METHOD FOR DETERMINING CLOSED-CONTROL PARAMETERS FOR A HYDRAULIC SYSTEM

Abstract

In order to carry out largely automated parameterisation of the closed-loop control parameters for closed-loop control of a hydraulic system comprising a servo drive, a method and a device for determining the closed-loop parameters of a closed-loop control unit of the hydraulic system are specified, wherein an actual system pressure of a hydraulic consumer of the hydraulic system is closed-loop controlled by means of a predefined set point rotational speed of a servo drive, wherein an actual rotational speed of the servo drive follows the predefined set point rotational speed, wherein an excitation signal is applied to the setpoint rotational speed, and the actual system pressure which is set here is measured, the dynamics of the hydraulic system are acquired from the actual rotational speed and/or the setpoint rotational speed and the actual system pressure, and the closed-loop control parameters are calculated from the acquired dynamics.

Claims

1. A method for determining control parameters (KP, Tt, TD, TF) of a control unit (6) of a hydraulic system (1), characterized in that an actual system pressure (p_ist) of a hydraulic load (8) of the hydraulic system (1) being controlled by a predetermined target speed (n_soll) of a servo drive (3) comprising a motor (31) and a pump (32), wherein an actual speed (n_ist) of the servo drive (3) follows the predetermined target speed (n_soll), wherein an excitation signal (n1) is applied to the target speed (n_soll) and the resulting actual system pressure (p_ist) is measured, in that the dynamics of the hydraulic system (1) is determined from the actual speed (n_ist) and/or the target speed (n_soll) and the actual system pressure (p_ist), and in that the control parameters (KP, Tt, TD, TF) are calculated from the determined dynamics.

2. The method according to claim 1, characterized in that a constant initial speed (n0) is predetermined for the target speed (n_soll) before the application of the excitation signal (n1), at which point an initial pressure (p0) results in the hydraulic system (1) as the actual system pressure (p_ist).

3. The method according to claim 3, characterized in that the actual system pressure (p_ist) represents a pressure of a hydraulic cylinder or a hydraulic motor.

4. The method according to claim 1, characterized in that a square wave signal, a harmonic signal, preferably with increasing frequency, pulses, or a mixed signal is used as the excitation signal (n1).

5. The method according to claim 1, characterized in that the dynamics of the hydraulic system (1) is described by a transfer function (Gp/n(z)) of the hydraulic system (1).

6. The method according to claim 5, characterized in that the transfer function (Gp/n(z)) is determined by a Fast Fourier Transformation (FFT).

7. The method according to claim 5, characterized in that the parameters of the transfer function (Gp/n(z)) are approximated by a method of least squares.

8. The method according to claim 1, characterized in that the control parameters (KP, Tt, TD, TF) are calculated by a frequency characteristic method.

9. The method according to claim 8, characterized in that default parameters of the step response, preferably rise time and overshoot, are predetermined for the frequency characteristic method, which preferably comprise an overshoot (ue) and a rise time (tr) of the step response.

10. The method according to claim 1, characterized in that, after determining the control parameters (KP, Tt, TD, TF), the control behavior of the control unit (4) is verified by step-shaped changes in the system pressure (p_soll) being predetermined and a step response being determined.

11. The method according to 10, characterized in that a user carries out the verification of the control behavior of the control unit (4).

12. The method according to claim 1, characterized in that a compensation filter for the control unit (6) is additionally parameterized from the determined dynamics.

13. A hydraulic system (1) comprising a servo drive (3) with a motor (31) and a pump (32), wherein the servo drive (3) has a target speed (n_soll) predetermined by a control unit (5), wherein an actual speed (n_ist) of the servo drive (3) follows the predetermined target speed (n_soll), and comprising a hydraulic load (8) which has an actual system pressure (p_ist) which is controlled by the predetermined target speed (n_soll), wherein a control unit (4) is provided which is configured to apply an excitation signal (n1) to the target speed (n_soll) and to measure the resulting actual system pressure (p_ist) of the hydraulic system (1), to determine the dynamics of the hydraulic system (1) from the actual speed (n) and/or the target speed (n_soll) and the actual system pressure (p_ist), and to calculate the control parameters (KP, Tt, TD, TF) of the control unit (4) from the determined dynamics.

14. The hydraulic system according to claim 13, characterized in that the control unit (4) is integrated in the servo drive (3).

15. The hydraulic system according to claim 13, characterized in that the control unit (4) is integrated in a operating unit (5) which is superordinate to the servo drive (3).

