JP2006304478A - Motor drive controller and electric power steering apparatus therewith - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a motor drive controller, including a rotor position estimation device, capable of accurately calculating a rotor position and an angle velocity even in a low velocity region and even if temperature change occurs. <P>SOLUTION: In the motor drive control device which includes a motor voltage/current detector, a position detector for detecting a rotor rough position, and the rotor position estimation portion for estimating the rotor position and the angle velocity on the basis of the voltage/current and the rotor rough position, and which controls a vector on the basis of the estimated rotor position and the angle velocity, and command value, the rotor position estimation portion is constituted by an angular velocity calculating potion for calculating the angle velocity on the basis of each phase counter electromotive voltage, an electrical angle calculating portion for calculating the electrical angle of the rotor on the basis of the angular velocity, a rotor phase detecting portion for detecting a rotor phase on the basis of the rotor rough position, a correction resistor calculating portion for calculating the correction resistor of each phase on the basis of the angle error of the electrical angle and the rotor phase, and the motor current, and a multiplication portion for multiplying the output of a motor model and the output of the correction resistor calculating portion. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a motor drive control device that can accurately estimate a rotor position and an angular velocity of a motor by correcting a motor model, and an electric power steering device using the motor drive control device.

  Conventionally, as a drive control method for a motor used in an electric power steering device, for example, a drive control method for a brushless DC motor, a rotating magnetic field is generated from the motor drive control device via an inverter based on the rotational position of the rotor, and the rotor A control method is adopted that controls the rotation of the motor. This drive control system controls the rotational drive of the rotor by sequentially switching the excitation of each excitation coil by a control circuit according to the rotor position to a plurality of excitation coils arranged at predetermined angular intervals inside the stator. It is.

  Vector control often used as a drive control method for a brushless DC motor is disclosed in, for example, Japanese Patent Application Laid-Open No. 2001-18822 (Patent Document 1). FIG. 10 is a configuration example showing a motor drive control device used in the electric power steering device.

  In FIG. 10, a command signal from the command current determining unit 51 that determines the motor control command value to the motor 56 via the PI control unit 52, the two-phase / three-phase coordinate conversion unit 53, the PWM control unit 54, and the inverter 55. The main path is formed. A current detector 57 is disposed between the inverter 55 and the motor 56, and the motor current value detected by the current detector 57 is disposed between the command current determination unit 51 and the PI control unit 52. A feedback path for feeding back to the subtracting circuit 58 is formed. In this feedback path, a three-phase / two-phase coordinate converter 59 is arranged.

  The command current determination unit 51 inputs the command value Tref calculated from the torque detected by the torque sensor, the electrical angle θ and the electrical angular velocity ω indicating the position of the rotor detected by the position detection sensor 11, and the command current Idref and Iqref is determined. The command currents Idref and Iqref are corrected by feedback currents that are detected by the current detector 57 and then converted into two phases by the three-phase / two-phase coordinate converter 59. That is, the deviations between the feedback currents Id and Iq and the current command values Idref and Iqref are calculated by the subtraction circuits 58d and 58q, respectively. Thereafter, signals indicating the duty of PWM control are calculated as Vd and Vq in the form of d and q components by the PI control units 52d and 52q, and the two-phase / three-phase coordinate conversion unit 53 performs two-phase d and q components. To 3 phase components Va, Vb, and Vc. The inverter 55 is PWM controlled via the PWM voltage generator 54 based on the command values Va, Vb, and Vc, and an inverter current is supplied to the motor 56 to control the rotation of the motor 56.

  In addition, 61 is a vehicle speed sensor, 62 is a sensitive area determination unit, 63 is a coefficient generation unit, 64 is a basic assist force calculation unit, 65 is a return force calculation unit, and 66 is an electrical angle conversion. , 67 is an angular velocity conversion unit, and 68 is a non-interference control correction value calculation unit.

  In such vector control, the command current determination unit 51 determines the current command values Idref and Iqref based on the torque command value Tref, the rotor angular velocity ω, and the rotor position θ. Further, the feedback currents Iu, Iv, and Iw of the motor 56 are converted into currents Id and Iq by the three-phase / two-phase coordinate conversion unit 59, and thereafter, the deviation between the currents Id and Iq and the current command values Idref and Iqref is subtracted. Circuits 58d and 58q are operated, and the deviations ΔId and ΔIq are subjected to current control by PI control in the PI control units 52d and 52q, whereby voltage command values Vd and Vq to the inverter are obtained. Then, the two-phase voltage command values Vd and Vq are again converted back to the three-phase voltage command values Va, Vb and Vc, the inverter 55 is controlled, and the drive control of the motor 56 is performed.

