A current control method of permanent magnet synchronous motor with adjustable switching frequency FCS-MPC

Abstract

The application relates to a permanent magnet synchronous motor current control method of an adjustable switching frequency FCS-MPC, and compared with the prior art, the method solves the defect that the switching frequency of an inverter is not fixed in the existing FCS-MPC permanent magnet synchronous motor system current control. The method comprises the following steps: collecting drive system data; constructing a discretized current prediction control model; determining a cost function and a delay compensation strategy; calculating a current error control band; selecting an optimal voltage vector; and outputting the voltage vector. The method retains the unique advantages of FCS-MPC, overcomes the inherent defect that the switching frequency is not fixed, enables FCS-MPC to operate at a specified switching frequency within a wide speed regulation range, and through the introduction of a PI type switching frequency controller, can realize the zero-error tracking of a switching frequency reference value. The control method realizes effective regulation and control of the switching frequency on the basis of guaranteeing the current control performance of the permanent magnet motor.

Description

TECHNICAL FIELD

[0001] The application relates to the field of permanent magnet synchronous motor current control, in particular to a permanent magnet synchronous motor current control method based on adjustable switching frequency FCS-MPC. BACKGROUND

[0002] Permanent magnet synchronous motors have the advantages of high efficiency and high power density, and are widely used in new energy vehicles, servo drives and other fields. In order to achieve high-performance control of permanent magnet synchronous motors, many scholars and engineering researchers have proposed various advanced control strategies. Among them, the finite control set model predictive control (FCS-MPC) algorithm, as a potential advanced control algorithm, has been applied in the field of motor drives. FCS-MPC can fully utilize the inherent discrete characteristics of the inverter, directly control the switching state of the inverter, and obtain superior current dynamic control performance. In addition, the cost function design of FCS-MPC has high design flexibility, and multiple control objectives and nonlinear optimization control under constraint conditions can be achieved through the design of the cost function.

[0003] However, the permanent magnet synchronous motor system based on FCS-MPC does not have a modulation link of the inverter, and the switching of the inverter switching state depends on the cost function value, which leads to the fact that the switching frequency of the inverter often changes with the operating conditions, such as the DC bus voltage, the operating speed of the motor, the load, and the sampling time, which will affect the switching frequency, resulting in the problems that the system cannot realize fixed switching frequency operation in a wide speed range, and the inverter switching loss and thermal stress are difficult to quantify, which brings technical challenges to the current control of the FCS-MPC permanent magnet synchronous motor system.

[0004] Currently, there are two technical solutions to solve this problem. The first technical solution is to reintroduce a modulation link and use multiple basic voltage vectors in one control cycle to solve the problem of variable switching frequency in the FCS-MPC permanent magnet synchronous motor system. The use of multiple basic voltage vectors also effectively improves the current control quality of the FCS-MPC permanent magnet synchronous motor system, but also weakens the technical advantages of high dynamic response and high flexibility of FCS-MPC. The second technical solution is to add an additional switching frequency control target to the cost function to realize fixed switching frequency control. This solution better preserves the inherent advantages of FCS-MPC, but requires the introduction of a weight factor to realize simultaneous control of current and inverter switching frequency. However, a fixed weight factor is difficult to obtain balanced control performance in a wide speed range, so the reasonable determination of the weight factor becomes a new technical problem.

[0005] In summary, how to overcome the problem of the variable switching frequency of the FCS-MPC permanent magnet synchronous motor system, skillfully avoid the technical problem of the difficulty in setting the weight factor, and ensure the current control performance of the permanent magnet synchronous motor has become a key technical problem that needs to be solved for the current control of the FCS-MPC permanent magnet synchronous motor system. SUMMARY

[0006] The purpose of the present application is to solve the problem of the variable switching frequency of the existing FCS-MPC permanent magnet synchronous motor system, and to provide a FCS-MPC permanent magnet synchronous motor current control method with adjustable switching frequency, which not only skillfully avoids the technical problem of the difficulty in setting the weight factor, but also ensures the current control performance while achieving precise control of the switching frequency.

[0007] In order to achieve the above-mentioned purpose, the technical scheme of the present application is as follows:

[0008] A FCS-MPC permanent magnet synchronous motor current control method with adjustable switching frequency, comprising the following steps:

[0009] 11) Acquisition of drive system data: acquisition of permanent magnet synchronous motor drive system DC bus voltage, three-phase current, and rotor position data;

[0010] 12) Construction of discretized current predictive control model: coordinate transformation according to rotor position, based on the synchronous speed rotating dq coordinate system of rotor field orientation and the acquired three-phase current, obtain the dq axis current in the synchronous speed rotating dq coordinate system, establish the discretized current predictive control model, substitute the eight basic voltage vectors of the two-level inverter into the discretized current predictive control model, and obtain the dq axis predicted current at k+2 time of each voltage vector;

[0011] 13) Determination of cost function and delay compensation strategy: design of cost function, determination of performance evaluation standard of finite control set model predictive control, substitution of the dq axis current prediction value obtained in step 12) into the designed cost function, and obtaining the cost function size corresponding to each voltage vector;

[0012] 14) Calculation of current error control bandwidth: calculation of the switching frequency of the finite control set model current predictive controller based on the sliding window method, establishment of the switching frequency controller, and determination of the current error control bandwidth;

[0013] 15) Selection of optimal voltage vector: selection of the switching state with the least switching frequency and within the current error control bandwidth determined in step 14) as the optimal voltage vector according to the cost function result of step 13);

[0014] 16) Voltage vector output: the inverter outputs the optimal voltage vector, generates the inverter driving signal according to the inverter switch state corresponding to the optimal voltage vector, completes the voltage vector update of the control period, and realizes the real-time tracking of the d-q axis currents of the permanent magnet synchronous motor.

