DE102017123953A1 - Method and apparatus for determining a transfer function in a powertrain component - Google Patents

Abstract

Method for determining a transfer function of a drive train component (102) and apparatus for carrying out the method, wherein in the method a drive of a drive train component (102) is dependent on information about a desired operating variable (SM) of the drive train component (102) at an output of a drive train component (102) information about an actual operating variable (s) of the drive train component (102) is detected, an output value of an infinite impulse response filter (500) is determined depending on the information about the target operating variable (SM), and at least one coefficient ( a, ... a, b, ... b) of the filter (500) is determined depending on a result of a comparison of the output value with the information about the actual operating variable (s).

Description

The invention relates to a method and a device for determining a transfer function of a drive train component, in particular an automated friction clutch in a motor vehicle drive train.

Out DE 102012217132 A1 a method for reducing chatter is known in which a Lockln controller from a transmission input speed determines a correction torque for a desired clutch torque. As a controlled system, the transmission behavior between nominal clutch torque and transmission input speed is considered. A torque error caused by geometrical coupling errors can be compensated for by virtue of the fact that the correction torque generates a setpoint clutch torque which oscillates in phase opposition to the juddering vibrations. Thus, the juddering vibrations are eradicated.

In order to determine a suitable correction torque, the transfer function for the Lockln controller must be known.

Out WO 2016070879 A1 a method for determining a transmission behavior of a drive train is known by which a transfer function can be determined by a measurement of the motor vehicle. However, this requires mathematically complicated and complex estimation functions and complete measurements to determine a frequency response.

The determination of the transfer function is very time-consuming due to the complexity of the estimator used and requires large storage space. An application in a control device in the motor vehicle and during operation of the motor vehicle is thus impossible.

It is the object of the present invention to provide an improved method and apparatus for determining a transfer function of a powertrain component.

This object is achieved by the method and apparatus according to the independent claims.

With respect to the method, the object is achieved in that a drive of a drive train component is dependent on information about a desired operating variable of the drive train component is detected at an output of a drive train component information about an actual operating variable of the drive train component, an output value of a filter with infinite impulse response dependent is determined by the information on the target operating quantity, and at least one coefficient of the filter is determined depending on a result of a comparison of the output value with the information about the actual operating variable. As a result, an error deviation between output value and actual operating variable is used to correct the coefficients. The filter with infinite impulse response requires little storage space and can be executed in a control device for a motor vehicle and during operation of the motor vehicle, in particular in real time. Real time here denotes computing processes with a process time which is, for example, less than 200 ms or, for example, 100 ms, 20 ms, 10 ms or 5 ms.

Preferably, the filter comprises a first filter for an input signal of the filter and a second filter for an output signal of the filter, wherein the output signal is fed back via the second filter and added to a filter output of the first filter. This implements the output signal by means of a particularly efficient discrete infinite impulse response filter.

Preferably, an adaptation of the at least one coefficient is determined depending on the result of a comparison of the output value with the information about the actual operating variable, wherein the adaptation or the coefficient is limited by a limit value.

Preferably, the filter comprises less than twenty or less than fifteen, in particular thirteen coefficients. This saves memory space and reduces the computational effort with a high quality of the determined transfer function.

Preferably, the powertrain component is an automated friction clutch, wherein as the information about the actual operating size information about a current speed of the automated friction clutch is detected, and wherein the automated friction clutch is driven depending on information about a desired clutch torque as information about the desired operating variable , This can be determined during operation of the motor vehicle, a transfer function for an automated friction clutch of the motor vehicle. The transfer function can be used to reduce juddering vibrations.

Preferably, a time profile of the desired operating variable is specified, the information about the actual operating variable of the drive train component is predefined during the time course acquired at a time, and the output value of the filter with infinite impulse response depends on a current value of the desired operating variable determined at this time. This allows a particularly good system identification of the powertrain component.

Preferably, the time profile of the desired operating variable is specified as a pseudo random bit stream for desired clutch torques. The pseudo random bit stream represents a kind of white noise and includes the frequency components necessary for the system identification.

Preferably, after a start of specification of the time profile of the desired operating variable, it is determined whether a transient process of the drivetrain component has ended, wherein the time at which the information about the actual operating variable of the drivetrain component is detected is after the end of the transient process. For example, a settling time period is awaited.

The information about the actual operating variable of the drivetrain component is detected repeatedly, the output value of the infinite impulse response filter is determined during the course of time of the nominal operating variable, and the at least one coefficient of the filter is determined. As a result, time profiles of the measured actual operating variable and specific output values are used to determine the at least one coefficient.

