CN113067514A - Rapid online rotational inertia identification method and system suitable for servo system - Google Patents

Rapid online rotational inertia identification method and system suitable for servo system Download PDF

Info

Publication number
CN113067514A
CN113067514A CN202110312137.0A CN202110312137A CN113067514A CN 113067514 A CN113067514 A CN 113067514A CN 202110312137 A CN202110312137 A CN 202110312137A CN 113067514 A CN113067514 A CN 113067514A
Authority
CN
China
Prior art keywords
motor
identification
angular velocity
rotational inertia
feedback current
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110312137.0A
Other languages
Chinese (zh)
Other versions
CN113067514B (en
Inventor
童乔凌
王旭宝
闵闰
刘涛
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huazhong University of Science and Technology
Original Assignee
Huazhong University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huazhong University of Science and Technology filed Critical Huazhong University of Science and Technology
Priority to CN202110312137.0A priority Critical patent/CN113067514B/en
Publication of CN113067514A publication Critical patent/CN113067514A/en
Application granted granted Critical
Publication of CN113067514B publication Critical patent/CN113067514B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
    • H02P21/14—Estimation or adaptation of machine parameters, e.g. flux, current or voltage
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
    • H02P21/14—Estimation or adaptation of machine parameters, e.g. flux, current or voltage
    • H02P21/16—Estimation of constants, e.g. the rotor time constant
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
    • H02P21/14—Estimation or adaptation of machine parameters, e.g. flux, current or voltage
    • H02P21/18—Estimation of position or speed
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
    • H02P21/14—Estimation or adaptation of machine parameters, e.g. flux, current or voltage
    • H02P21/20—Estimation of torque
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
    • H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
    • H02P21/22—Current control, e.g. using a current control loop

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Control Of Electric Motors In General (AREA)

Abstract

The invention discloses a quick online rotary inertia identification method and a quick online rotary inertia identification system suitable for a servo system. The current and the rotating speed are synchronously processed by considering the sampling delay of the current so as to ensure the validity of algorithm input data, the speed and the current are synchronously processed by adopting a digital moving averaging method, and even a digital processing chip with weak computing capability can meet the computing requirement.

