RU2486563C1 - System of control objects identification - Google Patents

Abstract

FIELD: information technologies.

SUBSTANCE: system comprises a model of a control object, the third block of comparison and a block of calculation of parameters of an object model, the inlet of which is connected with an output of the operator controller, the outlet of the model parameters calculation block is connected to the second output of the control object identification system and is connected to the first inlet of the comparison object model, the second and third inlets of which are connected, accordingly, with the inlet and outlet of the control object; the first, second and third outlets of the control object model are connected, accordingly, to the first inlet of the third block of comparison, to the second and third inlets of the operator controller, the second inlet of the third block of comparison is connected with the outlet of the control object, and the outlet of the third comparison block is connected to the second inlet of the second comparison block and to the inlet of the block of generation of control error properties.

EFFECT: higher accuracy of control objects identification.

3 cl, 3 dwg

Description

The invention relates to automatic control and can be used in automatic control systems of dynamic non-stationary objects, mathematical models of which contain variable operators and / or parameters.

An example of such objects is a vortex furnace for burning a mixture of coal and slurry fuel, the technological process of which is subject to uncontrolled disturbances caused, in particular, by changes in the quality characteristics of coal and slurry fuel, violations of the spraying process of the latter, aging of structural elements of the furnace , which leads to a change in the dynamics of thermal processes and, accordingly, the dynamics of the channels of conversion of material and energy flows. In addition, the object in question is an object with a variable structure, which changes, for example, when the installation is switched to work with and without coal supply, which, along with changing the structure of the control object, entails the need to change its mathematical model.

To identify control objects, an adaptive identifier [1] is known that contains the first model of the object, the first comparison unit, the adder, the first delay unit, the second object model, the second comparison unit, the controller, extrapolator, and the second delay unit connected in series to the first input of the second the model of the object and the second comparison unit, the second input of which is connected to the output of the controller, and the output to the first input of the first model of the object, the output of which is connected to the first input of the adder, the first and second input The s of the first comparison unit are connected respectively to the output of the adder and the first input of the identifier, the second input of which is connected through the first delay unit to the second input of the first and second models of the object, the outputs of which are connected respectively to the first and second inputs of the adder, the output of the first comparison unit is connected to the input of the controller .

When the identifier operates in a model-closed control loop composed of an adder, a first object model, a regulator, first and second comparison blocks, the estimate of the transmission coefficient of the object is retarded with delay. This estimate is extrapolated to the current point in time, and is also adjusted taking into account the extrapolation error, which is determined in the circuit containing the second delay block and the second comparison block.

A disadvantage of the known identifier is its low functionality, since it is oriented to determine only an estimate of the transfer coefficient of a dynamic object.

Closest to the technical nature of the proposed system for identifying control objects is a control system [2], containing a control object, sequentially connected control unit, a first comparison unit, a unit for generating properties of regulation errors, a second comparison unit, an operator controller, a coordinate controller, the second input of which is connected with the output of the first comparison unit, the output of the coordinate controller is connected to the input of the control object, the output of which is connected to the second input of the first comparison unit the output of which is connected to the second input of the second comparison unit, the output of the control object is connected to the output of the system.

During the operation of the control system, depending on the error signal, the regulating action is generated at the output of the first comparison unit by the coordinate controller, for example, in order to provide the specified properties of the control errors of the output effect of the control object subject to the influence of an unknown disturbance. Moreover, if the error of coordinate regulation does not correspond to its specified properties, a signal about which is generated at the output of the block for generating properties of regulation errors, then the operator regulator produces control actions for changes in the structure or values of the parameters of the regulation law implemented in the coordinate controller.

The disadvantage of this system is its low functionality, since it is not intended to perform the identification function of a non-stationary object, i.e. assessing the structure and parameter values of its mathematical model.

The objective of the invention is the expansion of the functionality of the system.