[0015] The collection of the driving system data comprises the following steps:

[0016] 21) Collecting the direct current bus voltage, three-phase current signal and rotor position signal of the permanent magnet synchronous motor driving system at the k moment;

[0017] 22) Transforming the three-phase current signal to the synchronous speed rotating dq coordinate system oriented by the rotor magnetic field, and representing as:

[0018]

[0019] Wherein, i d , i q are the d-q axis currents, i a , i b , i c are the three-phase currents in the three-phase stationary coordinate system, θ is the rotor position electrical angle, and sinθ and cosθ represent the sine value and cosine value of the rotor position;

[0020] 23) Obtaining eight kinds of switch states of the three-phase two-level inverter, and representing the different combinations [Sa Sb Sc] of the upper bridge arm switch states of the three-phase two-level inverter as {000, 100, 110, 010, 011, 001, 101, 111}. In the synchronous speed rotating dq coordinate system, the d-q axis voltages corresponding to each switch state are represented as:

[0021]

[0022] Wherein, u d , u q represent the d-q axis voltages output by the inverter, Sa, Sb and Sc represent the switch states of the upper bridge arm of the inverter, 1 represents turn-on, 0 represents turn-off, and V dc is the direct current bus voltage.

[0023] The construction of the discretized current predictive control model comprises the following steps:

[0024] 31) Setting that, in the synchronous speed rotating dq coordinate system oriented by the rotor magnetic field, the voltage equation of the permanent magnet synchronous motor is:

[0025]

[0026] In the formula, i d , i q , u d , uq is the dq-axis current, voltage, L s is the stator inductance, R s is the stator resistance, ψ f is the permanent magnet flux linkage, ω r is the rotor electrical angular velocity, denotes the d-axis current differential, denotes the q-axis current differential;

[0027] 32) The discrete current prediction model of the permanent magnet synchronous motor is obtained based on the Euler discretization method as follows:

[0028]

[0029] where k is the kth sampling period, T s is the sampling period, i d (k+1), i q (k+1) represents the dq-axis current prediction value at k+1 time, ω r (k) represents the rotor electrical angular velocity sampling value at k time, u d (k), u q (k) represents the dq-axis voltage output value output by the inverter at k time, L s is the stator inductance, R s is the stator resistance, ψ f is the permanent magnet flux linkage;

[0030] 33) The eight basic voltage vectors of the three-phase two-level inverter are substituted into formula (4) respectively to obtain the dq-axis prediction current values corresponding to each basic voltage vector.

[0031] The determination of the cost function and the time delay compensation strategy includes the following steps:

[0032] 41) To compensate for the timing mismatch caused by the time-consuming calculation, the updated voltage vector at k time and the sampling signal at k time are used to estimate the current at k+1 time, and the current estimation value at k+1 time is taken as the starting point of the current prediction at k+2 time, so as to predict the current at k+2 time, thereby realizing the compensation for the digital control "lagging one beat" through the control "leading one beat" above;

[0033] 42) The cost function is designed as the sum of the square of the dq-axis current tracking error at k+2 time, i.e.:

[0034]

[0035] where, is the dq-axis current reference value, i d (k+2), i q (k+2) represents the dq-axis current prediction value at k+2 time, C limfor preventing overcurrent of the motor, is expressed as:

[0036]

[0037] where k is the kth sampling period, is the dq-axis current reference value, i d (k+2), i q (k+2) represents the dq-axis current prediction value at k+2, I sm represents the maximum allowable current of the permanent magnet synchronous motor;

[0038] 43) Substitute the dq-axis current prediction value corresponding to each basic voltage vector in step 33) into formula (5) to obtain the cost function size corresponding to each basic voltage vector.

[0039] The calculation of the current error control band includes the following steps:

[0040] 51) Based on the idea of sliding window, the ratio of the total number of switching times of the upper bridge arm switches of the inverter in a certain time window to the duration of the time window represents the average switching frequency of the inverter,

[0041] The average switching frequency of the inverter at k is expressed as:

[0042]

[0043] where k is the kth sampling period, f sw (k) is the average switching frequency of the inverter calculated at k, T w is the duration of the sliding window, n represents the number of time periods of the sliding window, S = [Sa, Sb, Sc] represents the switching state of the upper bridge arm of the inverter, 1 represents turn-on, 0 represents turn-off, ‖‖1 represents the total number of switching times of the upper bridge arm switches of the three-phase two-level inverter, S(k-i) is the switching state of the upper bridge arm of the inverter at k-i, S(k-i-1) is the switching state of the upper bridge arm of the inverter at k-i-1;

[0044] To eliminate the quantization noise existing in the calculation of the switching frequency, the real-time calculation value of the switching frequency is further processed through a low-pass filter, and the final real-time feedback of the switching frequency is:

[0045]

