JP6362188B2 - Sensor misalignment learning method for motor speed estimation - Google Patents

Description

  CROSS REFERENCE TO RELATED APPLICATIONS This application claims the benefit of US Provisional Patent Application No. 62 / 301,566, filed February 29, 2016, which is hereby incorporated by reference in its entirety.

  The present disclosure relates to motor speed estimation. Brushless direct current (BLDC) motors and electric motors are generally controlled, often using several cascaded closed loop controllers. One of those loops controls the motor speed and compares the speed setpoint with the actual speed of the electric motor. Since the actual speed of the electric motor is not directly measured, the actual speed of the electric motor can be estimated using information such as motor position. The subject matter disclosed herein describes a method for estimating electric motor speed from Hall effect sensor information and taking into account misalignment in Hall effect sensor portions and ring magnet irregularities.

  Conventional BLDC motors typically include a stator with windings thereon and electromagnetic poles, and a rotor with permanent magnets that produce pairs of permanent magnet poles. The stator and rotor interact magnetically when current flows through the stator windings. The phase commutation of the current flowing through each of the stator windings is performed at the appropriate time to form a continuously rotating magnetic field. It can be achieved if the rotor position is recognized correctly.

  BLDC motors most commonly utilize a three-phase structure in which Hall effect sensors are embedded in the motor and define the commutation position for each phase. A conventional three-phase BLDC motor includes a rotor having a plurality of magnetic poles, and typically includes two to eight pairs of magnetic poles. As shown in FIG. 1, a three-phase BLDC motor has six states of commutation (also referred to herein as a Hall state). If all six states in the commutation sequence have been executed, the sequence is repeated to continue the rotation. The number of pairs of magnetic poles in the annular magnet determines the number of electrical rotations per mechanical rotation. For example, a BLDC having a rotor with two pairs of magnetic poles requires two electrical rotations to spin the motor once. In other words, two electrical rotations generate one mechanical rotation.

  Hall effect sensors in BLDC motors are typically used to sense the pole position of the rotor's annular magnet and commutate the motor based on changes in Hall effect sensor signals. In other words, the Hall effect sensor is used to control the current in the stator winding, thereby controlling the torque of the BLDC motor. Hall effect sensors are used because they are cost-effective position sensors.

  A conventional BLDC motor has three Hall effect sensors embedded in the stator at the non-drive end of the motor. When the rotor pole passes near the Hall effect sensor, the Hall effect sensor provides a high or low signal (ie, pulse) indicating that the N or S pole passes near the Hall effect sensor. Based on the combination of the three Hall effect sensors, the exact sequence of commutation can be determined. The Hall state is defined by a predetermined position of the rotor relative to one or more Hall effect sensors or by a continuous set of predetermined positions of the rotor.

  In typical BLDC motor operation, two of the three phases of the BLDC motor conduct current while the third phase has zero current (ie, an invalid phase) for the motor to rotate. A conventional three-phase BLDC motor uses a Hall effect sensor. Each Hall state indicates which two of the three phases are valid (ie not invalid). The Hall state can be used to create a one-to-one relationship between the rotor phase and the direction in which voltage needs to be applied. Here are six possible hole phase combinations, which cover exactly one electrical rotation. Therefore, the position resolution using the three-phase Hall effect sensor is limited to 1/6 of the electrical rotation.

  Conventional methods for estimating motor speed are based on complete Hall effect pulses, triggering on the rising edge of the Hall effect signal, and the corresponding mechanical power at the last Hall period (ie, the first Hall cycle). Dividing by the duration of the period between the pulse and the second pulse. However, as shown in FIG. 4, the information collected in the conventional manner is delayed by one full electrical rotation of the motor. Therefore, it is desirable to obtain the velocity estimate with a faster response through different methods.

  As shown in FIG. 4, the method described herein estimates the motor speed with a faster response by focusing on the duration of individual Hall states. This technique of estimating speed can introduce more noise into the estimated motor speed. Thus, the method reduces the effects of noise on the estimated motor speed by learning the motor characteristics. The method reduces noise in the BLDC motor speed estimation without introducing the additional delay produced by the conventional average method.

