Outlet Dust Concentration Prediction And Energy Consumption Optimization For Dry Electrostatic Precipitator Of Coal-Fired Unit

The dust emission from coal-fired units is the main source of dust pollutants in China’s atmosphere.With the increasing attention of the state to energy conservation and environmental protection,the dust emission standard of coal-fired units is becoming more and more strict.As a high-efficiency dust removal equipment,electrostatic precipitator is widely used in flue gas treatment of coal-fired power plants.At present,the electric precipitator of domestic coal-fired power plants generally adopts open-loop or manual control mode.In order to ensure that the outlet concentration of the electrostatic precipitator meets the environmental protection requirements under the wide load flexible operation mode of the unit,the electrostatic precipitator is usually set to operate at a high-power level,resulting in increased energy consumption.Therefore,establishing the prediction model of dust concentration at the outlet of electrostatic precipitator and optimizing the operation parameters of electrostatic precipitator(secondary current of high-frequency power supply)will help to reduce dust removal energy consumption,which is of great significance to the economic and safe operation of electrostatic precipitator of coal-fired units.This paper takes the dry electrostatic precipitator of 1000 MW ultra-supercritical coal-fired unit as the research object,and carries out the research on outlet concentration prediction and energy consumption optimization based on data-driven and artificial intelligence methods.The main research contents and results are as follows:(1)Based on the working principle of dry electrostatic precipitator,the factors affecting dust removal efficiency are analyzed.Under three typical load conditions,by changing the secondary current of electric fields at all stages,roughly adjusting the dust concentration by the first three electric fields and carefully adjusting the dust concentration by the last two electric fields,the dust concentration data at the outlet of dry electrostatic precipitator when the power supply parameters change is obtained,which lays a foundation for the research on data-driven prediction modeling of dust concentration at the outlet of dry electrostatic precipitator and energy consumption optimization.(2)Based on the concept of PID controller design,a nonlinear PID optimizer is proposed.By making full use of the information of current gradient,cumulative gradient and gradient change,the optimizer can effectively improve the speed of parameter adjustment of deep neural network,so as to speed up the training and learning process of deep neural network.Based on the deep neural network with nonlinear PID optimizer,a prediction model between the secondary current of all stages of electric field and the dust concentration at the outlet of dry electrostatic precipitator under typical load conditions is established.The results show that the prediction model can effectively reflect the relationship between outlet dust concentration and secondary current,and has high accuracy.(3)Aiming at the imbalance between exploration ability and exploitation ability in artificial bee colony algorithm,an evidence-driven artificial bee colony algorithm is proposed.The algorithm introduces the evidence fusion mechanism of evidence theory in the onlooker bee phase,and determines the search direction of the solution to be evolved by fusing the evidence provided by other solutions in its neighborhood.At the same time,the corresponding relationship between secondary current and secondary voltage of each electric field of dry electrostatic precipitator is constructed by polynomial fitting method,and the energy consumption model of secondary current and electrostatic precipitator is established.With the goal of energy consumption,under the constraints of outlet dust concentration and power supply parameters,the optimal secondary current parameter settings of all stages of electric field under typical load conditions are obtained based on evidence-driven artificial bee colony algorithm,and the optimal energy consumption under different outlet concentrations is obtained.The experimental results show that the proposed evidence-driven artificial bee colony algorithm improves the convergence speed and search accuracy,and is more effective in high-dimensional optimization problems.Moreover,by optimizing the secondary current setting value of each electric field of the dry electrostatic precipitator,the dust removal energy consumption can be effectively reduced on the premise of meeting the emission standard.