[0046] where k is the kth sampling period, T s represents the control period, a is the filter coefficient, and ω c is the cutoff frequency of the first-order low-pass filter, f sw (k) is the average switching frequency of the inverter calculated at k, f swL(k) is the filtered average switching frequency of the inverter at k, f swL (k-1) is the average switching frequency of the inverter calculated at k-1;

[0047] 52) Based on the feedback calculation of real-time switching frequency according to formula (7)-(8), a switching frequency control loop is established,

[0048] The switching frequency controller adopts a PI regulator, and the output of the PI regulator is the current control error bandwidth. The switching frequency controller is represented as:

[0049]

[0050] Where k is the kth sampling period, T s represents the control period, δ(k) is the current control error bandwidth at k, is the reference value of the average switching frequency of the inverter at k, f swL (k) is the filtered average switching frequency of the inverter at k, k p and k i is the proportional gain and integral gain of the switching frequency controller, e f (k) is the average switching frequency error of the inverter at k, e f (k-1) is the average switching frequency error of the inverter at k-1;

[0051] The range of the output variable δ of the switching frequency controller is limited to:

[0052] δ min <δ(k)≤δ max (10)

[0053] Where k is the kth sampling period, δ(k) is the current control error bandwidth at k, δ min is determined by the maximum value of the average switching frequency of the inverter allowed by the operating point, δ max is determined by the maximum value of the current ripple allowed by the operating point;

[0054] 53) The real-time average switching frequency of the inverter is realized by adjusting the bandwidth of the current control error band;

[0055] When the reference value of the average switching frequency of the inverter is lower than the actual switching frequency, the positive error is input into the PI regulator, and the PI gradually increases the width of the current control error band, so as to realize accurate tracking of the switching frequency reference value;

[0056] When the switching frequency reference value is higher than the actual switching frequency, the negative error is input into the PI regulator, and the PI gradually reduces the width of the current control error band;

[0057] Under the regulation of the integrator, the actual switching frequency of the inverter realizes the zero-error tracking of the switching frequency reference value.

[0058] The selecting the optimal voltage vector comprises the following steps:

[0059] 61) According to the voltage vector applied at k+1 moment, the candidate voltage vector at k+2 moment is divided into three categories according to the switching times:

[0060]

[0061] Wherein, k is the kth sampling period, S(k+1) and S(k+2) represent the inverter switching state at k+1 and k+2 moments, S=[Sa, Sb, Sc] represents the switching state of the upper bridge arm of the inverter, 1 represents on, 0 represents off, ‖‖1 represents the total switching times of the upper bridge arm of the three-phase two-level inverter, V I represents the switching state set without bridge arm switching, V II represents the switching state set of one bridge arm switching, V III represents the switching state set of two bridge arm switching;

[0062] If the voltage vector at the last moment is continued to be applied at (k+2) moment, the switching state does not change, at this time the switching times is zero, the voltage vector at the last moment is confirmed as V I group;

[0063] If a new voltage vector is applied at (k+2) moment and only the switching state of one bridge arm changes, the vector is confirmed as V II group, if a new voltage vector is applied at (k+2) moment and the switching states of two bridge arms change, the vector is confirmed as V III group;

[0064] 62) Based on the classification result of the candidate voltage vector defined by formula (11), the selection mode of the candidate voltage vector at k+2 moment is represented as:

[0065]

[0066] Wherein, k is the kth sampling period, V(k+2) represents the optimal voltage vector to be output at k+2 moment, V optI represents the voltage vector with the minimum cost function in V I , V optII represents the voltage vector with the minimum cost function in V I , V II represents the voltage vector with the minimum cost function in V optIII , V I represents the voltage vector with the minimum cost function in V II , V IIIThe voltage vector with the minimum cost function, g[] represents the cost function obtained in step 43, and delta is the current error control bandwidth determined in step 52.

[0067] That is, the V I group of voltage vectors is evaluated first, and if the cost function of the VI group of voltage vectors is less than delta 2 , it indicates that applying this voltage vector will not cause the actual current to exceed the current control error band, and the voltage vector of the previous moment is continued to be used, and the switching state of the inverter does not need to be changed.

[0068] If the cost function of the VI I group of voltage vectors is greater than delta 2 , the voltage vectors in the set of {V I ∪V II} are evaluated, and the voltage vector with the minimum cost function is selected. 2 If the cost function of the voltage vector is less than delta 2 , the voltage vector is used as the final output.

[0069] If the cost function of the voltage vector is still greater than delta 2 , the voltage vectors in the set of {V I ∪V II ∪V III} are evaluated, and the voltage vector with the minimum cost function is selected as the final output.

[0070] Advantages

[0071] The adjustable switching frequency FCS-MPC permanent magnet synchronous motor current control method of the application retains the unique advantages of FCS-MPC compared with the prior art, overcomes the inherent defect of the non-fixed switching frequency, enables FCS-MPC to operate at a specified switching frequency in a wide speed range, and realizes zero-error tracking of the switching frequency reference value by introducing a PI type switching frequency controller, which effectively regulates the switching frequency on the basis of ensuring the current control performance of the permanent magnet motor. BRIEF DESCRIPTION OF DRAWINGS

[0072] Figure 1 The method sequence diagram of the application;

[0073] Figure 2 The overall control structure diagram of the application;

[0074] Figure 3 The experimental effect diagram of the application when the 0A step is changed to 10A at 0.8kHz switching frequency and 100rpm;

