DE10062107C1 - Regulation device for electromagnetic actuator uses 3-point regulators for regulating field currents of lifting magnets between which armature of electromagnetic armature is displaced - Google Patents

Abstract

The regulation device uses a pair of 3-point regulators (61,62) for regulating the field currents for the lifting magnets (13,13s,13o) between which the armature (11) of the electromagnetic actuator (1) is displaced, with a measuring sensor (15) for detecting the actual armature position, coupled to a filter (2) providing a filtered armature stroke signal, armature velocity and armature acceleration, fed to a stroke regulator (3). The latter determines the required magnetic force of each lifting magnet, used for calculation of the corresponding field current via a magnetic force characteristic.

Description

The invention relates to a product with the features of the independent claim.

The actuators used in particular in the fully variable electromechanical valve train work on the principle of a weakly damped spring-mass transducer and exist essentially consisting of two solenoids, an armature (magnet armature) and two springs, where when using the actuator in an electromechanical valve control, the lower spring is the conventional valve spring. Controlled by the two solenoids and the pre tensioned springs, the anchor moves back and forth between its two end positions. at an electromagnetic valve control, the two end positions of the armature are the same meaningful with the valve positions "valve open" and "valve closed".

A problem with electromagnetic actuators is the so-called capture of the armature at least one of its end positions. The anchor is usually in its end positions run against a firm stop. This stop is usually the pole face of the fan lifting solenoids. The anchor plate is placed on this pole face when the anchor is caught in its final position. When the anchor plate approaches the pole face of the lifting magnet ten, d. H. with a decreasing air gap between the anchor plate and the pole surface of the Hubmag neten, the magnetic force acting on the armature increases exponentially, while the Ge force of the return spring only grows linearly, so that the armature always in its end positions is accelerated more and hits the pole face with increasing speed. In addition to the development of noise, bouncing processes can occur. H. the anchor first hits the pole face, but then takes off at least briefly until it finally comes fully to the plant. This can impair the function of the the actuator articulated actuator, which is particularly the case with actuators with high Switching frequency, such as. B. in gas exchange valves of internal combustion engines, too significant Can cause interference.  

DE 199 09 109 A1 therefore has a method for detecting the armature movement proposed on an electromagnetic actuator with which the Ankeraufreffgeschwindg speed can be regulated on the pole face of the solenoid. The actuator is special Suitable for driving gas exchange valves in motor vehicles. In the actuator whose type kerplatte between two end positions by means of two solenoids and two springs and can be moved here, a sensor which is dependent on the armature movement becomes Measurement signal obtained. The measurement signal is turned into a Po in a signal processing unit position signal for the current position of the anchor plate and a speed signal for the current speed of the anchor plate is calculated. The signals are used in a path control which compares the current system parameters with the target specifications and from which Comparison derives a control signal for the current regulator of the solenoids. By a verbes Serte spatial resolution of the sensor in the phase of the armature approaching its End position can be more precise control of the anchor speed, which also leads to Reduction of the speed of impact of the anchor plate on the pole face of the Hubmag can be used. The anchor impact speed is given at less than 0.1 m / s give.

DE 199 22 969 A1 is a continuation of DE 199 09 109 A1 by the same applicant Method for operating an electromagnetic valve train for actuating a gas exchange valve addressed to a piston internal combustion engine. With one at the end positions of the Anchor plate high-resolution position sensor is used exclusively on the lowest anchorge speed designed positive control during the approach phase of the anchor plate carried out to their end position. However, this forced control is only lower in the area Engine speeds are practical and must be switched off at higher speeds (see DE 199 22 969 A1, Column 5, lines 9-10). However, wear is especially at higher speeds the biggest.

From the applicant of the patent applications DE 198 50 687 A1, which belong together, DE 198 34 548 A1, DE 198 32 198 A1 and DE 198 32 196 A1 were used in particular in the DE 198 32 196 A1 already describes a method for reducing the impact speed of a Armature of an electromagnetic actuator proposed, with the help of an observer ter model of the main movement variables of the actuator by calculated by the observer Estimates are replaced. A detailed actuator control itself is not in the documents refer to. The observer is preferably designed as an extended Kalman filter and has the task of two of the three state variables required to characterize the movement condition of the anchor plate, namely speed and acceleration To replace estimates. This allows speed and acceleration sensors be saved. (Compare DE 198 32 198 A1, column 9, lines 35-39 and DE 198 32 196 A1, Column 7, lines 44-52). The estimated values for the real conditions in the actuator to be able to adapt, the observer has a correction function by means of which the estimated values can be adjusted to the measured values. (Compare DE 198 32 196 A1, Column 7, line 52-column 8, line 9) This correction function is used for checking of the actuator model that is stored in the observer. Dynamic disturbances that occur during the Actuator operation can occur neither with said observer nor with said correction door function are regulated.

