Study On Coal Mill Condition Monitoring And Performance Degradation

1.Based on the mechanism characteristics of the coal mill,the gray box modeling of a660 MW coal-fired coal-fired unit was carried out.An improved particle swarm optimization algorithm was used to identify multiple undetermined parameters in the model and use the field The operation data verifies the accuracy of the model,and then simulates the typical failure of the coal mill through the gray box model,and combines the statistical methods to monitor the typical failure of the coal mill,and provides guidance and suggestions for on-site maintenance work.2.An improved least squares support vector machine(DOSLPSO-LSSVM)modeling method based on dynamic opposing self-learning particle swarm optimization algorithm is proposed,and based on the analysis of key factors affecting coal consumption,gray correlation analysis is used to extract and grind The unit coal consumption is closely related to the operating characteristic parameters,and then a coal consumption unit model based on DOSLPSO-LSSVM is established.The comparison and verification of the model and the results of sensitivity analysis will provide guidance for the optimal operation of a single mill.3.Established the optimized load distribution model of the coal mill,divided the operation modes of two different coal mills,fixed the start-stop combination of the coal mill,and adopted the DOSLPSO algorithm,combined with the regularly updated coal consumption model to obtain the optimization The best load distribution plan.The optimization results show that under the condition of satisfying the output,not only can the coal mill be started and stopped frequently,but also the total power consumption of the coal mill can be reduced.4.A multi-period multi-model idea was used to establish a multi-period coal mill current model.The current model values of the time series were obtained using the input parameters under typical working conditions,and used as an indicator to measure the performance degradation of the coal mill;then The theoretical maintenance time of the coal mill is determined according to the change rule of the wear amount of the roller sleeve with time,and the corresponding maintenance current of the coal mill is obtained.Finally,a long-term and short-term memory neural network is used to establish a prediction model of the current model value to obtain the performance of the coal mill The degradation trend provides a basis for the on-site maintenance plan.5.Based on the above researches,designed and developed the coal mill condition monitoring and optimized operation software,which provides a basis for on-site optimized operation and maintenance guidance.