It is preferably checked whether an operating state of the drive train component, of a drive train in which the drive train component is arranged, or of a motor vehicle in which the drive train component is arranged, is a predetermined operating state, wherein the time profile of the desired operating variable is specified when the operating state corresponds to the specified operating condition. This ensures that the motor vehicle is in a suitable state, for example creep mode, before the system identification takes place.

With respect to the device, a micro-processor, an input device and an output device are provided, which are designed to execute the method in a motor vehicle during operation of the motor vehicle, wherein the output device is designed to control the driveline component as a function of the nominal operating variable for the driveline component, the input means is adapted to receive the information about the actual operating magnitude of the powertrain component detected at the output of the driveline component, and wherein the micro-processor is configured to determine the output value of the infinite impulse response filter depending on the desired operating magnitude , and to determine the at least one coefficient of the filter depending on the result of the comparison of the output value with the actual operating variable.

• 1 schematisch Teile eines Antriebsstrangs,
• 2 einen zeitlichen Verlauf einer Soll-Betriebsgröße,
• 3 Schritte in einem Verfahren zur Systemidentifikation,
• 4 zeitliche Verläufe der Soll-Betriebsgröße und einer Ist-Betriebsgröße,
• 5 schematisch Details einer Systemidentifikation.

Further advantageous embodiments will become apparent from the following description and the drawings. In the drawing shows

• 1 schematically parts of a powertrain,
• 2 a temporal course of a target operating variable,
• 3 Steps in a system identification procedure,
• 4 time profiles of the desired operating variable and an actual operating variable,
• 5 schematically details of a system identification.

1 schematically shows parts of a drive train of a motor vehicle 100 , The car 100 includes a drive component 102 , and a device 104 for determining a transfer function of a powertrain component 102 ,

The drive component 102 For example, an automated friction clutch, in particular a dual-clutch transmission in the drive train of the motor vehicle 100 ,

The device 104 for determining the transfer function comprises a micro-processor, an input device and an output device. The output device is configured, for example, information about a desired operating variable M for driving the drive component 102 issue. The input device is designed to information about an actual operating variable n of the drive component 102 attached to an output of the powertrain component 102 is detected, receive.

The device 104 may be formed, the drive component 102 driving. The device 104 is for example a control unit of the motor vehicle 100 , For example, a clutch control software runs in the control unit.

The clutch control software is designed in the example, depending on a predetermined clutch torque SM , for example, a driving strategy control unit as a target operating size SM pretends to determine a clutch desired torque M.

The output device is designed, the drive train component 102 depending on the nominal operating variable SM for the powertrain component 102 driving. The output device is designed, for example, to control the automated friction clutch with a desired clutch torque M.

The input device is designed, for example, to receive information about a transmission input rotational speed as the information about the actual operating variable n of the drive component 102 to recieve.

The device 104 is formed, the method described below in the motor vehicle 100 , in particular during operation of the motor vehicle 100 perform.

The micro-processor is configured to determine an output value of an infinite impulse response filter depending on the target operation amount M as described below, and to add at least one coefficient of the filter depending on a result of comparing the output value with the actual operation amount n determine.

The device 104 is designed to specify a time course SM (t) of the desired operating variable SM. Exemplary of a particularly suitable time course of the desired operating variable SM is in 2 a pseudo random bit stream is shown over a time axis t. In the example, the target clutch torque SM at random times t between a first value M1 , z. B. -1 Nm and a second value M2 , z. B. +1 Nm, switched. An at the same time the device 104 For example, the desired clutch torque given by the driving strategy unit is used to determine the first value and / or the second value. The time profile SM (t) of the setpoint operating variable SM is preferably predefined as a pseudo random bit stream for setpoint clutch torques. The pseudo random bit stream represents a kind of white noise. In the example, the pseudo random bit stream includes the frequency components necessary for the system identification of the automatic friction clutch of the motor vehicle.

Other frequencies and other time histories SM (t) may also be used. Conveniently, temporal courses SM (t) representing approximately white noise or comprising the frequency components required for the system identification.

The drive of the powertrain component 102 occurs depending on the information about the desired operating variable SM of the powertrain component 102 , At the output of the powertrain component 102 is the information about the actual operating size n of the powertrain component 102 detected. The output value of the infinite impulse response filter is determined as a function of the information about the set operating variable SM. At least one coefficient of the filter is determined depending on the result of the comparison of the output value with the information about the actual operating quantity n.

