CN1588793A

CN1588793A - Method or controlling permanent magnet synchronous motor-air conditioner compressor system without speed sensor

Info

Publication number: CN1588793A
Application number: CNA2004100781411A
Authority: CN
Inventors: 刘智超; 赵铁夫; 黄立培
Original assignee: Tsinghua University
Current assignee: Tsinghua University
Priority date: 2004-09-17
Filing date: 2004-09-17
Publication date: 2005-03-02
Anticipated expiration: 2024-09-17
Also published as: CN1283041C

Abstract

This invention relates to nonvelocity sensor permanent-magnet synchro-motor-air conditioning compressor system. It is characterized by that nonvelocity sensor vector control is used for permanent-magnet synchromotor, so, overcoming the shortcoming of the large rotating speed pulsation. At the same time, the torque command current composite control method is used to make further lowering rotating speed pulsation of the motor. It overcomes the problem of unsatability of the system during the procedure of regulating PI regulator parameter in traditional vector control system.

Description

Control method of permanent magnet synchronous motor-air conditioner compressor system without speed sensor

Technical Field

The invention belongs to the motor control technology, in particular to the technical field of application of a permanent magnet synchronous motor in a variable frequency air conditioner.

Background

With the coming of energy crisis worldwide, governments of various countries actively promote energy-saving and consumption-reducing technologies for the purpose of economic sustainable development. As one of the main appliances for home electricity, the conventional fixed frequency air conditioner is gradually coming out of the market due to its low operation efficiency. The new generation of inverter air conditioner has a series of advantages of obvious energy-saving effect, stable temperature regulation, low operation noise in the whole frequency range and the like, so that the inverter air conditioner is concerned by the market.

With the continuous improvement and perfection of the performance of the permanent magnet material and the gradual maturity of the research and development experience of the permanent magnet motor, the permanent magnet motor is more and more widely applied in the aspects of national defense, industrial and agricultural production, daily life and the like, and is developed in the aspects of high power, high functionalization and miniaturization. The permanent magnet motor has the characteristics of simple structure, small volume, light weight, low loss, high efficiency and the like, and is gradually valued by the industry.

Conventionally, the Permanent Magnet Motor can be divided into a Permanent Magnet Synchronous Motor (PMSM) and a brushless direct current Motor (BLDCM) according to the difference of the geometric shape of the rotor magnetic steel of the Permanent Magnet Motor. There are many similarities between the two, the biggest difference between them is: when the rotor rotates, the counter electromotive force generated in the stator has different waveforms, and the counter electromotive force of the magnetic synchronous motor is a sine wave, whereas the brushless dc motor is a trapezoidal wave. Therefore, the two motors differ in principle, model, and control method. A control system of a brushless dc motor is widely used in an air conditioner at present.

The vector control system of the permanent magnet synchronous motor can realize high-precision, high-dynamic performance and large-range speed and position control. Compared with a direct current motor, the motor has no mechanical commutator and brush; compared with asynchronous motor, it needs no reactive exciting current, so that it has high power factor, low stator resistance loss, measurable rotor parameter and good control performance. At present, a servo system composed of a permanent magnet synchronous motor is widely applied to the fields of flexible manufacturing systems, robots, office automation, numerical control machines and the like.

For compressor load in an air conditioner, the load torque of the motor periodically fluctuates due to changes in the pressure in the compressor cylinder during each rotation of the motor rotor. In a conventional control system using a brushless dc motor, the stator current of the brushless dc motor is a square wave, and each phase is turned on at 120 ° electrical angle and then turned off at 60 ° electrical angle. The brushless DC motor rotor position detector outputs a pulse every 60 DEG electrical angle because a switch changes state every 60 DEG electrical angle. In general, the rotor position angle of the brushless dc motor is unknown in each 60 ° electrical angle, so that the brushless dc motor correspondingly outputs six reference torque values to approximate the load torque in a six-step manner, and cannot accurately follow the change of the load torque. Fig. 1 is a schematic diagram of a method for controlling the load of a compressor and the torque of a brushless dc motor, and it can be seen that the set torque cannot track the load torque of the compressor well. This inevitably produces large rotational speed fluctuations, which affect the performance of the compressor.