  When such vector control is used, if the motor control is performed in a state where the rotor position θ cannot be detected correctly, the torque ripple of the motor 56 will increase, and the electric power steering device may give the driver steering feel uncomfortable, such as vibration. Undesirable phenomena such as increased motor noise occur. Therefore, in order to correctly detect the rotor position θ, as disclosed in Japanese Patent Application Laid-Open No. 2001-187578 (Patent Document 2), the rotor position detector 11 is an expensive but highly accurate resolver. It is necessary to use an encoder.

  Since resolvers and encoders are expensive, there is an example in which motor control is attempted using a Hall sensor that is an inexpensive rotor position detector. For example, as disclosed in Japanese Patent Application Laid-Open No. 2002-272163 (Patent Document 3), a Hall sensor signal is used to correct the starting point or midpoint of a phase for generating a sine wave used for PWM control of a motor. ing. However, in the example of Patent Document 2, the rotor rotation angle θ in the middle section until the next Hall sensor signal is obtained cannot be calculated, and the function as a rotor position detector such as an encoder is reached. Not in.

  In addition, even with a detector with good detection accuracy such as an encoder, when the rotational speed of the rotor becomes low, the number of detection points obtained from the encoder decreases, and there is a problem that the position detection accuracy of the rotor is lowered. Therefore, as disclosed in Non-Patent Document 1, improvement in accuracy is improved by using the moment of inertia of the motor load and the motor current. However, there is integration until the angular velocity ω is calculated from the angular acceleration obtained from the moment of inertia and the motor current, and there is also integration until the electrical angle θ of the rotor is calculated from the angular velocity ω, and the integral calculation is used twice. Therefore, the calculation accuracy is poor, and it is difficult to determine the moment of inertia of the load driven by the motor correctly.

Furthermore, a change in the temperature of the motor is one factor that increases the torque ripple of the motor. In other words, when calculating the electrical angle θ of the rotor, if a rotor position estimation circuit that uses the counter electromotive voltage of the motor is used, the resistance and inductance of the motor used to calculate the counter electromotive voltage are calculated. If the value changes due to a temperature change and the resistance or the like due to the temperature change is not corrected, the electrical angle θ of the rotor cannot be calculated accurately, resulting in a problem that the torque ripple becomes large. Japanese Patent No. 3104865 (Patent Document 4) is an example of calculating a resistance value of a motor in consideration of a temperature change. However, it is inconvenient because a special condition is imposed such that the rotational speed of the motor is zero.
JP 2001-18822 A JP 2001-187578 A JP 2002-272163 A Japanese Patent No. 3104865 Seung-Ho-Song, speed observer for low-speed control of AC motors (An Instantaneous Speed for Low speed Control of ac Machine), IEEE application power electronics 1998 (AP 1998 EPE AP -581-586

  As described above, in order to control the motor using vector control, it is necessary to correctly detect the rotor position of the motor. However, since the resolver and encoder are expensive, the electric power steering device is manufactured at low cost. It becomes an obstacle. Further, there is a problem that even if a high-precision position detector such as a resolver or an encoder is used, the rotor position cannot be detected correctly in the low-speed region of the motor. Furthermore, there has been a problem that the detection accuracy of the rotor position is deteriorated due to the temperature change of the motor.

  When using a Hall sensor as an inexpensive rotor position detector, accurate angle data can be obtained only when a Hall sensor signal is input. Therefore, angle data (angle estimated value) is generated from the estimated value of the back electromotive force of the motor. The Hall sensor signals are interpolated. And the back electromotive force is estimated from the current, voltage, and resistance model of each phase, but the actual resistance value of the motor fluctuates due to variations in manufacturing, aging, temperature change, etc., so an error occurs in the resistance model, An error also occurs in the estimated value of the back electromotive force. As a result, an error also occurs in the estimated angle value of the motor shaft, and problems such as current fluctuations and control sounds occur in the control of the electric power steering apparatus.

  The present invention has been made under the circumstances as described above, and the object of the present invention is to correct the error of the resistance model so that the rotor position and angular velocity can be accurately determined despite the use of an inexpensive rotor position detector. In addition to providing a motor drive control device including a rotor position estimating means that can accurately calculate the rotor position with software even in the low speed region of the motor or even when there is a temperature change or the like, the rotor calculated correctly By correctly executing motor vector control using a motor drive control device that uses position and angular velocity, even if it is a middle speed switching steering of a steering wheel such as slalom running, it is inexpensive and does not feel strange to steering the steering wheel. The object is to provide a high-performance electric power steering apparatus.

