Creating Clusters in a 5th Generation District Heating and Cooling Network

Abstract

The Dutch government has set a goal to phase out the use of natural gas by 2050. As a result, municipalities are focusing on implementing alternative heating strategies in certain neighbourhoods. However, districts with older building stocks are often neglected, raising concerns that city centers with older, poorly insulated buildings and high heat demands may be left until later stages of the transition, despite the need for a clear heating strategy in these areas. Additionally, urban areas face the challenge of the urban heat island effect, which leads to trapped heat and elevated summer temperatures. This effect, coupled with the consequences of climate change, will increase the demand for cooling.
To initiate a new heat transition, it is crucial to adopt new sustainable and locally generated heat systems. A potential solution is the implementation of a 5th District Generation Heating and Cooling (5GDHC) network, which utilizes low-temperature heat and cold, potentially sourced from waste heat. However, the design and implementation of a 5GDHC network presents numerous challenges, including the absence of comprehensive guidelines and limited knowledge regarding the deployment of these systems on a larger scale or in an area with an older building stock. This research focuses on developing a tool that can identify potential clusters for a 5GDHC system in densely populated urban areas.
To achieve the research objective, a methodology is developed to identify clusters for a 5GDHC network. Clusters are essential for the implementation of a 5GDHC system in a larger area, as they mitigate investment risks and facilitate a clearer implementation process. The methodology in this study integrates the Single Linkage clustering algorithm and Geometric Graph Theory, which are extended into a model. This model generates clusters based on building locations and energy profiles, and assesses their performance using metrics such as the aggregated hourly lack of supply throughout the year and the total length of the pipe network.
A case study of the inner city of Amsterdam, part of the 'high hanging fruit' project by the AMS Institute, is utilized to test the model. The model requires data on potential waste heat and retrofitted buildings as input. The developed model effectively identifies clusters within a large urban area based on building locations and energy profiles. Trade-offs between pipe network length and energy efficiency must be considered when evaluating the model's results. It is highly recommended to adopt a bottom-up approach and establish 5GDHC clusters incrementally within the city. The hourly disbalances, calculated by the model, can identify potential clusters ready for connection. Moreover, the performance metrics derived from the model can serve as valuable decision-making guides during the design phase of 5GDHC networks. To enhance the decision-making process further, it is crucial to integrate the model's information with urban planning considerations and engage relevant stakeholders. By combining these factors, a comprehensive and well-informed decision-making process can be facilitated, leading to more effective and efficient 5GDHC network designs and implementations.
The full model created within this research can be retrieved from: \url{https://github.com/svanburk/clustering5GDHC.git}