The problem is achieved in that in a system containing a control unit, a control object connected in series, a first comparison unit, a coordinate controller, a control error properties generating unit connected in series, a second comparison unit, an operator controller, and the coordinate controller output connected to the control object input, output which is the first output of the system, the setter output is connected to the second input of the first comparison unit, a model of the control object is introduced, the third comparison unit and a unit for calculating the parameters of the model of the object, the input of which is connected to the output of the operator controller, the output of the unit for calculating the parameters of the model is connected to the second output of the identification system of the control object and connected to the first input of the model of the control object, the second and third inputs of which are connected respectively to the input and output of the control object ; the first, second and third outputs of the control object model are connected respectively to the first input of the third comparison unit, to the second and third inputs of the operator controller, the second input of the third comparison unit is connected to the output of the control object, and the output of the third comparison unit is connected to the second input of the second comparison unit and to the input of the block forming the properties of regulation errors.

The control object model contains in series the first delay block, the fourth comparison block, the object model in increments and the fifth comparison block, the second delay block, the input of which is connected to the third input of the control object model, and the output of the second delay block is connected to the second input of the fifth comparison block, the output of which is connected to the first output of the model of the control object, the output of the model of the object in increments is connected to the second output of the model of the control object, the third output of which is connected to the first input of the mode whether the object is in increments, the second input of the fourth comparison unit is connected to the input of the first delay unit and the second input of the control object model, the second input of the object model in increments is connected to the first input of the control object model, which is connected to the output of the model parameter calculation block.

The operator controller contains in series the first signal redefinition unit, the first squaring unit, the adder, the first division unit, the first multiplication unit, the first integration unit and the first scaling unit, the differentiation unit connected in series, the second signal redefinition unit, the second division unit, the second unit multiplication, the second integration unit and the second scaling unit, the second squaring unit, the input of which is connected to the output of the second signal redefinition unit, and its output is connected to the second input of the adder, the first input of the operator controller connected to the output of the second comparison unit is connected to the second inputs of the first and second multiplication units, respectively, the second input of the operator controller connected to the second output of the model of the control object is connected to the input of the differentiation unit, the third input of the operator controller connected to the third output of the control object model is connected to the input of the first signal redefinition unit, the output of which is connected to the second input the first division unit, the first input of which is connected to the second input of the second division unit, the first input of which is connected to the input of the second squaring unit, the outputs of the first and second scaling units are connected to the output of the operator controller and are its output signal components connected to the input of the calculation unit model parameters.

Figure 1 shows a block diagram of a system for identifying a control object. Figure 1 shows the following notation:

u (t) - signal about the control action;

y (t) - signal about the output of the control object;

- сигнал об оценках коэффициентов модели объекта управления;k _j (t); $$j=\overline{\mathrm{one,}J}$$

- a signal about the estimates of the coefficients of the model of the control object;

J is the number of these coefficients;

t is continuous time.

Figure 2 presents a block diagram of a model of the control object, which is one of the options for implementing the model of the control object 5. Figure 2 shows the following notation:

y ^M (t) - signal about the output effect of the model of the control object;

δu (t) is the difference between the signals u (t) and u (t-τ);

δy (t) - signal about the output of the model of the control object in increments.

Figure 3 shows an example of a block diagram of an operator controller that implements the law of operation (16), (17), (10), (11). Figure 3 shows the following notation:

на свойства ошибки операторного регулирования;σ (t) is a signal about the deviation of the error of the object model ε _y (t) from the task $${\epsilon }_y^{*}(t)$$

on the error properties of operator regulation;

; $$k_2^{л}(t)$$

- сигналы о текущих оценках коэффициентов линейно-параметрической модели (9). $$k_{\mathrm{one}}^l(t)$$

; $$k_2^l(t)$$

- signals about current estimates of the coefficients of the linear-parametric model (9).

The control object identification system contains a control object 1, a coordinate controller 2, a first comparison unit 3, a setter 4, a control object model 5, a third comparison unit 6, a model parameter calculation unit 7, an operator controller 8, a second comparison unit 9, a property generating unit 10 regulation errors.

The control object model contains a first delay unit 11, a fourth comparison unit 12, an object model 13 in increments, a fifth comparison unit 14, and a second delay unit 15.