Description

Rapid online rotational inertia identification method and system suitable for servo system
Technical Field
The invention belongs to the field of parameter identification of servo parameter self-tuning, and particularly relates to a quick online rotational inertia identification method and system suitable for a servo system.
Background
The identification of the rotational inertia in the alternating current servo system has very important significance for the parameter self-tuning of the speed ring, and the fluctuation of the rotational inertia in the actual system can also deteriorate the dynamic performance of a servo driver and influence the bandwidth of the speed ring. The upper limit of the speed loop bandwidth of the alternating current system, namely the system limit bandwidth, is a basic parameter influencing the design of the speed loop PI controller, the change of the rotational inertia of the system is monitored in real time, and the parameters of the speed loop PI controller are adjusted in a matched mode, so that the system can be guaranteed to be in a good running state all the time.
At present, the common rotational inertia identification in the domestic market mostly uses offline acceleration and deceleration identification and online gradient correction-based identification algorithms, the offline acceleration and deceleration identification algorithm is simple, but is limited by factors such as delay of given instruction feedback calculation, external working condition change and the like, and the actual use scene is limited; the online inertia identification calculation amount based on gradient correction is small, so that the identification parameters can be iteratively converged towards the negative gradient direction of the criterion function, and the identification parameters are gradually converged to the optimal identification result that the criterion function reaches the minimum value, but because the iterative gain is a constant value, the rapid convergence can not be achieved by reflecting the fluctuation of input data, and the online convergence time is long; other methods for identifying the rotational inertia, such as a genetic algorithm, a kalman filtering algorithm and the like, cannot meet the actual industrial application requirements due to high requirements on the computing power of a chip of a digital processor.
Disclosure of Invention
Aiming at the defects or improvement requirements of the prior art, the invention provides a quick online rotational inertia identification method and system suitable for a servo system, so that the problems that the traditional rotational inertia identification method is limited under actual working conditions, long in identification convergence time and the like are solved.
To achieve the above object, according to one aspect of the present invention, there is provided a fast online rotational inertia identification method for a servo system, the method including:
s101: make the motor start rotating under the condition that the acceleration is not zeroObtaining the feedback current i of the motor in the preset rotation rangeFirst stageAnd angular velocity ωFirst stageBased on the feedback current iFirst stageAnd angular velocity ωFirst stageCalculating the initial value J of the moment of inertia of the motor0；
S102: under the condition that the rotation range of the motor is larger than the preset rotation range, sampling a feedback current i and an angular speed omega of the motor in real time, and synchronously processing the feedback current i and the angular speed omega;
s103: judging whether the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than a feedback current change rate threshold, an angular velocity change rate threshold and an angular velocity threshold, if so, performing a step S104, otherwise, returning to the step S102;
s104: according to the initial value J of the identification of the rotational inertia of the motor0The current i and the angular velocity omega are fed back, iterative calculation is carried out based on a motor rotational inertia identification algorithm, and an identification final value J of the motor rotational inertia is obtainedk。
Preferably, the initial identification value calculation formula of the rotational inertia of the motor is as follows:
Figure BDA0002989025690000021
wherein, TsFor identifying the period, KtIs the torque coefficient of the motor, can be obtained by the conversion of the AD sampling feedback of the servo driver, iBeginning qTo the feedback current iFirst stageQ-axis current, delta omega, obtained after CLARKE conversion, PARK conversion and per unit processingFirst stageIs the angular velocity variation in one recognition period.
Preferably, in step S102, after sampling the feedback current i and the angular velocity ω of the motor in real time and before performing synchronization processing on the feedback current i and the angular velocity ω, if the rotation range of the motor is greater than the preset rotation range, the method further includes:
and performing per unit processing on the feedback current i and the angular velocity omega.
Preferably, the motor rotational inertia identification algorithm is any one of a gaussian least square algorithm, a gradient correction algorithm or a model reference adaptive method.
Preferably, when the motor rotational inertia identification algorithm is a gaussian least square algorithm, the iterative calculation formula is as follows:
Figure BDA0002989025690000031
wherein the content of the first and second substances,
Figure BDA0002989025690000032
in order to estimate the variables, the variables are,
Figure BDA0002989025690000033
y (k) is an actual variable, and y (k) is ωm(k)-2ωm(k-1)+ωm(k-2); k (k) is a gain variable, and the iterative formula of K (k) is:
Figure BDA0002989025690000034
p (k) is an iterative covariance variable, and p (k) is:
Figure BDA0002989025690000035
mu is a forgetting factor, and mu is a forgetting factor,
Figure BDA0002989025690000037
is an identification variable;
according to the identification variable
Figure BDA0002989025690000038
And a recognition period TsCalculating to obtain a final value J of the moment of inertia identificationkThe calculation formula is as follows:
Figure BDA0002989025690000036
preferably, the initial value of the rotational inertia of the motor is calculated and the motor rotatesFinal identification value J of dynamic inertiakThe iterative computation of (2) all adopts an integer arithmetic mode.
According to another aspect of the present invention, there is provided a fast online rotational inertia identification system suitable for a servo system, the system comprising an identification initial estimation unit, a sampling and synchronization unit, an iterative update judgment unit and an identification algorithm calculation unit;
the identification initial estimation unit is used for enabling the motor to start rotating under the condition that the acceleration is not zero and acquiring the feedback current i of the motor in a preset rotating rangeFirst stageAnd angular velocity ωFirst stageBased on the feedback current iFirst stageAnd angular velocity ωFirst stageCalculating the initial value J of the moment of inertia of the motor0Identifying the initial value J0Outputting to a sampling and synchronization unit;
the sampling and synchronizing unit is used for sampling the feedback current i and the angular velocity omega of the motor in real time under the condition that the rotation range of the motor is larger than the preset rotation range, synchronously processing the feedback current i and the angular velocity omega to obtain the feedback current i and the angular velocity omega which are synchronously processed, and outputting the feedback current i and the angular velocity omega to the iteration updating and judging unit;
the iteration updating and judging unit is used for judging whether the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than a feedback current change rate threshold, an angular velocity change rate threshold and an angular velocity threshold:
if yes, the identification algorithm calculation unit identifies an initial value J according to the rotational inertia of the motor0The current i and the angular velocity omega are fed back, iterative calculation is carried out based on a motor rotational inertia identification algorithm, and an identification final value J of the motor rotational inertia is obtainedk；
If not, the sampling and synchronizing unit continues to sample the feedback current i and the angular speed omega of the motor in real time and performs synchronous processing on the feedback current i and the angular speed omega.
In general, compared with the prior art, the above technical solution contemplated by the present invention can achieve the following beneficial effects:
1. before the identification formula is calculated, a section of data is extracted on line by limiting the motion state of the motor, a more accurate identification initial value is estimated through the motion control model of the motor, the convergence time of the algorithm can be greatly accelerated by estimating the initial value, and the effect of quick on-line identification is achieved. The current and the rotating speed are synchronously processed by considering the sampling delay of the current so as to ensure the validity of algorithm input data, the speed and the current are synchronously processed by adopting a digital moving averaging method, and even a digital processing chip with weak computing capability can meet the computing requirement. The sampled current and angular velocity data can be the current and angular velocity data output by the servo system in an online position and velocity mode, and can also be the current and angular velocity data output by the servo system after given velocity excitation in an offline mode.
2. The identification is carried out based on the Gaussian least square algorithm, so that the online identification speed of the rotational inertia is further effectively improved; and the identification formula is a scalar formula, vector operation is not involved, the calculation requirement on the digital processing chip is not high, the identification algorithm can be completed in a control period of a plurality of alternating current servo systems, and the integrity of the main control logic and the communication of the motor is ensured.
3. The method adopts an integer arithmetic mode when calculation is carried out in a digital processor chip, floating point arithmetic is not involved, the integer arithmetic is expanded to be the original integer multiple for calculation, the time of one control period of a servo system is occupied as little as possible, and although some data precision is lost, the identification result still has high precision and does not influence the identification precision.
Drawings
FIG. 1 is a flowchart of a method for fast online rotational inertia identification for a servo system according to an embodiment of the present invention;
FIG. 2 is a control block diagram of a servo driver provided by an embodiment of the present invention;
FIG. 3 is a graph showing the comparison result between the speed setting and the speed feedback of the servo system in the online mode according to the method for identifying the fast online rotational inertia of the servo system provided by the embodiment of the present invention;
fig. 4 is a torque current variation diagram of the servo system in the online mode according to the fast online rotational inertia identification method for the servo system provided by the embodiment of the present invention;
FIG. 5 is a diagram illustrating a result of fast online rotational inertia identification for a servo system according to an embodiment of the present invention;
fig. 6 is a block diagram of a fast online rotational inertia identification system suitable for a servo system according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
The embodiment of the invention provides a quick online rotational inertia identification method suitable for a servo system, which comprises the following steps of:
s101: the motor starts to rotate under the condition that the acceleration is not zero, and the feedback current i of the motor in the preset rotation range is obtainedFirst stageAnd angular velocity ωFirst stageBased on the feedback current iFirst stageAnd angular velocity ωFirst stageAnd estimating an initial identification value J of the rotational inertia of the motor.
Specifically, a load motor starts to rotate under a speed change state with acceleration not being zero, and a section of feedback current and angular speed data of the motor are extracted on line under a transient acceleration and deceleration state of the motor to estimate an initial identification value of the moment of inertia. The preset rotation range can be n circles, and n is a real number larger than zero. For example, when n is 1, the motor is rotated for 1 circle through an online position mode in an initial state, and a section of effective data is extracted by limiting the motor to a motion state with acceleration not equal to zero to estimate an initial value of rotational inertia identification.
The given speed of the load motor is a waveform with acceleration not zero, including but not limited to triangular wave, trapezoidal wave, and sine wave.
Further, the initial value estimation of the moment of inertia is realized by calculating a motor control motion equation on the premise of neglecting the load torque, and the calculation formula is as follows:
Figure BDA0002989025690000061
wherein, TsFor identifying the period, KtIs the torque coefficient of the motor, can be obtained by the conversion of the AD sampling feedback of the servo driver, iBeginning qTo the feedback current iFirst stageQ-axis current, delta omega, obtained after CLARKE conversion, PARK conversion and per unit processingFirst stageIs the angular velocity variation over an identification period.
Delta omega is key data to be extracted, and the extraction interval TsAnd the change of the moment should be taken into consideration, and the output moment is always changed by setting the motor to rotate under the condition that the acceleration is not zero, so as to avoid the wrong estimation and prolong the convergence time instead. Even if fast online identification requires several identification cycles, the drive position control scheme should relax the limit on the given speed of the speed loop to extract more valid data within a small range of motor rotation.
Equation (1) ignores the load moment term compared to the commonly used motor motion control model, so J is an approximate estimate of the moment of inertia.
The estimation needs to be performed under the strict condition limitation, that is, the rotational acceleration of the motor is not equal to zero, so as to avoid longer iteration time caused by wrong estimation, and the initial value of the estimation can greatly reduce the subsequent identification iteration calculation time, thereby realizing the rapid online identification of the rotational inertia.
In order to ensure the effectiveness of estimation, the average value of multiple estimation results is taken as the initial identification value of the moment of inertia.
S102: and under the condition that the rotation range of the motor is larger than the preset rotation range, sampling the feedback current i and the angular speed omega of the motor in real time, and synchronously processing the feedback current i and the angular speed omega.
Specifically, under the condition that the rotation range of the motor is larger than the preset rotation range, the feedback current and the angular speed of the motor are sampled at sampling intervals of a plurality of identification periods. In order to ensure the synchronization of the angular velocity and the current, the collected feedback current i and the angular velocity omega are stored in a fixed length in a plurality of identification periods, the stored input data are averaged, the average value processing of the identification algorithm period is carried out by adopting a digital moving averaging method, and even a digital processing chip with weak computing capability can meet the computing requirement.
For example: the current and angular velocity are stored for 8 cycles of sampled data. And dividing the sum of the current sampling values of 8 periods by 8 to obtain an average value of 8 periods, namely obtaining the transient current value before 4 periods. The angular speed is calculated by taking the first value and the last value of the photoelectric encoder in 8 periods, calculating the difference value and then dividing by 8 to obtain the angular speed transient value before 4 periods.
Further, the angular velocity is obtained by measuring the velocity by M method using a photoelectric encoder of the motor, and other encoders, such as a rotary encoder and an absolute value encoder, may be used as long as the synchronization process of the velocity and the current can be ensured.
Further, if the ripple of hardware sampling is large, low-pass filtering processing can be performed on the current and the angular velocity at the same time, as long as the first-order inertia delay brought by the filter is the same, that is, the angular velocity and the current are still kept synchronous. The cut-off frequency of the filtering can not be lower than the bandwidth of the servo speed loop, otherwise, errors can occur in online rotational inertia identification in the position mode.
