Theoretical Research On Data Mining Based On Energy Saving And Consumption Diagnosis For Thermal Power Units

The cleaning use of coal is the key idea of China "13th Five-Year" energy plan.Electric coal accounts for more than 60%in total coal consumption.Power plants were usually deviated from the best operation conditions,resulting from increasing energy consumption,improper operation adjustment and changeable environment.Therefore,it is of great significance to do energy consumption analysis and energy saving diagnosis for units.Coal-fired units are with high heat flow density,the large scale system and operation parameters fluctuation range.energy conversion are in different levels in the system,the energy loss is highly nonlinear,and there is a strong dependence between energy consumption,external environment,load and other factors.In view of the above characteristics,based on the big data mining technology,the energy consumption characteristics of thermal power units were analyzed and studied and a set of relatively perfect energy consumption analysis and energy saving diagnosis system for large coal-fired units was established in this paper.Firstly,according to the complex operating characteristics of coal-fired power generation,a new combination intelligent algorithms between generalized neural network(GRNN)and the mean impact value(MIV)was proposed,which screened out the key energy consumption characteristics of coal-fired units determining the power supply coal consumption of the unit.Secondly,the improved k-means clustering algorithm was used to determine the benchmark values of energy consumption under different loads.Finally,the theory and method of energy consumption diagnosis for coal-fired units were introduced in detail.Based on the classification of energy consumption factors and establishment of key energy consumption variable selection models,the energy consumption characteristics and distribution diagnosis models of large coal-fired units were established.The energy consumption distribution of a 660MW supercritical coal-fired unit was analyzed and diagnosed.The results show that the model can accurately diagnose the energy loss of each component of the unit which can provide a reference for the optimization of the power plant.