The stator current of the permanent magnet synchronous motor is sine wave, and the position of a rotor needs to be continuously detected. The vector control system of the permanent magnet synchronous motor can realize high-precision, high-dynamic performance and large-range speed and position control, thereby achieving the effect of high-performance control in the variable frequency air conditioner. The vector control system of the permanent magnet synchronous motor can replace the original brushless direct current motor to quickly and accurately track load torque, reduce fluctuation of rotating speed and realize high-performance control. Meanwhile, because the compressor cannot be provided with a speed and position mechanical sensor, the method of using a speed sensor-free method to realize the control of the load of the high-performance compressor becomes a novel control mode of the air-conditioning compressor.

Vector control system of permanent magnet synchronous motor

In 1971, the vector control method of the alternating current motor proposed by Blaschke et al in Germany theoretically solves the problem of high-performance control of the torque of the alternating current motor. The basic idea of vector control stems from the rigorous simulation of dc motors. The direct current motor has good decoupling performance, and can achieve the purpose of controlling the torque of the motor by controlling the armature current and the exciting current of the direct current motor respectively. The ultimate goal of vector control is to improve the torque control performance of the motor, while the implementation still falls on controlling the stator current. The vector control divides the stator current into an excitation component and a torque component through the motor magnetic field orientation, and the excitation component and the torque component are respectively controlled, so that a good decoupling characteristic is obtained. The control method is firstly applied to the asynchronous motor and is quickly transplanted to the synchronous motor. In fact, vector control is easier to implement on permanent magnet synchronous machines. Since the machine does not induce slip frequency currents in the machine during vector control and control is less affected by parameters (mainly rotor parameters). Currently, the vector control technology is widely applied to permanent magnet synchronous motors.

Speed sensorless control

Vector control of PMSM typically controls stator current or voltage by detecting or estimating the position and magnitude of the motor rotor flux. Thus, the torque of the motor is only related to the magnetic flux and the current, and the control method is similar to the control method of the direct current motor, so that high control performance can be obtained. In the control of a conventional permanent magnet synchronous motor, in order to obtain the precise position and speed of the rotor, a mechanical sensor is generally required to be arranged on the shaft of the rotor to measure the speed and position of the motor. These mechanical sensors are often encoders, resolvers and tachogenerators. Mechanical sensors provide the rotor signals required by the motor, but also present problems to the governor system:

(1) the mechanical sensor increases the rotational inertia on the rotor shaft of the motor, increases the space size and the volume of the motor, increases a connecting wire and an interface circuit between the motor and a control system by using the mechanical sensor, enables the system to be easily interfered, and reduces the reliability.

(2) The system is not widely applicable to various occasions due to the limitation of the use conditions of the mechanical sensor, such as temperature, humidity, vibration and the like.

(3) The mechanical sensors and their auxiliary circuits increase the cost of the governor system, the price of some high-precision sensors can even be compared with the price of the motor itself.

In order to overcome the defects brought to the system by using the sensor, a plurality of scholars develop the research of the sensorless permanent magnet synchronous motor control system. The AC speed regulating system without mechanical sensor is to utilize the relevant electric signal in the motor winding to estimate the position and speed of the rotor by proper method to replace the mechanical sensor so as to realize the closed-loop control of the motor.

Control of PMSM-air conditioner compressor system

For a permanent magnet synchronous motor-air conditioner compressor load system without speed sensor vector control, speed fluctuation is caused due to the large-range periodical change of load torque. In the actual motor operation process of the system, parameters of the motor, including resistance, inductance, back electromotive force coefficient and the like, can change due to the change of the operation conditions (such as current, temperature, humidity and the like) of the motor, and meanwhile, due to errors in measurement, the parameters of the real motor cannot be accurately obtained. In addition, in order to reduce the cost, it is desirable to increase the length of the control cycle time as much as possible. The study on the motor parameter and the control period change shows that: for the compressor load in the air conditioner, in the process of each rotation of the motor rotor, due to the change of the pressure in the compressor cylinder, the load torque of the motor can fluctuate periodically, the identification of the rotating speed and the position and the operation performance of the system are easy to be subjected to parameter change, and similar problems also occur when the control period of the system is increased.

