Modelling and optimisation of integrated urban energy systems for both heating and power

Chen, Si (2022) Modelling and optimisation of integrated urban energy systems for both heating and power. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2022chenphd.pdf] PDF
Download (45MB)


Taking into account the rapid increase of renewable energy power generation in the UK, the electrified heating represents an attractive solution for decarbonisation of heat in the long term. However, this will significantly increase the peak power demand in winter and bring further challenges to the grid. Therefore, this work aims to model and optimise a district-level multi-vector integrated energy system for both heating and power through technical and market analysis of using a variety of local renewable energy resources for electricity and heat. In this thesis, the integrated urban energy system is modelled and optimised in multi processes. As a target system, the heating and electricity demand of the University of Glasgow is used as a case study. In order to model the heating and electricity demand under different weather profiles, the heat demand of the buildings is modelled in an engineering model and a statistical model respectively to predict the hourly heat demand according to weather conditions; while the electricity demand is modelled considering both the building baseload and occupancy rate. In heat demand modelling, in order to distinguish the heat demand of each building from the data of whole campus provided by the Energy Center when the detailed building parameters are unknown, this work uses a bottom-up building energy model, which uses physical process of heat transfer to simulate the space heating of buildings, and proposes a Bayesian-based calibration method to calibrate the building parameters in the model. The results show that the Bayesian approach-based emulator performs better with fewer calibration times to find the optimal point, which is relable and efficient to calibrate the thermal parameters in building energy models. Due to the complexity of building a bottom-up building energy model, it is not easy to expand the model to larger areas or add more building samples in the model. Therefore, this work also builds a more general statistical model that can predict the heat demand of different types of buildings simply by giving weather conditions and building characteristics. This work uses artificial neural networks (ANN) technology to simulate the nonlinear relationship between weather conditions, building characteristics and heat demand. In order to improve the training efficiency of ANN, a new sensitivity analysis method is proposed to analyse the correlation between input variables and detect and remove the variables with low importance and the variables that have high importance but contain duplicated features. The result shows the proposed method can re duce training time by around 20% while achieving the same training and prediction performance compared with the traditional sensitivity analysis method. In the electricity demand modelling, the impact of randomness of occupants’ activity on power demand forecasting for buildings has been a difficult problem. In order to solve this problem, this work proposes two approaches for fitting and predicting the electricity demand of office buildings by splitting the time horizon for different occupancy rates. The first proposed approach splits the electricity demand data into fixed time periods and using linear regression approach to fit the building baseload and occupancy rate. The second proposed approach uses the ANN and fuzzy logic techniques to fit the building baseload, peak load, and occupancy rate with multi-variables of weather variables. The result shows that the proposed methods reduce the prediction error of electricity demand by 30% and 55% compared with the conventional ANN approach. To study the impact of electrified heating on buildings and the grid, an Integrated Energy Network (IEN) is established that includes the heat and electric demands of buildings, as well as the generation of local renewable resources and energy storage techniques. In order to rationally plan this new type of IEN based on electric heat pump (HP), this work studies and develops a particle swarm optimisation (PSO) algorithm-based optimisation size method to maximize the decarbonisation on building heating under limited investment cost. According to different source of electric driven, the IEN can be designed as a grid powered HP based heating system and a grid independent renewable heating system (RHS). For the grid powered IEN, this work formulates an operating scheme based on different electricity tariffs to reduce the operational cost of grid power. For the grid independent RHS, this work uses the PSO algorithm to optimise the size of local renewable resources, heat pumps and storage equipment based on the annual investment cost to minimise the total CO2 emission and reduce the operational cost of natural gas. This work provides a feasible solution for how to invest in RHS to replace the existing gas boiler/CHP based heating system. In summary, the significance of this study is to use of local renewable energy sources in electric heating taking into account the local weather conditions and the demand of heat and electricity to reduce carbon emissions in heating and electricity supply.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Colleges/Schools: College of Science and Engineering > School of Engineering
Supervisor's Name: Yu, Prof. Zhibin
Date of Award: 2022
Depositing User: Theses Team
Unique ID: glathesis:2022-82770
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 30 Mar 2022 15:36
Last Modified: 08 Apr 2022 16:41
Thesis DOI: 10.5525/gla.thesis.82770
Related URLs:

Actions (login required)

View Item View Item


Downloads per month over past year