Method of controlling a low-controllability actuator, and corresponding system

Abstract

A method of controlling an actuator (31) is described. The value of the current position (x actuated ) of the actuator is sensed and the value of a target position (x target ) of the actuator is received. The value of the target position is subtracted (401) from the value of the current position to calculate a positioning error value (e) of the actuator. A value of actual velocity of the positioning error (e spd ) is computed (402) based on the positioning error (e). A value of target velocity of the positioning error (e spd_tgt ) is computed (403) based on the positioning error (e) and a time parameter (T s ) indicative of a target convergence time. The value of actual velocity of the positioning error is subtracted (404) from the value of target velocity of the positioning error to calculate a deviation value of the convergence velocity of the positioning error (Spd err ). The absolute value (e abs ) of the positioning error (e) is calculated (408). If the deviation value of the convergence velocity of the positioning error is not comprised (405, 406, 407) between a lower threshold velocity value (Error spd_thrld Lo) and an upper threshold velocity value (Error spd_thrld Hi) and simultaneously the absolute value (e abs ) of the positioning error (e) is greater (409) than an error threshold value (Error thrld Lo), an enable signal (ControlFlagEn) is asserted; otherwise, the enable signal (ControlFlagEn) is deasserted. If the enable signal is asserted, the current position of the actuator is continuously controlled (342, 344, Act_Dis) to minimize, instant by instant, the positioning error (e). If the enable signal is deasserted, the current position of the actuator is maintained unchanged (345, 344, Act_Dis), instant by instant, with respect to the previous instant.

Description

[0001] “Method of controlling a low-controllability actuator, and corresponding system” ****

[0002] TEXT OF THE DESCRIPTION

[0003] Field of the invention

[0004] The present invention relates to the control of actuators, in particular low-controllability actuators.

[0005] Prior art

[0006] In the field of numerical controls or digital controls, a low- controllability actuator is understood as an actuator whose position, even when controlled via a continuous or quasi-continuous signal (i.e., having very fine or dense quantization or discretization), can only assume a limited number of distinct values, with relatively coarse quantization or discretization (or in any case less fine than the quantization of the control signal).

[0007] Consider for example a continuous system controlled by a continuous or quasi-continuous control signal (e.g., an analog signal or a digital signal with very fine quantization, i.e., a high number of bits). If the system includes a high-controllability actuator (i.e., an actuator whose position can be finely adjusted, with dense quantization - for example, capable of assuming 100 distinct positions between a lower end position 0 and an upper end position 1 , with a granularity of 0.01 ), it can be assumed that in any operating condition the correct working point of the actuator is always reachable within negligible tolerance. If instead the system includes a low-controllability actuator (for example, compared to the previous case, an actuator that can only assume 10 distinct positions between a lower end position 0 and an upper end position 1 , with a granularity of 0.1 ), then it is likely that, at least in some conditions during system operation, the actuator cannot operate at the desired working point within the required tolerance. In such situation, unstable actuator operation may occur, with its position continuously oscillating between the two positions closest to the desired working point (one above and one below it) as qualitatively exemplified in the graph of Figure 1 , which shows the position AD of the actuator (here expressed qualitatively as a percentage) oscillating over time between two values, between which lies the position CWP corresponding to the desired working point.

[0008] Generally, if the controlled system is sufficiently linear, the oscillatory behavior of the actuator should still lead to cancellation (convergence to zero) of the error (i.e., it should allow obtaining an average value of the controlled quantity that, over time, equals the desired value), within negligible tolerance. On the other hand, actuators are often equipped with protection systems that prevent continuous operation at high frequency (as occurs in the oscillatory behavior exemplified in Figure 1 ) to safeguard actuator integrity and prevent premature aging, and thus it is not guaranteed that such error cancellation can actually be achieved. Moreover, limitations may be imposed on the controlled state, and it is possible that, being unable to precisely reach the desired working point, overshoots and / or undershoots of the controlled quantity may occur. For example, the graph in Figure 2 illustrates the time trend of the position XAct of a low-controllability actuator when the target position Xrarget changes over time (i.e., a change in actuator state is required), and also shows (with a dashed line) a possible limit displacement X / waxL / m / i that the actuator may need to respect.

