{"title":"利用机器学习预测原油提取和分离的能源需求和二氧化碳排放","authors":"Muhammad Abbas, Omar Naeem","doi":"10.2523/iptc-22801-ms","DOIUrl":null,"url":null,"abstract":"\n With the drop of oil reservoirs’ natural pressure, and injection of higher amounts of water, predicting energy consumption required to extract multiphase hydrocarbon product, and separate it into crude, gas, and water has become a challenging and more dynamic problem. This paper discusses a detailed technique to forecast energy demand for water injection and Gas-Oil Separation Plant (GOSP). Key elements of the method include identifying the energy, products, and feed streams, along with other parameters impacting the energy demand. The relationships among all independent and dependent variables are identified, along with the consideration of ambient conditions and equipment operating efficiencies. Machine Learning (ML) algorithms are then applied, using available industry software, to build and improve these relationships using the historical data. The best-fit forecast models, also called champion models, are selected that provide the least variance from actual data. These models can be updated, using the software, as the new data is received and variance between predicted and actual energy increases. The forecasted energy demand is converted to CO2 emissions using the conversion factors for fuel gas and power. The forecasting results and underlying process can be converted into dashboards for visualization and utilization by the users of operating plants. The method described in the paper is novel and first of a kind for predicting energy demand and CO2 emissions for a GOSP considering increases in water cut and water-injection.","PeriodicalId":153269,"journal":{"name":"Day 2 Thu, March 02, 2023","volume":"1 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2023-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Forecasting Energy Demand and CO2 Emissions for Crude Extraction and Separation Using Machine Learning\",\"authors\":\"Muhammad Abbas, Omar Naeem\",\"doi\":\"10.2523/iptc-22801-ms\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"\\n With the drop of oil reservoirs’ natural pressure, and injection of higher amounts of water, predicting energy consumption required to extract multiphase hydrocarbon product, and separate it into crude, gas, and water has become a challenging and more dynamic problem. This paper discusses a detailed technique to forecast energy demand for water injection and Gas-Oil Separation Plant (GOSP). Key elements of the method include identifying the energy, products, and feed streams, along with other parameters impacting the energy demand. The relationships among all independent and dependent variables are identified, along with the consideration of ambient conditions and equipment operating efficiencies. Machine Learning (ML) algorithms are then applied, using available industry software, to build and improve these relationships using the historical data. The best-fit forecast models, also called champion models, are selected that provide the least variance from actual data. These models can be updated, using the software, as the new data is received and variance between predicted and actual energy increases. The forecasted energy demand is converted to CO2 emissions using the conversion factors for fuel gas and power. The forecasting results and underlying process can be converted into dashboards for visualization and utilization by the users of operating plants. The method described in the paper is novel and first of a kind for predicting energy demand and CO2 emissions for a GOSP considering increases in water cut and water-injection.\",\"PeriodicalId\":153269,\"journal\":{\"name\":\"Day 2 Thu, March 02, 2023\",\"volume\":\"1 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2023-02-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Day 2 Thu, March 02, 2023\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.2523/iptc-22801-ms\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Day 2 Thu, March 02, 2023","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.2523/iptc-22801-ms","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Forecasting Energy Demand and CO2 Emissions for Crude Extraction and Separation Using Machine Learning
With the drop of oil reservoirs’ natural pressure, and injection of higher amounts of water, predicting energy consumption required to extract multiphase hydrocarbon product, and separate it into crude, gas, and water has become a challenging and more dynamic problem. This paper discusses a detailed technique to forecast energy demand for water injection and Gas-Oil Separation Plant (GOSP). Key elements of the method include identifying the energy, products, and feed streams, along with other parameters impacting the energy demand. The relationships among all independent and dependent variables are identified, along with the consideration of ambient conditions and equipment operating efficiencies. Machine Learning (ML) algorithms are then applied, using available industry software, to build and improve these relationships using the historical data. The best-fit forecast models, also called champion models, are selected that provide the least variance from actual data. These models can be updated, using the software, as the new data is received and variance between predicted and actual energy increases. The forecasted energy demand is converted to CO2 emissions using the conversion factors for fuel gas and power. The forecasting results and underlying process can be converted into dashboards for visualization and utilization by the users of operating plants. The method described in the paper is novel and first of a kind for predicting energy demand and CO2 emissions for a GOSP considering increases in water cut and water-injection.