[0075] Figure 4The experimental effect diagram of the application at 500 rpm under 0A to 10A step at 1 kHz switching frequency;

[0076] Figure 5 The experimental effect diagram of the application at 300 rpm-10A under switching frequency step from 1 kHz to 3 kHz;

[0077] Figure 6 The experimental effect diagram of the application at 500 rpm-20A under switching frequency step from 1 kHz to 2 kHz;

[0078] Figure 7 The algorithm implementation flowchart involved in the application. DETAILED DESCRIPTION

[0079] In order to have a further understanding and recognition of the structural features and the achieved effects of the application, the following detailed description is given in combination with the preferred embodiments and the accompanying drawings:

[0080] As shown in the accompanying drawings, Figure 1 The current control method of the adjustable switching frequency FCS-MPC permanent magnet synchronous motor of the application comprises the following steps:

[0081] Firstly, the data of the driving system is collected: the DC bus voltage, three-phase current and rotor position data of the permanent magnet synchronous motor driving system are collected.

[0082] (1) Collect the DC bus voltage, three-phase current signal and rotor position signal of the permanent magnet synchronous motor driving system at time k.

[0083] (2) Transform the three-phase current signal to the synchronous speed rotating dq coordinate system oriented by the rotor magnetic field, which is expressed as:

[0084]

[0085] Wherein, i d , i q are the dq axis currents, i a , i b , i c are the three-phase currents in the three-phase stationary coordinate system, θ is the rotor position electrical angle, and sinθ, cosθ represent the sine and cosine values of the rotor position.

[0086] (3) Obtain eight switching states of the three-phase two-level inverter, and the different combinations of the upper arm switching states of the three-phase two-level inverter [Sa Sb Sc] are expressed as {000, 100, 110, 010, 011, 001, 101, 111}. In the synchronous speed rotating dq axis coordinate system, the corresponding dq axis voltage of each switching state is expressed as:

[0087]

[0088] where u d , u q represents the dq-axis voltage output of the inverter, Sa, Sb, Sc represents the switching state of the upper arm of the inverter, 1 represents turn-on, 0 represents turn-off, V dc is the DC bus voltage.

[0089] Second step, the construction of the discretization current prediction control model: according to the rotor position, coordinate transformation is carried out, based on the synchronous speed rotating dq coordinate system of the rotor field orientation and the collected three-phase current, the dq-axis current in the synchronous speed rotating dq coordinate system is obtained, the discretization current prediction control model is established, the eight basic voltage vectors of the two-level inverter are substituted into the discretization current prediction control model, the dq-axis predicted current of each voltage vector at k+2 time is obtained, the discrete switching characteristics of the inverter can be effectively utilized, and the dynamic response of the current can be effectively improved.

[0090] (1) Set the voltage equation of the permanent magnet synchronous motor in the synchronous speed rotating dq coordinate system oriented by the rotor magnetic field as:

[0091]

[0092] In the formula, i d , i q , u d , u q is the dq-axis current, voltage, L s is the stator inductance, R s is the stator resistance, ψ f is the permanent magnet flux linkage, ω r is the rotor electric angular velocity, represents the d-axis current differential, represents the q-axis current differential.

[0093] (2) The discrete current prediction model of the permanent magnet synchronous motor is obtained based on the Euler discretization method as follows:

[0094]

[0095] Where k is the kth sampling period, T s is the sampling period, i d (k+1), i q (k+1) represents the dq-axis current prediction value at k+1 time, ω r (k) represents the rotor electric angular velocity sampling value at k time, u d (k), u q (k) represents the dq-axis voltage output of the inverter at k time, L s is the stator inductance, R sψ is the stator resistance, f is the permanent magnet flux linkage.

[0096] (3) Substitute the eight basic voltage vectors of the three-phase two-level inverter into formula (4) respectively to obtain the dq-axis predicted current value corresponding to each voltage vector.

[0097] Third step, determine the cost function and delay compensation strategy: design the cost function, determine the performance evaluation standard of the finite control set model predictive control, and substitute the dq-axis current prediction value obtained in the second step into the designed cost function, that is, obtain the cost function size corresponding to each voltage vector.

[0098] The cost function can include multiple control objectives, has the advantages of simplicity and intuition, and the FCS-MPC controller can directly select the optimal voltage vector without going through the voltage modulation link through enumeration evaluation of the cost function. The delay compensation strategy can accurately compensate for the timing mismatch caused by the calculation time delay, effectively improving the current control performance.

[0099] (1) To compensate for the timing mismatch caused by the calculation time delay, the updated voltage vector at time k is used to estimate the current at time k+1, and the current estimation value at time k+1 is used as the starting point for the current prediction at time k+2, so as to accurately compensate for the digital control "lag" by controlling the "advance" one beat.

[0100] (2) The cost function is designed as the sum of the square of the dq-axis current tracking error at time (k+2), that is:

[0101]

[0102] wherein, is the dq-axis current reference value, i d (k+2), i q (k+2) represents the dq-axis current prediction value C lim at time k+2, which is used to prevent the motor from overcurrent, and is represented as:

[0103]

[0104] wherein, k is the kth sampling period, is the dq-axis current reference value, i d (k+2), i q (k+2) represents the dq-axis current prediction value I sm at time k+2, which represents the maximum allowable current of the permanent magnet synchronous motor;

[0105] (3) The dq-axis current prediction value corresponding to each basic voltage vector in the third step is substituted into formula (5) to obtain the cost function size corresponding to each basic voltage vector. The cost function effectively compensates for the time delay effect caused by system calculation time through the lead evaluation of the current at (k+2) time, and solves the timing mismatch problem.

