Sliding mode searching method for inductance estimation of permanent magnet motor

Abstract

The sliding mode searching method for estimating the inductance of the permanent magnet motor has the advantages that the inductance of the motor can be accurately estimated on the premise that parameters such as prior resistance and flux linkage do not need to be accurately estimated. Firstly, a motor rotor flux linkage is estimated in real time by constructing a sliding-mode flux linkage observer, inductance is indirectly estimated by combining the buffeting effect of the sliding-mode observer and utilizing the fluctuation characteristic of the sliding-mode observer, and dependence of a traditional method on motor parameter accuracy is avoided; and secondly, a rapid search algorithm of a golden section method is adopted, and multiple iterative optimization is combined, so that accurate estimation of the inductance is realized, and the precision of an estimation result and the robustness of the system are remarkably improved. The method can perform inductance estimation under the actual working condition of the motor, has strong adaptability and practicability, can still provide a reliable inductance estimation result especially without depending on prior parameters, has wide application prospects, and provides a novel and efficient technical means for motor control, performance optimization and fault diagnosis.

Description

Technical Field

[0001] The present invention belongs to the field of inductance estimation of permanent magnet motors, and particularly relates to a sliding mode search method for inductance estimation of permanent magnet motors. Background Technique

[0002] Permanent magnet motors have been widely used in modern high-efficiency drive systems. Their excellent power density and high efficiency make them an important power source in fields such as unmanned aerial vehicles and electric vehicles. Inductance is one of the key parameters of permanent magnet motors, and in motor control, inductance directly affects the performance of the motor. In practical applications, accurate estimation of the inductance value is crucial for motor control and performance optimization. Existing inductance estimation methods are mainly divided into two categories: one is the direct off-line measurement method based on sensors, and the other is the on-line identification method. In contrast, the advantage of the on-line identification method is that it can estimate the inductance in real time without stopping the machine or interrupting the operation of the motor, and can accurately estimate under the actual working conditions of the motor, with stronger adaptability and practicability, thus attracting extensive attention.

[0003] Inductance on-line identification technology usually relies on the mathematical model of the motor to construct an observer for inductance estimation. During the process of inductance estimation, in addition to the inductance itself, other accurate parameters such as magnetic flux linkage and resistance are also required. If there are errors in these parameters, it will lead to inaccurate inductance estimation results. However, it is very difficult to ensure the accuracy of parameters such as resistance and magnetic flux linkage in practical applications, which challenges the application of traditional methods under actual working conditions, and existing research usually ignores this problem. For example, Ren Xiangqiang et al. proposed a method for estimating the inductance of a permanent magnet motor based on an observer in the patent "Automatic Identification Method for the Direct and Quadrature Axis Inductances of a Permanent Magnet Synchronous Motor", but regarded the motor resistance as known and accurate.

[0004] The present invention proposes a sliding mode search method for inductance estimation of permanent magnet motors. Its implementation principle is as follows: First, a sliding mode magnetic flux observer is constructed to estimate the rotor magnetic flux of the motor in real time; secondly, combined with the fast search algorithm of the golden section method, using the chattering effect of the sliding mode observer, the inductance value is searched through multiple iterations, and finally the inductance value that minimizes the fluctuation of the observed magnetic flux is found and used as the estimation result of the inductance. Different from traditional inductance estimation methods, this method does not directly rely on the estimation ability of the sliding mode observer, but realizes the inductance estimation through its chattering effect. The main advantage of this method is that it can accurately estimate the inductance under actual working conditions without obtaining accurate motor parameters in advance, thus significantly improving the estimation accuracy and robustness. Summary of the Invention

[0005] To avoid the dependence of the online identification of the inductance of a permanent magnet motor on parameters such as the accurate resistance and magnetic flux of the motor, the present invention proposes a sliding mode search method for estimating the inductance of a permanent magnet motor. This method constructs a sliding mode magnetic flux observer to estimate the rotor magnetic flux of the motor in real time, combines the fast search algorithm of the golden section method, and uses the chattering effect of the sliding mode observer to perform multiple iterative searches, finally determining the inductance value that minimizes the fluctuation of the observed magnetic flux. Through this method, the present invention can achieve accurate estimation of the motor inductance without precise prior parameters, improving the accuracy and robustness of the estimation.

