SU1291926A1 - Adaptive control system for potentially dangerous object - Google Patents

Abstract

The invention relates to the field of automation and is intended for use in chemical, energy, etc. region x. The purpose of the invention is to increase the efficiency and reliability of operation of a potentially dangerous object under conditions of uncontrolled external disturbances and changes in the properties of the object. The system includes a control object, a variable size sensor, a first error meter, a control signal generator, a variable value set optimizer, a unit for limiting the rate of change of a given value of a variable value, a second error meter, and a key. New in the system is the first low-pass filter and the third error meter, which determine the random component of the fluctuations of the controlled variable; a quad and a second low-pass filter defining an estimate of the variance of random oscillations of a controlled variable; a differentiator, a second quad, and a third low-pass filter, which determine the estimated variance of the derivative of random oscillations of a controlled variable over a sliding interval; a computing unit consisting of a division unit, a first and a second square root extraction unit, a logarithm calculation unit, a multiplication unit, a fourth error meter and a control unit alarm value storage unit, which determines the threshold set value of the control value depending on statistical characteristics random fluctuations of the controlled variable. The system can be fully implemented in commercially available equipment and satisfies the requirements of highly automated technology. 1 il. i (L 1C with co N9 O)

Description

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 efficiency and reliability of operation of a potentially hazardous facility. The achievement of the purpose of the invention is due to the fact that under conditions of uncontrolled external disturbances and changes in the parameters of the control object, the change in the threshold value of the controlled variable is implemented, depending on the statistical characteristics of random fluctuations of the controlled variable.

The drawing shows the functional scheme of the device. The device contains serially connected first error meter 1, control signal generator 2, control object 3, sensor 4, which together constitute the main control loop 5.

The output of sensor 4 through a serially connected optimizer 6 of a specified value of the controlled variable, unit 7 for limiting the rate of change and the second error meter 8 is connected to the control input of switch 9, blocks 7–9 constitute protection unit 10, the first input of switch 9 is connected to the output of the change rate control unit 7, the second input of the change rate control unit 7 is connected to the output ne of switch 9 and to the first input of the first I-error meter, the output of sensor 4 is connected to the second input of the first The third meter of the block is the limit of the rate of change. The device also contains three low-frequency filters 11-13, a third error meter 14, two quadrants 16 and 17, a computing unit 18, the output of sensor 4 being connected to the first input of the third error meter 14 and, through a low-frequency filter, to the second input of the third meter 14 mismatch, and the output of the third meter mismatch 14 through the first quad 16 connected in series and the second low frequency filter 12 is connected to the first input

0

five

0 Q 5 0 5

the numeral unit, as well as the third error meter 14 through the serially connected differentiator 15, the second quad 17 and the third low-frequency filter 13 are connected to the second input of the computing unit 18, the output of which is connected to the second input of the second error meter 8, to the second input of the switch 9 and to even-: to the right input of the unit 7 for limiting the rate of change.

In turn, the computing unit 18 contains the division unit 19, the first 20 and second 21 square root extraction units, the amplifier 22, the logarithm calculation unit 23, the multiplication unit 24, the fourth error meter 25, the alarm value storage unit 26, the first computing input unit is connected to the first input of division block 19 and the first input of multiplication unit 24, and the second input of the computing unit is connected to the second input of division unit 19, the output of which through serially connected first unit 20 is extracted and the square root, the amplifier 22 and the logarithm 23 calculating unit 23 are connected to the second input of multiplication unit 24, the output of which is connected via the second square root extraction unit 21 to the first input of the fourth error meter 25, the second input of which is connected to the output of the alarm storage unit 26 magnitude, and the output of the fourth error meter 25 is connected to the output of the computing unit. Blocks P-26 together constitute the blocks 27 for setting the threshold value of the regulated value. I

The device works as follows.

