SU1513416A1 - Adaptive control system for potentially hazardous object - Google Patents

Abstract

The invention relates to the automatic control and protection of potentially hazardous objects operating under conditions of uncontrolled disturbances, and can be used in the chemical, food and other industries. The aim of the invention is to increase the reliability and accuracy of the system under the conditions of an uncontrolled disturbance on an object. The system contains a main control loop 5, a protection block 10, a block 18 for estimating statistical parameters of a controlled variable, a block 19 for evaluating statistical parameters of a random and periodic components, a computing block 20, a block 26 for correcting a threshold set value for a constant component, a block 34 for correcting dynamic values system properties, adders 22, 35, optimizer 6. The system compensates for the delay and inertia in the main control loop, corrects the set value of the adjustable coordinate and accurately estimates The likelihood of trouble-free operation, which increases the reliability and efficiency of the system. 2 hp ff, 1 ill.

Description

(L

ate

00 4

ABOUT

3isn

 control tour 5, protection block 10, unit 1 for estimating statistical parameters of the controlled variable, unit 19 for estimating statistical parameters of random and periodic components, computational unit 20, block 2b for correcting the threshold set value for constant component, block 3 for correcting dynamic properties

16 V

systems, adders 22, 35, optimizer 6. The system compensates for the delay and inertia in the main control loop, corrects the set value of the adjustable coordinate and accurately estimates the probability of trouble-free operation, which increases the reliability and efficiency of the system. 2 hp f-ly, 1 ill.

The invention relates to the automatic control and protection of potentially dangerous objects operating under conditions of uncontrolled disturbances, and can be used in the chemical, food and other industries.

The aim of the invention is to increase the reliability and accuracy of the system under the conditions of an uncontrolled disturbance on an object.

The drawing shows a block diagram of the system.

The system includes the first error meter 1, the control signal generator 2, the control object 3, the adjustable variable 1 1 counter k, which constitute the main control loop 5, the optimizer 6 (the specified variable value), the speed limit block 7, the second 8 meter mismatch and key E., which make up protection block 10, first 11, second 12, third 13 low frequency filters, third error meter, differentiator 15, first 1b and second 17 quadrants, which make up statistical estimate block 18 x parameters of the controlled variable, unit 19 for estimating statistical parameters of the random and periodic components, computing unit 20, block 21 for correcting the threshold set value for the harmonic component, first adder 22, adder 23, low frequency filter 24 and meter 20 mismatch that a constant setpoint correction value block 26, a common latency model 27, a mismatch meter 28, a regulator model 29 (without delay), an object model 30 (without delay), a model Sensor 31 (without zapaz- dyvani) mismatch meter 32, the high pass filter 33, constituting the correction unit 3 the dynamic properties of the system, a second adder 35.

The system works as follows.

In the main control loop 5, the driver of the control signal 2 generates a control signal to the object 3 so that the error at the output of the first meter 1 of the error E X - X is minimum where X is the set value of the controlled variable, i.e. the main control loop 5 processes the change in the signal of setting the controller Xj and the change in the values of external disturbances | u according to the dependence X (p) W LR) -KhZ + W j (p) - p (p), (1) where,)) are transmission

the functions of the main control loop 5 on the setpoint and perturbation channel.

The optimizer 6 of the specified value of the regulated value by the values of the adjustable value X with compensation for the delay and inertia received from the error meter 32, and by the sign of the external parameters r, ..., q determines the optimal set value of the adjustable value Xo corresponding to the extremum of the target function Е f (X, г, ..., q), i.e.

Х argextr Е (Х, г, ..., q). (2) The calculated value of the signal of the optimal setpoint X goes to the protection unit 10, namely, to the input of the block 7, the rate of change of the setpoint of the controlled variable, to the other inputs of which signals are received that are adjustable by the value X with delay and inertia compensation from the output is

the error generator 32, the forced setpoint X of the controlled variable X from the output of the adder 35 and the threshold setpoint XH of the adjustable value from the output of the adder 22.

Limiting the rate of change of the optimal setpoint Xp at a certain calculated level Xp is necessary so that, taking into account the inertia of the control object 3, limit the speed of approaching the controlled variable OX to the threshold value X and thereby prevent such outputs X beyond X "that would exceed the alarm hd value Block 7 can perform this function, for example, by dependencies

{

X with ultrasound - Vo 7 0; . 0;

X

Xp

 with Vj - Vo

V.

K (X „- X);

Vo Xо - X, UZ + X,

where K is the proportionality coefficient.

