| The 64.5km Quzi-Tiexi oil pipeline has a designed transport capacity of 1.5 million tons per year.The pipeline was put into operation in 2011 and then gradually entered the stage of low-transport operation.In order to ensure the safe and stable operation of the pipeline,the thermal and hydraulic characteristics analysis and flow safety guarantee research are carried out for the pipeline,and the economic operation scheme of Quxi Pipeline is optimized.First of all,understand the basic parameters of the pipeline,obtain the field operation data,and collect the crude oil of Quzi Station and Tiexi Station in the laboratory to measure its basic physical properties,such as density,freezing point,viscosity,wax analysis characteristic,specific heat capacity,etc.the kernel principal component analysis method was used to analyze the weight of factors affecting the wax deposition rate,and it was concluded that the viscosity,wall shear stress,flow rate,radial temperature gradient and wax crystal solubility coefficient had great influence.The accuracy of the results was verified by grey correlation method.Combined with the long and short term memory network algorithm,the divided training set was deeply learned and the better wax deposition rate model was iterated constantly.The trained model was substituted into the test set to predict the wax deposition rate and then compared with the actual value to judge its accuracy.The wax deposition rate and thickness along the pipeline were calculated by programming,and the wax deposition conditions along the pipeline under different outbound temperature and different delivery volume were obtained.Secondly,After analyzing the experimental data,the regression model of the total heat transfer coefficient K value and the calculation formula of the hydraulic friction were established,and the thermodynamic hydraulic research of the pipeline was carried out.A new temperature drop and pressure drop model was established by inputting the factors that have a great influence on the temperature drop and pressure drop as the input layer by RBF neural network algorithm.Compared with the traditional model,the average absolute percentage error of the temperature drop model is reduced from 4.54% to 2.33%,and the average absolute percentage error of the pressure drop model is reduced from 6.68% to 3.80%.And the critical safe transmission is determined by the thermal minimum and hydraulic minimum.Finally,the safe and economical pigging cycle model was developed,and the calculation program of the paraffin cleaning cycle was prepared considering the power cost,thermal cost and pigging cost.The operation cost and the best pigging cycle under different working conditions were obtained by inputting field data,and the optimal operation scheme of Quxi oil pipeline was given. |