CN105763120A - Permanent magnet synchronous motor quasi dead-beat model prediction flux linkage control method - Google Patents

Abstract

The invention discloses a permanent magnet synchronous motor (PMSM) quasi dead-beat model prediction flux linkage control method comprising steps of: computing a stator flux linkage vector reference value at a next moment within a limited control framework, and obtaining a target voltage vector according to the reference value and a dead-model control idea; determining the sector of the target voltage vector in virtue of the position angle of the target voltage vector and selecting three valid voltage vectors by using the sector; predicting the stator flux linkage vector at the next moment according to the three valid voltage vectors; and obtaining the optimum on-off state of an inverter by optimizing a value function, wherein the inverter outputs voltage to a PMSM according to the optimum on-off state. The value function does not include weight calculation. When the value function is optimized, only the three valid voltage vectors are optimized so that algorithm operation time is shortened.

Description

A kind of quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor

Technical field

The present invention relates to a kind of quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor, belong to motor and drive and control technology.

Background technology

Limited domination set Model Predictive Control can solve optimization problem online according to the constraint of controlled device and discrete feature, its simple in construction and easily realizing, and drives field to be widely used at power electronics and motor in recent years.In Motor Control Field, the difference according to control variable, limited domination set Model Predictive Control has the control of model prediction electric current, model prediction direct torque and model prediction magnetic linkage control.Model prediction electric current controls with stator armature electric current for control object, and the pulsation of its steady state torque is bigger.The cost function of model prediction direct torque is made up of torque and stator magnetic linkage amplitude two parts, owing to the two dimension is different, it is necessary to being coupled by weights, the selection of current weights is still an open problem, it does not have unified theoretical direction.Model prediction magnetic linkage control, compared with model prediction direct torque, only with stator magnetic linkage vector for controlling target, it is to avoid weight computing, can simplify control algolithm.Driving in structure at three phase alternating current motor two-level inverter, model prediction magnetic linkage control is searching optimal voltage vector in 7 different basic voltage vectors.And more voltage vector object, optimization process can be aggravated system delay, cause PREDICTIVE CONTROL not accurate, particularly in polyphase machine PREDICTIVE CONTROL.Therefore, the object reduced in optimization process is conducive to improving PREDICTIVE CONTROL performance.

Summary of the invention

Goal of the invention: in view of above-mentioned background, the present invention is directed to optimal voltage vector selection course in permagnetic synchronous motor model prediction magnetic linkage control process and has done further optimization.According to track with zero error thought, utilize the flux linkage vector in model prediction magnetic linkage control with reference to calculating target voltage vector, it is provided that a kind of quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor；Owing to model prediction magnetic linkage control designs under limited domination set principle, therefore the inventive method can not obtain target voltage vector truly, optimum basic voltage vectors can only be selected from the basic voltage vectors in sector, target voltage vector place, namely can not realize track with zero error truly, and quasi-track with zero error can only be realized.

Technical scheme: for achieving the above object, the technical solution used in the present invention is:

A kind of quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor, comprises the steps:

(1) torque reference T is calculated_e ^*(k): by reference velocity ω^*K the difference e (k) of () and actual feedback speed omega (k) inputs PI controller, calculate torque reference T according to formula (1.1)_e ^*(k)；

${T_e}^{*}\left(k\right)=e\left(k\right)(K_P+\frac{K_I}{s})---\left(1.1\right)$

Wherein: K_PAnd K_IThe respectively proportional gain of PI controller and storage gain；

(2) (k+1) moment stator magnetic linkage vector is calculated with reference to ψ_s ^*(k+1): by stator magnetic linkage amplitude reference | ψ_s ^*(k) | with torque reference T_e ^*K () input torque angle computing module, examines δ according to formula (2.1) calculating torque JIAOSHEN^*(k)；By rotor position angle θ_rK () and actual feedback speed omega (k) input rotor position angle prediction module, predict (k+1) moment rotor position angle θ according to formula (2.2)_r(k+1)；By angle of torsion with reference to δ^*(k) and (k+1) moment rotor position angle θ_r(k+1) it is added, obtains (k+1) moment stator magnetic linkage angular position theta_s(k+1)；By stator magnetic linkage amplitude reference | ψ_s ^*(k) | with (k+1) moment stator magnetic linkage angular position theta_s(k+1) input stator magnetic linkage vector is with reference to computing module, calculates (k+1) moment stator magnetic linkage vector with reference to ψ according to formula (2.4)_s ^*(k+1)；

