The Use of Production Process Factors for the Short-Term Load Forecasting of the Compressor Station of the Main Gas Pipeline

The forecasting of the power consumption of the Gad Industry enterprises can allow to equalize the schedule of power consumption, optimize the operation modes of objects of own generation and storage systems of electrical energy, optimize the maintenance and repair schedule and processes of other supporting systems, minimize the cost of the use of power consumption. The object of this article is the compressor station of the main gas pipeline. The issue of forecasting the power consumption of the compressor station is more complicated than forecasting the load curve of a large system due to the less periodicity of the power consumption, its greater variance, and the need to consider factors of the production process. The novel usage of production process factors for the forecasting of the power consumption of the compressor station of the main gas pipeline was introduced. In the article, two experiments with different feature combinations were conducted. The need for consideration of factors of the production process was discussed. It was shown that using these factors reduces the mean absolute error from 105 kWh (mean absolute percentage error is 16.81%) obtained using expert prediction, up to 15.52 kWh (3.44%) obtained using the ensemble machine learning method. The application of the SHaply Additive exPlanation algorithm for the interpretation of the results of short-term load forecasting was presented.