Study On Critical Sedimentation Velocity Of Heterogeneous Slurry Pipelines

The critical nondeposit velocity is an important parameter of the slurry pipeline transportation for its complexity and particularity,it directly related to the safety and reliability of the pipeline operation and economic practicality.There are some defects in the critical velocity calculation fonnula and the network prediction model.In this paper,on the basis of theoretical analysis and the experimental study,the means of data driven analysis has been used to explore the problem.The main research work can be summarized as follows:(1)the main influence factors of the critical nondeposit velocity are the pipe diameter,particle size,concentration of slurry and solid density,etc.According to experiments,the solid density and particle size are the coupling influence variables.(2)Considering the problems of the multiple computing formula whose form and parameter have some difference and the big calculation error,the basic expression of the critical nondeposit velocity was derived on the assumption of suspended load energy consumption theory and the relationships between the critical nondeposit velocity and friction losses.And introducing the check coefficient--the relationships among the particle size,diameter and concentration after the normalization of data in the same function to check the formula,and making a comparison between the Liu Dezhong formula and the Wasp formula under the condition of tailings delivery,iron sulfide concentrate delivery,iron concentrate delivery.The basic formula of energy consumption of suitability conditions were determined as follows:the middle and low concentration slurry,and the conveying pipe diameter for small diameter pipe diameter.(3)Considering the problems of great difficulties and low accuracies in predicting the critical deposition velocity of slurry pipeline,this paper,a prediction model of the critical nondeposit velocity based on extreme learning machine(ELM)was proposed.The method uses the pipe diameter,particle size,concentration of slurry and solid density as the input factors,the value of the critical nondeposit velocity as the output factor of the model of ELM.It can effectively predict the critical velocity value.For the optimization problem of its weight and implicit yuan,a prediction model of the critical nondeposit velocity based on particle swarm optimization and extreme learning machine(ELM)was proposed.The method uses particle swarm optimization(PSO)to optimize the ELM model parameters such as the input weight matrix ?i and bias matrix bi,then the optimized ELM model is used to fit and predict large diameter slurry pipeline of critical deposition velocity.The maximum error of the measured values and predicted values is 5.73%by experimental simulation,which shows that the predicted critical deposition velocity by PSO-ELM is superior to the conventional ELM method and BP neural network model.And it is a feasible prediction method of the critical nondeposit velocity.The results above are forward-looking and full of challenges.This thesis made some breakthrough in theory formula.It makes full consideration to the relationships among the particle size,diameter and concentration in the process of introducing the checking parameter.And there are some innovations in the process of studying the prediction model.In essence,ELM feedback comes from a single hidden layer neural network.The prediction model of network architecture is more simple and effective compared with BP neural network.In addition,aiming at the shortcomings of the ELM itself exists,the PSO algorithm is used to optimizing the parameters of ELM.It puts forward a new solution for the critical deposition velocity.Finally,it makes the prospect to the follow-up work on the basis of the summery of the whole research.