${\mathrm{\δ}}^{*}\left(k\right)=\arcsin \frac{2L_s{T_e}^{*}\left(k\right)}{3P_r\left|{{\mathrm{\ψ}}_s}^{*}\left(k\right)\right|\left|{{\mathrm{\ψ}}_f}^{*}\left(k\right)\right|}---\left(2.1\right)$

θ_r(k+1)=θ_r(k)+ω(k)T_s(2.2)

θ_s(k+1)=δ^*(k)+θ_r(k+1)(2.3)

ψ_s ^*(k+1)=| ψ_s ^*(k)|exp(jθ_s(k+1))(2.4)

Wherein: L_sFor permagnetic synchronous motor synchronous inductance, P_rFor permagnetic synchronous motor number of pole-pairs, | ψ_f ^*(k) | for permanent magnet flux linkage amplitude, T_sSampling time for PREDICTIVE CONTROL；

(3) active voltage vector selects: by (k+1) moment stator magnetic linkage vector with reference to ψ_s ^*(k+1) input target vector computing module, calculates target voltage vector u according to formula (3.1)_obj(k+1)；By target voltage vector u_obj(k+1) input target voltage azimuth computing module, calculates target voltage vector and α axle angle theta according to formula (3.2)_u(k+1)；By target voltage vector and α axle angle theta_u(k+1) input sector judge module, according to target voltage vector and α axle angle theta_u(k+1) target voltage vector u is judged_obj(k+1) sector number N (k+1)；Sector number N (k+1) input is selected module, it is thus achieved that three basic voltage vectors u_i(k+1), arrow number i=2N-1,2N, 2N+1；

$u_{obj}(k+1)=\frac{{{\mathrm{\ψ}}_s}^{*}(k+1)-{\mathrm{\ψ}}_s\left(k\right)}{T_s}+R_s{i}_s\left(k\right)---\left(3.1\right)$

${\mathrm{\θ}}_u(k+1)=arctan\frac{u_{obj\mathrm{\β}}(k+1)}{u_{obj\mathrm{\α}}(k+1)}---\left(3.2\right)$

Wherein: ψ_sK () is k moment stator magnetic linkage vector, R_sFor stator resistance, i_sK () is stator current, u_objβAnd u (k+1)_objα(k+1) respectively u_obj(k+1) α axle and beta-axis component；

(4) (k+1) moment stator magnetic linkage vector predictor ψ is calculated_s(k+1): by stator current i_s(k) and three basic voltage vectors u_i(k+1) input stator magnetic linkage vector prediction module, calculates (k+1) moment stator magnetic linkage vector predictor ψ according to formula (4.1)_s(k+1)；

ψ_s(k+1)=ψ_s(k)+T_s(u_i(k+1)-R_si_s(k))(4.1)

(5) inverter optimized switching state is selected: by (k+1) moment stator magnetic linkage vector with reference to ψ_s ^*And (k+1) moment stator magnetic linkage vector predictor ψ (k+1)_s(k+1) input optimizes module, according to formula (5.1) given price value function g_i, cost function g_iWhen taking minima, the basic voltage vectors of its correspondence is defined as optimum basic voltage vectors u_opt, according to optimum basic voltage vectors u_optObtain optimized switching state S_a,b,c；

g_i=| ψ_s ^*(k+1)-ψ_s(k+1)|(5.1)

(6) output optimal voltage: inverter is by optimized switching state S_a,b,cIt is converted into optimal voltage and flows to permagnetic synchronous motor.

