{"title":"Machine-learning-based hydraulic fracturing flowback forecasting","authors":"Jinyuan Guo, Weisi Guo, Lixia Kang, Xiaowei Zhang, Jinliang Gao, Yuyang Liu, Ji Liu, Haiqing Yu","doi":"10.1115/1.4056993","DOIUrl":null,"url":null,"abstract":"Hydraulic fracturing is an indispensable procedure to the economic development of shale gas. The flowback of the hydraulic fracturing fluid is one of the most important parameters recorded after shale gas wells are put into production. Generally, the flowback ratio is used as the flowback indicator. The flowback ratio has a great influence on shale gas production. However, the flowback ratio is subjected to various affecting factors with their correlativity unclear. Based on a large amount of original geological, engineering, and dynamic data acquired from 373 hydraulically-fractured horizontal wells in the Weiyuan Shale Gas Field, the flowback characteristics were systematically studied based on machine learning. Based on the data analysis and random forest forecasting, a new indicator, single-cluster flowback ratio, was proposed, which can more effectively reflect the inherent relationship between flowback fluid volume and influencing factors. The results of training random forests show that this indicator has better learnability and predictability. A good linear relationship exists between single-cluster flowback ratios in different production stages. Accordingly, the 30-day single-cluster flowback ratio can be used to predict the 90-day and 180-day single-cluster flowback ratios. The main controlling factors of production and flowback ratio were also systematically analyzed. It is found that the main controlling factors of the flowback ratio include the number of fracturing clusters, the total amount of sand and number of fracturing stages. This study can provide a fundamental reference for analyzing the hydraulically fracturing fluid flowback for shale gas reservoirs.","PeriodicalId":15676,"journal":{"name":"Journal of Energy Resources Technology-transactions of The Asme","volume":" ","pages":""},"PeriodicalIF":2.6000,"publicationDate":"2023-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Energy Resources Technology-transactions of The Asme","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1115/1.4056993","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENERGY & FUELS","Score":null,"Total":0}
引用次数: 0
Abstract
Hydraulic fracturing is an indispensable procedure to the economic development of shale gas. The flowback of the hydraulic fracturing fluid is one of the most important parameters recorded after shale gas wells are put into production. Generally, the flowback ratio is used as the flowback indicator. The flowback ratio has a great influence on shale gas production. However, the flowback ratio is subjected to various affecting factors with their correlativity unclear. Based on a large amount of original geological, engineering, and dynamic data acquired from 373 hydraulically-fractured horizontal wells in the Weiyuan Shale Gas Field, the flowback characteristics were systematically studied based on machine learning. Based on the data analysis and random forest forecasting, a new indicator, single-cluster flowback ratio, was proposed, which can more effectively reflect the inherent relationship between flowback fluid volume and influencing factors. The results of training random forests show that this indicator has better learnability and predictability. A good linear relationship exists between single-cluster flowback ratios in different production stages. Accordingly, the 30-day single-cluster flowback ratio can be used to predict the 90-day and 180-day single-cluster flowback ratios. The main controlling factors of production and flowback ratio were also systematically analyzed. It is found that the main controlling factors of the flowback ratio include the number of fracturing clusters, the total amount of sand and number of fracturing stages. This study can provide a fundamental reference for analyzing the hydraulically fracturing fluid flowback for shale gas reservoirs.
期刊介绍:
Specific areas of importance including, but not limited to: Fundamentals of thermodynamics such as energy, entropy and exergy, laws of thermodynamics; Thermoeconomics; Alternative and renewable energy sources; Internal combustion engines; (Geo) thermal energy storage and conversion systems; Fundamental combustion of fuels; Energy resource recovery from biomass and solid wastes; Carbon capture; Land and offshore wells drilling; Production and reservoir engineering;, Economics of energy resource exploitation