Phd Title : Cost of energy optimisation for the operation of microgrids based on Demand Response techniques

Wednesday 1 April  2020, at: 13:00, link : https://meetings.webex.com/collabs/#/meetings/detail?uuid=M9X8PS8V4DNLN20T9ARHF58AG3-BNI5&rnd=593012.48259

Supervisor: Associate Professor  Dionysia Kolokotsa

Seven-membered Examination Committee:

1. Dionysia Kolokotsa
2. Konstantinos Kalaitzakis
3. Georgios Karatzas
4. Mihalis Lazaridis
5. Theocharis Tsoutsos
6. Fotios Kaneos
7. Eftihios Koutroulis

Abstract

This Ph.D. thesis focuses on the development and evaluation of advanced demand response techniques for Near-Zero Energy Buildings (NZEB) and microgrids.

In this context, the energy performance of a residential and an industrial (/office) NZEB was investigated and analysed. The Leaf House (residential) and Leaf Lab (industrial/office) buildings are characterised as NZEB as they effectively integrate energy management systems with a wide range of automation, renewable energy sources, and energy storage. For the evaluation of the energy performance of these buildings, a method was developed and deployed which involved the collection and exploitation of measurements concerning the indoor and outdoor environment, energy consumption and renewable energy production. In addition, dynamic Open Studio / EnergyPlus models of the energy performance of buildings were created and subsequently validated with the aid of the aforementioned measurements and data. The analysis highlighted the importance of evaluating the “performance gap” of buildings as the actual energy performance of buildings can significantly deviate from the “theoretical” values typically used when designing or renovating a building.

Creating validated and dynamic building energy models was a prerequisite for the development and testing of the advanced HVAC demand response methodology described hereafter.

In this context, a novel methodology, for investigating and evaluating the potential HVAC load shifting based on temperature set-point adjustment, was developed and deployed for the industrial building (Leaf Lab). This approach concerns the determination of the hourly temperature set point by a Genetic Algorithm optimisation model. The scenarios that were developed for testing the GA model take into account variable hourly electrical energy prices based on real data by the Day-Ahead market of the building’s region. The optimisation model takes into account variation of the cost of the HVAC’s electrical energy consumption and the Predicted Mean Vote (PMV) index of thermal comfort. Results revealed significant margins of energy and cost savings while comfort levels and temperature set-point drift are kept in line with regulations defined by well-established international standards.

In parallel, a method for short-term (24 hours ahead) prediction of the electrical consumption and Renewable Energy Sources’ production was developed based on Artificial Neural Network models. The method was effectively tested using various datasets to produce results of a high correlation between the real and predicted values, both at building and at the microgrid level, as justified by various indicators (Pearson’s coefficient, MBE, MAPE). Furthermore, a double goal Genetic Algorithm optimisation model of the electrical energy cost and load shifting for the day ahead was developed and thoroughly tested. Day-ahead ANN-based predicted data are used as input for the GA optimisation model to produce balanced solutions for cost savings and load shifting at both building and microgrid level.