The operator controller comprises a first signal redefining unit 16, a first squaring unit 17, a first division unit 18, a first multiplication unit 19, an adder 20, a first integration unit 21, a first scaling unit 22, a differentiation unit 23, a second signal redefining unit 24, a second a squaring unit 25, a second division unit 26, a second multiplication unit 27, a second integration unit 28 and a second scaling unit 29.

The system for identifying control objects works as follows. The signal y (t) about the output action of the control object 1 is supplied by the first input to the first comparison unit 3, where it is compared with the signal y ^* (t) about setting the object 1 to the output variable, which comes from the output of setter 4 to the second input of the first block 3 comparisons. The output signal ε (t) of the first comparison unit 3 enters the coordinate controller 2, at the output of which a signal u (t) about the control action appears, generated in accordance with the coordinate regulation algorithm f _R {·}

$$\text{u (t)}={\text{f}}_{\text{R}}\text{{}{\epsilon }_{\text{u}}(t)\}.(\mathrm{one})$$

The signal u (t) from the output of the coordinate controller 2 is fed to the second input of the model 5 of the control object and to the input of the control object 1, where it provides with the required accuracy compensation for deviations of the signals y (t) from y ^* (t). Simultaneously with the signal u (t), the signal y (t) is supplied to the input of the model 5 of the control object, using which the signal y ^M (t) is calculated about the output of the model 5 of the control object.

The functioning of the model 5 of the control object is as follows. The signal u (t) from the output of the coordinate controller 2 is delayed in the first delay unit 11 by the delay time τ and is subtracted from the signal u (t). The resulting signal difference δu (t) is fed to the first input of the model 13 of the object in increments, the second input of which is fed from the output of the block 7 for calculating the model parameters a signal about the current values of the estimates of its coefficients a ₁ (t). The signal δy (t) from the output of the object model 13 in increments, characterizing its response to the increment δu (t), is summed in the fifth block 14 of comparison with the output signal y (t-τ) of the control object previously delayed by time τ in the second delay block 15 1, forming a signal y ^M (t) at the first output of model 5 of the control object.

The law of functioning of model 5 of the control object in general is represented by the following expressions

$${\text{y}}^{\text{M}}\text{(t)}=\text{y (t-}\tau \text{)}+\delta \text{y (t);}\left(\text{2}\right)$$

$$y(t-\tau )=f_s\{y(t)\};\left(3\right)$$

$$\delta y(t)=\phi \{\delta u_j(t);k_j(t)\};j=\overline{\mathrm{one,}J};(\mathrm{four})$$

$$\delta u(t)=u(t)-u(t-\tau );(5)$$

$$u(t-\tau )=f_s\{u(t)\},\left(6\right)$$

where τ is the delay time of the signals u (t) and y (t);

f _s {·} is the signal delay operator for the time τ;

φ {·) is the operator of the object model in increments;

- коэффициенты модели объекта в приращениях; $$k_j(t);j=\overline{\mathrm{one,}J}$$

- coefficients of the model of the object in increments;

J is the number of coefficients.

In particular, model (4), including non-linear, can be presented in a linear-parametric form convenient for identification

$$\delta y(t)={\displaystyle \sum _{j=\mathrm{one}}^{J^l}{k}_j^l(t)\delta u_j(t),(7)}$$

; $$j=\overline{\mathrm{1,}{J}^{л}}$$

- коэффициенты линейно-параметрической модели объекта в приращениях;Where $$k_j^l$$

; $$j=\overline{\mathrm{one,}{J}^l}$$

- coefficients of the linear-parametric model of the object in increments;

J ^l - the number of coefficients of the linear parametric model.

