Switched reluctance motor model prediction control method based on optimized sector division
By optimizing the model predictive control method based on sector partitioning and candidate voltage vector selection table, the problems of large computational load and torque ripple in switched reluctance motors are solved, achieving more efficient motor control and dynamic response.
Patent Information
- Application Number
- CN202411014343.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-15
AI Technical Summary
Existing model predictive control methods for switched reluctance motors involve large computational loads, which negatively impacts the dynamic response capability of the motor control system and results in significant torque ripple.
A model predictive control method based on optimized sector partitioning is adopted. By reconstructing the sector and candidate voltage vector selection table, the amount of computation is reduced. The switching signal is optimized through a cost function minimization module, and the stability of the position sensor is improved by combining it with an optical encoder.
It effectively reduces motor torque ripple, enhances motor torque dynamic response performance, simplifies the calculation process, and improves system stability and response speed.
Smart Images

Figure CN118971710A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of motor control, specifically to the field of switched reluctance motors and their control, and in particular to a model predictive control method for switched reluctance motors. Background Art
[0002] The switched reluctance motor has the advantages of simple structure, no need for rare metals, high reliability, etc. Its rotor has no windings or permanent magnets, each phase winding has an independent drive element, and the damage of a single phase does not affect the operation of other phases. Therefore, it has extremely high reliability and is widely used in industrial robots, boring machines, household appliances, textile machinery and other fields. However, the torque ripple caused by the double salient pole structure and nonlinear electromagnetic characteristics of the switched reluctance motor limits its promotion and application in industry. Therefore, how to reduce the torque ripple of the switched reluctance motor has become an urgent problem to be solved in its electrical transmission and control.
[0003] The most commonly used control method for driving a switched reluctance motor is current chopping control, but the traditional current chopping control method usually reduces the operating efficiency of the system. At present, model predictive control has a good ability to handle time-varying and nonlinear problems, and is an excellent control scheme for driving a switched reluctance motor. For example, in the literature "L.Ge, J.Zhong, Q.Cheng, Z.Fan, S.Song and RWDe Doncker, Model Predictive Control of SwitchedReluctance Machines for Suppressing Torque and Source Current Ripples UnderBus Voltage Fluctuation, IEEE Transactions on Industrial Electronics, vol.70, no.11, pp.11013-11021, Nov.2023.", the technology disclosed therein is to apply model predictive control to a switched reluctance motor drive system, predict the motor torque and inverter current under all possible switching states, and construct a cost function to suppress torque pulsation and source current pulsation; however, this method needs to consider too many switching states, resulting in a large amount of calculation and a long calculation time, which affects the dynamic response capability of the motor control system. The document with Chinese patent application number 202310900434.6 discloses a robust finite set model predictive flux control method for a permanent magnet synchronous motor. According to the inverter voltage vector output characteristics of the model-predicted flux control, the current is sampled twice within a sampling period to eliminate the dead zone in the identification equation, and the flux prediction model is improved using a generalized proportional-integral observer, thereby further improving the robustness. However, the control method has a large computational burden, which will cause the motor control system to respond untimely. Summary of the invention
[0004] The purpose of the present invention is to address the above problems and propose a model predictive control method for a switched reluctance motor based on optimized sector division to alleviate the problems of large system calculation amount and heavy calculation burden, enhance the dynamic response performance of the motor torque, and effectively reduce the motor torque pulsation.