The regulation of the lifting movement along a path-speed trajectory leads in general to improved results in an effort to achieve anchor strike speed to reduce. With this control strategy, disturbance variables can also be used to a limited extent are processed. Disturbances that are quasi-stationary and motion-inhibiting, such as B. he increased damping, friction or gas counter forces caused by the gas back pressure a gas exchange valve articulated on the actuator, which acts on the anchor flight, lead to their occurrence in the known control strategies to no increased anchor meeting speed. Dynamic, movement-promoting disturbing forces, however, are known with the. Compensate control strategies only if they compare with the time constants for the re Sufficiently slow, quasi-stationary. Highly dynamic, dynamic interfering variables that occur during the approach of the armature in one of its end positions ten, cannot be processed and processed sufficiently quickly with the known control strategies be compared. With gas exchange valves that are articulated to an actuator, such are  highly dynamic, movement-promoting disturbing forces, for example in a combustion engine Tor the air flow at the intake valve in the compression phase when the valve is closed late Inlet valve. A deteriorated engine acoustics, a high wear of the valve seats, as well The consequences are a very bad catching of the anchor in the corresponding end positions.

Starting from the previously described state of the art, the inventive method emerges The task is to specify an actuator control system that also regulates highly dynamic interference forces capable, even if these disturbing forces when the actuator approaches one of its end positions occur in a movement-promoting manner.

According to the invention, this object is achieved by the features of the independent An entitlement. Further advantageous embodiments are in the subclaims and in Description included.

The solution is achieved with an actuator control for an electromagnetic actuator that works on the principle of a weakly damped spring-mass oscillator and an armature between the two lifting magnets, driven by the two lifting magnets and the preloaded springs of the spring-mass oscillator, moved back and forth between its two end positions, comprising:

• - two three-point controllers for regulating the field currents in the two solenoids,
• - A transducer for determining the position of the armature and for generating a Position signal as a measure of the current position of the anchor,
• - Including a state variable filter for calculating state variables of the armature a filtered armature stroke signal, armature speed and armature acceleration,
• - A stroke controller, which is designed as a PI controller and which is based on the state variables of the Armature and a predetermined target position of the armature, the currently required target magnet force calculated for the solenoids,
• A selection element which assigns the target magnetic force to a lifting magnet,
• - A nonlinear compensation, the cha from the target magnetic force via a solenoid characterizing magnetic force map the target magnetic force corresponding to the target magnetic force  Field current calculated,
• - One current sensor for each of the two solenoids for measuring the current coils currents in the solenoids,
• - An observer, which consists of the armature acceleration and the current, measured coils a target field current is calculated for each of the two solenoids,

in which,
The target field currents calculated by the observer, the measured coil currents and the target field currents calculated by the non-linear compensation are each applied as input variables to the two three-point controllers and an output voltage is regulated in the three-point controllers from these input variables and this output voltage is applied to the lifting magnets is.

The main advantages of the invention are as follows:
The actuator control reliably enables the magnetic armature plate to land softly on the armature counterplate or the corresponding magnetic yoke of the lifting magnets even when highly dynamic interference forces act on the armature flight. The impact speeds are below 0.02 m / s.

Interference forces that act on the anchor flight in a movement-promoting manner, such as possibly suction forces due to gas forces due to the mixture change in a cylinder Internal combustion engine, can be regulated effectively and reliably.

An observer detects the coil currents through the two lifting magnets of the armature and the current acceleration of the armature. The coil currents in the solenoids determine the magnetic force acting on the anchor flight and thus its acceleration. The magnetic field distribution is determined experimentally for each actuator and mapped in a computer-based model in the form of a so-called state space design. For a given and known magnetic field distribution, the resulting acceleration of the magnet armature is therefore known. An electromagnetic actuator can therefore be model-based in the form of a state space with the parameters armature acceleration, coil current of the closing magnet and coil current of the opening magnet, and can be implemented in the form of a software program in a microcontroller on the observer. If the current acceleration does not correspond to the value for the acceleration to be expected from the state space, the observer calculates a correction for the coil currents and these corrected coil _currents i _{fS_beob} , i _{fO_beob} via summation _elements directly to the input of the two three-point controllers for the lifting magnets given for the anchor flight, ie without going through the complete control loop. The stroke regulator, the selection element and the non-linear compensation are bypassed by the corrected coil currents of the observer. This makes it possible to correct deviations in the armature acceleration caused by disturbing forces very quickly without having to go through the complete control loop with the associated time delay.

