FR3035936A1 - METHOD FOR CONTROLLING THE POSITION OF A CLUTCH CONTROL MEMBER - Google Patents

Abstract

Method for controlling the movement of a clutch control element (4) by an irreversible actuating system consisting of an electric actuating motor (1) connected by mechanical transmission means (2, 3) to clutch control element, by sending a reference voltage electric motor (Uref), which allows each moment to move the control element, from an initial position (x) to a reference position ( xref), characterized in that the reference voltage (Uref) is obtained by integrating a calculated value of its own derivative in time.

Description

FIELD OF THE INVENTION The present invention relates to the control of robotized gearboxes, equipped with at least one clutch for interrupting the transmission of torque between a motor. drive and the gearshift mechanism of the gearbox. More specifically, it relates to a method for controlling the movement of a clutch control element by an irreversible actuation system consisting of an electric actuation motor connected by mechanical transmission means to the element. clutch control.

This method is based on sending to the electric motor, a reference voltage, which allows each moment to move the control element from an initial position to a reference position. In robotized gearboxes, clutch control can be provided by an electric or hydraulic control system controlled by a computer. FIG. 1 non-restrictively illustrates the irreversible actuating system, consisting of an electric actuating motor 1 and a motion transmission system composed of a set of gears (housed under a casing 2), which transmit the movement of the electric motor on a screw-nut system 3 which converts the rotational movement of the motor 1 into a translation movement of a clutch control cable 4, which "pulls" on a clutch lever ( not shown). Finally, an assistance spring 6 stores energy at closing the clutch, and restores it by helping to open the clutch. In order to ensure a controlled transmission of the torque of the heat engine to the gearbox, the position of the cable must be controlled with an accuracy of the order of 0.1 mm. However, the peculiarity of such an actuating system is its irreversibility, linked to the high level of friction between the screw and the nut. This feature keeps the system in position, in case of failure of the control of the electric motor (voltage drop for example). But it poses a problem of control, because its control is not linear: as long as the electric motor does not provide a level of effort superior to the friction, the system remains immobile, but once this level of effort exceeded, the system can get moving very quickly.

The level of friction can reach 70 to 80% of the maximum effort of the electric motor, and can vary depending on the wear of the clutch, as indicated by way of example, on the curves of FIG. This figure makes it possible to compare the force curves (in Nm) on the stroke of the clutch abutment (in millimeters), in the direction of its opening (disengaging) and its closing (clutch) for a clutch. new condition (curves in dashed lines) and for worn clutch (curves in solid lines). It is found that 20 provide with a clutch sense of closure because of the return spring, the clutch in the new state. When the maximum additional effort is sought to be used, it is more important to compensate in addition for the wear which facilitates the closing of the transmission of the torque of the engine to the gearbox, no overrun of the position of the engine. cable is not eligible. This requirement makes the use of conventional control techniques (PID type) unsuitable for controlling the position of the cable. An object of the present invention is to move the clutch cable in a completely transparent manner to the driver during clutch and clutch operations. The desired accuracy is 0.005 mm. It must be respected, without exceeding the position of the cable, nor know the effort that must exert the electric motor to overcome the friction. In addition to these mechanical constraints, the control must respect thermal constraints. To ensure the durability of the electrical system, it must in particular avoid overly disordered or sudden stresses of the motor 1, responding to repeated voltage jumps, or the passage of excessive currents. Finally, at the end of the closing of the clutch, the "licking" phase, where the torque of the heat engine is progressively transmitted to the gearbox, must be as transparent as possible, to avoid the jerking. couple. The movement of the clutch cable must therefore be controlled with an accuracy of 0.005, mainly during the licking phase, and without exceeding. The control must also be as continuous as possible, avoiding strong long current draws, in the motor. The control input is generally a reference voltage of the actuating motor. The known calculation methods for the reference voltage aim at minimizing at every moment the difference between the actual position of the control element of the clutch and the reference position which it must reach. However, they most often generate a discontinuous voltage command, which has the drawback of strongly biasing the actuator, as this may cause the electric motor to overheat. In summary, the main difficulties to overcome to control the movement of a control element such as a clutch cable by an irreversible actuating system consisting of an electric motor connected by mechanical transmission means to the control clutch, are: - to overcome the friction which is non-linear, - the desired static precision, which is 0.005 mm, - to ensure small displacements (of the order of 0.1 mm) without exceeding the position, for eliminate the main sources of torque surges in the clutch licking phase, and - limit the risk of overheating or breakage of the electrical system.