In the conventional vector control, the parameters of the PI regulator in the control block diagram are generally adjusted to make the response of the system meet the requirements. In the compressor load system, because of the periodic pulsation of the compressor load, the contradiction between the stability and the response speed in the PI parameter setting method is highlighted: when the PI gain is too small, the identified rotating speed is difficult to track the real rotating speed; the PI gain is too large, and the system is unstable and oscillates. Through the adjustment of the regulator parameters, the estimated speed can stably track the actual speed, however, the response speed of the system is reduced due to the reduction of the gain of the speed regulator. The output of the speed regulator, i.e., the torque command current, does not track the changes in load torque well, resulting in a significant lag of the motor electromagnetic torque with respect to the load torque. FIG. 2 shows the simulation result after the PI parameter of the speed regulator is set when the q-axis inductance fluctuates by 10%. Therefore, in each mechanical cycle of the rotor, the speed cannot be guaranteed to be a constant value, and large fluctuation occurs, so that the working performance of the compressor is influenced.

Disclosure of Invention

The invention aims to provide a control method of a permanent magnet synchronous motor-air conditioner compressor system without a speed sensor and with small fluctuation.

Compared with the brushless direct current motor control system which is widely used in the air conditioning system at present, the invention is characterized in that the speed sensorless vector control is applied to the permanent magnet synchronous motor-air conditioner compressor load system, thereby overcoming the defect that the output torque of the brushless direct current motor can not accurately follow the change of the load torque to cause larger rotating speed pulsation. Meanwhile, according to the characteristic of periodic pulsating load torque of the compressor, the invention provides an improved method for reducing the motor rotating speed pulsation in a permanent magnet synchronous motor-air conditioner compressor load system, namely a torque command current composite control method. The method can effectively overcome the contradiction between system stability and response speed in the parameter setting process of the PI regulator of the traditional vector control system, and improves the control performance of the permanent magnet synchronous motor in the variable frequency air conditioner.

Permanent magnet synchronous motor-air conditioner compressor system without speed sensor

The invention provides a permanent magnet synchronous motor-air conditioner compressor system, and a control schematic block diagram of the system is shown in figure 3. In the system, stator phase current of a permanent magnet synchronous motor is firstly input to a position estimation link through sampling and A/D conversion, the position and the speed of the motor are identified through a speed sensorless algorithm, and then the stator current is divided into an excitation component and a torque component by utilizing an existing vector control method and is respectively controlled. Wherein the exciting current i_dReference value i_drefAnd (0), the inverter adopts a space vector PWM modulation method, and the position estimation link adopts a counter-electromotive-based embedded permanent magnet synchronous motor speed sensorless and position identification algorithm. The system has the innovation points that a permanent magnet synchronous motor vector control system is adopted to replace a brushless direct current motor system, and the rotating speed and the position of the motor are estimated by adopting a speed-sensorless method, so that the load torque is quickly and accurately tracked, the fluctuation of the rotating speed is reduced, and the high-performance control is realized.

Torque command current composite control method

A torque command current composite control method is provided for the parameter setting problem of a PI regulator when parameters change in a permanent magnet synchronous motor-air conditioner compressor load system. Compared with the traditional vector control strategy, the method can effectively reduce the rotation speed pulsation and improve the control performance of the permanent magnet synchronous motor in the air conditioning system.

Because there is a relatively strict correspondence between the torque and the position angle of the compressor load, feed forward compensation for the torque current can be added based on the estimated position. I adopted in the system_dA control method of 0, thus at steady state:

T_em＝p_n(i_qψ_d-i_dψ_q)＝p_n(i_qψ₁+(L_d-l_q)i_di_q)＝p_ni_qψ_r

in the formula, T_emIs an electromagnetic torque, p_nIs the number of pole pairs, i_d，i_qFor stator winding d-q axis currents, #_d，ψ_qIs d-q axis flux linkage, psi_rFor permanent magnet flux linkage of rotor, L_d，L_qIs a d-q axis inductance.

