The Application Of Predictive Control In The Boiler System

Temperature control plays a very important role in the industrial automaticcontrol, it’s a important control parameter of process control areas, and widely used inthe field of electricity, petroleum, chemical, paper, brewing and food industrialautomation.The boiler is a typical temperature control object because the heating media isdifferent, and can be divided into two categories: one is burning heating methods, theother is electric heating. Due to the use of safety, environment protection, energysaving, using easily and high thermal efficiency, the latter is a ideal heating equipment.Boiler control is a typical temperature control case, its dynamic characteristics areoften with large inertia, large hysteresis and non-linear characteristics. If using thetraditional PID control mode with lower control precision and unsatisfactory controleffect, it’s hard to establish a precise mathematical model. Therefore, designing aneffective to make the system with good stability, small overshoot and adjustment timeperformance will be with practical significance if taking boiler temperature control asthe research object.In this paper, we analyzed the deficiency of the traditional PID control intemperature control of the boiler, On this basis, we proposed a predictive controlalgorithm based on neural network technology, we used the neural network aspredictive control model, it took full advantage of the capabilities of the neuralnetwork to in the strong nonlinear mapping approximation and the actual optimalcontrol of receding-horizon in predictive control. Quantitative analyzes the mainparameters of the neural network predictive controller and determine the effect fromeach parameter to controller.Finally, we took the laboratory simulation boiler as theresearch object, the neural network predictive control algorithm was used to the boilertemperature control system, being compared with PID control method, a new methodwas put forward that combing the neural network predictive control with the PID, and has achieved good control effect.The test results of the laboratory simulation boiler showed that the neuralnetwork predictive control is feasible in temperature control system of the boiler. If itcan be tested and verify in engineering practice of the practical boiler, it will bewidely used in this kind of process control.