Description

[0024] In the following, the present invention will be explained in more detail with reference to FIG. 1 to 5, which show exemplary advantageous embodiments of the invention in a schematic and non-limiting manner. In the drawings:

[0025] FIG. 1 shows a hydraulic system comprising a general hydraulic load,

[0026] FIG. 2a shows an overshoot and undershoot of the actual system pressure,

[0027] FIG. 2b shows an exemplary desired curve for the actual system pressure,

[0028] FIG. 3 shows a hydraulic system comprising a differential hydraulic cylinder as a hydraulic load,

[0029] FIG. 4a shows a progression over time of a target speed,

[0030] FIG. 4b shows a progression over time of an actual system pressure,

[0031] FIG. 5 shows a progression over time of an actual system pressure in response to a step-like progression over time of a target system pressure.

[0032] FIG. 1 shows an exemplary hydraulic system 1. A hydraulic load 8 has an actual system pressure p_ist, which is regulated by an actual speed n_ist of a servo drive 3. The servo drive 3 is composed of a motor 31 and a pump 32. FIG. 1 shows an open hydraulic circuit, i.e. that the pump 32 conveys a medium 11, for example hydraulic fluid, from a (mostly pressureless) tank 10 and the medium 11 passes from the hydraulic load 8 back into the tank 10. As a result, the actual system pressure p_ist in the hydraulic load 8 is influenced.

[0033] The target system pressure p_soll is predetermined for the control unit 5, for example a PLC, by a user or a program. The target speed n_soll is controlled from the actual system pressure p_ist and the target system pressure p_soll by a pressure regulator, which is implemented here on the control unit 5. For this purpose, the actual system pressure p_ist is likewise supplied to the control unit 5. Furthermore, the actual speed of the servo drive 3 is controlled by a speed regulator, which receives the target speed n_soll from the pressure regulator, i.e. from the operating unit 5 in this case. The speed regulator, however, is implemented on the control unit 6 in this case. The control unit 6 can be represented by an ACOPOS servo amplifier, for example, which provides the servo drive 3 with the necessary current required for following the predetermined target speed n_soll. Both the pressure regulator and the speed regulator can be implemented at any point in the hydraulic system, as already mentioned. For example, it is also possible for the pressure to be regulated on the control unit and only the speed to be regulated on the servo drive 3. The servo drive 3 conveys the medium 11 depending on the actual speed p_ist and thus regulates the actual system pressure p_ist in the hydraulic load 8. Until now, according to the prior art, the control parameters of the control unit 6 have been set manually and iteratively. By way of example, a possible manual setting of the control parameters of a control unit 6 designed as a PID controller shall be described, in which a gain factor K.sub.P, an integration time constant T.sub.I and a differentiation time constant T.sub.D are determined as control parameters of the PID controller in a known manner. In this case, at the start a pressure filter is suitably parameterized, i.e. a filter time T.sub.f is determined by the measured actual system pressure p_ist being filtered. Signals delivered by pressure sensors usually show superimposed noise, resulting e.g. from interference from the environment. This noise causes unpleasant noises when controlling the actual pressure, is transmitted directly to the speed regulator via the pressure regulator and thus has an effect on the actual speed. Therefore, the noise is not only audible, but also has other negative influences, for example on the service life of components. The pressure filter normally has the characteristics of a first order low-pass filter. If the cutoff frequency of this pressure filter is selected to be too low, high attenuation is achieved even at low frequencies, whereby information is lost in the event of rapid changes in pressure and the regulation becomes slower. If the filter time T.sub.f is selected to be too low, disturbances of the actual system pressure p_ist are not suppressed enough, which causes the servo drive 3 not to run smoothly. Too high a filter time T.sub.f, on the other hand, causes slower regulation of the actual system pressure p_ist. The determination of an initial value of the gain factor K.sub.P is further taken as a starting point. For this purpose, a desired target system pressure p_soll is set and the hydraulic load 8 is brought into an initial state, for example by a piston of the hydraulic load 8 being brought into a striking position. The gain factor K.sub.P is then gradually increased until an oscillation, for example a sinusoidal oscillation around the target system pressure p_soll, results for the actual system pressure p_ist.