Further, the real-time sampling of the feedback current i and the angular speed ω of the motor may be data sampled in a position and speed mode in a case that a rotation range of the motor is greater than the preset rotation range; or the data sampled under the condition that the rotation range of the motor is larger than the preset rotation range after the given speed excitation is excited to the servo system in the off-line mode, wherein the given speed excitation can be sine waves, triangular waves or trapezoidal waves.
S103: and judging whether the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than a feedback current change rate threshold, an angular velocity change rate threshold and an angular velocity threshold, if so, performing step S104, otherwise, returning to step S102.
Specifically, the sampled and synchronized data are screened, and when the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than a feedback current change rate threshold, an angular velocity change rate threshold and an angular velocity threshold, step S104 is performed to iteratively update the identification value of the moment of inertia; otherwise, keeping the identification intermediate quantity of the previous beat unchanged, returning to the step S102 to continue sampling and synchronizing the feedback current and the angular speed, and waiting for the next iteration identification processing to avoid the divergence of the identification result caused by wrong iteration calculation in the stable speed operation. The reason is that the rotational inertia and the acceleration in the motion equation of the motor are coupled together, and if the acceleration of the motor does not change, the parameter identification cannot be performed according to the parameter identification theory.
It will be appreciated that the current rate of change threshold, the angular velocity rate of change threshold, and the angular velocity threshold may be determined from a number of experiments.
S104: according to the initial value J of the identification of the rotational inertia of the motor0The feedback current i and the angular velocity omega are subjected to iterative calculation based on a motor rotational inertia identification algorithm to obtain an identification final value J of the motor rotational inertiak。
Specifically, when the change rate of the feedback current, the change rate of the angular velocity, and the absolute value of the angular velocity are respectively greater than the feedback current change rate threshold, the angular velocity change rate threshold, and the angular velocity threshold, the electromagnetic torque of the motor is calculated according to the feedback current, iterative calculation is performed based on a motor rotational inertia identification algorithm according to the initial identification value of the motor rotational inertia, the electromagnetic torque, and the angular velocity, iterative identification is performed on the motor rotational inertia on the basis of the initial identification value of the motor rotational inertia, the identification value of the output rotational inertia is updated in real time, and finally the final identification value of the motor rotational inertia is obtained.
According to the method provided by the embodiment of the invention, a section of data is extracted online by limiting the motion state of the motor before the identification formula is calculated, a more accurate identification initial value is estimated through the motion control model of the motor, the convergence time of the algorithm can be greatly accelerated by estimating the initial value, and the effect of rapid online identification is achieved. The current and the rotating speed are synchronously processed by considering the sampling delay of the current so as to ensure the validity of algorithm input data, the speed and the current are synchronously processed by adopting a digital moving averaging method, and even a digital processing chip with weak computing capability can meet the computing requirement. The sampled current and angular velocity data can be the current and angular velocity data output by the servo system in an online position and velocity mode, and can also be the current and angular velocity data output by the servo system after given velocity excitation in an offline mode.
Optionally, in step S102, after sampling the feedback current i and the angular velocity ω of the motor in real time and before performing synchronization processing on the feedback current i and the angular velocity ω when the rotation range of the motor is greater than the preset rotation range, the method further includes:
and performing per unit processing on the feedback current i and the angular velocity omega.
Specifically, CLARKE conversion and PARK conversion are carried out on the motor feedback current i to obtain a q-axis current iqFor the q-axis current iqPerforming per-unit treatment to obtain q-axis current after per-unit treatment; obtaining the per-unit electrical angle from the feedback angle of the photoelectric encoder, storing two variables in a buffer with fixed length in a digital control processor, calculating the speed at fixed intervals, replacing the midpoint instantaneous speed with the angular speed average speed of a period of time, moving and averaging the per-unit q-axis current to obtain the period of timeCurrent corresponding to the instantaneous speed at the midpoint of time; and then low-pass filtering the averaged q-axis current and angular velocity, and keeping the two synchronous under the same first-order inertia delay.
According to the method provided by the embodiment of the invention, per unit data processing is carried out on the sampled current and the angular speed and the rotational speed according to the motor nameplate parameters, so that the problem of precision loss caused by calculation overflow can be effectively avoided.
Optionally, the motor rotational inertia identification algorithm is any one of a gaussian least squares algorithm, a gradient correction algorithm or a model reference adaptive method.
When the motor rotational inertia identification algorithm is a Gaussian least square algorithm, the iterative calculation formula is as follows:
Figure BDA0002989025690000101
wherein the content of the first and second substances,
Figure BDA0002989025690000102
in order to estimate the variables, the variables are,
Figure BDA0002989025690000103
y (k) is an actual variable, and y (k) is ωm(k)-2ωm(k-1)+ωm(k-2); k (k) is a gain variable, and the iterative formula of K (k) is:
Figure BDA0002989025690000104
p (k) is an iterative covariance variable, and p (k) is:
Figure BDA0002989025690000105
mu is a forgetting factor, and mu is a forgetting factor,
Figure BDA0002989025690000106
is an identification variable;
specifically, fig. 2 is a control block diagram of a servo driver for a servo system according to an embodiment of the present invention, and a control strategy employs idThe magnetic field orientation control mode is 0,idthe method is d-axis current obtained by CLARKE and PARK conversion of three-phase current i fed back by a motor. The controlled permanent magnet synchronous motor motion equation can be expressed as formula (2) when the viscosity-friction coefficient is ignored:
Figure BDA0002989025690000107
wherein, TeIs the electromagnetic torque of the machine, Te＝Kt·iq，TlIs the load torque of the motor, J is the total moment of inertia of the motor body containing the load, omegamThe discretization equations of the k time and the k-1 time of the motor, which are the mechanical angular velocity of the motor, can be expressed as an equation (2) and an equation (3):
Figure BDA0002989025690000108
Figure BDA0002989025690000109
when the recognition period T issWhen the load torque is small, it is considered that the load torque at two adjacent times is kept unchanged, and equation (5) is obtained by subtracting equation (3) from equation (4):
Figure BDA00029890256900001010
in the formula (5), ω (k) -2 ω (k-1) + ω (k-2) and Te(k)-TeAnd (k-1) can be obtained from the feedback angle and sampling current of the photoelectric encoder of the servo driver.
When the motor rotational inertia identification algorithm is a Gaussian least square algorithm, the order is given
Figure BDA0002989025690000111
y(k)＝ωm(k)-2ωm(k-1)+ωm(k-2)，ΔTe(k-1)＝Te(k-1)-Te(k-2) performing a conversion using the recurrence formula of the formula (6)Identification of dynamic inertia:
Figure BDA0002989025690000112
wherein the content of the first and second substances,
Figure BDA0002989025690000113
in order to identify the variable(s),
Figure BDA0002989025690000114
initial value J of identification according to moment of inertia0And (4) calculating.
The recursive gain variable K (k) in the formula (6) is iterated in each recognition algorithm, and the rapidity of algorithm convergence is guaranteed.
The iteration of the recursive gain variable k (k) and the recursive covariance variable p (k) can be obtained by equation (7):
Figure BDA0002989025690000115
wherein mu is a forgetting factor, and the value of mu is a constant less than 1.