  The present disclosure provides a method for estimating electric motor speed. The method determines whether an electric motor sequence cycle has occurred. The method calculates a correction factor for each hole state. If an electric motor sequence cycle occurs, the method also calculates an updated quality factor for the set of sequence cycle correction factors. In addition, the method updates the correction factor using the updated quality factor. The method updates the speed signal of the electric motor with the updated correction factor.

  The accompanying drawings are incorporated herein as part of the specification. The drawings described herein illustrate embodiments of the presently disclosed subject matter and are illustrative of selected principles and teachings of the present disclosure. However, the drawings are not intended to illustrate all possible implementations of the presently disclosed subject matter and are not intended to limit the scope of the present disclosure in any way.

FIG. 6 is a schematic diagram of a Hall effect signal during a single electrical rotation, in accordance with an embodiment of the presently disclosed subject matter.

FIG. 6 is a graphic illustration of a Hall effect sensor misalignment as a possible source of noise in a velocity signal, in accordance with an embodiment of the presently disclosed subject matter.

FIG. 6 is a graphic illustration of an irregularity of an annular magnet as a possible source of noise in a velocity signal, in accordance with an embodiment of the presently disclosed subject matter.

FIG. 3 is a graphical illustration of a time delay result utilizing a conventional method of estimating a motor speed of a Hall effect sensor, wherein the Hall effect sensor error for motor speed estimation is in accordance with an embodiment of the presently disclosed subject matter. Compared with the alignment learning method.

FIG. 5 is a graphical diagram of frequency domain results utilizing a conventional method of estimating the motor speed of a Hall effect sensor, and the present method of misalignment learning of the Hall effect sensor for estimation of the subject motor speed disclosed herein. Compared with

FIG. 4 is a graphical diagram of the speed signal noise reduction result using the second implementation of Hall effect sensor misalignment learning method for motor speed estimation, uncorrected results corrected Hall effect sensor misalignment The result of correcting the misalignment of the Hall effect sensor and the irregularity of the annular magnet is compared.

6 is a flowchart illustrating a method for estimating a motor speed of a Hall effect sensor, according to an embodiment of the presently disclosed subject matter.

6 is a flowchart illustrating a method for estimating a motor speed of a Hall effect sensor according to another embodiment of the presently disclosed subject matter.

  It is to be understood that the invention may assume various alternative adaptations and sequences of steps, unless explicitly specified otherwise. Certain devices, assemblies, systems, and processes illustrated in the accompanying drawings and described in the following specification are merely exemplary embodiments of the inventive concepts defined herein, Should be understood. Thus, specific dimensions, orientations, or other physical characteristics related to the disclosed embodiments should not be considered limiting unless explicitly stated otherwise. Also, although not so, like elements in the various embodiments described herein may be generally referred to by like reference numbers within this section of the application.

  BLDC motors are used in many industrial applications including automotive, aerospace, consumer goods, medical, industrial automation equipment and instrumentation applications. The subject matter disclosed herein may be utilized in the operation of an all-wheel drive vehicle connect / disconnect system. However, those skilled in the art will recognize that the subject matter disclosed herein can be applied to any application that utilizes an electric motor and a Hall effect sensor.

  In an embodiment, the present methods 200, 300 update the BLDC motor speed signal when any one of a plurality of Hall effect sensors changes state. As shown in FIGS. 2 and 3, utilizing the method 200, 300 may generate noise in the signal produced by Hall effect sensor offsets and ring magnet irregularities. Thus, in order to minimize the effects of noise in the signal, the methods 200, 300 compare the duration of the individual Hall states with the duration of the entire sequence cycle. The entire sequence cycle is an electrical rotation when performing Hall effect sensor offset correction, as in Method 200, and when performing an annular magnet irregularity correction, as in Method 300. Is mechanical rotation. The first step 204, 304 of the method determines whether the entire sequence cycle has occurred. If it is determined at 204A, 304A that the entire sequence cycle has not occurred, the methods 200, 300 are corrected at the fifth step 212, 312 (illustrated in the figure) calculated over the previous sequence cycle. The electric motor speed is calculated using the coefficient. If it is determined at 204B, 304B that the entire sequence cycle has occurred, the method 200, 300 continues to the second step 206, 306 described in the figure. The methods 200, 300 may be an iterative process, and the method resumes at the first step 204, 304 after the motor speed is calculated at the fifth step 212, 312.

