Low-speed operation control method of permanent magnet direct-drive servo motor
Technical Field
The invention discloses a low-speed operation control method of a permanent magnet direct-drive servo motor, particularly relates to a method capable of inhibiting periodic and aperiodic torque disturbance at low speed even extremely low speed, and belongs to the technical field of power generation, power transformation and power distribution.
Background
Different from the traditional servo system, the permanent magnet direct-drive servo system cancels a gear transmission mechanism such as a speed reducer and directly drives a load by adopting a low-speed large-torque motor, and has a series of advantages of compact structure, high precision, high efficiency, fast dynamic response, high reliability, low noise and the like, so the permanent magnet direct-drive servo system is widely applied to high-performance servo equipment in industry, aerospace and the like.
However, the motor is typically subject to torque disturbances from the system itself (e.g., cogging torque, harmonic torque) and from external loads (e.g., load torque, friction torque), and these disturbances are transmitted directly between the load and the motor shaft due to the absence of a drive or reduction mechanism. Although the influence of some disturbances is negligible at medium and high speeds, the stability and performance of the rotational speed will be severely affected for low and even ultra-low speed operating conditions, thereby reducing the accuracy and performance of the servo system. Therefore, how to reduce the influence of complex and variable torque disturbance on the low-speed or even ultra-low-speed running motor is an important problem faced by the direct-drive servo system.
The conventional speed control strategy of the permanent magnet motor servo system can be briefly divided into three parts: a speed outer loop, a current inner loop and a pulse width modulation module. Among other things, speed outer loop controllers are typically based on simple, reliable proportional-integral regulators, but their interference immunity is not ideal in certain high performance applications. Therefore, various methods are proposed by numerous scholars to suppress torque interference, wherein proportional resonance control, repetitive control and iterative learning control can be used to suppress periodic interference such as cogging torque, magnetic flux harmonics, current measurement errors and phase imbalance, but the methods often require knowledge of specific frequency information of the disturbance. For aperiodic disturbances such as sudden load shedding and random disturbances, some researchers have also studied different disturbance observers including a reduced order observer, an extended state observer, a sliding mode disturbance observer, etc. to suppress by a direct current compensation method, but the accuracy of observation is usually affected by too low speed and inaccurate model, thereby reducing the compensation effect, and inevitably accompanied by an increase in algorithm and implementation complexity. In addition, the scholars can suppress the non-periodic disturbance by a sliding mode control method, but the buffeting problem is inevitably encountered, and the influence of the buffeting on the speed fluctuation is not negligible for the low-speed and even ultra-low-speed operation.
Generally, accurate speed feedback is a fundamental requirement for the above conventional control strategy with a speed outer loop. For medium and high speed, the speed information is usually easy to obtain, and the accuracy of the speed information can be ensured. However, at low or even ultra-low speeds, even if the error of the position measurement is small, the differential operation in the speed calculation causes a large error in the speed information, and if such an inaccurate speed signal is used as feedback, not only the speed stability is affected, but also the torque disturbance resistance is impaired.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a low-speed operation control method of a permanent magnet direct-drive servo motor, which converts speed control into position control according to a derivative relation between a rotating speed and a rotor position. Based on the position-current two-loop control framework, a rotor position tracking controller is designed to ensure smooth tracking of a real-time position track, so that stable speed is obtained. By using the strategy, the suppression effect of the permanent magnet direct drive servo motor on periodic and aperiodic torque disturbance during low-speed operation is improved, accurate speed feedback and specific torque disturbance information are not required, and the realization is simple. The technical problems that feedback control for tracking the rotating speed of the permanent magnet direct drive motor based on the position measurement value in the low-speed state influences speed stability and weakens the torque disturbance resistance are solved.