The model-based state space is in the form of a software program on a microcontroller of the observ implemented aft. By replacing the software program, the actuator control can be set to different Ak gates or be adapted to various actuators articulated to the actuator, e.g. B. on different mag net valves of different internal combustion engines. The rest of the loop, like that State variable filter, the stroke regulator, the selection element, the two three-point regulator u. s. w. are more advantageous wisely implemented as hardware. As a result, the remaining components of the control loop need to be adjusted not be changed to a new application. The aforementioned hardware components in an advantageous embodiment can be designed as analog hardware components. Analog hardware Components do not require any computing power, which is particularly true with regard to a higher control speed Offers advantages.

Embodiments of the invention are illustrated below with reference to drawings provides and explained in more detail. Show it:  

Fig. 1 shows schematically the overall structure of the Aktorregelung invention,

Fig. 2 shows schematically the structure of a possible known state variable filter for advantageous use in the Aktorregelung.

Fig. 1 shows the overall structure of the Aktorregelung invention. An electromagnetic shear actuator 1 contains a magnet armature 11 , the armature plate 12 of which moves back and forth between two lifting magnets 13 . The upper of the two solenoids 13 acts as a closing magnet 13 S, the lower of the two solenoids 13 acts as an opening magnet 13 o. The magnet armature 11 is additionally biased with a closing spring 14 S and an opening spring 14 o, so that the whole actuator is weakly damped Spring mass forms oscillator. The actuators that can be driven by the actuator are not shown. The actuators themselves do not form an inventive component. Actuators to be driven are articulated to the magnet armature 11 in a manner known per se. A preferred application of the actuator control according to the invention is the use in a fully variable electromagnetic valve control. In this application, the driven actuator is formed by a gas exchange valve of an internal combustion engine. The valve lifter or the valve needle of the gas exchange valve is then non-positively connected to the magnet armature and is driven by the two lifting magnets 13 and the two springs 14 .

The actuator also includes a position sensor 15 with which the current position of the kerplatte 12 can be determined. Inductive or magnetostrictive sensors are preferably used as position sensors. A current sensor 16 S measures the coil current i _{s_mess} through the closing _magnet , another current sensor 16 o measures the coil current i _{o_mess} through the opening magnet.

A state variable filter ZVF is used to filter the armature stroke signal measured with the position sensor 15 . The structure of the state variable filter is described in more detail in connection with FIG. 2. In the state variable filter, the measured position signal x _{mess is} a filtered position _signal x _ZVF as a measure of the current position of the magnet armature 11 , a first time derivative x ' _{ZVF of} the filtered position signal as a measure of the current speed of the magnet armature during the anchor flight and a second Time derivative x " _{ZVF of} the filtered position _{signal is formed} as a measure of the current acceleration of the magnet armature during the anchor flight.

The three arithmetically formed state variables for the armature position, the armature speed and the armature acceleration are fed to the stroke regulator 3 . The stroke controller 3 is implemented as a PI status controller (proportional integral status controller) with four adjustable control parameters r _i , r ₁ , r ₂ , r ₃ . With the control parameters, the task parameters of the overall system to be influenced are specifically strengthened. Analog operational amplifiers are preferably used for this. The control parameters of the stroke controller 3 are determined for the respective actuator, which is to be controlled with the actuator control, via a model-based state space design and contain the most important information about the dynamic behavior of the actuator, such as, for. B. inertia, spring stiffness, damping, time constants etc. This state space design is implemented in the stroke controller in the form of a software program on a microcontroller. The selection of the control parameters is advantageously carried out in such a way that the controlled overall system, ie essentially the armature flight of the magnet armature when approaching the end positions of the magnet armature, is free from overshoot processes. Overshoots would result in the magnetic armature being placed in its end position several times. In order to avoid this overshoot, the magnet armature is moved asymptotically to its end position. The asymptotic approach of the magnet armature to its end position is realized with a stroke controller whose controller parameters r _i , r ₁ , r ₂ , r ₃ are selected, for example, so that the overall system has a 4th order deceleration behavior with a 4-fold pole in the desired time constant T _set having the controlled overall system. The target time constant remains a freely selectable parameter and, within the framework of the basic controller design, enables the overall control to be adapted to the respective requirements of the actuator with regard to its temporal behavior.

