Permanent magnet synchronous motor control method for coping with parameter fluctuation and uncertain load

Abstract

The invention relates to the technical field of permanent magnet synchronous motor model predictive control, in particular to a permanent magnet synchronous motor control method for coping with parameter fluctuation and uncertain load, which comprises the step that a current loop of a motor adopts finite control set model predictive control. According to the permanent magnet synchronous motor control method for coping with the parameter fluctuation and the uncertain load, the core of a T-MPCC strategy is to select an optimal effective voltage vector to accurately control current, and the voltage vector is screened from seven options including six basic voltage vectors and a zero voltage vector. In order to achieve an ideal current control effect in each sector, the seven fundamental voltage vectors need to be predicted online seven times to determine the best selection. Compared with the prior art, the TV-MPCC strategy reduces the number of the alternative voltage vectors to six, so that the number of times of online prediction is correspondingly reduced to six, the calculation burden of the algorithm is remarkably reduced through the improvement, and the efficiency of the control strategy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of permanent magnet synchronous motor model predictive control, and specifically to a control method for permanent magnet synchronous motors to cope with parameter fluctuations and uncertain loads. Background Art

[0002] With the continuous development of industrial automation and precision manufacturing technologies, variable-frequency cooling control systems play an increasingly important role in various production equipment and processes. The variable-frequency cooling control system achieves precise control of the cooling effect by adjusting the operating frequency of the cooling equipment to ensure the stability of the production process and product quality. Existing variable-frequency cooling control systems, although they can meet the cooling requirements in conventional production environments to a certain extent, their performance is not satisfactory when faced with the complexity and variability in actual application scenarios. Especially when dealing with inevitable parameter fluctuations during the production process, such as changes in environmental factors like temperature, humidity, and pressure, and uncertain loads caused by factors such as production batch differences and equipment wear, these systems often appear powerless. These deficiencies are mainly reflected in the system's response speed, adjustment accuracy, and adaptability, resulting in unstable cooling effects and difficulty in ensuring the efficiency of the production process and the constancy of product quality. Therefore, the practical application of these systems is restricted to a certain extent, and there is an urgent need for further optimization and improvement. At the same time, existing control strategies are single and have high energy consumption, lacking effective adaptive adjustment capabilities. There are mainly constant voltage-frequency ratio control and vector control. Among them, constant voltage-frequency ratio control is a simple motor control method that mainly adjusts the motor speed by keeping the ratio of voltage to frequency constant. However, it is inefficient at low loads and high speeds, has limited torque control capabilities, and is sensitive to motor parameter changes, making it unsuitable for occasions requiring high-performance dynamic responses. Although vector control can provide excellent torque and speed control performance and achieve high-precision control of the motor, its disadvantages are that the control strategy is complex, requires high computational capabilities of the control system and parameter accuracy, and may lead to a decline in control performance when parameters change or the motor model is inaccurate. In addition, the design and debugging process of vector control systems are relatively cumbersome, increasing system costs and maintenance difficulties.

[0003] Therefore, it is particularly crucial to deeply study and develop an efficient and reliable control method for permanent magnet synchronous motors to cope with parameter fluctuations and uncertain load problems that may occur in variable-frequency cooling control systems. Summary of the Invention

[0004] The object of the present invention is to provide a control method for a permanent magnet synchronous motor to cope with parameter fluctuations and uncertain loads, so as to solve the problems proposed in the above background technology that the control strategy of the external permanent magnet synchronous motor is complex, has high requirements for the computing power of the control system and the accuracy of parameters, and may lead to a decline in control performance when the parameters change or the motor model is inaccurate. In addition, the design and debugging process of the vector control system is cumbersome, increasing the system cost and maintenance difficulty. To achieve the above object, the present invention provides the following technical solution: A control method for a permanent magnet synchronous motor to cope with parameter fluctuations and uncertain loads, including that the current loop of the motor adopts finite control set model predictive control and adds parameter error feedforward compensation. To cope with the change of the motor parameters with the working state and environment, the sliding mode control algorithm is used to calculate the parameter perturbation error in each cycle and predict the current in the next cycle. The three-vector model predictive current control provides a wide range of choices for the selection of voltage vectors. In terms of the motor speed loop control, an active disturbance rejection controller is adopted.

[0005] Further preferably, for the sliding mode control algorithm, the current equation of the permanent magnet synchronous motor is established in the three-phase abc stationary coordinate system and the three-phase working currents i a 、i b 、i c are collected. After Clark transformation, the equation can be transformed from the three-phase abc stationary coordinate system to the two-phase αβ stationary coordinate system, and the currents i α 、i β in the two-phase αβ stationary coordinate system are obtained. Then, Park transformation is performed to transform the equation from the two-phase αβ stationary coordinate system to the synchronous rotating coordinate system, so as to obtain the currents i d 、i q in the synchronous rotating coordinate system. The state equations of the direct-axis component i d and the quadrature-axis component i q of the motor stator current in the synchronous rotating coordinate system are respectively

[0006]

