Synthetical Energy-saving Optimization For Plant-level Automatic Generation Control Of Large Power Plant In Condition Of Complex Coal

At present,the feature of most domestic power plants is that they have numerous units and the capacities and performances of these units are different with each other.Since the traditional AGC system cannot take the economy of load distribution into consideration,the plant-level AGC system is becoming a new way of load distribution.However,in some power plants,the coal quality is complicated and changeable,and the methods of coal discharge are extensive,which damage the energy conservation effect of the plant-level AGC system a lot.And also,in existing plant level load distribution strategies,attention is mostly focused on mathematical model rather than the basis of load distribution--coal consumption curve,which leads to the poor accuracy of the curve.This paper takes some domestic power plant as an example.The coal quality of this plant is very complex,ranging from anthracite,lean coal and bituminous coal.As a result,problems such as deflagration in coal mill and high carbon content of fly ash occur in the unit operation,which seriously damage the safety and economy of the unit.Since plant-level AGC is a precise method of energy conservation,when one single unit is not running in an extensive way,the energy-saving space of AGC will shrink.Experiments are carried on to learn the running condition of units and the problem of coal blending mode,and a new coal blending scheme is proposed based on the unit adaptability to different coals.Optimization of air distribution is conducted on the basis of the new coal blending scheme,and it centers on the concentration of O2 at the furnace exit and valve opening of OFA.The optimization of coal blending and air distribution can improve the unit operating situation directly.Also,the optimization can "fix" the coal and operating condition,which reduces the difficulty of forecasting the unit coal consumption online.In this power plant,the coal consumption curve of plant-level AGC is obtained from the unit performance test.The curve is fit respectively by the coal consumption of 100%?75%and 50%of the unit load.This paper analyses the effect of pulverizing systems on coal consumption,on basis of which the load section is divided and each sub interval is represented by corresponding quadratic function,the new coal consumption curve is compared with the plant-level AGC coal consumption curve which is applied to most power plants.The results show that the new curve is better than traditional curves in terms of both accuracy and trend,and reliable as the basis of plant-level AGC load distribution.Influenced by the external environment and operating condition,the coal consumption of units remains in a flux,so that the coal consumption of units should be measured online.Carbon content of fly ash is the key to the coal consumption,and in this paper,it will be measured by neural network algorithm.A new concept "weighted burn-out characteristic" is proposed aimed at the complex coal quality and the method of "pulverizing in separate mills".The input variables of the neural network model are unit load?weighted burn-out characteristic?concentration of O2 at the furnace exit and valve opening of OFA.A model of neural network prediction model is achieved after training and testing of the sample data.