An adaptation of the at least one coefficient is determined, for example, depending on the result of a comparison of the output value with the information about the actual operating variable n. The adaptation or the coefficient is limited in the example by a limit. This makes the system identification more robust against interference.

Preferably, the filter comprises less than twenty or less than fifteen, in particular thirteen coefficients. This saves memory space and reduces the computational effort with sufficiently high quality of the determined transfer function. For example, for a third order infinite impulse response discrete filter, i. a filter with three delay elements, only four discrete readings used.

Preferably, only a predetermined number of measured values, for example only four discrete measured values, are used. Preferably, the actual operating variable n is sampled to generate the information about the actual operating variable with a sampling frequency of 50 Hz. This is sufficient for sampling in a frequency range between 0 and 25 Hz, in which the vibrations of the automated friction clutch usually move.

After the start of the procedure will be in an optional step 302 checks whether there is a suitable operating state for the system identification.

Preferably, it is checked whether an operating state of the powertrain component 102 , the powertrain in which the powertrain component 102 is arranged, or of the motor vehicle 100 in which the powertrain component is arranged is a predetermined operating state.

If the predetermined operating state is present, for example, the motor vehicle 100 is in an appropriate state, such as crawl mode, becomes an optional step 304 executed. Otherwise, the example in step 302 Wait until the appropriate condition is present.

In step 304 the coefficients of the filter are initialized. For example, the coefficients are initialized to a predetermined value for which the transfer function of the filter fits well with the friction clutch used. These values are determined, for example, in the design of the friction clutch. Other, for example random values for the coefficients can also be specified.

Subsequently, a step 306 executed.

In step 306 becomes the time course SM (t) the target operating quantity SM For example, given as the mentioned pseudo random bit stream for the desired clutch torque.

Subsequently, an optional step 308 carried out.

In step 308 it is checked whether the transient is over. When the transient is over, one step becomes 310 executed. Otherwise, in step 308 Wait until the transient is over. After a beginning of the specification of the time course SM (t) of the desired operating variable SM For example, it is determined whether the transient process of the powertrain component 102 is over. For example, a settling time period is awaited.

In step 310 At a discrete instant i, an error deviation e (i) = d (i) -y (i) between a discrete output value y (i) and a discrete actual operating variable d (i) is determined. The discrete output value y (i) is determined, for example, by means of a transfer function G (z) of the filter from a discrete actual operating variable d (i) acquired at the discrete instant i. In the example, the discrete actual operating variable d (i) is a value of the actual operating variable n at the discrete instant i, ie in the example a measured value of the instantaneous rotational speed of the automated friction clutch is used.

The information about the actual operating variable n of the powertrain component 102 is detected at time i while the timing SM (t) is set, and the output value of the infinite impulse response filter is detected depending on a current value of the target operation amount SM i determined at this time. This allows a particularly good system identification of the powertrain component 102 , The point in time at which the information about the actual operating variables of the powertrain component 102 is detected, for example, lies temporally after the end of the transient process.

Subsequently, a step 312 executed.

mit

• µ: Faktor für Lernrate
• x_f: Filterzustandsvektor.

In step 312 a correction of the coefficients is performed. The correction of the coefficients is carried out, for example, by means of the least mean square algorithm, wherein a new vector w (i + 1) of the coefficients is determined as a function of a current vector w (i) of the coefficients $$\text{w}\left(\text{i}+1\right)=\text{w}\left(\text{i}\right)+\mathrm{\mu }{\text{x}}_{\text{f}}\left(\text{i}\right)\text{e}\left(\text{i}\right).$$

With

• μ: factor for learning rate
• x _f : filter state vector.

mit $$\text{A}\left(\text{z}\right)={\text{a}}_0\left(\text{i}\right)+{\text{a}}_1\left(\text{i}\right){\text{z}}^{-1}+\dots +{\text{a}}_{\text{o}}\left(\text{i}\right){\text{z}}^{-\text{o}}$$

$$\text{B}\left(\text{z}\right)={\text{b}}_0\left(\text{i}\right)+{\text{b}}_1\left(\text{i}\right){\text{z}}^{-1}+\dots +{\text{b}}_{\text{p}}\left(\text{i}\right){\text{z}}^{-\text{p}}$$

a₀, ... a_o, Koeffizienten eines ersten Filters A für das Eingangssignal mit Filterordnung o
b₀, ... b_p, Koeffizienten eines zweiten Filters B für das rückgekoppelte Ausgangssignal mit Filterordnung p.The transfer function G (z) of the filter is $$\text{G}\left(\text{z}\right)=\frac{A\left(z\right)}{B\left(z\right)}$$

