A method for predicting winding current of a magnetic levitation motor

Abstract

This invention proposes a method for predicting the winding current of a magnetic levitation motor. By adding a current prediction algorithm to the original control algorithm, the actual current value at the output moment is predicted in advance based on the collected current value. This predicted current value is then used in the control algorithm calculations, ensuring that the current used for calculations is as close as possible to the actual output current. This avoids control anomalies caused by deviations between the current value collected at the time of acquisition and the current value at the control output moment, thereby improving the control stability of the magnetic levitation motor. This enhances the real-time performance and accuracy of the control algorithm, ultimately optimizing and improving the overall performance of the magnetic levitation motor.

Description

Technical Field

[0001] This invention relates to the field of magnetic levitation motor technology, and more specifically to a method for predicting the winding current of a magnetic levitation motor. Background Technology

[0002] A magnetic levitation motor is a type of motor that utilizes magnetic levitation technology to achieve contactless bearing support. Compared to traditional mechanical bearing motors, it offers numerous advantages, such as no mechanical wear, high efficiency, high speed, high precision, fast response, and low vibration. Therefore, it is widely used in demanding industrial and scientific research fields such as precision machine tools, magnetic levitation pumps, and aerospace. The operation of a magnetic levitation motor can be broadly divided into three parts: sensor signal acquisition, control algorithm calculation, and control output. High-precision, fast-response detection of the winding current is one of the core technologies for achieving stable and reliable operation of a magnetic levitation motor. Current methods for detecting winding current typically rely on sampling resistors or integrated current sensing ICs. The acquired current, along with other sensor signals, is then used to perform algorithmic calculations to update the control output. Since algorithmic calculations take time, the actual current at the final output stage may have changed significantly from the current at the time of acquisition. Using the current at the time of acquisition for control output would inevitably lead to a decrease in control performance. Therefore, if a winding current prediction algorithm can be added during the control algorithm calculation stage to predict the winding current at the control output moment, and the predicted current can be used for control calculation, the decrease in motor control accuracy due to the delay in current output can be avoided, the real-time performance and accuracy of the control algorithm can be improved, and the overall performance of the magnetic levitation motor can be optimized and improved.

[0003] Existing technologies for winding current detection mainly employ two schemes: one is a winding current acquisition scheme based on sampling current. This scheme indirectly measures the current magnitude by connecting a high-precision sampling resistor in series in the winding circuit and utilizing the voltage drop across the resistor. Its advantages include low cost, simple design, and the ability to meet accuracy requirements to a certain extent. The other scheme is a winding current detection scheme based on an integrated current sensing IC. Compared to the former, this scheme utilizes a dedicated current sensing chip, providing higher measurement accuracy and lower temperature drift, while reducing reliance on external components and simplifying circuit design. Furthermore, this scheme features fast response and strong anti-interference capabilities, contributing to improved overall system performance.

[0004] Both of the above schemes involve using the collected current to calculate the control algorithm. However, there is a significant time delay between the current acquisition time and the actual control output time. This delay means that the current information used for the control algorithm calculation is no longer the latest state. This current deviation will lead to a decrease in the motor control accuracy and reduce the overall performance of the magnetic levitation motor. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention proposes a method for predicting the winding current of a magnetic levitation motor. By predicting the actual current value within the current control cycle and applying this current value to the subsequent control algorithm of the magnetic levitation motor, the current used for calculation by the control algorithm is made as close as possible to the actual output current, thereby improving the real-time performance and accuracy of the control algorithm and optimizing and improving the overall performance of the magnetic levitation motor.

[0006] The motor winding is a typical RL circuit. The full-state current response of an RL circuit is as follows: ,in Where L is the winding inductance, R is the winding resistance, and u is the output voltage of the previous control cycle. Differentiating the above current yields its derivative: When the control cycle is controlled within a certain range, The calculation result is close to 0, meaning the current derivative is close to a constant. .

[0007] The present invention specifically includes the following steps: Within each control cycle T of the magnetic levitation motor, the control cycle is divided into a signal acquisition stage P1, a control algorithm calculation stage P2, and a control output stage P3.

[0008] In phase P1, the current control cycle T is collected. n The current value i n ; In phase P2, based on the previous control period T n-1 Output voltage value u n-1 Current control period T n The current value i collected internally n The predicted current for the current control cycle is estimated using the current derivative formula, which is: T n Periodic Predicted Current Value ;coefficient , The following system of equations was used to calculate the result over three consecutive control cycles, where t1 is the time difference between current acquisition and control output within one cycle, and t2 is the time difference between the current control output and the current acquisition time of the next control cycle: ; ; ; .

[0009] In the P3 phase, the output is the estimated predicted current. ; Repeat stages P1, P2, and P3. Based on the collected current values, predict the actual current value of the current cycle in advance and apply this current value to the subsequent control algorithm of the magnetic levitation motor to achieve stable control of the magnetic levitation motor. Furthermore, the control period T is set to be less than or equal to 0.1. This is to ensure that the current derivative is close to a constant, thereby improving the accuracy of current prediction.

[0010] By continuously collecting multiple sets of control cycle data, the calculation errors of k1 and k2 values ​​are reduced, resulting in smaller errors in the output predicted current value.