Concrete, select module after obtaining sector number N (k+1), select three basic voltage vectors u according to following relation_i(k+1):

1. during N (k+1)=1,Basic voltage vectors is (100,000,110), and three basic voltage vectors are u₁、u₂And u₃；

2. during N (k+1)=2,Basic voltage vectors is (110,000,010), and three basic voltage vectors are u₃、u₄And u₅；

3. during N (k+1)=3,Basic voltage vectors is (010,000,011), and three basic voltage vectors are u₅、u₆And u₇；

4. during N (k+1)=4,Basic voltage vectors is (011,000,001), and three basic voltage vectors are u₇、u₈And u₉；

5. during N (k+1)=5,Basic voltage vectors is (001,000,101), and three basic voltage vectors are u₉、u₁₀And u₁₁；

6. during N (k+1)=6,Basic voltage vectors is (101,000,100), and three basic voltage vectors are u₁₁、u₁₂And u₁₃。

Beneficial effect: the quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor provided by the invention, directly with stator magnetic linkage vector for control variable, compared with common Model Predictive Control, avoid the calculating of weights, simplify optimum basic voltage vectors selection course, reduce system delay, be conducive to improving the ageing of algorithm, be more suitable for practical application.Method provided by the present invention also provides a kind of new approaches for polyphase machine and multi-electrical level inverter Model Predictive Control policy optimization.

Accompanying drawing explanation

Fig. 1 is principles of the invention block diagram, selects module 3, stator magnetic linkage vector prediction module 4 including PI controller 1, stator magnetic linkage vector with reference to computing module 2, active voltage vector, optimizes module 5, inverter 6, permagnetic synchronous motor 7 and photoelectric encoder 8；

Fig. 2 is the stator magnetic linkage vector theory diagram with reference to computing module 2, including angle of torsion computing module 2-1, rotor position angle prediction module 2-2 and stator magnetic linkage vector reference computing module 2-3；

Fig. 3 is the theory diagram that active voltage vector selects module 3, including target voltage vectors calculation module 3-1, target voltage azimuth computing module 3-2, sector judge module 3-3 and selection module 3-4；

Fig. 4 is k moment permagnetic synchronous motor vector correlation figure；

Fig. 5 is k moment voltage vector graph of a relation；

Fig. 6 is quasi-dead beat model prediction magnetic linkage control algorithm flow chart.

Detailed description of the invention

Below in conjunction with accompanying drawing, the present invention is further described.

It is illustrated in figure 1 a kind of permagnetic synchronous motor quasi-dead beat model prediction magnetic linkage control system as it is shown in figure 1, include PI controller 1, stator magnetic linkage vector with reference to computing module 2, active voltage vector selection module 3, stator magnetic linkage vector prediction module 4, optimization module 5, inverter 6, permagnetic synchronous motor 7 and photoelectric encoder 8.

Realize the quasi-dead beat model prediction magnetic linkage control of described permagnetic synchronous motor, it is necessary to calculate stator magnetic linkage vector with reference to ψ_s ^*And stator magnetic linkage vector predictor ψ (k+1)_s(k+1), active voltage vector is selected according to quasi-track with zero error principle.Fig. 4 is the relation between k moment each vector, and Fig. 5 is target voltage vector and the algorithm flow chart that basic voltage vectors figure, Fig. 6 are the quasi-dead beat model prediction flux linkage control method of described permagnetic synchronous motor.Specifically include following steps:

(1) torque reference T is calculated_e ^*(k): by reference velocity ω^*K the difference e (k) of () and actual feedback speed omega (k) inputs PI controller 1, calculate torque reference T according to formula (1.1)_e ^*(k)；

${T_e}^{*}\left(k\right)=e\left(k\right)(K_P+\frac{K_I}{s})---\left(1.1\right)$

Wherein: K_PAnd K_IThe respectively proportional gain of PI controller 1 and storage gain；

