TW201234762A - Motor control method with no sensor - Google Patents

Abstract

A motor control method with no sensor is applied to a permanent magnetic synchronous motor by combining a high frequency signal injection manner with a reference model adaptive control manner. The high frequency signal injection manner is taken as a starting strategy to achieve the function of sinusoidal driving of a compressor. When the compressor reaches a predetermined rotational speed, the system is switched to the reference model adaptive control manner suitable for intermediate and high rotational speeds, such that the compressor can be operated to reach the higher rotational speed. In addition, the intelligent artificial neural proportional differential and integral controller is taken to replace the conventional proportional differential and integral controller inside the estimators of the high frequency signal injection manner and the reference model adaptive control manner, thereby further improving the estimation performance of the motor.

Description

201234762 VI. Description of the invention: [Technical field to which the invention pertains] The present invention relates to a motor (four) method without a turnover, which is replaced by a method for using a high age number, and (4) a Huilai nerve closed differential integral controller. Traditionally, the differential integral (4)$ is used, and the combined reference material is used to adapt the system speed estimation rule to the sensorless motor (4) method. [Prior Art]

When discussing the inverter variable frequency control technology applied to cold, east air conditioning and other occasions, 'because the compressor is used in high temperature, plus the refrigerant has miscellaneous! · Health', therefore, it is impossible to install the air gap between the pressure motor. The controller or the speed sensor 'has to use the sensorless control law to achieve variable frequency control. In the case of a DC inverter compressor, the internal motor is a permanent magnet synchronous motor, which is divided into surface-mounting, embedded and built-in permanent magnet synchronous motors according to the way of rotor magnetism. The difference between the mandrel inductance and the q-axis inductance makes a slight change in the motor model. 〃 For the built-in permanent magnet synchronous motor, the inductance Ld is not equal to the inductance Lq, and there is a clear salient effect. Therefore, there is no sensor control law to estimate the rotor flux angle based on the characteristics of the salient pole. position. However, the surface-mounting type 5 5 motor's inductance Ld is equal to the inductance Lq, so its salient pole effect is not as good as the built-in permanent magnet synchronous motor (4), so the non-sensing control law applied to the salient pole characteristics cannot be applied. On the surface-mounted permanent magnet synchronous motor. There are three common techniques for permanent magnet synchronous motor control technology without sensor. Anti-TF1004505 201234762 Electromotive force zero parent over-point method, rotor flux estimation method and reference model adaptive control method can be applied to salient or hidden On the pole motor, the techniques are briefly described below. 1. Back EMF Zero Crossing Detection Method: A non-sensing ϋ circuit is designed based on detecting the back electromotive zero crossing point of the compressor motor, thereby obtaining a commutation signal instead of the Hall sensor. 2. Rotor Flux Estimation Method: Estimate the motor's fixed Luzi flux angle by using the three-phase electric house and current of the motor, and obtain the rotor flux angle by calculating the compensation of (4). . 3. Reference model adaptive control method Based on the motor model, a modulation model is established, and an adaptive mechanism is used to estimate I; the motor speed 'directly obtains the motor rotor flux position by the integrator. The common defect of the above-mentioned sensorless control law is that when the motor is running at low speed or at rest, the back electromotive force cannot be measured, the measured back electromotive force is too small, or the angle is Problems such as the initial value make the above method unsuitable for the start-up state, and an additional start-up strategy must be used to help the press and the motor to start to the medium-high speed. Open circuit square wave start-up is a common starting gas for household air-conditioning compressors. The advantage is easy to achieve. The disadvantage is that when this type of start-up method is used to start the compressor ^ #巧达 will be accompanied by a large starting current, adding machinery Wear and shrink TF1004505 4 201234762 When the short compressor is running, and the square wave drive is switched to the sine wave drive, the current waveform at the switching point will change instantaneously. If it is not properly designed, it may produce a momentary violent The torque ripples and even causes the compressor to stop running. The high-frequency injection method proposed in recent years can be applied to the motor control technology of non-sensing state at zero rotation speed, and it has been extensively studied, and it can be effectively started whether using a salient-pole motor or a hidden-pole motor. The rotor can be controlled at a low speed range. By means of the rotor angle estimation technology without the sensor, even at zero speed, the rotor flux position can be obtained by means of high frequency signal modulation, avoiding a large start. The problem of current; however, when the motor reaches the set switching speed after starting, 'there is still the rotor following the up and down swing, which cannot be accurately followed. SUMMARY OF THE INVENTION The present invention provides a sensorless motor control method that replaces a conventional proportional-integral-derivative control module with a smart-type proportional-integral-integral-differential control (four) group to improve the high-frequency signal injection method. In order to improve the performance of the high-frequency signal injection knife at zero speed and low turnover, and (4) the proportional production model combined with the reference model adaptive control method 'according to the motor operation second reading current and motor electrical parameters The adjustable speed is obtained by the adjustable «type and adaptive mechanism, so as long as the current sampling frequency is high enough, the reference speed is not limited. (4) The operation speed of the adaptive control method is not limited, which is quite easy to implement in the hardware of the compressor. The drive control of the architecture. Tian Jin TF1004505 201234762 This: (4) The first aspect provides a motor-free method of sensorless operation. The side method operates in the motor as a static guard step. · w speed method includes the lower field 1 controller sends out - control signal To drive the controller to convert the high frequency signal coordinate into the control signal; the 〃 中 此 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 接收 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气 电气And a class of neural proportional integral differential operation is obtained to obtain an estimated rotational speed and an estimated measurement degree, wherein one of the proportional parameter, the differential parameter and the individual learning weight value according to the integral parameter in the neural proportional integral differential operation Each time the estimated error is changed to be larger than the side size, thereby lightly comparing the ratio, the differential parameter and the integral parameter; and the micro-control H compares the commanded speed with the _phase to obtain a command control signal; The high frequency signal and the command control signal are summed by the microcontroller to obtain the control signal for driving the motor. According to the method of the first aspect of the present invention, the following steps are performed by the k-number processing and the motor speed estimation module of the microprocessor: the electrical signal is subjected to a band pass filtering process to obtain a high frequency signal form. a surface frequency electrical signal; the high frequency electrical signal and the high frequency signal form of the high frequency signal adjustment value are multiplied to obtain a DC power supply form of the DC electrical signal τρ1004505 6 201234762 'where the high frequency signal is adjusted The value is a high-frequency voltage angular frequency, and the DC electrical signal is subjected to a low-pass filtering process to obtain an input signal related to the estimated error angle; the input signal is subjected to a neural-like proportional integral differential operation to obtain —车备速号' where the equation for this type of neural proportional integral differential operation is: (^) = Kp (N)e + Kd (N) ^ + Kj (A^)| edt p + = Kp (N)e + 7, S0e sgn(e) + 0 = Kd(N)e + ηλδ^ — dt (N + 1) = Kt (N)e + 77, J edt sgn(e) = e if e>〇sgn(e ) = -e if e <0