[0007] Since the algorithm of the present invention is aimed at an external permanent magnet synchronous motor, in the equation, L s is the stator inductance of the motor, u d 、u q are respectively the direct-axis component and the quadrature-axis component of the motor stator voltage, R s is the stator resistance of the motor, ω e is the electrical angular velocity of the motor rotor, ψ f is the permanent magnet flux linkage of the motor rotor. The current equations of the d-axis and q-axis are discretized by the Euler method. According to the d-axis and q-axis current values i d (k)、i q(k), the predicted current value i at the next sampling moment can be estimated d (k + 1), i q (k + 1), the discrete mathematical model is

[0008]

[0009] In the formula, T s is the sampling period of the current, u d (k), u q (k) are the voltage values of the d-axis and q-axis at the current moment, ω e (k) is the electrical angular velocity of the motor at the current moment. Under ideal conditions, assuming that the parameters of the permanent magnet synchronous motor are not affected by the external environment and operating conditions, the behavior of the motor will follow Equation (1-2). However, in reality, due to temperature changes, the resistance of the motor will change; at the same time, the inductance and magnetic flux will also fluctuate due to different degrees of magnetic field saturation. These actual factors will affect the accuracy of the model predictive current control and lead to a decrease in its performance. If the parameters of the actual motor are inconsistent with the model parameters, the predicted current value will deviate, which is not conducive to the control performance of the system and may even cause the motor to malfunction severely. Considering that the disturbance caused by parameter mismatch is limited, the current equation of the permanent magnet synchronous motor is re-expressed as

[0010]

[0011]

[0012] In the formula, f d and f q are the disturbances caused by the parameter mismatches of the d-axis and q-axis respectively, F d and F q are the change rates of the d-axis and q-axis disturbances respectively, ΔR s , ΔL s , Δψ f are the parameter errors of the motor resistance, inductance and magnetic flux respectively.

[0013] Furthermore, preferably, in order to effectively reduce the influence of parameter errors on the system performance, based on the design principles of Equation (1-3) and Equation (1-4), a sliding mode disturbance observer is developed. This observer compensates for these uncertain factors in the control strategy by continuously monitoring and estimating the parameter changes and external disturbances of the motor. The designed sliding mode observer is

[0014]

[0015] In the formula, are the estimated values of the d-axis and q-axis currents respectively, The estimated values of the parameter errors and disturbances on the d-axis and q-axis are d and q respectively, and K d_smo and K q_smo are the sliding mode control functions on the d-axis and q-axis respectively, and λ d and λ q are the sliding mode gain coefficients on the d-axis and q-axis respectively.

[0016] Subtracting Equation (1-5) from Equations (1-3) and (1-4), the estimated error equation can be obtained as

[0017]

[0018] In the formula, e id and e iq are the differences between the estimated currents and the actual currents on the d-axis and q-axis respectively, and e fd and e fq are the differences between the estimated error disturbances and the actual error disturbances on the d-axis and q-axis respectively. The expressions of e id , e iq , e fd , e fq are as follows

[0019]

[0020] Select e id and e iq as the sliding mode surfaces of the sliding mode disturbance observer, and adopt the equal-speed reaching law to design the sliding mode function. It can make the errors e id , e iq , e fd , e fq converge quickly. Therefore, the following expressions can be obtained:

[0021]

[0022] In the formula, s d and s q are the sliding mode surfaces on the d-axis and q-axis respectively. The coefficient k is the reaching law gain coefficient, and its value is positive. To solve the chattering phenomenon caused by the discontinuity of the sign function at the zero point, the present invention uses the sigmod function that is smooth and continuous at the zero point to replace the original sign function sign. The definition of the sigmod function is

[0023]

[0024] Then, from Equation (1-10), it can be deduced that

[0025]

[0026] Considering e fd and e fqis the error between the estimated disturbance and the actual disturbance, relative to e id and e iq exists as a disturbance. When e id and e iq quickly converge to 0, the disturbance will also converge to 0. Therefore, only e id and e iq need to be considered in the sliding mode control function. Substituting Equation (1-12) into Equation (1-6), the sliding mode control function can be obtained as

[0027]

[0028] From Equations (1-5) and (1-13), it can be seen that the designed sliding mode disturbance observer has three gain coefficients that need to be adjusted, namely λ d and λ q and k. In order to ensure that the model can stably reach the predetermined sliding mode surface, the designed sliding mode control function must meet the necessary conditions for sliding mode stability. Therefore, the gain coefficient k must satisfy

[0029]

[0030] To ensure that e id and e iq and their derivatives can quickly converge to 0, λ d and λ q must be positive. Combining Equations (1-5) and (1-13), and after discretizing the model, the model of the sliding mode disturbance observer is derived as follows:

[0031]

[0032] where are the estimated current values at the next sampling moment on the d and q axes respectively, are the parameter disturbance errors at the next sampling moment on the d and q axes respectively, are the parameter disturbance errors at the current moment on the d and q axes respectively.

