Sensorless Control with Switching Frequency Square Wave Voltage Injection for SPMSM with Low Rotor Magnetic Anisotropy

Question

1. What is the main drawback of using rotor position sensors in PMSMs and what are the two main groups of sensorless control methods?

2. What is the purpose of injecting high-frequency voltage in the ERF d-axis?

3. What is the proposed technique to increase the signal-to-noise ratio in current measurements for SPMSMs?

4. What hardware is used for experimental tests?

Accepted Answer

The main drawback of using rotor position sensors in Permanent Magnet Synchronous Motors (PMSMs) is the increase in cost, machine size, and reduction in reliability. Sensorless control methods can be divided into two main groups: passive methods and active methods. Passive methods are based on estimators and exploit the machine model to determine rotor angle, while active methods exploit rotor anisotropy by injecting a high-frequency (HF) voltage and demodulating the corresponding HF measured current.

Accepted Answer

Injecting high-frequency voltage in the ERF d-axis is essential for distinguishing between the Real rotor Reference Frame (RRF) and the Estimated rotor Reference Frame (ERF). This injection helps in calculating the rotor angle error between the RRF and the ERF, which is crucial for accurate motor control. The injected voltage allows for the demodulation of the r qh i component, enabling the implementation of a rotor angle estimator. By using the quantity in the square bracket, which is always positive and proportional to r, as feedback error, the system can effectively estimate and correct the rotor angle, ensuring optimal motor performance.

Accepted Answer

The proposed technique to increase the signal-to-noise ratio in current measurements for SPMSMs is current oversampling. By oversampling the current measurements and calculating the average over one switching period, the noise decreases. This is because the noise in current measurements can usually be assumed as Gaussian white noise with zero mean. Oversampling helps to reduce the impact of noise and improve the signal-to-noise ratio, making it easier to estimate the rotor angle accurately.

Accepted Answer

The control hardware used for experimental tests includes an FPGA and a microcontroller integrated in the Dspace MicrolabBox. Two currents are measured with 10A LEMs, and gate pulses are sent as single-ended digital signals to a custom board. The board converts signals to fiber optic signals for inverter control. A resolver interface measures the rotor angle. The first motor test bench uses a 2 Nm motor, while a 25.7 Nm motor simulates the load. A prototype inverter controls the tested motor, and a commercial inverter controls the load-motor with Modbus communication. The parameters of the first test motor and the inverter are reported in Table I. Control parameters are in Table II. A second test bench with a load cell measures motor torque, and the second test motor is coupled with another SPMSM. The injected voltage amplitude is 100V, with a maximum of 280V at standstill. This voltage is within the literature values for similar motors. The second test motor parameters are in Table III.

Accepted Answer

In open-loop, FPGA output is measured with a fixed estimated angle and a variable rotor angle, disabling the microcontroller algorithm. This test is performed for various angle ^r, and FPGA is proportional to the angle error. Theoretically, FPGA should be as in equation (21), where K is a constant dependent on motor parameters, HF voltage injection waveform, measurement hardware, and samples per period. The DS cannot be used for SPMSM due to low signal-to-noise ratio and high order harmonics, while the OS shows good correspondence with ideal behavior with a small offset. The angle error at the zero crossing affects the estimation algorithm.

Accepted Answer

The effect of current oscillations during switching phase is significant, as observed in Fig. 10. However, only a few current samples are performed during this phase, limiting their contribution in the overall integration process. The modified oversampling method, although tested, is more complicated to implement and not used for experimental tests. The instantaneous iq value is sent to a DAC and acquired with an oscilloscope to better understand the effect of current oscillations during the switching phase. The contribution of these samples in the integration process carried out over the entire switching period is limited due to the limited number of samples performed during this phase.

Accepted Answer

In both cases, the FOC control can face the load step and the estimation algorithm follows the rotor angle also during the load step. Reference and measured speed during load steps at 50 rpm are shown in Fig. 18, angle error in Fig. 19, and phase current in Fig. 20. The same quantities are shown for the load step at 250 rpm in Fig. 21, Fig. 22, and Fig. 23, respectively. This demonstrates the FOC control's ability to handle load steps effectively at different speeds.

