Prediction Model for Fluid Accumulation of Long distance Pipeline Based on Improved Particle Swarm Optimization Algorithm and Random Forest

Xinhai Zhang

doi:10.54097/vm9cy218

Authors

Xinhai Zhang

DOI:

https://doi.org/10.54097/vm9cy218

Keywords:

Random Forest, Particle Swarm Optimization Algorithm, Volume of Fluid Accumulation, Long Distance Pipeline

Abstract

At present, as most gas wells gradually enter the middle and late stages of production, the proportion of wet gas has greatly increased, leading to increasingly serious problems of pipeline fluid accumulation. In order to ensure the efficiency and economy of pipeline transportation, this paper optimizes the number and depth of trees in the Random Forest (RF) model for predicting volume of fluid accumulation of pipelines based on an improved Particle Swarm Optimization (PSO) algorithm. Firstly, 45330 sets of on-site data collected from a certain gas gathering station were used as the dataset. Among them, 45231 sets of data were used as the training set for model training, and the remaining 99 sets of data were used as the testing set for validation. At the same time, quantitative analysis was conducted on the influencing factors of pipeline volume of fluid accumulation based on SHAP (Shapley Additive ex Plans) to determine the five characteristics of oil pressure, daily gas production, casing pressure, cumulative gas production, and gas water ratio. Finally, the optimized random forest model is applied to predict the volume of fluid accumulation of the gathering and transportation pipeline. The prediction results show that the MAE of the optimized random forest model based on the improved PSO algorithm reached 0.003168, the MSE reached 0.0004089, and the MAPE reached 1.045%. The accuracy, universality, and feasibility of the optimized model in this paper have been verified, and it can provide some reference for the design and maintenance of natural gas pipelines.

Downloads

Download data is not yet available.

References

[1] Yuan Jingling, Zhong Luo, Li Xiaoyan Research and Development of Grey Neural Networks [J]. Journal of Wuhan University of Technology, 2009, 31 (03): 91-93.

[2] Luan Guohua, He Shunli, Shu Shaoyi, etc Application of artificial neural network method to predict gas well fluid accumulation [J]. Block oil and gas fields,2010, 17(05): 575-578.

[3] CHEN Min, LIU Jun. The improvement of BP neural network and its applications[J] Journal of Hunan University of Arts and Science (Natural Science Edition), 2005, 17( 2) : 68-71.

[4] OSMAN E S. Artificial neural network models for identifying flow regimes and predicting liquid holdup in horizontal multiphase flow[J]. SPE Production & Facilities, 2004, 19 (1) : 29-34.

[5] Zhang Yang, Ke Yong, Wang Luxin, etc Multi factor comprehensive evaluation method based on fuzzy mathematics and grey relational analysis for optimizing well completion methods [J]. Liaoning Chemical Industry, 2012, 41 (10): 1075-1076+1093.

[6] Zhao Binjuan, Liu Qi, Fu Yanxia, etc Prediction of Vortex Flow at the Boundary of Centrifugal Pump Impeller Based on Artificial Neural Network [J]. Journal of Irrigation and Drainage Machinery Engineering, 2020, 38 (02): 127-132.

[7] Fan Zhenyu BP neural network model and learning algorithm [J]. Software Guide, 2011, 10 (7): 66-68. FAN Zhenyu. BP neural network model and learning algorithm [J]. Software Guide, 2011, 10 (7): 66-68.

[8] Chongchong Qi, Xiaolin Tang. Slope stability prediction using integrated metaheuristic and machine learning approaches: A comparative study[J]. Computers & Industrial Engineering, 2018, 118.

[9] SHIPPEN M E, SCOTT S L. A neural network model forprediction of liquid holdup in two-phase horizontal flow [C]. SPE Annual Technical Conference and Exhibition, 2002.

[10] Tian Rui, Meng Haidong, Chen Shijiang, etc RF-AHP Cloud Model for Predicting Rockburst Intensity Grading [J]. Chinese Journal of Safety Science, 2020, 30 (07): 166-172.

[11] Breiman L. Random Forests[J]. Mach Learn, 2001, 45(01): 5-32.

[12] Teacher Dong, Huang Philosophy Analysis of Random Forest Theory [J]. Ensemble Technology, 2013, 2 (01): 1-7.

[13] Guo Chao, Yang Yuxuan, Hu Ronglei, etc WSN Coverage Optimization Based on Particle Swarm Optimization [J]. Research on Computer Applications,2020,37(4):1170-1173+ 1182.

[14] Zhou Lijun, Peng Wei, Zou Fang, etc Adaptive Mutation Particle Swarm Optimization Algorithm [J]. Computer Engineering and Applications, 2016, 52 (7): 50-55+149.

[15] Cheng Jiawei Prediction Model and Application of Grain Storage Temperature Based on BP Neural Network [D]. An Xue, 2021. DOI: 10.26917/d.cnki.ganhu.2021.000357.

[16] Yan Qunmin, Dong Xinzhou, Mu Jiahao, Ma Yongxiang. 2022. Optimization of Energy Storage Configuration in Active Distribution Networks Based on Improved Multi Objective Particle Swarm Optimization Algorithm [J]. Power System Protection and Control, 50 (10): 11-19.

[17] Chen Xinguo, Leng Xulin, An Yunpeng, etc Hydraulic Prediction Model for Gas Pipeline Based on Deep Learning Structural Network [J]. Oil and Gas Field Surface Engineering, 2018, 37 (08): 52-57.

[18] Zhang Yu Research on wax deposition prediction algorithm for waxy crude oil pipelines based on machine learning [D]. Northeast Petroleum University, 2019: 36-65.