Energy Consumption Modeling and Forecasting for Commercial Industrial Manufacturing Applications

In the United Kingdom, industry accounts for roughly a quarter of greenhouse gas emissions. The UK Government has set ambitious net-zero targets committed to the decarbonisation of heavy industry, and the Industrial Clusters mission aims to establish the world's first net-zero carbon industrial cluster by 2040. To reduce the energy costs and carbon footprint of industry, one of the most effective solutions is the use of digital tools enabling businesses to monitor and visualize their energy consumption in real-time. Due to recent advancements in industrial digitalization, many industrial sites already generate data, including energy monitoring data, with varying degrees of digital maturity. However, a major challenge with this data is a lack of commercial tools for modeling, predicting, and visualizing industrial manufacturing energy data for efficiency improvement and emissions reduction. This paper describes efforts in a recently funded project to develop a prototype flexible, industrial energy efficiency, and visualization profiling Toolbox (I-CAT). The toolbox embeds energy analytics and Machine Learning (ML) capabilities into an existing commercial SCADA platform for industrial manufacturing operations. This approach allows the creation of an energy Digital Twin. The paper describes requirements of the toolbox, and experimental analysis of the toolbox in a case study, an operational sawmill in Carlisle, UK. Data-driven modeling allows the creation of a predictive model of the energy consumption of the facility from a forecasted production schedule. Mean average modeling errors of less than 10% were obtained. The paper concludes by highlighting areas of future development work.