The present invention aims to overcome these difficulties, to control the position of a robotized box clutch. For this purpose, it proposes that the reference voltage be obtained by integrating a calculated value of its own derivative over time. Preferably, the derivative of the reference voltage includes a first term of difference between the state and the position setpoint of the control element, a second term of difference between the state and the setpoint of its speed, and a third term of difference between the state and the setpoint of the derivative of its speed. Integrating the first gap term gives a compensation term representing the resistive torque of the system; integrating the second gap term gives a proportional term on a position error; the integration of the third term gives a proportional term on a speed error. The present invention will be better understood on reading the following description of a non-limiting embodiment thereof, with reference to the accompanying drawings, in which: FIG. 1 represents an electric clutch control system. FIG. 2 shows the deviations of forces involved in a clutch opening and closing stroke according to its state, and FIG. 3 shows position, control voltage and motor current curves. when closing a clutch. An electric motor such as that illustrated in FIG. 1 is represented by the following equations: Ldl = brrel-kW-R /: electrical equation of the motor dt 5 = T-Td = kl-Td: mechanical equation of displacement of dt the clutch, with cc): Regime of the electric motor, T: Torque of the electric motor, 10 Td: Resistant torque of the electric motor (unknown exogenous input, regarded as a disturbance) J: Inertia of the electric motor; /: Current of the armature of the electric motor; k: Coefficient of the voltage induced by the electric motor; R: Resistance of the armature of the electric motor; L: Inductance of the armature of the electric motor; tinf: Reference voltage that we want to apply to the terminals of the electric motor, it is between U'n and [Fax 20 (typically, [Fax will be the voltage of the on-board network); The resisting torque of the electric motor Td = f (x) is a function of the position x of the clutch cable. It depends on the force applied by the screw-nut system, the force of the return spring, and the force applied by the diaphragm of the clutch. This effort also varies according to the degree of wear of the clutch. Different sensors measure the position of the cable x, which is between two stops (open and closed position), as well as the speed -dx = V of a point of the clutch cable. A current sensor 30 is also available to provide a measurement of the armature current of the DC motor. The proposed control strategy is to move the clutch cable from an initial position x to a reference position xref, to satisfy the conditions of a perfect shift in the gearbox, while ensuring the maintenance of the electrical system. The cable displacement control input is the reference voltage of the electric motor Uref which provides this displacement. From the two equations above, the system with respect to the position to be regulated can be written: / 0 1 0 \ kx 10 0 0 7 (to + 0 -1 (Urn 19), 1 Tà I k R \ In the known regulator synthesis methods, for example, a proportional type regulator (PID) is used, in which case such a regulator is not optimal with respect to the orders of magnitude of the system, because the resisting torque to be compensated can be very large and imposes in these conditions gains too important.To obtain a more continuous control, the invention proposes to take into consideration an additional physical quantity, such as the derivative The system can then be written in the following way: - (o 1 ooo` () c 'oooo () c 00 0 - 0 0 0 -1 0 - - 20 v 0 0 0 0 -kv + 0 0 0 1 oooo J 000 o - oo - k 0) 0 - 0 0 LL With such an extended system, the main control becomes the derivative of the voltage to be applied to the According to the invention, the reference voltage Uref is then obtained by integrating a calculated value of its own derivative in time. This ensures a greater continuity of the control synthesized by the regulator. In addition, the disturbance to be compensated is no longer 3035936 - 7 - Uref, but its derivative, which has the advantage of reaching zero when stopping moving the clutch. We take advantage of the property of the function Td = f (x), which can not vary at constant x.

