Research On Key Equipments Fault Diagnosis Of Coal Gangue Power Generation

China is a country with high yield ofcoal,as an annual raw coal yield of 32,4 billon tons,washing coal quantity reached one third of the raw coal yield by coal washing plant,and the percentage of poor quality coal is 15%to 20%,such as coal gangue and coal slime.There are about 200 coal gangue power plants at present.Coal gangue burning power generation is a form of power generation,safe is the base of operation and sustainable development in power plant.Directed to the intrinsically safe coal gangue power generation,The research of key equipments fault diagnosis is taken in this paper.Based on the analysis of main factors of influence of coal gangue safety power generation,the circulating fluidized bed boiler in coal ganrgue power plant is established to be one of the key equipments for safety power generation,which safe operation level is directly related to the safe production of power plant,and the correlation diagrams of the safe operation factors of the boiler are given.The correlation diagram reflects research contents of the key equipment of coal gangue power generation in this paper cearly.As the coal gangue conveyed is block form,the running condition of belt eonveyance machine has characteristics and nonlinear;directed towards the conveyaunce machine in coal gangue power plants,a method of fault recognition and prediction based on fuzzy multilevel support vector machine(FMSVM)is proposed;ten kinds of fault samples are obtained from the analysis and statisties of belt conveyance machines in coal gangue power plants,the multilevel statements and support vector machine are combined,thus constructing the FMSVM inerratic kernel funetion to identify the problems with random and nonlinear.By using random statisical data as training samples and the kernel function as a criterion,fault recognition and prediction are efficacious,according to the simulation results,and the coupling relationship of every fault support vector is found oot,Compared with the conventional SVM fault identification,this method not only overcomes the disadvantage of SVM which is sensitive to noise but also takes less time.As power plant circulating fluidized bed boiler,a method based on improved maximum likelihood estimation(IMLE)used in random noise and nonlinear system recognition is proposed,the algorithm finds minima in the implicit function of the loss function of the maximum likelihood recognition,so as to recognize the parameters of the nonlinear systems with global random noise perturbation accurately.As the simulation results of a primary air fan in Huangling coal gangue power plant show that the IMLE algorithm can accurately recognize the parameters of the nonlinear systems with colored noise perturbation;compared with the general recursive maximum likelihood method,this algorithm has less error of the recognized parameters,and has a higher convergence speed.According to the recognized parameters,the primary air fan has been transformed in the form of high voltage frequency conversion,and the operating records show that it can save more than 30%of energy.The boiler is another key equipment in safe coal gangue power generation,after analyzing of the typical fault and hidden fault of main components of boiler system,directed towards the nonlinear characteristics of the oxygen content of flue gas of the induced draft fan has many factors influence,a method based on least squares support vector machine(LS-SVM)used in flue gas oxygen content model recognition is proposed.By using nonlinear mapping,the input vector is mapped from the original space to a high dimensional Hilbert space;in the high dimension space,by using the concept of minimum of loss function,and using the SVM kernel function in the original space to replace the inner product mathematical operation in the high dimensional feature space,so as to transform the problem of nonlinear function estimation into the problem of linear function in the high dimension feature space,the crucial parameters measured of influence boiler stable operation are used to recognition and prediction,including the oxygen content of the flue gas,coal gangue flow and wind pressure of material return,the results show that this method has higher precision(the error is less than 6%o);Compared with general SVM,this method reduces the complexity of the calculation;and the least squares,SVM and adjustment of neural network RBF characteristics are united organically in this method.Based on above,the synergistic control method based on immune algorithm is established,then the method to predict and recognize the implied faults in power generating process is proposed,and the effectiveness of this method is proved by a large amount of measured simulation results.Through an analyze of typical failure cases of electrical equipment in thermal power plants,an improved CMAC(ICMAC)temporary capacitive current network identification method for the power supply system ground fault feature is proposed,simulation test and evaluation are conducted.For several emerging issues of power supply,the Plant-power supply operating value has been recalculated and set up.This would avoid the failure cases occurred again caused by value inaccurate,and also guarantee the auxiliary power system's security and reliability.