Sustainable municipality modelling: Clustering-based bi-level optimization of a decentralized municipality energy and resource treatment infrastructure portfolio

Abstract

Sustainable municipality modelling: Clustering-based bi-level optimization of a decentralized municipality energy and resource treatment infrastructure portfolio Matthias Maldet, Technical University Vienna Energy Economics Group, maldet@eeg.tuwien.ac.at Georg Lettner, Technical University Vienna Energy Economics Group, lettner@eeg.tuwien.ac.at Christoph Loschan, Technical University Vienna Energy Economics Group, loschan@eeg.tuwien.ac.at Daniel Schwabeneder, Technical University Vienna Energy Economics Group, schwabeneder@eeg.tuwien.ac.at Motivation Reaching European climate targets requires interaction between centralized and decentralized generation and conversion technologies. Community operation in the form of renewable energy communities or sustainable communities is often mentioned as a possibility to promote consumer engagement in decentralization. However, these communities often miss a driving force in community formation and technology investments. Municipalities could emerge as an operator for decentralized local energy and resource markets by providing technologies and a local marketplace. The efficient operation of local markets in municipalities should consider resources like waste and sewage. Energy recovery from waste treatment and marketing of such in a municipality could play a fundamental role in decentralized local municipality markets. Furthermore, sewage treatment requires high electricity input energy that should be considered in analyses. Investment decisions in treatment plants and consumer technologies such as PV and heatpumps are investigated. Analyses are performed by developing an optimization model that considers local market operations and investment decisions. However, this leads to high computation time in solving the problem. Therefore, a clustering-based bi-level optimization model is developed to assess the investment and operation of a sustainable municipality. Methods A municipality in Lower Austria is set up as a sustainable municipality. Investigations on the local market operation and portfolio optimization are performed in the municipality. Investment decisions are performed for PV, heat pumps and batteries. Additionally, investment decisions in waste and sewage treatment plants are carried out. District heat investment is further analyzed to use recovered heat from waste incineration in the municipality. Residents in the municipality are aggregated into four representative consumers. Additionally, public buildings and office buildings of local government are considered as own aggregated consumers. Electricity trading between consumers is enabled, and consumers can procure heat from the municipality. Investment decisions and operational analyses are decoupled in a bi-level optimization framework to solve the problem. A portfolio optimization in the municipality is performed in the first step of the optimisation. Annual input data in hourly resolution is clustered into two representative weeks by applying a K-means algorithm. The clustering is done for each month with 30 representative hours for monthly variation. Additional data processing is required for the optimization, whereas K-means algorithms consider mean values as cluster centers. However, for determining treatment capacities, maximum values of resource occurrence must be considered, and a weighting of the cluster centers by maximum values is required. Evaluated capacities in the first step are given as input parameters in the second step. The second step considers a detailed operational analysis of the municipality for a whole year in hourly resolution. The bi-level approach aims to evaluate investment decisions and detailed operational analyses while still keeping the computational time at an acceptable level. Results

Results of the optimization model showed that a sustainable municipality can provide a marketplace for residents. Trading is carried out between consumers to reduce electricity grid procurement. Moreover, recovered energy from waste treatment can be marketed to residents. Consumers carry out district heat investment to purchase the required heat. Investment decisions in waste and sewage treatment plants are based on resource transportation distance and energy availability. A comparison with four-hour mean value optimization is performed to assess the impact. The proposed bi-level optimization framework was efficient. PV investment was carried out to the same capacity in both approaches. District heat systems were installed at the same positions in both methods but at slightly different allocations. The investment of battery storages was undermined in the bi-level practice compared to the four-hour mean-value approach by 36%. Heat pump investment was carried out at 1,6% more but with similar ratios in different positions. Waste and sewage treatment plants were installed at the same capacities at the exact locations in both optimization approaches. A significant advantage was seen in the reduction of the computation time. Computation time decreased from over ten hours in the four-hour mean value approach to 50 minutes in the bi-level method, thus resulting in a computational time reduction of 92%. However, the municipality operation was less cost-efficient in the decoupled modelling, leading to cost increase of 11,5%. Overall, the developed bi-level optimization framework was efficient by means of technology allocation and computational time reduction. A slight increase in costs should be accepted due to all the other advantages of the approach. Since only one option for waste treatment in the portfolio is analysed, the bi-level approach can have an even higher impact in more complicated portfolio optimization scenarios. Acknowledgement Acknowledgement: This work is done in the “Hybrid Local Sustainable Communities” project and is supported with the funds from the Climate and Energy Fund and implemented in the framework of the RTI-initiative “Flagship region Energy” within Green Energy Lab.