A simplified model for the prediction of energy consumption in large-scale commercial activities

Energy saving has gained increasing attention because it represents a strategy for the reduction of pollutant emissions and dependence on fossil fuels. Nevertheless, many human activities are still energy intensive, thus requiring suitable energy efficiency measures. For this purpose, Energy Service COmpanies (ESCOs) propose solutions for industrial, commercial and residential activities with the aim to reduce energy consumptions. In terms of energy saving, the contribution of ESCO activity is often estimated as the difference between the measured consumptions and a reference one that would have occurred in the same operating conditions, if no modifications were made. When the ESCO intervention leads to a saving, both ESCO and customer share the earnings; otherwise, the ESCO must pay some penalties. As a result, the remuneration system is affected by the reference consumption modeling. This aspect is crucial especially for supermarkets, whose energy consumption might depend on many external factors, such as: climatic variables, thermal performance of the building, customer attendance, etc. Supermarkets require a consumption model for the estimation of the energy saving that must be based on the data available at the contract signature. In attrition, the saving estimation procedure has to be readily understandable by the customer, who often has little experience in the specific field. In this paper, three alternative models are proposed to perform the consumption prediction on monthly basis on 9 case studies. As a result, the maximum average annual error was around 8%, while other supermarkets often show average errors between 4% and 5%. Finally, robustness of provided guidelines is tested on an expanded case study, confirming the obtained results.