Process Simulation And Process Parameter Optimization For Air Separation Unit Based On Data Mining

With the continuous development of intelligent manufacturing technology,big data technology is all the rage.For air separation devices,the use of data mining technology for process simulation and parameter optimization is of great significance.The purpose of this research is to analyze the data and conduct field research on the process simulation and parameter optimization methods of air separation plants at home and abroad.Establish a suitable simulation model to guide the production of the air separation device through data mining technology by using the historical data of the air separation device.It provides a reliable theoretical and practical basis for process simulation in the air separation industry in the future.A process simulation modeling method for air separation plant based on artificial neural network technology is proposed,and the process parameters are optimized by using the BAS algorithm.And use the model to obtain the best process parameters and guide the selection of processing parameters during production.Avoid the blindness of process parameter selection,improve the efficiency of the production process,reduce production costs,and reduce energy consumption.In this paper,the air separation device is used as the object,and the process simulation and parameter optimization of the air separation device are based on historical data.main tasks as follows:(1)Breaking the traditional simulation model of air separation unit based on mathematical model.For the first time,the introduction of big data analysis and BP neural network into air separation unit process simulation has opened up new ideas for air separation unit simulation.The BP neural network is optimized by using the BAS algorithm to provide a new method for BP neural network optimization.(2)The historical data of 20 parameters are collected,and the missing values and outliers are processed.The normalization of the data was carried out.Through the correlation analysis,nine parameters such as the inlet temperature of the air compressor were determined as the input parameters of the oxygen production prediction model.(3)The three-layer BP neural network of 9-25-1 is constructed.The oxygen production is predicted based on historical data.The average relative error of the trained neural network is 0.0109.The average relative error is optimized by the BAS algorithm.Dropped to 0.0038,which is reduced.65.1%.(4)Through the BP neural network,the energy consumption of the device is predicted based on historical data.The average relative error of the trained neural network is 0.0108.After optimization by the BAS algorithm,the average relative error is reduced to-0.005 and 0.025.With the goal of minimizing energy consumption,the process parameters were optimized,and the optimal values of the corresponding parameters were obtained,which provided a reference for energy saving and consumption reduction of the device.