The modelling of future energy scenarios for Denmark

The contributions of this PhD dissertation are in three main areas. In the first main area, existing energy scenarios - The Danish Society of Engineer’s IDA 2050, the large research project CEESA, and the Danish Climate Commission’s CC2050 are compared. Second, energy system analyses for the important but uncertain areas biomass and flexible demand are performed. Thirdly, modelling-related issues are investigated with a focus on the effect of future forecasting assumption and differences between a predefined priority order and order determined by given efficiencies and constraints.
Transformation from a conventional fossil-fueled energy system to a 100% renewable energy system is not an easy task, however, it is a necessary goal to be attained. Academic and governmental bodies in Denmark have released several projects on future scenarios free from fossil fuel by year 2050.
Three existing energy scenarios commonly conclude that a 100% renewable energy system in 2050 is feasible both technologically and economically. The three scenarios assume similar technology options, but different rates of usage. For example, the IDA 2050 scenario establishes an energy system dependent on more biomass than the other scenarios but also a system more self-reliant in terms of electricity with the least international interconnection capacity. The CC2050 scenario builds a more electrified energy system through producing weather dependent, thus intermittent RES rather than biomass and integrating more closely with neighboring countries with the largest interconnection capacity among the three. The CEESA scenario is located between the two scenarios in that sense.
Two scenarios, CEESA and IDA 2050, are compared by methodologies and results. Compared to IDA 2050, CEESA adopts three models external from the overall energy system model for analyzing three subjects which are important but uncertain areas in the future. The first model is a consequential LCA analysis for biomass potential. The second model targets transport demand due to uncertain technology development in the future transport sector. The third model addresses grid stability with a high time resolution. As a result of the consequential LCA, the potential of biomass is less than that of IDA2050. The reduced biomass potential in turn requires larger non-biomass RES capacity, which necessitates a larger capacity of flexible means as a chain effect. Especially, CEESA assumes a new technology more prevalent than IDA 2050 as a mean of reducing biomass in transport sector.
The potential of biomass is regarded as one of the most influential factor in future energy system as explained. However the availability of biomass is uncertain since there might be limits on the environmental capacity and opposition against biomass usage for societal and political reasons in the future. For clarifying better usage of biomass in a 100% renewable energy system, biomass usage for either heat production or electricity generation is compared in this thesis by assuming a decrease of biomass availability for the electricity and heat sectors respectively. The base line scenario for reducing biomass is set to be IDA 2050. The results show that reduction of biomass use in the heat sector could reduce biomass consumption with less electricity exported than the reduction of biomass use in the electricity sector. The less cost solution is achieved through reducing biomass for heat production compared to reducing biomass for electricity production.
Among the technologies supplying flexibility to the system in the future, this PhD dissertation focuses on flexible demand since this subject is less studied in a coherent energy system level than the other flexible means. The potential of flexible demand is assessed in two approaches. The first approach is a technical and bottom-up approach decomposing the electricity consumption into several processes and assesses the potential of flexibility within individual processes. The second approach is a top-down approach to find the level of flexible demand that makes a significant impact on the energy system. For the bottom-up approach the processes are categorized with three criteria; storability, controllability and whether process is interlocked with other processes or independent. The first two criteria are used for assessing flexible demand in residential and commercial sector, and the last criterion is added for the assessment for the industry processes. The results show that 24% of the electricity demand can be moved within a time frame of two hours and approx. 7% of the electricity demand can be moved within a time frame of 24 hours. The system benefit at the assessed amount of flexible demand is limited however. Results from the other analysis indicate that in order to have a significant impact on the energy system performance, more than a quarter of the classic electricity demand would need to be flexible within a month, which is highly unlikely to happen.
For the investigation of the energy system model, EnergyPLAN, which is used for two scenario analyses, two questions are asked; “what is the value of future forecasting assumption in the model?”, and “what is the difference between descriptive and prescriptive model?” For answering these questions, a linear optimization model is created with the ambition to emulate the EnergyPLAN model. The IDA 2050 scenario is used for common scenario for the emulation. Two of the main differences between EnergyPLAN and the created linear optimization model are a) that EnergyPLAN uses an endogen priority of technologies while priority in the linear model is given by efficiencies and b) the linear model could make use of the full set of exogenously given future hourly production and demand values if permitted by user.
Regarding the comparison results of these two models, total fuel consumption for dispatchable plants in the linear optimization model is less than that of EnergyPLAN by 11%. Most of the difference comes from the difference in condensing mode power plants. Except in the condensing mode power plants, the fuel consumptions of CHPs and boilers, irrespective of which district heating grid, are similar in both models.
Despite of the differences in the dispatch order for heat production of CHPs between two models, the overall results are similar since the electricity dispatch order override the heat generation dispatch order, and the electricity dispatch order is similar in the two models due to the chosen efficiencies.
The linear optimization model assumes perfect knowledge about the future in any given situation, and the demands can hence to a greater extent be meet by efficient plants compared with non-perfect knowledge, which results in a reduced fuel consumption. EnergyPLAN bases its production on the specific conditions in each hour, and only uses a limited knowledge about future situations. When the time frame of future forecasting is reduced to a week from a year, the fuel consumption is increased from the one year time frame but less than the case of no future forecasting assumption.