[0005] To achieve the above object, the technical solution adopted by the model predictive control method of a switched reluctance motor based on optimized sector division of the present invention includes the following steps:
[0006] Step 1): The flux calculation module receives the phase current i output by the voltage and current sampling module P and phase voltage u P , calculate the phase flux ψ PThe rotor position angle prediction module receives the actual rotor position angle θ output by the position sensor module. P The predicted phase rotor position angle θ is obtained by comparing the actual motor speed ω output by the speed calculation module Pest And the predicted A phase rotor position angle θ taking phase A as an example Aest ;
[0007] Step 2): an asymmetric half-bridge power converter is used to control the switched reluctance motor. When the stator winding is in the excitation state, both the upper and lower switch tubes are turned on, and the state is 1; when the stator winding is in the demagnetization state, both the upper and lower switch tubes are turned off, and the state is -1; when the stator winding is in the zero-voltage freewheeling state, the state is 0;
[0008] The rotor sector determination module receives the predicted A-phase rotor position angle θ Aest , according to the predicted A-phase rotor position angle θ Aest The sector is judged in the interval, when 0°<θ Aest ≤17°, phase A adopts switch state 1, 0 or -1; when 17°<θ Aest ≤22.5°, phase A adopts switch state 0 or -1; when 22.5°<θ Aest When ≤45°, phase A adopts switch state -1 and outputs rotor sector signal S x ;
[0009] Step 3): The phase voltage estimation module receives the bus voltage u dc , winding inductance voltage drop u r , switch tube voltage drop u S , freewheeling diode voltage drop u D And the rotor sector signal S x , get the estimated phase voltage u Pest :
[0010]
[0011] Step 4): The flux prediction module receives the phase flux ψ P , phase current i P And all estimated phase voltages u Pest , predict the predicted phase flux ψ Pest :The lookup table prediction module receives the predicted phase flux ψ Pest and predicted phase rotor position angle θ Pest According to the preset current-flux-rotor position i-ψ-θ three-dimensional model table and torque-current-rotor position Ti-θ three-dimensional model table, the predicted phase torque T is obtained by querying Pest ;
[0012] Step 5): The cost function minimization module receives the reference torque T' and the predicted phase torque T Pest , calculate the cost function, m is the number of motor phases, and select the switching signal that minimizes the cost function G as the input of the power converter module to control the switched reluctance motor.
[0013] Furthermore, the predicted phase rotor position angle θ Pest =θ P +T s ω, T s is the sampling step length.
[0014] Furthermore, the predicted phase flux ψ Pest =(u Pest -R P i P )T s +ψ P , R P is the phase resistance, T s is the sampling step length.
[0015] Furthermore, the cost function m is the number of motor phases.
[0016] The beneficial effects of the present invention are:
[0017] (1) The model prediction module in the present invention adopts a reconstructed sector and a candidate voltage vector selection table, which effectively solves the problem of large amount of calculation and heavy calculation burden of the model prediction control system, and finally outputs a switching signal, thereby enhancing the dynamic response performance of the motor torque.
[0018] (2) The cost function of the model prediction module in the present invention is composed only of the reference torque and the predicted phase torque, without the influence of other parameter variables, and can effectively reduce the motor torque pulsation.
[0019] (3) The switched reluctance motor signals required in the present invention are easy to obtain and measure in practical applications. The position sensor module has a separate photoelectric encoder, which can improve the stability of the switched reluctance motor operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to explain the present invention more clearly, the following is a brief introduction to the drawings required for describing the specific embodiments:
[0021] Figure 1 A structural block diagram of a switched reluctance motor model predictive control system based on optimized sector division used to implement the control method of the present invention;
[0022] Figure 2 for Figure 1 Schematic diagram of the composition of the model prediction module in;
[0023] Figure 3 for Figure 2 A schematic diagram of the structure of the lookup table module in FIG.