(2) (k+1) moment stator magnetic linkage vector is calculated with reference to ψ_s ^*(k+1): by stator magnetic linkage amplitude reference | ψ_s ^*(k) | with torque reference T_e ^*K () input torque angle computing module 2-1, examines δ according to formula (2.1) calculating torque JIAOSHEN^*(k)；By rotor position angle θ_rK () and actual feedback speed omega (k) input rotor position angle prediction module 2-2, predict (k+1) moment rotor position angle θ according to formula (2.2)_r(k+1)；By angle of torsion with reference to δ^*(k) and (k+1) moment rotor position angle θ_r(k+1) it is added, obtains (k+1) moment stator magnetic linkage angular position theta_s(k+1)；By stator magnetic linkage amplitude reference | ψ_s ^*(k) | with (k+1) moment stator magnetic linkage angular position theta_s(k+1) input stator magnetic linkage vector is with reference to computing module 2-3, calculates (k+1) moment stator magnetic linkage vector with reference to ψ according to formula (2.4)_s ^*(k+1)；

${\mathrm{\δ}}^{*}\left(k\right)=\arcsin \frac{2L_s{T_e}^{*}\left(k\right)}{3P_r\left|{{\mathrm{\ψ}}_s}^{*}\left(k\right)\right|\left|{{\mathrm{\ψ}}_f}^{*}\left(k\right)\right|}---\left(2.1\right)$

θ_r(k+1)=θ_r(k)+ω(k)T_s(2.2)

θ_s(k+1)=δ^*(k)+θ_r(k+1)(2.3)

ψ_s ^*(k+1)=| ψ_s ^*(k)|exp(jθ_s(k+1))(2.4)

Wherein: L_sFor permagnetic synchronous motor synchronous inductance, P_rFor permagnetic synchronous motor number of pole-pairs, | ψ_f ^*(k) | for permanent magnet flux linkage amplitude, T_sSampling time for PREDICTIVE CONTROL；

(3) active voltage vector selects: by (k+1) moment stator magnetic linkage vector with reference to ψ_s ^*(k+1) input target vector computing module 3-1, calculates target voltage vector u according to formula (3.1)_obj(k+1)；By target voltage vector u_obj(k+1) input target voltage azimuth computing module 3-2, calculates target voltage vector and α axle angle theta according to formula (3.2)_u(k+1)；By target voltage vector and α axle angle theta_u(k+1) input sector judge module 3-3, according to target voltage vector and α axle angle theta_u(k+1) target voltage vector u is judged_obj(k+1) sector number N (k+1)；Sector number N (k+1) input is selected module 3-4 (with reference to table 1), it is thus achieved that three basic voltage vectors u_i(k+1), arrow number i=2N-1,2N, 2N+1；

$u_{obj}(k+1)=\frac{{{\mathrm{\ψ}}_s}^{*}(k+1)-{\mathrm{\ψ}}_s\left(k\right)}{T_s}+R_s{i}_s\left(k\right)---\left(3.1\right)$

${\mathrm{\θ}}_u(k+1)=arctan\frac{u_{obj\mathrm{\β}}(k+1)}{u_{obj\mathrm{\α}}(k+1)}---\left(3.2\right)$

Wherein: ψ_sK () is k moment stator magnetic linkage vector, R_sFor stator resistance, i_sK () is stator current, u_objβAnd u (k+1)_objα(k+1) respectively u_obj(k+1) α axle and beta-axis component；

(4) (k+1) moment stator magnetic linkage vector predictor ψ is calculated_s(k+1): by stator current i_s(k) and three basic voltage vectors u_i(k+1) input stator magnetic linkage vector prediction module 4, calculates (k+1) moment stator magnetic linkage vector predictor ψ according to formula (4.1)_s(k+1)；

ψ_s(k+1)=ψ_s(k)+T_s(u_i(k+1)-R_si_s(k))(4.1)

(5) inverter optimized switching state is selected: by (k+1) moment stator magnetic linkage vector with reference to ψ_s ^*And (k+1) moment stator magnetic linkage vector predictor ψ (k+1)_s(k+1) input optimizes module 5, according to formula (5.1) given price value function g_i, cost function g_iWhen taking minima, the basic voltage vectors of its correspondence is defined as optimum basic voltage vectors u_opt, according to optimum basic voltage vectors u_optObtain optimized switching state S_a,b,c；

g_i=| ψ_s ^*(k+1)-ψ_s(k+1)|(5.1)

(6) output optimal voltage: inverter is by optimized switching state S_a,b,cIt is converted into optimal voltage and flows to permagnetic synchronous motor.