[0033] Furthermore, preferably, in the operation process of the motor, if the system prediction model is invaded by parameter disturbances or deviations, the sliding mode disturbance observer proposed by the present invention has the ability to monitor and identify in real time the disturbances caused by parameter errors, and the disturbance value is compensated into the prediction model, and can also predict the current value Importantly, the current values output by the sliding mode disturbance observer have undergone disturbance compensation processing. Applying these compensated current values and disturbance values to the model predictive current control model can achieve one-step delay compensation and parameter error compensation of the system, thereby effectively reducing the impact of system delay and disturbance on the operating performance of the permanent magnet synchronous motor. Therefore, the current values output by the sliding mode disturbance observer will be used as the current input values of the model predictive current control model, while the disturbance values will be used as the feedforward compensation of the prediction model. Referring to Equation (1-15), the following outlines the specific implementation steps of the d-axis sliding mode disturbance observer. The observer operation steps for the q-axis follow the same logic:

[0034] S1: Calculate the estimated current and the error value e d between the actual current i id (k), and substitute this error value e id into Equation (1-15) to calculate the sliding mode control function K d_smo ;

[0035] S2: Substitute the value of the variable K d_smo back into Equation (1-15) to calculate the disturbance value at the next sampling moment and record this value. In subsequent calculations, use this as the new for the disturbance value calculation at the next moment;

[0036] S3: Substitute the calculated and K d_smo in the sliding mode disturbance observer model to calculate the current value at the next sampling moment and record this value. In subsequent calculations, use this as the new for the disturbance value calculation at the next moment.

[0037] Further preferably, the three-vector model predictive current control strategy involves selecting two adjacent effective voltage vectors and a zero voltage vector to form an optimal voltage vector combination. These two adjacent effective voltage vectors are selected from six feasible voltage vectors. Through the effective combination of these three voltage vectors, the synthesis region of the optimal voltage vector in each sector can be depicted. In six different sectors, six optimal voltage vectors with variable amplitudes and directions are formed. These vectors together constitute a candidate voltage vector set, which covers all possible directions and magnitudes, providing a wide range of choices for the selection of voltage vectors. The candidate voltage vector set is shown in the following table:

[0038]

[0039]

[0040] During the operation of a permanent magnet synchronous motor, too high a switching frequency will cause the switching devices to act frequently, resulting in a large switching loss, increasing the thermal burden, shortening the service life of the switching devices, and at the same time will also exacerbate electromagnetic interference, affecting the stability of the system and the normal operation of surrounding electronic devices. In addition, a high switching frequency will also increase the noise level of the system, reducing the overall energy efficiency and reliability. In order to fix and reduce the switching frequency of the switching tube, the present invention stipulates the zero voltage vectors selected for each sector as shown in the following table:

[0041]

[0042] Further preferably, in order to accurately select the target voltage vectors u evv_Ⅰ to u evv_Ⅵ in each cycle, it is necessary to use the current deadbeat control technology to accurately calculate the action duration of each candidate voltage vector set in each voltage vector, that is, t i 、t j 、t z . When calculating the time of two adjacent effective voltage vectors and zero voltage vectors, it is necessary to consider the deadbeat tracking of d and q axis currents. Therefore, it is necessary to first calculate the current slopes of the zero voltage vector and two adjacent effective voltage vectors on the d and q axes. Since the present invention adds a sliding mode disturbance observer to the model of model predictive current control to compensate for parameter error disturbances, therefore, when calculating the current slope, disturbance compensation is also required. According to Equation (1-3), Equation (1-4) and Equation (1-15), the current slope calculation formulas for the zero voltage vector and two adjacent effective voltage vectors can be obtained respectively as

[0043]

[0044] In the formula, s d0 、s q0 are the current slopes on the d and q axes when the zero voltage vector acts respectively, s d1 、s q1 、s d2 、s q2 are the current slopes on the d and q axes when two adjacent voltage vectors act respectively, u d1 、u q1 、u d2 、u q2 are the components of two adjacent voltage vectors on the d and q axes respectively.

[0045] Further preferably, after calculating the current slopes when the zero voltage vector and two adjacent effective voltage vectors act, the action times t1 and t2 of the two adjacent effective voltage vectors can be further calculated respectively. Their expressions are

[0046]

[0047] In the formula are the reference current values of the d-axis and q-axis at the next moment respectively. After obtaining the action times t1 and t2 of the two adjacent effective voltage vectors, the action time t0 of the zero voltage vector can be obtained according to the relationship between t1, t2 and the sampling time T s . Since the sum of the action times t0, t1, and t2 of the three voltage vectors is always equal to the sampling period T s , that is, T s = t0 + t1 + t2, the action time t0 of the zero voltage vector can be calculated in 7 cases as follows:

[0048] If t1 ≤ 0 and t2 ≤ 0, then let t1 = 0 and t2 = 0. At this time, t0 = T s ;

[0049] If t1 ≥ 0, t2 ≥ 0 and t1 + t2 ≥ T s , then at this time t0 = T s - t1 - t2;

[0050] If t1 ≥ 0, t2 ≥ 0 but t1 + t2 ≤ T s , then t1 and t2 need to be reduced proportionally before calculating t0. Therefore, t1 = t1 × T s / (t1 + t2), t2 = t2 × T s / (t1 + t2), t0 = 0;