In practice, for the vast majority of control objects, uncontrolled external disturbances have many independent sources, therefore changes of the controlled variable over time are random, i.e. are random processes. Therefore, the output of the regulated value for the emergency value will be a random event, the probability of occurrence of which depends on

statistical properties of these random processes. In this case, the reliability of the operation of the object of control must also be assessed by a probabilistic measure, in particular, by the probability of the controlled value not exceeding the emergency value. Random fluctuations of the controlled variable X over time t, i.e. random processes X (t) arise as a result of the impact on the inertial control object of a multitude of sources of uncontrolled external disturbances, and because of this, the distribution law of these processes is close to normal. The correlation function of random processes X (t) characterizes the quality of operation of the main control loop (local stabilization system). In particular, the time T, its decay to negligibly small values can be interpreted as the average time of compensation of disturbances.

The actual processes X (t) are not stationary due to the nonstationary nature of the parameters, the control object, and external disturbances. The rate of change of these parameters can also be estimated by the decay times T, T of the corresponding correlation functions of the parameters to negligibly small values.

Since, for a large number of control objects, the rate of change of their parameters and parameters of external perturbations is relatively low and the decay times, T is much longer

TC TC T.T., this allows one to accept the hypothesis about the quasistationarity of the parameters and, at a certain time interval, consider the processes X (t) to be stationary.

Considering random processes X (t) on a sliding time interval T can be promptly (with a delay of t) determining the estimates of their statistical characteristics — expectation, variance, etc. The recommended value is c (1-5) T, i.e. is negligible compared to .. "and T ..,.

l. (J CM

Operation of a potentially hazardous object makes sense when an accident occurs, i.e. the outputs (outliers) of the regulated value X (t) for the emergency value Xd are extremely rare events. And in this case, from the standpoint of the theory of probability, these events are independent and the probability of their occurrence can be described by the Poisson law. In particular, the probability of an event X (t) X being canceled over a time interval for a normal distribution law is equal to

R. p ft). "P - 11-

V.

where b, b is the standard deviation of the controlled variable X (t) and its derivative over time s;

ha - the mathematical expectation of the controlled variable X (t) during time C.

This probability determines the probability of trouble-free operation on the interval and is the measure that characterizes the reliability (safety) of operation of a potentially dangerous object.

Theoretically, since the normal distribution law is defined on infinite intervals of variation of the argument, the probability of P. is always less than 1, i.e. the probability of an accident (,. - Pgp O; However, the actual values of X (t) cannot vary infinitely. Therefore, it is assumed that the values of X (t) m ,, t 36j, the theoretical probability of occurrence em 0.003, almost impossible.

During the operation of a potentially hazardous object, changes in its parameters and parameters of external disturbances lead to corresponding changes in the statistical characteristics of the random process X (t). An operative estimate of m, (j, Cj jj on a sliding time interval b) allows one to determine from (1) the probability of trouble-free operation (T) almost in real time, in tempo with the technological process in the object.

Information about t, j, G, Pgp (t) can be used not only to assess reliability, but also to control a potentially dangerous object.

So from (1) it follows that in order to ensure trouble-free operation of a potentially dangerous object for a time interval (i.e. to ensure

RBD () 5) De DOPU - the probability of trouble-free operation

on this interval) mathematical 1

the expectation of the regulated t must not exceed the threshold set value

Xp X,

{2Gl ((o). (2

With Rlr, the actual probability of a trouble-free operation of a Rdvr. The value of Rxp depends on the type of real law of distribution of the controlled variable X (t).

Operatively estimating the variations b, 6 jj of the controlled variable X (t) over the sliding time interval t when the object parameters and external disturbances are named, the proposed system adjusts the threshold set value X for the main control loop (local stabilization system) by (2) which maintains the mean value1 (mathematical expectation mj () at the required level,

In the main control loop 5, the driver of the control signal 2 generates a control signal to the object 3 so that the mismatch is

the output of the first meter I mismatch XT - X is minimal, where Xi is the specified value of the controlled variable, i.e. The main control loop 5 processes the change in the signal of the task to the controller XB and the change in the values of external disturbances U according to

X (P) W, (P) X (P) + W (P) | rj (P), (3) where W, (P) and W (P) are the transfer functions of the main control loop 5 over the reference and perturbation channel .

