Designing LQG controller based on neural network estimator for boiler system

Due to demands for the economical operations of power plants, performance control of a boiler-turbine unit has great importance. Besides, the multi-input multi-output (MIMO) structure of boiler systems has some challenges that make their control some problems. In this research, the control of boiler systems is performed based on the neural network algorithm. In the boiler system, the liquid (usually water) reaches its desired temperature and its output steam is employed for making electricity, locomotive movement, environmental heating, health domain and etc. Since the water level in the tank has a great effect on the stability of the boiler, controlling dram water level significantly affects the system's performance. In this paper, controlling the water level of the boiler system is applied utilizing LQG and a neural network algorithm. For the system variables estimation, neural networks are employed instead of system conditions, due to their high ability in system identification in various conditions. The simulation results of the proposed method are compared with the Kalman-filter-based LQG controller.