Research On Data Driven Based Energy Consumption Diagnosis Method And Analysis For Coal-fired Power Plants

At present,the energy industry is in a transition period.China’s resource endowment determines that coal-fired power generation will still occupy a key position in the energy structure for a period of time.At a time when the linkage between big data,artificial intelligence and other new technologies and traditional industries is becoming mature,it will become the booster for the coal power industry to tap the energy saving potential and complete the transformation of flexible power generation.Based on data-driven research ideas,the diagnosis and analysis of energy consumption of coal-fired power plants were carried out in the dissertation.It does not need to over-explore the mechanism,but makes use of intelligent algorithms to fully mine the effective information in the historical operation data to promote the research.First of all,the collected data were initially processed to improve the data quality,and a random forest combined with analytic hierarchy process method was proposed to screen all the variables,retaining 20 variables that have important effects on power generation energy consumption to improve the modeling efficiency.Secondly,in view of the non-linearity and strong interaction of related variables in the energy consumption of coal-fired power plants,support vector regression was adopted to conduct data-driven modeling for coal-fired power plants,and the relationship between each variable and power generation energy consumption was established.Based on the model,the model parameters were optimized by Grey Wolf optimization algorithm,and the training efficiency of the model was improved.Finally,according to the established data model and E-Fast method(Extended Fourier Amplitude Sensitivity Test),the global energy consumption Sensitivity analysis was conducted for each load operating segment of the unit.On the basis of considering the interaction between variables,the influence degree of each variable on power generation energy consumption is obtained.The analysis results in the dissertation show that maintaining the stability of operating load in all load ranges is the most important for energy saving,and reasonable control of exhaust steam flow rate,main steam flow rate and regulating steam flow rate is also very important for energy saving.In addition to paying attention to the above variables,through the specific analysis of each load range,it can be concluded that the unit should closely monitor the seventh stage extraction temperature to maintain it within a certain range when operating in the 60-90%load range;When the load range is 70-90%,the change of the left condenser pressure has a great influence on the energy consumption.When running in the 90-100%load range,the supply of circulating water should be guaranteed to make the unit at a higher energy saving level.In the dissertation,based on the historical operation data of coal-fired power plants and the intelligent algorithm,a data model is established,and the global energy consumption sensitivity analysis is carried out in each load interval of the unit.According to the analysis results,the adjustment strategy is proposed,which provides decision support for the operators to make energy saving plans.