[0024] Figure 4 This is a schematic diagram of sector division of the three-phase 12 / 8 switched reluctance motor proposed in the present invention;
[0025] Figure 5 The corresponding candidate voltage vector table after the optimized sector division proposed by the present invention;
[0026] In the figure: 1. switched reluctance motor; 2. position sensor module; 3. speed calculation module; 4. speed setting module; 5. PI controller module; 6. voltage and current sampling module; 7. flux calculation module; 8. model prediction control module; 9. power converter module; 81. rotor position angle prediction module; 82. rotor sector judgment module; 83. phase voltage estimation module; 84. flux prediction module; 85. lookup table prediction module; 86. cost function minimization module; 851. current prediction module; 852. torque prediction module. DETAILED DESCRIPTION
[0027] like Figure 1 The switch reluctance motor model predictive control system shown is connected to the switch reluctance motor 1, and the switch reluctance motor 1 is connected to the load. The position sensor module 2 is used to detect the actual rotor position angle θ of the switch reluctance motor 1. P , the actual rotor position angle θ P The actual rotor position angle θ is input to the speed calculation module 3 and the model prediction control module 8 respectively. P The actual motor speed ω is calculated by the speed calculation module 3 ; the actual motor speed ω is also input into the model prediction control module 8 .
[0028] The speed setting module 4 sets the motor reference speed ω', compares the reference speed ω' with the actual speed ω, and the difference between the two is the speed error ω err , speed error ω err The reference torque T' is input to the PI controller module 5 and outputted through PI control. The reference torque T' is also input to the model prediction control module 8.
[0029] The voltage and current sampling module 6 is used to collect the phase current i of the motor P and phase voltage u P ; The phase current i P and phase voltage u P Input flux calculation module 7, the phase current i P The input model prediction control module 8, the flux calculation module 7 calculates the phase current i P and phase voltage u P Processing, the output is phase flux ψP , and the phase flux ψ P Input into the model predictive control module 8.
[0030] The output of the model predictive control module 8 is the phase switch signal ST P , phase switch signal ST P The power converter module 9 is input to the power converter module 9, and the power converter module 9 is based on the phase switch signal ST P Control the state of the motor phase switch device to adjust the motor torque output. At the same time, the power converter module 9 converts the bus voltage u dc , Winding inductance voltage drop u r , Switching tube voltage drop u S And the freewheeling diode voltage drop u D Feedback to the model predictive control module 8.
[0031] The inputs of the model predictive control module 8 are the reference torque T', the actual motor speed ω, and the actual rotor position angle θ P , phase flux ψ P , phase current i P , bus voltage u dc , winding inductance voltage drop u r , switch tube voltage drop u S And the freewheeling diode voltage drop u D , the output is the phase switch signal ST P .
[0032] like Figure 2 The model prediction control module 8 shown is composed of a rotor position angle prediction module 81, a rotor sector judgment module 82, a phase voltage estimation module 83, a flux prediction module 84, a lookup table prediction module 85 and a cost function minimization module 86. Among them, the output of the rotor position angle prediction module 81 is connected to the rotor sector judgment module 82 and the lookup table prediction module 85 respectively, the rotor sector judgment module 82, the phase voltage estimation module 83, the flux prediction module 84, and the lookup table prediction module 85 are connected in sequence, and the output end of the lookup table prediction module 85 is connected to the cost function minimization module 86, and the output of the cost function minimization module 86 is also the output of the model prediction control module 8. The lookup table prediction module 85 has a pre-established three-dimensional model table of the switched reluctance motor current-flux-rotor position i-ψ-θ and a three-dimensional model table of torque-current-rotor position Ti-θ.
[0033] like Figure 3 The lookup table prediction module 85 shown in the figure is composed of a torque prediction module 852 and a current prediction module 851. The current prediction module 851 receives the predicted phase flux ψ output by the flux prediction module 84. Pest The predicted phase rotor position angle θ output by the rotor position angle prediction module 81Pest , based on the current-flux-rotor position i-ψ-θ three-dimensional model table data query, obtain the predicted phase current i Pest , as the output of the current prediction module 851. The torque prediction module 852 converts the predicted phase current i output by the current prediction module 851 into Pest The predicted phase rotor position angle θ output by the rotor position angle prediction module 81 Pest As input, the predicted phase torque T is obtained based on the motor torque-current-rotor position Ti-θ three-dimensional model table data query. Pest , as the output of the torque prediction module 852.