With $$\text{A}\left(\text{z}\right)={\text{a}}_0\left(\text{i}\right)+{\text{a}}_1\left(\text{i}\right){\text{z}}^{-1}+...+{\text{a}}_{\text{O}}\left(\text{i}\right){\text{z}}^{-\text{O}}$$

$$\text{B}\left(\text{z}\right)={\text{b}}_0\left(\text{i}\right)+{\text{b}}_1\left(\text{i}\right){\text{z}}^{-1}+...+{\text{b}}_{\text{p}}\left(\text{i}\right){\text{z}}^{-\text{p}}$$

a ₀ , ... a _o , coefficients of a first filter A for the input signal with filter order o
b ₀ , ... b _p , coefficients of a second filter B for the feedback output signal with filter order p.

A filter output of the first filter A is added to the output signal coupled back via the second filter B.

The vector of the coefficients w is formed from the coefficients a ₀ ,... A _o , b ₀ ,... B _p .

Subsequently, the filter is adapted with the new coefficients, i. the new coefficients will be used in the next step.

Subsequently, a step 314 executed.

In step 314 It is checked whether an adaptation time has expired or if a driving strategy requires an abort of the adaptation. If the adaptation time has expired or a driving strategy requires an abort of the adaptation, a step will be taken 316 executed. Otherwise, the step becomes 310 executed. During the time course SM (t) the target operating quantity SM is specified, the information about the actual operating variable n of the drive train component is repeated 102 detected. The output value of the infinite impulse response filter and the at least one coefficient of the filter are determined several times. As a result, time profiles of the measured actual operating variable and specific output values are used to determine the at least one coefficient. Through the repeated execution of the step 310 and 312 the coefficients converge and the error deviation is reduced. This improves the model quality. Ideally, the transfer function G (z) of the filter is identical to a transfer function of the real powertrain component 102 , The model quality of this transfer function G (z) is sufficiently good even with a small number of coefficients a ₀ ,... A _o , b _{0 ,.} By a small number of coefficients a _0, ... a _o, b _0, ... b _p, the discrete output value y (i) can be calculated with less amount of computing resources. In particular, a small number of measured values is sufficient for a small number of coefficients. This reduces a memory requirement and a number of arithmetic operations to be performed.

In step 316 the specification of the time course SM (t) of the set operating variable SM is terminated.

Subsequently, a step 318 executed.

In step 318 the coefficients of the filter are stored. Optionally, the coefficients or the transfer function determined by this system identification may be the powertrain component 102 also be provided to other functions.

Then the process ends.

Such an implemented filter with infinite impulse response can in the control unit for the motor vehicle 100 and during operation of the motor vehicle 100 , especially in real time. Real time here denotes computing processes with a process time which is, for example, less than 200 ms or, for example, 100 ms, 20 ms, 10 ms or 5 ms.

In the example, the powertrain component 102 the automated friction clutch, wherein as the information about the actual operating quantity n information about a current speed of the automated friction clutch, as discrete actual operating variables d (i) is detected at discrete times i. Thus, the transfer function G (z) for the automated friction clutch of the motor vehicle during operation of the motor vehicle 100 determined. The transfer function G (z) can be used to reduce juddering vibrations. For this purpose, for example, the above-mentioned Lockln controller is used.

4 sets the time course SM (t) the target operating variable SM and the course of the actual operating variable n over the time t. The time course SM (t) starts at t = 0 with the first value M1. At a time t1, an activation of the drive train component starts in the example 102 with the pseudo random bit stream or some other kind of white noise. In the example, the pseudo random bit stream is turned off 2 for a period of time t _{PRBS_Seq} between the time t1 and a time t2 and alternately predetermined with moments with the first value M1 and the second value M2. From the time t2, the first value M1 is specified.