  As shown in FIGS. 7 and 8, the second step 206, 306 of the method 200, 300 includes calculating a new correction factor. In the case of Hall effect sensor offset correction as in method 200 (see FIG. 7), there is one correction factor for each Hall state. In an example where there are three Hall effect sensor combinations, there are n = 6 correction factors. In the method 300, in the example with a combination of three Hall effect sensors, the irregularities of the annular magnet are also corrected (see FIG. 8). In method 300, there is a correction factor of n = 6 * p, where p is the number of magnetic poles on the annular magnet. Referring now to FIG. 1, an annular magnet 100 having a pair of magnetic poles (ie, two magnetic poles) is illustrated in each of the six hole states, and the change in each hole state is represented by points 102, 104, 106. , 108, 110, 112.

Reference is now made to a method 200 directed to offset correction of a Hall effect sensor, as shown in FIGS. Regardless of the initial Hall condition, a new correction factor can be calculated so that if a full electrical rotation occurs, the motor direction does not change during the full electrical rotation. If the BLDC motor is moving in such a way that the Hall condition proceeds in the same direction (eg, | 123456 | 123456 |) so that the BLDC motor completes electrical rotation in one direction, Proceed to calculate In other words, the method is that each of the change in hole state 102, 104, 106, 108, 110, 112 when the previous electrical rotation was moving in a single direction regardless of the initial hole state. Later, the process proceeds to second steps 206 and 306 for calculating a new correction coefficient. For example, the following sequence illustrates a complete electrical rotation in one direction (ie, 6 hole states), indicating that the initial hole state is irrelevant. The six hole states in a single electrical rotation moving in a single direction are underlined.
1 | 234456 | 23456; 12 | 345612 | 3456; 123 | 456123 | 456 etc. The BLDC motor speed is calculated after each Hall state change 102, 104, 106, 108, 110, 112. However, the correction factor for the hall condition described in the second steps 206, 306 can only be updated if the BLDC motor was moving in a single direction in the last complete sequence cycle.

  As described above with respect to t, if the entire sequence cycle has not occurred, the third step 208, 308 and the fourth step 210, 310 of the method 200, 300 described below are omitted, and the method 200, 300 calculates the motor speed using the correction vector of the previous sequence cycle (see the fifth steps 212 and 312).

を利用して算出され得る。 The correction coefficient c _i for each hole state _i is

Can be calculated using.

Here, n is the number of correction coefficients, t _i is the time spent in one hole state, and t _cycle is the time spent in the entire sequence cycle. For Hall effect sensor correction only, there are three Hall effect sensor offsets and six correction factors. The resulting correction factor does not directly describe the physical Hall effect sensor offset. Although the physical Hall effect sensor offset may be derived from the correction factor, the derivation of the Hall effect sensor offset is not further described herein. However, as illustrated in the figure, the relationship between the correction factor and the physical Hall effect sensor offset can be used to evaluate the quality of a new set of correction factors from the entire sequence cycle. .

  The third step 208, 308 of the method 200, 300 involves calculating an update quality factor to determine the quality of the current set of correction factors from the entire sequence cycle, and quality at a value between 0 and 1 Including the step of capturing information about. The quality of the current correction factor depends in part on the acceleration effect during the entire sequence cycle in which the correction factor was calculated. Thus, if the speed is kept approximately constant, the method obtains an accurate estimate of the Hall effect sensor offset and the irregularity of the annular magnet, and if the speed increases, the acceleration effect becomes the Hall effect sensor offset. And distort the estimation of the irregularity of the annular magnet. The electrical rotation is shorter than the mechanical rotation, so the possibility of obtaining a favorable condition during correction of the Hall effect sensor is more than the possibility of obtaining a favorable condition during correction of the irregularity of the annular magnet. high.