Among them, 'nine is the output of the proportional integral differential operation of the neuron, which is the proportional parameter, (for the differential parameter, (for the integral parameter, the error is estimated by the error angle, ~ is the weighting parameter, and 4 is set to e, 7| is the weight Learning rate; ^ low-pass filtering the speed signal and multiplying the parameter related to the number of teas of the motor to obtain the estimated speed; and performing an integral operation on the speed signal to obtain the estimated angle e Executing the method according to the first aspect of the present invention, wherein the micro-control is performed by the following steps: by the one of the microcontrollers, the switching wheel set refers to the estimated angle to the two-phase stationary coordinate of the electrical Reed etch ~ gas signal ~ a 1 #U converted into synchronous rotating coordinates - dq electric TF1004505 201234762 The signal processing and motor speed estimation module performs the above signal processing and a class of neural proportional integral differential operation steps, The dq axis electrical signal is calculated to obtain the estimated rotational speed and the estimated angle; and the speed control module of the microcontroller compares the commanded rotational speed with the estimated rotational speed to obtain a speed difference, and performing a proportional integration operation on the speed difference to obtain a current command signal; wherein the current control module of the microcontroller compares the dq axis electrical signal with the current command signal to obtain a current difference And performing the proportional integral operation on the current difference to obtain a voltage signal; decoupling the voltage signal by the voltage decoupling module of the microcontroller to obtain a decoupling voltage signal, wherein the decoupling operation system Compensating for the interference of the estimated rotational speed on the electrical signal of the dq axis and the coupling between the two; and summing the high frequency signal generated by the high frequency signal injection module of the microcontroller and the combined surface voltage signal To obtain a voltage command signal, wherein the south frequency signal injection module generates the south frequency signal of the south frequency voltage, and the expression is as follows:

Vdl, 乂丨, 〇-idh _ 0 zql, _ where vrf/l, respectively, the d-axis and q-axis high-frequency power components ' &, ◊' of the synchronous reference coordinates are respectively d of the synchronous reference coordinates The south frequency current component of the shaft and the q sleeve, Ζί/Λ, 'the high frequency impedance of the d-axis and the q-axis of the synchronous reference coordinate of the motor respectively; TF1004505 8 201234762 same, the anti-coordinate conversion module refers to the estimation The angle will be σ/疋^ which is the command signal coordinate converted into one of the three three-phase control signals; and the call is made by the (four) 11-chord-wave width modulation module. The two-phase control signal performs sine wave pulse width modulation to obtain = according to the method of the present invention, and further comprises the following steps: providing a DC power supply to the converter by a power supply; is isolated by an isolation amplifier The microcontroller transmits an electronic signal to the electronic control signal, and the control signal is amplified into an electrical signal and an electrical control signal; and the fly=inverter converts the DC power into a three-phase power supply, and generates a signal. Control the size of the three (four) community, read the person to the expected. ‘ = brother's method, in which the microcontroller is driven to a predetermined low speed according to • 2: = the motor is controlled by σ. This high frequency signal is not generated. The fourth aspect of the present invention provides a motor control method without a sensor, and the following steps are performed on the motor for the speed or high rotation of the towel: ι Each is sent by the microcontroller - the control signal is Driving the motor; =: control: receiving an electrical signal of the motor when it is driven; according to - command electric, signal, an estimated electrical speed and the motor's electric rolling parameters to calculate 丨 - J estimated Current signal, and according to the estimated TF1004505 9 201234762 = signal, the electrical signal, the motor's electrical parameters and - the type of neural ratio f differential operation (four) of the nuclear return, (four) to - (10) speed and a measured angle 'the _ Through the proportional integral differential operation - the proportional reference W-parameter and the individual learning weight value according to the integral parameter with each time: the change of the measurement error is the size of the case, and then the _ parameter, 5 Hai differential parameter and The integration parameter; and driving the motor by comparing a commanded speed with the estimated speed to obtain the control signal. According to the second aspect of the present invention, t, the processing of the motor and the motor speed estimation module are performed by the following steps: According to the command (4) money, the electrical parameter of the speed touch motor The adjustable variable model is calculated according to the adjustable variable model, wherein the adjustable variable model is ··· d_ ~dt hi L Ld ώ.

Ss_ Ld L,

Ld, where & 9 is the estimated current signal, ~ is the estimated electrical speed, the resistance parameter is reached, ^, 々 is the motor of the thunder v. · ......... pressure signal; cover one · the inductive parameters of the transport And the command circuit generates an adaptive mechanism according to the estimated current signal, the electrical signal and the electrical parameter of the motor, and calculates according to the adaptive mechanism to obtain the estimated electrical speed, wherein the adaptive mechanism includes: Neuron: Example ^ points: TF1004505 10 201234762 The difference is the following: ●) Η · 'This kind of nerve is better than the one, the Λ is the analogy of the proportional-integral differential operation. The integral equation of the differential operation is: Λ (N) = Kp (N)e + ΚΛΝ)^ + Κ/ (Λ,} j edt Kp{N + l) = Kp^e + JlAesgn{e)

