Model Identification for Industrial Coal Fired Boiler Based on Linear Parameter Varying Method

ABSTRACT: System or process identification is a mathematical modeling of systems (processes) from test or experimental data. Process models obtained from identification process can be used for process simulation, analysis, design of safety systems and control systems for the process. This paper presents the Linear Parameter Varying (LPV) modeling of 210MW Industrial Coal Fired Boiler which is commonly used in thermal power plants. LPV model is the interpolation of linear transfer function models at different operating conditions. The LPV model is adopted by considering the fact that the Industrial Coal Fired Boiler in the thermal power plant has several operating conditions due to the fluctuations in steam flow based on demands. By assuming that at every operating condition, there are changes in parameters, the LPV model is suitable for covering all operating conditions. The Industrial Coal Fired Boiler is modeled using the mass and energy balance equation in MATLAB / SIMULINK. Data needed for identification of transfer function models is taken from first principle model of the process with sampling time of 1 second. LPV model is obtained for selected physical quantities of the process. At first, linear transfer function models are identified using the data at every operation conditions using Prediction error method and then the Linear Parameter Varying model is obtained by interpolating the linear models of different operating conditions using weighting functions. The simulation result of Linear Parameter Varying model shows reasonable fit with the First principle model response. Keywords210MW Coal Fired Boiler, modeling equations, Linear Parameter Varying Model, Model Performance.