Process Calculation And Operation Scheme Optimization Of Oil Pipeline Based On Neural Network

“Three high” crude oil transportation in cold regions requires higher temperature and pressure,resulting in a large amount of carbon emissions.In order to reduce carbon emissions,it is necessary to optimize the production and operation scheme of oil pipeline considering pipeline characteristics,oil properties,environment and other factors.Among them,the calculation accuracy of pipeline process parameters plays a key role in the success of the operation scheme.Due to many influencing factors,the traditional process calculation model has a large deviation,which has a certain impact on the rationality of the production operation scheme.With the wide application of machine learning in many industries and good results have been achieved,coupled with the popularization of digital oil pipeline,a large number of production and operation data have been accumulated,which lays the foundation for the study of oil pipeline process calculation by machine learning method.Taking an oil pipeline in Northeast China as the research object,the calculation of pipeline operation parameters,equipment energy consumption and production operation optimization are carried out based on neural network.The research results are as follows:Firstly,based on the historical operation data of pipeline,the BP temperature drop calculation model is established by Python language.Compared with the traditional fitting temperature drop model,the calculation accuracy is improved.In order to speed up the convergence speed of the model and optimize the model structure,the GA-BP temperature drop model is established by combining the advantages of BP neural network and genetic algorithm(GA).The measured data are selected for calculation and model comparison.The results show that the accuracy of GA-BP temperature drop calculation model is the highest,and the average absolute error is 0.151°C.Compared with the traditional model,the average relative error is reduced by 7.64%,and the calculation accuracy of the model is significantly improved.Secondly,the GA-BP pressure drop calculation model is trained for pipeline pressure drop calculation.Compared with the traditional pressure drop calculation model,the accuracy is slightly improved.In order to further improve the calculation accuracy,the particle swarm optimization(PSO)algorithm combined with adaptive BP neural network is used to establish an improved PSO-BP pressure drop calculation model.The traditional,GA-BP and improved PSO-BP models are selected to calculate the measured data,and the average absolute error of the improved PSO-BP model is the smallest,which is 0.02 MPa.Compared with the traditional pressure drop model,the average relative error is reduced by 13.16%.Thirdly,the principal component analysis method is used to analyze the influencing factors of energy consumption of heating furnace and oil pump,and the calculation models of fuel consumption of heating furnace and power consumption of oil pump based on improved PSO-BP are established respectively.The verification results of measured data show that compared with the traditional calculation model of fuel consumption,the average relative errors of the improved PSO-BP calculation model are reduced by 3.81 % and 3.5 %,respectively.Finally,based on the neural network process calculation model,the mathematical model of production and operation scheme optimization is established with the minimum carbon emission as the objective function and the pipeline oil transportation task,operation parameters and equipment characteristics as constraints.The genetic algorithm is used to solve the problem.Under typical working conditions,the average carbon emission per day after optimization is reduced by 9.75 %.On this basis,considering the equipment reliability curve and the average fault-free operation time,the annual pipeline operation and maintenance scheme is given,which can reduce CO2 emissions by 1095.4 tons per year while ensuring the safe and reasonable operation of task traffic and equipment.Compared with the original scheme,it reduces carbon emissions by 8.22%,and the emission reduction effect is remarkable.