Kd(N + \) = Kd(N)e + n color dt K(m) = K(N)e + #〇jedt

Sgn(e) = e if e>〇sgn(e) = -e if e<Q where,, "4 electric money, (for the back electromotive force constant, parameter ' (for the differential parameter is the integral parameter, - error angle, ^ Weighting parameter, and (4) is , , ^ is the estimated rate; 1 is the learning speed of the weighting, the estimated electrical speed is integrated to obtain the estimated electrical speed multiplied by the motor and Obtaining the estimated speed. The parameter of counting money according to the second aspect of the present invention is as follows: ,, the microcontroller performs the electrical signal signal of the stationary coordinate of one of the microcontrollers; The standard conversion module refers to the estimated angle to convert one of the three coordinates of the dq axis into a synchronous rotation TF1004505 201234762. The signal processing and the motor speed estimation module perform the above signal processing and a step of the neural proportional integral differential operation, The dq axis electrical signal is calculated to obtain the estimated rotational speed and the estimated angle; and the speed control module of the microcontroller compares the commanded rotational speed with the estimated rotational speed to obtain a speed difference, and the speed is obtained Difference Performing a proportional integral operation to obtain a current command signal; comparing, by the current control module of the microcontroller, the dq axis electrical signal and the current command signal to obtain a current difference, and performing the current difference Integrating to obtain a voltage signal; decoupling the voltage signal by the voltage decoupling module of the microcontroller to obtain a voltage command signal, wherein the decoupling operation cancels the estimated rotational speed of the dq axis electrical signal Interference and the integration between the two; the anti-coordinate conversion module of the microcontroller refers to the estimated angle, and the voltage command signal of the synchronous rotating coordinate is converted into a three-phase static coordinate by the coordinate And controlling, by the sine wave pulse width modulation module of the microcontroller, the three-phase control signal to perform sine wave pulse width modulation on the estimated angle to obtain the control signal. According to the second aspect of the present invention The method of the aspect further includes the following steps: providing a DC power supply from a power supply to an inverter; being isolated by the isolation amplifier by the microcontroller The control signal in the form of an electronic signal 'enlarges the control signal into an electrical control signal in the form of an electrical signal; and TF1004505 12 201234762 converts the DC power source into a two-phase power source by the converter' and controls three according to the electrical control signal The size of the phase power supply is input to the motor. [Embodiment] The following drawings are used to illustrate the preferred embodiment of the present invention. Fig. 1 is a system block diagram of a motor control method without a sensor of the present invention. In a power supply 22, a power supply 22 supplies DC power to a voltage source converter 24. An isolation amplifier 26 isolates a control signal transmitted by a microcontroller 28 into an electronic signal and amplifies the control signal to One of the electrical signal forms is an electrical control signal for output to the voltage source inverter 24. The purpose of the isolation amplifier 26 is to prevent the large power supply of the voltage source converter 24 from entering the microcontroller 28, causing electrical signal noise to affect the operation of the microcontroller 28 and even damage the microcontroller 28. The voltage source converter 24 converts the DC power into a three-phase power, and controls the size of the three-phase power according to the electrical control signal, and inputs the three-phase power to a permanent magnet synchronous motor 30 to drive the φ permanent magnet synchronization. The motor 30 operates. The microcontroller 28 receives the current signal of the permanent magnet synchronous motor 30 during operation, and the current embodiment uses a current signal as an electrical signal, but may also use a voltage signal as an electrical signal, and via the microcontroller 28 The operation of the module generates a new control signal whereby the new control signal drives the operation of the permanent magnet synchronous motor 30. Referring to the flow chart attached below and the system block diagram of Fig. 1, the steps of implementing the motor control method of the sensorless sensor of the present invention will be described. 2 is a flow chart of a motor control method for a sensorless sensor of the present invention. In the figure TF1004505 13 201234762 2, DC power is supplied from the power supply 22 to the voltage source converter 24 (step S50). The control signal in the form of an electronic signal transmitted by the microcontroller 28 is isolated by the isolation amplifier 26, and the control signal is amplified into an electrical control signal in the form of an electrical signal (step S52). The voltage source converter 24 converts the DC power source into a two-phase power source' and controls the size of the two-phase power source according to the electrical control signal, and inputs the three-phase power source to the permanent magnet synchronous motor 30 to drive the permanent magnet synchronous motor 30 is operated (step S54). When the permanent magnet synchronous motor 30 is in a stationary state or a low rotational speed, the microcontroller 28 controls the operation of the permanent magnet synchronous motor 30 by means of a high frequency signal injection method (step S56). The embodiment of the micro-controller 28 uses the high-frequency signal injection method to control the operation of the permanent magnet synchronous motor 30. FIG. 3 is a flowchart of the operation of the microcontroller using the high-frequency signal injection method to control the operation of the permanent magnet synchronous motor. The figure shows. In FIG. 3, a current signal ~ is received by the microcontroller 28, and a positive coordinate conversion module 32 of the microcontroller 28 refers to an estimated angle & a current of a three-phase stationary coordinate (ie, a phase difference of 120 degrees per phase) The signal ^ is converted into one of the synchronous rotation coordinates dq axis current signal ~, Η step S60). The estimated angle is used to estimate the rotor flux angle of one of the permanent magnet synchronous motors 30. Wherein, the current signal G, 卩 of the three-phase stationary coordinate is converted into one of the synchronous rotating coordinates, and the conversion matrix of the dq axis current signal ~~ is as follows:

Cos 0re cosd -120. ) cosd +120. ) -sin^re -s\n(9re -120°) -sin(^e +120°) TF1004505 14 201234762 Since the permanent magnet synchronous motor 30 is operated at a standstill or at a low speed, it is output by the microcontroller 28. The control signal has a component of a high frequency signal, that is, a high frequency signal injection module 34 of the microcontroller 28 generates a high frequency voltage, which is expressed as follows: X, 0 'ldh 0 } φ. where '% is synchronous The high-frequency voltage component of the d-axis and the q-axis of the reference coordinate is the peak of the synchronous reference coordinate and the high-frequency current component of the ❻, 4, the heart, respectively, the synchronous reference frame of the permanent-magnet synchronous motor 3 〇 High frequency impedance of the axis and q axis. Ld Thus, the dq-axis current signal received by the microcontroller 28 and converted by the coordinates has a high frequency signal component, which is expressed as follows: ^arg c〇s coscv V2Zdifrsm2d

