ES2903009B2 - VIBRATION CORRECTION PROCEDURE FOR ROTARY ELECTRIC MOTORS WITH FIELD-ORIENTED CONTROL - Google Patents

Description

DESCRIPTION

VIBRATION CORRECTION PROCEDURE FOR ENGINES

ROTARY ELECTRIC WITH FIELD ORIENTED CONTROL

TECHNICAL SECTOR

The present invention finds application within the area of mechanics, in the field of rotary motors and the control of undesired effects in them, in particular, of vibrations.

BACKGROUND OF THE INVENTION

In applications that use motors, vibrations may occur due to causes inherent to the system in which the motor is coupled (such as, for example, incorrect shaft alignment) or those due to the conditions of use of said system (fluctuations in the load that must be move the engine). These vibrations spread throughout the structure of the system and can cause discomfort to people and defects in the machinery, sometimes rendering it useless.

There are various studies and patents that address the problem of detection and correction of vibrations that occur in systems coupled to rotating electrical machines. The control techniques that allow their correction are called active vibration control. In AJ Fleming, SOR Moheimani, and S. Member, "Inertial Vibration Control Using a Shunted Electromagnetic Transducer," IEEE/ASME Trans. Mechatronics, vol. 11, no. 1, pp. 84-92, 2006, discusses the feasibility of incorporating electromagnets for structural vibration cancellation. This system has the advantage of not requiring additional sensors. On the other hand, in AG Wills, D. Bates, AJ Fleming, B. Ninness, and SOR Moheimani, "Model predictive control applied to constraint handling in active noise and vibration control,” IEEE Trans. Control Syst. Technol., vol. 16, no. 1, p. 3-12, 2008, the operating limits of predictive control techniques in this field are explored, which allow the inclusion of said limits in the control action, achieving better results.

In the case of automobiles, there is great interest in correcting vibrations, since these directly influence the driving experience. Reference W. Sun, H. Gao, and B. Yao, "Adaptive robust vibration control of full-car active suspensions with electrohydraulic actuators," IEEE Trans. Control Syst. Technol., vol. 21, no. 6, pp. 2417 2422, 2013, develops a complete model of the vehicle where the suspension is acted upon, implementing controllers that incorporate the forces of the actuators as virtual inputs, resulting in robust controllers in the face of parameter uncertainty In C. Gohrle, A. Schindler , A. Wagner, and O. Sawodny, “Design and vehicle implementation of preview active suspension controllers,” IEEE Trans. Control Syst. Technol., vol. 22, no. 3, p. 1135-1142, 2014, the vehicle suspension is also acted on, but this time data on the road from sensors already present in the vehicle are incorporated into the model, in order to anticipate to improve the response. CN109941120 describes a control capable of attenuating the torsional vibrations generated by changes in the vehicle's engine torque reference, which is also used to generate a signal that cancels the vibrations.

The vibrations produced by rotating electrical machines are not exclusive to electric vehicles, they are also found in industrial and even domestic environments. Clear examples of this are the patents ES2354982T3 and ES2240260T5, which deal with the detection and damping of vibrations produced in washing units by the high speeds of rotation of the machines in the presence of unbalanced loads. In W. He, X. He, and C. Sun, "Vibration Control of an Industrial Moving Strip in the Presence of Input Deadzone," IEEE Trans. Ind. Electron., vol. 64, no. 6, pp. 4680- 4689, 2017, the effect of vibrations and their active attenuation in an industrial application such as the control of the movement of mobile sheets in metal, paper or plastic lamination processes is studied.

Application WO2020067482 describes an active vibration attenuation system, which has the same objective as the present invention. In said request, the information about the vibrations to be eliminated is taken from the counter electromotive force of the rotating machine and is transformed into an additional torque command responsible for the elimination of said vibrations. The main problem that this form of vibration elimination entails is the need to introduce new sensors to measure the voltage at the machine terminals, necessary for the calculation of the counter electromotive force, which otherwise would not be present in a control oriented to field as in which it is implemented.

In addition, according to the form of vibration cancellation described in the previous patent, the modification of the initially calculated corrections for the complete elimination of vibrations is not contemplated.

SUMMARY OF THE INVENTION

The object of the present invention is to provide a method for correcting vibrations in an electric motor by using transducers already present in said motor. To do this, the invention makes use of the position measurements of the motor rotor (or speed, since both are closely related) to obtain the information on the vibrations to cancel. Such measurements are always present in traditional field-oriented control, which is the most widely used control in industry for rotating electrical machines. Subsequently, by means of an iterative process, the invention determines the necessary modifications to the torque setpoint for the effective cancellation of the vibrations.

Optionally, what is described in the present invention could be applied in the event that other measures were used to obtain the information on the vibrations to be corrected, such as, for example, the motor currents. This would make it extrapolated to other types of existing controls.