The invention adopts the following technical scheme for realizing the aim of the invention:
a low-speed operation control method of a permanent magnet direct-drive servo motor comprises the following steps:
A. converting the speed command into a real-time rotor position trajectory by introducing an integrator based on a derivative relationship between rotor speed and position;
B. designing a rotor position tracking controller on the basis of position-current two-loop control, and outputting quadrature axis given current;
C. designing a filter structure in the rotor position tracking controller according to the real-time characteristic of the position track to realize unsteady-state error tracking;
D. and the determined controller structure is subjected to parameter design from the aspects of stability and interference resistance, so that a better position smoothing effect is achieved.
Further, the step a of the low-speed operation control method of the permanent magnet direct drive servo motor specifically comprises the following steps: according to the derivative relation between the rotating speed and the rotor position, constructing a real-time position track as follows: θ ═ ω ═ t, θ ═ is the position command, ω ═ is the given speed, and t is time. The presence of the integrator makes the position trajectory tend to be continuous.
Further, the step B of the low-speed operation control method of the permanent magnet direct drive servo motor is specifically as follows: in order to realize smooth tracking of the real-time position track, the structure of the rotor position tracking controller is designed as
J is moment of inertia, B
aAs damping coefficient, K
TF(s) is a torque current constant, and f(s) is a filter. Thus, the quadrature current specification of the current loop is obtained.
Further, the step C of the low-speed operation control method of the permanent magnet direct drive servo motor is specifically as follows: aiming at the particularity of the real-time position track, in order to ensure no steady-state error tracking, the filter structure is designed as follows:
r is the filter order and λ is the filter time constant.
And finally, the step D of the low-speed operation control method of the permanent magnet direct drive servo motor specifically comprises the following steps: according to control system sensitivity function
The order r is determined to be 2 according to the interference resistance reflected by the amplitude-frequency characteristic; by robust stability conditions:
to determine a lower bound for λ; at a certain order r, by a sensitivity function
And load sensitivity function
To determine the upper bound of lambda and thus obtain a preferred value range of lambda, tau
dDelay time, τ, due to sampling, modulation, or the like
iThe inverse of the current loop bandwidth.
By adopting the technical scheme, the invention has the following beneficial effects:
(1) the low-speed control method disclosed by the invention converts the speed instruction into a real-time position track in the low-speed or even extremely low-speed running process of the motor, avoids the influence of position errors as much as possible, and can further improve the gain of the controller, thereby improving the tracking capability of the system on the speed instruction.
(2) According to the low-speed control method disclosed by the invention, a torque disturbance observation link with a position as an input exists in the rotor position tracking controller, periodic and aperiodic disturbances can be observed at the same time, the output of the link is introduced to a quadrature axis current given end for current compensation, so that the suppression of complex torque disturbance is realized, and the periodic and aperiodic torque disturbances are suppressed at the same time at low speed even very low speed.
(3) The low-speed control method disclosed by the invention has the advantages that the step speed instruction is converted into a continuous position instruction, and the overshoot of the speed given to step change can be weakened to a certain extent; uncertain factors of the model are considered in parameter design, and the method has strong robustness on system parameter change.
(4) The low-speed operation control method provided by the invention is simple and convenient to realize, is particularly suitable for a digital control system, can reduce the precision requirement on the position sensor to a certain extent, and saves the cost.
Drawings
Fig. 1 is a block diagram of rotor position tracking control.
FIG. 2 is a diagram of a speed command versus a real-time position trajectory.
FIG. 3 is a control block diagram of a position trajectory control loop.
Fig. 4 is a graph showing the amplitude-frequency characteristics when the parameter r takes different values.
Fig. 5 is a graph of amplitude-frequency characteristics for determining the lower bound of λ.
Fig. 6(a) and 6(b) are amplitude-frequency characteristics and gain-time characteristics for determining the upper limit of λ, respectively.
Detailed Description
The technical scheme of the invention is explained in detail in the following with reference to the attached drawings.
Usually because of the acquisitionThe precision is higher, and the position error is very little. However, as can be seen from equation (1), the error will be amplified by approximately 1/T during the velocity acquisition processcMultiple, which means that if the speed loop control frequency is 2kHz, the error will be up to 2000 times.
