DE102011006227A1 - Method for optimizing adjusting process in control system of internal combustion engine of hybrid motor vehicle, involves continuing adjustment of control system after transient oscillation process and after reaching of reference value - Google Patents

Abstract

The method involves using stimulation (10) in the form of high gradient stimulation (10.1) in different phases by using a servo control according to clearance principle till to reaching of a reference value during transient oscillation process. Distribution of the phases is adapted to a control system in a characteristics-dependent and automatic manner, where optimal distribution of pre-control values is determined during the adaptation. Adjustment of the control system is continued after the transient oscillation process and after reaching of the reference value. An independent claim is also included for a device for optimizing adjusting process in a control system.

Description

The invention relates to a method for optimizing a control process in a control system, which may be combined with an adaptively shaped pilot control.

The invention further relates to a corresponding device for carrying out the method.

State of the art

In the field of engine control development for motor vehicles, it is necessary to tailor the ECU software to the customer's specifications. The ECU software contains different software functions that need to be parameterized or adapted at different locations. Part of the software functions are controllers that regulate specifications (setpoints) to subsystems, such as the air system or to the high-pressure system in diesel injection systems, dynamically as well as stationary.

The transient response of controlled systems, as they occur in the engine-typical systems, is usually controlled with PID controllers, often in combination with a feedforward control, which can be adapted. Examples of this are boost pressure controls, rail pressure regulators, regulations for exhaust gas recirculation, idling control etc.

For example, in the published patent application DE 10 2008 030 480 A1 a method for operating a starting clutch in a hybrid vehicle with an internal combustion engine having a crankshaft and an electric machine with a rotor described, wherein for a start of the internal combustion engine, the electric machine accelerates to a predetermined speed and the effective between the electric machine and the engine starting clutch as a function of the above Starting clutch is delivered to be transmitted moment. It is provided that in a first starting phase, a pilot torque is set to a predetermined value and in a second starting phase when detected start of the internal combustion engine, a target torque, depending on the acceleration of the crankshaft, is regulated. Furthermore, the pre-control torque can be increased in the first starting phase in dependence on the time from the starting phase starting until a start of the internal combustion engine is detected. The pilot torque can be increased in a ramp and can also be stored as a time-dependent characteristic in a control unit and set as a function of time from the start of the start phase as a time-dependent pilot torque.

• a.) Vorgeben einer Sollposition durch einen Sollwert und Auswählen eines Vorsteuerwerts für das Stellglied zum Erreichen der Sollposition aus einem Kennfeld mit jeweils den Sollpositionen zugeordneten Vorsteuerwerten,
• b.) Einstellen einer Vorsteuerposition des Stellgliedes entsprechend dem Vorsteuerwert und Vergleichen der Vorsteuerposition des Stellglieds mit der Sollposition,
• c.) Regeln der Vorsteuerposition bis zum Erreichen der Sollposition, der ein adaptierter Steuerwert entspricht, und Messen des adaptierten Steuerwerts,
• d.) Überschreiben des Vorsteuerwertes der Sollposition in dem Kennfeld mit dem der Sollposition entsprechenden adaptierten Steuerwert und Abspeichern des überschriebenen Vorsteuerwerts als neuer Vorsteuerwert in dem Kennfeld und
• e.) Invertieren der Korrelation zwischen de Sollposition und dem Vorsteuerwert aus dem Kennfeld, so dass der neue Vorsteuerwert in Abhängigkeit von dem vorgegebenen Sollwert aufgetragen ist, und Anwenden der invertierten Korrelation auf die vorgegebene Sollposition.

Another example of an adaptable feedforward control is known from the patent specification DE 10 2006 008 051 B3 known. Then, an adaptive positioning method of an actuator, in particular a throttle valve of an internal combustion engine is described, comprising the following steps:

• a.) Specifying a desired position by a desired value and selecting a pre-control value for the actuator to reach the desired position from a map with pre-control values respectively associated with the desired positions,
• b.) setting a pilot position of the actuator according to the pilot value and comparing the pilot position of the actuator with the target position,
• c.) controlling the pilot position until reaching the target position corresponding to an adapted control value, and measuring the adapted control value,
• d.) overwriting the pre-control value of the target position in the map with the adaptive control value corresponding to the target position and storing the overwritten pre-control value as new pre-control value in the map and
• e.) inverting the correlation between the desired position and the pre-control value from the map such that the new pre-control value is plotted against the predetermined target value, and applying the inverted correlation to the predetermined target position.