[0106] Fourthly, the current error control band is calculated: the switching frequency of the limited control set model current prediction controller is calculated based on the sliding window method, a switching frequency controller is established, and the current error control band is determined.

[0107] This step innovatively establishes a real-time control loop of average switching frequency, and decomposes the coupled control of current control and switching frequency into two parts of switching frequency outer loop and current control inner loop through the bridge of current error band. The introduced switching frequency controller can realize errorless tracking of switching frequency command, overcoming the characteristics that the static error of traditional switching frequency control is difficult to eliminate. This step fully utilizes the advantages of feedback control system, has very small calculation load, and is easy to implement.

[0108] (1) Based on the idea of sliding window, the ratio of the total number of switching times of the upper arm of the inverter in a certain time window to the duration of the time window is expressed as the average switching frequency of the inverter,

[0109] The average switching frequency of the inverter at k time is expressed as:

[0110]

[0111] Where k is the kth sampling period, f sw (k) is the average switching frequency of the inverter calculated at k time, T w is the duration of the sliding window, n represents the time period number of the sliding window, S=[Sa, Sb, Sc] represents the switching state of the upper arm of the inverter, 1 represents on, 0 represents off, ‖‖1 represents the total number of switching times of the upper arm of the three-phase two-level inverter, S(k-i) is the switching state of the upper arm of the inverter at k-i time, and S(k-i-1) is the switching state of the upper arm of the inverter at k-i-1 time.

[0112] In order to eliminate the quantization noise in the calculation of switching frequency, the real-time calculation value of switching frequency is further processed through a low-pass filter, and the real-time feedback of switching frequency is finally:

[0113]

[0114] Where k is the kth sampling period, T s represents the control period, and α is the filter coefficient. c is the cut-off frequency of the first-order low-pass filter, and f sw(k) is the filtered inverter average switching frequency at kth moment, f swL (k) is the filtered inverter average switching frequency at kth moment, f swL (k-1) is the inverter average switching frequency calculated at k-1th moment.

[0115] (2) Feedback calculation based on formula (7)-(8) for real-time switching frequency, to establish switching frequency control loop,

[0116] The switching frequency controller adopts PI regulator, and the output of the PI regulator is current control error bandwidth. The switching frequency controller is represented as:

[0117]

[0118] Where k is the kth sampling period, T s represents the control period, and δ(k) is the current control error bandwidth at kth moment, is the reference value of the inverter average switching frequency at kth moment, f swL (k) is the filtered inverter average switching frequency at kth moment, k p and k i is the proportional gain and integral gain of the switching frequency controller, e f (k) is the inverter average switching frequency error at kth moment, e f (k-1) is the inverter average switching frequency error at k-1th moment;

[0119] The range of the output variable δ of the switching frequency controller is limited as:

[0120] δ min <δ(k)≤δ max (10)

[0121] Where k is the kth sampling period, δ(k) is the current control error bandwidth at kth moment, and δ min is determined by the maximum value of the inverter average switching frequency allowed by the operating working point, δ max is determined by the maximum value of the current ripple allowed by the operating working point.

[0122] (3) Real-time average switching frequency of the inverter is realized by adjusting the bandwidth of the current control error band;

[0123] When the inverter average switching frequency reference value is lower than the actual switching frequency, the positive error is input into the PI regulator, and the PI gradually increases the width of the current control error band to realize accurate tracking of the switching frequency reference value;

[0124] When the switching frequency reference value is higher than the actual switching frequency, the negative error is input into the PI regulator, and the PI regulator gradually reduces the width of the current control error band.

[0125] Under the regulation of the integrator, the actual switching frequency of the inverter realizes the zero-error tracking of the switching frequency reference value.

[0126] Fifth step, selecting the optimal voltage vector: according to the results of the cost function in the third step, the switching state with the cost function within the current error control width determined in the fourth step and the least switching times is selected as the optimal voltage vector. This step takes the current error band as a hard constraint condition, and realizes the quantitative control of the current performance. The candidate voltage vectors are sequentially evaluated according to different priorities, and the voltage vector that meets the current control performance and has the least switching times is evaluated with the least calculation load.

[0127] (1) According to the voltage vector applied at k+1 time, the candidate voltage vector at k+2 time is divided into three categories according to the switching times:

[0128]

[0129] Wherein, k is the kth sampling period, S(k+1) and S(k+2) represent the inverter switching state at k+1 and k+2 time, S=[Sa,Sb,Sc] represents the switching state of the upper bridge arm of the inverter, 1 represents on, 0 represents off, ‖‖1 represents the total switching times of the upper bridge arm of the three-phase two-level inverter, V I represents the switching state set without bridge arm switching, V II represents the switching state set of one bridge arm switching, V III represents the switching state set of two bridge arm switching;

[0130] If the voltage vector at (k+2) time continues to follow the voltage vector at the previous time, the switching state does not change, and the switching times is zero, and the voltage vector at the previous time is confirmed as V I group;

[0131] If a new voltage vector is applied at (k+2) time and only the switching state of one bridge arm changes, the vector is confirmed as V II group, and if a new voltage vector is applied at (k+2) time and the switching states of two bridge arms change, the vector is confirmed as V III group.

