Embedded permanent magnet synchronous motor (IPMSM) has the characteristics of high torque density, wide constant power area, and low eddy current loss of permanent magnet, and has been widely used in many fields. Due to the asymmetry of the AC-DC axis magnetic circuit, both permanent magnet torque and reluctance torque exist in the electromagnetic torque of the embedded permanent magnet synchronous motor. In order to improve the utilization rate of reluctance torque, the maximum torque-to-current ratio (MTPA) control has become the preferred control strategy for embedded permanent magnet synchronous motors. MTPA control minimizes the armature current required to maintain the operation of the motor by distributing the dq shaft current to a certain load torque, thereby reducing copper consumption and improving efficiency.
In order to realize MTPA control, scholars at home and abroad have mainly proposed four methods: analytical method, table lookup method, search method and signal injection method.
The analytical method directly calculates the optimal vector angle of MTPA control according to the mathematical model of the embedded permanent magnet synchronous motor. The look-up table method is based on a large number of offline test data to find the optimal solution of the current working conditions, which has good engineering practicability. The search method does not rely on the mathematical model of the motor and the prior data, and can search for the optimal solution of MTPA control online, but there is a periodic oscillation phenomenon in the steady state. The signal injection method is a relatively novel method that combines the mathematical model solving method and the search method, including the actual signal injection method and the virtual signal injection method.
The existing signal injection methods all extract the required control signal to achieve MTPA control by injecting high-frequency signals and demodulating signals to obtain specific high-frequency components in the response signals. The actual signal injection method is not affected by the change of motor parameters and has high control accuracy, but it will bring additional losses. The Virtual Signal Injection (VSI) method overcomes the problems existing in the actual signal injection method, but the partial derivative term of the motor parameter has an impact on the control accuracy of the algorithm. Both signal injection methods use cascading filtering technology to achieve signal demodulation, which limits the dynamic tracking speed of MTPA state to a certain extent.
The Multiple Virtual Signal Injection (MVSI) method extracts the required control signals through mathematical operations, which accelerates the convergence speed of the algorithm. In addition to the sinusoidal signal, the injection of real/virtual square wave signals also helps to improve the design of the signal demodulation mechanism and shorten the response time of the system to converge to the MTPA state. However, the existing improvement measures for the signal injection method all increase the complexity of the algorithm, which doubles the computing burden of the processor. In addition, the convergence speed of the existing signal injection method is inconsistent under different working conditions, that is, the convergence speed is sensitive to the load condition, which makes it difficult to select the MTPA control gain and limits the practical application of the algorithm.
In order to solve the problems of large amount of calculation, high complexity and low practicability in MTPA control methods, it is difficult to study and improve the methods to optimize the control performance of the algorithm, simplify the solution process, reduce the amount of computation and improve the practicability of MTPA control in engineering applications.
In order to realize the high-performance operation of the embedded permanent magnet synchronous motor, Fu Xinghe, Chen Rui and others from the School of Electrical Engineering of Southeast University and other units wrote an article in the 19th issue of Transactions of China Electrotechnical Society in 2023, proposing a decision-making and control algorithm for the maximum torque-to-current ratio, aiming to simplify the control structure, accelerate the dynamic response of the system, and enhance the robustness of the system.
Table 1 Comparison of the performance of different MTPA methods
Table 2 Comparison of algorithm complexity
The algorithm directly calculates the differential term of electromagnetic torque according to the DQ axis current, and uses it as the decision criterion for the maximum torque-to-current ratio state. The current reference value of the D-axis is compensated according to the criterion value, and the maximum torque-to-current ratio status can be tracked online. The proposed algorithm does not need to inject and solve real or virtual signals, and the dynamic performance of the system can be improved. The D-axis current compensation control solves the problem that the convergence speed of the algorithm is sensitive to the load conditions, and improves the robustness of the control system.
Figure 1 Test platform
The researchers note that this method does not require any form of signal injection process. Compared with the virtual signal injection method, the convergence time of the proposed algorithm to the maximum torque-to-current ratio state is reduced to 1/5~1/2 under different loads.
The results of this work were published in the 19th issue of Transactions of China Electrotechnical Society in 2023, and the title of the paper is "Linear Axis Current Compensation IPMSM Maximum Torque to Current Ratio Control Algorithm Based on Direct Criterion Extraction Method". This project is supported by the Natural Science Foundation of Jiangsu Province, the National Natural Science Foundation of China and the Key Laboratory of Special Electrical Appliances and High Voltage Electrical Appliances of the Ministry of Education.