Preferably, μ ═ 0.99. When μ is 0.99, the fluctuation of the recognition result is reduced, and the weight of the previous input data in the new moment of inertia recognition is gradually eliminated.
The initial value P (0) of the covariance matrix is a constant with a value of 100.
The specific iterative update is represented as: when the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than the feedback current change rate threshold, the angular velocity change rate threshold and the angular velocity threshold, the iteration of the covariance variable and the identification formula in the identification formula is performed, the recursive gain variable K (k) is calculated, and the initial iteration is performed according to the initial identification value J obtained in the step S1010Reiterative identification variables
Figure BDA0002989025690000116
The covariance variable is then updated for the next passAnd calculating a pushing gain variable.
Finally according to the identification variable
Figure BDA0002989025690000117
And a recognition period TsCalculating to obtain a final value J of the moment of inertia identificationk：
Figure BDA0002989025690000118
To maintain the validity of the recognition algorithm calculation, it is preferable that the recognition period T issThe time is 1ms to deal with the situation that identification cannot be carried out due to low-speed change in the online inertia identification in the position mode, and meanwhile, more sampling data can meet the requirement of an iteration updating judgment unit, and the identification speed is increased.
Preferably, to ensure the accuracy of data calculation and reduce the overflow problem of data calculation, all data are calculated by shifting left by 17 bits, and the step processing calculation is performed on step S103 under the condition that the requirement that the recognition period is 1ms is satisfied in consideration of the limit problem of the switching period of the servo driver.
According to the method provided by the embodiment of the invention, the identification is carried out based on the Gaussian least square algorithm, so that the online identification speed of the rotational inertia is further effectively improved; and the identification formula is a scalar formula, vector operation is not involved, the calculation requirement on the digital processing chip is not high, the identification algorithm can be completed in a control period of a plurality of alternating current servo systems, and the integrity of the main control logic and the communication of the motor is ensured.
Optionally, the initial identification value of the rotational inertia of the motor and the iterative calculation of the estimated identification value of the rotational inertia of the motor both adopt an integer operation mode.
The method provided by the embodiment of the invention adopts an integer arithmetic mode when the calculation is carried out in a digital processor chip, floating point arithmetic is not involved, the integer arithmetic is expanded to the original integer multiple for calculation, the time of one control cycle of a servo system is occupied as little as possible, and although some data precision is lost, the identification result still has high precision and does not influence the identification precision.
The method for identifying the rapid online rotational inertia applicable to the servo system provided by the embodiment of the invention is simulated, wherein the rotational inertia value of the adopted permanent magnet synchronous motor is 0.000263 kg-m2. Fig. 3 is a graph of speed setting and speed feedback comparison results of a servo system in an online mode according to the method for identifying a fast online rotational inertia of the servo system provided by the embodiment of the present invention; fig. 4 is a torque current waveform diagram of the servo system in the online mode according to the fast online rotational inertia identification method for the servo system provided by the embodiment of the invention; fig. 5 is a diagram illustrating a result of fast online rotational inertia identification for a servo system according to an embodiment of the present invention. As shown in fig. 3 to 5, the iteration of the identification result is not performed all the time, and when one of the three parameters of the change rate of the feedback current, the change rate of the angular velocity, and the absolute value of the angular velocity does not satisfy the critical condition, the identification result is stable and unchanged, so that the phenomenon that the convergence time of the identification result is prolonged due to the recursive invalid iterative computation is avoided. In addition, the online identification convergence time in the embodiment of the invention is only 2.2s, and the relative error of the identification result is only 1.43%, so that the reliability of the method provided by the invention is fully proved.
The following describes a fast online rotational inertia identification system suitable for a servo system, and the fast online rotational inertia identification system suitable for a servo system described below and the fast online rotational inertia identification method suitable for a servo system described above may be referred to in correspondence with each other.
The embodiment of the invention provides a fast online rotational inertia identification system suitable for a servo system, as shown in fig. 6, the system comprises an identification initial estimation unit 601, a sampling and synchronization unit 602, an iteration update judgment unit 603 and an identification algorithm calculation unit 604;
the identification initial estimation unit 601 is configured to start rotation of the motor under a condition that the acceleration is not zero, and obtain a feedback current i of the motor within a preset rotation rangeFirst stageAnd angular velocity ωFirst stageBased on the feedback current iFirst stageAnd angular velocity ωFirst stageCalculating the initial value J of the moment of inertia of the motor0Identifying the initial value J0Output to the sampling and synchronization unit 602;
the sampling and synchronizing unit 602 is configured to sample the feedback current i and the angular velocity ω of the motor in real time when the rotation range of the motor is greater than the preset rotation range, perform synchronization processing on the feedback current i and the angular velocity ω to obtain a feedback current i and an angular velocity ω after the synchronization processing, and output the feedback current i and the angular velocity ω to the iteration update determining unit 603;
the iterative update determination unit 603 is configured to determine whether the change rate of the feedback current, the change rate of the angular velocity, and the absolute value of the angular velocity are greater than a feedback current change rate threshold, an angular velocity change rate threshold, and an angular velocity threshold, respectively:
if yes, the identification algorithm calculation unit 604 identifies the initial value J of the rotational inertia of the motor0The current i and the angular velocity omega are fed back, iterative calculation is carried out based on a motor rotational inertia identification algorithm, and an identification final value J of the motor rotational inertia is obtainedk；
If not, the sampling and synchronizing unit 602 continues to sample the feedback current i and the angular velocity ω of the motor in real time, and performs synchronization processing on the feedback current i and the angular velocity ω.
According to the system provided by the embodiment of the invention, a section of data is extracted online by limiting the motion state of the motor before the identification formula is calculated, a more accurate identification initial value is estimated through the motion control model of the motor, the convergence time of the algorithm can be greatly accelerated by estimating the initial value, and the effect of rapid online identification is achieved. The current and the rotating speed are synchronously processed by considering the sampling delay of the current so as to ensure the validity of algorithm input data, the speed and the current are synchronously processed by adopting a digital moving averaging method, and even a digital processing chip with weak computing capability can meet the computing requirement. The sampled current and angular velocity data may be current and angular velocity data output by the servo system in a position and velocity mode, or current and angular velocity data output by the servo system after excitation at a given velocity in an offline mode.
Optionally, the system further comprises a rotational inertia identification output unit for outputting an identification final value of the rotational inertia of the motor.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims (7)