To use this extended system, one preferably chooses a method that takes into account the gains on the differences between the states and the setpoints on the position of the cable x -x'f, its speed v - v'f. The derivative of the reference voltage [Tref then preferably includes a first term of difference between the state and the position reference of the control element (x'f - x), a second term of difference between the state and the setpoint of its velocity (v'f - v), and a third term of difference between the state and the setpoint of the derivative of its velocity (19 "f -1)). The first deviation term gives a compensation term representing the resistive torque of the system, the integration of the second discrepancy term gives a proportional term on a position error, and the integration of the third term gives a proportional term on an error. The reference voltage [Tref is then equal to the sum of the first three proportional terms, and a fourth additional term proportional to the current I. The Lyapunov theory makes it possible, for example, to produce a regulator that meets these criteria. , and which presents in addition the advantage of ir a single setting parameter, which greatly simplifies the debugging phase (MAP) of the regulator. This theory is therefore particularly appropriate. However, without departing from the scope of the invention, other powerful calculation methods may be used, provided that all deviation terms and proportional terms depend only on a single adjustment parameter. By naming À, the tuning parameter in the Lyapunov equation, the following regulator is obtained on the derivative of the control voltage, in which gains over the position, velocity and acceleration (vref_ i)), depend on the single parameter X: - JL uref = (Â2 + (1+ Â2) 2) (xref _ + (3 + 4A.2) (Yl'ef - 1)) - LIÂT I) By integration, we deduce the voltage command [Tref to be applied: uref = L (λ2 + (1 + λ2) 2) f (xref x) Â (7 22) (xref x) + (3 + 4A.2 In this command, the first term is proportional to the integral of the positional deviation. It makes it possible to compensate for the resisting torque of the system Id (whether it is in the worn or new state). The second term is proportional to the position error. The third term is proportional to a speed error. The last term, referred to as feed-forward, serves to increase the current dynamics. In summary, the method makes it possible to control the movement of a clutch control element by an irreversible actuation system composed of an electric motor 20 actuated by mechanical transmission means to the control element of the clutch. clutch, by sending to the electric motor a reference voltage Uref, which allows each moment to move the control element, from an initial position x to a reference position x, f.

The various stages of calculation of the command at time t are for example the following: a) We receive the set of measurements and setpoints: I (t), x (t), xref (t) and v ( t), b) The speed reference trajectory 30 is calculated by deriving the reference position, vref (t) = xre f (t) xref (t-1). x) + 2L (7 + L + 22L2) (Yr el - y) 3035936 - 9 - This choice of reference trajectory calculation is simple, and has different advantages. Firstly, when the position set varies abruptly, we have a speed pulse vref (t), which means that we aim instantaneously at a high speed which vanishes immediately after, if the position setpoint does not vary. , thus limiting the risk of overtaking. This is the case during the first phase of clutch closure, between t = 0.1 and t = 0.15 on the first curve of FIG. 3. The second advantage appears when xref (t) varies slightly for example in the clutch licking phase that follows. In this case, we have a velocity pulse vref (t) which makes it possible to follow exactly the trajectory corresponding to xref (t). If xref (t) varies linearly, we have a constant target speed 15 which makes it possible to better follow the evolution of the position reference. c) The integral control term is incremented: Are / int (t) = 3'ref int (t-1) + (xref (t) x (t)) - This integral term serves to compensate the evolution of the torque 20 resistant electric motor. It is essential to separate the compensation from this torque, with the torque necessary for tracking. Since the resistive torque is a function of the position (and not of the speed), the position error must be integrated to best compensate for this term. The torque necessary to maintain the speed of the motor can be calculated by the loop on the speed error, as follows: d) We update the errors in position and speed: Arefx (t) = xref (t) ) - x (t) and Are '"v (t) = Vref. (t) - v (t) 30 e) We calculate the new command to be applied: Uref (t) = IL (À.2 + (1 + À2) 2) relint (t) À (7 + À + 2À2) Aref x (t) + (3 + 4A2) Aref v (t) - 4A -k1 (t) 3035936 - 10 - The setting of the parameter A is obtained by a coherence of the orders of magnitude of the terms of the equation of the Step 5. By setting parameter λ to 50, the implementation of this command on a clutch system gives the performance illustrated by way of example by the curves of FIG. 3. On the first graph, the instruction of position of the clutch is in broken line, and the position measured in solid line. The case presented is a case of licking the clutch where it is necessary to be able to regulate as finely as possible the position to ensure the most comfortable passage for the driver. The rapid response of the actuator can be seen when the set point suddenly rises from 0 to 4.5 mm to start the licking and then to a follow-up within 0.1 mm of the set point, when this varies around the licking point. The Uref command applied for this type of displacement is represented on the second graph, for a case of available voltage of 10V. We see the fast transition to 10V to respond quickly to the set point. During the monitoring phase, very little voltage is used and very little current as can be seen in the last graph which avoids the heating of the motor and the power electronics during the licking phase. Since this can last for several seconds, the regulator obtained has the objective of limiting the heating of the control device. This calculation uses only one setting parameter. It offers the advantage of always having the right ratio between the evolution of the different terms of the equation. The three control terms taking into account the position and velocity errors are thus automatically brought into coherence, if it is desired to vary the movement dynamics of the actuator. This makes it possible to always ensure the separation of the calculations from the trajectory tracking and resistive torque compensation terms. If it is desired to increase the movement dynamics of the actuator, it is possible, for example, to increase the value of the parameter λ. This increases the dynamics of evolution of the error terms in position and speed. The dynamics of evolution of the integral term is also increased to the point of fairness, to overcome the fact that, if the actuator moves faster, the resistive torque can also evolve more quickly. By keeping a great stability of regulation over a wide range of value of the parameter λ, it is possible to overcome the constraints of stability for the calibration of the regulator. Under these conditions, the only criterion remaining to be optimized is the level of current and voltage used during the licking phase. The objective is then to calibrate the term as low as possible while checking the fact of always staying in a range of +/- 0.005mm around the setpoint. The objective may be, for example, to never exceed 10A in the engine during licking phases, which can last up to several seconds. A suitable method is to start by setting the term il to a high value, and then to decrease this value until it is possible to respect the stress on the current used during the licking phase. In conclusion, the synthesis of the regulator according to the Lyapunov theory offers appreciable advantages for the development. However, without departing from the scope of the invention, other theories can be applied, in that they make it possible to synthesize a corrector for a system extended to the derivative of the control voltage.