[0051] If 0 ≤ t1 ≤ T s and t2 ≤ 0, then let t2 = 0, t0 = T s - t1;

[0052] If t1 ≥ T s and t2 ≤ 0, then let t1 = T s , t2 = 0, t0 = 0;

[0053] If 0 ≤ t2 ≤ T s and t1 ≤ 0, then let t1 = 0, t0 = T s - t2;

[0054] If t2 ≥ T s and t1 ≤ 0, then let t2 = T s , t1 = 0, t0 = 0;

[0055] Traditional three-vector model predictive current control requires first using t1 and t2 to calculate the voltage values on the d and q axes, and then using Equation (1-2) to calculate the current values at the next moment. This will increase the burden on the data processing chip. Therefore, the present invention simplifies this part. After calculating the action times t1 and t2 of two adjacent effective voltage vectors and the action time t0 of the zero voltage vector, and combining the current slope in Equation (2-1), the currents on the d and q axes at the next moment can be directly calculated. The specific formula is

[0056]

[0057] In the formula, i d (k + 2), i q (k + 2) respectively represent the currents on the d and q axes calculated through the model predictive current control model in each sector. Since the motor control is divided into six different sectors, six groups of i d (k + 2), i q (k + 2) prediction values are obtained. Substituting these six groups of predicted currents into the cost function can screen out the optimal set of data. The action times t1 and t2 of two adjacent effective voltage vectors and the action time t0 corresponding to this set of optimal predicted currents will directly determine the working state of the permanent magnet synchronous motor switching tubes. The value function adopted by the present invention is

[0058]

[0059] When screening the optimal data, substitute the six groups of predicted current data corresponding to the six sectors into Equation (2-4) respectively to calculate six e values, and then compare the magnitudes of the six groups of e values. The smallest e value corresponds to the optimal predicted current.

[0060] Further preferably, after screening out the action times t1 and t2 of two adjacent effective voltage vectors and the action time t0 corresponding to the optimal predicted current, it is necessary to calculate the voltages on the d and q axes at the next sampling moment to provide the input voltage for the sliding mode disturbance observer in the next sampling period. Combining the current slope and voltage vector action time provided by Equation (2-1) and Equation (2-2), the components of the expected voltage vector on the d and q axes in the next sampling period can be obtained. The formula is

[0061]

[0062] The working principle of the entire current loop is as follows:

[0063] The working currents i d (k), i q (k) on the d and q axes of the motor are obtained through ADC sampling, and the electrical angular velocity ω e(k);

[0064] Substitute the collected i d (k), i q (k), ω e (k) and the calculated u d (k), u q (k) into the sliding mode disturbance observer of Equation (1-15), and the current value can be calculated and obtained and the parameter error disturbance The current value obtained here is substituted into the model predictive current control model again, which is equivalent to performing one-step delay compensation;

[0065] Substitute into Equation (2-1) to find the current slopes when the zero voltage vector and two adjacent effective voltage vectors act, and then substitute the obtained current slopes into Equation (2-2) to find the action times t1 and t2 of the two adjacent effective voltage vectors. Then, by judging and determining the action time data t1, t2, and t0 of the final six groups of voltage vectors;

[0066] Substitute the obtained six groups of t1, t2, and t0 into Equation (2-3) respectively, and then find the six groups of predicted current values i d (k + 2), i q (k + 2), and then substitute these six groups of predicted current values into the cost function of Equation (2-4) to screen the optimal predicted current value;

[0067] Use the t1, t2, and t0 corresponding to the optimal predicted current value to control the motor switching tubes.

[0068] Further preferably, the active disturbance rejection controller includes a tracking differentiator (TD), an extended state observer (ESO), and a nonlinear state error feedback (NLSEF). The TD link can achieve non-overshoot fast tracking of the target, ensuring the smoothness of the system response. The ESO link is responsible for observing and compensating for the internal and external disturbances of the system, as well as the errors of the unmodeled part, and regarding these as the total disturbance. The NLSEF link generates a control signal by comparing the difference between the target signal and TD and the difference between the system output and ESO, realizing the effective suppression of the disturbance. The working principle of the speed loop active disturbance rejection controller is as follows:

[0069] Sampling period starts: At the beginning of each sampling period, input the set reference speed ω * into the differential tracker.

[0070] Tracking value acquisition: The differential tracker outputs two tracking values v1 and v2;

[0071] Observer input: actual rotational speed ω and reference current i processed with b gain * As the input of the extended state observer;

[0072] Observed value output: The observer calculates and outputs the estimated value x1 of the actual rotational speed, the differential value x2 of the rotational speed estimation, and the estimated value x3 of the total disturbance;

[0073] Error signal calculation: By comparing v1 with x1 and v2 with x2, error signals e1 and e2 are calculated;

[0074] Nonlinear state error feedback: Error signals e1 and e2 are input into the nonlinear state error feedback module, and after processing, the control quantity u0 is obtained;

[0075] Reference current determination: Subtract the estimated value x3 of the total disturbance adjusted with 1 / b gain from u0 to obtain the reference current value for the next sampling period;