[0034] This oscillation is e.g. recognizable by louder engine noise of the servomotor 3 than previously, since an oscillation of the actual system pressure p_ist is also noticeable in the actual speed n_ist. An oscillation of the actual system pressure p_ist is primarily visible in the actual pressure signal, which is measured by a sensor. Since the actual system pressure is directly related to the actual speed via the system dynamics, this will then also oscillate. The gain factor K.sub.P is then reduced by 20%. If, furthermore, an oscillation of the actual system pressure p_ist is noticeable, then the gain factor K.sub.P is reduced gradually until no oscillation is noticeable. An initial value of the integration time constant T.sub.I of the pressure regulator is then adjusted. To do this, a desired target system pressure p_soll is set, once again the hydraulic load 8 is brought into an initial position and jumps in the target system pressure are applied. The integration time constant T.sub.I is then gradually reduced until an undershoot of the actual system pressure p_ist, shown by way of example in FIG. 2a, results and then the integration time constant T.sub.I is increased by 10%. If an undershoot already results with the initial value, then integration time constant T.sub.I is gradually increased until the undershoot at the actual speed n_ist disappears. A suitable starting value for the integration time constant T.sub.I would be T.sub.I=0.1 sec. for example. For the differentiation time constant T.sub.D, too, an initial value is set and then the differentiation time constant T.sub.D is gradually increased until barely any overshoot of the actual system pressure p_ist can be identified. If the overshoot of the actual system pressure p_ist is already too heavily suppressed with the initial value of the differentiation time constant T.sub.D, the differentiation time constant T.sub.D is gradually reduced until the overshoot is barely identifiable, as shown by way of example in FIG. 2b. If oscillations occur when increasing the differentiation time constant T.sub.D, these can also be attenuated by adjusting a filter time constant of the differentiator T.sub.1. Thus, a first determination of the control parameters takes place. During operation, however, readjustments must be made frequently, e.g. in accordance with the pattern just mentioned. Apart from the large amount of time required for setting the control parameters, a disadvantage is that, despite detailed documentation, the specific parameterization of the control parameters depends heavily on the user.

[0035] FIG. 3 shows a specific embodiment of a hydraulic system 1. In this figure, the hydraulic load 8 is designed as a differential hydraulic cylinder 7, which is connected to a switching valve 2. The switching valve 2 is further connected both to a servo drive 3 (consisting of a motor 31 and a pump 32) and to a tank 10. As already mentioned with regard to the general embodiment from FIG. 1, an open loop of the hydraulic system 1 can also be seen in FIG. 3. In an open loop, a medium 11 is conveyed from a tank 4 (which usually has atmospheric pressure) by means of the pump 32 of the servo drive 4, at which point the medium 11 is returned into the tank 4 from the hydraulic load 8. However, the present method is not limited only to hydraulic systems 1 having open hydraulic circuits, but is also applicable to hydraulic systems 1 having closed hydraulic circuits, i.e. closed tanks 11 which constitute pressure accumulators. The differential hydraulic cylinder 7 includes a piston 70 of which the position 70 in the differential hydraulic cylinder 7 is dependent on the actual system pressure p_ist. The actual system pressure p_ist of the differential hydraulic cylinder 7 is regulated by the actual speed n_ist of the servo drive 3. The target speed n_soll is in turn predetermined for the servo drive 3 by a control unit 5. The control unit 6 in turn regulates the actual speed n_ist to the target speed n_soll provided by the control unit 5. Using the example of the differential hydraulic cylinder 7, the precise regulation of the actual system pressure p_ist for example allows a defined force to be applied by means of the piston rod 72 connected to the piston 70. As an alternative to a differential hydraulic cylinder 7, a hydraulic motor could for example also serve as a hydraulic load 8, which provides a certain moment. The control unit 6 of the hydraulic load 8, in this case this servohydraulic drive train, can be mounted both centrally, i.e. directly on the servo drive 3, as shown in FIG. 3, as well as decentrally in the operating unit 5.

[0036] By means of the operating unit 5, the actual system pressure p_ist is detected in the differential hydraulic cylinder 7 and the target speed n_soll for the servo drive 4 is calculated by means of the control unit 6, in order to set a desired target system pressure p_soll in the differential hydraulic cylinder 7. Of course, the actual system pressure p_ist could also be detected and processed directly using the control unit 6.

