Detection And Control Technology Research Of Powder System Optimization In Thermal Power Unit

Coal pulverizing system of thermal power unit is necessary link in the coal-fired power plant, ball mill pulverizing system is the core of the equipment. As acontrolled object, the nonlinear, strong coupling, large delay, large inertiacharacteristics make the difficulty of optimization control for ball mill. For along time coal pulverizing system basically conservative operation in the coal-fired power plant, result in the coal pulverizing system power consumption.Detection and control the material level in ball mill is the key of the coalpulverizing system optimization, but also more difficult to achieve. We need tocontrol the material level in the most reasonable position, in order to ensure thecoal pulverizing system in normal operation, so ball mill level measurement andoptimization of control has become a problem urgently to be solved in coal-firedpower plant.In the past, the way of material level detection of ball mill is single, italmost depend on human experience or using a single sensor to get information,this cannot accurately measure coal quantity in order to reduce powerconsumption. Using the theory and method of multi-sensor fusion technology tointelligent process and control the material level of ball mill is an importantresearch direction in the future. Data fusion technology become a system oneself,it is a various and multi-level process of diverse information used to complete thetarget detection, correlation, and evaluation.This article in view of the ball mill material level detection and the researchof the problem of optimal control first analyze the structure of ball mill, runningmechanism and operating characteristics. Combining mechanism characteristicanalysis of the main factors influencing the coal mill material level, and establishthe model of the mechanism of ball mill. Further study multivariate data fusion theory. We make ball mill level measurement and control solutions, will be soundsignal of ball mill, inward and outward differential pressure and temperature andinlet negative pressure as fuse object, the CMAC(the cerebellum model jointcontroller) neural network as fusion algorithm research is discussed and a fusionof CMAC neural network simulation was carried out. The simulation resultsshow that this scheme can achieve more accurate detection of material level forball mill, reduces the power consumption of the ball mill, and thus laid thefoundation for optimal control of the coal pulverizing system.