[0132] For example, if the voltage vector applied at (k+1) time is V5(101), the candidate voltage vector V5(101) at (k+2) time belongs to V I group, V1(001), V4(100), V7(111) belong to V II group, V3(011), V6(110) belong to V III group.

[0133] (2) Based on the candidate voltage vector classification results defined by formula (11), the selection method of the candidate voltage vector at time k+2 is expressed as follows:

[0134]

[0135] Where k is the kth sampling period, and V(k+2) represents the optimal voltage vector to be output at time k+2. optI V represents I The voltage vector with the minimum cost function, V optII Represents {V I ∪V II The voltage vector with the minimum cost function, V optIII Represents {V I ∪V II ∪V III The voltage vector with the minimum cost function, g[] represents the cost function, and δ is the current error control band width;

[0136] That is, to prioritize the evaluation of V I The voltage vector of group VI, if the cost function of the voltage vector of group VI is less than δ 2 If the voltage vector is applied, it means that applying the voltage vector will not cause the actual current to exceed the current control error band. The voltage vector from the previous moment will continue to be used, and the switching state of the inverter does not need to be changed.

[0137] If V I The cost function of the group voltage vector is greater than δ 2 Then for {V I ∪V II The voltage vectors within the set are evaluated, and the voltage vector with the minimum cost function is selected. If the cost function of this voltage vector is less than δ... 2 Then the voltage vector is taken as the final output;

[0138] If the cost function of the voltage vector is still greater than δ 2 Then for {V I ∪V II ∪V III The voltage vectors in the set are evaluated, and the voltage vector with the minimum cost function is selected as the final output. Current is the target of permanent magnet motor control, voltage is the control input, and the cost function is the performance function; the smaller the value, the better the current control. The idea behind predictive current control is to select the voltage vector that minimizes the cost function at future moments as the final output. The process of selecting the voltage vector is the process of current control; current control is implemented here, and the final output is performed in the sixth step.

[0139] The sixth step is voltage vector output: the inverter outputs the optimal voltage vector, generates the inverter driving signal according to the inverter switching state corresponding to the optimal voltage vector, completes the voltage vector update of the control period, and realizes the real-time tracking of the d-q axis current of the permanent magnet synchronous motor.

[0140] The specific implementation flow chart of the control method is shown in Figure 7 The control structure diagram of the overall scheme is shown in Figure 2

[0141] Table 1 Permanent magnet synchronous motor nominal parameters

[0142]

[0143] The controlled motor nominal parameters are shown in Table 1, the inverter is powered by a programmable DC power supply, the output voltage is set to 48V, and the controlled motor works in torque control mode and is powered by a three-phase two-level inverter. The dSPACE / DS1007 is used as the measured motor controller on the controlled motor side, a pair of rotating transformers is used to detect the rotor position of the measured motor, and the motor phase current sensor is LEM LA25-P. The current sensor samples the three-phase current value in real time, and the sampling time is set to 50us, and the inverter dead time is set to 2us. The dynamometer is a 2.2kW three-phase asynchronous motor, which works in speed control mode and is driven by an ACS 800 frequency converter.

[0144] Figure 3 The measured waveform of the q-axis current command changing from 0A to 10A when the permanent magnet motor runs at 100rpm and the switching frequency reference value is set to 0.8kH.

[0145] Figure 4 The measured waveform of the q-axis current command changing from 0A to 10A when the permanent magnet motor runs at 500rpm and the switching frequency reference value is set to 1.0kH.

[0146] Figure 5 The measured waveform of the q-axis current command changing from 0A to 10A when the permanent magnet motor runs at 300rpm and the switching frequency reference value is set to 1.0kH.

[0147] Figure 6 The measured waveform of the q-axis current command changing from 0A to 10A when the permanent magnet motor runs at 500rpm and the switching frequency reference value is set to 1.0kH.

[0148] From Figure 3 and Figure 4 ​It can be seen from the figures that when the dq-axis current command changes, the FCS-MPC current controller can achieve accurate tracking of the current command, and the switching frequency can be stably maintained at the reference value and will not change with the change of the operating point.

[0149] From Figure 5 and Figure 6 It can be seen that when the dq-axis current command remains constant, the actual switching frequency can accurately track the step change of the reference value, and there is no steady-state error, which shows that the strategy can effectively control the switching frequency on the basis of ensuring the current control performance.

[0150] Comprehensive test waveforms, the application can simultaneously achieve accurate tracking of the current command and the switching frequency command, and operate at a fixed switching frequency in a wide speed regulation range, overcoming the inherent problem of the switching frequency of the limited control set model predictive control changing with the operating point. In addition, the application does not require a weight factor, and the complex weight factor setting process is saved.

[0151] The basic principles, main features and advantages of the application are shown and described above. Those skilled in the art should understand that the application is not limited by the above examples, and the above examples and descriptions in the specification are only the principles of the application. Without departing from the spirit and scope of the application, various changes and improvements can be made to the application, and these changes and improvements all fall within the scope of the claimed application. The scope of protection claimed by the application is defined by the appended claims and their equivalents.