When designing the transient response in such pilot controls, the conflict of objectives arises between the reduction of overshoot or undershoot and the achievement of maximum dynamics taking into account the stability criteria. An overshoot should be minimized or eliminated as far as possible, the time to reach the setpoint should also be minimized as possible. The overshoot behavior and the time until the setpoint is reached thereby represent the properties or criteria for the design and evaluation of the transient response of the controller. However, it is not apparent from the prior art cited above that an extinction of vibrations is achieved by appropriate shaping of the precontrol can be.

The object of the invention is therefore to represent an alternative way to simplify the Einregelvorgang of control systems and thus to improve the control behavior, especially with regard to avoiding overshoots.

It is a further object of the invention to provide a device for carrying out the method.

Disclosure of the invention

The object of the method is solved by the features of claims 1 to 11.

According to the invention it is provided that during the transient process until reaching a desired value excitation by means of a pilot control according to the principle of eradication, d. H. Extinguishing vibrations, is used with a shaped excitation in different phases, the division of the phases depending on the property and automatically adapts to the control system and after the transient and after reaching the setpoint, the control is continued by means of a conventional regulator.

The device-related object of the invention provides for carrying out the method that the device is part of a control unit, wherein the control unit function of the control device has process features according to the inventive method, wherein the control device function software or hardware-based is implemented in the control unit. It is provided that the calculation of the phases is implemented in the context of the control unit function, the height and the shape and the duration of the first phase are calculated from the waveform of the corresponding internal control variables.

With the method according to the invention and the corresponding device, an application simplification can be achieved by reducing the measurement and parameterization effort. Due to its adaptive behavior, adjustments to system tolerances and aging of components of the control system and operating point or mode-dependent adjustments can be made independently, which is conducive to an increase in the system robustness and thus an increase in quality.

A preferred variant of the method provides that the shaped excitation is divided into two time-shifted phases, wherein in the first phase the proportion of a total jump height of a step-like excitation is selected such that the maximum of a system response just reaches the stationary reference value and when changing from the first phase to the second phase for the second phase, a second jump height is selected which corresponds to the difference between a total jump height and the first jump height. In this case, the phase phase T _{s of} the first phase used is a rise time T _a until the setpoint is reached or an overshoot time T _m until the maximum overshoot is reached.

The supplied energy in phase 1 generates a vibration in the system. The jump-like excitation of the second phase at the maximum, a striking point according to the repayment principle, generates a counter-vibration to the natural vibration of the system. Only with a counter-vibration generated at the right time and with the necessary or required energy, the vibrations cancel each other out. In the optimal case, there are no longer any vibrations that a controller needs to correct. In the best case, the controller is only used for dynamic changes in order to correct stationary deviations and disturbance variables. Thus, an optimal and rapid setting of the target value can be achieved without the control system tends to overshoot. In addition to an improvement in the quality of the transient process (ie faster control, without overshoot), the application effort can be reduced.

The adaptation of the properties in the first phase, ie the first jump height and / or the phase duration T _{s of} the first phase, can by means of a comparison of an actual value with a predetermined overshoot Δh, ie Δh = 0 or Δh = maximum allowable value or predetermined Setpoint, are performed, when exceeding the default value for the overshoot Δh the excitation with respect to the height and / or duration, ie the energy is reduced and increases below the default value, the excitation in terms of height and / or duration. Thus, an efficient adaptation of the control system can be achieved.

The adaptation of the second jump height can be carried out by means of an evaluation of an actual value of the temporal gradient Δy (signal change within a specific time interval) after the phase duration T _{s of} the first phase, wherein at a value Δy> 0 the excitation is reduced in terms of height and at a value Δy <0 the excitation is increased. This can be achieved that after the stimulation a steady state can be reached very quickly. Any excesses from the excitation of the system can be avoided.

With regard to a control behavior with great dynamics can also be provided that selected in the first phase for the first jump height, the total jump height of the jump-like excitation value and compared to the original phase duration T _{s of} the first phase, which the rise time T _a to reach the set value or the overshoot time T _m corresponds to the achievement of the maximum overshoot, is prematurely changed to a lower value. In order to avoid overshoots of this step-like excitation, this lower value is precontrolled until no or a defined overshoot occurs. The second phase will be after reaching the Maximum jump started with the stationary required jump-shaped excitation to maintain the setpoint.