[0034] First, it is necessary to perform offline experimental measurements and calculate the model information of the switched reluctance motor 1 for the implementation of the control strategy. The information of the switched reluctance motor 1 that needs to be pre-acquired is as follows:
[0035] (1) A three-dimensional model table of current-flux-rotor position i-ψ-θ of the switched reluctance motor 1, built into the model prediction module 8, specifically a current prediction module 851 built into the lookup table prediction module 85, for predicting the phase flux value ψ Pest and the rotor position angle θ P Look up the table to obtain the predicted value of phase current i Pest .
[0036] (2) A torque-current-rotor position Ti-θ three-dimensional model table of the switched reluctance motor 1, which is built into the model prediction module 8, specifically a torque prediction module 852 built into the lookup table prediction module 85, for predicting the phase current value i Pest and the rotor position angle θ P Look up the table to obtain the phase torque prediction value T Pest .
[0037] The torque balance method and Fourier series are used to obtain the current-flux-rotor position i-ψ-θ three-dimensional model table and the torque-current-rotor position Ti-θ three-dimensional model table of the switched reluctance motor 1. When the switched reluctance motor 1 is in a single-phase or multi-phase excitation state, the total torque at several special rotor position angles is zero. At this time, the motor rotor remains stationary, and the rotor position angle is defined as the torque balance position angle.
[0038] Taking phase A as an example, if phases BC are excited at the same time, phase A will be in a non-aligned position. When the motor rotor is balanced, the rotor position angle of phase A is θ Ar =0, defined as the first special position;
[0039] When the motor rotor is balanced, the A phase rotor position angle θ is Ar =τ r / 2=π / N r=π / 8, defined as the second special position, where N r is the number of rotor poles, τ r is the rotor mechanical angle period, for the switched reluctance motor 1, N r =8,τ r =2π / N r .
[0040] When the C phase is excited alone, the rotor position angle of the A phase when the motor rotor is balanced is θ Ar =τ r / 6, defined as the third special position;
[0041] When the AC phases are excited at the same time, the rotor position angle θ of phase A is when the motor rotor is balanced. Ar =τ r / 3, defined as the fourth special position.
[0042] The phase voltage equation of the switched reluctance motor 1 can be expressed as:
[0043]
[0044] In the formula, u p is the phase voltage value, R p is the phase resistance value, i P is the phase current value, ψ P is the phase flux value, P=A,B,C.
[0045] By transforming formula (1), the phase flux expression can be obtained:
[0046]
[0047] In the formula, ψ A is the A-phase flux, which is related to the rotor position angle θ and the A-phase current i A Related, T is the flux integration time, u A is the phase A voltage, R A is the A phase resistance.
[0048] After collecting the phase voltage and phase current of the switched reluctance motor 1, the phase flux ψ of the switched reluctance motor 1 can be calculated by equation (2): A Multiple groups of equal phase currents of different magnitudes are injected into the stator winding to obtain the flux characteristics at this special position. The Fourier series expression is used to describe the inductance characteristics of phase A of the motor:
[0049]
[0050] Where, L A is the inductance of phase A, which is related to the rotor position angle θ and the current i of phase A A Related, L n (iA ) are the coefficients of the Fourier series.
[0051] The phase flux equation of the switched reluctance motor 1 can be expressed as:
[0052] ψ P =L P ·i P (4)
[0053] Combining formula (3) and formula (4), the flux characteristics of motor phase A can be described as:
[0054]
[0055] Where N r is the number of rotor poles of the switched reluctance motor, h n (i A ) is the coefficient of the Fourier series. In order to ensure the accuracy of the nonlinear model while reducing the amount of calculation, the fourth-order Fourier series expansion is used to represent the motor A phase flux characteristics:
[0056]
[0057] Linear approximation τ r / 6 and τ r / 3 flux characteristics at the rotor position, A phase rotor position angle θ Ar =τ r The magnetic flux characteristics at / 4 can be expressed as:
[0058] ψ A (τ r / 4, i A )=[ψ A (τ r / 6, i A )+ψ A (τ r / 3, i A )] / 2 (7)
[0059] In the formula, ψ A (τ r / 4,i A ) is the rotor position angle θ Ar =τ r The phase flux value at / 4 defines this rotor position as the fifth special position.