The reaction of the actual operating variable n, in the example of the transmission input speed, is in the upper part of 4 shown. At the time t1 starts a transient, the t after Einschwingzeitdauer _transient ends. Immediately thereafter, the adaptation starts in an adaptation period t _adaptation , which lasts until time t2. During the Einschwingzeitdauer _settling t is no adaptation takes. During the adaptation period t _adaptation , the adaptation takes place at least during a minimum time t _{min_adaption} . This may coincide with the adaptation period t _adaption , or be shorter. As long as the time course SM (t) of the desired operating variable SM is constantly specified as torque M1, the reaction of the actual operating variable n is smooth. While the changing time course SM (t) of the desired operating variable SM is specified between the times t1 and t2, the actual operating variable n oscillates. Immediately after the time t2, this oscillation stops and the course of the actual operating variable n becomes smooth again ,

5 schematically shows details of the system identification with a filter 500 with infinite impulse response, in which the first filter A (z) and the second filter B (z) by an adaptation device 502 with respect to their coefficients a _0, ..., _a, b _0, ... b _p in which the adaptation device 502 the new coefficients are determined as described.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 102012217132 A1 [0002]
• WO 2016070879 A1 [0004]

Claims (11)

A method for determining a transfer function of a drive train component (102), characterized in that a drive of a drive train component (102) is dependent on information about a desired operating variable (SM) of the drive train component (102), at an output of a drive train component (102) information about a Actual operating variable (s) of the powertrain component (102) is detected, an output value of an infinite impulse response filter is determined as a function of the information about the desired operating variable (SM), and at least one coefficient (a ₀ , ... a _o , b ₀ , ... b _p ) of the filter is determined as a function of a result of a comparison of the output value with the information about the actual operating variable (s). Method according to Claim 1 , characterized in that the filter (500) comprises a first filter (A) for an input signal of the filter (500) and a second filter (B) for an output signal of the filter (500), wherein the output signal via the second filter (B ) and added to a filter output of the first filter (A). Method according to Claim 1 or 2 , characterized in that an adaptation of the at least one coefficient (a ₀ , ... a _o , b ₀ , ... b _p ) determined depending on the result of a comparison of the output value with the information about the actual operating variable (s) , wherein the adaptation or the at least one coefficient (a ₀ , ... a _o , b ₀ , ... b _p ) is limited by a limit. Method according to one of the preceding claims, characterized in that the filter (500) comprises less than twenty or less than fifteen, in particular thirteen coefficients (a ₀ , ... a _o , b ₀ , ... b _p ). Method according to one of Claims 1 to 4 , characterized in that the drivetrain component (102) is an automated friction clutch, wherein as the information about the actual operating variable (s) information about a current rotational speed of the automated friction clutch is detected, and wherein the automated friction clutch is dependent on information about a target Clutch torque is driven as information about the target operating variable (SM). Method according to one of Claims 1 to 5 , characterized in that a time course (SM (t)) of the desired operating variable (SM) is specified, the information about the actual operating variable (s) of the drive train component (102) during the time course (SM (t)) specified is detected at a time, and the output value of the infinite impulse response filter (500) is determined depending on an instantaneous value of the target operation amount (SM) at that time. Method according to Claim 6 , characterized in that the time profile (SM (t)) of the desired operating variable is specified as a pseudo random bit stream for nominal clutch torques. Method according to Claim 6 or 7 , characterized in that after a start of specification of the time course (SM (t)) of the set operating variable (SM) it is determined whether a transient process of the driveline component is over, the time at which the information about the actual Operating variable (s) of the powertrain component (102) detected is temporally after the end of the transient. Method according to one of Claims 6 to 8th , characterized in that during the course of time (SM (t)) of the desired operating variable (SM) is specified repeatedly the information about the actual operating variable (s) of the drive train component (102) detected, the output value of the filter (500) determined with infinite impulse response, and the at least one coefficient (a ₀ , ... a _o , b ₀ , ... b _p ) of the filter (500) is determined. Method according to one of the preceding claims, characterized in that it is checked whether an operating state of the drive train component (102), a drive train in which the drive train component (102) is arranged, or a motor vehicle (100) in which the drive train component (102) is arranged, a predetermined operating state, wherein the time course (SM (t)) of the desired operating variable (SM) is specified when the operating state corresponds to the predetermined operating state. Device for determining a transfer function of a drive train component (102), characterized in that the device comprises a micro-processor, an input device and an output device, which are designed to perform the method according to one of Claims 1 to 10 in an automotive vehicle (100) during operation of the motor vehicle (100), wherein the output device is adapted to drive the powertrain component (102) responsive to the desired operating size (SM) for the powertrain component (102), the input device is configured to provide the information about receive the actual operating variable (s) of the powertrain component (102) detected at the output of the powertrain component (102), and wherein the micro-processor is configured to adjust the output value of the infinite impulse response filter (500) in accordance with FIG Target operating size (SM) to determine, and at least a coefficient (a ₀ , ... a _o , b ₀ , ... b _p ) of the filter (500) depending on the result of the comparison of the output value with the actual operating variable (SM) to determine.

Images