Then, a position-torque table is formed based on the correspondence between the compressor torque and the motor rotor position. In each control period, the torque of the load is found according to the estimated position lookup table, and the required electromagnetic torque is obtained, so that the required torque current is directly calculated as follows:

FIG. 4 is a schematic block diagram of a torque command current compound control method. As shown in the figure, the method introduces the estimated angle into a torque current control link, namely a torque current i through a compressor load torque reference table_{q_ref}' and then calculating the torque current i_{q_ref}' sum with the output of the speed PI regulator as the torque command current value (i)_{q_ref}+i_{q_ref}') the rest of the system is identical to the existing vector control system. FIG. 3 is a block diagram of a system control modified by the introduction of a torque command current compound control method.

The invention is characterized in that it comprises the following steps in sequence:

(1) system initialization

Inputting a user-provided compressor load torque-rotor position curve table to a Digital Signal Processor (DSP) in the system;

in the DSP, the following are set:

reference value omega of the rotational speed _ref0, excitation current reference value i_{d_ref}＝0；

Back emf estimation constant K_eConstant of position estimation K_θIs a set value;

the control period T is a set value;

setting proportional and integral constants K of speed regulator_p1，K_i1Proportional and integral constants K of torque current regulators_p2，K_i2Proportional and integral constants K of excitation current regulator_p3，K_i3；

Parameters of the permanent magnet synchronous machine provided by the user: number of pole pairs p_nPermanent magnet linkage psi of rotor_rDq axis inductance L_d，L_qStator resistance R, back emf coefficient K_E(ii) a At the same time, the initial value of the dq-axis voltage of the motor is set to zero, i.e., u_d(0)＝0，u_q(0)＝0；

(2) DSP detects stator three-phase current and calculates dq axle voltage of motor

DSP detects stator three-phase current i coming from the current transformer, the filter capacitor and the A/D converter on the stator side of the motor in sequence_a，i_b，i_c；

For u, n.gtoreq.1_d，u_qTaking the actual reference dq axis voltage calculated in the last digital control period T in the DSP, namely u_d(n)＝u_{d_ref}(n-1)，u_q(n)＝u_{q_ref}(n-1)；

(3) The DSP adopts a counter-electromotive force-based rotating speed and position identification method of an embedded permanent magnet synchronous motor non-speed sensor to identify the rotating speed and position of the motor in the nth periodAnd position theta_M(n)；

Setting: a model motor with exactly the same parameters is built on the y δ axis, which is the estimated rotor axis and has an angular error Δ θ from the dq axis, Δ θ being θ - θ_MTheta is the angle between the d axis and the reference axis + A, theta_MIs the included angle between the gamma axis and the reference axis + A:

setting initial time, rotation speed

Position theta_M(0)＝0，

Then:

e_M(n)＝e_M(n-1)-K_eΔi_δ(n)

wherein,

Δθ(n)＝θ(n)-θ_M(n)

p is a differential operator;

when the estimation reaches the steady state, Δ θ ≈ 0, so that the formula L_γ，L_δ，L_γδAre respectively approximately equal to L_d，L _q0, said L_γ，L_δ，L_γδAre respectively

Self-inductance of the gamma axis, self-inductance of the delta axis and mutual inductance between the gamma axis and the delta axis;

wherein,

n is the number of time periods after discretization,

θ_Mis the estimated rotor position angle and is,

e_Mis the back-emf of the model motor,

e is the back-emf of the actual motor,

ω is the rotor speed, i.e.

K_eIs a back-emf estimation constant, K_θThe position estimation constants are set values;

u_γ，u_δthe voltages on the gamma, delta axes for a real motor,

i_γ(n-1)，i_δ(n-1) is the current response of the real motor on the gamma and delta axes,

i_Mγ(n)，i_Mδ(n) is the current response of the model motor established by taking gamma and delta shafts as rotor shafts,

Δi_γ(n)，Δi_δ(n) is the difference between the real current and the current of the model motor on the gamma and delta axes in each period;

(4) the DSP adopts a torque command current composite control method to calculate a dq axis voltage reference value, and the method sequentially comprises the following steps;

(4.1) comparing i in the current coordinate system in the abc coordinate system_a，i_b，i_cMultiplying by a coordinate transformation matrix to obtain a current value i under dq axis_d、i_q、i₀：

