Study On Optimal Control Of Coal Pulverizing System Based On Support Vector Machine

Electric power industry is a basic industry in China's economic development, and promoting the healthy development of economy. It is necessary to research the optimal control of coal pulverizing system, since which is influential in reducing the energy consumption and improving the economy of power plant. The thermal power plants of small and medium-sized units in our country are used to use middle storage pulverizing system in the past, with disadvantages of large occupation area, high power consumption, high steel consumption etc. The direct blowing pulverizing system with medium speed mill is widely used in 300 MW and above units, since which is simple, convenient and flexible to operate, easy to realize automatic control, compact layout, low steel consumption, small occupation area, low power consumption etc.The mill capacity of direct blowing pulverizing system is amount of coal boiler, and the medium speed mill is a large inertia, high order, multi input multi output coupled nonlinear system. The input variables of pulverizing system are the volume of hot air, wind quantity, the quantity of coal. The main control object of direct blowing pulverizing system is coal mill. The output variables of pulverizing system are the outlet temperature of the coal mill, primary air flow, the differential pressure signal. By using the front bearing vertical vibration component of ball mill instead of the differential pressure signal to characterize the amount of stored coal in this paper, the air will not affect the storage coal quantity signal, the quantity of coal of the negative pressure signal and temperature signal can be ignored. So the mill object is decomposed into a coupling double input and double output object and a single input and single output object. The least squares support vector machine and PID control algorithm, which is a combination of support vector machine and PID control, is used for the simulation of the direct blowing pulverizing system. The simulations of coal pulverizing system by conventional PID control and the simulations of coal pulverizing system by least square support vector machine PID control are analyzed. We analyze the dynamic performance, robustness and anti – interference of simulation results by comparing the maximum overshoot, stability time and steady error. The simulations results show that the least squares support vector machine PID control has achieved a good effect on the control of time-varying, coupling and time delay of direct blowing pulverizing system.