  The present invention relates to a voltage detection means for detecting a phase voltage or a line voltage of a brushless DC motor having three or more phases, a current detection means for detecting a motor current of the motor, and a position detection for detecting a rough rotor position of the motor. And a rotor position estimating unit that estimates a rotor position and an angular velocity of the motor based on the phase voltage or line voltage, the motor current, and a rotor rough position, and a rotor position estimated by the rotor position estimating unit, and The present invention relates to a motor drive control device that performs vector control of the motor based on an angular velocity and a command value, and the object of the present invention is to calculate the angular velocity based on each phase counter electromotive voltage. An angular velocity calculator; an electrical angle calculator that calculates an electrical angle of the rotor based on the angular velocity; and a rotor phase of the motor based on the rough rotor position. A rotor phase detector that outputs, a correction resistance calculator that calculates a correction resistance for each phase based on the motor angle and the electrical angle and the angle error of the rotor phase, and the motor of the motor that receives the motor current A model and a multiplication unit that multiplies the output of the motor model and the output of the correction resistance calculation unit, and the difference between the phase voltage or line voltage and the output of the multiplication unit is the counter-electromotive voltage of each phase Is achieved.

Further, the object of the present invention is that the position detecting means is a Hall sensor, or that an addition value obtained by adding the correction resistor to the motor resistance is input to the multiplication unit, or the angular velocity is within a predetermined range. By executing the correction when at least one of the motor currents is within a predetermined range, or by setting the motor resistance as R, the correction resistance as ΔR, the inductance as L, and the Laplace operator as s, By setting the model to (R + ΔR) (L / R · s + 1), or setting the time constant to T ₁ and setting the motor model to (L · s + R) / (T ₁ · s + 1), or the motor model Is more effectively achieved by setting (L / R · s + 1) / (T ₁ · s + 1).

  According to the motor drive control device of the present invention, even if an inexpensive rotor position detection sensor is used, by combining the electrical angle of the rotor calculated from the motor voltage, current, etc. with an inexpensive rotor position detection sensor, It is possible to provide a motor drive control device that can detect a highly accurate electrical angle or angular velocity of a rotor including a low rotational speed region and compensates for a resistance value and a temperature of a motor winding.

  According to the motor drive control device of the present invention, the motor model can be corrected without increasing the processing load of software by multiplying the motor resistance value under the condition that the motor angular velocity is within the predetermined range or the motor current is within the predetermined range. This is simple, and the rotor position of the motor can be accurately estimated even by using a cheap and relatively coarse rotor position detecting means (for example, a hall sensor).

  Further, when the motor drive control device is used in an electric power steering device such as a vehicle, a highly accurate electrical angle of the rotor can be detected even by using an inexpensive rotor position detection sensor such as a hall sensor. A low-cost and high-performance electric power steering device that can smoothly follow the middle-speed steering of the steering wheel with less motor control can be provided.

  First, a technique that is a premise of the present invention will be described.

An example in which the phase voltages Va, Vb, and Vc are detected as the motor voltages will be described, but this is true even if the voltages are line voltages Vab, Vbc, and Vca. The line voltage is, for example, as shown in Equation 13 of Japanese Patent Application Laid-Open No. 2004-312834 and FIG. Motor currents ia, ib, and ic are detected in addition to the voltage. Motor winding resistances Ra, Rb, and Rc and inductances La, Lb, and Lc can be obtained in advance from the characteristics of the motor. Each phase back electromotive voltage ea, eb, ec of the motor is as follows.
(Equation 1)
ea = Va− (Ra + s · La) · ia
eb = Vb− (Rb + s · Lb) · ib
ec = Vc− (Rc + s · Lc) · ic
Note that s is a Laplace operator.

In general, the relationship between the back electromotive force e of the motor and the angular velocity ω is as follows.
(Equation 2)
e = Ke · ω
Ke is a back electromotive force constant [V / rpm] of the motor.
The above formulas 1 and 2 are for a brush motor. In a brushless motor, it is necessary to rectify the counter electromotive voltages ea, eb, and ec of each phase. In trapezoidal wave current control and rectangular wave current control, rectification is the same as taking the maximum value, and this is expressed as follows.
(Equation 3)
ω = 2 × {max (| ea |, | eb |, | ec |)} / Ke
Here, a waveform example of the back electromotive voltages ea, eb, ec is shown in FIG. Rectifying means taking the envelope of the counter electromotive voltages ea, eb, ec and taking the maximum value. The reason why the numerator of Equation 3 is doubled is that the negative value is superimposed on the positive side by taking the absolute values of the back electromotive voltages ea, eb, and ec.

Next, the electrical angle θ can be obtained from the following equation (4). Note that θ _i in Equation 4 is an initial value of the integration interval.
(Equation 4)
θ = θ _i + ∫ωdt
Then, when digital expression 4 is performed, the following expression 5 is obtained.
(Equation 5)
θ = θ _i + n · ω · Ts
Note that n is obtained by dividing the time ΔTs waiting for the next Hall sensor signal by the sampling time Ts, and ΔTs = T ₁₂₀ −T ₆₀ in FIGS. 3, 4, and 5.

