广义更新过程下生产系统最优维修策略  

Optimal maintenance strategy of production system under general renewal process

作　　者：金海波 陈赛东[1] Jin Haibo;Chen Saidong(College of Software,Liaoning Technical University,Huludao Liaoning 125105,China;National&Local Joint Laboratory of Energy-saving Control Technology for Industrial Equipment,Dalian University of Technology,Dalian Liaoning 116024,China)

机构地区：[1]辽宁工程技术大学软件学院,辽宁葫芦岛125105 [2]大连理工大学工业装备节能控制技术国家地方联合实验室,辽宁大连116024

出　　处：《计算机应用研究》2025年第4期1150-1157,共8页Application Research of Computers

基　　金：国家自然科学基金资助项目(62173171);辽宁省科技厅资助项目(2022-MS-397)。

摘　　要：当生产系统的部件发生故障时,选择何种维修策略进行维修对生产制造企业具有重要意义。以往对维修策略的研究多集中于将维修成本最低作为研究目标,而忽略了维修活动对系统净收益产生的潜在价值。为此,针对如何最大化系统生命周期内获得的净收益问题,设计了基于广义更新过程的维修/更换策略,该策略通过引入维修次数和虚拟年龄刻画系统可靠度,利用模拟技术实现相邻两次维修策略时的可靠度转移概率。通过定义维修成本、更换成本、运行收益、运行成本给出每个策略周期内的系统净收益,以此构造系统生命周期内净收益的函数表示,以该净收益的最大值为目标函数建立维修策略模型。为求解该模型,提出了基于半Markov策略过程、参数估计、模拟技术和动态规划技术的求解算法。为了验证模型的正确性和求解算法的有效性,以数控机床主轴为研究对象,对维修策略进行数值实验并与Rosqvist模型结果进行了对比,对比实验结果验证了维修模型的正确性和高效性。敏感性分析结果表明,关键参数对最优维修活动均具有显著影响。该模型能够为企业在系统生命周期内实现净收益最大化提供有效维修决策。When a component in a production system fails,selecting an appropriate maintenance strategy is crucial for manufacturing enterprises.Most previous studies focus on minimizing maintenance costs,overlooking the potential impact on the system’s net profit.To maximize net profit over the system’s life cycle,this paper designed a maintenance/replacement strate-gy based on the generalized renewal process.This strategy used maintenance count and virtual age to model system reliability,and simulation techniques to capture reliability transitions between maintenance decisions.This paper established a maintenance model that maximized net profit by defining maintenance and replacement costs,as well as operating profit and cost.This paper solved the model using a semi-Markov decision process,parameter estimation,and dynamic programming.To verify the correctness of the model and the effectiveness of the algorithm,humerical experiments on maintenance strategies were conducted with the CNC machine tool spindle as the research object and compared with the results of the Rosqvist model.The comparison of the experimental results verify the correctness and efficiency of maintenance model.Sensitivity analysis reveals that key parameters significantly affect optimal maintenance decisions.In conclusion,this model provides an effective approach for maximizing net profit in the system’s life cycle.

关 键 词：维修策略 可靠性 动态规划 WEIBULL分布 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]