For example, if the model 13 of the object in increments is presented as a first-order linear differential equation

$$T(t)\frac{d\delta y(t)}{dt}+\delta y(t)=k(t)\delta u(t),(8)$$

where T (t), k (t) are the current values of the estimates of the model parameters in increments: the time constant and transmission coefficient, respectively, which are the components of the signal k _j (t), then expression (7) will have the form

$$\delta y(t)=k_{\mathrm{one}}^l(t)\delta u_{\mathrm{one}}(t)+k_2^l(t)\delta u_2(t);\left(9\right)$$

$$\delta u_{\mathrm{one}}(t)=\delta u(t);(10)$$

$$\delta u_2(t)=\frac{d\delta y(t)}{dt}.(\mathrm{eleven})$$

The signal y ^M (t) from the first output of model 5 of the control object is supplied at the first input to the third comparison unit 6, where it is subtracted from the signal y (t) received at the second input of the third block 6 of comparison from the output of the control object 1. The output signal ε _y (t) of the third comparison unit 6, proportional to the difference of the signals from the output of the control object 1 y (t) and from the first output of the model 5 of the control object y ^M (t), i.e.

$${\epsilon }_y(t)=y(t)-y^M(t),(12)$$

arrives at the second input to the second comparison unit 9 and to the input of the control error properties generating unit 10, in which the operator S {ε _y (t)} is implemented, which reflects in general terms the task on the error control properties of the operator control.

Alternatively, this task can be related to the nature of the transient error ε _y (t) and, in particular, is expressed using a relation of the following form

$${\epsilon }_y^{*}(t)={\epsilon }_y(t)\cdot e^{-\alpha t},\left(13\right)$$

where α is a constant coefficient.

поступает по первому входу во второй блок 9 сравнения, где он сравнивается с сигналом ε_y(t), вырабатывая на выходе второго блока 9 сравнения сигналSignal $${\epsilon }_y^{*}(t)$$

arrives at the first input in the second block 9 comparison, where it is compared with the signal ε _y (t), generating at the output of the second block 9 comparison

$$\sigma (t)={\epsilon }_y^{*}(t)-{\epsilon }_y(t),(\mathrm{fourteen})$$

which arrives at the first input of the operator controller 8, the second and third inputs of which receive signals δy (t) and δu (t), respectively, from the second and third outputs of the model 5 of the control object.

The law of operation of the operator controller 8, which can implement in a particular case the functions of the parametric controller, is presented in general terms using the following expressions

$$k_j(t)=f_{\mathrm{and}}\{\sigma (t);[\delta u_j(t)]\};j=\overline{\mathrm{one,}J};(\mathrm{fifteen})$$

;

;

; $$\sigma (t)={\epsilon }_y^{*}(t)-{\epsilon }_y(t)$$

;

; $${\epsilon }_y^{*}(t)=S\{{\epsilon }_y(t)\}$$

;

ε _y (t) = y (t) -y ^M (t),

where f _and {·} is the operator of the current estimation of the model coefficients in increments.

If the model 13 of the object in increments is presented in the form of expression (9), then expression (15) will take the following form

$$k_{\mathrm{one}}^l(t)=\frac{\mathrm{one}}{\theta }{\displaystyle \underset{t-\theta }{\overset{t}{\int }}\frac{\delta u_{\mathrm{one}}(t)}{{\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}}}\cdot \sigma (t)dt;\left(16\right)$$

$$k_2^l(t)=\frac{\mathrm{one}}{\theta }{\displaystyle \underset{t-\theta }{\overset{t}{\int }}\frac{\delta u_2(t)}{{\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}}}\cdot \sigma (t)dt;\left(17\right)$$

δu ₁ (t) = δu (t);

$$\delta u_2(t)=\frac{d\delta y(t)}{dt}$$

where θ is the integration interval.

, который по первому входу поступает в сумматор 20.The operation of the operator controller is as follows. The signal δu (t) from the third output of the model of the control object 5 is supplied via the third input of the operator controller to the first signal redefining unit 16 and, after conversion in accordance with expression (10) to the signal δu ₁ (t), is supplied to the first block 17 of square, forming at its output a signal proportional to the value $$\delta u_{\mathrm{one}}^2(t)$$

, which at the first input enters the adder 20.

