1. Introduction
In modern rail traction, asynchronous motors have gradually replaced DC motors as the main type of motor used for rail traction. Vector control is a relatively ideal control method to achieve better AC speed regulation performance. For mining locomotive traction systems, to obtain better control performance, the traditional vector control system should be improved to suit the working characteristics of mining locomotives. Locomotives operate under harsh conditions, with strong vibrations and a dusty working environment, which can significantly damage speed sensors. Therefore, speed sensors cannot be used in certain situations, and the system design must consider achieving sensorless operation, estimating the motor speed using electrical information such as voltage and current. Furthermore, in many cases, the locomotive motor is required to maintain sufficient torque output even when operating above its rated speed. Therefore, the system design must incorporate a suitable field-weakening control algorithm to ensure the motor operates normally above its rated speed.
This paper first designs a corresponding vector control system for mining locomotives based on their working characteristics. Then, it analyzes and studies the speed estimation stage based on the Model Reference Adaptive (MRAS) method within the system. Subsequently, it focuses on the field weakening stage, comparing and analyzing traditional field weakening control strategies with optimized field weakening control strategies considering motor voltage and current constraints. A motor control system based on the TMS320F2812 controller was designed and fabricated, and experimental verification of the speed estimation and field weakening stages based on the MRAS method was completed on this system.
2. Vector Control Principles and System Design
An asynchronous motor is essentially a high-order, nonlinear, and strongly coupled multivariable system. Due to the coupling of various quantities within the motor, directly controlling the three-phase input voltage and current yields unsatisfactory results. Through appropriate coordinate transformation, the internal currents of the motor can be decoupled, greatly simplifying the motor model. In commonly used field-oriented vector control systems, the d-axis of the dq coordinate system is typically aligned with the direction of the rotor magnetic field. In this case, the q-axis component of the rotor flux is zero. The motor model in this coordinate system is as follows:
As can be seen from the torque equation, this method of choosing a coordinate system can achieve current decoupling, thereby controlling the torque and excitation current separately and obtaining better control performance.
As can be seen from the principle of vector control above, the vector control system of an asynchronous motor first transforms the physical quantities in the actual three-phase coordinate system of the motor, then controls the motor's flux linkage and torque in the transformed coordinate system to determine the control setpoint, and finally transforms it back to the actual three-phase coordinate system to control the motor. Considering the working characteristics of the traction motor in mining locomotives mentioned earlier, speed estimation and field weakening control should be added to the traditional vector control. Taking into account the special requirements of the mining locomotive system, the system structure diagram is shown in Figure 1.
3. Rotational speed estimation based on the Model Reference Adaptive (MRAS) method
Sensorless vector control requires accurate estimation of flux linkage and speed information. Traditional open-loop flux linkage observation methods, such as voltage model methods and current model methods, suffer from insufficient observation accuracy and stability. Moreover, these observers can only estimate flux linkage information; speed information requires other estimation methods, making the system more complex. To address this issue, the Model Reference Adaptive (MRAS) method offers a good approach. By selecting an appropriate reference model and a parameter-adjustable model, both the motor speed and flux linkage information can be observed simultaneously.
For vector control systems, based on the mathematical model of the asynchronous motor in a two-phase stationary coordinate system, two forms of flux linkage estimation models can be obtained:
Voltage model
It is evident that the output of both models is the motor flux linkage. Since the voltage model lacks a speed term, it can be chosen as the reference model. The current model is used as the parameter-adjustable model, and the speed is used as the adjustable parameter. Based on Popov's hyperstability theory, the error is selected as...
By selecting the proportional-integral adaptive rate, the angular velocity can be obtained. The identification formula is as follows:
This model-referenced adaptive method allows for the simultaneous acquisition of motor flux linkage and speed information, simplifying the system structure to some extent. MRAS is an identification method based on observer stability design, ensuring asymptotic convergence of parameter estimation. Since the observer in MRAS is based on a reference model, the accuracy of the reference model's parameters directly affects the accuracy of speed and flux linkage identification. To eliminate the influence of pure integrals in the voltage model, it can be improved: a first-order inertial element replaces the pure integral element for the back electromotive force, and the phase lag in state estimation introduced by the inertial element is compensated by the filtered signal of the reference rotor flux linkage. The improved voltage model used in this system is shown in Figure 2.
The improved model still suffers from low estimation accuracy at low speeds. Since low-speed operation is rare in mining locomotive traction systems, this model is only suitable for locomotive traction motor control systems.
4. Magnetic field weakening control
In actual system operation, the magnetic field current and torque current cannot be arbitrarily given. Various constraints must be considered to ensure the normal operation of the motor and avoid problems such as overheating, overcurrent, and operational instability. The inverter's output voltage in the system has extreme values. Under the SVPWM modulation strategy, its maximum output phase voltage amplitude is , where is the DC bus voltage. The inverter's output current and the current the motor can withstand are also limited, with a maximum output current limit. Voltage and current limits must be ensured during control. The steady-state voltage relationship can be derived from the vector control principle, where is the synchronous speed, is the motor's transient inductance, and are the stator, rotor, and magnetizing inductances of the motor, respectively. It can be seen that when the resistance voltage drop is ignored, the motor's terminal voltage is approximately proportional to the product of the flux linkage and the speed. When the motor speed exceeds the rated speed, increasing the motor speed while maintaining the rated flux linkage will lead to a decrease in the maximum torque current, affecting the motor's torque output capability. The most direct way to avoid this is to increase the speed while weakening the internal magnetic field strength of the motor, reducing the magnetizing current, and thus controlling the motor's terminal voltage to maintain its rated value.
Basic field weakening control
Because in the voltage constraint expression, after the speed exceeds the rated speed, let where is the rated synchronous speed, is the current synchronous speed of the motor, and is the rated excitation current. This ensures that the voltage required to maintain the current remains basically constant and will not increase significantly with the increase of speed. For ease of engineering implementation, the rotor speed of the motor is often used instead. Since the motor torque is proportional to the product of , as the speed increases, the output torque of the motor will decrease inversely proportional to the speed, and the output power of the motor will remain constant. This is the basic field weakening control. Under this control method, the torque and power output of the motor are shown in Figure 3.
This method is not affected by motor parameters and is easy to implement in engineering. However, it does not take into account all the constraints of motor operation and cannot reasonably allocate current to maximize torque. Moreover, due to the internal structure of the motor, this algorithm cannot keep the excitation voltage constant. The excitation voltage will increase continuously as the speed increases, which may cause the current to fail to follow effectively and affect the stability of operation.
Optimize field weakening control
6. Conclusion
Due to the unique working environment of mining locomotives, speed estimation and field weakening control must be considered when designing the vector control system for the traction motor. The speed estimation method based on MRAS can simultaneously estimate the motor's speed and flux linkage information, simplifying the system structure and making it well-suited for traction motor vector control systems. Optimized field weakening control ensures the motor's maximum torque output, giving it better load-carrying capacity and a wider operating range, making it suitable for traction motor vector control systems with high torque output requirements.
References:
[1] Li Yongdong. Digital Control System for AC Motors [M]. 1st ed. Beijing: China Machine Press, 2003.
[2]Schauder C. Adaptive Speed Identification for Vector Control of Induction Motors without Rotational Transducer.[J] IEEE-IAS Meeting, 1989:493～499.
[3] Liu Yang. Research on high-performance spindle induction motor drive technology [D]. Wuhan: Huazhong University of Science and Technology, 2008.
[4]Kim Sang-Hoon, Sul Seung-Ki. Maximum torque control of an induction machine in the field weakening region.[J] IEEE Trans. On Ind. Appli., 1995, 31(4):787～794