V/J

—)%C0SiV > which is the synchronous reference d-axis and the q-axis called the axis, the frequency current money 'which is called the axis current signal & the age component, ~ is the high-frequency voltage ' ^ for permanent magnet synchronization The average value of the high frequency impedance of the motor 30 'Z• is the high frequency impedance difference of the permanent magnet synchronous motor 30, and the % is the high frequency voltage angular frequency. Where a, g

Zdh+ z

Zdh-h group 36 will be subjected to signal processing and motor speed estimation TF1004505 15 201234762 dq axis current signal ~~ for signal processing and a kind of neural proportional integral differential operation to obtain an estimated mechanical speed and estimate Angle seven (step S62). The signal processing and motor speed estimation module 36 implements the signal processing and the neural-like proportional integral differential operation of the dq-axis current signal~~, and the high-frequency signal is implemented in the signal processing and motor speed estimation module of the present invention. The block diagram of the injection method and FIG. 5 are the flow chart of the signal processing and the motor-like proportional integral differential operation of the signal processing and motor speed estimation module of the present invention. In FIGS. 4 and 5, a band pass filter 80 of the signal processing and motor speed estimation module 36 performs a band pass filtering process on the q-axis current signal of the dq axis current signal having a high frequency signal. The frequency other than the frequency of the high frequency signal is filtered out to obtain a high frequency current signal b which is one of the high frequency signal forms (step S100). Among them, the high-frequency current signal & expression is as follows: V = vm / sin 20re Ί 2r r cos -% 厶哳Sin fiV) ^h^dh^qh =ere~^re where Α/λ, ^ respectively d The high-frequency stator inductance of the shaft and the q-axis is the high-frequency resistance difference between the d-axis and the q-axis of the permanent-magnet synchronous motor 30, and the high-frequency inductance between the d-axis and the q-axis of the permanent-magnet synchronous motor 30. The difference, t is the estimated error angle of the rotor flux angle position, ~ is the actual angle of the rotor flux angle position, and & is the estimated angle of the rotor flux angle position. A multiplier 80 of the signal processing and motor speed estimation module 36 multiplies the high frequency current signal & and the high frequency signal adjustment value sin~ in the form of a high frequency signal by a TF1004505 16 201234762 multiplication operation to obtain a DC power supply. One of the forms of the direct current signal Usin " V (step S102). The purpose of multiplying the high-frequency current signal & by the high-frequency signal adjustment value sin_ is to eliminate the shadow of cos(7)"/-sin(7)〆) in the high-frequency current signal. The low-pass filter 84 of the signal processing and motor speed estimation module 36 converts the DC current signal ksin (v to perform a low-pass filtering process to filter out the high-frequency signal to obtain a current related to the estimated error angle & Input signal ~ (step φ step S104) where -sin 21 k] = -4 from ", ζ, if the estimated error of the rotor position is very small, Sin2& will approach to < then current input signal & Linearization to Κ Θ error re Figure 6 is a network architecture diagram of the PIDNN controller of the present invention. PIDN (proportional-integral-derivative) is one of the signal processing and motor speed estimation modules 36 in Figures 4 and 6. The neural network, the controller 86 performs a neural-integral-integral-differential operation on the current input signal k to obtain a rotational speed signal A (step S106), wherein the neural proportional integral differential The equation for the operation is: y0(N) = M^] +w202+M>303=Kp(N)e + Kd(N)^ + K,(N)\edt TF1004505 Μ 201234762 In order to derive the weights M;1 , W2,% adjustment formula, first define the energy function as follows: E = X - (dm-df = X-e2 by The energy function defined above is derived by the inverse transfer method. The formula for adjusting the weight is as follows: △W, - = -7/, dE dw, ' = -7, dE dy0(N) dy0(N) ^ =7ΑΘ, where, 77 'Represents the learning rate of the weight of the i-th item. Further discussion of the weighting parameter & is derived as follows: „ _ dE _ dE de _ dE de dd ° dy〇(N) de dy0(N) de dd dy0(N) The partial differential must accurately determine the sensitivity of the system. Considering the sensitivity of the actual system is not easy to obtain, the weighting parameter & is directly assumed to be the error e, so the weight correction formula is as follows: wi (A^ +1) = (N) + Awf The output and adjustment formula of the proportional-integral differential operation can be derived by the above formula: ^0(iV) = Kp(N)e + Kd(N)^ + ΚΧΝ)1 edt Kp(N + \) = Kp(N)e + 7hSQe

Kd(N + l) = Kd(N)e + ^S0^- at (N + \) = Ki (N)e + η, δ0\ edt Since the weighting parameter & is assumed to be the error e, causing the heart The correction term is always greater than zero, and V2 causes the heart to accumulate upwards, so the heart (〃 + 1) is corrected. TF1004505 18 201234762 is as follows: Κρ{ΝΛΛ) = Κρ{Ν)β + η,δϋε sgn(e) sgn( e) = e if e>Q sgn(e) = -e if e <0 - where y. For the analogy of the proportional-integral differential operation, the heart is the proportional parameter, the heart is the differential parameter, and the & is the integral parameter. In Figure 4, the error e is proportional to the estimated error angle & relationship. The purpose of the design controller 86 is to ensure that when the error e is increased, the parameter value of the Lu PIDNN controller 86 is 夂"夂", and the force can be increased relative to the booster port. The response of the synchronous motor 30, and as the error e decreases, the parameter value heart & of the PIDNN controller 86 can be relatively reduced, causing the microcontroller 28 to control the permanent magnet synchronous motor 30 to smoothly enter a steady state. Since the PIDNN controller 86 does not require a complicated calculation process, writing the PIDNN controller 86 in C can be easily implemented and applied to the implementation φ, and does not increase too much compared to the conventional proportional integral derivative controller. The computing time also has the ability to adapt to the parameters of self-adaptation. Since the high-frequency signal injection method requires a complicated signal processing process to obtain the rotor flux position, and the parameter design of the controller is quite difficult, if the PIDNN controller 86 is used, the self-adaptation capability can be utilized to optimize the parameters of the controller. The time for designing the controller parameters can be greatly reduced, and in addition to effectively shortening the transient time of the angle estimation, the efficiency of the high frequency signal injection method for estimating the angle can be effectively improved. TF1004505 19 201234762 Next, low pass filtering by one of the signal processing and motor speed estimation module 36