The optimizer 6 of the specified value of the regulated value, by the values of adjustable value X, coming from sensor A, and by the values of external parameters r, .. ,, q determines the optimal set value of the adjustable value X, corresponding to the extremum of the target function. The calculated value of the signal of the optimal setpoint X is fed to the protection unit 10, namely to the input of the block 7, the rate of change of the setpoint 30

40

moreover, due to the presence of uncontrolled external disturbances and the inertia of the main control loop 5, X must be less than X. By the sign of the difference Xp - Xp, which is calculated on the second error meter 8 and fed to the control input of the switch 9, the latter switches its output to one from the inputs to which the signals X are received, from the output of block 35 and 27 and Xp from the output of block 7, i.e. dependence is realized.

XP at X - Xr O,

X "at Xj, - Xp 0. Changes in the intensity, spectral composition of uncontrolled external disturbances (U and parameters of the object 3 controls in time cause changes in the statistical characteristics of the random process X (t), which requires appropriate adjustment of the threshold set value X" according to (2).

(five)

The implementation of the dependence (2) after a long time of the value of the regulated value, on the 50th unit, the threshold setting block 27

 set value of the regulated value. The values X (t) from the output of sensor 4 of the variable size are fed to the input of the first low frequency filter 1I 55, which averages them over a sliding time interval. Technically, it is easiest to perform a low-pass filter in the form of an inertial link of the first order-integral, the other inputs of which receive signals of an adjustable value X from the output of sensor 4, the set value Xi. adjustable value from the output of the switch 9 and the threshold set value X "adjustable value from the output of the block 27. Limit the rate of change of the optimal set

X values on a specific calculation

O

50

five

0

0

At a given level, Xp is necessary in order to take into account the inertia of the control object 3, to limit the rate at which the controlled variable X approaches the threshold value X and thereby prevent such outputs X beyond X, which exceed the emergency value Xd. Block 7 can perform this function, for example, by dependencies

GC with V. - Vo 0; Р 1Х when V - V, 0; V K (Xn - X); X,

(four)

Vo Ho: v

-k) where K - coeff () is the patient of proportionality.

Since the optimizer of the set value of the regulated value 6 determines the optimal set value XQ by (4) without taking into account the limitation X (t) Xg, the set value X of the controlled value must be limited at the threshold value X,

moreover, due to the presence of uncontrolled external disturbances and the inertia of the main control loop 5, X must be less than X. By the sign of the difference Xp - Xp, which is calculated on the second error meter 8 and fed to the control input of the switch 9, the latter switches its output to one from the inputs to which signals X are received, from the output of block 5 and 27 and Xp from the output of block 7, i.e. dependence is realized.

XP at X - Xr O,

X "at Xj, - Xp 0. Changes in the intensity, spectral composition of uncontrolled external disturbances (U and parameters of the object 3 controls in time cause changes in the statistical characteristics of the random process X (t), which requires appropriate adjustment of the threshold set value X" according to (2).

(five)

of a torus covered by a unit negative feedback, which realizes an exponentially weighted averaging over the dependence 1 X (t, t) i - f exp (- -)) i

X x (t-c;) dt + x (t 0),

(6)

where X (t, o) is the signal at the output of the filter averaged over a moving time interval;

T is the integration constant (T, 0.5C);

X (t, t;) is the signal at the input of the filter;

X () is the value of the initial conditions on the integrator.

The third error meter 14 determines the random process Xii (t) centered with respect to X (t, o) as the difference between the signals X (t) from the output of sensor 4 and X (t, t) from the output of the first low frequency filter I1, i.e. .

Xy (t) X (t) - X (t, r). The signal XuCt) from the output of the third error meter 14 is squared by the first quad 16 and averaged by the formula (6) on the second low-pass filter 12, the signal from the output of which is an estimate of the variance of the controlled variable on the interval D.