, который, поступая во второй блок 24 переопределения сигнала, преобразуется в соответствии с выражением (11) в сигнал δu₂(t). Последний, поступая во второй блок 25 возведения в квадрат, формирует на его выходе сигнал, пропорциональный значению $$\delta u_2^2(t)$$

, который через второй вход поступает в сумматор 20, где, суммируясь с сигналом $$\delta u_1^2(t)$$

, вырабатывает на выходе блока 20 сигнал, пропорциональный $${\displaystyle \sum _{j=1}^2\delta u_j^2(t)}$$

. Этот сигнал по вторым входам поступает в блоки 18 и 26 деления, на выходе которых вырабатываются сигналы, пропорциональные значениям $$\frac{\delta u_1(t)}{{\displaystyle \sum _{j=1}^2\delta u_j^2(t)}}$$

и $$\frac{\delta u_2(t)}{{\displaystyle \sum _{j=1}^2\delta u_j^2(t)}}$$

соответственно.Simultaneously with the signal δu (t), the signal δy (t) from the second output of the control object model 5 enters the differentiation unit 23 at the second input of the operator controller, the output of which forms a signal proportional to the value of the derivative $$\frac{d\delta y(t)}{dt}$$

, which, entering the second block 24 of the redefinition of the signal, is converted in accordance with expression (11) into a signal δu ₂ (t). The latter, entering the second block 25 squaring, forms at its output a signal proportional to the value $$\delta u_2^2(t)$$

, which through the second input enters the adder 20, where, summing up with the signal $$\delta u_{\mathrm{one}}^2(t)$$

generates a signal proportional to the output of block 20 $${\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}$$

. This signal at the second inputs enters the blocks 18 and 26 of the division, the output of which produces signals proportional to the values $$\frac{\delta u_{\mathrm{one}}(t)}{{\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}}$$

and $$\frac{\delta u_2(t)}{{\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}}$$

respectively.

о задании на свойства ошибки операторного регулирования, т.е.Simultaneously with the signals δu (t) and δy (t), through the first input of the operator controller, the signal σ (t) proportional to the deviation of the model 5 error of the object ε _y (t ) from the signal $${\epsilon }_y^{*}(t)$$

about setting the error properties of operator regulation, i.e.

$$\sigma (t)={\epsilon }_y^{*}(t)-{\epsilon }_y(t).(\mathrm{eighteen})$$

определяются в блоке 10 формирования свойств ошибок регулирования. Выходные сигналы блоков 19 и 27 умножения поступают соответственно в первый 21 и второй 28 блоки интегрирования, а затем в первый 22 и второй 29 масштабирующие блоки, где умножаются на величину $$\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$\theta $}\right.$$

.Signal Values $${\epsilon }_y^{*}(t)$$

are determined in block 10 of the formation of the properties of regulation errors. The output signals of the multiplication blocks 19 and 27 are received respectively in the first 21 and second 28 integration blocks, and then in the first 22 and second 29 scaling blocks, where they are multiplied by $$\raisebox{1ex}{$\mathrm{one}$}\!\left/ \!\raisebox{-1ex}{$\theta $}\right.$$

.

Thus, the output of the first scaling unit 22 produces a signal proportional to the current coefficient estimate

, $$k_{\mathrm{one}}^l(t)=\frac{\mathrm{one}}{\theta }{\displaystyle \underset{t-\theta }{\overset{t}{\int }}\frac{\delta u_{\mathrm{one}}(t)}{{\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}}}\cdot \sigma (t)dt$$

,

which corresponds to expression (16), and the output of the second scaling unit 29 is a signal proportional to

, $$k_2^l(t)=\frac{\mathrm{one}}{\theta }{\displaystyle \underset{t-\theta }{\overset{t}{\int }}\frac{\delta u_2(t)}{{\displaystyle \sum _{j=\mathrm{one}}^2\delta u_j^2(t)}}}\cdot \sigma (t)dt$$

,

поступают с выхода операторного регулятора 8 на вход блока 7 расчета параметров модели, в котором текущие оценки коэффициентов $$k_1^{л}(t)$$