(4.2) reference value ω of rotation speed of DSP_refAnd (4) the estimated rotating speed obtained in the step (3)

The difference Δ ω (n) ═ ω_ref- ω (n) input speed regulator, finding q-axis referenceCurrent i_{q_ref}(n)，

Wherein, K_p1，K_i1Proportional and integral constants of the speed regulator;

(4.3) estimating the rotor position θ according to the step 3_M(n) in each control cycle, looking up the rotor position-load torque curve table according to the estimated position, while the electromagnetic torque of the motor is equal to the load torque, thus obtaining the electromagnetic torque T_emThereby calculating a required torque current i_{q_ref}′：

(4.4) DSP to i_{q_ref}′+i_{q_ref}With actual q-axis current i_qDifference value Δ i of_q(n)＝i_{q_ref}′(n)+i_{q_ref}(n)-i_q(n) inputting the torque current regulator to obtain a q-axis reference voltage u_{q_ref}(n)，

Wherein, K_p2，K_i2Proportional and integral constants of the torque current regulator;

(4.5) DSP to i_{d_ref}0 and i_dDifference value Δ i of_d(n)＝i_dref(n)-i_d(n) inputting the excitation current regulator to obtain d-axis reference voltage u_{d_ref}(n)，

Wherein, K_p3，K_i3Proportional and integral constants of the excitation current regulator;

(5) DSP according to u_{q_ref}、u_{q_ref}The voltage vector in the dq0 coordinate system is calculated,

(6) the DSP obtains switching signals of the inverter by using a space vector PWM method, obtains the on-off time of each switching tube of the inverter, and sends out corresponding PWM pulses to control the output voltage of the inverter, so as to realize the control of the permanent magnet synchronous motor: in each control period T, the on-off time T of each switching tube in three bridge arms of the inverter_a，T_a，T₀：

T₀＝T-T_a-T_b

The space of voltage vector distribution is divided into 6 regions, and the voltage u in each control period_α，u_βThe resultant space vector magnitude under the α β axis is V_sThe argument is gamma; and the action time of the switching vector is determined by V in each region_sT，V_aT_a，V_bT_bThe triangle formed is determined, where V_a＝V_b＝E_dcT is a control period, E_dcIs a dc bus voltage; meanwhile, gamma also determines the area where the switching vector, i.e. the space vector, is located, i.e. the current operating state of the inverter.

By applying the torque command current composite control method, because the feedforward control of the torque command current is added, the torque current required by the torque changing in a large range can be calculated in advance according to the estimated position of the rotor, and meanwhile, the traditional PI controller can also adjust the feedback speed signal, so that the obtained electromagnetic torque can track the load torque quite well. FIG. 9 shows the system simulation result after the torque command current composite control method is adopted, wherein the q-axis inductance fluctuates by 10%, the control period is increased from 100us to 400 us. Therefore, the speed fluctuation of the motor in a steady state is greatly reduced (within +/-5%), and the system estimation algorithm is not influenced by parameters and can be continuously and stably carried out, so that the effectiveness of the method in improving the control performance of the permanent magnet synchronous motor in the variable frequency air conditioner is proved.

Drawings

FIG. 1 is a schematic diagram of a compressor load and brushless DC motor torque control method;

FIG. 2 is a simulation result after a PI parameter of a speed regulator is set when q-axis inductance fluctuates by 10%;

FIG. 3 is a diagram of a PMSM-air conditioner compressor control system;

FIG. 4. Torque command current compound control method;

FIG. 5 actual motor d-q axes and model motor γ - δ axes;

FIG. 6 software system control flow;

FIG. 7. software computation flow chart;

FIG. 8 is a schematic diagram of a space vector PWM method: a. inverter voltage vector and sector schematic; b, vector synthesis schematic diagram;

FIG. 9 shows that the q-axis inductance fluctuates by 10%, the control period is increased from 100us to 400us, and the system simulation result is obtained after the torque command current composite control method is adopted;

FIG. 10 is a block diagram of a hardware system architecture.