Claims

1. A current control method for a permanent magnet synchronous motor with adjustable switching frequency FCS-MPC, characterized in that, Includes the following steps: 11) Acquisition of drive system data: Acquire DC bus voltage, three-phase current, and rotor position data of the permanent magnet synchronous motor drive system; 12) Construction of Discrete Current Predictive Control Model: Based on the rotor position, coordinate transformation is performed. Based on the synchronous speed rotating dq coordinate system oriented by the rotor magnetic field and the collected three-phase current, the dq axis current in the synchronous speed rotating dq coordinate system is obtained. A discrete current predictive control model is established. The eight basic voltage vectors of the two-level inverter are substituted into the discrete current predictive control model to obtain the dq axis predicted current of each voltage vector at time k+2. 13) Determine the cost function and delay compensation strategy: Design the cost function, determine the performance evaluation criteria of the finite control set model predictive control, and substitute the dq axis current prediction values ​​obtained in step 12) into the designed cost function to obtain the size of the cost function corresponding to each voltage vector; 14) Calculate the current error control band: Calculate the switching frequency of the current prediction controller of the finite control set model based on the sliding window method, establish the switching frequency controller, and determine the current error control band. 51) Based on the sliding window concept, the ratio of the total number of times the inverter's upper bridge arm switches are switched within a certain time window to the duration of the time window is expressed as the inverter's average switching frequency. The average switching frequency of the inverter at time k is expressed as: Where k is the kth sampling period, f sw (k) is the average switching frequency of the inverter calculated at time k, T w Let n be the duration of the sliding window, n be the number of time cycles of the sliding window, S = [Sa, Sb, Sc] be the switching state of the upper arm of the inverter, 1 represents on, 0 represents off, ||||1 represents the total number of switching times of the upper arm of the three-phase two-level inverter, S(ki) is the switching state of the upper arm of the inverter at time ki, and S(ki-1) is the switching state of the upper arm of the inverter at time ki-1. To eliminate quantization noise in the switching frequency calculation, the real-time calculated switching frequency value is further processed by a low-pass filter, and the final real-time feedback of the switching frequency is: Where k is the kth sampling period, T s This represents the control period, α is the filter coefficient, and ω is... c f is the cutoff frequency of a first-order low-pass filter. sw (k) is the average switching frequency of the inverter calculated at time k, f swL (k) represents the average switching frequency of the inverter after filtering at time k, f swL (k-1) is the average switching frequency of the inverter calculated at time k-1; 52) Based on the feedback calculation of the real-time switching frequency using formulas (7)-(8), a switching frequency control loop is established. The switching frequency controller uses a PI regulator. The output of the PI regulator is the current control error band. The switching frequency controller is represented as follows: Where k is the kth sampling period, T s This represents the control period, and δ(k) is the current control error band width at time k. f is a reference value for the average switching frequency of the inverter at time k. swL (k) represents the average switching frequency of the inverter after filtering at time k, where k is the frequency of the inverter. p With k i e represents the proportional gain and integral gain of the switching frequency controller. f (k) represents the average switching frequency error of the inverter at time k, e f (k-1) represents the average switching frequency error of the inverter at time k-1; The range of the output variable δ of the switching frequency controller is limited as follows: d min <δ(k)≤δ max (10) Where k is the kth sampling period, δ(k) is the current control error band width at time k, and δ min The value of the average switching frequency of the inverter, δ, is determined by the maximum value allowed by the operating point. max The maximum allowable current ripple at the operating point; 15) Select the optimal voltage vector: Based on the cost function result of step 13), select the switching state with the least number of switching operations that is within the current error control width determined in step 14) and has the lowest cost function as the optimal voltage vector. 16) Voltage Vector Output: The inverter outputs the optimal voltage vector, and generates an inverter drive signal based on the inverter switching state corresponding to the optimal voltage vector to complete the voltage vector update of the control cycle, thereby realizing the real-time tracking of the dq axis current by the permanent magnet synchronous motor.

2. The method for current control of a permanent magnet synchronous motor with adjustable switching frequency FCS-MPC according to claim 1, characterized in that, The acquisition of drive system data includes the following steps: 21) Collect the DC bus voltage, three-phase current signal, and rotor position signal of the permanent magnet synchronous motor drive system at time k; 22) The three-phase current signal is transformed into a synchronously rotating dq coordinate system oriented by the rotor magnetic field, and expressed as: Among them, i d i q Let i be the dq-axis current. a i b i c Let θ be the three-phase current in a three-phase stationary coordinate system, θ be the electrical angle of the rotor position, and sinθ and cosθ be the sine and cosine values ​​of the rotor position, respectively. 23) Obtain the eight switching states of the three-phase two-level inverter. Different combinations of the upper arm switching states of the three-phase two-level inverter [Sa Sb Sc] are represented as {000, 100, 110, 010, 011, 001, 101, 111}. In the synchronous speed rotating dq axis coordinate system, the dq axis voltage corresponding to each switching state is represented as: Among them, u d ,u q This represents the dq-axis voltage output by the inverter. Sa, Sb, and Sc represent the switching states of the upper bridge arms of the inverter, where 1 indicates on and 0 indicates off. dc This is the DC bus voltage.

