Implementation and validation of experimental test bench for laboratory-scale microgrid

Abstract

Due to the depletion of non-renewable energy sources, the severity of environmental pollution, and their impact on the global climate, local communities were looking for a reliable and efficient electricity solution. As a result, the notion of microgrid has emerged as the times dictate. With the advancement of society and the rapid development of science and technology, microgrid systems have become an integral part of the power sector, since it can efficiently combine several distributed power generation systems and has a high renewable energy penetration rate. Currently, every country in the world is constructing microgrid experimental platforms and demonstration projects, and vigorously researching and applying microgrids. The primary objective of this thesis is to establish a microgrid experimental platform and conduct experiments and verifications on this test bench, including microgrid power coordination control, real-time calculation, short-term load forecasting, and energy optimization scheduling strategies, to achieve peak load shaving and improve the economic benefits of distributed storage. First of all, this paper provides a background on microgrid research and reviews important experimental and pilot projects related to microgrids. It then studies the microgrid system design and develops a complete physical test platform for microgrids, which includes a battery bank, a load pack, an inverter, and a power meter. Secondly, the overall design of this microgrid test bench is developed utilizing a hierarchical control concept in accordance with the traditional grid-connected microgrid system structure. This includes a microgrid gateway cabinet, power meter, load, and energy storage system. To incorporate optimization algorithms, LabVIEW and SQLite were used concurrently as development tools to create a centralized microgrid control system with features such as realtime data monitoring, intelligent control, power quality analysis, and database management. Additionally, the communication among the controller, energy storage devices, and the power meters is facilitated using the Modbus protocol. Ultimately, this thesis concludes with performance tests of this grid-connected microgrid test bench, which confirm its reliability and accuracy. By assessing changes in load demands and combining the theoretical short-term load forecasting based on neural network algorithms, the economic operation algorithms and energy management techniques have been validated. The performance test was conducted on a weekend day in 2019, whose power consumption was relatively lower than that on a weekday and the system operation is minimally affected by the algorithm. However, the validation tests are mainly scheduled on two days that are considered more representative in 2019, specifically winter and summer. Those two days have the highest power usage and are also the most crucial for load power shaving operation. The experimental results indicate that this system has strong scalability and a user-friendly interface, providing reliable support for comprehensive evaluations, problem diagnostics, and economic analysis.

Date	1 May 2023
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Louis-A. Dessaint (Supervisor)