Since the optimizer 6 of the specified value of the controlled value determines the optimal specified value XQ by (2) without taking into account the constraint X (t) Xj, |, then the specified value X of the controlled value must be limited at the threshold value X, moreover, due to the presence of uncontrolled external These disturbances and inertia of the main circuit 5 of the control X must be less than X. By the sign of the difference X "- Xp, which is calculated on the error meter 8 and goes to the control input of the key 9, the latter switches its output to one of the inputs wherereceive the corrected signal X р Xfl + d m; (from the output of block 26 and the signal from the output of block 7, i.e. realizes the dependence, Г Хр at Х п - Хр 0; at - Хр 0. (

Changes in the intensity, spectral composition of uncontrolled external disturbances f and the parameters of the object 3 of the control in time cause changes in the statistical characteristics of the random process X (t), which requires the corresponding X

the threshold value of the set value Xp according to the formula

,to

Хп Х -6.V21nCl.6 / 2tr6 1n ((t)) 4. + +%, (5)

 t

Q 5 20

sixteen

where P, fp (C) is the probability of trouble-free operation on the interval T;

C is the contribution of the periodic component to the definition

P

Considering C as a function of C f (6l, AS, Wg, (0) i;) in 2, we obtained the specification of a regression equation that approximates this functional dependence for the ranges of argument measurements that cover a significant part of the control objects encountered in practice and the conditions of their functioning.

Having determined the parameter estimates for a random process X on a sliding interval and by dependencies

(, 1 6 - A, V2; (6) 6 6 - (7)

Hell V2 17 ((6) - Hx; (8)

five

0 5 0 w

0

five

s.

- | 3i4i) .. AV e b (b) s: Us

where p, pj (9)

s

L,

u) X S

estimates of the fourth central moments of the controlled variable X (1) and its derivative X (t) on the interval, the amplitude of the derivative of the harmonic oscillation,.

It is possible to establish such a threshold setpoint X, which will provide the necessary reliability of operation of a potentially dangerous object if there are changes in the controlled variable, in addition to a random harmonic component.

The implementation of these dependencies is performed by the block 18 for estimating statistical parameters of the controlled variable, the block 10 for evaluating the statistical parameters of the random and periodic components, the computing block 20, the block 21 for correcting the threshold set value for the harmonic component, the block 26 for correcting the threshold set value for the constant composition 26.

Block 18 on a sliding time interval O evaluates the statistical parameters of the predicted value Hu of an adjustable value X (t) coming from the output of the error meter 32 to the input of the low-frequency filter 11, which averages them, over the interval t.

715134

Gauge And mismatch on-i

is centered relative to

process (X (t) s. date output (t, i;) random process X, j (t) as

signal difference

"JMKa 4 and X (t, C) from the output of filter 11

low frequency i.e.

Hu (1 :) X (t) - X (t, C); (YU)

-t

x | () - exp {-) X (t-i;) dt t

X (t 0)

(eleven)

The signal Xy (t) from the output of the error meter is squared by the squared 1b and averaged over (11) on the low-frequency filter 12, the signal from the output of which is a dispersion scene of the controlled variable ff in the interval Г

The same signal X, (t) is fed to the input of the differentiator 15, which determines its derivative Xy (t), and the quadrant 17 and the low-pass filter 13 determine the estimate of the dispersion of the intelligent variable value on the transducer.

I The signals Xy (t),, 6, and X ,, (t) are respectively 4plugged to the inputs of the block 19, the evaluation of the statistical parameters of the random and periodic components.

Block 19 on the sliding interval C according to dependencies (6) i- (9) determines the ft estimate of the statistical parameters of the flux n (t) and periodic s (t) delivering a controlled variable

: t (t).

from the outputs of block 19 signals 6 and b. equal to the estimate of the variance and the estimate; the version of the derivative of the random element n (.t) on the sliding interval .Ie, go to the inputs of the computing unit 20, which, given the values of the Emergency, adjusts the value of the Averaging of trouble-free (care P d |; p (C) according to

K, X „-il 2,21p (1.6 / 2,, 1n (1 / Pd5p (1

K12)

where Xj, is the emergency value of the regulated value;

 , are the rms values of the controlled variable X (t) and its derivative C (t) estimated over a sliding time interval of-C,

calculate the contribution of the random component uMy Rdb to the determination of the threshold

values of X ,, according to (5).

eight

0

about 5

,

35

,

Q

55

Signals A are the estimates of the amplitude of the harmonic component s (t) at | the slidable interval €, „,„, 60 - the harmonic component frequency estimate s (t) on the interval C from the outputs of the block 1-9 is fed to the inputs of the correction setpoint correction unit 21 for the harmonic component for the set time averages C and admissible probability of trouble-free operation (c) according to

g-w l. to AS,, 2- С b, + b ,, ----

 1n (1 / RA1Gr ({)}, (13)

which is an approximation of the functional dependence C f (6., 6 l, Ad, (lSd, PftgpCf), f), calculates the contribution of the periodic component s (t) to the definition of the threshold setpoint X "according to (5).

The final generation of the threshold setpoint X performs the adder 22 of the block of correction of the threshold setpoint on: The constant component on which the algebraic addition of the output signals from blocks 20 and 21 occurs.

The signal Hz is fed to the input of the low-pass filter 2k, which calculates the value of the mathematical expectation of a given value of the controlled variable X. The error meter 25 subtracts from the signal m, the signal m from the output of the low-pass filter 17, and the difference is fed to the input of the adder 23, where the difference flmjj tx "- m is summed with the signal Xp in the event that Xj X". In the absence of this equality, the summation does not occur and the signal dm X is used only to form a tolerance for the signal X. At the output of block 26, when entering the correction dm X, the signal Xn Xx + dt, appears. Thus, block 26 adjusts the threshold given constant voltage flag X by the value of & t.