BRIEF DESCRIPTION OF THE FIGURES

In order to help a better understanding of the characteristics of the invention and to complement this description, the following figures are attached as an integral part of the same, whose nature is illustrative and not limiting:

Figure 1 shows the general scheme of a system in which the method of the invention can be applied.

Figure 2 shows the operating process of the method of the invention.

Figures 3 and 4 show an example of application of the invention to a system in which the speed of rotation of a motor presents high-frequency variations and, therefore, vibrations. Figure 3 shows the corrections that are calculated at each moment and figure 4 shows how these variations are reduced accordingly.

Figure 5 shows the detail of the control and how it is implemented within the general control of the system.

DETAILED DESCRIPTION

Referring to Figure 1, a signal acquisition system records the necessary variables for the characterization of the vibrations to be corrected, that is, the position or speed of the motor. This signal acquisition system comprises current and speed transducers that are already present in the device, for example, hall effect sensors 2 and a resolver or encoder 3 coupled to the shaft of the electric motor 1. Optionally, the vibrations can be measured with a piezoelectric transducer or accelerometer 4 also present in the system.

The measurements are passed through a signal filtering stage 5, to eliminate ambient and switching noise. For this, low-pass filters and voltage buffers are used to decouple the resistance of the filter from that of the analog-digital converter of the control platform 6. In said control platform 6, the analog signals are transformed into digital signals. Once the signals have been transformed by the control platform, they are available for use in the system's digital controller.

The classification systems 7 and attenuation 8 of figure 1 can be seen once again included in the field-oriented control described in figure 5, in addition to the previously obtained digital signals ( u ^m , i ^{a and} i ^b ). The characterization of the vibrations can be carried out by means of a fast Fourier transform (FFT) of the position or speed of the motor 1, using a variable time window or by means of an asynchronous reference system (ASRF). The vibrations observed by means of the velocity or vibration transducer are given by the following expression in the time domain, since they are periodic vibrations:

Where v ^h is the vibration of order h, which will be a multiple of the motor rotation frequency f ^m , A ^h is the amplitude of the vibration, t is the variable that represents the time and is the phase of said vibration. These parameters are the ones that will provide us with the characterization of the vibrations and, once the vibration to be corrected has been selected, these initial values are They will be saved with a subscript 0 to later compare with those obtained when corrections are introduced (figure 2, ref. 20).

The sampling frequency is closely related to the frequency of the vibration to be corrected. To ensure a correct reconstruction of the vibration wave from the sampled values, at least 8 points per wave period are necessary. In this way, if the turning frequency of the machine is 5 hertz, and it is desired to correct a 5th order vibration, which would correspond to a vibration with a frequency of 25 hertz, it is necessary to sample at 200 hertz for the system to be able to identify it correctly.

Using the initial extracted data, the necessary corrections to the current are determined and introduced at the observed frequency, with preliminary amplitude and phase values 30, based on the observed vibration. In this sense, the phase of the correction will be equal to that of the vibration and the amplitude will be half the amplitude of the vibration, so the initial correction is defined by the following expression:

^carr iTii = 1 s in ( '1 ^' ^ ^' f ,n ^' * ⁺ <Ph)

Where icor_ini is the instantaneous initial correction current and I is the amplitude of said correction, which is given by the expression I = Ah¡2. The rest of the parameters are those extracted in the vibration analysis. From this point, increments are first introduced in phase 5<¡, modifying it in one direction (by adding or subtracting it) 40.

To check if the vibration 50 has been reduced, an analysis of the signal is carried out again and it is compared with the initial stored value of step 20. If it has been reduced, the phase is modified again in the same direction and it is carried out in the same way. again the check, but this time comparing with the previous value instead of the initial one. Otherwise, it proceeds to modify it in the opposite direction 60, checking again if the observed vibration 70 decreases.

Upon reaching the minimum point of the vibration, the algorithm will keep oscillating around it. At this point, it is possible to refine the correction modification until an optional convergence criterion is reached. An example would be reducing the amplitude of the phase increment to 20% of its original value and going back to step 40 again.

Next, the amplitude of the correction 80 is modified, proceeding analogously to the previous case, that is, adding amplitude increments 5 ^to . The amplitude of the correction is increased and it is checked if the amplitude of the vibration decreases by 90. If it does not decrease, the amplitude is reduced by 100 until the minimum is found and optionally pass to a convergence criterion 110, such as defined in the phase modification step. At that moment, the complete attenuation of the selected harmonic component of vibration will have been reached. In general, the correction introduced to the main control system in iteration n of the algorithm is given by the expression:

Where the subscript n refers to iteration n of the procedure. Thus, in amplitude correction, In-1 is the amplitude value of the previous iteration, while in phase correction, $n-1 is the phase value of the previous iteration. On the other hand, k is a variable that takes the value 1 or -1 depending on whether we are effectively correcting the vibration or not. Thus, if the amplitude of the vibration in the previous iteration, Ah_n- ¹ , is greater than the amplitude in the current iteration, Ah_n, k takes the value 1, being -1 otherwise. The rest of the parameters are those obtained from the initial vibration analysis and the values of the phase and amplitude increments.

