Long−Term Predictive Maintenance: A Study of Optimal Cleaning of Biomass Boilers

Abstract Combustion in a biomass-fired boiler causes build-up of soot, which reduces the heat transfer and decreases the efficiency of operation. In order to mitigate this natural occurrence, cleaning via soot blowing is an important maintenance action. The objective of this study is to develop long-term optimal maintenance strategies, which are model-based and specifically employ the dynamics of boiler efficiency and of anticipated heating demand, both of which are identified from empirical data. An approximate dynamic programming algorithm is set up, resulting in the optimal maintenance actions over time, so that the total operational costs of the biomass boiler plus the cleaning costs are minimized. A practical case study with real data is used to elucidate the benefits of the new approach.