back to list

Dr.-Ing. Ali Montazeri

Contact

Motivation

The attempts to control electrical machines have led to development of different control strategies where Direct Torque Control and Direct Self Control as well as control strategies based on Pulse Width Modulation (PWM) are standard controller schemes today.

Recently, Model Predictive Control (MPC) schemes have gained more attention by the research community in order to further increase the performance of electrical drives. The MPC controllers are advantageous due to their multi-objective control and optimization feature. Nevertheless, as in the nature of MPC strategies an underlying prediction and optimization problem has to be solved, they are computationally demanding which is a major drawback for practical implementation.

Description

The core idea of the research is to integrate dynamic programming algorithm in a finite control set model predictive controller (Figure 1), which is beneficial regarding reduction in computational effort as well as finding global optimum of a given cost function.

Dynamic programming is a mathematical optimization algorithm, which deals with problems where decision has to be made in stages. In this research, a time discrete-state dynamic programming model with finite horizon, which follows a recursive analytical approach in order to find the best sequence of decisions, will be used.

Research Objectives

The main objective in this research is to develop a finite control set model predictive controller based on dynamic programming, which further enhances the dynamic performance of asynchronous machine compared to standard controllers, while decreases the inverter’s switching and conduction losses. In addition, in order to decrease the copper and core losses in asynchronous machine, an optimized stator flux trajectory -in steady state and transient conditions is a matter of interest.

Furthermore, it will be investigated how to implement dynamic programming (as optimization algorithm) parallel with prediction algorithm which aims to avoid exponential increase of the prediction effort with increasing prediction horizon (Figure 2). Hence, it would be feasible to implement higher prediction horizons on standard computational hardware available for control of drives. A compromise between reasonable length of prediction horizon regarding future machine behavior and available computational platform has to be found.

A System on Chip (SoC) platform where the both processing system (microcontroller) and programmable logic (FPGA) are available can serve as an ideal computational platform for MPC controller. Because, the parallel computational power available on FPGAs can be utilized perfectly for parallel computation of prediction.

Project Partner

The project is funded by the German Research Foundation.

  • Bachelor Thesis: Modular Hardware and Software Development of a Current Measurement Platform for Motor Control Applications
  • Master Thesis: Implementation of Particle Swarm and Pontryagin’s Minimum Principle Optimization Methods on a Developed Model Predictive Controller of an Induction Machine
  • Master Thesis:Dynamic Programming-Based Model Predictive Control of Active Front End
  • Master Thesis:Model Predictive Control of a 2x3 Phase Inverter with Interleaving Chokes
  • Bachelor Thesis: Simulation and Implementation of Continuous Control Set Model Predictive Control of an Asynchronous Machine
  • Bachelor Thesis: Investigating the performance of FCS-MPC with different prediction horizons on single phase passive load
  • Master Thesis: Self-sensing control of a high speed PM synchronous machine
  • A. Montazeri, G. Griepentrog
    Dynamic Programming-Based Optimal Torque Control of Induction Machine
    in IEEE Workshop on Electrical Machines Design, Control and Diagnosis (WEMDCD), Torino, Italy, 26-27 March, 2015
  • A. Montazeri, O. König, G. Griepentrog
    Investigating effects of Delta-Sigma modulation for current measurement on FCS-MPC of IM: Hardware in the loop results
    iin IEEE International Symposium on Predictive Control of Electrical Drives and Power Electronics (PRECEDE), Pilsen, Czech Republic, 4 – 6 September 2017
  • A. Montazeri, G. Griepentrog
    Explicit Consideration of Inverter Losses in the Cost Function for Finite Control Set Model Predictive Control of Induction Machine, Experimental Results
    in 2019 IEEE 28th International Symposium on Industrial Electronics (ISIE), Vancouver, BC, Canada, 2019, pp. 521–526.
  • A. Montazeri, G. Griepentrog
    New Approach for Optimizing Inverter Losses in Finite Control Set Model Predictive Control of Induction Machine, Experimental Evaluation
    in IECON 2019 – 45th Annual Conference of the IEEE Industrial Electronics Society, Lisbon, Portugal, 2019, pp. 4001–4006.