The signal X from the output of the protection unit 10 is fed to the input of the block 3 for the correction of the dynamic properties of the system, where the inertia and lag of the main control loop are compensated. For this purpose, the signal Xj occurs through the main control loop model, which is a series connected

models of controller 29, object 30 and sensor 31 without delay, and model 27 of total delay, which are closed by a negative feedback circuit. At the output of the adder, a signal X is generated., Equal to the sum of the signal of the increment prediction due to the change in X by the forward lag time and the signal of the controlled variable X from the output of the main control loop.

To the input of the main circuit model, in addition to the signal X, a signal comes from the output of the high-pass filter 33, which provides real acceleration of the transient processes in the object. The signal from the output of the high-pass filter 33 is also fed to the input of the adder 35, which forms an accelerated setpoint X, which arrives at the input of the block 7 for limiting the rate of change, the input of the main control loop.

Thus, block 3 eliminates the adverse effect of the delay by the predictive correction and the effect of the inertia by introducing a forcing element accelerating the transient processes in the object.

At the initial moment of time when the control object 3 is turned on, the threshold set value of the controlled variable X J is set at a safe level, which excludes the occurrence of an emergency situation with the worst actual external disturbances by setting the initial conditions of the low-frequency filters 12 and 13. With the accumulation of information about a random process X (t) over a sliding interval 1}, the value Xp is corrected by blocks 20 and 21 taking into account the probabilistic properties of this random process.

The inputs of block 3 are supplied with initial conditions from the output of the sensor k, the object 3 and the controller 2, which provide a smooth output | of the controlled coordinate when the control object is turned on for a given value.

The control system of a potentially dangerous object can compensate for the delay and inertia in the main control loop, which increases the control efficiency, as well as correct the set value of the adjustable coordinate and accurately estimate the probability of failure-free

operation of the adaptive control system that increases the reliability and efficiency of the system.

Claims (3)

1. Adaptive control system of a potentially dangerous object, containing the first error meter, the output of which is connected to the first input of the first error meter through a serially connected control signal generator, the control object, a variable-value sensor, the second input of which is connected connected to the output of the second error meter, the first input of which is connected to the first information input of the key and to the output of the limiting unit soon the first input of which is connected to the second input of the second error meter and to the output of the first adder, the first input of which is connected to the output of the correction unit of the threshold set value for the harmonic component, the first, second, third and fourth inputs of which are connected respectively to the first, second, the third and fourth outputs of the unit for estimating statistical parameters of random and periodic components, the second and third outputs of which are connected to the first and second inputs of the calculator, respectively The second block, the output of which is connected to the second input of the first adder, the second input of the speed limiting block is connected to the output of the optimizer connected by the input to the third input of the speed limiting block to the first input of the third error meter and to the input of the first low-frequency filter whose output is connected to the second the input of the third error meter connected by the output to the input of the first quad, to the input of the differentiator and to the first input of the evaluation unit of statistical parameter and periodic components The first, second, third and fourth inputs of which are connected respectively to the outputs of the second low-pass filter, the third low-pass filter and the differentiator, the output of the differentiator is connected to the input of the second quadr connected by the output to the input The tpetbero low-pass filter, the output of the quad quater is connected to the input of the second low-pass filter, which means that, in order to increase the reliability and accuracy of the system under conditions of an uncontrollable disturbance on the object, the correction unit of the threshold laden value for the constant connection, the unit for the analysis of the dynamic properties of the system and the second summer, the output of the first adder: connected to the first input of the correction unit of the threshold setpoint and constant component , the second input and output, which go connected respectively with the output of the low-pass filter and the second information | key stroke 6m, the output of which is connected to the first input of the dynamic properties correction unit of the system and to the first input of the second adder, the output of which is connected to the fourth input () of the speed limit local, the third input of which is connected to the first output (Systems, the second output of which is connected to the second input of the second sum-., Fopa and the third input of the first one - (mismatch forerunner, the first move of which is connected to the second input of the system dynamic correction unit). : 2. The system of claim 1, wherein i | l and so that, the correction unit of the threshold setpoint value by the constant component contains the sum 5 0 five mater, error meter, low frequency filter, the first block input connected to the first input of the adder, the output of which is connected to the block output, the first input of the block connected via a low frequency filter to the first input of the error meter, the second input of which is connected to the second input block, and the output is connected to the second input of the adder. 3. The system according to claim 1, characterized in that the dynamic properties correction block of the system contains a high-frequency filter, two error meters, a controller model, an object model, a sensor model, and a general delay model, the first input of the block is connected to the first input of the first error meter and the input of the high-frequency filter, the output of which is connected to the second output of the unit and the second input of the first error meter, the output of which through a serially connected model is a regula tor, the object model and the sensor model is connected to the first input of the second error meter and to the input of the general delay model, the output of which is connected to the third input of the first error meter and to the second input of the second error meter, output and the third input of which are connected respectively to the first output and the second input of the unit,