Data-driven control of voltage source converters for AC/DC power grids

Abstract

The use of Voltage Source Converters (VSC) is gaining ground in recent yearsmdue to the many benefits it offers. A wide range of electronic devices are using thisntechnology, including systems like renewable energy sources (RES), direct current transmission, Flexible AC transmission system (FACTS), are some of which can be mentioned. The energy matrixes of the countries around the world are changing and the electrical power systems are experiencing new challenges in control, dispatch, design, etc. This thesis involves the design of a new type of controller for VSC. This new controller is based on data gathered from the simulation of multiple setpoints of operations of a model-based control used for a VSC connected to the grid. Before the design of this new data-driven control, an model-based controller is modelled and tuned according to grid parameters and dynamic characteristics. The design of the new controller starts from the data obtained from the previous simulations of the model-based controller. The data is used as input to the control design which is based on the measured variables (time-discrete data). The control design follows many particular criteria defined by the control theory and can be solved by an optimization approach in order to comply with the stability function. The new controller must meet certain parameters such as robustness, speed, stability criterion, and capturing the grid dynamic characteristics within itself. The tests to compare the performance between the data-driven controller and model-based controller consist of using three types of devices that are VSC-based. These tests include step response, voltage recovery under transient conditions due to fault and load shedding. With these comparisons, the perks and disadvantages of using the new control approach were presented and proven using a two-area, four generator electric power system.