The corresponding relation of table 1 sector number, sector angle, basic voltage vectors and arrow number

Fig. 4 is k moment permagnetic synchronous motor voltage vector-diagram, and (k+1) moment stator magnetic linkage vector is with reference to ψ_s ^*(k+1) obtained by formula (2.4) calculating；Target voltage vector u_obj(k+1) obtained by formula (3.1) calculating, utilize the voltage vector relation in Fig. 5 to determine target voltage vector u_obj(k+1) sector, place is No. 3 sectors, then be now likely to become being only possible to as u of optimum basic voltage vectors₅、u₆And u₇, utilize formula (4.1) to calculate the stator magnetic linkage vector predictor ψ that 3 basic voltage vectors are corresponding_s(k+1), as shown in phantom in Figure 4；Again through optimizing module according to formula (5.1) to the stator magnetic linkage vector predictor ψ under 3 basic voltage vectors effects_s(k+1) it is optimized, obtains optimum basic voltage vectors.In this optimization process, only use 3 basic voltage vectors, compared with common model prediction magnetic linkage, decreased optimization number of times, reduce system delay, be conducive to improving the ageing of algorithm.

Described control method is to design for the three-phase permanent magnet synchronous motor driven by two-level inverter, and the method can also be extended in multiphase permanent magnet synchronous motor control

The above is only the preferred embodiment of the present invention; it is noted that, for those skilled in the art; under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improvements and modifications also should be regarded as protection scope of the present invention.

Claims (2)

1. the quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor, it is characterised in that: comprise the steps:

(1) torque reference T is calculated_e ^*(k): by reference velocity ω^*K the difference e (k) of () and actual feedback speed omega (k) inputs PI controller (1), calculate torque reference T according to formula (1.1)_e ^*(k)；

${T_e}^{*}\left(k\right)=e\left(k\right)(K_P+\frac{K_I}{s})---\left(1.1\right)$

Wherein: K_PAnd K_IThe respectively proportional gain of PI controller (1) and storage gain；

(2) (k+1) moment stator magnetic linkage vector is calculated with reference to ψ_s ^*(k+1): by stator magnetic linkage amplitude reference | ψ_s ^*(k) | with torque reference T_e ^*K () input torque angle computing module (2-1), examines δ according to formula (2.1) calculating torque JIAOSHEN^*(k)；By rotor position angle θ_rK () and actual feedback speed omega (k) input rotor position angle prediction module (2-2), predict (k+1) moment rotor position angle θ according to formula (2.2)_r(k+1)；By angle of torsion with reference to δ^*(k) and (k+1) moment rotor position angle θ_r(k+1) it is added, obtains (k+1) moment stator magnetic linkage angular position theta_s(k+1)；By stator magnetic linkage amplitude reference | ψ_s ^*(k) | with (k+1) moment stator magnetic linkage angular position theta_s(k+1) input stator magnetic linkage vector is with reference to computing module (2-3), calculates (k+1) moment stator magnetic linkage vector with reference to ψ according to formula (2.4)_s ^*(k+1)；

${\mathrm{\δ}}^{*}\left(k\right)=\arcsin \frac{2L_s{T_e}^{*}\left(k\right)}{3P_r\left|{{\mathrm{\ψ}}_s}^{*}\left(k\right)\right|\left|{{\mathrm{\ψ}}_f}^{*}\left(k\right)\right|}---\left(2.1\right)$

θ_r(k+1)=θ_r(k)+ω(k)T_s(2.2)

θ_s(k+1)=δ^*(k)+θ_r(k+1)(2.3)

