Real-time Optimal Control Design Of Alum Dosing Process In Waterworks

The alum dosing process is an important part of water purification in water plants.The main function of alum dosing is to remove impurities in the raw water.According to the US Federal Environmental Protection Agency’s Drinking Water Virus Removal Technical Guidelines,the lower the turbidity,the safer the water quality.Therefore,during the new crown virus epidemic,effective control of the operation of the alum dosing process is strengthened,which is conducive to the effective removal of the virus and can ensure the safety of the factory water quality.At present,most domestic water plants use traditional feedback control or manual control based on the fixed ratio of water flow in the control of alum dosing.The biggest problem lies in the lag time of at least two hours from the dosing of alum,coagulation and sedimentation to filtration.Therefore,for such a large lag process,when the raw water quality changes suddenly(such as caused by heavy rains,windy weather and upstream enterprise sewage),the existing control methods are difficult to make timely and accurate adjustments to the amount of alum dosing,which greatly affects the process of alum dosing.The turbidity removal effect may cause the turbidity of the effluent to exceed the standard for a period of time,and even endanger the drinking water safety.Thus,it is very significant to establish a control method for alum dosing that can adapt to the sudden change of raw water quality and the large hysteresis of the alum dosing process.In order to solve the above problems,this paper proposes a real-time optimization control method for the alum dosing process in a water plant.The method is based on a feedforward-feedback compound control method of random forest and second-order sliding mode control.The main work is as follows:(1)In this paper,the step response curve method is used to establish the feedback control model of alum adding process,and an online identification method for the parameters of the feedback control model of alum adding process is designed.This method can automatically reconstruct the feedback control model of alum adding process when the actual controlled process changes.(2)According to the large lag characteristics of alum addition process,a prediction model of effluent turbidity based on random forest is established,which can predict the corresponding effluent turbidity according to the current raw water quality index data and alum addition amount;A model corrector was designed to correct the output of the effluent turbidity prediction model according to the comparison between the predicted value and the actual value.(3)Based on the feedback control model of alum adding process established in(1),a special second-order sliding mode controller is designed with the deviation between the predicted value and the set value of effluent turbidity in(2)as the input.The simulation and experimental results show that the special second-order sliding mode controller is superior to the ordinary sliding mode controller based on exponential reaching law in control accuracy and response speed.(4)Aiming at the influence of multiple raw water quality index changes on alum dosage,a feed-forward prediction model of alum dosage based on random forest is designed.The model can predict the required feed-forward alum dosage in real time according to the current raw water quality index,which can effectively make up for the problem that the existing alum dosage control methods of waterworks can not deal with the raw water quality changes in a timely and accurate manner.