и $$k_2^{л}(t)$$

линейно-параметрической модели (9) пересчитываются в коэффициенты модели объекта в приращениях (4). В частности, если эта модель представлена выражением (8), то их пересчет осуществляется по формуламwhich corresponds to expression (17). These signals, being components of the signal $$k_j^l(t)=\{k_{\mathrm{one}}^l(t);k_2^l(t)\},$$

come from the output of the operator controller 8 to the input of the block 7 for calculating the model parameters, in which the current estimates of the coefficients $$k_{\mathrm{one}}^l(t)$$

and $$k_2^l(t)$$

linear-parametric models (9) are converted into the coefficients of the model of the object in increments (4). In particular, if this model is represented by expression (8), then they are recalculated by the formulas

$$k(t)=k_2^l(t);T(t)=-k_2^l(t).(19)$$

Current estimates of the coefficients of model (8), which is a special case of the model of the object in increments (4), are received as signal components k _j (t) = {k (t); T (t)} from block 7 for calculating model parameters into model 13 of the object in increments, as one of the elements of the control object model. This ensures continuous adjustment of the estimates of the coefficients of this model.

The introduction of new blocks and relationships allows you to expand the functionality of the identification system of control objects, i.e. evaluate the structure and parameter values of a mathematical model of an unsteady control object. It also makes it possible to use this system to identify linear and nonlinear control objects, the models of which can be reduced to linear-parametric form.

INFORMATION SOURCES

1.SU 1365047 A1. 07/07/1986.

2. Emelyanov S.V., Korovin S.K. New types of feedback: Management under uncertainty. - M .: Science. Fizmatlit, 1997 .-- 352 p., P. 143, Fig. 3.16.

Claims (3)

1. A control object identification system comprising a master, a control object connected in series, a first comparison unit, a coordinate controller, a control error properties generating unit connected in series, a second comparison unit, an operator controller, the coordinate controller output connected to the input of the control object, the output of which is the first system output, the setter output is connected to the second input of the first comparison unit, characterized in that a control object model is introduced into it, the third a comparison lock and a block for calculating the parameters of the model of the object, the input of which is connected to the output of the operator controller, the output of the block for calculating the parameters of the model is connected to the second output of the identification system of the control object and connected to the first input of the model of the control object, the second and third inputs of which are connected respectively to the input and output management object; the first, second and third outputs of the control object model are connected respectively to the first input of the third comparison unit, to the second and third inputs of the operator controller, the second input of the third comparison unit is connected to the output of the control object, and the output of the third comparison unit is connected to the second input of the second comparison unit and to the input of the block forming the properties of regulation errors. 2. The system according to claim 1, characterized in that the control object model comprises a first delay unit, a fourth comparison unit, an object model in increments and a fifth comparison unit, a second delay unit, the input of which is connected to the third input of the control object model, and the output of the second delay unit is connected to the second input of the fifth comparison unit, the output of which is connected to the first output of the control object model, the output of the object model in increments is connected to the second output of the control object model, the third output which is connected to the first input of the object model in increments, the second input of the fourth comparison unit is connected to the input of the first delay unit and the second input of the control object model, the second input of the object model in increments is connected to the first input of the control object model, which is connected to the output of the model parameter calculation block . 3. The system according to claim 1, characterized in that the operator controller comprises a first connected signal redefining unit, a first squaring unit, an adder, a first division unit, a first multiplication unit, a first integration unit and a first scaling unit, connected in series to the differentiation unit , a second signal redefinition unit, a second division unit, a second multiplication unit, a second integration unit and a second scaling unit, a second squaring unit, the input of which is connected to the WTO output of the second signal override unit, and its output is connected to the second input of the adder, the first input of the operator controller connected to the output of the second comparison unit is connected to the second inputs of the first and second multiplication blocks, respectively, the second input of the operator controller connected to the second output of the control object model, connected to the input of the differentiation unit, the third input of the operator controller connected to the third output of the control object model is connected to the input of the first signal redefinition unit, in the stroke of which is connected to the second input of the first division block, the first input of which is connected to the second input of the second division block, the first input of which is connected to the input of the second squaring block, the outputs of the first and second scaling blocks are connected to the output of the operator controller and are components of its output signal connected to the input of the model parameter calculation block.

Images