Procedures for calculating the quality factor for the current set of Hall effect sensor correction factors include, but are not limited to, the following three procedures. The first procedure is to determine whether the acquired correction vector (ie, the set of correction coefficients for the sequence cycle) is physically feasible, and use the acquired correction vector as a physical constraint on the Hall effect sensor position. Confirming by comparing. In the first procedure, the accelerations of the BLDC motors provide Hall effect sensor signals that indicate Hall effect sensor positions that are different from their predetermined positions. For example, if one of the Hall effect sensors is offset by a certain factor, the effect of full electrical rotation on the six Hall states takes the following form:
[1 + ε, 1,1-ε, 1 + ε, 1,1-ε]

  The second procedure involves using an external signal, such as the motor duty ratio, to determine the actual BLDC motor speed and to provide a secondary source for detecting acceleration. . During motor commutation, the duty ratio controls the voltage in the commutation phase. A duty ratio of 0% indicates that the three-phase inverter switch is always closed, and a duty ratio of 100% indicates that the three-phase inverter switch is always open. Setting the duty ratio between 0% and 100% means that the commutation phase of the electric motor is controlled to a desired voltage between zero and maximum voltage using a pulse width modulation signal. To do.

  A third procedure for obtaining a quality factor related to the correction factor of the Hall effect sensor includes detecting a tendency of acceleration using the correction factor itself. If all of the previous correction factors were clearly increased or decreased, the BLDC motor was accelerated. In each procedure, the higher the acceleration level, the lower the quality factor of the update. A predetermined value between 0 and 1 is assigned to the acceleration range of the BLDC motor.

The fourth step 210, 310 of the method 200, 300 includes updating the correction factor using the quality factor calculated in the third step 208, 308 of the method 200, 300. The correction coefficient vector may be updated using a quality factor as follows.

Where c _updated is the updated correction vector, c _new is the new correction vector, w _new is the quality factor, c _old is the correction vector previously calculated from the last sequence cycle, and w _total is The sum of the previous quality factors. In the fourth step 210, 310, the correction factor may be updated utilizing other methods, including but not limited to incorporating a form of forgetting factor.

The fifth steps 212, 312 of the methods 200, 300 include calculating the BLDC motor speed using the updated correction factor as follows.

は、１秒あたりの度（度／秒）のモータ速度、ｃ_ｉは、先のホール状態に関する補正係数、αは、ホール状態毎の機械的度数、ｔ_ｐｒｅｖは、先のホール状態において費やされる時間、である。方法２００、３００は、最低モータ速度の実装のため、方向反転の場合のｔ_ｐｒｅｖの外れ値の確認のため、及び／又は、現在のタイマと比較して予測的な手法で減速を検出するため、の更なるステップも含み得る。例えば、第１のホール状態において費やされる時間が１ｍｓで、且つ、モータが、現在第２のホール状態である場合、方法２００、３００は、第１のホール状態においてモータが費やした時間を利用して、通常、モータ速度を算出する。しかしながら、第２のホール状態において費やされる時間が、第１のホール状態の持続時間（すなわち、この例においては１ｍｓ）より長いことが明らかで、且つ、モータが、まだ第３のホール状態ではない場合、方法２００、３００は、第２のホール状態を利用してモータ速度を算出し得る。それは、電気モータ速度が減少していることが明らかであるからである。 here,

Is the motor speed in degrees per second (degrees / second), c _i is the correction factor for the previous hole state, α is the mechanical power for each hole state, and t _prev is spent in the previous hole state Time. The methods 200, 300 are for implementing a minimum motor speed, for checking outliers in t _prev in case of direction reversal, and / or for detecting deceleration in a predictive manner compared to the current timer. , Further steps may be included. For example, if the time spent in the first hall state is 1 ms and the motor is currently in the second hall state, the methods 200 and 300 use the time spent by the motor in the first hall state. Normally, the motor speed is calculated. However, it is clear that the time spent in the second Hall state is longer than the duration of the first Hall state (ie 1 ms in this example) and the motor is not yet in the third Hall state. If so, the methods 200, 300 may use the second Hall state to calculate the motor speed. This is because it is clear that the electric motor speed is decreasing.