Claims (7)

REVENDICATIONS1. Method for controlling the displacement of a clutch control element (4) by an irreversible actuating system consisting of an electric actuating motor (1) connected by mechanical transmission means (2, 3) to the clutch control element (4), by sending a reference voltage (U'ef) to the electric motor, which enables the control element (4) to be displaced at any moment from a position initial (x) at a reference position (x ,, f), characterized in that the reference voltage (Uref) is obtained by integrating a calculated value of its own derivative in time. 2. Control method according to claim 1, characterized in that the derivative of the reference voltage (Uref) includes a first term of difference between the state and the position reference of the control element (x_ref - x), a second term of difference between the state and the setpoint of its speed (vref - v), and a third term of difference between the state and the setpoint of the derivative of its speed (19 "f -1.9). 3. Control method according to claim 1 or 2, characterized in that the integration of the first discrepancy term gives a compensation term representing the resistive torque of the system, the integration of the second term of difference gives a term proportional to a position error, and the integration of the third term gives a proportional term on a speed error. 4. Control method according to claim 3, characterized in that the reference voltage (U, f) is equal to the sum of the first three proportional terms and a fourth additional term proportional to the current (/). 5. Control method according to claim 2, 3 or 4, characterized in that all the deviation terms and the proportional terms depend only on a single adjustment parameter (X). 3035936 - 13 - 6. Control method according to one of the preceding claims, characterized in that it comprises the following steps: - one receives all the measurements and instructions of current, reference position position and speed: / ( t), x (t), xref (t) and vM, the velocity reference trajectory is calculated by deriving the reference position, ie vref (t) = xref (t) - xref (t -1), - we increment the integral control term: Are '' int (t) =, Arefint (t - (xref (t) - X (0), and - we update the errors in position and speed: Arefx (t) = xref (t) - x (t) and Arefv (t) = vref (t) - v (t). 7. Control method according to claim 6, characterized in that it comprises the following step of calculating the new control voltage (Uref) (t) to be applied: JL Uref. (t) = -k (A2 + (1 + A2) 2) relint (t) to (7 + to + 2A2) Arefx (t) k) + (3 + 4A2) ev (t) - 4A-1 (t )