FIG. 3 shows an example of the electrical angle θ calculated by the rotor phase estimation unit, and the error accumulates as time elapses due to the calculation error. The amount of the error, the true value in the example the time T ₆₀ whereas the 60 degrees, the calculated value θe is 65 degrees. An error of 5 degrees has occurred between time T ₀ and time T ₆₀ . However, for example, if three Hall sensors are attached to a 4-pole motor, the electrical angle θ ₀ can be detected every 60 degrees, so that the calculated electrical angle θ can be corrected. FIG. 4 shows the state of the correction. Calculating the electrical angle θe at time T ₆₀ in FIG. 3 becomes 65 degrees, but is modified to detected electrical angle theta ₀ by 60 degrees. That is, θ ₀ = 60 degrees is substituted for θ _i and is calculated as the following formula 6.

よって、時点Ｔ_６０から時点Ｔ_１２０の区間では、電気角θの初期値を６５度ではなく６０度に補正して算出するので、誤差が累積することはない。

Thus, in a section of the time T ₁₂₀ from time T _60, since the initial value of the electrical angle θ calculated by correcting 60 degrees rather than 65 degrees, not an error is accumulated.

  On the other hand, the electrical angle θ can be corrected every 60 degrees using the detection value of the Hall sensor, but during that period, the electrical angle θ accumulates errors. In order to improve the error of the electrical angle θ between the error accumulations, the following is performed. The error in the calculated electrical angle θ is caused mainly by a rise in the winding temperature and a change in the winding resistance Rm (m = a, b, c). An angular velocity error Δω is determined from the electrical angle error Δθ, a counter electromotive voltage Δe is determined from the angular velocity error Δω, and a winding resistance correction resistor ΔRm (m = a, b, c) is determined from the counter electromotive voltage Δe. . Then, the electrical angle θ is calculated with the resistance Rm obtained by replacing the winding resistance Rm of Equation 1 with (Rm + ΔRm). This content can be expressed as follows.

First, the electrical angle error angle Δθ is (Equation 7).
Δθ = θ−θ ₀
In digital processing (Equation 8)
Δθ = n · Ts · Δω
The angular velocity error Δω is obtained from this equation 8, and the counter electromotive voltage Δe is obtained according to the following equation 9.
(Equation 9)
Δe = Δω · Ke / 2
Next, the correction resistance ΔRm is obtained according to the following formula 10.
(Equation 10)
ΔRm (m = a, b, c) = Δe / i m
The correct motor winding resistance Rm in consideration of the calculated correction resistance ΔRm with respect to the motor winding resistance Rm before correcting the correction resistance ΔRm is expressed by the following equation (11).
(Equation 11)
Rm (new) = Rm + ΔRm
Substituting the winding resistances Ra, Rb, and Rc obtained in Equation 11 into Equation 1 and correcting the electrical angle θ with the electrical angle θ ₀ detected every 60 degrees according to Equation 6, the result is as shown in FIG. The error is reduced. The reason why the error still remains is that the cause of the temperature change is not only the winding resistance but also the error of the detected voltage and current.

Since the temperature coefficient α (Ω / ° C.) of the motor winding resistance is known from the material, shape, etc., the temperature change ΔT of the winding resistance can be obtained by the following formula 12.
(Equation 12)
ΔT = ΔR / α
An example in which the motor drive control device based on the above theory is applied to an electric power steering device will be described with reference to FIGS.

  In FIG. 6, the motor 1 is a four-pole three-phase brushless DC motor. The motor 1 has a rotor (not shown), and a Hall sensor 48 (48-1, 48-) serving as a position detection sensor for detecting the electrical angle of the rotor. 2, 48-3) are arranged every 120 degrees. As a result, the electrical angle θ of the rotor of the motor 1 can be detected as a rough rotor position at intervals of 60 degrees. In order to vector-control the motor 1 so that the torque ripple is small, it is necessary to correctly calculate the electrical angle θ of the rotor correctly. In order to calculate the current command values Iavref, Ibvref, Icvref from the torque command value Tref, Calculation is performed using the electrical angle θ and the angular velocity ω of the rotor. The current command values Iavref, Ibvref, and Icvref calculated by the vector control unit 100 are used as reference values, and the currents ia, ib, and ic detected by the current detection circuits 32-1, 32-2, and 32-3 are fed back and subtracted. The error current is obtained by the units 20-1, 20-2, 20-3, and the voltage command values Vpa, Vpb, Vpc are obtained by the proportional integration (PI) unit 21 using the error current as an input, and the PWM control unit 30 is an inverter. 31 is PWM controlled based on the voltage command values Vpa, Vpb, and Vpc. The electrical angle θ and the angular velocity ω with high accuracy are calculated by the rotor position estimation unit 200.

  Details of the rotor position estimation unit 200 will be described with reference to FIG.