The dL unit 90 performs a low-pass filtering process on the rotational speed signal &amp; to filter out the high frequency signal to obtain an estimated electrical rotational speed of the permanent magnet synchronous motor 30, and is multiplied by one of the signal processing and motor speed estimation module 36. The controller 92 estimates the electric speed &lt; multiply 2_ the parameter 7 relating to the number of poles of the permanent magnet synchronous motor 30 to obtain the estimated mechanical rotational speed of the permanent magnet synchronous motor 30 (step S108). The integrator 88, which is one of the signal processing and motor speed estimation module 36, performs an integral operation on the rotational speed signal & to obtain an estimated angle 匕 (step S109). Referring again to FIGS. 1 and 3, a speed control module 38 of the microcontroller 28 compares a command mechanical speed ω'&quot; with an estimated mechanical speed ω&quot; to obtain a speed difference, and is controlled by the speed control module 38. A proportional integrator (not shown) performs a proportional integration operation on the speed difference to obtain a current command signal < (step S64). Wherein, the command mechanical speed &lt; is a preset value. The current control module 40 of the microcontroller 28 compares the dq axis current signals ~, ~ and the current command signals &lt;, &lt; respectively to obtain two current difference values, and integrates the two ratios of the current control module 40. The two current difference values are proportionally integrated to obtain a voltage signal 〜, Μ step S66). Wherein, the current command signal &lt; is preset to be 0. The d-axis voltage decoupling module 42 of the microcontroller 28 decouples the voltage signal ν&quot; from a d-axis decoupling value (ie, subtraction) to obtain a TD1004505 20 201234762-d-axis decoupling voltage command signal. The q-axis voltage decoupling module 44 of the microcontroller 28 decouples the voltage signal & with a q-axis decoupling value 4 &amp; + + « to perform a decoupling operation (ie, addition) to obtain a voltage command signal &lt; S68). Among them, the decoupling operation cancels the estimated mechanical rotation speed &lt;the interference to the dq axis current signal ~~ and the coupling between the two. The high frequency signal generated by the high frequency signal injection module 34 and the d-axis decoupling voltage command signal are summed to obtain a voltage command signal &lt; (step S70). _ The anti-coordinate conversion module 46 of the microcontroller 28 refers to the estimated angle t which will be the voltage command signal of the synchronous rotating coordinate &lt;, v; coordinate converted into a three-phase control signal of one of the three-phase stationary coordinates (step S72). Wherein, the conversion matrix of the three-phase control signal of the three-phase static coordinate of the voltage command signal of the synchronous rotating coordinate is converted into the inverse matrix of the three-phase static coordinate, which is the inverse matrix of the conversion matrix, that is, by the microcontroller. The 8 sine wave pulse width modulation module 4 8 refers to the estimated mechanical angle ^&quot; 'the sine wave pulse width modulation of the three-phase control signal to obtain a control signal output to the isolation amplifier 26 (step S74), wherein the sine wave pulse width modulation module 48 performs the sine wave pulse width modulation technique using the well-known sine wave pulse width modulation technique. The high frequency of the embodiment is illustrated by an operation example. The signal injection method combines the performance difference between the PIDNN controller and the high-frequency signal injection method in combination with the conventional PID controller, as shown in Fig. 7 is the high-frequency signal injection method combined with the conventional PID controller for speed measurement and follow-up in the no-load condition. The waveform of the response is shown in Fig. 7. The high-frequency signal injection method is combined with the ΡID ΝΝ controller to measure the speed of the controller in the no-load condition and the waveform of the TF1004505 21 201234762 follow-up response, and FIG. 7C is the high-frequency signal method combined with the conventional pi. Controller is adding 10 The waveform of the speed measurement and follow-up response of the load condition of kg-cm is shown in Fig. 7D, the frequency-frequency signal injection method combined with the piDNN control, and the waveform of the speed measurement and follow-up response of the load of 10 kg-cm is shown. The permanent magnet synchronous motor 30 is started by the high frequency signal input method. The commanded mechanical speed of the low speed is set to &lt; 15 rpm, the injected voltage is 15 V, and the high frequency signal frequency is 280 Hz. The permanent magnet synchronous motor 30 In the no-load situation, it can be understood from the waveforms of FIGS. 7A and 7B that the PIDNN controller 86 can instantly adjust the setting of the parameter value center, A, ' with the error e, thereby shortening the estimated mechanical speed &amp;'&quot; The transient time and the reduction of the steady-state error, the speed follow-up response of the system can also be better, and compensate for the small error of the rotor position error signal obtained by the signal modulation when the amplitude of the injected signal is low. The system still has good control performance. Consider the case where the permanent magnet synchronous motor 30 uses the load of 1 〇kg-cm provided by the magnetic powder brake. From the waveforms of Fig. 7C and 7D, it can be understood that 'the high frequency signal injection method combined with PIDNN The controller can have better robustness for driving the permanent magnet synchronous motor 30 than the high frequency signal injection method in combination with the conventional PID controller. Referring again to Figures 1, 2, the microcontroller 28 is based on the estimated machine. The rotational speed ώ&quot; 'determines that the permanent magnet synchronous motor 30 is driven to a predetermined low rotational speed, the high frequency signal injection module 34 does not generate a high frequency signal. The permanent magnet synchronous motor TFI004505 22 201234762 30 is driven from the predetermined low rotational speed. When operating at a medium to high rotational speed, the microcontroller 28 controls the operation of the permanent magnet synchronous motor 30 using the reference model adaptive control method (step S58). The microcontroller 28 is controlled by the reference model adaptive control method - an embodiment of the operation of the permanent magnet synchronous motor 30 is as shown in Fig. 8. The controller uses the reference model adaptive control method to control the permanent magnet synchronous motor The flow chart of the operation is shown. The description of step S110 of Fig. 8 is the same as that of step S60 of Fig. 3, and will not be described here. The signal processing and motor speed estimation module 36 is based on the command voltage signal &lt;, &lt;, estimating the electrical mechanical speed &lt; and the resistance parameter of the permanent magnet synchronous motor 30, the inductance parameter L, &amp; Calculate the estimated current signal ~~, and calculate according to the estimated current signal ~dq axis current signal ~~, the electrical parameters of the permanent magnet synchronous motor 30 and the parameters required for a class of neural proportional integral differential operation to obtain an estimate Mechanical speed 6 &quot;, and estimated angle • &amp; (step S112). FIG. 9 is a block diagram of a reference model adaptive control method for signal processing and motor speed estimation module of the present invention, and FIG. 10 is a signal processing of the present invention. And the motor speed estimation module implements a flow chart of the reference model adaptive control method. In FIGS. 9 and 10, the adjustable module 140 of the signal processing and motor speed estimating module 36 is based on the command voltage signal &lt;, v "estimating the electrical speed 夂 · and the resistance of the permanent magnet synchronous motor 30 Parameter ~, inductance parameter heart, ^ and other electrical parameters TF1004505 23 201234762 According to the adjustable variable model to calculate a number to establish a variable model (step S150 ·), get the estimated current signal B, where the 'adjustable variable model is: d 7t ld

