Bogotá - Ingeniería - Doctorado en Ingeniería - Ingeniería Eléctrica · 2022
Data-driven control of interconnected energy systems
This research proposed several algorithms for the identification and control of microgrids based on the Koopman operator. The contributions presented in this manuscript are focused on the control of voltage and reactive power. We have considered five control scenarios based on the Koopman operator: (i) a centralized algorithm that regulates the microgrid voltage without sharing information using MPC. (ii) a non-cooperative distributed control, with a consensus term in the restrictions, that regulates the voltage based on the Koopman model of the inverters. (iii) a cooperative distributed MPC that uses the microgrid Koopman model, where the agents share their control inputs to generate the control signals. Here, we identify the input matrices by using data. (iv) a distributed control that uses data to identify the system error to design an ADMM algorithm. (v) an online data-driven controller that regulates the microgrid voltage and an analysis of the eigenvalues of the system and the effects of noisy measurements.
Full text 121 pages with text out of 123
Read the full thesis Record in the repository
Contents
- Online EDMD Simulationsp. 101
- MPC Algorithm with Disturbances and Noise Rejectionp. 105
- Algorithm with Noise Rejectionp. 107
- Simulation of the System with Perturbationsp. 107
- Conclusions of the Chapterp. 95
- RMS voltage after connecting a load at t = 5s using online EDMMp. 102
- Reactive power at each inverter after connecting the loads at t = 5sp. 103
- Online control signal at each inverter for load changingp. 103
- Voltage at inverter one, ideal, and with noise addedp. 106
- Eigenvalues approximation for two algorithmsp. 106
- Identification of a system with noise using EDMDp. 108
- RMS voltage for the case with disturbancesp. 108
- 4.10 Control signals when there are disturbancesp. 109