[0060] Therefore, the relationship between the Fourier series coefficients at five special positions and their flux characteristics can be derived from formula (6) as follows:
[0061]
[0062] The Fourier series coefficient h0(i A ), h1(i A ), h2(i A ), h3(i A ), h4(i A ), the unknown variables in formula (6) are only the rotor position angle θ and the phase current i A , by substituting different values of θ into the arithmetic progression and i A The magnetic flux ψ at this position can be obtained by A For example, if θ = 1.35°, i A =2.5A Substitute into formula (6), and we can get ψ A =0.00221. Similarly, the phase flux values at different rotor positions and currents shown in Table 1 can be calculated. The vertical variable in Table 1 is the phase rotor position θ, the horizontal variable is the phase current i, and the variable in the cell is the phase flux ψ:
[0063] Table 1 Switched reluctance motor phase current-flux-rotor position i-ψ-θ three-dimensional model table
[0064]
[0065] According to the virtual displacement principle, the magnetic energy of phase A can be expressed as:
[0066]
[0067] Where W Aco is the magnetic energy of phase A.
[0068] Then the instantaneous torque of phase A can be expressed as:
[0069]
[0070] Based on the obtained fourth-order Fourier series flux characteristics, substituting equation (6) into equation (10), the torque characteristic expression of phase A can be obtained:
[0071] T A =l1(i A )sin(N r θ)+l2(i A )sin(N r θ)+l3(i A )sin(N r θ)+l4(i A )sin(N r θ) (11)
[0072] In the formula, Since h1(i A ), h2(i A), h3(i A ), h4(i A ) are known quantities to be obtained, so l k (i A )(k=1,2,3,4) are also known variables. The only unknown variables in formula (11) are the rotor position angle θ and the phase current i A , by substituting different values of θ into the arithmetic progression and i A The torque T at this position can be obtained by A For example, if θ = 1.35°, i A =2.5A Substitute into formula (11), and we can get T A =0.00183. Similarly, the phase torque values at different rotor positions and currents shown in Table 2 can be calculated. The vertical variable in Table 2 is the phase rotor position θ, the horizontal variable is the phase current i, and the variable in the cell is the phase torque T:
[0073] Table 2 Switched reluctance motor phase torque-current-rotor position Ti-θ three-dimensional model table
[0074]
[0075] The flux linkage ψ described by equation (6) A The torque T described by the characteristics and formula (11) A The characteristics together constitute the electromagnetic characteristics of the switched reluctance motor 1, and the three-dimensional model table of each phase current-flux-rotor position i-ψ-θ and the three-dimensional model table of each phase torque-current-rotor position Ti-θ of the switched reluctance motor 1 can be constructed therefrom, as shown in Table 1 and Table 2.
[0076] Before control, Table 1 needs to be pre-built into the current prediction module 851 in the model prediction module 8, and Table 2 needs to be pre-built into the torque prediction module 852 in the model prediction module 8. The specific control method is:
[0077] The position sensor module 2 receives the photoelectric signal of the switched reluctance motor 1 detected by the position encoder as input, and outputs the actual rotor position angle θ P .
[0078] The speed calculation module 3 converts the rotor position angle θ output by the position sensor module 2 into P As input, the differential of the rotor position signal is calculated to obtain the actual motor speed ω, which is used as the output of the speed calculation module 3.
[0079] The speed setting module 4 generates a reference speed ω′ according to a preset working condition as an output of the speed setting module 4 .
[0080] Compare the actual motor speed ω with the reference speed ω' to get the speed error ω err, as the input of PI controller module 5, the speed error ω err After PI regulation, the motor reference torque T' is obtained and serves as the output of the PI controller module 5.