[0006] A sliding mode search method for estimating the inductance of a permanent magnet motor, characterized by the following steps:

[0007] Step 1: Collect current, voltage, speed, and position information: Use a current sensor to collect the three-phase current of the permanent magnet motor, denoted as i a , i b , i c , use a voltage sensor to collect the three-phase voltage of the permanent magnet motor, denoted as u a , u b , u c , use an optical encoder to collect the speed and rotor position information of the permanent magnet motor, denoted as ω m and θ;

[0008] Step 2: Information transformation: According to the rotor position information θ of the permanent magnet motor, transform the collected three-phase voltages u a , u b , u c and three-phase currents i a , i b , i c into voltage and current information in the d, q-axis coordinate systems using the following calculation formulas. The voltage is denoted as u d , u q , and the current is denoted as i d , i q :

[0009]

[0010] Step 3: Construct a sliding mode magnetic flux observer: Use the permanent magnet motor speed ω m , q-axis current i q , d-axis current i d , q-axis voltage u q to construct a sliding mode speed observer (analytical expression) as follows:

[0011]

[0012] Among them, R sis a resistor whose value is randomly set and does not need to be an accurate resistance value, L s is an inductor, i q * is the estimated q-axis current in the sliding-mode speed observer, p is the number of pole pairs of the permanent magnet motor, is the estimated current i q * and the sampled current i q The difference is:

[0013]

[0014] F( ) is the switching function, expressed as:

[0015]

[0016] λ is the scaling factor and satisfies the condition λ ≥ 1;

[0017] Step 4: Set the search range: Set the search range [L smin , L smax for the inductor of the permanent magnet motor, where L smin is the lower limit value of the search range, and L smax is the upper limit value of the search range;

[0018] Step 5: Calculate the characteristic points: Based on the golden section theory, use L smin and L smax to calculate the following characteristic points L q_g1 , L q_m and L q_g2 :

[0019]

[0020] Step 6: Flux linkage observation: Take L smin , L q_g1 , L q_m 、 L q_g2 and L smax as the values of the inductor L s respectively, substitute them into the sliding-mode flux linkage observer constructed in Step 3, and estimate the corresponding flux linkage information ψ _min , ψ _g1 , ψ _m 、 ψ _g2 and ψ _max ;

[0021] Step 7: Flux linkage ripple evaluation: Compare the flux linkage information ψ _min , ψ _g1 , ψ _m 、 ψ _g2 and ψ _maxThe pulsation amplitude, select the flux linkage with the smallest pulsation amplitude, and record the inductance value L corresponding to this flux linkage s as the inductance value P to be optimized 1 , and its value is equal to L smin 、L q_g1 , L q_m 、 L q_g2 and L smax one of them;

[0022] Step 8: Calculate new feature points: Based on the golden section theory and the inductance value P to be optimized 1 , generate new feature points P 0 and P 2 :

[0023]

[0024] Step 9: Flux linkage re-observation: Take P 0 and P 1 as the values of the inductance L s respectively, substitute them into the sliding mode flux linkage observer constructed in Step 3, and estimate the corresponding flux linkage information ψ _ p0 and ψ _p2 ;

[0025] Step 10: Flux linkage pulsation re-evaluation: Compare the pulsation amplitudes of the flux linkage information ψ _ p0 and ψ _p1 , select the flux linkage with the smallest pulsation amplitude, and record the inductance value L corresponding to this flux linkage s as the inductance value P to be optimized 1 , and its value is equal to P 0 and P 2 one of them;

[0026] Step 11: Inductance calculation: Repeat Steps 8 to 10 until the difference between the new feature points P 0 and P 1 is less than 1%. At this time, calculate the inductance value of the permanent magnet motor as:

[0027]

[0028] Among them, L s * is the estimated inductance value of the motor.

[0029] The beneficial effects of the present invention are as follows: The present invention proposes a sliding mode search method for permanent magnet motor inductance estimation. Its main advantage lies in being able to accurately estimate the motor inductance without the need for accurate prior parameters such as resistance and magnetic flux. First, a sliding mode magnetic flux observer is constructed to estimate the rotor magnetic flux of the motor in real time. Combining with the chattering effect of the sliding mode observer, its fluctuation characteristics are used to indirectly estimate the inductance, avoiding the dependence on the accuracy of motor parameters in traditional methods. Second, a fast search algorithm based on the golden section method is adopted, combined with multiple iterative optimizations, to achieve accurate estimation of the inductance, significantly improving the accuracy of the estimation results and the robustness of the system. This method can estimate the inductance under the actual working conditions of the motor, has strong adaptability and practicability. Especially when not relying on prior parameters, it can still provide reliable inductance estimation results, and has a wide range of application prospects, providing a new and efficient technical means for motor control, performance optimization and fault diagnosis. Brief Description of the Drawings

[0030] Figure 1 It is a flowchart of a sliding mode search method for permanent magnet motor inductance estimation of the present invention. Detailed Embodiment

[0031] The present invention will be further described below in conjunction with the drawings and embodiments. The present invention includes but is not limited to the following embodiments.