[0076] Control loop: As a new sampling period begins, the system repeats steps 1 to 7 to form a continuous control loop;

[0077] The tracking differentiator plays a key pre-processing role in the active disturbance rejection controller. It is equivalent to introducing a differentiating link. Its main function is to soften the input signal, thereby generating a high-speed and overshoot-free control signal. TD realizes the smooth approximation of the generalized derivative of the input signal through the nonlinear function in ADRC, ensuring the accuracy of control and the stability of the system. Specifically, the discrete form of the second-order TD algorithm effectively completes this process, making the signal processing more delicate and the control response faster without being excessive. Its discrete equation can be expressed as:

[0078]

[0079] In the formula, h is the integration step size, r is the velocity operator, λ is the filtering factor, v1 is the reference rotational speed ω * The processed tracking signal, v2 is the tracking differential signal of v1, and fhan(x1,x2,r,λ) is the nonlinear fast control optimal synthesis function, which can avoid overshoot in the transient process. Its specific expression is

[0080]

[0081] The performance of the tracking differentiator is deeply affected by its performance parameters h, r, λ. For example, an increase in the parameter r will lead to a shorter system tracking time for the target, while an increase in the parameter λ will extend the system's response time, indicating that the system cannot reach the stable state quickly. Therefore, to ensure that the tracking differentiator exhibits excellent performance, the relevant parameters must be carefully adjusted to find the most suitable setting values;

[0082] The extended state observer is the core component of the active disturbance rejection controller. Its function is to accurately estimate unknown disturbances and unmodeled dynamics through a feedback mechanism, thereby obtaining the estimated values of each state variable in the system. This process not only realizes the reconstruction of the target state but also ensures that the system can effectively cope with external and internal uncertainties. As a device specifically used to observe system state variables, the ESO accurately determines each target state quantity by processing the signal information output by the tracking differentiator, providing a solid foundation for optimizing the control strategy. In order to further optimize the control performance of the active disturbance rejection controller, the present invention optimizes the third-order extended state observer, and the expression of its discrete algorithm is

[0083]

[0084] In the formula, ω is the actual rotational speed of the permanent magnet synchronous motor, x1 is the observed value of the actual rotational speed, x2 is the differential variable of x1, x3 is the observed value of the total disturbance, e is the observed error of the rotational speed, b is the compensation factor, h0 is the sampling step, i * is the reference current value of the current loop of the permanent magnet synchronous motor, β1, β2, and β3 are the non-linear adjustable gain coefficients of the extended state observer respectively, α1 and α2 are the non-linear factors of the extended state observer respectively, δ and γ are the filtering factors of the extended state observer respectively, and cfal(e, α, δ, γ) is the improved non-linear function;

[0085] Selecting an appropriate non-linear function is crucial for improving the observation performance of the controller. An ideal non-linear function should have the following characteristics: it should be symmetric about the origin and differentiable at all points. Its curve should have good continuity, convergence, and smoothness. In addition, the parameters to be tuned of the function should be as simplified as possible, and the rule of "providing a larger gain when the error is small and a smaller gain when the error is large" should be followed during design. Traditional non-linear functions often have deficiencies in continuity and smoothness and are prone to jitter phenomena near the origin. To solve these problems, the present invention designs a new continuous and smooth non-linear function cfal(e, α, δ, γ), and the application of this function significantly improves the overall performance of the controller. Its expression is

[0086]

[0087] In the formula, the expressions of l1, l2, and l3 are

[0088]

[0089] Nonlinear State Error Feedback (NLSEF) is an innovative state error feedback control strategy that does not rely on the mathematical model of the controlled object. By combining an improved nonlinear function, the design of the NLSEF algorithm achieves more precise control. In the active disturbance rejection control system, the nonlinear signal error feedback control plays a role similar to the P module in traditional PID control. It can adaptively adjust and output the current value according to the difference between the commanded speed and the estimated speed of the extended observer. This form of nonlinear error feedback control enables the system to respond more flexibly and effectively to various complex control requirements. The specific expression of the NLSEF algorithm is

[0090]

[0091] where u0 is the nonlinear compensation term, is the disturbance compensation term, β4 and β5 are the nonlinear adjustable gain coefficients of the extended state observer respectively, and α3 and α4 are the nonlinear factors of the extended state observer respectively.

[0092] Compared with the prior art, the beneficial effects of the present invention are:

[0093] In the present invention, the core of the T-MPCC strategy is to select an optimal effective voltage vector to precisely control the current. This voltage vector is selected from seven options including six basic voltage vectors and a zero voltage vector. To achieve an ideal current control effect in each sector, seven online predictions of these seven basic voltage vectors are required to determine the best choice. In contrast, the TV-MPCC strategy reduces the number of alternative voltage vectors to six, so the number of online predictions is correspondingly reduced to six. This improvement significantly reduces the computational burden of the algorithm and improves the efficiency of the control strategy.

