Enhanced Energy Efficiency Method in Oil and Gas Industry Using Hybrid Machine Learning Approach

Abstract

In the OGI, there are multiple opportunities to reduce total production energy consumption to produce the equivalent or more unrefined petroleum and flammable gas, and improvements in efficiency may be influenced by mechanical, chemical factors, what's more, other actual boundaries. The most widely recognized strategies for further developing creation energy efficiency include replacing proficient creation hardware and further developing creation processes. In addition, energy consumption forecasts can help managers scientifically plan energy use for energy generation and transition energy use to non-peak hours. However, this remains a challenging issue due to inherent complexities and vulnerabilities. Different mixes of energy utilization and apparent performance of OGI are presented to this end; our work focuses on predicting the energy utilization of OGI. First, four different prediction models Support Vector Machines, Linear Regression, Extreme Learning Machines and Artificial Neural Networks are prepared on the preparation of informational collection and then evaluated using the test data set. The second is to work on the precision of energy utilization expectations, the combination of these four models was tested, and the value of the predictor variable was taken as the normal of the outcomes of the two models. The results show that four different model predetermines energy consumption versus accuracy, as well as artificial hybrids and models of red neurons for maximum accuracy. Furthermore, hybrid models are installed in energy management systems in the OGI to oversee energy utilization in oil fields and further develop proficiency.