The rotor position estimation unit 200 includes motor phase voltages Va, Vb, Vc, motor currents ia, ib, ic, Hall sensors 48-1, 48 detected by the voltage detection circuits 33-1, 33-2, 33-3. The electrical angle θ ₀ = _0, 60, 120, 180, 240, 300 degrees as the rotor coarse position from −2, 48-3 is input. The attachment position of the electrical angle θ ₀ does not have to be 0 degrees. For example, when the hall sensor 48-1 is installed at a position of 30 degrees, θ ₀ = 30, 90, 150, 210, 330 degrees. Motor currents ia, ib, ic are input to transfer function units 201-1, 201-2, 201-3. Here, the transfer functions of the transfer function units 201-1, 201-2, and 201-3 are expressed by the following Equation 13, which corresponds to Equation 1.
(Equation 13)
Z = (Rm + s · Lm) / (s · T _f +1)
Note that m = a, b, c, and s is a Laplace operator.
The numerator of Equation 13 is the impedance (Rm + s · Lm) multiplied by the motor current of Equation 1. The impedance is multiplied by 1 / (s · T _f +1), which is a transfer function of a low-pass filter that does not exist in Equation 1. Yes. The reason why the low-pass filter is used is to remove noise because the currents ia, ib, and ic include noise, and has a practical meaning rather than a theoretical meaning.

  The motor voltages Va, Vb, Vc and the output of the transfer function units 201-1, 201-2, 201-3 are input to the subtracting units 202-1, 202-2, 202-3, and when the difference between them is taken, Phase back electromotive voltages ea, eb, ec are calculated. That is, Equation 1 is executed to calculate the back electromotive voltages ea, eb, ec of each phase.

Next, the counter electromotive voltages ea, eb, and ec of the respective phases are input to the angular velocity calculation unit 203, and the mathematical expression 3 is executed to calculate the angular velocity ω. As a method of calculating the maximum value of the counter electromotive voltage required by the angular velocity calculating unit 203, a method of calculating the maximum value of the counter electromotive voltages ea, eb, ec by taking the absolute value and multiplying it as shown in Equation 3 However, as can be seen from FIG. 2, the phase of the back electromotive force voltage is determined by the electrical angle θ, so that every 60 degrees detected by the Hall sensor 48 (48-1 to 48-3). Is derived from the following equation using the electrical angle θ ₀ of:
(Equation 14)
ω = (ea × Ca + eb × Cb + ec × Cc) / Ke
Ca, Cb, and Cc are parameters representing commutation. The values of “1”, “0”, and “−1” are assumed for trapezoidal wave current and rectangular wave current, and the electric angle θ ₀ The section in which the parameters Ca, Cb, and Cc are “1”, the section that is “0”, and the section that is “−1” are determined. The section is determined by the detection signal Shall of the Hall sensor 48. can do.

When calculating the angular velocity ω using Equation 14, the counter electromotive voltages ea, eb, ec and the electrical angle θ ₀ from the rotor phase detecting unit 205 are input to the angular velocity calculating unit 203, and the parameters Ca, Cb are based on the input. , Cc are determined, and angular velocity calculation unit 203 executes equation 14 to calculate angular velocity ω.

  Next, the electrical angle calculation unit 204 for obtaining the electrical angle θ from the angular velocity ω is the integration expressed by Equation 6, and the electrical angle θ is calculated by inputting the angular velocity ω. If there is an error in the electrical angle θ calculated by the electrical angle calculation unit 204, the error is accumulated by integration, resulting in the result shown in FIG. 3, and the accurate electrical angle θ is not calculated.

Therefore, as shown in FIG. 7, a rotor phase detection unit 205 is provided in the rotor position estimation unit 200. The rotor phase detection unit 205 receives the Hall sensor signal Shall from the Hall sensor 48 and detects the detected electrical angle. θ ₀ = _0, 60, 120, 180, 240, and 300 degrees are output. The detected electrical angle θ ₀ is input to the electrical angle calculation unit 204, and the initial value θ _i is reset to the electrical angle θ ₀ in Equations 4 and 6. As a result, at time T _60, as shown in FIG. 4, but 65 degrees and the error 5 degrees calculated value theta to the true value 60 ° of electrical angle is generated, the detection value theta _{0 =} 60 in the rotor phase detector 205 Since the calculated value θ is reset to 60 degrees in degrees, since the initial value θ _i is set to 60 degrees and the error is reset during the next integration interval, that is, from the time point T ₆₀ to the time point T ₁₂₀ , the error is There is no accumulation. The same applies to the other sections.

