US-12626204-B2 - Forecasting energy demand and CO2 emissions for a gas processing plant integrated with power generation facilities

US12626204B2US 12626204 B2US12626204 B2US 12626204B2US-12626204-B2

Abstract

Systems and methods include a computer-implemented method for determining energy efficiency and emissions. Real-time energy stream data for a gas processing plant is received. Correlations of energy streams between demand-side energy demands for meeting production requirements of the gas processing plant and fuel requirements for supply-side equipment of the gas processing plant are generated. Machine learning algorithms are trained using the correlations of the energy streams to identify relationships among dependent variables and independent variables of demand-side energy consumers and supply-side energy sources. Forecasted values of total energy consumption of the gas processing plant are determined using the machine learning algorithms and real-time energy stream data. Forecasting models are re-trained using new data if an error between the forecasted values and actual energy demand exceeds a threshold. An energy intensity (EI) for the gas processing plant is generated. CO 2 emissions for the gas processing plant are determined.

Inventors

Mussa Hadi Alamri
MUHAMMAD ABBAS
Ali H. Al-Qahtani

Assignees

SAUDI ARABIAN OIL COMPANY

Dates

Publication Date: 20260512
Application Date: 20221205

Claims (16)

1 . A gas processing plant, comprising a computer system configured to operate the gas processing plant, the computer system configured to perform operations comprising: receiving real-time energy stream data for the gas processing plant; generating, using historical energy and production stream data, correlations of energy streams between demand-side energy demands for meeting production requirements of the gas processing plant and fuel requirements for supply-side equipment of the gas processing plant; training, using the correlations of the energy streams, machine learning algorithms to identify relationships among dependent variables and independent variables of demand-side energy consumers and supply-side energy sources with an objective to achieve a best fit forecasted value of a dependent variable with minimum error compared to actual values associated with supply and demand; determining, using the machine learning algorithms and the real-time energy stream data, forecasted values of total energy consumption of the gas processing plant; re-training forecasting models of the machine learning algorithms using new data if an error between the forecasted values and actual energy demand exceeds a threshold; generating, using the forecasted values of the total energy consumption of the gas processing plant, an energy intensity (EI) for the gas processing plant, wherein the EI is defined as a ratio of net energy consumption to total production and indicates an assessment of energy efficiency of the gas processing plant; determining, based on the forecasted values of total energy consumption, a predicted energy consumption by the gas processing plant; and determining, using the predicted energy consumption by the gas processing plant and by applying emission factors for fuel gas consumed by the gas processing plant, forecasted CO 2 emissions for the gas processing plant.
2 . The gas processing plant of claim 1 , wherein the real-time energy stream data comprises: feed streams for wet gas from gas fields and gas oil separation plants (GOSPs); product streams for sales gas (SG), hydrocarbon condensate (HC Cond), sulfur (Sul) from sulfur recovery units (SRUs); energy streams from energy sources comprising power (P), steam(S), fuel gas for boilers (FGB), fuel gas for cogeneration plant (FGC), and fuel gas for other processes (FGP); and ambient temperatures (T).
3 . The gas processing plant of claim 2 , wherein the computer system is further configured to perform operations further comprising, based on the forecasted CO 2 emissions for the gas processing plant, changing one or more of the feed streams for wet gas from gas fields and gas oil separation plants (GOSPs); product streams for sales gas (SG), hydrocarbon condensate (HC Cond), sulfur (Sul) from sulfur recovery units (SRUs); and energy streams from energy sources comprising power (P), steam(S), fuel gas for boilers (FGB), fuel gas for cogeneration plant (FGC), and fuel gas for other processes (FGP) to change the CO 2 emissions.
4 . The gas processing plant of claim 1 , wherein the supply-side equipment comprises cogeneration plant and boilers.
5 . The gas processing plant of claim 1 , wherein the computer system is further configured to perform operations comprising: generating, using the forecasted CO 2 emissions and CO 2 emissions for the gas processing plant, a dashboard that comprises energy efficiency key process indicator (KPI) information.
6 . The gas processing plant of claim 5 , wherein the dashboard comprises an actual EI KPI reading, a predicted EI KPI reading, a computed different between the actual EI KPI reading, and the predicted EI KPI reading, a total production, and a total energy.
7 . The gas processing plant of claim 1 , wherein the computer system is further configured to perform operations comprising: performing data cleaning on the real-time energy stream data.
8 . The gas processing plant of claim 1 , further comprising: executing, using the real-time energy stream data, multiple different ML algorithms; scoring the multiple different ML algorithms based on accuracy of results; and selecting, from the multiple different ML algorithms and based on the scoring, a specific algorithm that best produces an acceptable accuracy result.
9 . The gas processing plant of claim 8 , wherein the multiple different ML algorithms include a decision tree algorithm, a forest algorithm, a gradient boosting algorithm, and a multiple linear regression algorithm.
10 . A computer system for a gas processing plant, the computer system comprising a non-transitory, computer-readable medium storing one or more instructions executable by the computer system to perform operations of the gas processing plant comprising: receiving real-time energy stream data for the gas processing plant; generating, using historical energy and production stream data, correlations of energy streams between demand-side energy demands for meeting production requirements of the gas processing plant and fuel requirements for supply-side equipment of the gas processing plant; training, using the correlations of the energy streams, machine learning algorithms to identify relationships among dependent variables and independent variables of demand-side energy consumers and supply-side energy sources with an objective to achieve a best fit forecasted value of a dependent variable with minimum error compared to actual values associated with supply and demand; determining, using the machine learning algorithms and the real-time energy stream data, forecasted values of total energy consumption of the gas processing plant; re-training forecasting models of the machine learning algorithms using new data if an error between the forecasted values and actual energy demand exceeds a threshold; generating, using the forecasted values of the total energy consumption of the gas processing plant, an energy intensity (EI) for the gas processing plant, wherein the EI is defined as a ratio of net energy consumption to total production and indicates an assessment of energy efficiency of the gas processing plant; determining, based on the forecasted values of total energy consumption, a predicted energy consumption by the gas processing plant; and determining, using the predicted energy consumption by the gas processing plant and by applying emission factors for fuel gas consumed by the gas processing plant, forecasted CO 2 emissions for the gas processing plant.
11 . The computer system of claim 10 , wherein the real-time energy stream data comprises: feed streams for wet gas from gas fields and gas oil separation plants (GOSPs); product streams for sales gas (SG), hydrocarbon condensate (HC Cond), sulfur (Sul) from sulfur recovery units (SRUs); energy streams from energy sources comprising power (P), steam(S), fuel gas for boilers (FGB), fuel gas for cogeneration plant (FGC), and fuel gas for other processes (FGP); and ambient temperatures (T).
12 . The computer system of claim 10 , wherein supply-side equipment comprises cogeneration plant and boilers.
13 . The computer system of claim 10 , the operations further comprising: generating, using the CO 2 emissions for the gas processing plant, a dashboard that comprises energy efficiency key process indicator (KPI) information.
14 . The computer system of claim 13 , wherein the dashboard comprises an actual EI KPI reading, a predicted EI KPI reading, a computed different between the actual EI KPI reading, and the predicted EI KPI reading, a total production, and a total energy.
15 . The computer system of claim 10 , the operations further comprising: performing data cleaning on the real-time energy stream data.
16 . The computer system of claim 10 , the operations further comprising: executing, using the real-time energy stream data, multiple different ML algorithms; scoring the multiple different ML algorithms based on accuracy of results; and selecting, from the multiple different ML algorithms and based on the scoring, a specific algorithm that best produces an acceptable accuracy result.