[0081] The voltage and current sampling module 6 converts the collected electrical signals of the switched reluctance motor 1 into voltage and current signals required for calculation, and outputs the phase current i of the switched reluctance motor 1. P , phase voltage u P ;
[0082] The flux calculation module 7 receives the phase current i output by the voltage and current sampling module 6 P , phase voltage u P As input, the phase flux is calculated by the following formula:
[0083]
[0084] In the formula, u P is the motor phase voltage, R P is the phase resistance, i P is the motor phase current, ψ P is the phase flux, as the output of the flux calculation module 7, P = A, B, C, T is the flux integration time.
[0085] The switch reluctance motor 1 control adopts an asymmetric half-bridge power converter with three working states. When the stator winding is in the excitation state, the upper and lower switch tubes are both turned on, and the excitation current flows through the stator winding to generate a magnetic field to drive the rotor to rotate. This state is defined as "1"; when the stator winding is in the demagnetization state, the upper and lower switch tubes are both turned off. This state can quickly eliminate the residual current in the circuit and avoid the continuous effect of negative torque. This state is defined as "-1"; when the stator winding is in the zero-voltage freewheeling state, the winding current flows through the single switch tube and the single diode for freewheeling, and this state is defined as "0".
[0086] The model predictive control module 8 receives the reference torque T', the motor speed ω, and the rotor position angle θ P , phase flux ψ P , phase current i P , bus voltage u dc , winding inductance voltage drop u r , switch tube voltage drop u S And the freewheeling diode voltage drop u D Signals are used as input for computational processing.
[0087] like Figure 2 As shown, the rotor position angle prediction module 81 receives the phase rotor position angle θ output by the position sensor module 2. P The actual motor speed ω output by the speed calculation module 3 is used to predict the phase rotor position angle θPest It can be calculated by the following formula:
[0088] θ Pest =θ P +T s ω (13)
[0089] In the formula, θ Pest To predict the phase rotor position angle, P = A, B, C, take phase A as an example, when P = A, predict the phase A rotor position angle θ Aest =θ A +T s ω, the B-phase rotor position angle and the C-phase rotor position angle can be obtained by the same method. Predicted phase rotor position angle θ Pest And predict the A-phase rotor position angle θ Aest As the output of the rotor position angle prediction module 81, the predicted A-phase rotor position angle is output to the rotor sector judgment module 82, and the predicted phase rotor position angle θ Pest Output to the lookup table prediction module 85.
[0090] The rotor sector determination module 82 receives the predicted A-phase rotor position angle θ output by the rotor position angle prediction module 81 Aest , the sector division follows the torque distribution rule, the phase close to the motor stator and rotor misalignment position produces more torque, and the phase close to the motor stator and rotor alignment position produces less torque. Therefore, near the stator and rotor misalignment position, that is, 0°<θ Aest ≤17°, phase A adopts switch state 1, 0 or -1; near the stator and rotor alignment position, that is, 17°<θ Aest ≤22.5°, phase A adopts switch state 0 or -1; when the rotor enters the negative torque generation area, that is, 22.5°<θ Aest When ≤45°, phase A adopts switch state -1. In summary, the optimization results of sector division are as follows: Figure 4 and Figure 5 As shown in the figure, the number of candidate voltage vectors for model prediction control at the same time is reduced, thereby effectively shortening the calculation time required for control and reducing the calculation burden of the system. Aest The sector is judged in the interval and the rotor sector signal S is output. x , as the output of the rotor sector judgment module 82.
[0091] The phase voltage estimation module 83 receives the bus voltage u dc , winding inductance voltage drop u r , switch tube voltage drop u S , freewheeling diode voltage drop u D And the rotor sector signal S output by the rotor sector judgment module 82 x As input, the estimated phase voltage u is obtained byPest :
[0092]
[0093] In the formula, u Pest To estimate the phase voltage, as the output of the phase voltage estimation module 83.