  Methods 200, 300 may be implemented in a number of ways. In the first embodiment, the methods 200, 300 are performed under controlled conditions, such as in end-of-line testing, to map individual BLDC motor characteristics. This implementation assumes that Hall effect sensor misalignment and annular magnet irregularities do not change over the life of the motor. However, in the case of annular magnet correction, the motor learns the position of the annular magnet again each time the motor is initialized.

  In a second embodiment, the methods 200, 300 are performed online during operation of the associated system (eg, an automobile). This second implementation can provide value in harsh environments where the alignment within the motor can change over the life of the motor.

While various embodiments of the disclosed subject matter have been described above, it should be understood that they have been presented for purposes of illustration and not limitation. It will be apparent to those skilled in the art that the disclosed subject matter can be embodied in other specific forms without departing from its spirit or essential characteristics. The embodiments described above are thus considered in all respects as illustrative and not restrictive.
According to a first aspect of the present invention, there is provided a method for estimating electric motor speed, comprising: a stator having at least one Hall effect sensor coupled thereto; and a rotor having at least one pair of magnetic poles. Providing a step, calculating a correction factor for each Hall condition of the electric motor, determining whether a sequence cycle of the electric motor has occurred, and updating when the electric motor has gone through the sequence cycle Calculating a quality factor, updating the correction factor using the updated quality factor when the entire sequence occurs, and calculating a speed signal of the electric motor using the correction factor And a method comprising:
As a second aspect of the present invention, the first aspect further comprises the step of calculating the electric motor speed using a correction vector from a previous sequence cycle when the second sequence cycle has not occurred. A method for estimating the electric motor speed is provided.
As a third aspect of the present invention, there is provided a method for estimating an electric motor speed according to the first or second aspect, wherein the sequence cycle is one electric rotation.
According to a fourth aspect of the present invention, the step of calculating a correction coefficient for one of the hole states includes determining the number of hole states in the sequence cycle and determining the duration of each of the hole states. Determining a duration of the sequence cycle; multiplying the number of hole states in the sequence cycle by the duration of one of the hole states to generate a first product; And a method of estimating the electric motor speed according to any one of the first to third aspects, comprising: dividing the first product by the duration of the sequence cycle.
As a fifth aspect of the present invention, the step of calculating the update quality factor compares the correction factor of the sequence cycle with a physical constraint of the at least one Hall effect sensor to determine the accuracy of the correction factor. A method for estimating the electric motor speed according to any one of the first to fourth aspects, comprising the steps of: determining a value from the accuracy.
As a sixth aspect of the present invention, the step of calculating the update quality factor includes the steps of determining the speed of the electric motor from the duty ratio of the electric motor, and determining the acceleration of the electric motor from the duty ratio. And determining a value from the acceleration of the electric motor. 5. A method for estimating an electric motor speed according to any one of the first to fourth aspects is provided.
As a seventh aspect of the present invention, the step of calculating the update quality factor compares the correction factor of each of the sequence cycles to determine the acceleration of the electric motor, and the value from the acceleration of the electric motor Determining an electric motor speed according to any one of the first to fourth aspects.
According to an eighth aspect of the present invention, the step of updating the correction factor using the update quality factor includes the step of multiplying a new correction vector by a new update quality factor to generate a first product, Multiplying the previously calculated correction vector by the sum of the updated quality factors to generate a second product, and adding the first product to the second product, Generating, adding the new update quality factor to the sum of the update quality factors to generate a second sum, and dividing the first sum by the second sum The method for estimating the electric motor speed according to any one of the first to seventh aspects is provided.
According to a ninth aspect of the present invention, the step of calculating the speed signal of the electric motor using the updated correction coefficient includes setting one of the correction coefficients to the Hall state associated with the correction coefficient. Multiplying by the mechanical power moved by the rotor during the period to generate a first product, and dividing the first product by the duration of the hole state prior to the hole state; A method for estimating an electric motor speed according to any one of the first to eighth aspects is provided.
As a tenth aspect of the present invention, there is provided a method for estimating an electric motor speed according to any one of the first to ninth aspects, wherein the sequence cycle is one mechanical rotation.
As an eleventh aspect of the present invention, there is provided a method for estimating an electric motor speed according to any one of the first to tenth aspects, wherein the electric motor is controlled by a cascaded closed loop controller. .
As a twelfth aspect of the present invention, there is provided the method of estimating an electric motor speed according to any one of the first to eleventh aspects, wherein the step of updating the correction coefficient includes the step of incorporating a forgetting coefficient. The
As a thirteenth aspect of the present invention, the method for estimating the electric motor speed according to any one of the first to twelfth aspects, further comprising mapping the characteristics of the electric motor during an end-of-line test. Is provided.
As a fourteenth aspect of the present invention, there is provided a method for estimating an electric motor speed according to any one of the first to thirteenth aspects, wherein the electric motor is coupled to an all-wheel drive cutoff device of an automobile. The