[0094] In the present invention, the ODC-MPCC strategy constructs an optimal voltage vector combination by selecting an effective voltage vector from six feasible voltage vectors and combining it with a zero voltage vector. In this strategy, the direction of the alternative vector is fixed, but its amplitude can be adjusted to ensure that the selected voltage vector and duty cycle combination are optimal. Nevertheless, the ODC-MPCC strategy is limited to the direction of the basic voltage vector when selecting the voltage vector, which limits its application range. In contrast, although the TV-MPCC strategy also provides six alternative voltage vectors, it allows the voltage vector to be selected in any direction and amplitude, greatly expanding the selection space of the voltage vector and improving the flexibility and efficiency of control.

[0095] In the present invention, in terms of three-vector model predictive current control, the ODC-MPCC strategy only achieves deadbeat control on the q-axis current, while the d-axis current does not achieve it. The TV-MPCC strategy proposed in this paper calculates the vector action time using deadbeat control for both the d-axis and q-axis currents simultaneously, achieving simultaneous deadbeat control of the d-axis and q-axis currents and further improving the control accuracy. Description of the Drawings

[0096] Figure 1 It is a schematic diagram of the process structure of the present invention;

[0097] Figure 2 It is a schematic diagram of the principle of the sliding mode disturbance observer for the d-axis of the present invention;

[0098] Figure 3 It is a schematic diagram of the principle of the sliding mode disturbance observer for the q-axis of the present invention;

[0099] Figure 4 It is a principle block diagram of three-vector model predictive current control of the present invention;

[0100] Figure 5 It is a working principle block diagram of the active disturbance rejection controller of the present invention;

[0101] Figure 6 It is a schematic diagram of the working conditions of the d-axis and q-axis currents of the two strategies when the parameters of the present invention are normal;

[0102] Figure 7 It is a schematic diagram of the working conditions of the d-axis current of the two strategies when a 2 N·m load is applied and the parameters are mismatched in the present invention;

[0103] Figure 8 It is a schematic diagram of the working conditions of the q-axis current of the two strategies when a 2 N·m load is applied and the parameters are mismatched in the present invention;

[0104] Figure 9 It is a schematic diagram of the working conditions of the d-axis current of the two strategies when a 4 N·m load is applied and the parameters are mismatched in the present invention;

[0105] Figure 10 It is a schematic diagram of the working conditions of the q-axis current of the two strategies when a 4 N·m load is applied and the parameters are mismatched in the present invention;

[0106] Figure 11 It is a schematic diagram of the comparison of the rotational speed waveforms between the active disturbance rejection and PID of the present invention. Detailed Implementation Modes

[0107] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0108] Embodiment 1

[0109] Based on the control principle of the invention, a detailed analysis is carried out on three control strategies: single-vector traditional model predictive current control (T-MPCC), double-vector duty ratio model predictive current control (ODC-MPCC), and three-vector model predictive current control (TV-MPCC). The analysis mainly starts from four aspects: the number of vectors, the number of predictions, the voltage vector selection range, and the calculation method of the vector action time. The specific details are as follows:

[0110]

[0111]

[0112] The three-vector model predictive current control adopted by the present invention realizes the efficient control of the motor current through accurate prediction and optimization of the selection and action time of the voltage vector. Its advantages lie in reducing the computational burden of online prediction, expanding the voltage vector selection range, and simultaneously achieving deadbeat control of multi-axis current, thus significantly improving the dynamic response speed and control accuracy of the system and optimizing the motor operation performance.

[0113] Embodiment 2

[0114] In order to verify the effectiveness of the improved sliding mode disturbance observer combined with the three-vector model predictive current control proposed by the present invention in reducing current fluctuations and enhancing the system robustness, the present invention conducts a simulation experiment using the MATLAB / Simulink platform. The following are the relevant parameters of the permanent magnet synchronous motor (PMSM) used in the experiment, as shown in the following table

[0115]

[0116]

[0117] In this invention, through experiments, the traditional one - beat delay three - vector model predictive current control strategy is compared with the three - vector model predictive current control strategy with a sliding - mode disturbance observer to confirm the advantages of the three - vector model predictive current control strategy with a sliding - mode disturbance observer in enhancing system stability. The experiment takes a permanent - magnet synchronous motor driven by a three - phase two - level inverter as the control object. During the experiment, the motor starts from no - load state to 800 r / min, and an additional load of 2 N·m is applied at 0.20 s and 0.40 s. Throughout the experiment, the speed loop adopts the PID control algorithm, and the PI control parameters are kept consistent. The sampling frequency is set to 15 kHz. The parameters of the sliding - mode disturbance observer are configured as k = 8000, λ d and λ q are both 1000.

[0118] When there is no mismatch in the parameters of the permanent - magnet synchronous motor, the operating currents of the traditional one - beat delay three - vector model predictive current control and the three - vector MPCC with a sliding - mode disturbance observer on the d - axis and q - axis are respectively as shown in Figure 6 (a) and (b). It can be clearly seen from the figure that when the parameters of the permanent - magnet synchronous motor are normal, the performance of these two control strategies is almost the same.