The same signal X (t) is fed to the input of the differentiator 16, which, together with the second quad 17 and third low-frequency filter 13, determine the estimate of the dispersion of the derivative of the controlled variable GJ in the interval o

The signals GX and (5 from the outputs of the second 1 2 and third 13 low-frequency filters arrive at the corresponding inputs of the computing unit 18, which for a given alarm value of the controlled variable Xd. And the permissible probability of a trouble-free operation Pgg determines the threshold setpoint adjustable value Xr.

The amplifier 22 introduces a coefficient

corresponding l / 21G1p (-), and

„Emergency value memory unit

adjustable value 26 - the value of Hd. The functions performed by dividing unit 19, first 20 and second 21 square root extractor units, logarithm calculating unit 23,

com 24 multiply and the fourth mismatch gauge 25 correspond to their names, and the sequence of their connection follows from (2),

At the initial moment of time at

the inclusion of the control object 3 into operation, the threshold setpoint of the adjustable value Xp, is set at a safe level, excluding

the occurrence of an emergency with the worst real external disturbances by setting the initial conditions on the integrators of the second 12 and third 13 filters low often, you .. When

the accumulation of information about a random process X (t) during a sliding time interval, the value of X ,, is corrected by block 27 taking into account the probability properties of this random

process.

The device can be implemented on the basis of the means of the State system of devices and automation equipment, for example, an aggregate complex of electric means of regulation in microelectronic design (AKESR).

The effectiveness of the adaptive automatic control system of a potentially dangerous object is that it improves the accuracy of reaching the extremum of the target function by 20-30% by increasing the permissible range of change of the controlled variable and, accordingly, by 5-10% technical and economic performance indicators of the control object. . In addition, the system eliminates, with a probability of no less than a given, the occurrence of

situations, taking into account the probabilistic properties of the variable and the change in the parameters of the control object.

Claims (2)

1. An adaptive control system of a potentially hazardous object, containing a serially connected first error meter, control signal generator, control object, sensor, adjustable value set optimizer, change rate limiting unit and second meter the error, the output of which is connected to the control input of the switch, the first INPUT of which is connected to the output of the limiting unit measuring speed, the second input of which is connected to the output of the switch and to the first input of the first error meter, the output of the sensor is connected to the second input of the first error meter and the third input of the measuring speed limiting unit, which in order to increase the efficiency and reliability of operation Remembering the emergency value of the regulator, three low-pass filters, a third error meter, a differentiator, two quadrants and a computing unit are introduced, and the output of the sensor is dinene with the first input of the third error meter and through the first low frequency filter with the second input of the third error meter, and the output of the third error meter through serially connected first quad and second low frequency filter is connected to the first input of the computing unit, and the third error meter through serially connected differentiator, the second quad and the third low-frequency filter are connected to the second input of the computing unit, the output of which is dklyuchen vho- to the second row of the second meter rassoglasova Compiled by V, Peshkov Editor V. Ivanov Tehred I.Popovich Proofreader M. Order 261 / A5 Circulation 864 Subscription VNIIPI USSR State Committee for inventions and discoveries 113035, Moscow, Zh-35, Raushsk nab., 4/5 Production and printing company, Uzhgorod, ul, Proektna, 4 nor, to the second input of the switch and to the fourth input of the measuring rate limiting unit. 2. The system of claim 1, wherein the deductive unit comprises a division unit, the first and second square root extraction units, an amplifier, a logarithm calculation unit, a multiplication unit, a unit for variable size, the fourth error meter, the first input of the computing unit connected to the first input of the division unit and the first input of the multiplication unit, and the second input of the computing unit connected to the second input of the division unit, the output of which is through a series-connected first square root extraction unit and the logarithm calculation unit is connected to the second input of the multiplication unit, the output of which is connected to the first input of the fourth meter through the second extraction unit of the square root popping a second input coupled to the output of the emergency storing values of the regulated quantity, and the output of the fourth mismatch meter connected to the output calculation unit.