Accepted Answer

The proposed HF method integrates with the Rotor Flux Observer (RFO) proposed in [10] for medium and high speed applications. The HF method is used until 350 rpm, after which the algorithm switches to the RFO method. This integration allows for high performance in a wide speed range, as proven in [31]. The RFO method is able to perform a full-load step starting from 300 rpm, providing a safety margin for load step disturbance rejection. The speed ripple is almost the same when using the HF method at 350 rpm and the RFO method at 355 rpm, despite differences in angle error.

Accepted Answer

The proposed algorithm shows a significantly lower speed ripple at both low and high speeds compared to the sensorless torque control in [32]. In full-load tests, both algorithms react well to the load step at 250 rpm, showing comparable speed ripple at full load. However, the proposed algorithm demonstrates a more stable performance at very low speed, with smaller speed and torque ripples when the rated torque is applied. Additionally, the speed overshoot obtained with the proposed algorithm when the load is removed is much lower than in the case of the sensorless control torque in [32]. These results indicate that the proposed algorithm outperforms the state-of-the-art sensorless torque control in terms of speed and torque ripple.

Accepted Answer

The high-frequency voltage injection sensorless algorithm offers several advantages. Firstly, it minimizes the disadvantages of voltage injection, such as acoustic noise, by injecting a square wave at the switching frequency. This approach is particularly beneficial for Switched Permanent Magnet Synchronous Motors (SPMSMs) with low anisotropy ratios, as it improves the signal-to-noise ratio. Experimental tests have shown that the proposed algorithm provides accurate angle estimation and facilitates Field Oriented Control (FOC) implementation. Additionally, the algorithm demonstrates a good response under both no-load and full-load conditions, maintaining stability during load steps. The speed loop bandwidth is approximately 5 Hz, and a parallel comparison with a state-of-the-art injection sensorless algorithm has shown superior torque and speed performance in most tests. Overall, the high-frequency voltage injection sensorless algorithm offers improved performance and stability in SPMSM applications.

Sensorless Control with Switching Frequency Square Wave Voltage Injection for SPMSM with Low Rotor Magnetic Anisotropy

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What is the main drawback of using rotor position sensors in PMSMs and what are the two main groups of sensorless control methods?

2. What is the purpose of injecting high-frequency voltage in the ERF d-axis?

3. What is the proposed technique to increase the signal-to-noise ratio in current measurements for SPMSMs?

4. What hardware is used for experimental tests?

5. How does FPGA output in open-loop compare to traditional DS?

6. What is the effect of current oscillations during switching phase?

7. How does FOC control handle load steps at 50 rpm and 250 rpm?

8. How does the proposed HF method integrate with RFO?

9. How does the proposed algorithm compare to the state-of-the-art sensorless torque control in terms of speed and torque ripple?

10. What are the advantages of using a high-frequency voltage injection sensorless algorithm?

Citations

Adaptive Step-Size Predictive PLL based Rotor Position Estimation Method for Sensorless IPMSM Drives

An Improved Adaptive Finite-Time Super-Twisting Sliding Mode Observer for the Sensorless Control of Permanent Magnet Synchronous Motors

Full Speed Range Sensorless Control of Surface Permanent Magnet Synchronous Motor Based on Novel Nonlinear Flux Observer

FPGA-Based High-Frequency Voltage Injection Sensorless Control with Novel Rotor Position Estimation Extraction for Permanent Magnet Synchronous Motor

A sensorless control scheme based on fractional order super-twisting algorithm and improved FLL for PMSM

References

Extended Kalman filter tuning in sensorless PMSM drives

Sensorless drive of surface-mounted permanent-magnet motor by high-frequency signal injection based on magnetic saliency

High-Bandwidth Sensorless Algorithm for AC Machines Based on Square-Wave-Type Voltage Injection

Sensorless Sliding-Mode MTPA Control of an IPM Synchronous Motor Drive Using a Sliding-Mode Observer and HF Signal Injection

FPGA-Based Sensorless PMSM Speed Control Using Reduced-Order Extended Kalman Filters

Related Papers (5)

Square rooter using saw tooth wave-based time division technique

Iron losses of silicon steel under square wave voltage excitation

A Flexible High Voltage Square Wave Generator

Noise predictions for rotor/inlet-probe wake interactions

Electrical Tree Growth under Square Wave Voltages with DC Bias