In this variant of the method, it can further be provided that a value for the shortening of the phase duration is determined by adapting the system response and learned adaptively, this being reduced if the overshoot height Δh <0, and increased if the overshoot height Δh> 0 ,

If the parameters of the excitations are determined as a function of a weighting between minimizing overshoots and maximizing the dynamics of the control system, as provided by a further advantageous variant of the method, it can be determined individually whether as short as possible rise time T _a or avoidance of control deviations for the respective control task it is asked for.

Furthermore, it is advantageous if, in addition, the influence of system dead times is taken into account in the parameterization of the excitations, which improves the quality of the control behavior.

The excitation can be carried out in discrete steps and / or continuously. For example, a sawtooth excitation with a jump-like excitation at the beginning and a continuous drop after it is possible. Thus, an optimal adaptation to the corresponding application is possible. It should be noted that a continuous shaping of the excitation also quasi-continuous, d. H. can be done in small, discrete steps. Furthermore, a shaping of the excitation may be possible, resulting in the superposition of several independent suggestions.

In a variant of the method it is provided that the adaptation of the first phase is realized with a map-based precontrol. In this case, the precontrol values of the excitations are determined by stationary measurements and stored in predetermined characteristic maps, during which the optimal distribution of the stored precontrol values is determined as a function of the operating point or mode depending on the first phase.

In a further variant of the method, an adaptation of the first and second phases is provided as pure adaptive precontrol. In this case, the pilot control values of the excitations for the second phase are determined as a function of operating point or mode-dependent.

The principle of eradication works optimally only with a defined distribution of the excitation and when triggered at the maximum, as described above in the various variants. But even slightly different from the repayment principle formations do not lead to optimal vibration damping, but can still lead to good or desired results. It may therefore be necessary, for example with regard to a compromise between settling time and the occurrence of overshoots, for certain applications, to deviate slightly from the optimal excitation variants described above according to the principle of repayment.

A preferred use of the method with its previously described method variants provides for use within an engine control system for controlling and regulating an internal combustion engine and / or an exhaust aftertreatment system of the internal combustion engine, whereby the customer or project-specific application effort in the adaptation of the control unit software over the prior art, such as mentioned in the beginning, can be reduced. Preferred applications are, for example, boost pressure controls, rail pressure regulators, regulations for exhaust gas recirculation, idling control, etc.

Brief description of the drawings

The invention will be explained in more detail below with reference to an embodiment shown in FIGS. Show it:

1 in a waveform diagram a jump-like excitation and the corresponding system response,

2 in a second waveform diagram a stepped shaped excitation with the corresponding system response,

3 in a third waveform diagram a further shaped excitation with the corresponding system response,

4 in a fourth waveform diagram an alternatively shaped excitation with the corresponding system response,

5 various waveform diagrams as simulation results with self-adapting jump parameters and

6 different waveform diagrams as simulation results with a stepped excitation.

1 shows in a waveform diagram 1 depending on the time 30 a suggestion 10 in the form of a jump-like excitation 10.1 (top Diagram) and a system response 20 using the example of a controlled system with a PT2 behavior. A PT2 element is a transmission element in control engineering which has a proportional transmission behavior with a delay 2 , Has order. The transient response of the step response can, inter alia, with the parameter overshoot 24 (Δh), the time to reach the maximum overshoot 26 (Maximum 23 ), the overshoot time 22 (T _m ) and the time to reach the setpoint, the startup time 21 (T _a ). Another parameter to characterize the system response 20 can additionally the gradient 25 (Δy), ie the time derivative of the signal height.

In 2 is in another waveform diagram 1 a suggestion 10 in the form of a shaped stimulus 10.2 shown, which in the example shown in two staggered phases 12.1 . 12.2 is divided. In the first phase 12.1 is the proportion of the total erratic stimulus 10.1 out 1 chosen such that the maximum 23 the system response 20 just reached the stationary end value (setpoint) and the height of the overshoot, ie the overshoot height 24 Δh = 0. Likewise, the gradient 25 Δy = 0. The phase duration 12 (T _s ) of the first phase 12.1 corresponds to the time to reach the maximum 23 , the overshoot time 22 (T _m ), as well as the rise time 21 (T _a ). When changing from the first phase 12.1 to the second phase 12.2 is for a second jump height 11.2 a difference value between the pre-control value for the setpoint and the first jump height 11.1 the first jump-like excitation at the beginning of the first phase 12.1 set. The dynamic process of precontrol is with this 2-stage shaped excitation 10.2 completed. The regulation for the stationary part after the transient process is then carried out by a conventional controller (eg a PI controller).

