Dynamic load change operation education in air separation processes using a multivariable and nonlinear model

An operator training system (OTS) for dynamic load change operation education in air separation processes is developed. A linear parameter varying (LPV) dynamic model based on an iterative optimization strategy is identified to represent the nonlinear characteristics of the air separation process. First, the local models at typical working points are identified. Second, the weighting functions between the local models are designed and estimated. Finally, an iterative optimization strategy optimizes the local models and weighting functions. The identified LPV model is used as the core training model of the OTS. Model validation with actual data shows that the identified OTS model has reasonable simulation accuracy. By designing a model predictive control (MPC) algorithm, a multidimensional skill evaluation algorithm based on MPC operation data is proposed. The developed OTS and skill evaluation algorithm are applied in a vocational skill competition. Results show that the operation skills of different operators can be reflected and distinguished.