Effect of raw material selection on equipment replacement time: case of biomass power plants

In this study, we address inventory management problems of pumps used for material transfer in biomass plants in Japan. Biomass plants should contain spare parts such that pump problems can be addressed immediately. However, because spare parts are expensive and accurate, the "just-in-time inventory" strategy is used in those plants. Furthermore, several types of raw materials are mixed in the pump and flow, and the damage to the pump differs depending on the raw materials used. Because the inside of the pump is invisible, the facility manager must dynamically predict the inside of the pump based on experience while observing changes in indirect observation data pertaining to pump failures. Herein, we propose a method to estimate pump failure probability from past equipment operation data and raw material mixture ratio as well as to manage pump inventory. This paper primarily describes the procedure for determining the pump replacement timing using raw material mixture ratio and equipment operation data.