Currently, restricting the utilisation of fossil fuels and thereby limiting global warming to remain below 2ºC stands as one of the most crucial challenges confronting us. The electricity sector is one of the main contributors of CO2 emissions, but it is changing in a rapid pace with a decarbonizing rate which is faster compared to all other fossil sectors. To facilitate the decarbonizing of the electricity sector, optimisation models can provide a valuable framework to gather information about the futuristics of the electricity market. As optimisation models can handle all sort of characteristics like demand and supply which should always be the same, certain policies, energy security, economic development and costs they play an important role in the transition toward more renewables and less fossil fuels. However, these optimisation models do not always present the right solution as societal factors are mostly missing, which can lead to misleading results.
In this paper we will specifically look at the D-EXPANSE optimisation model from the University of Geneva and incorporate two societal aspects. This will be implemented as a hindcasting exercise to examine whether or not it will improve the model compared to the regular model where no societal factors are implemented. This is applied on 31 European countries from 1990 until 2019. The societal aspects that are included in the D-EXPANSE model are public acceptance and heterogeneity of actors. Public acceptance is incorporated in the optimisation model with specifically limiting the CO2 emissions per country with the help of survey data provided from 2009 until 2023 in combination with the set global European emission targets. Heterogeneity of actors is implemented by specifically adjusting the weighted average cost of capital per technology per country per year.
The main results are that it is still unclear whether or not the implementation of societal factors improves the accuracy of the model as a whole. For the implementation of public acceptance 9 out of the 18 countries experience a positive change regarding the error compared to the model where no societal factors are implemented. For the implementation of heterogeneity of actors 13 out of the 26 countries experienced an improvement, and for the combination of both factors 12 out of the 22 countries showed improvements. With this in mind, it is not justifiable that the implementation of public acceptance and/or heterogeneity of actors in this way improves the model which is shown as a hindcasting exercise.
This thesis fails to provide evidence supporting the idea that the inclusion of societal factors enhances the capabilities of optimisation models. This contradicts existing literature, which emphasizes that the incorporation of societal factors is a primary reason why optimisation models struggle to accurately predict the future. One potential explanation for this discrepancy in our findings may lie in the specific methods used to implement actor heterogeneity and public acceptance in the model. For the public acceptance model, it is shown that there is still room for improvement with a different upper limit for the amount of CO2 emissions per country. This can increase accuracy up to 5 percentage points. Therefore, future research should focus on refining the implementation of societal factors, especially considering the accelerating pace of decarbonisation in the electricity sector. Factors such as supply and demand, electricity costs, and energy security remain crucial features that cannot be underestimated. Moreover, with the increasing integration of renewables into the electricity generation, societal factors will continue to exert a growing influence on the progress and implementation.