Performance Analysis Of Hydraulic Cyclone Filter Based On Neural Network

The hydrocyclone is widely used in the industry,a cyclone filter is proposed on the basis of this combined with the characteristics of centrifugal separation and filtration separation,so that it can adapt to the more complex working environment.However,due to the complexity of the equipment within the flow,there is still no separation efficiency and pressure drop of universal method can accurately predict the hydraulic cyclone filter.The BP neural network because of its highly nonlinear are increasingly used in the engineering prediction.There are many parameters that affect the performance of the equipment,this study establishes a multivariate nonlinear prediction model based on BP neural network.The research of BP neural network showed that order of training samples of BP neural network performance are little;training sample number,BP neural network prediction performance is better;input node number,network model forecast precision is reduced;when using more than six parameters,BP neural network prediction error is more than 10% respectively on the input layer parameters and BP neural network weights and threshold value of the two parts are optimized using the genetic algorithm.Using numerical model method to qualitative analysis of flow field in the cyclone filter flow and separation mechanism,and compared the BP network prediction values and experimental results,the study found that using BP neural network prediction error results in less than 10%,to meet the industrial requirements.On the basis of cylinder length,cone angle,underflow rate,effects of inlet flow rate and inlet concentration on separation efficiency and pressure drop and determine the optimal structure and operation parameters.With the range of parameter variation and the increase of numerical value,the accuracy of prediction is higher.The research work of this paper has a certain theoretical and engineering significance for the development of new type of solid liquid separation equipment and perfect separation theory.