[0009] In some known solutions, the actuator control signal is dampened (or smoothed, filtered) to reduce the actuator oscillation frequency. However, damping (or smoothing) the actuation does not account for the overall desired system performance: it typically involves applying low-pass or hysteresis filters, which are not calibrated with physical values (and thus usually not adapted to the specific application).

[0010] Other known solutions in the field are described in documents US 7891180 B2, JP H08-235606, and US 2023 / 0391301 A1.

[0011] Therefore, there is a need in the art to develop a method of controlling an actuator, particularly a low-controllability actuator, that allows regulating the value of a certain quantity precisely within negligible tolerance, without causing oscillatory actuator behavior at excessively high frequency and avoiding overshoot and / or undershoot phenomena of the controlled quantity.

[0012] Object of the invention

[0013] The object of the invention is to solve the above-mentioned technical problem. In particular, the object of the invention is providing a method (or control algorithm) and system for discretely controlling an actuator, allowing stable achievement of the required regulation performance particularly regarding the convergence velocity of the error, obtaining a compromise between meeting the system’s required performance and limiting the number of actuator oscillations per unit time (avoiding the establishment of high-frequency oscillatory operation and thus safeguarding the component itself from premature aging).

[0014] Summary of the invention

[0015] The object of the invention is achieved by a method having the features forming the subject of the claims that follow, which constitute an integral part of the technical teaching provided herein in relation to the invention.

[0016] The method can be implemented, for example, by one or more electronic control units of a vehicle to control one or more actuators arranged in the vehicle (e.g., a valve).

[0017] One or more embodiments relate to a corresponding system.

[0018] Brief description of the figures

[0019] The invention will now be described with reference to the accompanying figures, provided by way of non-limiting example, in which:

[0020] - Figure 1 , previously described, is a graph qualitatively illustrating the position of an actuator oscillating between two values, between which lies the position corresponding to a desired working point;

[0021] - Figure 2, also previously described, is a graph qualitatively illustrating the trend of the position of an actuator when the actuator’s target position changes over time, and shows a limit curve of the displacement of the actuator;

[0022] - Figure 3 is a block diagram illustrating some components of a system according to one or more embodiments of the present description, including an actuator controlled by a method according to one or more embodiments of the present description;

[0023] - Figure 4 is a block diagram illustrating the steps of a method of controlling an actuator according to one or more embodiments of the present description;

[0024] - Figure 5 is a graph qualitatively illustrating the time trend of physical quantities and signals in a system according to one or more embodiments of the present description, in particular the actual actuator position, the target (or desired) actuator position, the actuator limit position, the actuator command signal, an enable signal for an actuator control logic, and the error of the convergence velocity of the actuator positioning error; and

[0025] - Figure 6 is a graph qualitatively illustrating the time trend of the physical quantities and signals of Figure 5 in a system according to one or more embodiments of the present description, in a different application scenario.

[0026] Detailed description

[0027] As anticipated, the invention relates to control of actuators, particularly low-controllability actuators, and can be applied in a system 30 as exemplified in Figure 3. In particular, the method according to the invention is based on an error convergence velocity calculation algorithm (i.e., an algorithm that determines the velocity at which the difference between the current actuator position and the target actuator position tends to zero), and provides for implementing different actuator control strategies based on the calculated convergence velocity.

[0028] The objective of the control strategy is to improve (e.g., optimize) the ratio between control power (i.e., the impact on the controlled quantity of an added percentage point on the actuator command - e.g., if controlling a pressure with a valve, the control power can be understood as the pressure variation obtained by changing the valve opening by one percentage point) and actuator movements and, if present, ensure that the upper and / or lower limits of the physical quantity whose value is determined by the actuator position are respected. Since it is not always possible to predict the resulting control power as a function of actuator position variation, the criterion underlying the control method according to the present invention is to move the actuator only when the performance in terms of error convergence velocity of the actuator is unsatisfactory (i.e., when the difference between the current error convergence velocity and the target error convergence velocity is outside a certain desired value range). To do this, the control algorithm substantially decides when to control the actuator continuously (conventionally) and when instead to “freeze” the actuator position, maintaining at each iteration of the control cycle the position from the previous cycle.