For an actuator control of an electromagnetic valve train for gas exchange valve, the following control parameters have proven to be advantageous for amplifying the task variables of the stroke controller 3 :

In it are:
m: anchor mass + valve mass
T _red : largest total time constant of both solenoids
d: damping coefficient
c: resulting spring constant
λ = -1 / T _to: target eigenvalue (pole) of the controlled system
T _set: target time constant of the controlled system

The task size of the stroke control is the difference between a position setpoint x_soll specified by an external control (not shown in more detail) and the filtered anchor position x _ZVF . The difference between the two variables is formed in a summation element 32 . The difference signal is sent to the PI state controller and amplified by the factor r _i at its output. The amplified signals of the state variable filter ZVF are subtracted from the amplified difference signal in a summation element 33 . Here are the signals supplied by the state variable _filter 2 , quantities for the _armature position x _ZVF , the _{armature speed} x ' _ZVF and the armature acceleration x " _ZVF . The _armature position is amplified with the factor r ₁ , the armature speed is amplified with the factor r ₂ and the armature acceleration is amplified with the factor r _3. All three amplified variables of the state variable filter are summed with two summation terms 31 .

The stroke controller 3 is characterized as a PI state controller by its robustness against changeable. System parameters, such as friction, damping or time-variant time constants of the electromagnetic subsystem, and occurring disturbance variables, such as flow forces in a gas exchange valve. From the system states of the actuator 1 , given by armature position, armature speed and armature acceleration, as well as from the armature target position x_soll, the stroke controller 3 calculates the magnetic force required at the armature, the target magnetic force F _{M_Soll} , for asymptotic from the state space design of the actuator and rapid adjustment of the anchor target position x_soll. This target magnetic force F _{M_soll} is given as the output variable of the stroke controller to the input of a selection element 4 . Since both lifting magnets of the actuator can only generate attractive forces, the target magnetic force must be assigned to a lifting magnet by the selection element 4 . The selection element 4 works according to the following selection criterion:
If the target magnetic force F _{M_soll} <0, then the closer _{magnet is activated} .
If the target magnetic force F _{M_soll} <0, then the opening magnet is activated.
If the target magnetic force F _{M_soll} = 0, then no lifting _{magnet is activated} .

A non-linear compensation _element 5 arranged downstream of the selection _element calculates the required target field current i _{fS_soll} , i _{fO_soll} for the respective lifting _magnet from the target magnetic force F _{M_soll} for each individual lifting _magnet via the corresponding inverse magnetic force _map of the lifting magnets. The target field currents are adjusted via two three-point controllers 61 , 62 , which are arranged after the nonlinear compensation, on the solenoids. At the output of the three-point controller 61, the driving voltage U _S is located on the closing magnet and the output of the three-point controller 62 is the driving voltage U _O of the opening magnet. In the exemplary embodiment described here, the control voltage was -42 V, 0 V or +42 V. Of course, the actuator control is not limited to this control voltage.

The task variable of the two three-point controllers 61 , 62 is in each case the difference between the respective target field current i _{fS_soll} , or i _{fO_soll} and a sum of measured coil current i _{S_mess} , i _{O_mess weighted} with control parameters K _i , K _if and calculated by an observer net coil current i _{fO_beob} , i _{fS_beob} .

For this purpose, the measured coil current i _{O_mess is amplified} on the input side with an adjustable control parameter K _i on the three-point controller 62 , which controls the opening magnet 13 o. A corrected coil current i _{fO_beob} calculated by the nonlinear observer is also _{amplified on the} input side with an adjustable controller parameter K _if . Measured, weighted coil current and calculated, corrected, weighted coil current are summed in a summation element 621 and the sum signal is subtracted from the desired field current in a further summation element 622 .

The same procedure is followed for the three-point controller 61 , which controls the NO magnet 13 S. Here, too, the coil current i _{S_mess} measured in the make _{magnet is} amplified on the controller _input side with an adjustable controller parameter K _i . A corrected coil current i _{fs_beob} calculated by the nonlinear observer for the make magnet is also _amplified on the input side with an adjustable controller parameter K _if . The weighted, measured coil current and the weighted calculated corrected coil current are mationsglied in a Sum summed 611 and the sum signal is subtracted in a further summing circuit 612 from the target field current i _{FS_soll.}

The controller parameters K _i , K _if are preferably implemented as adjustable analog operational amplifiers and enable the controller behavior to be adapted. The practical setting values for the operational amplifiers and thus for the controller parameters K _i . Depending on the purpose of the actuator control, K _if must be determined and set in tests.

