Forecast Modeling of Demand-Side Flexibility - Prognosmodellering av Flexibilitetsstrategier

Abstract

In a push to combat anthropomorphic climate change and fossil fuel dependence, the electrification of human society is in full swing. Sweden aims to more than double their power consumption by 2040 as an effect. However, with only 14 years left and lead times of building new transmission lines averaging around 10 years, alternatives to grid expansion need to be explored. The use of Demand-Side Management (DSM) strategies would allow for a more effective use of the power grid and allow for further electrification without expanding the grid. The Distribution System Operators (DSO’s) in Sweden have for the past 5-6 years run DSM pilot projects to gauge their efficacy in the Swedish power grid, mainly Local Flexibility Markets (LFM’s). A power tariff has been in the works since 2022 and was set for mandatory implementation by start of 2027. Göteborg Energi (GE), DSO in the city of Gothenburg, have implemented voluntary power tariffs and run an LFM, called Effekthandel Väst, together with Mölndal Energi for five seasons. In order to accurately predict how the load on their power grid will develop in the future, these strategies need to be included in their forecasting model. Modeling systems greatly dependent on human interactions is both complex and complicated. Combined with data scarcity, the need for simplification is significant. A forecast model of DSM would require both a depiction of behavioral response as well as projected future growth. This thesis aims to implement working models of aforementioned strategies into GE’s forecast model. As well as identify simplifications and assumptions necessary to create the models and understand their impact on the models. The models are created via combining existing data over the GE grid with results from prior studies and projected technology developments. The results show varying sensitivity to different parameters with LFM’s being most sensitive to duration of flex deliveries and power tariffs to change in power consumption. Different aspects of the models include varying levels of uncertainty, reflected in the difference in high and low load scenarios. The results further show that DSM strategies can be used to help keep the load below grid constraints but not eliminate the need for further grid expansion completely.