The symbols and variables in the figures are illustrated as follows:

a PMSM permanent magnet synchronous motor;

theta estimated rotor position angle;

the motor speed of ω estimation;

ω_refsetting the rotating speed of the motor;

i_dd-axis electricity of motor under dq0 coordinate systemA current, i.e. an excitation current;

i_qthe q-axis current, namely the torque current, of the motor under a dq0 coordinate system;

i_{q_ref}a reference torque current;

i_{q_ref}a reference torque current obtained by a torque command current composite control method;

i_{d_ref}an excitation current reference value;

u_{d_ref}，u_{q_ref}d-axis reference voltage and q-axis reference voltage of the motor under a dq0 coordinate system respectively;

a conversion module from an abc-dq abc coordinate system to a dq0 coordinate system;

a conversion module from dq-alpha beta dq0 coordinate system to alpha beta 0 coordinate system;

SVPWM is a voltage space vector PWM modulation module, and PI is a proportional integrator module;

u_d，u_qstator winding d-q axis voltage;

i_d，i_qstator winding d-q axis current;

ψ_d，ψ_qd-q axis flux linkage;

ψ_ra rotor permanent magnet flux linkage;

r stator armature resistance;

L_d，L_qa d-q axis inductance;

T_em，T_Lelectromagnetic torque and load torque;

p_nthe number of pole pairs;

moment of inertia of the J-rotor

p differential operator

ω rotor speed (also notedθ is the estimated rotor position angle);

e_Mis the back emf of the model motor;

e is the back emf of the actual motor.

Introduction of the coordinate system of the motors d, q, 0, α, β, 0:

the d, q, 0 coordinate system is a coordinate system which is placed on the rotor and rotates along with the rotor, the d axis is in the direction of the longitudinal axis of the rotor, the q axis leads the d axis by 90 degrees in electrical angle, and the 0 axis is an imaginary axis for ensuring the reversibility of transformation. The d, q, 0 coordinate system is a stationary coordinate system established on the stator. α, β, 0 can be seen as a d, q, 0 coordinate system with zero rotational speed.

The conversion relation of the voltage, the current and the flux linkage of the d, q, 0 coordinate system and the a, b, c coordinate system is as follows:

where θ is the rotor position angle.

The conversion relation of the voltage, the current and the flux linkage of the alpha, beta, 0 coordinate system and the a, b, c coordinate system is as follows:

Detailed Description

FIG. 10 is a block diagram of the hardware system of the present invention. The experimental hardware of the invention adopts a motor control development system PE-Expert of Myway corporation of Japan, and the platform utilizes a DSP chip TMS320C32 of TI corporation and adopts C language programming. The hardware system mainly comprises a PC, a DSP board, an A/D, D/A converter, a PWM generator and a two-level voltage type inverter specially designed for vector control of the AC motor. The system of the invention detects the current of a PMSM stator loop and the direct current bus voltage of an inverter through sensors, carries out AD conversion by utilizing a development system PE-Expert, and carries out velocity-sensor-free vector control in a DSP thereof, and modules related to coordinate transformation, PI regulation and the like in figure 6. PWM pulses are formed by using a space vector modulation method to control an inverter, so that high-performance control of a permanent magnet synchronous motor in the variable-frequency air conditioner is realized.

The system control flow of the present invention is shown in fig. 6, and can be divided into the following steps:

1. firstly, initializing a software system, and setting a reference value omega of a rotating speed_refEstimation of rotational speedCalculation constant K_e，K_θProportional and integral constants K of respective regulators of control period T_p，K_iAnd inputting a user-provided compressor load torque-position curve table. Initializing motor parameters: number of pole pairs p_nPermanent magnet linkage psi of rotor_rD-q axis inductance L_d，L_qStator resistance R.

2. Detecting three-phase currents of the stator and calculating the dq-axis voltage of the motor.

In a practical system, the stator current can be detected by a current sensor, and after being conditioned by filtering, the stator current is subjected to A/D conversion and then input into a Digital Signal Processor (DSP) for utilization. The initial value of the dq-axis voltage of the motor is zero, i.e. u_d(0)＝0，u_q(0) 0; for u with n ≧ 1_d，u_qActual reference dq-axis voltage, i.e. u, calculated for the last digital control period in the DSP_d(n)＝u_{d_ref}(n-1)，u_q(n)＝u_{q_ref}(n-1)，n≥1。

3. And identifying the rotating speed and the position of the motor through a position estimation link.

The system adopts a counter-electromotive-force-based embedded permanent magnet synchronous motor speed sensorless rotation speed and position identification algorithm.

