SU526856A1 - System control method with nonlinear correction - Google Patents

Description

one

The invention relates to the field of automation and can be widely used in automatic control systems.

Methods are known in which initial conditions for constructing high-speed processes are taken into account by measuring the velocity and accelerating the output coordinate. 1J However, this formation of a correction signal in order to reduce oscillatory processes in a nonlinear system is not advisable, since the velocity measurement in systems with an output value in the form of a shaft is carried out using tachogenerators and these measurements are significantly distorted by the presence of nonlinear characteristics of the meters themselves (tachogenerator, gearbox, junction a tachogenerator with an object, etc.).

Therefore, it is most effective to reduce oscillation in nonlinear systems to take into account the initial conditions for the system error signal.

A known method for correcting a nonlinear automatic control system (ACS) is based on extracting the difference signal between the system error signal and the smoothed signal of this error and then generating a corrective pulse at the moment the difference signal η arises by summing the correction pulse with the error signal until the signal proportional to the speed from .change signal difference and reference voltage. 2 However, this method of generating a correction signal is not sufficiently effective in high-order systems (3rd and higher). This is due to the fact that the processes caused by the correction signal in these systems are more complex than in systems below the third order, therefore the correction signal formed by comparing the rate of change of sig) {ala difference with the reference signal is not effective enough eliminates undesirable oscillatory processes.

The aim of the invention is to eliminate this drawback, as a result of which the accuracy of non-linear ACS is effectively increased and their range is expanded.

This goal is achieved due to the fact that the duration of the first part of the correction pulse is determined by the time of change of the difference signal to the value of the signal, which with a delay determined by the parameters of the system is formed according to the law that ensures the maximum response rate of the decrease of the difference signal, the slope of this law is inversely proportional to the rate of change of the difference signal and the sign corresponds to the sign of the difference signal, and the moment of switching off another part of the correction signal opposite in sign howl determined time rate of change of the difference signal to a predetermined value, and the sign of the first part of the correction pulse coincides with the sign of the difference signal. FIG. 1 shows a block diagram of a device implementing the proposed method; in fig. Figure 2 shows the time diagrams of the system, where: a is the diagram of the driving curve, a (t); b - diagrams of the dynamic error curve e (0 of a nonlinear system and a linear system smoothed error of a nonlinear system Ел (01; c is a curve diagram of the difference signal between the system error signal and a smoothed signal of this error Ер (/); г is a diagram of the difference signal curve e.skor (About the corrected system, where and is the signal generated according to the law, which provides the maximum response rate for reducing the difference signal with a steepness, back propositional rate of change of the difference signal e (0; t is the delay determined by the parameters d - corrective sign-switching pulses, e is the curve of the signal of the difference sp.cKop (0 corrected system, where f / l is the signal generated according to the linearly increasing law (simplified law); and is the diagram of the dynamic error curve of skor- (0 nonlinear system, corrected by the proposed method. The system contains a driver unit 1, a correction system 2, an object 3, a difference signal generating unit 4, a differentiating link 5, a zero-body 6, a functional unit 7, a delay unit 8, a comparison unit 9, form l 10 corrective imnulsov. The system works as follows. Let, for example, the control object 3 should track the movement of the specifying unit 1 according to the a () law with high accuracy (Fig. 2, a). The system is closed in angle and tracking has a dynamic error E (t) (Fig. 2, b). The magnitude and form of the dynamic error depends on the type of control action and the system parameters. If the system is linear, i.e., there is no backlash, dead zone, etc., then the dynamic error has a smooth EL character (O (Fig. 2, b)) with a smooth control action. However, electromechanical linear systems do not exist in practice, therefore in nonlinear auto-oscillations due to non-linearities of the system are superimposed on the dynamic error. 0 65 scrap of this error is extracted by block 4 for shaping the difference signal and at the moment of occurrence of the difference signal, which is fixed by the null organ, the driver 10 of correction pulses feeds to the input of the correction system 2 sign-impulse Correction (0 (Fig. 2)) which is predominant for the deviation of the control object 3 caused by the oscillatory process. To determine the duration of the first part of the correction pulse, the rate of change of the difference signal is measured using differentiating element 5, and in block 9 of the comparison The difference signal Kp (t} and the signal from the output of the functional unit 7 and T (0 are shown (Fig. 2, d). The functional unit 7 includes a null-body 6 through the delay unit 8 when the sign of the difference signal changes. The delay unit 8 is necessary in order to prevent the negative effect of the correction pulse: overshoot in the case when the slope of the output from the function block 7 is greater than the slope of the difference signal at the moment of its occurrence, i.e. the switch command (or, in the simplified case, disconnection) the correction pulse will not be supplied by the compare unit 9 to the driver 10 correction impulses. The delay value τ is chosen as the minimum permissible for each system and depends on the frequency response of the automatic control system (pulse, but duration equal to the delay τ, should not pass to the output of the automatic control system). At the moment when the signal from the output of the function block 7 becomes equal to the difference signal, the corrector impulse generator 10 changes the sign of the corrective pulse to the opposite one (in the ynponiei version turns off) and turns off the corrective pulse when the signal of the rate of change of the difference signal from the output of differentiator 5 becomes equal given value. Functional block 7 generates a signal of 1 LES according to the law, which ensures the maximum speed of operation of the reduction signal of the difference, and the slope of this law is inversely proportional to the rate of change of the signal of difference. The sign of the signal from the output of the function block 7 and the first part of the correction pulse corresponds to the sign of the difference signal. In the simplified scheme of the device, there are no differentiated link of its link 5 and connections from the differentiating link. The correction pulse in the simplified scheme is formed in: 1 moment of occurrence of the difference signal, which is fixed by the null authority 6. To determine the duration of the correction 1; , 1 in Fig. 2, e). The slope of this signal is determined by the parameters of the system and is adjusted to provide the minimum value of the difference signal of the correction system. The sign of the linearly rising signal and the correction pulse is set equal to the sign of the difference signal.

Obviously, if the initial conditions of the adjustable coordinate contribute to the development of oscillation (the directions of the velocity and acceleration vectors of the control object 3 coincide with the velocity vector and y, the oscillation process), then the difference (oscillation) curve reaches the level of the fixed curve (linearly increasing law, parabola, and . e.) for a longer time than if the initial conditions and counteract the development of oscillations. Therefore, in the first case, the pulse opposing the oscillatory process will be longer than in the second case.

Thus, in the proposed method, the initial conditions of the adjustable coordinate are taken into account when forming a correction pulse.

The law of the curve formed in the functional block 7 is selected for a particular system depending on its structure and parameters. As the order of the system increases, this law becomes more complex.

The introduction of a correction signal by the proposed method allowed to reduce

oscillatory processes in the electromechanical tracking system 3–5 times. A decrease in the vibrational component of the dynamic error by a factor of 4–5 in turn made it possible to increase the Q factor of the system by a factor of 3–4.

Claims (2)

1. Chumakov N. М. Aviation cybernetic systems. Publishing house KVAVU Air Force, Kiev, 1968. with. 234-236. 2. Auth. St. Up to 451047, L1. Cl. G OOB 5/00, 18.12.72. R pcKop ij / 0 (1 (f PA r Pr.oplt) (, Ai soon ()