Research On Dynamic Data-driven Prediction System Of Air Input Of Bio-oxidation Tank

Bio-oxidation gold extraction technology is one of the most important gold extraction process technology,the technology investment low,little effect on the environment,and has broad prospects for development.The air input amount is an important indicator of bio-oxidation pretreatment,the number of gas into the oxidation tank directly affects the degree of chemical reaction and activity of the bacteria,thereby affecting the ratio of gold in the gold ore.At present,the air input system of most gold extraction plants is open-loop control,the amount of air input is controlled by manual valve adjustment,and the strategy of "preferably more and less" is often used to provide air,which causes a lot of energy waste and reduces economic benefits.Therefore,accurate prediction of air input volume,plays an important role in improving gold extraction rate and economic benefits.The bio-oxidation pretreatment process is complex,and there are many factors that affect the oxygen demand,including factors such as the ambient temperature in the tank,the flow speed of the slurry,etc.,which make the air input system highly non-linear,with the characteristics of uncertainty and dynamics,affecting establishment of air input mechanism model.At the same time,most of the gold extraction plants in Xinjiang are located in a harsh environment,making the measurement data susceptible to noise interference during the actual production process,and there are deviations,making the accurate prediction of the air input volume a difficult problem.In this paper,the dynamic data-driven concept is used to carry out research on the accurate prediction of air input volume,and the measured data and model prediction values are merged to improve the prediction accuracy.The main research contents are as follows:(1)A method for establishing the state space model of the air input and pressure of the oxidation tank is proposed.Analyze the principle of gas pipe flow,take the one-element pipe flow equation as the control equation,take the air input volume and pressure as the state variables,and establish the state space model of the state variables by using control theory ideas.(2)Research on data assimilation algorithms,introduce commonly used filtering algorithms,and compare the pros and cons of different algorithms.Select the most suitable algorithm,build a workflow for data assimilation of air input volume and pressure,and correct and update model predictions.(3)The concept of dynamic data-driven theory is studied,and the framework and application process of the dynamic data-driven bio-oxidation tank air input prediction system are constructed.Through computer simulation,the established oxidation tank air input prediction model was preliminarily verified.at the same time,the accuracy of the bio-oxidation tank air input prediction system driven by dynamic data was verified.