Analysis Of Key Parameters Of Cyclone Separator Based On RBF Neural Network

The cyclone separator is widely used in the field of oil and gas.The performance of the cyclone separator is mainly determined by the separation efficiency and production capacity.Most of the current research focuses on the separation efficiency of the cyclone separator,the research on the production capacity is very few.In order to meet the needs of reality,the researchers chose some models to calculating the production capacity of the cyclone separator through theoretical and experimental ways.But the obtained models isn't universal and can't describe the internal flow field of the cyclone separator accurately.With the improvement of gas mining technology and construction conditions,the results obtained by the traditional calculation models of the production capacity have large error with the reality.It brought great difficulties to cyclone selection in actual production.Therefore,it is essential to establish a new model for calculating the production capacity of cyclone separator to calculate the actual value.The production capacity of the cyclone separator is affected by the working conditions and the size of the cyclone separator,which has many parameters and complex relationships.After a comparative analysis,this paper attempts to use radial basis function(RBF)neural network to establish a calculation model for the production capacity of the cyclone separator.The RBF neural network can analyze the relationship between various parameters directly,and the response surface method can be used to calculate the quadratic polynomial regression equation about the production capacity of the calculation model.The response surface method can also reflect the impact of the key parameters of the cyclone separator on the production capacity and build an optimization model.The research work in this paper has practical significance for improving the quality of oil and gas mining.The main research work is as follows:1)Simulating and analyzing the internal flow field of the cyclone separator based on the theory of computational fluid mechanics to obtain the preliminary qualitative relationship between the structure size and production capacity of the cyclone separator;2)Training the RBF neural network by using the production capacity data of 220 sets of the cyclone separator with different working condition parameters and structural size parameters and then establishing a calculation model of the cyclone separator production capacity.Comparing the established calculation model with the theoretical model and the BP neural network calculation model to analyze the advantages of the RBF neural network in calculating the production capacity of the cyclone separator;3)Using the response surface analysis to analyze the obtained RBF neural network calculation model and establish a quadratic regression equation of the parameters and production capacity of the cyclone separator.Optimizing the design of the cyclone separator based on the obtained equations to obtain the structural parameters and working condition parameters of the cyclone separator under the optimal conditions.Simulating and analyzing the cyclone separator before and after optimization and then comparing their performance.The results show that the error between the calculation result of the RBF neural network model's production capacity of the cyclone separator and the actual value is very small,and the determination coefficient reaches 0.99993.Compared with the calculation results of the cyclone separator production capacity formula given in the current literature,the RBF neural network calculation results are closer to the actual production of the cyclone separator production capacity value.Compared with the BP neural network,the RBF neural network has only increased by 0.01% in the determination coefficient value,but the mean square error has been reduced by 95.7%,and the training time is shorter,which shows that RBF can get accurate calculation results more quickly.The response surface analysis of the RBF calculation model can get the quadratic regression equation about the production capacity of the cyclone separator.Then proposing a set of optimization parameters and comparing the cyclone separator before and after optimization.Studies have shown that the optimized separator can reduce the pressure drop while keeping the production capacity unchanged,and reduce the degree of wear of the cyclone separator,at the same time,extend the service life of the cyclone separator,and also improve the quality of oil and gas mining.