Depending on whether you focus on minimizing the overshoot 26 or maximizing the dynamics of the system response 20 lays, becomes the first jump height 11.1 in the first phase 12.1 selected. So can by choosing a larger jump height in the first phase 12.1 the startup time 21 (T _a ) at the expense of a higher overshoot 26 (Overshooting 24 Δh> 0) are shortened.

3 shows an alternative controller pre-control to avoid overshoots 26 , In the waveform diagram 1 is a two-stage shaped stimulus 10.2 shown, in which the shaping on the first phase 12.1 refers. The precontrol value is set to a value that is increased in this case, the value increased for the pre-control value required for the stationary state (first jump height 11.1 ) for maximum dynamics. To the overshoot 26 To avoid this sudden excitation, the duration of this excessive feedforward control is shortened, in which case as long as a lower pilot control value is applied to the input side, to no or a defined overshoot 26 arises. The time required to reduce the system input can be via the system response 20 calculated or learned adaptively. It is reduced if the overshoot height 24 Δh <0 and increases when the overshoot height 24 Δh> 0. The triggering of the second phase 12.2 done as in 2 represented, at the maximum 23 (Gradient 25 Δy = 0), wherein the precontrol value required for the steady state, that is, the value required to keep the system at the target value, is set abruptly.

4 shows as a further possibility a shaped excitation 10.2 in which the suggestion 10 in discrete jumps (jump-like excitation 10.1 ) and / or continuously (dashed line) is performed. In the example shown, after setting a first jump height 11.1 within the first phase 12.1 the pre-control value is linearly reduced, so that in the system response 20 an overshoot 26 is avoided. After that, at the beginning of the second phase 12.2 leaps and bounds with a second jump height 11.2 the pre-control value is set to the pre-control value required for setting the stationary state.

5 shows in different waveform diagrams 1 Simulation results with self-adapting jump parameters. The left partial image (Figure 1) shows as a suggestion 10 a jump-like excitation 10.1 as well as the corresponding system response 20 in a typical PT2 controller with the overshoots 26 , According to the method approach according to the invention, the occurring oscillation of the step response is canceled by means of a stepped excitation. This is the total jump height 11 in two partial jumps with the first jump heights 11.1 and the second jump height 11.2 divided up. First, a partial excitation in the first phase 12.1 with the first jump height 11.1 according to picture 2 in 5 to which then in the second phase 12.2 a second jump height 11.2 is added (Figure 3 in 5 ). The distribution of jump heights 11.1 . 11.2 and the time offset is such that the oscillation is almost completely eliminated (Fig. 4 in 5 ). As a result of the addition of the two pilot jumps, the shaped excitation is obtained 10.2 ,

In 6 is this superimposition of two temporally staggered jump-shaped excitations with the different jump heights 11.1 . 11.2 shown as picture above left. As a result, the shaped stimulus is obtained 10.2 , which is also shown in the picture above right. For comparison is also the jump-like excitation 10.1 , which is performed in one step, shown. The system response to the first suggestion 20.1 and the system response to the second stimulus 20.2 is in 6 shown in the picture below left. It can be seen that by overlaying the two system answers 20.1 . 20.2 the vibrations are so out of phase that the maxima of the excursions are compensated. The resulting system response 20.3 is shown in the picture below right. For comparison is also the system response 20 at the jump-like excitation 10.1 shown in one step.

Basically, the shaped stimulus 10.2 be composed as an overlay of different partial excitations. As system response 20 results in a superposition of corresponding to the partial excitations resulting system responses 20.1 . 20.2 , wherein according to the object of the invention overshoot 26 be avoided. The partial suggestions can be adjusted individually or together.