1. A quick online rotational inertia identification method suitable for a servo system is characterized by comprising the following steps:
s101: the motor starts to rotate under the condition that the acceleration is not zero, and the feedback current i of the motor in the preset rotation range is obtainedFirst stageAnd angular velocity ωFirst stageBased on the feedback current iFirst stageAnd angular velocity ωFirst stageCalculating the initial value J of the moment of inertia of the motor0；
S102: under the condition that the rotation range of the motor is larger than the preset rotation range, sampling a feedback current i and an angular speed omega of the motor in real time, and synchronously processing the feedback current i and the angular speed omega;
s103: judging whether the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than a feedback current change rate threshold, an angular velocity change rate threshold and an angular velocity threshold, if so, performing a step S104, otherwise, returning to the step S102;
s104: according to the initial value J of the identification of the rotational inertia of the motor0The current i and the angular velocity omega are fed back, iterative calculation is carried out based on a motor rotational inertia identification algorithm, and an identification final value J of the motor rotational inertia is obtainedk。
2. The method as claimed in claim 1, wherein the initial value of the motor rotational inertia is calculated by:
Figure FDA0002989025680000011
wherein, TsFor identifying the period, KtIs the torque coefficient of the motor, can be obtained by the conversion of the AD sampling feedback of the servo driver, iBeginning qTo the feedback current iFirst stageQ-axis current, delta omega, obtained after CLARKE conversion, PARK conversion and per unit processingFirst stageIs the angular velocity variation in one recognition period.
3. The method for fast online identifying rotational inertia of a servo system as claimed in claim 2, wherein in step S102, after sampling the feedback current i and the angular velocity ω of the motor in real time and before performing the synchronization process on the feedback current i and the angular velocity ω, when the rotation range of the motor is greater than the preset rotation range, the method further comprises:
and performing per unit processing on the feedback current i and the angular velocity omega.
4. The fast online rotational inertia identification method for the servo system as claimed in claim 1, wherein the motor rotational inertia identification algorithm is any one of a gaussian least squares algorithm, a gradient correction algorithm or a model reference adaptive method.
5. The method for fast online identification of rotational inertia suitable for servo system of claim 4, wherein when the motor rotational inertia identification algorithm is a Gaussian least squares algorithm, the iterative calculation formula is:
Figure FDA0002989025680000021
wherein, y(k-1) is an estimation variable,
Figure FDA0002989025680000022
y (k) is an actual variable, and y (k) is ωm(k)-2ωm(k-1)+ωm(k-2); k (k) is a gain variable, and the iterative formula of K (k) is:
Figure FDA0002989025680000023
p (k) is an iterative covariance variable, and p (k) is:
Figure FDA0002989025680000024
mu is a forgetting factor, and mu is a forgetting factor,
Figure FDA0002989025680000025
is an identification variable;
according to the identification variable
Figure FDA0002989025680000026
And a recognition period TsCalculating to obtain a final value J of the moment of inertia identificationkThe calculation formula is as follows:
Figure FDA0002989025680000027
6. the fast online rotational inertia identification method for the servo system as claimed in any one of claims 1 to 5, wherein the initial identification value of the rotational inertia of the motor is calculated and the final identification value J of the rotational inertia of the motor is calculatedkThe iterative computation of (2) all adopts an integer arithmetic mode.
7. A fast online rotational inertia identification system suitable for a servo system is characterized by comprising an identification initial estimation unit, a sampling and synchronization unit, an iteration updating judgment unit and an identification algorithm calculation unit;
wherein the identificationThe initial estimation unit is used for enabling the motor to start rotating under the condition that the acceleration is not zero, and obtaining the feedback current i of the motor in a preset rotating rangeFirst stageAnd angular velocity ωFirst stageBased on the feedback current iFirst stageAnd angular velocity ωFirst stageCalculating the initial value J of the moment of inertia of the motor0Identifying the initial value J0Outputting to a sampling and synchronization unit;
the sampling and synchronizing unit is used for sampling the feedback current i and the angular velocity omega of the motor in real time under the condition that the rotation range of the motor is larger than the preset rotation range, synchronously processing the feedback current i and the angular velocity omega to obtain the feedback current i and the angular velocity omega which are synchronously processed, and outputting the feedback current i and the angular velocity omega to the iteration updating and judging unit;
the iteration updating and judging unit is used for judging whether the change rate of the feedback current, the change rate of the angular velocity and the absolute value of the angular velocity are respectively greater than a feedback current change rate threshold, an angular velocity change rate threshold and an angular velocity threshold:
if yes, the identification algorithm calculation unit identifies an initial value J according to the rotational inertia of the motor0The current i and the angular velocity omega are fed back, iterative calculation is carried out based on a motor rotational inertia identification algorithm, and an identification final value J of the motor rotational inertia is obtainedk；
If not, the sampling and synchronizing unit continues to sample the feedback current i and the angular speed omega of the motor in real time and performs synchronous processing on the feedback current i and the angular speed omega.
CN202110312137.0A 2021-03-23 2021-03-23 Rapid online rotational inertia identification method and system suitable for servo system Active CN113067514B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110312137.0A CN113067514B (en) 2021-03-23 2021-03-23 Rapid online rotational inertia identification method and system suitable for servo system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110312137.0A CN113067514B (en) 2021-03-23 2021-03-23 Rapid online rotational inertia identification method and system suitable for servo system