[0119] To reproduce the situation of parameter mismatch of the permanent - magnet synchronous motor, the resistance parameter is increased to four times the actual value, the inductance parameter is increased to twice the actual value, and the flux - linkage parameter is set to 0.15 Wb. When the experiment reaches 0.2 s, an additional load of 2 N·m is applied. At this moment, the current responses of the traditional one - beat delay three - vector MPCC and the three - vector MPCC with a sliding - mode disturbance observer on the d - axis and q - axis are respectively shown in the (a) and (b) regions of Figure 7 and the (a) and (b) regions of Figure 8 . Subsequently, when the experiment reaches 0.4 s, another load of 2 N·m is added. Under this new load condition, the current dynamic changes of the traditional one - beat delay three - vector MPCC and the three - vector MPCC with a sliding - mode disturbance observer on the d - axis and q - axis are respectively presented in detail in the (a) and (b) parts of Figure 9 and the (a) and (b) parts of Figure 10 .

[0120] By comparing Figure 7 , Figure 8 , Figure 9 and Figure 10By analyzing and comparing the data therein, the following conclusions can be drawn: When there are parameter mismatches in the permanent magnet synchronous motor, whether it is the d-axis or the q-axis, the current fluctuation amplitude of the motor using the traditional one-beat delay three-vector model predictive current control strategy is generally greater than that of the motor using the three-vector MPCC control strategy with a sliding mode disturbance observer. This phenomenon indicates that under the condition of motor parameter mismatch, the three-vector MPCC control strategy with a sliding mode disturbance observer adopted in the present invention can more effectively suppress current pulsation, thereby maintaining the control performance of the motor at a relatively high level.

[0121] Furthermore, this discovery highlights the superiority and practicality of the control strategy of the present invention. During the operation of the motor, parameter mismatch is an inevitable problem, and the control strategy provided by the present invention can still ensure the smoothness and efficiency of the motor operation under such adverse conditions, which is of great significance for improving the overall performance and reliability of the permanent magnet synchronous motor. Therefore, the present invention not only provides a new solution for the field of motor control, but also provides a strong theoretical basis and practical guidance for the optimization of motor control in practical applications.

[0122] Embodiment III

[0123] The present invention experimentally compares the optimized active disturbance rejection control algorithm with the traditional PID control algorithm to verify that the optimized active disturbance rejection control algorithm can significantly improve the response speed of the system speed loop, enhance stability, and improve robustness. During the experiment, the motor accelerates from no-load to 1500 rad / min and suddenly adds a load of 3 N·m at 0.2 s, and then reduces the speed to 1000 rad / min at 0.4 s. Throughout the experiment, the current loop adopts the improved three-vector MPCC control strategy proposed by the present invention, which combines a sliding mode disturbance observer, and keeps the parameters of the sliding mode disturbance observer constant. The sampling frequency is set to 15 kHz to ensure the accuracy and real-time performance of the control. The parameter configuration of the PID control algorithm is P = 0.18 and I = 18, and the parameter configuration of the active disturbance rejection control algorithm is shown in the following table.

[0124]

[0125] Figure 11 Shows the performance comparison between the optimized active disturbance rejection control algorithm and the conventional PID control algorithm in the present invention for speed control. Observe Figure 10It is clear that when using PID control, the permanent magnet synchronous motor will experience significant overshoot during the starting phase, and it takes a relatively long time to reach the stable state, accompanied by significant steady-state error and slight steady-state fluctuations. On the contrary, in the case of using active disturbance rejection control, the motor hardly shows overshoot during startup and can enter stable operation quickly and smoothly, with both steady-state fluctuations and errors effectively suppressed. In the test of applying load at 0.2 seconds, the adjustment process of PID control is relatively slow after being disturbed, with frequent speed fluctuations and still existing fluctuations and errors after stabilization. While in the face of disturbances, active disturbance rejection control significantly shortens the adjustment time and can achieve a stable state without error at the set working point. In the test of reducing speed at 0.4 seconds, the response of active disturbance rejection control is more rapid and the overshoot is relatively small.

[0126] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of the present invention claimed is defined by the appended claims and their equivalents.

Claims

1. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads, characterized in that: The current loop of the motor adopts finite control set model predictive control and adds parameter error feedforward compensation. In order to cope with the changes in the motor parameters with the working state and environment, a sliding mode control algorithm is used to calculate the parameter disturbance error in each cycle and predict the current of the next cycle. The three-vector model predictive current control provides a wide range of options for the selection of voltage vectors. In terms of the motor speed loop control, an anti-disturbance controller is used.

2. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 1, characterized in that: The sliding mode control algorithm establishes the current equation of the permanent magnet synchronous motor in the three-phase abc stationary coordinate system and collects the three-phase working current i a 、i b 、i c After Clark transformation, the equation can be transformed from the three-phase abc stationary coordinate system to the two-phase αβ stationary coordinate system, and the current i in the two-phase αβ stationary coordinate system can be obtained. α 、i β , and then perform Park transformation to transform the equation from the two-phase αβ stationary coordinate system to the synchronous rotating coordinate system, thereby obtaining the current i in the synchronous rotating coordinate system d 、i q , establish the direct axis component i of the motor stator current in the synchronous rotating coordinate system d and the quadrature axis component i q The state equations are Where, L s is the motor stator inductance, u d 、u q are the direct-axis component and quadrature-axis component of the motor stator voltage, R s is the stator resistance of the motor, ω e is the electrical angular velocity of the motor rotor, ψ f is the permanent magnet flux of the motor rotor. The current equations of the d and q axes are discretized by the Euler method. According to the d and q axis current values ​​i collected at the current moment, d (k), i q (k), the current estimate i at the next sampling moment can be estimated d (k+1), i q (k+1), the discretized mathematical model is Where, T s is the current sampling period, u d (k) and u q (k) is the voltage value of d and q axes at the current moment, ω e (k) is the electrical angular velocity of the motor at the current moment. The motor is subject to temperature changes, causing fluctuations in various parameters. When the disturbance caused by parameter mismatch is limited, the current equation of the permanent magnet synchronous motor is restated as In the formula, f d 、f q are the disturbances caused by the mismatch of d-axis and q-axis parameters, respectively. d 、F q are the rates of change of the d-axis and q-axis disturbances, ΔR s , ΔL s , Δψ f They are the parameter errors of motor resistance, inductance and flux linkage respectively.

3. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 2, characterized in that: In order to effectively reduce the impact of the parameter error on the system performance, these uncertain factors are compensated in the control strategy through the sliding mode disturbance observer. The result is used as the sliding surface of the sliding mode disturbance observer and the sliding function is designed using the constant velocity reaching law. The sigmod function that is smooth and continuous at the zero point is used to replace the original sign function sign. The sliding mode control function must meet the necessary conditions for sliding mode stability. The model of the sliding mode disturbance observer is as follows: In the formula, are the estimated current values ​​of the next sampling moment of the d and q axes, are the parameter disturbance errors of the d-axis and q-axis at the next sampling moment, are the parameter disturbance errors of the d and q axes at the current moment respectively.

4. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 3, characterized in that: The specific execution steps of the d-axis sliding mode disturbance observer are: S1: Calculate estimated current The actual current i d The error value e between (k) id , and its error value e id Substitute into formula (1-15) to calculate the sliding mode control function K d_smo ; S2: Set the variable K d_smo Substitute the value of into formula (1-15) to calculate the disturbance value at the next sampling time And record this value. In subsequent calculations, As new Used for calculating the disturbance value at the next moment; S3: The calculated value of S1 and S2 and K d_smo In the sliding mode disturbance observer model, the current value at the next sampling moment is calculated And record this value. In subsequent calculations, As new Used to calculate the disturbance value at the next moment.

5. The method for controlling a permanent magnet synchronous motor for coping with parameter fluctuations and uncertain loads according to claim 1, characterized in that: The three-vector model predictive current control strategy involves selecting two adjacent effective voltage vectors and a zero voltage vector to form an optimal voltage vector combination, and the two adjacent effective voltage vectors are selected from six feasible voltage vectors. Through the effective combination of these three voltage vectors, the synthesis area of ​​the optimal voltage vector in each sector can be depicted. In six different sectors, six optimal voltage vectors with variable amplitudes and directions are formed. These vectors together constitute a set of candidate voltage vectors. This set covers all possible directions and sizes, providing a wide range of choices for the selection of voltage vectors.

6. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 5, characterized in that: According to the sliding mode disturbance observer model and the reformulation of the current equation of the permanent magnet synchronous motor, the current slope calculation formulas when the zero voltage vector and two adjacent effective voltage vectors act can be obtained respectively: In the formula, s d0 、s q0 are the current slopes on the d and q axes when the zero voltage vector acts, s d1 、s q1 、s d2 、s q2 are the current slopes on the d and q axes when two adjacent voltage vectors act, u d1 、u q1 、u d2 、u q2 They are the components of two adjacent voltage vectors on the d and q axes respectively.

7. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 6, characterized in that: After calculating the current slope when the zero voltage vector and two adjacent effective voltage vectors act, the action time t1 and t2 of the two adjacent effective voltage vectors can be further calculated respectively. The expression is: In the formula are the reference current values ​​of the d and q axes at the next moment, respectively. After the action time t1 and t2 of two adjacent effective voltage vectors are calculated, the sampling time T s The relationship between the zero voltage vector action time t0 is calculated. The sum of the three voltage vector action times t0, t1, and t2 is always equal to the sampling period T s , that is, T s =t0+t1+t2, the action time t0 of the zero voltage vector can be calculated into 7 cases.

8. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 7, characterized in that: After selecting the action times t1 and t2 of the two adjacent effective voltage vectors and the action time t0 of the zero voltage vector corresponding to the optimal predicted current, it is necessary to calculate the voltage of the d and q axes at the next sampling time in order to provide the input voltage for the sliding mode disturbance observer of the next sampling period. The formula is:

9. A permanent magnet synchronous motor control method for coping with parameter fluctuations and uncertain loads according to claim 1, characterized in that: The active disturbance rejection controller includes a tracking differentiator, an extended state observer and a nonlinear state error feedback.

Images