Self-optimizing and model-adaptive control of electrical drive systems with predictive planning of pulse patterns

• Applicant: Professor Dr.-Ing. Joachim Böcker
• Subject Area: Electrical Energy Systems, Power Management, Power Electronics, Electrical Machines and Drives
• Term: from 2018 to 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 405351394

Electric drives are used in many purposes, e.g. in automated production lines, power plants, centrifuges, hoists and cranes as well as in road and rail vehicles. Meanwhile, most drives are operated in a controlled fashion where, besides the primary control objective as e.g. the speed, further aspects as minimization of power losses or wear play a role. In the past decades, great research efforts have been made dedicated to issues e.g. of control strategies, modulation and identification methods. Most approaches, however, are focusing on isolated problems or specific applications. A generalizable drive or control concept, which independently configures itself to the respective application-specific requirements and adapts continuously, does not exist.This proposal presents a concept of a multi-level control and operational management for electric drives which is based on the principle of self-optimization. The concept focuses on permanent-magnet synchronous motors with 2-level inverters. However, its modular structure allows to be extended to other configurations with low effort (e.g. induction motors or multi-level inverters). The concept is subdivided into three levels. At the lowest level, the classic control for currents, torque, speed and position is allocated. The parameters of this control are continuously adapted by a superimposed optimization level which makes use of electrical, thermal and mechanical models in order to predict the drive behavior and to intervene in the subordinate control level in case of need. The top level includes functions of self-optimization and adaptation. The latter examines the residuals between model estimation and measurements within the subordinate layer and matches at runtime, i.e. both during commissioning and over the entire drive life, the model parameters. Additionally, by means of self-optimization, the optimization goals are dynamically adjusted due to changing external or internal conditions. In that way it will be ensured that the control concept can be applied for a wide range of applications without manual adjustment.After simulative studies, the proposed control concept will be validated on several test benches for three exemplary applications: 1st: traction drive (rail and road vehicles), 2nd: high-speed continuous drives (pumps, fans, etc.), and 3rd: servo drives (e.g. machine tools). As the validation scenarios include very different requirement profiles, the generalizability of the proposed concept should be demonstrated. After successful completion of the project, the transfer to further drive configurations is planned.

• DFG Programme: Research Grants