3. The method for current control of a permanent magnet synchronous motor with adjustable switching frequency FCS-MPC according to claim 2, characterized in that, The construction of the discretized current prediction control model includes the following steps: 31) In a synchronously rotating dq coordinate system oriented by the rotor magnetic field, the voltage equation of the permanent magnet synchronous motor is: In the formula, i d i q u d u q For dq axis current and voltage, L s For stator inductance, R s For stator resistance, ψ f For permanent magnet flux linkage, ω r The rotor's electric angular velocity, This represents the differential of the d-axis current. This represents the q-axis current differential; 32) The discrete current prediction model for permanent magnet synchronous motors obtained based on the Euler discretization method is as follows: Where k is the kth sampling period, T s For the sampling period, i d (k+1),i q (k+1) represents the predicted dq-axis current at time k+1, ω r (k) represents the sampled value of the rotor's electric angular velocity at time k, u d (k),u q (k) represents the dq-axis voltage output value of the inverter at time k, L s For stator inductance, R s For stator resistance, ψ f For permanent magnet flux linkage; 33) Substitute the eight basic voltage vectors of the three-phase two-level inverter into formula (4) to obtain the predicted dq axis current value corresponding to each basic voltage vector.

4. The method for current control of a permanent magnet synchronous motor with adjustable switching frequency FCS-MPC according to claim 3, characterized in that, The determination of the cost function and delay compensation strategy includes the following steps: 41) To compensate for the timing mismatch caused by the calculation time, the current at time k+1 is estimated by using the voltage vector updated at time k and the sampled signal at time k. The estimated current at time k+1 is used as the starting point for the current prediction at time k+2. The current at time k+2 is predicted, thereby compensating for the "lagging" of digital control by "leading" the control. 42) The design cost function is the sum of squared dq current tracking errors g at time k+2, which is: in, i is the reference value for the dq-axis current. d (k+2), i q (k+2) represents the predicted dq-axis current at time k+2, C lim This is used to prevent overcurrent in the motor, and is represented as: Where k is the kth sampling period, i is the reference value for the dq-axis current. d (k+2), i q (k+2) represents the predicted dq-axis current value at time k+2, I sm This indicates the maximum allowable current of the permanent magnet synchronous motor; 43) Substitute the predicted dq-axis current value corresponding to each basic voltage vector in step 33) into formula (5) to obtain the cost function size corresponding to each basic voltage vector.

5. The method for current control of a permanent magnet synchronous motor with adjustable switching frequency FCS-MPC according to claim 4, characterized in that, The calculation of the current error control band includes the following steps: The real-time average switching frequency of the inverter is achieved by adjusting the bandwidth of the current control error band; When the average switching frequency reference value of the inverter is lower than the actual switching frequency, a positive error is input to the PI regulator, and the PI gradually increases the width of the current control error band to achieve accurate tracking of the switching frequency reference value. When the reference value of the switching frequency is higher than the actual switching frequency, a negative error is input to the PI regulator, and the PI regulator gradually reduces the width of the current control error band. Under the regulation of the integrator, the actual switching frequency of the inverter achieves zero steady-state error tracking of the switching frequency reference value.

6. The method for current control of a permanent magnet synchronous motor with adjustable switching frequency FCS-MPC according to claim 4, characterized in that, The selection of the optimal voltage vector includes the following steps: 61) Based on the voltage vector applied at time k+1, the candidate voltage vectors at time k+2 are divided into three categories according to the number of switching operations: Where k is the kth sampling period, S(k+1) and S(k+2) represent the inverter switching states at times k+1 and k+2, respectively, S = [Sa, Sb, Sc] represents the switching state of the inverter upper arm, 1 represents on, 0 represents off, and ||||1 represents the total number of switching operations of the upper arm of the three-phase two-level inverter. I V represents the set of switch states for switching without bridge arm switching states. II V represents the set of switch states for a bridge arm switch state transition. III The set of switch states representing the switching state transition between the two bridge arms; If the voltage vector from the previous time step is used at time (k+2), the switch state remains unchanged, the number of switch switching operations is zero, and the voltage vector from the previous time step is confirmed as V. I Group; If a new voltage vector is applied at time (k+2) and the switching state of only one bridge arm changes, then this vector is identified as V. II If a new voltage vector is applied at time (k+2) and the switching states of two bridge arms change, then this vector is identified as V. III Group; 62) Based on the candidate voltage vector classification results defined in formula (11), the selection method of the candidate voltage vector at time k+2 is expressed as follows: Where k is the kth sampling period, and V(k+2) represents the optimal voltage vector to be output at time k+2. optI V represents I The voltage vector with the minimum cost function, V optII Represents {V I ∪V II The voltage vector with the minimum cost function, V optIII Represents {V I ∪V II ∪V III The voltage vector with the minimum cost function, g[] represents the cost function obtained in step 43), and δ is the current error control band width determined in step 52; That is, to prioritize the evaluation of V I The voltage vector of group VI, if the cost function of the voltage vector of group VI is less than δ 2 If the voltage vector is applied, it means that applying the voltage vector will not cause the actual current to exceed the current control error band. The voltage vector from the previous moment will continue to be used, and the switching state of the inverter does not need to be changed. If V I The cost function of the group voltage vector is greater than δ 2 Then for {V I ∪V II The voltage vectors within the set are evaluated, and the voltage vector with the minimum cost function is selected. If the cost function of this voltage vector is less than δ... 2 Then the voltage vector is taken as the final output; If the cost function of the voltage vector is still greater than δ 2 Then for {V I ∪V II ∪V III The voltage vectors of the set are evaluated, and the voltage vector with the minimum cost function is selected as the final output.

Images