In this paper a new approach to scheduling of online and offline compressor washing has been presented.
The new method is based on a Model Predictive Control (MPC) optimisation that takes expected future fuel
and power prices into account. The MPC optimiser is built on a hybrid dynamic systems model that
describes the natural degradation of compressor efficiency and the discrete jumps in efficiency after
washing operations. The hybrid dynamic model contains discrete states for the operational mode of the
compressor and continuous states for the efficiency levels. The parameters of the efficiency model are
estimated from thermodynamic models and the extended Kalman filter based on historic plant data. The
proposed scheduling approach clearly shows the achievable economical benefits compared to the traditional
way of scheduling, especially for plants that are in continuous operation.