_£s_ L

L

Ld ώ„ t-L,

IX

The adaptive mechanism module 142 of the signal processing and motor speed estimation module 36 is based on the estimated current money ddq axis current synchronous motor 30 of the electronic 夂 | 俛 4-s ^ , magnetic type; ^ number and - class The neural proportional integral differential operation requires a number 〃, "if calculated, to establish a _ adapt to the adaptive mechanism to calculate to obtain an estimated electrical speed, step S152." wherein the adaptive mechanism includes a neural-like proportional integral Micro-arms, the adaptive mechanism is:

La L· )) where /. For the case of the analogy of the proportional-integral differential operation, the system of integral differential operation is based on the reference of the 6th, and the equation of the (10)-class-integral-integral-differential operation is: 'Nerve ratio (7^) = Kp (N)e + Kd (N ) ~ + K( (N)j edt p {N + \) = Kp (N)e + 7, i50esgn(e)

Kd (N+ 1) = Kd (N)e + ^S0~ dt

Kt (TV +1) = Ki (N)e + δ0 J edt sgn(e) = e if e&gt;〇sgn(e) = ~e if e &lt;〇TF1004505 24 201234762 where ~, &amp; is an electrical signal , &amp; is the back electromotive force constant, the heart is the proportional parameter 'heart is the differential mic 5 ' for the integral parameter 5 error e is set to the estimated error angle, & is the weighting parameter, and &amp; is set to e, % is the weight Learning speed. Rate. The integrator 144, which is one of the signal processing and motor speed estimation module 36, performs an integral operation to obtain an estimated angle & (step S15 4). The multiplier 146 of the signal processing and motor speed estimation module 36 estimates the φ electrical rotation speed &lt; multiplies the parameter 2_ 有关 related to the number of poles of the permanent magnet synchronous motor 30 to obtain an estimated mechanical rotational speed ώ&quot; (Step S156). Referring again to Figures 1 and 8, the description of steps S114 and S116 of Figure 8 is the same as that of steps S64 and S66 of Figure 3, and will not be further described herein. The d-axis voltage decoupling module 42 of the microcontroller 28 decouples the voltage signal ν&quot; from a d-axis decoupling value (ie, subtraction) to obtain a voltage command signal &lt;RTIgt; The q-axis voltage decoupling module 44 performs a decompression operation (ie, addition) on the voltage signal 〜 and a q-axis decompression value +~ force to obtain a voltage command signal (step S118). Wherein, the decoupling operation cancels the interference of the estimated mechanical rotation speed ώ'" on the dq axis current signal L~ and the engagement between the two. ' The description of steps S120 and S122 in FIG. 8 is the same as the step in FIG. _ S72, S74 description, no more details here. Permanent magnet synchronous motor 30 uses the reference model adaptive control method to run at TF1004505 25 201234762 high speed. Figure 11 shows the simulated permanent magnet synchronous motor operation of the present invention. Illustration at 50 (h.pm to 2000 rpm to 500 rpm, where () is a graphical representation of the actual mechanical speed and estimated mechanical speed ώ '&quot; and a graphical representation of the stomach heart and estimated angle & 12 is a diagram showing the case where the pseudo permanent magnet type motor of the present invention is operated at 2000 rpm, wherein (a) is the actual engine speed % and the estimated mechanical speed &lt; and (b) is the actual angle As shown in the figure of the estimated angle I. In Fig. 11, the permanent magnet synchronous motor 30 is operated at a rotational speed of 500 rpm to 20 rpm to 500 rpm, and the permanent magnet type synchronization using the reference model adaptive control method is employed. Motor 30 has good speed, angle estimation and follow-up In Fig. 12, the permanent magnet synchronous motor 3 is operated at a steady state speed of 2 jaws, and the actual mechanical speed of the permanent magnet synchronous motor 3 of the reference model adaptability (four) method is estimated and estimated. Measuring the mechanical speed ώ„, the maximum error amount is =Pm' The actual angle I and the estimated angle, the error is within $degree, and there is also a good speed and angle estimation effect. The advantage of the present invention is to provide a no The motor control method of the sensor replaces the traditional proportional product knife U dn group with a neural proportional integral differential control module such as Qi Hui 5L to improve the high frequency signal injection, thereby improving the high frequency signal injection method It is difficult to estimate the zero-speed and low-speed, and the neuron proportional-knife micro-knife Uj is combined with the reference mode (four) Dependability (four) method, and the information is based on the voltage and current of the motor and the electrical parameters of the motor. Variable model A should be ί! mechanism to obtain the estimated speed' so as long as the current sampling frequency is high enough TF1004505 26 201234762, the reference model adaptive control method will not be limited in speed, so it is quite easy to implement the hardware architecture of the compressor. The present invention has been described above with reference to the preferred embodiments and the accompanying drawings, and should not be considered as a limitation. The skilled person will make various modifications and omissions to the form and specific examples thereof. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a system for controlling a motor without a sensor according to the present invention; FIG. 2 is a non-inductive method of the present invention. FIG. 3 is a flow chart of controlling the operation of the permanent magnet synchronous motor by using the high frequency signal injection method of the microcontroller of the present invention; FIG. 4 is a signal processing and motor speed estimation of the present invention; Figure 5 is a block diagram of the test module; Figure 5 is a flow chart of signal processing and motor-like proportional integral differential operation of the signal processing and motor speed estimation module of the present invention; Figure 6 is a network architecture diagram of the PIDNN controller of the present invention; Figure 7A is a waveform diagram of the high-frequency signal injection method combined with the conventional PID controller for speed measurement and follow-up response in the no-load condition; Figure 7B shows the high-frequency signal injection method combined with the PIDNN controller in the no-load Figure 7C is a waveform diagram of the high-frequency signal injection method combined with the conventional PID controller with a speed measurement of 1 Okg-cm and a follow-up response; Figure 7D shows the waveform of the high-frequency signal injection method; High-frequency signal injection method combined with the PID signal of the PIDNN controller plus the load condition of TF1004505 27 201234762 lOkg-cm. Figure 8 is the microcontroller of the present invention using the reference model adaptive control j ^ control FIG. 9 is a block diagram of a spring model adaptive control method for signal processing and motor speed estimation module of the present invention; FIG. 1 is a signal processing and motor of the present invention. Flow chart of the speed estimation model-expansion model adaptive control method; 4 Figure η is a graphical representation of the simulated permanent magnet synchronous motor running in the range of 5〇〇2_ah to 5_m, where (a) A graphical representation of the actual machine % to % and the estimated mechanical speed % and (b) a plot of the actual angular center and the estimated angle; and the degree 12 is the simulated permanent magnet synchronous motor operation of the present invention. In the case of 2000 rpm, it (A) the actual mechanical speed and estimated mechanical speed nine% of the drawings and (b) to estimate the actual angle and the angle I ^ illustration. [Main component symbol description] 22 Power supply 24 Voltage source converter 26 Isolation amplifier 28 Microcontroller 30 Permanent magnet synchronous motor 32 Positive coordinate conversion module 34 High frequency signal injection module TF1004505 28 201234762 36 Signal processing and motor Speed estimation module 38 speed control module 40 current control module 42 d-axis voltage decoupling module 44 q-axis voltage decoupling module 46 anti-coordinate conversion module 48 sine wave pulse width modulation module φ 80 band pass Filter 82 Multiplier 84 Low Pass Filter 86 PIDNN Controller 88 Integrator 90 Low Pass Filter 92 Multiplier φ 140 Adjustable Module 142 Adaptive Mechanism Module 144 Integrator 146 Multiplier TF1004505 29