Since the rotor position angle θ is unknown, the precise d-q axis of the machine is not available, and all estimation algorithms can only be based on the estimated position angle θ_MI.e. assuming a model motor with exactly the same parameters is built on the estimated rotor axis (set as the y-delta axis), as shown in fig. 5, the initial moment

Due to the existence of the rotor angle error delta theta, the equation of the actual motor on the gamma-delta axis is changed as follows:

wherein,

Δθ(n)＝θ(n)-θ_M(n)

when the steady state is estimated to be reached, Δ θ is close to 0, so L in the formula (3)_γ，L_δ，L_γδAre respectively approximately equal to L_d，L_q，0。

In digital control, a discrete model of the motor is required to be used, a control period is recorded as T, and then the current response of the real motor on a gamma-delta axis in each control period is deduced as follows:

meanwhile, the current response of the model motor established by taking the gamma-delta shaft as the rotor shaft is as follows,

subtracting the formulas (4) and (5) to obtain the difference between the currents of the real motor and the model motor in each period,

in each control period, a constant K is given_eAnd K_θThe back emf is recurred and the rotor position angle is corrected as follows.

e_M(n)＝e_M(n-1)-K_eΔi_δ(n)

Wherein

The motor speed can be estimated as follows:

4. and calculating a dq axis voltage reference value by a torque command current composite control method.

According to the estimated rotating speed and position obtained in the step 3, i in a current coordinate system under an abc coordinate system_a，i_b，i_cMultiplying by a coordinate transformation matrix to obtain a current value i under dq axis_d、i_q、i₀。

In the present system, the current and speed regulators both take the form of typical limiting plus PI (proportional integrator) regulators. The mathematical expression for the input and output of a conventional PI regulator is:

e (n) is the regulator input, y (n) is the regulator output, K_p，K_iRespectively, proportional and integral constants. Reference value omega of the rotational speed_refDifference Δ ω (n) from the estimated rotation speed to ω_ref- ω (n) input speedA degree regulator for obtaining q-axis reference current i by proportional-integral operation_{q_ref}(n)，

And (3) finding the magnitude of the torque of the load according to the estimated position lookup table in each control period by the rotor position estimated in the step (3), and obtaining the magnitude of the required electromagnetic torque, thereby directly calculating the magnitude of the required torque current as follows:

will i_{q_ref}′+i_{q_ref}And i_qDifference value Δ i of_q(n)＝i_{q_ref}′(n)+i_{q_ref}(n)-i_q(n) inputting the torque current regulator, outputting as q-axis reference voltage u by proportional-integral operation_{q_ref}(n) that is

i_{d_ref}0 and i_dDifference value Δ i of_d(n)＝i_{d_ref}(n)-i_d(n) inputting an excitation current regulator, and outputting as a d-axis reference voltage u through proportional-integral operation_{d_ref}(n)，

5. The voltage vector in the α β axis is calculated. Conversion from dq0 coordinate system to α β 0 coordinate system according to u_{q_ref}、u_{q_ref}The voltage vector in the dq0 coordinate system is calculated.

6. And obtaining a switching signal of the inverter by using a space vector PWM method.

The switching states of the power devices in the three bridge arms of the inverter are respectively three switching states S_A，S_B，S_CAnd (4) showing. Let S_i(i ═ a, B, C) ═ 1 denotes inverters a, B, with the power devices on the C arm conducting, S_iAnd (i ═ a, B, C) ═ 0 indicates that the power devices below the arms of the inverters a, B, C are on. According to S_A，S_B，S_CThe inverter has 8 operating states in total. Corresponding to 8 operating states, there are 8 voltage space vectors V0-V7, wherein V1-V6 are non-zero vectors and have amplitude E_dc(DC bus voltage), V0, V7 are zero vectors and the amplitude is 0. The spatial distribution of V0-V7 is shown in FIG. 8. a.