[0032] As a key parameter for permanent magnet motor performance optimization and control, inductance directly affects the efficiency, stability and control accuracy of the motor. Traditional online inductance identification methods usually rely on the mathematical model of the motor, but these methods usually require accurate motor parameters such as resistance and magnetic flux. In practical applications, the accurate acquisition of these parameters is often restricted to a certain extent, thus affecting the accuracy of inductance estimation. To solve this problem, the present invention proposes a permanent magnet motor inductance estimation method based on sliding mode search. First, a sliding mode magnetic flux observer is constructed to estimate the rotor magnetic flux of the motor in real time, making full use of its chattering effect to avoid the dependence on accurate motor parameters. Second, combined with the fast search algorithm of the golden section method, through multiple iterative optimizations, the inductance value is accurately determined and the magnetic flux fluctuation is minimized. Finally, through several rounds of optimization iterations, the optimal inductance value is finally obtained, realizing high-precision estimation of the motor inductance. This method can accurately estimate the inductance under actual working conditions, significantly improving the estimation accuracy and system robustness, while avoiding the dependence on accurate prior motor parameters in traditional methods. It has a wide range of application prospects. Especially when not relying on high-precision sensors and complex hardware, it provides an efficient and reliable solution for motor control and performance optimization. The flowchart of a sliding mode search method for permanent magnet motor inductance estimation proposed by the present invention is as Figure 1As shown. In this embodiment, the electrical parameters of the permanent magnet motor are as follows: the number of pairs of permanent magnets is 3, the rated speed is 300 rad / s, and the process of estimating the inductance of the permanent magnet motor by using the sliding mode search method is as follows:

[0033] Step 1: Collect current, voltage, speed, and position information

[0034] Use a current sensor to collect the three-phase current of the permanent magnet motor, denoted as i a , i b , i c , and use a voltage sensor to collect the three-phase voltage of the permanent magnet motor, denoted as u a , u b , u c , and use an optical encoder to collect the speed and rotor position information of the permanent magnet motor, denoted as ω m and θ.

[0035] Step 2: Information transformation

[0036] According to the rotor position information θ of the permanent magnet motor, transform the collected three-phase voltages u a , u b , u c and the three-phase currents i a , i b , i c into voltage and current information in the d, q-axis coordinate system by using the following calculation formula. The voltage is denoted as u d , u q , and the current is denoted as i d , i q :

[0037]

[0038] Step 3: Construct a sliding mode flux observer

[0039] Use the speed ω m of the permanent magnet motor, the q-axis current i q , the d-axis current i d , and the q-axis voltage u q to construct the following sliding mode speed observer (analytical expression):

[0040]

[0041] Among them, R s is the resistance, which is artificially set to 0.2 in this embodiment, L s is the inductance, and i q * is the q-axis current estimated in the sliding mode speed observer. p = 3 is the number of pole pairs of the permanent magnet motor, To estimate the current i q * and collect the difference of the current i q , that is:

[0042]

[0043] F( ) is the switching function, expressed as:

[0044]

[0045] λ is the scaling factor and satisfies λ = 2.

[0046] Step 4: Set the search range

[0047] Set the search range [L smin , L smax = [0.001, 1] for the inductance of the permanent magnet motor, where L smin is the lower limit value of the search range and L smax is the upper limit value of the search range.

[0048] Step 5: Calculate the characteristic points

[0049] Based on the golden section theory, use L smin and L smax to calculate the following characteristic points L q_g1 , L q_m and L q_g2 :

[0050]

[0051] Step 6: Flux linkage observation

[0052] Take L smin , L q_g1 , L q_m 、 L q_g2 and L smax as the values of the inductance L s respectively, substitute them into the sliding mode flux linkage observer constructed in Step 3, and estimate the corresponding flux linkage information ψ _min , ψ _g1 , ψ _m 、 ψ _g2 and ψ _max .