FIG. 8 shows a configuration example of the rotor position estimating unit 200 when correcting the motor winding resistance. The subtracting unit 206 subtracts the error between the detected electrical angle θ ₀ and the calculated electrical angle θ at that time. Ask for in. For example, the error angle Δθ between the electrical angle θ and the electrical angle θ ₀ = 60 degrees at the time point T ₆₀ is calculated, which means that Expression 7 is executed. Next, the error angular velocity detection unit 207 executes Expression 8. That is, the error angular velocity detection unit 207 executes Δω = Δθ / (n · Ts), and calculates the angular velocity error Δω from the electrical angle error angle Δθ. Next, the error angular velocity Δω is input by the error counter electromotive voltage calculation unit 208 for executing the equation 9, and the error counter electromotive voltage Δe is calculated according to the equation 9. Then, the error resistance calculation circuit 209 calculates the winding resistance correction resistance ΔRm (m = a, b, c) from ΔRm = Δe / im according to the equation (10).

Next, in the resistance correction unit 210, the winding resistance Rm is replaced with (Rm + ΔRm) in consideration of the correction resistance ΔRm calculated by the error resistance calculation unit 209. Finally, the new resistance values Ra = Ra + ΔRa, Rb = Rb + ΔRb, Rc = Rc + ΔRc calculated by the resistance correction unit 210 are substituted into the resistances Ra, Rb, Rc of the transfer function units 201-1, 201-2, 201-3. Then, the angular velocity ω or the electrical angle θ is newly calculated using a correct resistance value considering the temperature change. As a result, theta = 65 degrees at point A at time T _60, as shown in FIG. 4 as the error was 5 degrees, for example, theta = 61 degrees at the point A 'at time T _60, as shown in FIG. 5 Thus, the error can be improved once. That is, the accuracy of the calculated electrical angle θ in the section between the discrete electrical angles θ ₀ detected by the rotor position detection sensor can be greatly improved.

Then, the correction resistance ΔRm calculated by the error resistance calculation unit 209 is input to the change temperature calculation unit 211. The change temperature calculation unit 211 calculates the temperature change ΔT by executing the temperature change ΔT = ΔRm / α shown in Equation 13. Here, since there are three types of correction resistors ΔRm, ΔRa, ΔRb, and ΔRc, three types of calculated temperature change ΔT may also occur. Whether the maximum value is used or the minimum value is used. Alternatively, whether to use the average value is selected in consideration of the usage target or the entire apparatus. If the temperature change ΔT is known, the temperature Tc = Ta + ΔT of the winding resistance can be calculated by adding ΔT to the initial temperature Ta.

As described above, the motor back electromotive force is estimated from the current, voltage, and resistance model of each phase, but the actual resistance value of the motor fluctuates due to manufacturing variations, aging, temperature changes, etc. An error occurs, and an error also occurs in the estimated value of the back electromotive force. As a result, an error also occurs in the estimated angle value of the motor shaft, and problems such as current fluctuation and control sound occur in advanced control such as electric power steering. Therefore, the present invention suppresses current fluctuation and control sound. For this reason, the resistance model error is corrected.

  That is, since the back electromotive voltage is small during low speed rotation, the error is large, and the correction resistance ΔRm cannot be calculated correctly. In addition, since the change timing of the Hall sensor signal cannot be accurately detected during high-speed rotation, the correction resistance ΔRm cannot be calculated correctly. For example, when the motor angular speed is 1200 rpm, the change interval of the Hall sensor signal is 4.167 ms. When the motor angular speed is 1800 rpm, the change interval of the Hall sensor signal is normally 2.778 ms, and the sampling cycle is usually about 1 ms, which is a considerable error. . Further, at the time of high speed rotation, the influence of the delay of the low-pass filter for removing noise also becomes an error, and the correct correction resistance ΔRm cannot be calculated. Since R · I is also calculated when calculating the back electromotive voltage, the correction resistance ΔRm cannot be calculated correctly even when the motor current is small. Therefore, in the present invention, the correction resistor ΔRm is calculated and the motor model is corrected under the condition that the motor angular velocity is within a predetermined range or the motor current is within a predetermined range.

  Here, the example of FIG. 8 is a fundamental motor model correction, and there is no problem in principle. However, when implemented as software, there are three constants to be changed, and the processing load is large, such as requiring an intermediate variable reset operation.

  Since there is an error in the resistance model, the angle estimation value also causes an error. However, when the Hall sensor signal changes, the angle error between the actual angle and the angle estimation value is known. The equivalent is calculated and fed back to adjust the resistance model value to reduce the back electromotive voltage estimation error. By correcting the resistance error in this way, the angle estimation accuracy between the Hall sensor signals can be improved, current fluctuation can be suppressed, and the control sound can be reduced.