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 102008030480 A1 [0005]
• DE 102006008051 B3 [0006]

Claims (13)

Method for optimizing an adjustment process in a control system, which can be combined with an adaptively shaped precontrol, characterized in that during the transient process, until a desired value is reached, an excitation ( 10 ) by means of a precontrol according to the principle of eradication with a shaped excitation ( 10.2 ) in different phases ( 12.1 . 12.2 ), the division of the phases ( 12.1 . 12.2 ) depending on the property and independently adapted to the control system and after the transient and after reaching the setpoint, the control is continued by means of a conventional controller. Method according to claim 1, characterized in that the shaped excitation ( 10.2 ) into two staggered phases ( 12.1 . 12.2 ), whereby in the first phase ( 12.1 ) for a first jump height ( 11.1 ) the proportion of a total jump height ( 11 ) of a jump-like excitation ( 10.1 ) is chosen such that the maximum ( 23 ) of a system response ( 20 ) just reached the steady-state setpoint and when changing from the first phase ( 12.1 ) to the second phase ( 12.2 ) for the second phase ( 12.2 ) a second jump height ( 11.2 ), which is the difference between a total jump height ( 11 ) and the first jump height ( 11.1 ), where as the phase duration ( 12 ) T _{s of} the first phase ( 12.1 ) a rise time ( 21 ) T _a until the setpoint is reached or an overshoot time ( 22 ) T _m until the maximum overshoot ( 26 ) is used. Method according to Claim 2, characterized in that the adaptation of the first jump height ( 11.1 ) and / or the phase duration ( 12 ) T _{s of} the first phase ( 12.1 ) by means of a comparison of an actual value with a predefinable overshoot height ( 24 ) Δh is performed, whereby when exceeding the preset value for the overshoot height ( 24 ) Δh the excitation ( 10 ) is reduced in terms of height and / or duration, and if it falls below the default value, the suggestion ( 10 ) is increased in terms of amount and / or duration. Method according to one of claims 2 or 3, characterized in that the adaptation of the second jump height ( 11.2 ) by means of an evaluation of an actual value of the temporal gradient ( 25 ) Δy after the phase duration ( 12 ) T _{s of} the first phase ( 12.1 ), where at a value Δy> 0 the excitation ( 10 ) is reduced in terms of height and at a value Δy <0, the excitation ( 10 ) is increased. Method according to one of claims 1 to 4, characterized in that in the first phase ( 12.1 ) for the first jump height ( 11.1 ), the total jump height ( 11 ) of the jump-like excitation ( 10.1 ) and compared to the original phase duration ( 12 ) T _{s of} the first phase ( 12.1 ), which of the start-up time ( 21 ) T _a until reaching the setpoint or the overshoot time ( 22 ) T _m until the maximum overshoot ( 26 ), is prematurely switched to a lower value, whereby as long as this lower value is precontrolled until no or a defined overshoot ( 26 ), and the second phase ( 12.2 ) after reaching the maximum ( 23 ) jump with the stationary required jump-like excitation ( 10.1 ) is started to maintain the setpoint. Method according to Claim 5, characterized in that a value for the shortening of the phase duration ( 12 ) by evaluating the system response ( 20 ) and is learned adaptively, this being reduced when the overshoot height ( 24 ) Δh <0, and is increased when the overshoot height ( 24 ) Δh> 0. Method according to one of claims 1 to 6, characterized in that the parameters of the suggestions ( 10 . 10.1 . 10.2 ) depending on a weighting between minimizing overshoots ( 26 ) and maximizing the dynamics of the control system. Method according to one of claims 1 to 7, characterized in that in addition the influence of system dead times in the parameterization of the suggestions ( 10 . 10.1 . 10.2 ). Method according to one of claims 1 to 8, characterized in that the excitation ( 10 . 10.1 . 10.2 ) is performed in discrete steps and / or continuously. Method according to one of claims 1 to 10, characterized in that the pre-control values of the suggestions ( 10 . 10.1 . 10.2 ) are determined by stationary measurements and stored in predetermined maps, whereby during the adaptation within the first phase ( 12.1 ) the optimal distribution of the stored pre-control values is determined depending on the operating point or mode-dependent. Method according to one of claims 1 to 10, characterized in that the pre-control values of the suggestions ( 10 . 10.1 . 10.2 ) for the second phase ( 12.2 ) depending on operating point or operating mode dependent. Use of the method according to one of claims 1 to 11 within an engine control system for controlling and regulating an internal combustion engine and / or an exhaust aftertreatment system of the internal combustion engine. Device for optimizing a Einregelvorganges a control system, which may be combined with an adaptively shaped pilot control, wherein the device is part of a control unit, characterized in that the controller function of the control device has method features according to claims 1 to 10, wherein the controller function software or hardware -based implemented in the control unit.

Images