Parameter-independent estimation strategy combining HF signal injection and sliding mode methodology for AC self-sensing control machines.

Transcription

1 Parameter-independent estimation strategy combining HF signal injection and sliding mode methodology for AC self-sensing control machines. Application to EV/HEV powertrains Renault - Ecole Centrale de Nantes JAMACS, 16/11/2018 Amir Messali, Malek Ghanes, M. Koteich, M.A. Hamida

2 Contents: Problem Statement IPMSM Mathematical Model Proposed Estimation strategy Results Conclusions and Perspectives

3 Contents: Problem Statement IPMSM Mathematical Model Proposed Estimation strategy Results Conclusions and Perspectives

4 Problem Statement The electrical machines are widely used in industrial application for various reasons (economic and ecological.). Automotive field

5 Problem Statement Position Sensor Limitations of use the position sensor Cost Volume Reliability Sensitivity

6 Problem Statement Controller

7 Estimation solutions Model based method Saliency based method Observers HFI (High Frequency Signal Injection) EMF Advantages Kalman 1- Observability Condition condition is guaranteed.. at in all Zero/Low Speed ranges Speed ranges Limitations 1-Observability condition is not guaranteed at Zero/Low Speed ranges 2- Observers depend on the machine parameters 2- HFI don t depends on the model

8 Contents: Problem Statement IPMSM Mathematical Model Proposed Estimation strategy Results Conclusions and Perspectives

9 IPMSM Mathematical Model Classical Model HF Model The voltages Model The Voltage Model The flux-current Model The Flux-Current Model

10 Contents: Problem Statement IPMSM Mathematical Model Proposed Estimation strategy Results Conclusions and Perspectives

11 Position Estimation Techniques

12 Position Estimation Techniques Square-wave Method Measurement axis Injected Voltage

13 3. Signal Injection: General control scheme Injected Voltage form

14 Position Estimation Techniques Current resulting from the injected voltage

15 Position Estimation Techniques Signal processing and demodulation Proposed tracking algorithm Error extraction

16 Classical Position Estimation Techniques Where

17 Proposed Estimation Techniques

18

19 Stability Analysis Based on Sign Propagation : Consider, and the position, speed and the acceleration estimation errors between the virtual model and the observer: (1) whose dynamics are given by: (2)

20 Stability Analysis Based on Sign Propagation : (3) (4)

21 Stability Analysis Based on Sign Propagation : (8) (9) Because

22 Stability Analysis Based on Sign Propagation : (10) (11)

23 Stability Analysis Based on Sign Propagation : (12) (13) (14)

24 Contents: Problem Statement IPMSM Mathematical Model Proposed Estimation strategy Results Conclusions and Perspectives

25 Machine : IPMSM Test bench PMSM motor Load machine

26 Cycle of drive benchmark

27 Robustness Comparative study PLL is evaluated by two tests Nominal Case (without considering Icn variation) By Considering Icn Variation Divergence

28 Simulation results of the proposed robust rotor position, speed and acceleration estimation Injected Frequency = 5 khz Injected magnitude = 2 V Speed Range: all speed ranges Torque: different torque values

29 Experimental results of the proposed robust rotor position, speed and acceleration estimation Injected Frequency = 5 khz Injected magnitude = 2 V Speed Range: all speed ranges Torque: different torque values

30 Contents: Problem Statement IPMSM Mathematical Model Proposed Estimation strategy Results Conclusions and Perspectives

31 Conclusions New Self-sensing control strategy based on sliding mode theory is proposed : + : Can be applied to any salient machines + : can be associated to any HF signal injection technique ++ : Get over the knowledge of machine parameters ++ : Estimate the acceleration