Dividing the voltage vector distribution space into 6 regions according to the voltage u in each control period_α，u_βThe resultant space vector (amplitude of:. alpha. beta.) under the α. beta. axis is obtained

The argument is gamma, the angle of the lobe,

) The sector in which the switching vector is selected is suitable. Within each zone V_sTwo non-zero vectors V of the region boundary can be utilized_a，V_bAnd zero vector V0 or V7 by properly controlling their action time T_a，T_b，T₀A reference voltage vector is obtained. As shown in fig. 8.b, V may be included in each region_sT，V_aT_a，V_bT_bThe action time T is obtained according to the sine theorem of the triangle_a，T_b，T₀：

T₀＝T-T_a-T_b

Wherein T is a control period, E_dcIs the dc bus voltage.

Therefore, the on-off time of each switching tube is obtained, and the corresponding PWM pulse is sent out to control the output voltage of the inverter, so that the control of the permanent magnet synchronous motor is realized.

7. It is determined whether the control routine is suspended. If not, returning to the step 2 to start execution again; if yes, the data display waveform is recorded, and the control program is stopped.

The software calculation flow of the above system is shown in fig. 6.

Claims

1. The control method of the speed sensorless permanent magnet synchronous motor-air conditioner compressor system is characterized by sequentially comprising the following steps of:

(1) system initialization

in the DSP, the following are set:

reference value omega of the rotational speed_ref0, excitation current reference value i_{d_ref}＝0；

the control period T is a set value;

Parameters of the permanent magnet synchronous machine provided by the user: number of pole pairs p_nPermanent magnet linkage psi of rotor_r，d_qShaft inductance L_d，L_qStator resistance R, back emf coefficient K_E(ii) a At the same time, the initial value of the dq-axis voltage of the motor is set to zero, i.e., u_d(0)＝0，u_q(0)＝0；

For u with n ≧ 1_d，u_qTaking the actual reference dq axis voltage calculated in the last digital control period T in the DSP, namely u_d(n)＝u_{d_ref}(n-1)，u_q(n)＝u_{q_ref}(n-1)；

Setting: a model motor with exactly the same parameters is built on the y δ axis, which is the estimated rotor axis and has an angular error Δ θ from the dq axis, Δ θ being θ - θ_MTheta is the angle between the d axis and the reference axis + A, theta_MIs the included angle between the gamma axis and the reference axis + A;

setting initial time, rotation speed

Position theta_M(0)＝0，

Then:

e_M(n)＝e_M(n-1)-K_eΔi_δ(n)

wherein,

Δθ(n)＝θ(n)-θ_M(n)

p is a differential operator;

when the estimation reaches the steady state, Δ θ ≈ 0, so that the formula L_γ，L_δ，L_γδAre respectively approximately equal to L_d，L_q0, said L_γ，L_δ，L_γδAre respectively

wherein,

n is the number of time periods after discretization,

θ_Mis the estimated rotor position angle and is,

e_Mis the back-emf of the model motor,

e is the back-emf of the actual motor,

ω is the rotor speed, i.e.

u_γ，u_δthe voltages on the gamma, delta axes for a real motor,

The difference Δ ω (n) ═ ω_ref- ω (n) input to the speed regulator, determining the q-axis reference current i_{q_ref}(n)，

Wherein, K_p1，K_i1Proportional and integral constants of the speed regulator;

(6) the DSP obtains the switching signal of the inverter by using a space vector PWM method, obtains the on-off time of each switching tube of the inverter and sends out corresponding signalsThe PWM pulse controls the output voltage of the inverter to realize the control of the permanent magnet synchronous motor: in each control period T, the on-off time T of each switching tube in three bridge arms of the inverter_a，T_a，T₀：

T₀＝T-T_a-T_b

The space of voltage vector distribution is divided into 6 regions, and the voltage u in each control period_α，u_βThe resultant space vector magnitude under the α β axis is V_sThe argument is gamma; and the action time of the switching vector is determined by V in each region_sT，V_αT_a，V_bT_bThe triangle formed is determined, where V_α＝V_b＝E_dcT is a control period, E_dcIs a dc bus voltage; meanwhile, gamma also determines the area where the switching vector, i.e. the space vector, is located, i.e. the current operating state of the inverter.