In the formula (1), ω iserr、θerrRespectively a velocity error and a position error. T iscIs the speed loop control period.
That is, since the derivative calculation is difficult to avoid the above-mentioned large speed acquisition error, in the case of medium and high speeds, the influence on the overall control is not large because the error is small compared to the speed command, but it is not negligible at low speeds and even ultra-low speeds. If the speed signal is used as an outer loop feedback at low speed, not only is speed stability difficult to ensure, but also torque disturbance resistance is greatly reduced, and although this problem can be alleviated by improving position measurement accuracy as much as possible, the effect is limited and the cost is greatly increased.
From the derivative relationship in equation (1), the smoothness of the rotor position determines the speed stability, and the position error is much smaller than the speed error. Therefore, smooth tracking of the rotor position trajectory enables the permanent magnet motor to achieve better speed control performance at low speeds. Based on the above idea, the present invention proposes a novel low speed control strategy, called rotor position tracking control strategy, as shown in fig. 1.
The position track generation part in fig. 1 means that a speed command passes through an integrator to generate a real-time position track, the corresponding relation between the speed command and an actual position track is shown in fig. 2, and at this time, the step set of the speed is eliminated, and then the speed is converted into a continuous position track, so that the speed overshoot can be reduced to a certain extent. Because the position track is directly controlled, the influence of the speed error brought by the formula (1) is avoided as much as possible, and under the same position precision, better speed performance can be obtained compared with the traditional speed control method. In other words, the present invention requires less accuracy of the position sensor at the same speed accuracy output.
To ensure the tracking performance of the rotor position, the rotor position tracking controller of fig. 1 is configured as G in the position trajectory control loop shown in fig. 3
RPTC,G
RPTCThe expression of(s) is formula (2). Wherein G is
m(s) is a nominal model of the actual object, i.e. G
m(s)=K
T/[s(Js+B
a)];
Under the structure of the formula (2), in order to realize the real-time position track without steady-state error tracking, the filter f(s) is designed as follows:
at this time, the final transfer function of the rotor position tracking controller can be obtained by substituting equation (3) into equation (2). And the position track control loop shown in fig. 3 can obtain an equation (4), namely, the steady-state error is theoretically avoided, and the smoothness of position tracking is improved.
After the transfer function of the rotor position tracking controller is determined, only the parameters r and λ in equation (3) are unknown. The parameter r may be determined by the sensitivity function SRPTC(s) to determine the values, the expression is as follows:
the amplitude-frequency characteristic curve of equation (5) is shown in fig. 4. As r increases, the noise immunity decreases, plus r cannot be 1 (f(s) is a constant value of 1 when r is equal to 1), so the optimal value r is selected to be 2 in the present invention.
The lower bound of the parameter lambda is determined by the equation (6) and the amplitude-frequency characteristic curve shown in fig. 5, and is used for judging the parameter value range in which the controller can ensure stability when the system has uncertain factors. That is, the value of the parameter λ cannot be too small, so that F (j ω) is setf) The ordinate values at all frequency points are at 1/| | Δ (j ω)f)||∞The following. Wherein, ω isfAt angular frequency, Δ (j ω)f) Is the uncertainty factor of the model. The lower bound is thus defined as λ 0.2Tc。
The upper bound of the parameter λ is determined by equations (7) and (8), the amplitude-frequency characteristic curve shown in fig. 6(a), and the gain-time characteristic shown in fig. 6(b), and the noise immunity decreases as the value of λ increases. Therefore, in order to ensure sufficient torque disturbance suppression capability, the upper limit is set to 2Tc. Meanwhile, compared with the traditional Proportional Integral (PI) speed control strategy, the method has better disturbance suppression capability in amplitude-frequency characteristics and gain-time characteristics.
According to the method, a complete rotor position tracking control strategy can be realized.