[0094] For example, taking phase A as an example, when predicting the rotor position angle θ of phase A Aest =20°, the rotor sector determination module 82 outputs the rotor sector signal S x is S5, and there are 6 candidate voltage vectors in total. A ,S B ,S C )" are "(0,-1,1)", "(0,-1,0)", "(0,-1,-1)", "(-1,-1,1)", "(-1,-1,0)", "(-1,-1,-1)", where S A ,S B ,S C Represent the switching states of motor phase A, phase B, and phase C respectively. Substituting the candidate voltage vector switching state into formula (14), we can get the estimated phase voltage u Pest , and as the output of the phase voltage estimation module 83, output to the flux prediction module 84.
[0095] The flux prediction module 84 receives the phase flux ψ P , phase current i P And the estimated phase voltage u under all candidate voltage vectors Pest As input, the motor phase flux is predicted in the flux prediction module 84:
[0096] ψ Pest =(u Pest -R P i P )T s +ψ P (15)
[0097] In the formula, ψ Pest To predict the phase flux, T s is the sampling step length, ψ Pest As the output of the flux prediction module 84.
[0098] The lookup table prediction module 85 has built-in pre-established switched reluctance motor current-flux-rotor position i-ψ-θ three-dimensional model table and torque-current-rotor position Ti-θ three-dimensional model table. Figure 3 As shown, the current prediction module 851 receives the predicted phase flux ψ output by the flux prediction module 84. PestThe predicted phase rotor position angle θ output by the rotor position angle prediction module 81 Pest As input, based on the motor current-flux-rotor position i-ψ-θ three-dimensional model table data query, obtain the predicted phase current i Pest , as the output of the current prediction module 851.
[0099] The torque prediction module 852 receives the predicted phase current i output by the current prediction module 851 Pest The predicted phase rotor position angle θ output by the rotor position angle prediction module 81 Pest As input, based on the motor torque-current-rotor position Ti-θ three-dimensional model table data query, obtain the predicted phase torque T Pest , as the output of the torque prediction module 852, that is, the output of the lookup table prediction module 85.
[0100] The cost function minimization module 86 receives the reference torque T′ output by the PI controller module 5 and the predicted phase torque T output by the lookup table prediction module 85. Pest As input, the predicted phase torque T under all candidate voltage vectors in the sector Pest Substitute into the cost function G for calculation. The cost function G is composed of the reference torque T' and the predicted phase torque T Pest constitute:
[0101]
[0102] Wherein m is the number of phases of the motor, and in the example of the present invention, m=3.
[0103] According to the rotor sector signal S x Select the candidate voltage vector range and substitute all candidate voltage vectors in the sector into formula (14) to obtain all possible estimated phase voltages u Pest , output to the flux prediction module 84, and obtain all possible predicted phase flux ψ Pest , output to the lookup table prediction module 85, and obtain all possible predicted phase torques T Pest , output to the cost function minimization module 86, substitute into the cost function formula (16) for operation and comparison, calculate the cost function with different values, and select the switch signal ST that minimizes the cost function value P As the output of the model prediction module 8.
[0104] Finally, the switching signal ST P As the input of the power converter module 9 , the on and off of the switch device is controlled to adjust the torque output of the switched reluctance motor 1 .
[0105] The present invention obtains the received reference torque T', motor speed ω, rotor position angle θ while keeping the torque mean unchanged.P , phase flux ψ P , phase current i P Bus voltage u dc , winding inductance voltage drop u r , switch tube voltage drop u S And the freewheeling diode voltage drop u D Perform summary processing and output the total logic switch signal ST P , improve the dynamic response speed of the switched reluctance motor during operation and reduce its torque pulsation.
