Demand-side improvement of short-term load forecasting using a proactive load management – a supermarket use case

Abstract

The future electricity grid will highly depend on cooperation between the supply- and demand-side. One possibility for cooperation is to involve demand-side end-users in a day-ahead load planning process. The end-users could provide their short-term load forecast (STLF) by incorporating local knowledge, which is usually not available to the load aggregators. Moreover, local load management capabilities could even provide a means to decrease the difference between the previously announced load profile and actual consumption. The article demonstrates how the STLF and proactive load management approach could be effectively combined for the supermarket use case. The use case shows that even with the simple STLF predictor, which incorporates the seasonal and regression model parts, a reasonably accurate model can be provided. Moreover, the demand shifting capabilities of the refrigeration systems enable for the deviations from the forecasted demand to be effectively reduced, which consequently increases the accuracy of the previously announced demand plan. Such an approach could be extended to various end-users for which the high STLF accuracy would provide a means to either directly participate in the electricity market, participate in some novel (non)cooperative trading schemes, or to directly support the load aggregators.