This paper describes the development of an innovative tool to help operators to select a series of corrective measures and optimally manage an Oil&Gas production plant. These action are implemented to reduce energy consumption and CO2 emissions from stationary combustion, enhancing hydrocarbon production and HSE sustainability. This in-house developed tool is based on Big Data Analytics techniques combining machine-learning algorithms with advanced statistical analysis. The tool consists of a machine learning based forecasting model and a series of aggregated analytics.
The machine learning model, based on a Gradient Boosting regression algorithm, allows operators to estimate future Energy Efficiency performances by the predicting the Stationary Combustion CO2 Emission Index KPI [tCO2/kboe]. Moreover, the tool shows aggregated stats of the related influencing equipment and variables. Such information are used by plants operators to evaluate corrective actions to undertake to better manage the plant. The modelling workflow includes a series of pre-processing tasks to manage a large amount of historical field data, using operator experience in order to train and validate the selected analytics. Then, the tool is fed with field real time data, predicting future KPI value and prescribing possible actions to reduce energy consumption and GHG emission in atmosphere. Results shows that it is possible to accurately forecast the 3 hours ahead value of the Emission Index KPI and, through the actions undertaken by operators, to reduce plant's CO2 emissions.
Improve Energy Efficiency of an Upstream production asset is fundamental to reduce energy consumption cost and it is a necessary step towards a low-carbon and sustainable future, reducing CO2 emission and environmental impact. In an upstream plant, the Energy Efficiency performances and Stationary Combustion CO2 Emission can be described by a global KPI [tCO2/kboe] relating the consumed energy and the relative CO2 emission with the total production.
This paper describes the development of an innovative tool, the "Energy Efficiency Predicting and Optimizing Digital System", to help operators select a series of corrective measures and optimally manage an Oil&Gas production plant. These actions are implemented to reduce energy consumption and CO2 emissions from stationary combustion, enhancing hydrocarbon production and HSE sustainability. The tool is obtained combining Oil&Gas process operation & energy management with statistic and data science know how and it is important to underline that a holistic approach is mandatory to achieve a better design of the tool and more effective operational result. Specifically, the tool consists of a machine learning based forecasting model and a series of aggregated analytics. The machine learning model, based on a Gradient Boosting regression algorithm, allows operators to estimate future Energy Efficiency performances by the predicting the Stationary Combustion CO2 Emission Index KPI [tCO2/kboe]. Moreover, the tool shows aggregated stats of the related influencing equipment and variables. Starting by the knowledge of the Emission Index KPI in advance, exploiting the aggregated statistics provided by the tool, the plant operators and the operational engineer can check and coordinate the actions necessary for the real time management of the plant. The analysis shows how the global performance of the whole site can be described by a global KPIs and depends by the energy performance of the main plant sections, units and equipment which can be classified by more detailed and tailored variables and more specific KPIs "per equipment".