Claims (1)

201234762 VII. Patent application scope: 1. A motor control method without sense sensor, the method is operated when the motor is at a static or low speed, and the method comprises the following steps: A control signal is sent by the -micro control n to drive the motor, wherein the microcontroller converts a high frequency signal coordinate into the control signal; and the electronic controller receives an electrical signal when the motor is driven to operate; The micro-laid touches the electric money series signal processing; ^ 仃-class-like proportional-integral differential operation to obtain an estimated rotational speed and an estimated angle', wherein the proportional-integral differential operation of the neuro-proportional-integral-differential operation, the differential parameter And the individual learning weight value according to the integral parameter is proportional to the change of each estimation error, thereby adjusting the proportional parameter, the differential parameter and the integral parameter; 'Comparing a command speed with the microcontroller Estimating the rotational speed to obtain a command control signal; and P summing the high frequency signal and the command control signal by the microcontroller to obtain the control signal for driving the motor. 2. If you apply for a patent scope! The method of the item wherein the microprocessor-k processing and motor speed estimation module performs the following steps: performing a band pass filtering process on the electrical signal to obtain a high frequency electrical signal as a high frequency signal step The high-frequency electrical signal and the high-frequency signal in the form of the high-frequency signal adjustment value are multiplied to obtain a DC power supply type TF1004505 201234762, wherein the high-frequency signal adjustment value is sin~, 乂a high-frequency voltage angular frequency; to obtain an input signal related to the low-pass filtering processing error angle of the DC electrical signal with the estimated number; and the input signal is subjected to a neural-like proportional integral differential operation to obtain a fast signal, The equation for the differential integral operation of this kind of neuron is (N) = KP (N)e + Kd (N) ^ + (N) { edt + \) = Kp{N)e + i-ix5Oesgn{e) Kd(N + \) = Kd(N)e + jhS0~ ruler, (7V + 1) = ruler, _(Ar)e + 77i(5 s sgn(e) = e if e&gt;Q sgn(e) = ~e if e &lt; 〇 Among them, eight is the round of the proportional-integral differential operation of the neuron, and the heart is the proportional parameter 'the heart is the differential parameter, 3 is Integral parameter, the error e is the estimated error angle 'degree, &amp; is the weighting parameter, and A is set to e, and % is the learning rate of the weight; the speed (four) is entered into the Weitong filter and the electric motor of the motor a parameter related to the parameter to obtain the estimated rotational speed; and 进行 performing an integral operation on the rotational speed signal to obtain the estimated angular production. The method of claim 2, wherein the following is performed by the wheel controller Step: By the microcontroller 々-----! You have to change the module to refer to the estimated angle to convert the electrical signal of the two-phase stationary coordinate into the synchronous rotary coordinate gas signal; q glaze TT1004505 31 201234762 The signal processing and motor speed estimation module performs the steps of the second item of the patent application scope, and the dq axis electrical signal is calculated to obtain the estimated rotation speed and The speed control module compares the commanded speed with the estimated speed to obtain a speed difference, and performs a proportional integral operation on the speed difference to obtain a current command signal; Comparing the dq axis electrical signal with the current command signal by a current control module of the microcontroller to obtain a current difference, and performing the proportional integral operation on the current difference to obtain a voltage signal; The voltage decoupling module decouples the voltage signal to obtain a decoupling voltage signal, wherein the decoupling operation cancels the interference of the estimated speed on the dq axis electrical signal and the engagement between the two; And summing the high frequency signal generated by the high frequency signal injection module of the microcontroller and the decoupling voltage signal to obtain a voltage command signal, wherein the south frequency signal injection module is generated as a surface frequency voltage The local frequency signal has the following expression: 乂—Zdh 0 — ^dh -V _0 V jqh _ where ν(/Λ is the d-axis and q-axis high-frequency power of the synchronous reference coordinate respectively The components, &, & are the high-frequency current components of the d-axis and q-axis of the synchronous reference coordinate, respectively, Ζ(/Λ,Ζ&lt;?Λ are the high-frequency of the d-axis and q-axis of the synchronous reference coordinate of the motor, respectively. Impedance; TF1004505 32 201234762 The estimated angle of the anti-coordinate conversion module of the microcontroller will be converted into a two-phase static control coordinate two-phase control signal for the voltage of the synchronous rotary money; The sine wave pulse width modulation module of the microcontroller performs the sine wave pulse width view on the three-phase control signal with reference to the estimated angle to obtain the control signal. 4 · The method of applying for the scope of the patent scope to the third item - the package includes the following steps: providing a DC power supply from a power supply to an inverter, and is isolated by the isolation amplifier by the microcontroller Transmitting the control money in the form of an electronic signal, and amplifying the control signal into an electrical control signal in the form of an electrical signal; and converting the money power source into a three-phase power source by the commutation H, and controlling according to the electrical control signal The size of the three-phase power supply is input to the motor. 