This paper addresses the implementation and optimization of an Extended Kalman Filter (EKF) for the Permanent Magnet Synchronous Motor (PMSM) sensorless control using an ARM Cortex-M3 microcontroller. A various optimization levels based on arithmetic calculation reduction was implemented in ARM Cortex-M3 microcontroller. The execution time of EKF estimator was reduced from 260.4 μs to 37.7 μs without loss of accuracy. To further reduce EKF execution time, the separation of a Kalman gain and covariance matrices calculation from prediction and measurement state update, a novel method was proposed, and the performance of it an EKF estimator with separation of a Kalman gain and covariance matrices calculation from prediction and measurement state update was analyzed. Simulation and experiments results validate that the proposed technique could provide the same accuracy with less computation time. A tendency of minimum Kalman gain and covariance matrices calculation frequency from rotor electrical frequency was analyzed and are presented in the paper.
The possibility of applying CMR-B-scalar sensors made from thin manganite films exhibiting the colossal magnetoresistance effect as a fast-nondestructive method for the evaluation of the quality of the magnetic pulse welding (MPW) process is investigated in this paper. This method based on magnetic field magnitude measurements in the vicinity of the tools and joining parts was tested during the electromagnetic compression and MPW of an aluminum flyer tube with a steel parent. The testing setup used for the investigation allowed the simultaneous measurement of the flyer displacement, its velocity, and the magnitude of the magnetic field close to the flyer. The experimental results and simulations showed that, during the welding of the aluminum tube with the steel parent, the maximum magnetic field in the gap between the field shaper and the flyer is achieved much earlier than the maximum of the current pulse of the coil and that the first half-wave pulse of the magnetic field has two peaks. It was also found that the time instant of the minimum between these peaks depends on the charging energy of the capacitors and is associated with the collision of the flyer with the parent. Together with the first peak maximum and its time-position, this characteristic could be an indication of the welding quality. These results were confirmed by simultaneous measurements of the flyer displacement and velocity, as well as a numerical simulation of the magnetic field dynamics. The relationship between the peculiarities of the magnetic field pulse and the quality of the welding process is discussed. It was demonstrated that the proposed method of magnetic field measurement during magnetic pulse welding in combination with subsequent peel testing could be used as a nondestructive method for the monitoring of the quality of the welding process.
A measurement system based on the colossal magnetoresistance CMR-B-scalar sensor was developed for the measurement of short-duration high-amplitude magnetic fields. The system consists of a magnetic field sensor made from thin nanostructured manganite film with minimized memory effect, and a magnetic field recording module. The memory effect of the La1−xSrx(Mn1−yCoy)zO3 manganite films doped with different amounts of Co and Mn was investigated by measuring the magnetoresistance (MR) and resistance relaxation in pulsed magnetic fields up to 20 T in the temperature range of 80–365 K. It was found that for low-temperature applications, films doped with Co (LSMCO) are preferable due to the minimized magnetic memory effect at these temperatures, compared with LSMO films without Co. For applications at temperatures higher than room temperature, nanostructured manganite LSMO films with increased Mn content above the stoichiometric level have to be used. These films do not exhibit magnetic memory effects and have higher MR values. To avoid parasitic signal due to electromotive forces appearing in the transmission line of the sensor during measurement of short-pulsed magnetic fields, a bipolar-pulsed voltage supply for the sensor was used. For signal recording, a measurement module consisting of a pulsed voltage generator with a frequency up to 12.5 MHz, a 16-bit ADC with a sampling rate of 25 MHz, and a microprocessor was proposed. The circuit of the measurement module was shielded against low- and high-frequency electromagnetic noise, and the recorded signal was transmitted to a personal computer using a fiber optic link. The system was tested using magnetic field generators, generating magnetic fields with pulse durations ranging from 3 to 20 μs. The developed magnetic field measurement system can be used for the measurement of high-pulsed magnetic fields with pulse durations in the order of microseconds in different fields of science and industry.
9 July 2022 in International Journal of Robotics and Control Systems
An accurate stator resistance and inductance are necessary for high-performance permanent magnet synchronous motor (PMSM) control. The stator resistance and inductance can be estimated during motor standstill operation. This study proposes a standstill estimation method for the determination of dq-axis inductances and resistance of a PMSM drive system fed by a conventional voltage source inverter (VSI). The proposed method estimates both inductance and the rotor's position using the same algorithm, and knowledge of its initial position is not required. The d- and q-axis inductances were estimated by applying three short-time voltage pulses and measuring phase current peak values. The stator's resistance is estimated by monitoring the exponential decay process of the direct axis current. The method was verified by simulation and experiments conducted on two different PM synchronous motors. A good agreement of simulation and experimental results was obtained. Moreover, the proposed method is relatively simple and can identify stator resistance and inductance at any motor load condition. Compared to the existing parameter estimation strategies, the proposed estimation scheme has a relatively faster estimation time. Additionally, it is shown that the method accounts for the dead-time effect as well.