[0029] In particular, system 30 includes a (low-controllability) actuator 31 that acts on a component 32 to control a certain variable or physical quantity thereof. For example, the actuator may be an electronically-controlled valve that determines the pressure of a fluid within component 32, or an electronically-controlled actuator that controls the position of component 32. One or more sensors 33 are operatively coupled to component 32 (e.g., mounted on or within it) to detect a certain physical quantity whose value is controlled by actuation of actuator 31 (e.g., one or more pressure sensors, in case actuator 31 is a valve). System 30 further includes an electronic control unit 34 (e.g., a vehicle control unit) that receives data detected by sensor(s) 33 and, based on them, produces the control signal for actuator 31 according to the logic described in the following.

[0030] In particular, control unit 34 includes a logic 341 that calculates the positioning error e of the actuator (i.e. , the difference between the current position Xactuated and the target position xtarget of actuator 31 : e = Xactuated - xtarget), the actual velocity of the positioning error esPd, defined as the first derivative of the positioning error e with respect to time (esPd = e or esPd = de / dt according to another notation), the target velocity of the positioning error esPd_tgt defined as the ratio between the positioning error e and a parameter Ts(having the dimensions of a time unit, and optionally calibratable or adjustable) representing the target convergence time of the positioning error (esPd_tgt = e / Ts), and the error (or deviation) of the convergence velocity Spderr defined as the difference between the target velocity of the positioning error esPd_tgt and the actual velocity of the positioning error esPd (Spderr = esPd_tgt - esPd). Logic 341 outputs the value of the positioning error e and the value of the convergence velocity deviation Spderr.

[0031] As exemplified in Figure 3, control unit 34 includes a logic 342 coupled to the output of logic 341 and functioning substantially as a conventional continuous control logic, i.e., determining at each iteration of the algorithm a required displacement for actuator 31 so that its current position Xactuated approaches the target position Xtarget (i.e., substantially trying to minimize the actuation error e). Moreover, control unit 34 includes a logic 343 that, based on the value of the actuation error e and the value of the convergence velocity error Spderr, asserts or deasserts an enable signal (flag) ControlFlagEn, as further described in the following, which controls a selector 344. When signal ControlFlagEn is asserted, selector 344 provides to actuator 31 , as command signal Act_Dis, a target position from the conventional continuous control logic 342, while when signal ControlFlagEn is deasserted, selector 344 provides to actuator 31 , as command signal Act_Dis, the target position stored at the previous iteration in a memory element 345, i.e., substantially “freezes” the movement of actuator 31 .

[0032] Further details regarding the operation of logics 341 and 343 are described with reference to the block diagram in Figure 5. In a step 401 , logic 341 receives the values Xactuated and xtarget and calculates the difference between them to determine the actuation error e = Xactuated - xtarget. In a step 402, logic 341 receives the value of the actuation error e and calculates its first time derivative, i.e., e = de / dt, to determine the actual velocity of the actuation error esPd. In a step 403, logic 341 receives the value of the actuation error e and the value of a parameter Ts(e.g., calibratable) indicative of the target convergence time, and calculates the target velocity of the actuation error esPd_tgt, i.e., the ratio esPd_tgt = e / Ts. In a step 404, logic 341 receives the target velocity of the actuation error esPd_tgt and the actual velocity of the actuation error esPd and calculates the difference between them to determine the convergence velocity error Spderr. In a step 405, logic 343 receives the value Spderr, compares it with an upper threshold value ErrorsPd_thridHi, asserts a first activation signal if the value Spderr is greater than the upper threshold value ErrorsPd_thridHi, and deasserts the first activation signal if the value Spderr is less than the upper threshold value ErrorsPd_thridHi. In a step 406, logic 343 receives the value Spderr, compares it with a lower threshold value ErrorsPd_thridLo, asserts a second activation signal if the value Spderr is less than the lower threshold value ErrorsPd_thridLo, and deasserts the second activation signal if the value Spderr is greater than the lower threshold value ErrorsPd_thridLo. In a step 407, an OR logic gate asserts a respective output signal when at least one of the activation signals output from blocks 405 and 406 is asserted. In a step 408, logic 343 receives the value of the actuation error e and calculates its absolute value eabs. In a step 409, logic 343 receives the value eabs, compares it with a further threshold value ErrorthridLo, asserts a third activation signal if the value eabs is greater than the threshold value ErrorthridLo, and deasserts the third activation signal if the value eabs is less than the threshold value ErrorthridLo. In a step 410, an AND logic gate asserts a respective output signal when both activation signals output from the OR logic gate 407 and from block 409 are asserted. The output signal of the AND logic gate 410 corresponds to the enable signal (flag) ControlFlagEn.

