Research team:

• Prof. Teresa Orłowska-Kowalska PhD, DSc, Eng (project manager),
• Maciej Skowron, PhD, Eng.,
• Michał Adamczyk, PhD, Eng.,  Krystian Teler, MSc, Eng.,
• Magdalena Miniach, MSc, Eng.,
• Maciej Skuriat, Eng.

The project's subject matter is related to the significant increase in interest in electric drive systems, not only in industrial automation and robotics, as previously seen, but also in the broadly defined transport of people and goods (carts, drones, aircraft, ships, passenger cars, and trucks). This is driven by environmental concerns and the widespread demand to reduce liquid fuel consumption and harmful gas emissions into the atmosphere, as well as noise and operating costs. Therefore, new requirements for electric drive systems have emerged, such as: low weight and compactness, low cost, high efficiency, maintenance-free operation, and reliability. Modern designs and technological solutions for induction motors (IM), and particularly permanent magnet synchronous motors (PMSM), meet these requirements, and thanks to the use of power electronics and modern control methods, they create electric drive systems ideally suited for the aforementioned applications. However, like all technical systems, electric drives can also be subject to failure. One of the weaker links in such systems are current sensors. Without current information, automatic control systems for the electromagnetic torque and angular speed of AC motors cannot operate. Furthermore, the recovery of inaccessible state variables, which is essential for implementing so-called vector control algorithms used for precise torque control of induction motors or synchronous motors with permanent magnets, cannot be achieved. Therefore, recent years have seen a search not only for effective methods for detecting stator current sensor faults, but above all, for methods for compensating their faults so that the drive system can maintain its full functionality despite the damage and enable safe stopping of the driven device (e.g., a bus arriving at a bus stop, a car arriving at a parking lot, etc.) or the controlled industrial process (e.g., a production line).

The aim of the project is to develop and test, in simulation and experimental studies, drive systems with AC motors (IMs and PMSMs) controlled with vector methods, tolerant to stator current sensors faults.

The implementation of this goal will be possible thanks to the development of new methods for detection and compensation of damage to current sensors in the stator windings of the tested motors, using modified estimators of the state variables and parameters of the motor, including neural networks. The developed methods for compensation of the stator current sensors damages will enable uninterrupted operation of the vector control structures of the IM or PMSM drives in the "current-sensorless" mode until, due to the safety requirements of the controlled process, it will be possible to smoothly stop the drive system.

Various failures of the current sensors will be considered, including complete signal loss, which in the case of only two current sensors used for the three-phase motor drives (which is now common practice in industrial solutions) is the most serious problem because it precludes the operation of the vector control methods of AC motor torque and rotor speed/position. Until recently, the approach used was to switch the control structure to the so-called scalar control, which does not ensure full functionality of the drive system. The innovative solutions proposed in this project will use state variable observers, Kalman filters and neural networks (including deep-NN), equipped with algorithms for adapting selected parameters to ensure good quality of the stator current reconstruction in the event of failure of one or even both stator current sensors. The planned research works are part of the current world research and development trends related to the issues of diagnostics and fault tolerant control in drive automation systems and fit well into the rapidly developing field of fault-tolerant control methods of complex mechatronic systems.