[0037] At the start of the method according to the invention for determining the control parameters, in FIG. 3 the piston 70 of the differential hydraulic cylinder 7 is moved into a desired position 70, e.g. an end position, by applying a constant initial speed n0 to the target speed n_soll. Therefore, in the differential hydraulic cylinder 7 of the hydraulic system 1, the actual system pressure p_ist reaches an initial pressure p0, which effectively sets an operating point. It should again be noted that instead of the differential hydraulic cylinder 7 described here, any other hydraulic load is of course conceivable. The hydraulic system 1 is excited by applying an excitation signal n1 to the target speed n_soll. A possible progression over time of the actual speed n_ist after applying the target speed with an excitation signal n1 is shown in FIG. 4a, in which the time t is plotted on the horizontal axis in seconds and the actual speed n_ist is plotted on the vertical axis. The excitation signal is intended to stimulate the dynamics of the control loop, i.e. of the hydraulic system 1. During this excitation, the actual system pressure p_ist is measured in the differential hydraulic cylinder. A possible resulting progression over time of the actual system pressure p_ist is shown in FIG. 4b, in which the time t is plotted on the horizontal axis in seconds and the actual system pressure n_ist is plotted on the vertical axis. From the progression over time of the actual system pressure p_ist and the actual speed n_ist or the target speed n_soll, which can also be measured or provided by the servo drive 3, a transfer function Gp/n(z) for describing the system dynamics is estimated (in the present case, by way of example, a time-discrete (z-range) transfer function, e.g. of the fifth order. Of course, time-continuous (s-range) transfer functions can also be used), for example using the method of least squares. For example, the underlying transfer function Gp/n(z) is:

[00001]$G_{\frac{p}{n}}\left(z\right)=\frac{2.399*{10}^{-4}.Math.z^3-6.894*{10}^{-5}.Math.z^2-1.079*{10}^{-4}.Math.z-5.007*{10}^{-6}}{z^5-2.647.Math..Math.z^4+2.197.Math..Math.z^3-0.3359.Math..Math.z^2-0.3038.Math..Math.z+0.08941}$

[0038] Now, for example, the structure of the transfer function Gp/n(z) is predetermined, in this case of the fifth order, but the parameters of the transfer function Gp/n(z) are unknown. The optimization problem is modeled with the error squares (deviation between the measured values and the values calculated from the parameters of the transfer function Gp/n(z)) as the objective function (error is minimized) in order to determine the parameters of the transfer function Gp/n(z). The unknown parameters are estimated e.g. by means of the method of least squares, which is adequately described in the literature. An abort criterion is not needed here because it is not an iterative method.

[0039] The dynamics of the system is thus described by the input-output behavior of the system. The dynamics of the pressure regulator describes the behavior of the target system pressure to the actual system pressure.

[0040] Based on this transfer function Gp/n(z), a Pl controller having a gain factor Kp and an integration time constant T.sub.I is parameterized as a control parameter, using a known frequency characteristic method, with no further filtering being carried out. Requirements for the closed loop are thus translated to requirements in the frequency range and the control parameters are calculated there. Specifying these conditions to the closed loop is clearer to the user than directly specifying the control parameters. In the above-mentioned embodiment, the rise time t, was thus set as a requirement to the closed loop. This means that after applying a jump in the target system pressure, the actual system pressure must reach the level change value of the target system pressure within the rise time t, e.g. 0.05 sec. As an additional requirement, an overshoot ue=0% was specified. As is standard in the context of the frequency characteristic method, the requirements for the closed loop are translated into requirements for the open loop (without regulation) by means of approximate relationships. In the frequency response of the open loop, the control parameters of gain factor K.sub.P and integration time constant T.sub.I are calculated and result, for example, in

[00002]$\mathrm{Kp}-42.42.Math.\frac{\mathrm{rpm}}{\mathrm{bar}}$

and T.sub.I=0.085 sec. After calculating the control parameters, a successful parameterization of the control parameters can be verified by specifying jumps in the target system pressure p_soll. As can be seen in FIG. 5, the actual system pressure p_ist follows the predetermined target system pressure p_soll. Since approximate relationships between a closed and open loop are used in the frequency characteristic method, deviations may occur in the actual behavior of the closed loop. Thus, even with a desired low overshoot or no overshoot, e.g. when using a PI controller in combination with a short rise time, a higher overshoot may be evident in the actual system, as can be seen in FIG. 5. In order to reduce or completely suppress this overshoot, instead of the PI controller a PID controller can be parameterized according to the invention, for example.

[0041] Due to the generally formulated transfer function Gp/n(z) with multiple poles and zeros, more complex controllers can also be designed as PID controllers. In this way, oscillatory transfer functions with complex conjugate poles can be identified. Such identified resonance frequencies can then be compensated for, by means of a notch filter for example. Anti-resonance frequencies can thus also be identified, which can be compensated for subsequently, e.g. using a biquad filter.