SU1290360A1 - Device for predicting production parameters - Google Patents

Abstract

The invention relates to the field of automatic control of technological processes and, in particular, to automatic prediction of technological parameters necessary for optimal control of objects and randomly varying in time. The purpose of the invention is to improve the accuracy of the device. The essence of the invention is that due to the hardware implementation of a complete mathematical model of a random process, including both stationary and non-stationary components, and adaptation to the actual process, the accuracy of the prediction of technological parameters is improved, performed by a single realization of a random process with the criterion Fisher 2 Il. i (L jiND QD O CO

Description

one

The invention relates to automatic control of technological processes and, in particular, to automatic prediction of technological parameters necessary for optimal control of objects and randomly varying with time.

The purpose of the invention is to improve the accuracy of the device.

The essence of the invention is that due to the hardware implementation of the complete mathematical model of a random process, including both stationary and non-stationary components, a more accurate simulation of the stationary component and adaptation to the real process increases the accuracy of the prediction of technological parameters, based on one implementations checked by the Fisher criterion. In this case, the coefficients in the autoregressive model of the stationary component of the random process of the nth order are defined as roots of a system of linear algebraic equations using a well-known digital-numerical device.

1 and 2 illustrate the proposed device.

The device (Fig. 1) contains a data input unit 1, a unit 2 for calculating predicted values, a display unit-3, a two-position switch 4, the first generator 5 clock pulses, the first frequency divider 6, the first meter 7 the average value of a random process, the first algebraic adder 8 , the correlator 9, the first memory block 10, the second clock generator 11, the calculator 12 coefficients, the calculating criterion calculator 13, the comparison block 14, the second memory block 15, the first 16 and the second 17 elements OR.

Node 2 for calculating predicted values (ftsg.2) consists of a second meter 18, the average value of a random process, a second algebraic adder 19, a multiplication unit 20, a accumulating adder 21, a third clock generator 22, and a second frequency divider 23.

The device works as follows.

15

2903602

The signals from the object sensors are received in block 1.. Reading from block 1 is carried out by the generator 5 and frequency divider 6 when the signal is received 5 (if necessary, calculate the new values of the autoregressive coefficients a (n) from the output of block 4. Otherwise, the The output signal from the output of block 4 enters the O node 2.

The meter 7 receives signals from the outputs of block 1 and the frequency divider 6. An estimate of the trend is made, as the arithmetic mean of the sequence of values at each of the k given intervals of constancy using known means.

The input of the adder 8 signals from the outputs of block 1 and measurement 20 rk 7, by subtracting from the elements of the considered sequence the corresponding trend values form a stationary component for k time intervals.

The output signal from the adder 8 enters the correlator 9. The correlation functions obtained in it are stored in memory block 15 and from there go to the calculator 12, where the coefficients are determined and the author of the steady-state model of the stationary component of the process is using the roots of the system of algebraic equations known device. For determination, the next 35 regimen of autoregression, i.e. the numbers n, the Fisher criterion F (n) is used. In a serially connected calculator 13, made according to a known scheme, and block 14, the value of this criterion is determined and the condition F (n) 0 is checked. If the latter condition is met, the coefficients av25

thirty

40

50

The regressions a are fed to block 10. Otherwise, the signal from output 5 of block 14 is fed to the input of generator 11, from which a signal is issued to calculate the autoregressive coefficients a.,.

The control of mixing information from the elements of node 2 and block 1 is carried out by the generator 22 and the frequency divider 23 by the signal from the output of switch 4.

In meter 18, the input of which receives signals from blocks 1 and 23, determines the trend estimate for a tecup; i.e. the interval of constancy:

55

™ u

k.I.n.

t-t

L -. "

In the adder 19, the signals to which come from blocks 1, 23 and 18, the stationary component of the random process is determined for the current constant interval:

 A t block 20 is calculated by the n terms of the autoregression equation

for 1,:.

  one

 + ... 1 a „t.,.,

for which the input of block 20 receives signals from block 10 and adder 10. Signals from the outputs of blocks 18 and 20 go to adder 21, where it is the calculation of the predicted value, technological parameter, which is the sum of the trend estimate Year in the current interval and stationary component,

f t +, in + AD + ... + a „.„. .

The signal from block 21 is fed to the input of block 3 ,. which gives the operator the predicted value of the process.

Claims (1)

Invention Formula A device for predicting technological parameters, comprising a serially connected clock pulse generator and a frequency divider, as well as a data input unit, a prediction value calculation unit, a first memory unit and a display unit, characterized in that, in order to improve the accuracy of the device, it contains position switch, second clock generator, two OR elements and a first meter of a random process connected in series, first algebraic adder, correl A torus, a second block of memory, a coefficient calculator, a calculator of the termination criterion and a comparison unit connected by the first and second outputs respectively to the input of the first memory block and to the first input of the first 2903604 OR, the second input and output of which are associated respectively with the command output of the second memory block and with the start input of the second clock generator 5, connected to the read control input of the second memory block, the information input of the first meter of the average value randomly - The first process and the second input of the first algebraic adder are combined and connected to the information output of the data input unit; the control input of the first meter of the average 5 value of the random process is connected to The first control output of the first frequency divider, the second control output of which is connected via the second element OR to the clock input of the data input unit, the first contacts of the two-position switch are connected to the start input of the first clock generator, the predictive values calculation node contains the second meter the mean value of the random process, the second .algebraic adder, the multiplication unit, the accumulating adder, the third generator of clock pulses and 25 thirty second frequency divider, first up0 five equal output of which is connected to the clock inputs of the second meter of the average value of a random process and with the clock input of the second algebraic adder, the second control output of the second frequency divider 1 is connected with the second input of the second OR element and the second memory control input of the second memory block , the information output of the data input unit is connected to the information input of the second meter of the average value of the random process and to the first information input of the second algebraic adder, the second information input and output of which are connected respectively to the output of the second meter of the average value of the random  process and, through multiplication unit - to ; the first information input on the accumulator is a connected adder the second information input and output, respectively, with the output of the second meter of the average value of a random process and with the input of the display unit, the second input of the multiplication unit is connected with the output of the first block This switch is connected to the start input of the third generator move with the input of the second frequency divider.