ψ_s ^*(k+1)=| ψ_s ^*(k)|exp(jθ_s(k+1))(2.4)

Wherein: L_sFor permagnetic synchronous motor synchronous inductance, P_rFor permagnetic synchronous motor number of pole-pairs, | ψ_f ^*(k) | for permanent magnet flux linkage amplitude, T_sSampling time for PREDICTIVE CONTROL；

(3) active voltage vector selects: by (k+1) moment stator magnetic linkage vector with reference to ψ_s ^*(k+1) input target vector computing module (3-1), calculates target voltage vector u according to formula (3.1)_obj(k+1)；By target voltage vector u_obj(k+1) input target voltage azimuth computing module (3-2), calculates target voltage vector and α axle angle theta according to formula (3.2)_u(k+1)；By target voltage vector and α axle angle theta_u(k+1) input sector judge module (3-3), according to target voltage vector and α axle angle theta_u(k+1) target voltage vector u is judged_obj(k+1) sector number N (k+1)；Sector number N (k+1) input is selected module (3-4), it is thus achieved that three basic voltage vectors u_i(k+1), arrow number

$u_{obj}(k+1)=\frac{{{\mathrm{\ψ}}_s}^{*}(k+1)-{\mathrm{\ψ}}_s\left(k\right)}{T_s}+R_s{i}_s\left(k\right)---\left(3.1\right)$

${\mathrm{\θ}}_u(k+1)=\mathrm{arc}tan\frac{u_{obj\mathrm{\β}}(k+1)}{u_{obj\mathrm{\α}}(k+1)}---\left(3.2\right)$

Wherein: ψ_sK () is k moment stator magnetic linkage vector, R_sFor stator resistance, i_sK () is stator current, u_objβAnd u (k+1)_objα(k+1) respectively u_obj(k+1) α axle and beta-axis component；

(4) (k+1) moment stator magnetic linkage vector predictor ψ is calculated_s(k+1): by stator current i_s(k) and three basic voltage vectors u_i(k+1) input stator magnetic linkage vector prediction module (4), calculates (k+1) moment stator magnetic linkage vector predictor ψ according to formula (4.1)_s(k+1)；

ψ_s(k+1)=ψ_s(k)+T_s(u_i(k+1)-R_si_s(k))(4.1)

(5) inverter optimized switching state is selected: by (k+1) moment stator magnetic linkage vector with reference to ψ_s ^*And (k+1) moment stator magnetic linkage vector predictor ψ (k+1)_s(k+1) input optimizes module (5), according to formula (5.1) given price value function g_i, cost function g_iWhen taking minima, the basic voltage vectors of its correspondence is defined as optimum basic voltage vectors u_opt, according to optimum basic voltage vectors u_optObtain optimized switching state S_a,b,c；

g_i=| ψ_s ^*(k+1)-ψ_s(k+1)|(5.1)

(6) output optimal voltage: inverter is by optimized switching state S_a,b,cIt is converted into optimal voltage and flows to permagnetic synchronous motor.

2. the quasi-dead beat model prediction flux linkage control method of permagnetic synchronous motor according to claim 1, it is characterised in that: select module (3-4) after obtaining sector number N (k+1), select three basic voltage vectors u according to following relation_i(k+1):

1. during N (k+1)=1,Basic voltage vectors is (100,000,110), and three basic voltage vectors are u₁、u₂And u₃；

2. during N (k+1)=2,Basic voltage vectors is (110,000,010), and three basic voltage vectors are u₃、u₄And u₅；

3. during N (k+1)=3,Basic voltage vectors is (010,000,011), and three basic voltage vectors are u₅、u₆And u₇；

4. during N (k+1)=4,Basic voltage vectors is (011,000,001), and three basic voltage vectors are u₇、u₈And u₉；

5. during N (k+1)=5,Basic voltage vectors is (001,000,101), and three basic voltage vectors are u₉、u₁₀And u₁₁；

6. during N (k+1)=6,Basic voltage vectors is (101,000,100), and three basic voltage vectors are u₁₁、u₁₂And u₁₃。