[0033] Thus, substantially, method 30 allows controlling the actuation of actuator 31 according to the continuous (conventional) logic 342 when the value of the convergence velocity error Spderr is greater than a certain upper threshold ErrorsPd_thridHi (optionally adjustable, settable) or less than a certain lower threshold ErrorSpd_thridLo (also optionally adjustable, settable), while at the same time the absolute value of the error eabs is greater than a certain threshold ErrorthridLo. On the other hand, method 30 “freezes” the position of actuator 31 (i.e. , prevents its movement) when the absolute value of the error eabs is less than the threshold ErrorthridLo or when the value of the convergence velocity error Spderr is between the thresholds Error sPd_thridHi and Error sPd_thridLo.

[0034] Deasserting the signal ControlFlagEn (and thus “freezing” the position of actuator 31 ) when the absolute value of the error eabs is less than the threshold ErrorthridLo advantageously reduces or completely avoids system instabilities around the zero-crossing point (i.e., avoids frequent switching between the continuous control algorithm and the “freezing” condition when the actuator positioning error is, in absolute value, very small).

[0035] Deasserting the signal ControlFlagEn (and thus “freezing” the position of actuator 31 ) when the value of the convergence velocity error Spderr is between the thresholds ErrorsPd_thridHi and ErrorsPd_thridLo advantageously avoids an oscillation condition between overshoot and undershoot, “turning off” the continuous control of the actuator when the convergence velocity is satisfactory, i.e., within the range defined by the two thresholds. Moreover, it should be noted that since the target velocity of the controller is defined as the ratio between the actual error and the target convergence time, the velocity to be tracked becomes smaller as the controlled quantity approaches its target value. This allows obtaining stable behavior using a fixed band. It should also be noted that if the convergence velocity error is within the allowed band (which can be appropriately calibrated based on system characteristics), the system remains in a convergence zone with satisfactory performance and maintains a margin of controllability of the continuous system when the velocity error exits the allowed band.

[0036] To adapt the control strategy of actuator 31 based on the properties and / or characteristics of the dynamic system in which actuator 31 operates (e.g., component 32 or the system in which component 32 is incorporated), and / or based on the desired control performance, the parameters ErrorspdjhridHi, ErrorspdjhridLo, ErrortMdLo and Ts can be adjusted (e.g., during design or subsequently). Setting the value of the threshold ErrorspdjhridHi determines the upper value of the convergence velocity error Spderr that activates the continuous control logic 342. Setting the value of the threshold ErrorspdjhridLo determines the lower value of the convergence velocity error Spderr that activates the continuous control logic 342. Setting the value of the threshold ErrorthridLo determines the threshold value of the absolute error eabs below which the position of actuator 31 is frozen. Setting the value of the parameter Tsdetermines the desired error convergence velocity. Thus, by setting the values of these parameters, the overall behavior of the control system can be adjusted. For example, reducing the target convergence time Tsand / or narrowing the tolerance interval defined between the thresholds ErrorspdjhridHi and ErrorspdjhridLo increases the dynamic response of the control system.