Description

TECHNICAL FIELD The present disclosure applies to energy requirements and emissions of a facility. BACKGROUND A facility, such as a gas or oil facility, includes supply and demand sides of energy streams. Energy consumption demand and carbon dioxide (CO2) emissions are affected by raw gas feeds in a gas plant, including process units configuration, final products' specifications, and ambient conditions. SUMMARY The present disclosure describes techniques that can be used for forecasting energy requirements for, and emissions from, a facility. In some implementations, a computer-implemented method includes the following. Real-time energy stream data for a gas processing plant is received. Correlations of energy streams between demand-side energy demands for meeting production requirements of the gas processing plant and fuel requirements for supply-side equipment of the gas processing plant are generated using historical energy and production stream data. Machine learning algorithms are trained using the correlations of the energy streams to identify relationships among dependent variables and independent variables of demand-side energy consumers and supply-side energy sources with an objective to achieve best fit forecasted value of a dependent variable with minimum error compared to actual values associated with supply and demand. Forecasted values of total energy consumption of the gas processing plant are determined using the machine learning algorithms and the real-time energy stream data. Forecasting models of the machine learning algorithms are re-trained using new data if an error between the forecasted values and actual energy demand exceeds a threshold. An energy intensity (EI) for the gas processing plant is generated using the forecasted values of the total energy consumption of the gas processing plant. The EI is an assessment of energy efficiency of the gas processing plant. CO2 emissions for the gas processing plant are determined using the predicted energy consumption by the gas processing plant and by applying emission factors for fuel gas consumed by the gas processing plant. Techniques of the present disclosure are implementable using a computer-implemented method; a non-transitory, computer-readable medium storing computer-readable instructions to perform the computer-implemented method; and a computer-implemented system including a computer memory interoperably coupled with a hardware processor configured to perform the computer-implemented method, the instructions stored on the non-transitory, computer-readable medium. The use of particular implementations of the subject matter described in this specification can realize or result in one or more of the following advantages. Techniques of the present disclosure can overcome the technical problem of forecasting energy consumption demand for a gas processing facility, especially a facility that has a cogeneration (cogen) facility (e.g., producing electric power and steam) as part of the operational boundary. With minor adjustments, the techniques can be applied to all types of gas processing plants. The techniques can lead to proactive monitoring and optimization of energy efficiency performance of a gas plant, and establishing challenging energy performance targets based on future production plans. The details of one or more implementations of the subject matter of this specification are set forth in the Detailed Description, the accompanying drawings, and the claims. Other features, aspects, and advantages of the subject matter will become apparent from the Detailed Description, the claims, and the accompanying drawings. DESCRIPTION OF DRAWINGS FIG. 1 is a diagram showing an example of system of supply and demand sides energy streams, according to some implementations of the present disclosure. FIG. 2 illustrates an example of a data analysis cycle applicable to each node represented in FIG. 1, according to some implementations of the present disclosure. FIG. 3 is a diagram showing an example architecture of typical data models, according to some implementations of the present disclosure. FIG. 4 is a screenshot showing an example of an energy demand forecasting screen, according to some implementations of the present disclosure. FIG. 5 is a screenshot showing an example of an energy intensity KPI daily monitor dashboard, according to some implementations of the present disclosure. FIG. 6 is a flowchart of an example of a method for generating a dashboard for a gas processing plant that displays energy efficiency KPI information, according to some implementations of the present disclosure. FIG. 7 is a block diagram illustrating an example computer system used to provide computational functionalities associated with described algorithms, methods, functions, processes, flows, and procedures as described in the present disclosure, according to some implementations of the present disclosure. Like reference numbers and designations in the vari