Publications (2)

Publication Number Publication Date
CN113067514A true CN113067514A (en) 2021-07-02
CN113067514B CN113067514B (en) 2022-03-18

Family

ID=76561740

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110312137.0A Active CN113067514B (en) 2021-03-23 2021-03-23 Rapid online rotational inertia identification method and system suitable for servo system

Country Status (1)

Country Link
CN (1) CN113067514B (en)

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH06195128A (en) * 1992-12-25 1994-07-15 Mitsubishi Electric Corp Device and method for controlling position of motor
JP2008148402A (en) * 2006-12-07 2008-06-26 Toshiba Mitsubishi-Electric Industrial System Corp Power conversion device
CN103219939A (en) * 2013-03-19 2013-07-24 哈尔滨工业大学 Rotational inertia on-line identification method for alternating current (AC) permanent magnet synchronous motor servo system
CN103338003A (en) * 2013-06-28 2013-10-02 西安交通大学 Online simultaneous identifying method for load torque and inertia of motor
CN109660169A (en) * 2018-11-26 2019-04-19 浙江浙能技术研究院有限公司 A kind of rotary inertia transient state discrimination method of induction machine
CN109861279A (en) * 2019-01-24 2019-06-07 太原理工大学 A kind of rotary inertia self-adaptation control method suitable for virtual synchronous generator
CN110149081A (en) * 2019-05-08 2019-08-20 固高科技(深圳)有限公司 The control method and on-line parameter identification system of permanent magnet synchronous motor
CN111464097A (en) * 2020-04-03 2020-07-28 华南理工大学 Method for identifying rotational inertia of permanent magnet synchronous motor offline
CN112332738A (en) * 2020-11-09 2021-02-05 上海英恒电子有限公司 Rotational inertia determination method, system, computer device and storage medium