5. If you apply for a patent scope! The method of any one of the preceding claims, wherein the microcontroller is not generating the high frequency signal by urging the microcontroller to determine that the motor is driven to a predetermined low speed based on the electrical signal. 6. A motor control method without a sensor, the method operating when the motor is at a medium or high speed, the method comprising the steps of: sending a control signal from a microcontroller to drive the motor; The controller receives an electrical signal of the motor when it is driven; and calculates the electrical speed and the electrical parameter of the motor to calculate the current cost, and according to The estimate TF1004505 33 201234762 = signal, the electrical signal, the electrical parameters of the motor and the parameters required for the -integration of the neural integral, the "obtained-estimated rotational speed and an estimated angle" In the integral differential operation, one of the proportional inflammation number, the -differential parameter and the integral value of the integral parameter are difficult to scale with each change of the estimated (four) difference, and then the proportional parameter, the delta differential parameter and the An integral parameter; and the motor is driven by the micro-control (10) to command the rotational speed and the estimated rotational speed to obtain the control finger. 7. The method of claim 6, wherein the signal processing and the motor speed estimation module of the microprocessor perform the following steps: estimating an electrical speed and an electrical parameter of the motor based on the command electrical signal In order to establish a tunable model, the 姑 ' 、, ώ μ & moduli branch is calculated according to the tunable model, wherein the tunable model is: d ώ ~ A - h - against Ld .' •-CO Lq n Plant A - id &quot;1 .1 TVd Λ. ——ώ A/一JL· 4” A u” + q 1 . ld, h where t is the estimated current signal, &lt;To estimate the electrical speed, for the resistance parameter of the motor, A/, 々 民, to 4 A ^ · pressure signal; one is the inductance parameter of the motor, v&quot;, \ is the command electric according to the estimated current The signal, the electrical signal and the electrical parameter of the motor are adapted to the H mechanism, and are calculated according to the adaptive mechanism to obtain the estimated electrical speed, wherein the adaptive mechanism comprises a neuro-proportional integral micro TF1004505 34 201234762 The adaptive mechanism is v·kji A/ · ? Y〇^idiq-~iqid~ Ld \ L where Λ is the analogy of the proportional-integral differential operation of the class of 类^ The integral equation of the integral differential operation is ···, the neurological ratio (TV) = Kp (N)e + Kd {N) ^ + Kj (Νψώ KP (N+\) = KP (N)e + ;7,^〇esgn(e) dt Kd(^ + 1) = Kd(N)e + η{δ〇(jV +1) = Kt (N)e + 7;, s0 j edt sgn(e) = e if e&gt;〇sgn(e) = -e if e&lt;〇 error β is set to the learning speed at which the estimated % is weight Where "" is the electrical signal, &amp; is the anti-electrical parameter, (for the differential parameter, the integral is pure, 帛P is the proportional error angle '^ weighting parameter, and ^ is set to e, rate; the estimated electrical speed is integrated The operation is performed to obtain the estimated electrical speed multiplied by the motor, and to obtain the estimated speed. The parameters related to the number of turns are as follows: 8. As claimed in the patent scope, the following steps are performed: The microcontroller is replaced by one of the microcontroller's positive seat rods ", r, a, τ, and the electrical signal of one of the two dq axes is referenced to the estimated angle; the electrical signal of the phase stationary coordinate is converted Synchronization Rotating seat TF1004505 35 201234762 The signal processing and motor speed estimating module performs the steps of item 7 of the patent application scope, and the dq axis electrical signal is calculated to obtain the estimated rotating speed and the estimated angle; One speed control module of the controller compares the commanded speed with the estimated speed to obtain a speed difference, and performs a proportional integral operation on the speed difference to obtain a current command signal; a current from the microcontroller The control module compares the dq axis electrical signal with the current command signal to obtain a current difference, and performs the proportional integral operation on the current difference to obtain a voltage signal; the voltage decoupling module of the microcontroller will The voltage signal is decoupled to obtain a voltage command signal, wherein the decoupling operation cancels the interference of the estimated rotational speed on the dq axis electrical signal and the face-to-face relationship therebetween; The conversion module refers to the estimated angle, and the voltage command signal of the synchronous rotating coordinate is converted into a three-phase control signal by one coordinate of the three-phase stationary coordinate. And the sine wave pulse width modulation module of the one of the microcontrollers performs the sine wave pulse width modulation on the three-phase control signal with reference to the estimated angle to obtain the control signal. 9. The method of any one of claims 6 to 8, further comprising the steps of: providing a DC power supply from a power supply to an inverter; isolating the transmission by the microcontroller by an isolation amplifier The control signal in the form of an electronic signal and amplifying the control signal into an electrical control signal in the form of an electrical signal TF1004505 36 201234762; and converting the DC power source into a two-phase power supply* by the converter and controlling according to the electrical control signal The size of the three-phase power supply is input to the motor. TF1004505 37