The motor model is Ls + R, and when the motor resistance value R changes from R → R + ΔR, the motor model is Ls + R + ΔR. However, it is inconvenient to implement this motor model with software for the above reasons. Therefore, when the motor resistance value is multiplied by L / R · s + 1 divided by the nominal resistance value R, the following equation 15 is obtained.
(Equation 15)
(R + ΔR) (L / R · s + 1)
In Equation 15, since L / R · s + 1 does not change, constant change and intermediate variable reset operation are not required. Although L · s + R + ΔR and Equation 15 are not exactly equal, there is no practical problem because they are almost equal. Further, since pure differentiation cannot be realized by software, it may be implemented as (L · s + R) / (T ₁ · s + 1)). As a model, (L / R · s + 1) / (T ₁ · s + 1) may be used.

  The motor counter electromotive voltage is expressed by the following equation (1), and the motor neutral point-a phase counter electromotive voltage Ean is expressed by the following equation (16), where the voltage between the a phase terminal and the motor neutral point is Van.

インダクタンスＬのモデル化誤差は静的な電流には影響されないので、インダクタンスＬの誤差を無視して考える。抵抗値各相によってノミナル値と誤差があり、この誤差量を推定し、抵抗値とすることで角度推定の誤差をなくすようにする。抵抗Ｒａについては下記数１７で表わされる。抵抗Ｒｂ，Ｒｃについても同様である。
（数１７）
Ｒａ＝Ｒｍ＋ΔＲａ
数１６に数１７を代入すると下記数１８になる。なお、Ｔ_２＝Ｌ／Ｒｍである。

Since the modeling error of the inductance L is not affected by the static current, the error of the inductance L is ignored. Each phase of resistance value has a nominal value and an error. This error amount is estimated, and the resistance value is used to eliminate the angle estimation error. The resistance Ra is expressed by the following Expression 17. The same applies to the resistors Rb and Rc.
(Equation 17)
Ra = Rm + ΔRa
Substituting Equation 17 into Equation 16 yields Equation 18 below. Note that T ₂ = L / Rm.

数１８の右辺第２項は伝達関数として実現できないため、フィルタを考慮して下記数１９のように実現する。

Since the second term on the right side of Equation 18 cannot be realized as a transfer function, it is realized as shown in Equation 19 in consideration of the filter.

また、電気角θは数４で表わされ、数４に数３を代入すると共に数１６を代入すると、下記数２０が成立する。他相のｂ相及びｃ相についても同様である。

Further, the electrical angle θ is expressed by the following equation (4). When the equation (3) is substituted into the equation (4) and the equation (16) is substituted, the following equation (20) is established. The same applies to the b phase and c phase of the other phases.

ここで、誤差角度Δθを電気角θに加算し、補正抵抗ΔＲａを考慮すると下記数２１が成り立つ。そして、角度の誤差が全てモータモデル内の抵抗によって発生していると考えると、下記数２１が成り立つ。

Here, when the error angle Δθ is added to the electrical angle θ and the correction resistance ΔRa is taken into consideration, the following equation 21 is established. When it is considered that all the angle errors are caused by the resistance in the motor model, the following equation 21 holds.

上記数２１より、誤差角度Δθは下記数２２となる。

From the above equation 21, the error angle Δθ is expressed by the following equation 22.

上記数２２より、下記数２３が求まる。他相のｂ相及びｃ相についても同様である。

From the above equation 22, the following equation 23 is obtained. The same applies to the b phase and c phase of the other phases.

なお、低速時には逆起電圧が小で誤差が大きくなり、抵抗変化分ΔＲｍを正確に算出することができず、また、高速時はホールセンサ信号の変化タイミングがずれ、ローパスフィルタによる誤差が含まれ、抵抗変化分ΔＲｍを正確に算出することができない。そのため、本発明では誤差の小さい範囲で抵抗変化分ΔＲｍの算出を行う。

Note that the back electromotive voltage is small and the error is large at low speed, and the resistance change ΔRm cannot be accurately calculated. At high speed, the change timing of the Hall sensor signal is shifted, and an error due to a low-pass filter is included. The resistance change ΔRm cannot be calculated accurately. Therefore, in the present invention, the resistance change ΔRm is calculated within a small error range.

  FIG. 9 shows an embodiment of the present invention corresponding to FIG. 8. The electrical angle error angle Δθ obtained by the subtraction unit 206 and the currents ia, ib, ic are input to the correction resistance calculation unit 220. Based on the error angle Δθ and the currents ia, ib, ic, the equation 23 is calculated for each phase, and the correction resistance ΔRm (m = a, b, c) is calculated. This calculation is performed when the motor angular velocity ω is within a predetermined range and when the motor currents ia, ib, ic are within the predetermined range, that is, when at least one of the conditions is satisfied. The correction resistance ΔRm is input to the change temperature calculation unit 211 and is input to the resistance correction unit 221. The resistance correction unit 221 calculates the corrected resistors Ra ', Rb', and Rc 'by executing the equation 16, and inputs the resistors Ra', Rb ', and Rc' to the multipliers 223a, 223b, and 223c, respectively.