[0011] Furthermore, the current prediction control method is implemented through a software algorithm and integrated into the control system of the magnetic levitation motor.

[0012] This invention adds a current prediction algorithm to the original control algorithm, which predicts the actual current value at the output time based on the collected current value, and uses the current value to calculate the control algorithm. This can avoid control anomalies caused by the deviation between the current value at the time of collection and the current value at the time of control output, thereby improving the control stability of the magnetic levitation motor. Attached Figure Description

[0013] Figure 1 A schematic diagram of one control cycle for a magnetic levitation motor; Figure 2 This is a schematic diagram of three consecutive control cycles of a magnetic levitation motor. Figure 3 A schematic diagram for simulation verification using circuit simulation software; Figure 4 To control the output voltage and simulate the current output line graph. Detailed Implementation

[0014] The present invention will be further described below with reference to the accompanying drawings.

[0015] The motor winding is a typical RL circuit. The full-state current response of an RL circuit is as follows: ,in Where L is the winding inductance, R is the winding resistance, and u is the output voltage of the previous control cycle. Taking the derivative of the above current, we can obtain its derivative: When the control cycle is controlled within a certain range, The calculation result is close to 0, meaning the current derivative is close to a constant. .

[0016] like Figure 1As shown, the control cycle T of a magnetic levitation motor is divided into three stages: signal acquisition stage P1, control algorithm calculation stage P2, and control output stage P3.

[0017] To demonstrate the effectiveness of the method described in this invention, this embodiment sets the parameters as follows: winding inductance L = 5.5mH, winding resistance R = 1Ω, control period T = 100µs, time interval t1 between current acquisition and control output in this cycle = 80µs, and time interval t2 between control output and current acquisition in the next cycle = 20µs; Figure 3 As shown in Table 1, circuit simulation software was used to verify the circuit. The control output voltage for each cycle is shown in Table 1. Figure 4 The specific output behavior of the controlled output voltage and simulated current is demonstrated. Based on the current prediction formula: Through such Figure 2 The calculated coefficients for the three periods T1, T2, and T3 shown are as follows. , Where t1 is the time difference between current acquisition and control output within one cycle, and t2 is the time difference between the current control output and the current acquisition time of the next control cycle; u0, u1, and u2 are the output voltage values ​​of the previous cycle, cycle T1, and cycle T2, respectively; i1, i2, and i3 are the current values ​​acquired within cycles T1, T2, and T3, respectively; i T1 i T2 i T3 The predicted current values ​​are for cycles T1, T2, and T3, respectively. Four equations are formed through three consecutive control cycles, yielding coefficients k1≈0.1289 and k2≈-0.0066. This algorithm is then repeatedly applied to predict the current. Table 2 shows the actual sampled current Isamp, the predicted current Iest, the current Isim calculated by circuit simulation software at the control output moment, the percentage absolute error between Isamp and Isim, and the percentage absolute error between Iest and Isim. As can be seen from the table, the deviation between the current predicted by the algorithm and the simulated output current is very small, indicating that the actual current of the winding at the control output moment can be accurately predicted. Using this prediction algorithm for control algorithm calculations can make the control of the magnetic levitation motor more precise, thereby significantly improving the working performance of the magnetic levitation motor.

[0018] Table 1. Control Output <![CDATA[1 st ]]> 6 <![CDATA[2 nd ]]> 8 <![CDATA[3 rd ]]> 5 <![CDATA[4 th ]]> -5 <![CDATA[5 th ]]> 3 <![CDATA[6 th ]]> -3 <![CDATA[7 th ]]> 4 <![CDATA[8 th ]]> 0 Table 2. Current and Error Data

Claims

1. A method for predicting the winding current of a magnetic levitation motor, specifically including the following steps: Within each control cycle T of the magnetic levitation motor, the control cycle is divided into a signal acquisition stage P1, a control algorithm calculation stage P2, and a control output stage P3. In phase P1, the current control cycle T is collected. n The current value i n ; In phase P2, based on the previous control period T n-1 Output voltage value u n-1 Current control period T n The current value i collected internally n The predicted current for the current control cycle is estimated using the current derivative formula, which is: T n Periodic Predicted Current Value ;coefficient , The following system of equations was used to calculate the result over three consecutive control cycles, where t1 is the time difference between current acquisition and control output within one cycle, and t2 is the time difference between the current control output and the current acquisition time of the next control cycle: ； ； ； ； In the P3 stage, the output is the estimated predicted current. ; Repeat stages P1, P2, and P3. Based on the collected current values, predict the actual current value of the current cycle in advance and apply this current value to the subsequent control algorithm of the magnetic levitation motor to achieve stable control of the magnetic levitation motor.

2. The method for predicting the winding current of a magnetic levitation motor as described in claim 1, characterized in that: The control period T is set to be less than or equal to 0.

1. This is to ensure that the current derivative is close to a constant, thereby improving the accuracy of current prediction.

3. The method for predicting the winding current of a magnetic levitation motor as described in claim 1, characterized in that: The current prediction control method is implemented through a software algorithm and integrated into the control system of the magnetic levitation motor.

Images