Low-pass filters of the 4th order have proven themselves as state variable filters ZVF for the actuator control according to the invention. The structure of such a known low-pass filter is shown by way of example in FIG. 2. The noisy measurement signal x _{mess of} the position sensor 15 is filtered with the filter with the aim of obtaining a noise-free measurement signal. For this purpose, the filter 4 contains integrators 21 , 22 , 23 , 24 connected in series , which integrate the measurement signal x _{mess on} the input side four times in total and thereby provide a largely noise-free filtered position _signal x _ZVF for the position of the magnet armature from FIG. 1. The output of each integrating element is fed back to the input-side summing element 25 with negative feedback. Apart from the position signal x _ZVF the first derivative of the position signal, so the speed signal x _'ZVF is tapped at the filter before the last integrator 24th Preceding the penultimate integrator 23 is the second derivative of the position signal, that is, the acceleration x _"ZVF tapped. All three tapped signals are represented schematically as in Fig. 1 shown in the present invention Aktorregelung used. By selecting the filter parameter f _0, f _1, f ₂ , f ₃ , f ₄ the characteristics of the filter can be changed and set, for a fully variable electromagnetic valve control a filter with a so-called Butterworth characteristic has proven to be particularly advantageous.

Claims (10)

1. Actuator control for an electromagnetic actuator ( 1 ) that works on the principle of a weakly damped spring-mass oscillator and between its two solenoids ( 13 , 13 S, 13 o) there is an armature ( 11 ), driven by the two lifting solenoids Neten and the preloaded springs ( 14 , 14 S, 14 o) of the spring-mass transducer, moved between its two end positions, comprising:
two three-point controllers ( 61 , 62 ) for regulating the field currents in the two lifting magnets,
a transducer ( 15 ) for determining the position of the armature ( 11 ) and for generating a position signal as a measure of the current position of the armature,
a state variable filter ( 2 ) for generating state variables of the armature comprising a filtered armature stroke signal, the armature speed and the armature acceleration,
a stroke controller ( 3 ), which generates the currently required target magnetic force for the stroke magnets ( 13 S, 13 o) from the state variables of the armature and a predetermined target position of the armature,
a selection element ( 4 ) which assigns the desired magnetic force to a lifting magnet ( 13 S, 13 o),
a non-linear compensation ( 5 ), which calculates the desired field current corresponding to the desired magnetic force from the desired magnetic force via a magnetic force map characterizing the lifting magnets ( 13 , 13 S, 13 o),
one current sensor ( 16 S, 16 o) for each of the two solenoids ( 13 S, 13 o) for measuring the current coil currents in the solenoids,
an observer ( 7 ) who calculates a corrected target field current for each of the two solenoids ( 13 S, 13 o) from the armature acceleration and the current measured coil currents,
in which,
the corrected target field currents calculated by the observer ( 7 ), the measured coil currents and the target field currents calculated by the nonlinear compensation are each applied in the form of a signal value as input variables to the two three-point controllers ( 61 , 62 ) and in the three-point controllers an output voltage is generated from these input variables and this output voltage is applied to the solenoids ( 13 S, 13 o). 2. Actuator control according to claim 1, in which the state variable filter ( 2 ) is designed as a 4th order low-pass filter. 3. Actuator control according to claim 1 or 2, in which the state variable filter ( 2 ) with variable filter parameters (f ₀ , f ₁ , f ₂ , f ₃ ) is adjustable to different applications. 4. Actuator control according to claim 1, 2 or 3, in which the state variable filter ( 2 ) is designed as a low-pass filter with Butterworth characteristic. 5. Actuator control according to claim 1, wherein the stroke controller ( 3 ) is designed as a proportional-integral controller (PI controller). 6. Actuator control according to claim 1, wherein the stroke controller ( 3 ) with variable controller parameters (r _i , r ₁ , r ₂ , r ₃ ) is adjustable to different applications. 7. Actuator control according to claim 1 with a fourfold pole position in the target time constant (Tsoll) of the regulated overall system.   8. Actuator control according to claim 7, wherein the time constant (Tsoll) in variable Wei se can be specified. 9. Actuator control according to claim 1, in which the observer ( 7 ) implements a model-based state space in the form of a software program on a microcontroller. 10. Actuator control according to claim 1, in which the stroke controller ( 3 ), the selection element ( 4 ), the state variable filter ( 2 ) and the two three-point controllers ( 61 , 62 ) are designed as analog hardware.