Data-driven approaches for modelling of sub-critical coal-fired boiler

Due to increasing shares of renewable electricity sources in the grid, thermal power plants need operate a more flexible manner future. This will involve frequent startups, shutdowns, and load changes. A central part plant analysed this study is coal-fired boiler. In previous study, first-principle model sub-critical coalfired boiler has been developed validated with operational data from Polish plant. Based on model, work aims develop computationally efficient sufficiently accurate data-driven that easy implement new software. selection multi-output algorithms was first compared using nonoptimised parameters, very few adaptations set. Then, each algorithm had undergone three different optimisation routines tune hyper-parameters. The results models were optimised ones, then reference average Mean Absolute Percentage Error as score. methods used comprise six base learners ensemble methods. based Powell conjugate direction method, Bayesian evolutionary algorithm. All shown lower percentage error than principle resulted improved prediction capacity for every learner, but not method-based algorithms.