The currents ia, ib, and ic are respectively input to the motor models 222a, 222b, and 222c described above, and the currents from which high-frequency noise has been removed are input to the multipliers 223a, 223b, and 223c, respectively, and corrected resistors Ra ′, Multiplication with Rb ′ and Rc ′ is performed. The motor model 222a, 222b, _{T 2} in 222c shows the L / R. The multiplication results of the multiplication units 223a, 223b, and 223c are input to the subtraction units 202-1, 202-2, and 202-3, respectively, and are subtracted from the motor phase voltages Va, Vb, and Vc, respectively, so that the equation 19 is executed. Is done. In FIG. 9, the motor phase voltages Va, Vb, and Vc are directly input to the subtraction units 202-1, 202-2, and 202-3, but a low-pass filter (1 / s · T _f +1) may be inserted. .

  By executing such a calculation on the software, the motor back electromotive voltages Ean, Ebn, Ecn of the respective phases shown in Expression 19 can be calculated, whereby the angular velocity calculation unit 203 correctly calculates the angular velocity ω. In addition, the electrical angle θ can be calculated correctly.

It is a principle figure explaining the counter electromotive voltage of a motor. It is a figure which shows the principle which calculates the electrical angular velocity of a rotor from a back electromotive force. It is a figure which shows the example of a calculation result of electrical angle (theta). It is a figure which shows the example of a calculation result of electrical angle (theta). It is a figure which shows the example of a calculation result of electrical angle (theta). It is a block diagram which shows the structural example of a motor drive control apparatus. It is a block diagram which shows the structural example of a rotor position estimation part. It is a block diagram which shows the apparatus structural example which calculates an electrical angle. It is a block diagram which shows the Example of this invention. It is a block diagram which shows the structural example of the vector control apparatus of the motor using the conventional electrical angle (theta).

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Motor 11 Rotor position detection sensor 20-1, 20-2, 20-3 Subtraction part 21 PI control part 22 2 phase / 3 phase conversion part 23 3 phase / 2 phase conversion part 24 PWM control part 31 Inverter 32-1, 32-2, 32-3 Current detection circuit 33-1, 33-2, 33-3 Voltage detection circuit 48-1, 48-2, 48-3 Hall sensor 100 Vector control unit 200 Rotor position estimation unit 201 Transfer function unit 202 Subtractor 203, 203a Angular velocity calculator 204 Electrical angle calculator 205 Rotor phase detector 206 Subtractor 207 Error angular velocity calculator 208 Error back electromotive force calculator 209 Error resistance calculator 210 Resistance correction unit 211 Change temperature calculator 212 Low pass Filter 220 Correction resistance calculation unit 221 Resistance correction unit

Claims (8)

Voltage detection means for detecting a phase voltage or line voltage of a brushless DC motor of three or more phases, current detection means for detecting a motor current of the motor, position detection means for detecting a rotor rough position of the motor, A rotor position estimating unit that estimates a rotor position and an angular velocity of the motor based on a phase voltage or a line voltage, the motor current, and a rough rotor position, and the rotor position, the angular velocity, and a command value estimated by the rotor position estimating unit In the motor drive control device that performs vector control of the motor based on the above, the rotor position estimation unit calculates an angular velocity based on each phase counter-electromotive voltage, an electrical speed of the rotor based on the angular velocity An electrical angle calculator that calculates an angle; a rotor phase detector that detects a rotor phase of the motor based on the rotor coarse position; and A correction resistance calculation unit that calculates a correction resistance of each phase based on an angle error of the rotor phase and the motor current, a motor model of the motor that receives the motor current, an output of the motor model, and the correction resistance A motor drive comprising: a multiplier for multiplying the output of the calculator; and a difference between the phase voltage or line voltage and the output of the multiplier is the back electromotive voltage for each phase. Control device. The motor drive control device according to claim 1, wherein the position detection means is a Hall sensor. The motor drive control device according to claim 1 or 2, wherein an addition value obtained by adding the correction resistor to the motor resistance is input to the multiplication unit. 4. The motor drive control device according to claim 1, wherein the correction is executed when at least one of the angular velocity is within a predetermined range and the motor current is within a predetermined range. 5. 5. The motor according to claim 1, wherein R is a motor resistance, ΔR is the correction resistance, L is an inductance, s is a Laplace operator, and the motor model is (R + ΔR) (L / R · s + 1). Drive control device. 5. The motor drive control device according to claim 1, wherein a time constant is T ₁ and the motor model is (L · s + R) / (T ₁ · s + 1). The motor drive control device according to claim 6, wherein the motor model is (L / R · s + 1) / (T ₁ · s + 1). An electric power steering apparatus using the motor drive control device according to any one of claims 1 to 7.

Images