Passive Fault-Tolerant Model Predictive Control of AC/DC PWM Converter in a Hybrid Microgrid Saeedreza Jadidi *, Hamed Badihi * , **, and Youmin Zhang * § * Department of Mechanical, Industrial and Aerospace Engineering, Concordia University, Montreal, Quebec, Canada (e-mails: saeedreza.jadidi@concordia.ca; youmin.zhang@concordia.ca) ** College of Automation Engineering, Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, Jiangsu, China (e-mail: hamed.badihi@nuaa.edu.cn) Abstract: This paper aims at presenting a novel fault-tolerant control (FTC) scheme for an AC/DC pulse- width modulation (PWM) converter operating in a microgrid framework. A group of interconnected loads and distributed renewable energy resources such as wind farm, solar photovoltaic (PV) farm, and a battery energy storage are considered to form a microgrid. The control system for the AC/DC PWM converter aims at tolerating the fault effects due to power-loss malfunctions in the solar system. A passive fault- tolerant control scheme based on model predictive control (MPC) is proposed and the effectiveness of the designed scheme is demonstrated in an advanced microgrid benchmark model implemented in MATLAB/Simulink environment. Keywords: microgrid, renewable sources, fault-tolerant control, model predictive control, PWM converter. 1. INTRODUCTION One of the key challenges facing the electricity sector is to meet the growing demand for electricity in a safe, secure and environmentally friendly way. For many years, utility companies have relied upon fossil fuels such as coal, oil and natural gas as the main sources of electricity production. However, such conventional sources suffer from several disadvantages such as adverse environmental impacts and limited resources. In addition to the mentioned issues, the existing power grid has several limitations. For instance, only about one third of fuel energy can be converted into electricity in traditional power plants. Also, a significant amount of electricity is lost along the extended transmission lines. Some other challenges relate to the aging infrastructures, nonoptimal usage of the assets, domino-effect failures and widespread blackouts due to the hierarchical topology of the system (Jadidi et al., 2019a). One key solution to overcome the mentioned shortcomings of the conventional power grid is to create an ‘intelligent’ or ‘smart’ grid which integrates information technology and communication systems into the existing power grid. More precisely, smart grid is a cyber-enabled power grid which provides a bidirectional power and information flow to enable a comprehensive control and wide-area monitoring/protection over all distributed grid components. In addition, it facilitates the more efficient integration of intermittent renewable energies such as wind and solar into the power grid. A smart grid is basically a network of smaller grid components called smart microgrids. As initially introduced in Lasseter (2001), a microgrid is a group of distributed energy resources (DERs) and loads which constitute a single controllable entity with the ability to operate in grid-connected and/or islanded modes. Certainly, the microgrid protection is one of the most critical issues regarding the reliability of microgrids. A well- designed protection system is necessary to detect and handle any fault conditions in microgrids. Indeed, protection systems suitable for use in microgrids are expected to be much more complicated and involved comparing with the currently available protection systems. This is particularly true since the energy in a microgrid can flow in different directions. In microgrids, although different types of faults are possible to occur in different components, some faults can be effectively tolerated at the control system level. Generally speaking, conventional control methods cannot guarantee the stability of the system or a desirable performance under fault conditions in components such as actuators, sensors or other subsystems. To solve this problem, fault-tolerant control (FTC) systems are introduced which can maintain the overall performance (under fault conditions) by handling the fault effects. From a control system design point of view, there are two different types of FTC: active (AFTC) and passive (PFTC). AFTC systems use real-time fault detection and diagnosis (FDD) information and control reconfiguration to maintain the entire system stable and achieve an acceptable performance in the presence of faults (Badihi et al., 2019). On the contrary, PFTC systems are fixed controllers which are robust against some levels of faults in the system without using any FDD information or explicit control reconfiguration. The specific application of FTC methods in microgrids is a relatively new topic of research which needs more investigation. In Gholami et al. (2018), a linear state-space model of a DER in grid-connected microgrids is described, and then, an AFTC system for sensor faults is designed using a sliding mode observer (SMO) as an FDD unit. In Youssef and Sbita (2017), adaptive observers are used for fault detection and isolation of the sensors in a three-phase inverter, and an AFTC method is introduced against faults in a photovoltaic system. An FTC strategy based on model predictive control (MPC) approach is presented in Prodan et al. (2015) for reliable operation of microgrid energy management unit. The problem of sensor failures in a wide-area measurement system due to Preprints of the 21st IFAC World Congress (Virtual) Berlin, Germany, July 12-17, 2020 Copyright lies with the authors 12272