[0053] Step 7: Flux linkage ripple evaluation

[0054] Compare the flux linkage information ψ _min , ψ _g1 , ψ _m 、 ψ _g2 and ψ _maxThe pulsation amplitude, select the flux linkage with the smallest pulsation amplitude, and record the inductance value L corresponding to this flux linkage s as the inductance value P to be optimized 1 , and its value is equal to L smin 、L q_g1 , L q_m 、 L q_g2 and L smax one of them

[0055] Step 8: Calculate new feature points

[0056] Calculate new feature points: Based on the golden section theory and the inductance value P to be optimized 1 , generate new feature points P 0 and P 2 :

[0057]

[0058] Step 9: Flux linkage re-observation

[0059] Take P 0 and P 1 as the values of the inductance L s respectively, substitute them into the sliding mode flux linkage observer constructed in Step 3, and estimate the corresponding flux linkage information ψ _ p0 and ψ _p2 .

[0060] Step 10: Flux linkage pulsation re-evaluation

[0061] Compare the pulsation amplitudes of the flux linkage information ψ _ p0 and ψ _p1 , select the flux linkage with the smallest pulsation amplitude, and record the inductance value L corresponding to this flux linkage s as the inductance value P to be optimized 1 , and its value is equal to P 0 and P 2 one of them

[0062] Step 11: Inductance calculation

[0063] Repeat Steps 8 to 10 until the difference between the new feature points P 0 and P 1 generated in Step 8 is less than 1%. At this time, calculate the inductance value L of the permanent magnet motor s * as:

[0064]

Claims

1. A sliding mode search method for permanent magnet motor inductance estimation, characterized in that Here are the steps: Step 1: Collect current, voltage, speed and position information: Use current sensor to collect the three-phase current of permanent magnet motor, denoted as i a ,i b ,i c , use the voltage sensor to collect the three-phase voltage of the permanent magnet motor, denoted as u a ,u b ,u c , the photoelectric encoder is used to collect the permanent magnet motor speed and rotor position information, which are denoted as ω m and θ; Step 2: Information transformation: According to the permanent magnet motor rotor position information θ, the collected three-phase voltage u a ,u b ,u c and three-phase current i a ,i b ,i c Use the following calculation formula to transform the voltage and current information in the d and q axis coordinate system, with voltage being recorded as u d ,u q , the current is recorded as i d ,i q : Step 3: Construct a sliding mode flux observer: Using the permanent magnet motor speed ω m , q-axis current i q , d-axis current i d , q-axis voltage u q The sliding mode speed observer (analytical expression) is constructed as follows: Among them, R s The resistance is set randomly and does not need to be an accurate resistance value. s is the inductance, i q * is the estimated q-axis current in the sliding mode speed observer, p is the number of permanent magnet motor pole pairs, To estimate the current i q * And the acquisition current i q The difference is: F( ) is the switching function, expressed as: λ is the proportionality factor and satisfies the condition λ ≥ 1; Step 4: Set the search interval: Set the search interval of the permanent magnet motor inductance [L smin , L smax ], where L smin is the lower limit of the search interval, L smax is the upper limit of the search interval; Step 5: Calculate feature points: Based on the golden section theory, use L smin and L smax Calculate the following feature points L q_g1 , L q_m and L q_g2 : Step 6: Magnetic flux observation: Set L smin , L q_g1 , L q_m 、 L q_g2 and L smax As inductor L s Substitute the value of into the sliding mode flux observer constructed in step 3 and estimate the corresponding flux information ψ _min , _g1 , ψ _m 、 ψ _g2 and ψ _max ; Step 7: Flux pulsation evaluation: Compare flux information ψ _min , _g1 , ψ _m 、 ψ _g2 and ψ _max The pulsation amplitude is selected, and the flux with the smallest pulsation amplitude is selected, and the inductance value L corresponding to the flux is s The inductance value to be optimized is P1, which is equal to L smin , L q_g1 , L q_m 、 L q_g2 and L smax one of the; Step 8: Calculate new feature points: Based on the golden section theory and the inductance value P1 to be optimized, generate new feature points P0 and P2: Step 9: Observe the flux again: Take P0 and P1 as inductance L respectively s Substitute the value of into the sliding mode flux observer constructed in step 3 and estimate the corresponding flux information ψ _ p0 and ψ _p2 ; Step 10: Re-evaluation of flux pulsation: Compare flux information ψ _ p0 and ψ _p1 The pulsation amplitude is selected, and the flux with the smallest pulsation amplitude is selected, and the inductance value L corresponding to the flux is s is recorded as the inductance value to be optimized P1, which is equal to one of P0 and P2; Step 11: Inductance calculation: Repeat steps 8 to 10 until the difference between the new characteristic points P0 and P1 generated in step 8 is less than 1%. At this time, the inductance value of the permanent magnet motor is calculated as: Among them, L s * Estimate the inductance value for the motor.

Images