SU987575A1 - Speed-wise quasioptimal control system - Google Patents

Description

The disadvantages of this system are low stability and speed. Closest to the proposed technical entity is a quasi-optimal speed control system, containing a sequentially connected master, comparison, shaper optimality criterion, first smoothing filter, first multiplication unit, first adder, amplifier, second block; Multiplication and integrator block, output which is connected to the control input, the controller, the first inputs of the auxiliary operator blocks are connected to the output of the setpoint generator, and the outputs through the corresponding third multiplication blocks - to the corresponding inputs of the second multiplication unit and inputs of the respective quadrants, the outlets of which are connected to the corresponding inputs of the second with a telmator, the output of which is connected to the second inputs of the third through the second smoothing filter to the second input of the first multiplication unit multipliers, and through the regulator to the input of the control object, the output of which is connected to the second input of the comparator unit, the output of the first smoothing filter through the differentiator One is connected to the second input of the first adder. The disadvantage of the known system is a low level of organization and speed. The aim of the invention is to increase the speed and responsiveness of the system. The goal is achieved by the fact that in the system the output of the block of integrators is connected to the second inputs of the blocks of auxiliary operators. In FIG. 1 is a graph of the transition time; FIG. 2 is a functional diagram of a speed-optimized control system. The control system contains a comparison unit 1, a controller 2, a control object 3, an integrator unit 4, an optimal tee generator 5, a first smoothing filter b, a first multiplication unit 7, a first adder 8, an amplifier 9, a second multiplication unit 10, a unit 11 forming a partial derivative, differentiator 12, blocks of 13 auxiliary operators, third multiplication blocks 14, quadrants 15, second adder 16, second smoothing filter 17 and setting unit 18, Consider the performance of the control system with respect to speed, From the theory of self-tuning from systems it is known that the full derivative of the optimality criterion Q depends on the rate of change of the adjustable parameters oL- (i 1, 2, ...,) and the rate of change of the non-adjustable parameters | ij (i 1, 2, .., m, as well as on the characteristics input signal and can be represented in the form: aa 3 "t, jHL. (, Pj t From (1) it is clear that the requirement is natural, and O to ensure the normal functioning of the tuning loop and the robustness of the entire system with gradient-hype search procedure) where const О, а-§, - component, caused by the influence of the input signal and is not taken into account perturbation factors). To ensure it is necessary that. rj apr fr; 3d / dG-t Pj, since the components in the left and right parts have different signs, If i and (3 POSITION, J.rO, is determined only by the input signal x (-b), the contour problem settings in this case - compensation of the dynamic error. Increasing the speed and value of the component on the left, part of equation (3) can provide a small dependence of -4x- on the right side, and hence the dependence of the transient on the characteristics of the input signal xCtJ, Proposed The technical solution solves the problem of providing a fast-optimal quay optimal performance. The transient process is aperiodic in the class of gradient self-adjusting systems. We require that the equality a - // ly "a, (4) a 0.5 e, where e x - y, y is the output value of the control system. We define the value of A in the equation (2) such that “the condition L-1“ () (5)

If the expression for D from (5) is substituted into (1), then we get:

, ,one

before

(fe)

at

l + uJlAQlr 1 K || daG

where K is the gain selected from the stability conditions

From scaling L) it is clear that at the moment of a sharp change in the input signal, a. means. and the value of the criteria Q ,. which will cause significant values of the gradient norm l / iQ || 2, the value of the total derivative Q will be:.

and when working out the mismatch e X - y, the magnitude is a O, and therefore also the lv is a small quantity, which will provide the value Q - fcj | d. .one%)

From (7) and (8) it follows that c. the moment of occurrence of the mismatch, the algorithm for adjusting the value A according to the law C 5) will ensure a sharp increase in the intensity of the adjustment of the coefficients and then, as the error develops, the rate of change in оС; tends to zero, to implement the law of the process of building close (CL4).

The structure of the speed-optimal control system in the class of gradient structures without overshoot implements equations C2) and (5) (Fig. 2).

The system works as follows. . .

The signal comes from the output of the set 18 to the input of the comparison unit 1, passes through the controller 2 and the control object 3, from whose output the output signal y is compared in the comparison unit 1 with the signal x, forming the error signal of the system e X - y. Simultaneously, the signal x is fed to the input of blocks 13 with

Noah function where Ф (ЦА;, / Ь, -) oo-i

the transfer function of the main control loop, consisting of blocks 1-3.

In blocks 14 of multiplication, the signals from the outputs of blocks 13 are multiplied by the values of §§ received at the second inputs of blocks 14 from the output of block 11 of the formation of the partial derivative. From the outputs of the multiplication unit 14, the signals arrive at the corresponding inputs of the multiplication unit 10, and through the corresponding quadrants and inputs of the adder 16, from the output of which the resulting signal | (dQII, which is smoothed in the filter 17) is obtained.

The multiplication unit 10 in the general case, when adjusting n parameters, represents n multipliers, in each of which the signal from the outputs of blocks 13 is multiplied by a signal from

the output of amplifier 9 with a gain of K. Blocks 5, b, 7, 8, and 12 form an error signal from the signal; system bk is the value | ft "t; au fl + Q .. Error signal after passing through the driver

5, the optimality criterion a -0.5e, which is a quad with a gain of 0.5, is fed through a smoothing filter 6 to multiplication unit 7, which is multiplied by IffIQI from the output of filter 17, and then to the input of adder 8. The signal from the output of the filter b is fed to the differentiator 12, from the output of which the signal Q is supplied

to the input of the adder 8. The output signal of the adder 8 is fed to the inputs n of the multipliers of the multiplication unit 10. From each multiplier, the signal arrives at the input (corresponding n integrators of block 4 of integrators / signals from which the signal comes to adjust the parameters otj in controller 2 and blocks 13.

In this way,

setting parameters ot .; with the aim of

ensuring a minimum of time. the flow process T in the class of kwai gradient structures by aperiodic law, i.e. the formation of transients of type 1 (Fig. 2-), and

not P and lit. .

The application of the invention makes it possible to increase the accuracy of the control system by increasing the speed of the tuning loop six times, to ensure aperiodic adjustment of the parameters, which will affect the control loop and, therefore, the control system. System can be with

It has been successfully applied in flow control systems where it is necessary to provide (fast aperiodic aromatic procedures).

50

Claims (3)

1.Smolnikov L.P. Synthesis of quasi-minimal automatic control systems. L., Energie, 1967, with. 23-24. 2. Boychuk L.M. The method of structural synthesis of nonlinear automatic control systems. M., Energie, 1971, p. 16-17. - 3. USSR author's certificate in accordance with the application 3002951 / 18-24, cl. G 05 13/00, 1980 (prototype).