•  
  •  
 

Document Type

Original Study

Subject Areas

Electrical Engineering

Keywords

Battery Energy Storage System (BESS), Gaussian Process Regression (GPR), Load Prediction, Peak Shaving, Oil Drilling Microgrids, Generator Efficiency, Depth of Discharge (DOD), C-Rate Validation, Specific Fuel Consumption (SFC).

Abstract

Diesel-based microgrids serving oil drilling operations are subject to sharp and unpredictable load fluctuations that lead to generator underloading, elevating specific fuel consumption (SFC), and reducing system reliability. Conventional battery energy storage system (BESS) sizing methods typically depend on long-term field measurements and static load assumptions, which fail to capture the real transient nature of drilling processes. This paper develops a framework for BESS sizing that directly integrates a measurement-based machine learning (ML) load prediction model. The adopted model, built using the exponential Gaussian process regression (GPR) technique, has been trained and validated on six months of high-resolution field measurements collected from an operating oil drilling rig. These predicted load sequences are then utilized to calculate instantaneous BESS power requirements, cumulative energy capacity, and discharge-rate validation under real operational constraints. The methodology further incorporates generator switching limits, depth-Of-discharge (DOD) margins, battery C-Factor and degradation safety factors to ensure practical deployment feasibility. Simulation results have shown that the highest predicted load peak of 1.546 MW, lasting 7 minutes, required a BESS capacity of 162 kWh. Incorporating operational constraints such as minimum generator switching intervals, an 80% DOD limit, and a 15% aging safety margin, the final BESS energy requirement is determined to be 232 kWh. The proposed approach eliminates the need for long-term field data collection while providing a physically feasible, prediction-driven tool for sizing energy storage systems in dynamic industrial microgrids.

Download

Share

COinS