Claims
1. A model predictive control method for switched reluctance motor based on optimized sector division, Its characteristics are The following steps are involved: Step 1): The flux calculation module (7) receives the phase current i output by the voltage and current sampling module (6) P and phase voltage u P , calculate the phase flux ψ P The rotor position angle prediction module (81) receives the actual rotor position angle θ output by the position sensor module (2). P The predicted phase rotor position angle θ is obtained by comparing the actual motor speed ω output by the speed calculation module (3) Pest And the predicted A phase rotor position angle θ taking phase A as an example Aest ; Step 2): an asymmetric half-bridge power converter is used to control the switched reluctance motor. When the stator winding is in the excitation state, both the upper and lower switch tubes are turned on, and the state is 1; when the stator winding is in the demagnetization state, both the upper and lower switch tubes are turned off, and the state is -1; when the stator winding is in the zero-voltage freewheeling state, the state is 0; The rotor sector determination module (82) receives the predicted A-phase rotor position angle θ Aest , according to the predicted A-phase rotor position angle θ Aest The sector is judged in the interval, when 0°<θ Aest ≤17°, phase A adopts switch state 1, 0 or -1; when 17°<θ Aest ≤22.5°, phase A adopts switch state 0 or -1; when 22.5°<θ Aest When ≤45°, phase A adopts switch state -1 and outputs rotor sector signal S x ; Step 3): The phase voltage estimation module (83) receives the bus voltage u dc , winding inductance voltage drop u r , switch tube voltage drop u S , freewheeling diode voltage drop u D And the rotor sector signal S x , get the estimated phase voltage u Pest : Step 4): The flux prediction module (84) receives the phase flux ψ P , phase current i P And all estimated phase voltages u Pest , predict the predicted phase flux ψ Pest :The lookup table prediction module (85) receives the predicted phase flux ψ Pest and predicted phase rotor position angle θ Pest According to the preset current-flux-rotor position i-ψ-θ three-dimensional model table and torque-current-rotor position Ti-θ three-dimensional model table, the predicted phase torque T is obtained by querying Pest ; Step 5): The cost function minimization module (86) receives the reference torque T' and the predicted phase torque T Pest , calculate the cost function, m is the number of motor phases, and select the switching signal that minimizes the cost function G as the input of the power converter module (9) to control the switched reluctance motor.
2. The control method according to claim 1, Its characteristics are: In step 1), the flux calculation module 7 is calculated by the formula Calculate the phase flux ψ P , R P is the phase resistance, P=A, B, C, T is the flux integration time.
3. The control method according to claim 1, Its characteristics are: In step 2), the rotor sector determination module (82) divides the sectors according to the torque distribution rule, the phase close to the motor stator and rotor misalignment position generates more torque, and the phase close to the motor stator and rotor alignment position generates less torque.
4. The control method according to claim 1, Its characteristics are: In step 4), the predicted phase rotor position angle θ Pest =θ P +T s ω, T s is the sampling step length.
5. The control method according to claim 1, Its characteristics are: In step 4), the predicted phase flux ψ Pest =(u Pest -R P i P )T s +ψ P , R P is the phase resistance, T s is the sampling step length.
6. The control method according to claim 1, Its characteristics are: In step 4), the current prediction module (851) receives the predicted phase flux ψ Pest and predicted phase rotor position angle θ Pest Based on the current-flux-rotor position i-ψ-θ three-dimensional model table query, the predicted phase current i is obtained. Pest The torque prediction module (852) receives the predicted phase current i Pest and predicted phase rotor position angle θ Pest , based on the motor torque-current-rotor position Ti-θ three-dimensional model table query to obtain the predicted phase torque T Pest .
7. The control method according to claim 1, Its characteristics are: In step 5), the actual motor speed ω is compared with the reference speed ω' to obtain the speed error ω err , after PI adjustment, the reference torque T' is obtained.
8. The control method according to claim 1, Its characteristics are: In step 5), the cost function m is the number of motor phases.
9. The control method according to any one of claims 1 to 8, Its characteristics are: The switched reluctance motor is measured in an offline experiment, and the torque balance method and Fourier series are used to obtain the current-flux-rotor position i-ψ-θ three-dimensional model table and the motor torque-current-rotor position Ti-θ three-dimensional model table.
