针对批量生产环境下带缓冲区的串并联系统，提出一种基于可变时间窗的机会维护决策模型，旨在提高系统的运行效率，减少因设备故障造成的停机成本。首先，考虑设备故障率随批量动态变化的特性，基于分解法建立不同批量下缓冲区平均在制品数量的估算模型，以反映缓冲区对设备差异化的缓冲效果。在此基础上，结合设备故障率、产能配比和缓冲区状态对维护时间窗的动态作用机制，通过Sigmoid函数调整时间窗的伸缩率，构建基于可变时间窗的系统动态机会维护策略。实例分析表明，相较于传统的固定时间窗模型，该模型能有效提高带缓冲系统正常工作的概率，降低系统维护总成本。

This paper proposes an opportunity maintenance decision model based on variable time windows for serial-parallel systems with buffer in batch production environments, aiming to enhance system operating efficiency and reduce downtime costs caused by equipment failures. First, considering the dynamic variation of equipment failure rates across batches, an estimation model for the average work-in-progress (WIP) quantity in buffers under different batches is established using the decomposition method, reflecting the differentiated buffering effects of buffers on equipment failures. Building on this, a dynamic opportunistic maintenance strategy is developed by integrating the dynamic interaction mechanism of equipment failure rates, capacity allocation ratios, and buffer states on maintenance time windows. The elasticity of time windows is adjusted through a Sigmoid function to optimize maintenance scheduling. Case studies demonstrate that compared to traditional fixed time window models, the proposed model effectively increases the probability of normal system operation and reduces total maintenance costs.