[0037] Figure 5 includes graphs exemplifying the time trend of some signals and physical quantities of the control system in a first case. In particular, the trend of signal Xrarget (target or desired position of the actuator) is shown, which in this case is set equal to the limit XMaxLim, the trend of the actual actuator position XActuator, the actuator control signal Act_Dis, the binary signal ControlFlagEn enabling the continuous control logic 342, the value of the convergence velocity error Spderr, and the values of the thresholds ErrorspdjhridHi and ErrorspdjhddLo. In the example of Figure 5, there are two distinct time intervals in which the convergence velocity error Spderr is greater than the upper threshold ErrorspdjhridHi a first interval between to and ti, and a second interval between t2 and ts. To avoid an overshoot of the target, the method parameters can be adjusted based on the properties of the system in which actuator 31 is integrated. In general, when it is possible (and convenient) to tolerate slight oscillation of the controlled quantity between a minimum and a maximum value, the behavior exemplified in Figure 6 can be obtained, depending on the parameters used and the chosen continuous control algorithm 342.

[0038] The actuator control method described here thus allows increasing control performance without having to use a more precise (and therefore more expensive) actuator, while also reducing component aging. Of course, implementation details and embodiments may vary widely from what has been described and illustrated without thereby departing from the scope of the invention as defined by the attached claims.

Claims

CLAIMS1. A method of controlling an actuator (31 ), the method comprising:- sensing the value of the current position (xactuated) of said actuator (31 ) and receiving the value of a target position (xtarget) of said actuator (31 );- subtracting (401 ) the value of the target position (xtarget) of the actuator (31 ) from the value of the current position (xactuated) of the actuator (31 ) to compute a value of the positioning error (e) of the actuator (31 );- computing (402) an actual value of velocity of the positioning error (esPd) based on the positioning error (e) of the actuator (31 );- computing (403) a target value of velocity of the positioning error (eSpd_tgt) based on the positioning error (e) of the actuator (31 ) and on a time parameter ( Ts) indicative of a target convergence time;- subtracting (404) the actual value of velocity of the positioning error (esPd) from the target value of velocity of the positioning error (espdjgt) to compute a difference value of the convergence velocity of the positioning error (Spderr)',- computing (408) the absolute value (eabs) of the positioning error (e) of the actuator (31 );- in response to the difference value of the convergence velocity of the positioning error (Spderr) not being comprised (405, 406, 407) between a lower threshold velocity value (ErrorsPd_thridLo) and an upper threshold velocity value (ErrorsPd_thridHi) and simultaneously the absolute value (eabs) of the positioning error (e) of the actuator (31 ) being higher (409) than a threshold error value (ErrorthridLo), asserting an enable signal (ControlFlagEn)’,- in response to the difference value of the convergence velocity of the positioning error (Spderr) being comprised (405, 406, 407) between the lower threshold velocity value (ErrorsPd_thridLo) and the upper threshold velocity value (ErrorSpd_thridHi), or the absolute value (eabs) of the positioning error (e) of the actuator (31 ) being lower (409) than the threshold error value (ErrorthridLo), deasserting said enable signal (ControlFlagEn)',- in response to said enable signal ( ControlFlagEn) being asserted, providing (344) to the actuator (31 ), as a drive signal (Act_Dis), a signal thatcontinuously controls (342) the current position (xactuated) of the actuator (31 ) to minimize, at each control iteration, the positioning error (e) of the actuator (31 ); and- in response to said enable signal (ControlFlagEn) being deasserted, providing (344) to the actuator (31 ), as a drive signal (Act_Dis), a signal that maintains unchanged (345) the current position (xactuated) of the actuator (31 ), at each control iteration, with respect to the previous control iteration.

2. The method of claim 1 , wherein said step of computing (402) the actual value of velocity of the positioning error (esPd) comprises computing the first time derivative of the positioning error (e) of the actuator (31 ).

3. The method of claim 1 or claim 2, wherein said step of computing (403) the target value of velocity of the positioning error (esPd_tgt) comprises computing the ratio between the positioning error (e) of the actuator (31 ) and the time parameter (Ts) indicative of the target convergence time.

4. A system (30) comprising:- an actuator (31 ) activatable by a drive signal (Act_Dis) to control an operational physical quantity of a component (32) of said system (30);- one or more sensors (33) configured to sense one or more values of said operational physical quantity;- an electronic control unit (34) configured to receive said one or more values of said operational physical quantity from said one or more sensors (33) and configured to produce said drive signal (Act_Dis) for the actuator (31 ) according to the method of any of the previous claims.

Images