The corrections calculated with the algorithms of figure 2 are introduced in the field oriented control (FOC, for its acronym in English), which is encompassed in the control platform 6 of figure 1 and detailed in figure 5. Said control platform calculates the necessary signals for the power stage 10, which feeds the electric motor. The corrections are introduced into the quadrature current. In this way you have

^ q j ' e f ^{^} q_PI ^{3 "^} q_corr

In other words, the reference current in quadrature will be the sum of the current of the proportional integral (PI) controller iq_pi, in charge of maintaining the speed equal to the reference speed, plus the correction current iq_ corr.

For this example, consider the system in Figure 1. The motor is rotated at a rate of 300 revolutions per minute, equivalent to 31.41 radians per second. A load with an average value of 4 Nm has been introduced into the system with a high-frequency variable component, which introduces vibrations into the system. These vibrations can be observed in the speed variations shown in Figure 4. The operation of the invention is detailed below:

1. The vibration classification system 7, using the speed measurement from the position encoder/resolver sensor, detects by means of a fast Fourier transform (FFT analysis) a 60 hertz component with amplitude equal to 2. To start the correction of said component, introduces a correction in the quadrature current of amplitude 1 and initial phase 0, as shown in Figure 3 for the moment of time 0.23 seconds.

2. The attenuation system 8 begins by modifying the phase of the initial correction at the instant 0.65 seconds, introducing a phase shift equal to n/4 radians. It can be seen in figure 4 that the velocity component at 60 hertz decreases in amplitude, so the following disturbances will continue to introduce phase shift in the same direction.

3. When the phase shift changes from n radians to 5n/4 radians in 1.64 seconds, an increase in the amplitude of the speed variation is observed, so the system changes direction and stays with the phase shift value of the last step.

4. At this moment, the attenuation system begins to modify the amplitude of the correction introduced in the quadrature current, increasing in each step a value of 0.5 A of current.

5. The variation in speed decreases with each increase in the amplitude of the correction, as shown in figures 3 and 4, until at the instant of time 2.715 seconds, when the value of said amplitude increases to 3.5 A, the variation in speed increases instead of decreasing. The algorithm reverses the correction modification direction, and keeps the previous amplitude value equal to 3 A.

6. At this time, the classification system checks that the amplitude of the speed variation has reduced its value until it almost completely disappears and considers the correction of this component to be complete. Subsequently, if additional components to the 60 hertz described were detected in the vibration characterization process, correction of said components could be carried out. The process in this case would be repeated for that component.

In view of this description and figures, the person skilled in the art will be able to understand that the invention has been described according to some preferred embodiments thereof, but that multiple variations can be introduced in said preferred embodiments, without exceeding the object of the invention as claimed.

Claims (2)

1. Procedure for correcting vibrations in rotary electric motors with field-oriented control, characterized in that it comprises the following steps: -a. characterization of the vibrations by means of a fast Fourier transform (FFT) of the position or speed of the motor, where the vibrations are expressed by the following expression:

where vh is the vibration of order h, which will be a multiple of the motor rotational frequency fm, Ah is the amplitude of the vibration, t is the variable that represents time and $h is the phase of said vibration; -b. storage, with subscript 0, of the amplitude and phase of each vibration in each order; -c. selection of the order to be corrected and carrying out an initial correction on the current that feeds the motor defined by the expression:

/ (íl - í ^ J _j where ¡corjm is the instantaneous initial correction current and I is the amplitude of said correction, which is given by the expression

entering the field-oriented control quadrature current; -d. introduction of increments in phase 5<¡, of the current modifying it in one direction; -and. check that step d. you have reduced the values with subscript 0 saved in step b; -F. otherwise, enter the increments in the opposite direction, -g. repetition of steps d, e and f, comparing with the value of the previous iteration, until a convergence criterion is reached; -h. introduction of increments in the amplitude Ah of the correction current modifying it in one direction; -Yo. check that step h. has reduced the values by comparing them with those stored in the previous step; -j. otherwise, input increments in the opposite direction, -k. repetition of steps h, i and j until reaching a convergence criterion; -l. Repeat step c for the rest of the orders to be corrected. 2. Method according to claim 1, characterized in that the characterization of the vibrations is carried out using a variable time window or by means of an asynchronous reference system (ASRF).