Claims (14)

A method for estimating electric motor speed, comprising:
Providing an electric motor including a stator having at least one Hall effect sensor coupled thereto and a rotor having at least one pair of magnetic poles;
Calculating a correction factor for each hall state of the electric motor by comparing a plurality of hall state periods with a sequence cycle period ;
Determining whether a sequence cycle of the electric motor has occurred;
Calculating an updated quality factor when the electric motor has undergone the sequence cycle;
A step of updating a plurality of the correction coefficient by using the quality factor of the update when the entire of the sequence cycle occurs,
Calculating a speed signal of the electric motor using the plurality of correction factors.   The method of estimating an electric motor speed according to claim 1, further comprising calculating the electric motor speed using a correction vector from a previous sequence cycle when a second sequence cycle has not occurred.   The method for estimating an electric motor speed according to claim 1, wherein the sequence cycle is one electric rotation. Calculating a correction factor for one of the hole states;
Determining the number of hole states in the sequence cycle;
Determining the duration of each said hole state;
Determining the duration of the sequence cycle;
Multiplying the number of hole states in the sequence cycle by the duration of one of the hole states to produce a first product;
4. The method of estimating an electric motor speed according to claim 1, comprising dividing the first product by the duration of the sequence cycle. 5. The stage of calculating the update quality factor is
Comparing the correction factor of the sequence cycle with physical constraints of the at least one Hall effect sensor to determine the accuracy of the correction factor;
The method of estimating an electric motor speed according to claim 1, comprising determining a value from the accuracy. The stage of calculating the update quality factor is
Determining a speed of the electric motor from a duty ratio of the electric motor;
Determining an acceleration of the electric motor from the duty ratio;
The method of estimating an electric motor speed according to claim 1, comprising determining a value from an acceleration of the electric motor. The stage of calculating the update quality factor is
Comparing the correction factors for each of the sequence cycles to determine the acceleration of the electric motor; and
The method of estimating an electric motor speed according to claim 1, comprising determining a value from an acceleration of the electric motor. Updating the correction factor using the update quality factor comprises:
Multiplying a new correction vector by a new updated quality factor to produce a first product; and
Multiplying the previously calculated correction vector by the sum of the updated quality factors to generate a second product;
Adding the first product to the second product to generate a first sum; and
Adding the new update quality factor to the sum of the update quality factors to generate a second sum;
Dividing the first sum by the second sum. 8. A method of estimating an electric motor speed according to any one of the preceding claims. The step of calculating the speed signal of the electric motor using the updated correction factor includes:
Multiplying one of the correction factors by the mechanical power moved by the rotor during the Hall condition associated with the correction factor to produce a first product;
9. The method of estimating an electric motor speed according to any one of claims 1 to 8, comprising dividing the first product by a duration of a hall state prior to the hall state.   The method for estimating an electric motor speed according to claim 1, wherein the sequence cycle is one mechanical rotation.   11. A method for estimating electric motor speed according to any one of the preceding claims, wherein the electric motor is controlled by a cascaded closed loop controller.   12. A method for estimating electric motor speed according to any one of the preceding claims, wherein updating the correction factor comprises incorporating a forgetting factor.   13. A method for estimating electric motor speed according to any one of the preceding claims, further comprising mapping the characteristics of the electric motor during an end-of-line test.   14. A method for estimating an electric motor speed according to any one of claims 1 to 13, wherein the electric motor is coupled to an all-wheel drive shut-off device of an automobile.

Images