Application of discrete random forest algorithm in multi-person asynchronous parallel disassembly sequence planning for hydropower station equipment maintenance

Working procedure is the key link of equipment maintenance in hydropower stations. Excellent working procedure is beneficial to improve maintenance quality and efficiency. However, for the complex hydropower station equipment with multi-part, it is difficult to obtain the multi-person collaborative working procedure of maintenance tasks. To overcome this problem, this paper establishes a multi-person asynchronous parallel disassembly sequence planning (DSP) model and solves it using a discrete random forest algorithm (DRFA). Firstly, based on the actual operation environment of the hydropower station, a disassembly model is constructed by considering the disassembly tools, disassembly direction, part distance, and working space. Secondly, the DRFA is proposed. Meanwhile, a sub-tree fitting strategy and a sub-tree iteration mechanism are introduced to increase the probability of finding the optimal sequence. Thirdly, four sets of experiments were conducted using three devices as examples, and compared with the Simplified Discrete Gravity Search Algorithm (SDGSA), Genetic Algorithm (GA), Simplified Cluster Optimization (SSO), and Discrete Whale Optimization Algorithm (DWOA). The experimental results showed that in the four experiments, the proportion of DRFA finding the optimal sequence reached 100%, 100%, 87%, and 74%, respectively, which was higher than the comparison algorithms. In addition, the stability of DRFA and the quality of the obtained sequences were significantly higher than those of other algorithms. Finally, the proposed method is used to automatically generate the maintenance process of the main servomotor in the virtual simulation system, which reflects the applicability of the method.