSMM)及需要解决的评估、解码和学习的三个问题所对应的前向后向算法、Viterbi算法和最大似然估计法进行了研究;在随机滤波理论的基础上,针对设备衰退因子的确定与选择,提出了一种考虑衰退因子的HSMM模型早期缺陷发生时刻识别及剩余寿命计算方法,通过仿真验证了该方法的有效性和可行性。最后,分析了维修行为及维修间隔期的决策及优化问题。研究了基于单位时间维修费用最少的维修决策优化目标,并给出了求解方法;针对目前维修方式优化目标单一的问题,给出了基于设备可靠性最高及单位时间内维修费用最少相融合的多目标决策优化,建立了相应的设备维修决策与目标优化模型,通过仿真验证了该方法的有效性。
[Abstract]:Condition-based maintenance (CBM) is a kind of execution method based on equipment running condition monitoring. The condition parameters of monitoring and various analytical means are used to judge the state of equipment and arrange the proper maintenance time and maintenance mode. In this paper, the decision making and decision process, an important part of state maintenance, is deeply studied, and the method of selecting maintenance mode based on the feature of equipment importance and fault mode is studied. Based on these models, the maintenance decision optimization model is established, and the feasibility and validity of these models are verified by computer simulation and case study. Firstly, the selection method of equipment maintenance mode is studied from the aspects of equipment importance and fault mode characteristics. The importance ranking method of equipment is studied, the characteristics and applicable conditions of AHP and entropy method are analyzed, and a combined weight analysis method combining them is put forward. The objectivity of the method is ensured by Monte Carlo simulation. The recognition method of fault pattern feature of equipment is studied, and an improved algorithm of cuckoo bird search and least square support vector machine is proposed to accurately identify the fault pattern of equipment. Finally, the ship power equipment is taken as an example. The method of determining equipment maintenance mode based on importance ranking and fault mode feature is verified. Secondly, the characteristics of equipment running state are analyzed, and the early identification method of equipment defect state is studied. From the point of view of equipment state deterioration, the forward and backward Viterbi algorithm and the maximum likelihood estimation method for the evaluation, decoding and learning of the hidden semi-Markov model (HSMMM) and the three problems needed to be solved are studied. Based on the theory of random filtering, a method for identifying the occurrence time of early defects and calculating the residual life of HSMM model considering the decay factor is proposed. The effectiveness and feasibility of the method are verified by simulation. Finally, the decision and optimization of maintenance behavior and interval are analyzed. This paper studies the maintenance decision optimization goal based on the minimum maintenance cost per unit time, and gives the solution method, aiming at the problem of the single target of the current maintenance mode optimization. The multi-objective decision optimization based on the combination of the highest reliability of equipment and the least maintenance cost per unit time is presented. The corresponding model of equipment maintenance decision and target optimization is established. The effectiveness of the method is verified by simulation.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TB11

【参考文献】