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH06195128A (en) * 1992-12-25 1994-07-15 Mitsubishi Electric Corp Device and method for controlling position of motor
JP2008148402A (en) * 2006-12-07 2008-06-26 Toshiba Mitsubishi-Electric Industrial System Corp Power conversion device
CN103219939A (en) * 2013-03-19 2013-07-24 哈尔滨工业大学 Rotational inertia on-line identification method for alternating current (AC) permanent magnet synchronous motor servo system
CN103338003A (en) * 2013-06-28 2013-10-02 西安交通大学 Online simultaneous identifying method for load torque and inertia of motor
CN109660169A (en) * 2018-11-26 2019-04-19 浙江浙能技术研究院有限公司 A kind of rotary inertia transient state discrimination method of induction machine
CN109861279A (en) * 2019-01-24 2019-06-07 太原理工大学 A kind of rotary inertia self-adaptation control method suitable for virtual synchronous generator
CN110149081A (en) * 2019-05-08 2019-08-20 固高科技(深圳)有限公司 The control method and on-line parameter identification system of permanent magnet synchronous motor
CN111464097A (en) * 2020-04-03 2020-07-28 华南理工大学 Method for identifying rotational inertia of permanent magnet synchronous motor offline
CN112332738A (en) * 2020-11-09 2021-02-05 上海英恒电子有限公司 Rotational inertia determination method, system, computer device and storage medium

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
YIFAN SHEN ET AL.: "Application of an Improved ADRC Controller Based on the Double Closed Loop Dynamic Disturbance Compensation in PMSM", 《2018 21ST INTERNATIONAL CONFERENCE ON ELECTRICAL MACHINES AND SYSTEMS (ICEMS)》 *
林巨广 等: "基于改进RLS算法的永磁同步电机参数辨识", 《合肥工业大学学报（自然科学版）》 *

Also Published As

Publication number Publication date
CN113067514B (en) 2022-03-18

Similar Documents

Publication Publication Date Title
Su et al. A simple improved velocity estimation for low-speed regions based on position measurements only
KR101087581B1 (en) Sensorless control method of permanent magnet synchronous motor
CN109586645B (en) Permanent magnet synchronous motor inertia identification method and equipment
CN109787524A (en) A kind of permanent magnet synchronous motor on-line parameter identification method
CN111181458A (en) Surface-mounted permanent magnet synchronous motor rotor flux linkage observation method based on extended Kalman filter
CN112332738B (en) Rotational inertia determination method, system, computer device and storage medium
CN116488514B (en) Sensorless control method and system for permanent magnet synchronous motor based on reduced order EKF
CN111293947B (en) Improved permanent magnet synchronous motor speed sensorless control method
CN110569767B (en) Motion parameter identification method, system, equipment and computer readable storage medium
CN116247996A (en) Sensorless control method and related device for permanent magnet synchronous linear motor
CN112910328B (en) Permanent magnet synchronous motor acceleration arrangement method based on torque observation compensation
CN114944801A (en) PMSM (permanent magnet synchronous motor) position sensorless control method based on innovation self-adaptive extended Kalman
CN110768600A (en) PMSM speed sensorless rotor detection method
CN113067514B (en) Rapid online rotational inertia identification method and system suitable for servo system
CN108521246B (en) Method and device for predictive control of current of single current sensor of permanent magnet synchronous motor
CN112003527B (en) Improved method of iterative extended Kalman filtering for asynchronous motor rotation speed identification
CN113037161A (en) Model reference self-adaptive permanent magnet synchronous motor position-sensorless vector control method based on super-distortion sliding mode algorithm
CN113224991A (en) Method, system, terminal and readable storage medium for identifying inductance of synchronous reluctance motor based on unscented Kalman filtering
CN110765717A (en) FPGA-based extended Kalman filter circuit structure design method
Ciabattoni et al. Adaptive extended Kalman filter for robust sensorless control of PMSM drives
CN114722528A (en) Induction motor sensorless vector control method based on adaptive extended Kalman filtering
Derdiyok Simple method for speed and rotor resistance estimation of induction machines
Chang et al. Improved rotor position detection speed based on optimized BP neural network for permanent magnet synchronous motor
Wang et al. Design of Full-Order State Observer for Motor Drive Systems Based on the Fixed Gain Filter
CN112886889B (en) Modular multi-winding permanent magnet motor system parameter immune prediction control method and system

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant