WO2009154586A1

WO2009154586A1 - Process for prediction of parameter sensor response

Info

Publication number: WO2009154586A1
Application number: PCT/UA2009/000012
Authority: WO
Inventors: Volodimir Mikolajovich Butsenko
Original assignee: Butsenko Volodimir Mikolajovic
Priority date: 2008-06-19
Filing date: 2009-03-30
Publication date: 2009-12-23
Also published as: UA37415U

Abstract

The utility model relates to measurement technology and can be used in different areas of science and technology, especially in energy industry, for example, for temperature measurement in nuclear reactors, steam generators, shut-off and regulating valves, turbine cylinders having great metal weight. Process for prediction of parameter sensor response involves creating of several correcting circuits, flexible control over input signal of lag elements and determining time constants of first order lag element and second order lag element included into transfer function of parameters of sensor. As a result higher speed and better accuracy of parameters measuring are achieved, which ensures well-timed regulation and control over parameters in different production facilities in service.

Description

PROCESS FOR PREDICTION OF PARAMETER SENSOR RESPONSE

The utility model relates to measurement technology and can be used in different areas of science and technology, especially in energy industry, for example, for temperature measurement in nuclear reactors, steam generators, shut-off and regulating valves, turbine cylinders having great metal weight.

Determination of prediction of sensor response for parameter measured is currently of great interest. With that it is important to determine prediction of sensor response for parameter without delay, for example, to define accurate temperature used in temperature control valve feedback during operation of nuclear power station and by that allow increasing of processing equipment operation reliability of nuclear power station in whole. Besides, it is essential to shorten delay time of parameters of sensor when defining parameters of medical thermometers, power meters and other modulating transducers as well as during examination and setup of automatic control systems. The closest process according to the declared technical solution is the process for prediction of parameter sensor response [1], which includes determination of the first derivative and the second derivative of the parameter measured. Herewith average of the first derivative and the second derivative of the parameter measured is determined on the basis of a definite number of temperature samples with account of additives on each average of determined temperature. It takes about two seconds to predict parameter sensor response. Though this method allows to receive predictions of parameter sensor response, the disadvantage of this method is that predictions of parameter sensor response are determined with delay and therefore accuracy of measurements is not that exact. The reason is that time constants included into transfer functions of parameters of sensor are not considered by this method and consequently speed of parameter measuring is lower.

The present utility model aims at creating of such process for prediction of parameter sensor response, where by creating of several correcting circuits, with the help of flexible control over input signal of lag elements and by determining time constants of first order lag element and second order lag element included into transfer function of parameters of sensor, better accuracy and higher speed of parameter measuring are achieved.

The assigned task is solved such that by means of known process for prediction of parameter sensor response, which includes determination of the first derivative and the second derivative of the parameter measured in accordance to the utility model, time constants of elements of parameter sensor (DO) are determined, at that in advance the first correcting circuit (DOO), comprising at least two elements, is built , output signals of each circuit element are defined, then repetitive structure (for) is set and time from the beginning of calculation (Si) is defined, afterwards, the second correcting circuit (DOK), comprising at least two elements, is built, output signals of each circuit element are defined and then the third correcting circuit (P), comprising one element, is built, after that, the first derivative (DM) of the output signal of the third correcting circuit (P) is determined and constant time of the element (T) of sensor parameter (DO) is calculated, then the calculated time (T) is recorded and it is assigned to the first time (TcII), which is entered into the first element of the second correcting circuit (DOK), then the next time parameter (T) is defined and it is assigned to the second time (Td2), which is afterwards entered into the second element of the second correcting circuit (DOK), herewith, the current level of the calculated time (T) is being determined without interruption, and if it is lower than the minimal boundary of dead zone (Tz1) and the derivative

(DM) is lower than zero, then degree counter (n) subtracts one, but if (T) is more than the maximal boundary of dead zone (Tz2) and the derivative (DM) is more than zero, then degree counter (n) sums up one, at that new values of time of inquiry of parameter sensor (TO) are calculated by the current time (T) and the current minimal boundary of dead zone (Tz1) is defined by its value, when degree counter (n) subtracts one and the current maximal boundary of dead zone (Tz2), when degree counter (n) sums up one, the accuracy boundaries of the calculations (Ch) of the calculated time (T) are defined by reading showed on degree counter (n) without interruptions, the current parameters of the counter are constantly recorded and the time of the closing of the measuring cycle is defined after the calculated times are recorded, at that absolute value of the first derivative (DBO) is defined from the signal of the second correcting circuit (DOK), and if time is less than the value of the set coefficient (K24), then the current value of control variable ( i ) is assigned to cycle continuation (Tl), and after that, output signal of parameter sensor equivalent (DO) is created for the purpose of checking the counter of parameter sensor time constant.

At that output signal of the first correcting element (P01) of the first correcting circuit (DOO) is calculated by formula below:

P01 = D0 + C01 * Kt01, where:

P01 - output signal of the first correcting element; DO - output signal of parameter sensor; C01 - derivative of parameter sensor signal;

KtO1 - first coefficient of amplification from the first time TdO1 ; C01 = D0 - E01; E01 = E01 + C01; E01 - integral from derivative C01; KtO1 = TkO1 * ((TdOI / TO) - K3 + K4 * (TO / TdOI)); TkO1 - coefficient of protection; Td01 - first time of the first correcting element; TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445.

Besides, output signal of the second correcting element of the first correcting circuit (DOO) is calculated by formula below:

DOO = P01 + C02 * KtO2, where: DOO - output signal of the first correcting circuit, which corresponds to output signal of the second correcting element;

P01 - output signal of the first correcting element; C02 - derivative of the first correcting element signal; KtO2 - second coefficient of amplification from the time TdO2; C02 = P01 - E02;

E02 = E02 + C02; E02 - integral from derivative C02; KtO2 = TkO2 *((TdO2 / TO) -K3 + K4 * (TO / TdO2)); TkO2 - coefficient of protection;

TdO2 - second time of the second correcting element; TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445. At that time from the beginning of calculation (Si) up to time for record (T) is calculated by formula below:

Si = Si + TO.

It makes sense to calculate output signal of the first correcting element (P1) of the second correcting circuit (DOK) by formula below:

P1 = D0 + C1 * Kt1, where:

P1 - output signal of the first correcting element; DO - output signal of parameter sensor; C1 - derivative of parameter sensor signal; Kt1 - first coefficient of amplification from the first time Td 1 ; C1 = D0 - E1; E1 = E1 + C1;

E1 - integral from derivative C1; Kt1 = Tk1 *((Td1 / TO) - K3 + K4 * (TO / TdI)); Tk1 - coefficient of protection; Td 1 - first time of the first correcting element; TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445.

Besides, output signal of the second correcting element of the second correcting circuit (DOK) is calculated by formula below:

D0K = P1 + C2 * Kt2, where: DOK - output signal of the second correcting circuit, which corresponds to output signal of the second correcting element;

P1 - output signal of the first correcting element; C2 - derivative of the first correcting element signal;

Kt2 - second coefficient of amplification from the second time Td2;^' C2 = P1 - E2; E2 = E2 + C2; E2 - integral from derivative C2;

Kt2 = Tk2 * ((Td2 / TO) - K3 + K4 * (TO / Td2)); Tk2 - coefficient of protection; Td2 - second time of the second correcting element;

TO - time of inquiry of parameter sensor; K3 -coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445.

It is preferably to calculate output signal of the third correcting circuit (P) by formula below: P = D0K+ DE0 * Q, where:

P - output signal of the third correcting circuit; DOK - output signal of the second correcting circuit; DEO - first derivative from output signal of the second correcting circuit DOK; Q- integral, which is proportional to calculated time T; DEO = DOK- F; F = F + DEO;

F - integral from derivative DEO; Q = Q + A1 * Y* W; A1 - temporary variable, which manipulates integral;

Y - coefficient, which allows/disallows to calculate integral that is defined through the second derivative DK from the signal of the second correcting circuit DOK;

W - coefficient of stability of integral calculation, which changes every tact from zero into one and vice versa; AI = DM *KU * Q ²; DM - first derivative of output signal of the third correcting circuit (P);

KU - coefficient of amplification dependable from the level of the output signal of the second correcting circuit DOK and actuating signal KO of parameter sensor.

Besides, element time constant (T) of parameter sensor (DO) is defined in accordance to time of inquiry of parameter sensor (TO) and is calculated by formula below: T = T0 * ((Q + K3) - K7/ (Q + K3)), where: T - element time constant of parameter sensor DO; TO - time of inquiry of parameter sensor; Q - integral;

K3 - coefficient, which is equal to 0.5; K7 - coefficient, which is equal to 0.0833338887.

It makes sense to calculate time of inquiry of parameter sensor (TO) by formula below:

T0= pow (K, n)*K10, where: TO - time of inquiry of parameter sensor; pow (K, n) - number K raised to n power;

K - coefficient, which is equal to 2; K10 - coefficient, which is equal to 0.000001.

At that value of the minimal boundary of dead zone (Tz1) is calculated by formula below: Tz1 = T0 * K14, where: Tz1 - minimal boundary of dead zone; K14 - coefficient, which is equal to 250.

Besides, value of the maximal boundary of dead zone (Tz2) is calculated by formula below:

Tz2 = T0 * K16, where: Tz2 - maximal boundary of dead zone; K16 - coefficient, which is equal to 750.

It is preferably to calculate the value of the first output signal (D1) of parameter sensor equivalent (DO) by formula below:

D1 = D1 + (K0- D1) /t1, where: D1 - first output signal of parameter sensor equivalent (DO) (output signal of the first lag element); KO - actuating signal of parameter sensor; t1 - time of the first lag element from the first time to be set (T1 ); t1 = (T1 / T0) + K3 + K4 * (T0 / T1); T1 - first time to be set of the first element; TO - time of inquiry of parameter sensor;

K3 - coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445.

Besides, value of the second output signal (D2) of parameter sensor equivalent (DO) is calculated by formula below: D2 = D2 + (D1 - D2) /t2, where: D2 - second output signal of parameter sensor equivalent (DO) (corresponds to output signal of the second lag element);

D1 - output signal of the first lag element; t2 - time of the second lag element from the second time to be set (T2); t2 = (T2/T0) + K3 + K4 * (T0 /T2);

T2 - second time to be set of the second element; TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445. At that, checking of the counter is performed by output signal of parameter sensor equivalent (DO):

DO = DI * d1 + D2 * d2, where: D1 - output signal of the first lag element; d1 - coefficient of choice, if it is equal to one, then coefficient of choice d2 is equal to zero and vise versa; D2 - output signal of the second lag element, which corresponds to two elements; d2 - coefficient of choice.

For improvement of the process for prediction of sensor response for parameter measured there were sequentially formed three correcting circuits, comprising elementary lag elements and high-speed elements, generated especially for a flexible control over output signal and which are also included into transfer function of parameter sensor equivalent. At that coefficient of proportionality of the elements is admitted to be equal to one.

Output signal of the 1 st-order lag element (LE) is usually calculated at the output signal from zero to one and original zero time by the following formula:

Xoutput = 1 -exp (-t/T), (1) where: Xoutput - output signal of LE; t- summarized current time; T - time constant of LE.

Output signal of LE, calculated by formula (1) is replaced with output signal of developed element of LE and calculated by the following formula: Xoutput = Xoutput + (Xinput - Xoutput) /Tie, (2) where: Xoutput - output signal of LE;

Xinput - input signal changes from zero to one at actuating signal of staircase type; Tie - time constant of new LE.

Output signals of LE represented in formula (1) and formula (2) differ in accuracy by their transfer characteristics. Experimentally and taking into account entered constant coefficients new time Tie, which corresponds to time T has been corrected and it is calculated by the following formula:

Tle=(T/T0) + K3 + K4 * (T0 / T), (3) where: Tie - corrected time constant of LE;

T - time constant of LE, which should not be equal to zero; TO - time of inquiry of parameter sensor;

Corrected time constant of LE received by such formula improves precision of transfer between transient characteristics of LE represented in formula (1) and formula (2) and allows fifteen decimal places of accuracy. The algorithm of formation of output signal of LE represented in formula (2) and formula (3) is used further during scheduled inspection of the entire calculation algorithm for time constants of parameter sensor elements as one element of parameter sensor equivalent.

Output signal of high-speed element (HSE) is usually calculated at the input signal from zero to one and at the original zero time by the following formula: Xoutput = exp (-t/ Td), (4) where: Xoutput - output signal of HSE; t - summarized current time; Td -time constant of HSE. Signal of high-speed element defined by formula (4) is replaced with signal of developed highspeed element and calculated by the following formula:

Xoutput = C* KT, (5) where: Xoutput - output signal of high-speed element (HSE); C- derivative of input signal Xinput;

KT - coefficient of amplification, which corresponds to time constant Td. At that derivative C of input signal Xinput is calculated by formula:

C = Xinput -E, (5.1) where: Xinput - input signal changes from zero to one at actuating signal of staircase type; E- integral from derivative C.

At that integral E from derivative C is calculated by the following formula:

E = E + C. (5.2)

Output signals of high-speed elements represented in formula (4) and formula (5) differ in accuracy by their transfer characteristics. There was experimentally defined corrected coefficient of amplification of derivative of output signal KT with account of entered time coefficients and time constant of high-speed Td, which is calculated by the following formula:

KT = Kz*((Td / TO) - K3 + K4 * (TO / Td)), (6) where: KT - corrected coefficient of amplification;

Kz - coefficient of protection; Td -time constant of HSE;

TO - time of inquiry of parameter sensor;

K3 - coefficient, which is equal to 0.5;

K4 - coefficient, which is equal to 0.083333194445.

Corrected coefficient of amplification of derivative of output signal KT received in such a way improves precision of transfer between transient processes of output signals of high-speed elements represented in formula (4) and formula (5) and allows 15 decimal places of accuracy.

The algorithm of formation of output signal of high-speed element defined by formula (5) is used further in the sequential correcting device (SCD) directly in the circuit of parameter sensor and in counter of times of parameter sensor. Output signal of sequential correcting device (SCD) is defined in accordance to signal of parameter sensor and output high-speed signal multiplied by coefficient of amplification, which corresponds to time of one high-speed element, and is calculated by following formula:

Xoutput = Xinput + C * KTK, (7) where: Xoutput - output signal of SCD; Xinput - input signal;

C - derivative of input signal Xinput; KTk - corrected coefficient of amplification (calculated in a similar way to formula 6). In such a way time constants of the 1st-order lag element and the 2nd-order lag element were defined, which are included into transfer function of parameter sensor and as a result higher speed and better accuracy of parameter measuring were achieved.

The main point of the utility model is described on the figures, wherein fig.1 is the structure of the connection between elements of time counter; fig. 2 is an algorithmic block diagram of the formation of coefficients of protection of the first correcting circuit depending on level of times entered; fig. 3 is an algorithmic block diagram of the first correcting circuit located sequentially with parameter sensor with input up to two times so far as the times are calculated; fig.4 is an algorithmic block diagram of the formation of coefficients of protection of the second correcting circuit depending on level of coefficients, which define record of calculated times of parameter sensor; fig. 5 is an algorithmic block diagram of the second correcting circuit of counter of times of parameter sensor located sequentially with parameter sensor with input up to two times into them so far as the times are calculated; fig. 6 is an algorithmic block diagram of the third correcting circuit located sequentially with the second correcting circuit with input of the signal of the calculable integral therein instead of coefficient of amplification; fig.7 is an algorithmic block diagram of the calculation of integrand, coefficient of amplification of integrand and time of the element of parameter sensor; fig. 8 is an algorithmic block diagram of sequential operating record for measuring of time constants of parameter sensor; fig. 9 is an algorithmic block diagram of measuring of numerical value of the counter of time of inquiry of parameter sensor, its maximal and minimal boundaries of dead zone; fig. 10 is an algorithmic block diagram of determining boundary of feasible accuracy of measuring of parameter sensor element time; fig. 11 is an algorithmic block diagram of determining end of cycle of calculation of parameter sensor element time constants; fig. 12 is an algorithmic block diagram of determining the moment for current parameters record of the counter when parameter sensor time is being measured; fig. 13 is an algorithmic block diagram of determination of parameter sensor equivalent.

Process for prediction of parameter sensor response is implemented in the following way. Algorithm of parameter sensor response measurement is shown on fig. 1.

Input signal of parameter sensor DO, for example, comes on one element of sequential correcting circuit, comprising of the first summator ∑1, on the first input of which the input signal of parameter sensor DO comes and integrator signal E1 with reversed sign from the first derivative C1 comes on the second input.

Input signal of parameter sensor DO comes on the first input and on the second summator 12, but the first derivative of sensor C1 multiplied by coefficient of amplification Kt1 corresponding to calculated element time of parameter sensor Td1 comes on the second input.

Corrected coefficient of amplification Kt1 is calculated in a similar way to formula 6 described above. At that if coefficient of amplification KT is defined as Kt1 and coefficient of protection Kz is defined as Tk1, then coefficient of amplification Kt1 is calculated in accordance to formula below: KM = Tk1 * ((TdI / TO) - K3 + K4 * (TO / TdI)). (8)

Output of the second summator DOK depending on time Td1, which is entered or not entered, is corrected or non-corrected signal of parameter sensor DO. Then, it is considered that time Td1 is not entered, coefficient of protection Tk1 is equal to zero, and so, output signal DOK will correspond to input signal of parameter sensor DO.

Measuring of time constant of parameter sensor consists of sequential correcting circuit with integrator F instead of E1 and the first derivative of sensor DEO instead of C1 multiplied by value of integrator Q instead of Kt1 and output signal P instead of DOK.

Level of integral Q corresponding to calculable time constant of parameter sensor T is calculated. At first, derivative DM of signal P is calculated, sign of this derivative is defined depending on the sign of the first derivative DEO of input signal of parameter sensor DO.

Then, the second derivative DK is calculated; this derivative is necessary for calculation of coefficient Y - this is the coefficient, which allows/disallows to calculate integral that is defined through the second derivative DK from the signal of the second correcting circuit DOK. If the second derivative DK of the input signal DOK is less than or equal to zero, level of one is assigned to coefficient Y and this allows to change integrand Q, but if the second derivative DK is more than zero, level of zero is assigned to coefficient Y and this disallows to change integrand Q.

Afterwards there is calculated signal of integrand A1 of integral Q, which is equal to derivative DM multiplied by coefficient of amplification KU and by integral Q raised to the second power. Before input of signal A1 into integrand Q, signal A1 is limited by level from two sides.

Then there is calculated integral Q, which is equal to sum of integral Q and signal A1 multiplied by coefficients W and Y.

After that element time constant of parameter sensor T is calculated with a glance to corresponding constants of the coefficients by formula below:

T = T0 * ((Q + K3) - K7 /(Q + K3)) (9) where: T - constant time of parameter sensor element DO; TO - time of inquiry of parameter sensor; Q - signal of integral; K3 - coefficient, which is equal to 0.5;

K7 - coefficient, which is equal to 0.0833338887.

Algorithm of the formation of coefficients of protection TkO1 and TkO2 of the first correcting circuit depending on level of times entered is shown on fig. 2 in the following way: Start 1.1 - End 1.1, comprising of the first correcting circuit DOO, which includes determinant of level of the first time TdO1 (1), and if it is less than zero or equal to zero, level of zero is assigned to coefficient of protection TkO1 (2), otherwise level of one is assigned to coefficient TkO1 (3). Regarding determinant of the second time TdO2 (4), if it is less than zero or equal to zero, level of zero is assigned to coefficient of protection TkO2 (5), otherwise level of one is assigned to coefficient TkO2

(6).

Fig.3 shows algorithm of the first correcting circuit DOO located sequentially with parameter sensor DO with input up to two times so far as the times are calculated: Start 1.2 - End 1.2, derivative C01 (7) of sensor signal DO and integral E01 (7) from derivative C01 are calculated. Coefficient of amplification KtO1 (8) is calculated from the first time TdO1, comprising of coefficient of protection TkO1 multiplied by bracket, in which coefficient K3 is subtracted from quotient of the first time TdO1 divided into time of inquiry of sensor signal TO, and then coefficient K4, multiplied by quotient of time of inquiry of sensor signal TO divided into the first time TdO1, is added.

Calculation of signal of the first correcting circuit P01 (9) is equal to the sum of signal of parameter sensor DO and derivative C01 multiplied by coefficient of amplification KtOL There is calculated the second correcting element comprising of calculations: derivative signal C02 (10) of the first correcting element P01 and integral E02 (10) from derivative C02. Coefficient of amplification KtO2 (11) is calculated from the second time TdO2, comprising of coefficient of protection TkO2 multiplied by bracket, in which coefficient K3 is subtracted from quotient of the second time TdO2 divided into time of inquiry of sensor signal TO, and then coefficient K4, multiplied by quotient of time of inquiry of sensor signal TO divided into the second time TdO2, is added. Calculation of output signal of the first correcting circuit DOO (12) is equal to the sum of signal of the first correcting element P01 and its derivative C02 multiplied by coefficient of amplification KtO2.

Then coefficients of protection Tk1 and Tk2 of the second correcting circuit are defined. Algorithm of the formation of coefficients of protection of the second correcting circuit depending on level of coefficients, which define record of calculated times of parameter sensor is shown on fig. 4 and it consists of: Start 2.1 - End 2.1, determinant of level of the coefficient a1 (13), and if it is equal to zero, level of zero is assigned to coefficient of protection Tk1 (14), otherwise level of one is assigned to coefficient Tk1 (15); determinant of level of coefficient a2 (16), and if it is equal to zero, level of zero is assigned to coefficient of protection Tk2 (17), otherwise level of one is assigned to coefficient Tk2 (18).

Algorithm of the second correcting circuit of counter of times of parameter sensor located sequentially with parameter sensor with input up to two times into them so far as the times are calculated is shown on fig. 5 and consists of Start 2.2 - End 2.2, repetitive structure for (19), which includes control variable i , conditions of cycle continuation Tl, sum of control variable i++. Total current time from the beginning of calculation Si (19) is calculated. There are calculated the first correcting element P1 and the second correcting circuit DOK comprising of calculations: derivative C1 (20) of sensor signal DO and integral E1 (20) from derivative of signal C1. Coefficient of amplification Kt1 (21) is calculated from the first time Td 1, comprising of coefficient of protection Tk1 multiplied by bracket, in which coefficient K3 is subtracted from quotient of the first time Td 1 divided into time of inquiry of sensor signal TO, and then coefficient K4, multiplied by quotient of time of inquiry of sensor signal TO divided into the first time Td1, is added. Calculation of signal of the first correcting element P1 (22) is equal to the sum of signal of parameter sensor DO and derivative C1 multiplied by coefficient of amplification KtL There is calculated the second correcting circuit DOK comprising of calculations: derivative C2 (23) of signal P1 and integral E2 (23) from derivative C2. Coefficient of amplification Kt2 (24) is calculated from the second time Td2, comprising of coefficient of protection Tk2 multiplied by bracket, in which coefficient K3 is subtracted from quotient of the second time Td2 divided into time of inquiry of sensor TO, and then coefficient K4, multiplied by quotient of time of inquiry of sensor signal TO divided into the second time Td2, is added. Calculation of output signal of the second correcting circuit DOK (25) is equal to the sum of signal of the first correcting circuit P1 and its derivative C2 multiplied by coefficient of amplification Kt2. Level of signal DOK (26) is defined, if it is less than or equal to zero, then level of zero is assigned to signal DOK (27).

Algorithm of the third correcting circuit located sequentially with the second correcting circuit with input of the signal of the calculable integral therein instead of coefficient of amplification is shown on fig. 6, comprising of: Start 3 - End 3, calculation of the first derivative DEO (28) of the signal of the second correcting circuit DOK and integral F (28) from derivative DEO. The third correcting circuit P (28) is calculated and it is equal to the sum of the signal of the second correcting circuit DOK and its derivative DEO multiplied by level of integral Q. Sign of derivative DEO (29) is defined and if it is more than or equal to zero, then derivative DM (30) of the third correcting circuit P is calculated and it is equal to difference of signal P and its previous value PZ; the second derivative DK

(30) of the signal of the second correcting circuit DOK is calculated and it is equal to difference of derivative DEO and its previous value DEZ. If the signal of derivative DEO (29) is less than zero, then derivative DM (31) is calculated and it is equal to difference of previous values of derivative PZ and signal P; the second derivative DK

(31) of the signal of the second correcting circuit DOK is calculated and it is equal to difference of previous values of derivative DEZ and derivative DEO. Assignment is performed: value of signal of the third correcting circuit P is assigned to signal PZ (32); value of derivative DEO is assigned to signal DEZ (32). Sign of the second derivative DK (33) is defined, and if it is less than or equal to zero, then level of one is assigned to coefficient Y (34), which allows to change integrand Q, otherwise level of zero is assigned to coefficient Y (35), which disallows to change integrand Q. Then coefficient of amplification KU of integrand A1 , integral Q and time T is calculated.

Algorithm of the calculation is represented on fig. 7 in the following way: coefficient of amplification KU is formed from: Start 4 - End 4, assignment of level of signal DOK to signal C (36), which limits the maximal coefficient of amplification KU. Sign of derivative DEO (37) is defined, and if it is less than zero, then temporary coefficient k (38) is calculated and it is equal to the sum of actuating signal KO of parameter sensor and coefficient K5. Level of signal C (39) is defined, if it is less than or equal to coefficient k, then level of signal k is assigned to signal C (40). Coefficient of amplification KU (41) is calculated and it is equal to one divided into difference of signal C and actuating signal KO. If signal of derivative DEO (37) is more than or equal to zero, then temporary coefficient k (42) is calculated and it is equal to difference of actuating signal KO of parameter sensor and coefficient K5. Level of signal C (43) is defined, if it is more than or equal to coefficient k, then level of signal k is assigned to signal C (44). Coefficient of amplification KU (45) is calculated and it is equal to one divided into difference of actuating signal KO and signal C. Signal of integrand A1 (46) is calculated and it is equal to derivative DM multiplied by coefficient of amplification KU and integral Q raised to the second power. Signal A2 (46) is calculated and this signal is aimed at limiting of level of signal A1 , which is equal to level of integral Q multiplied by coefficient K6. Level of signal A1 (47) is defined, if it is more than or equal to signal A2, level of signal A2 is assigned to A1 (48); if signal A1 (49) is less than negative signal A2, then negative level of signal A2 is assigned to A1 (50). Integral Q (51) is calculated and it is equal to sum of level of integral Q and signal A1 multiplied by coefficients Y and W, in its turn, coefficient W is entered into integrand Q for more accurate calculation. Element time constant T (51) is calculated and it comprises of time of inquiry of parameter sensor TO multiplied by bracket, in which coefficient K7 divided into sum of level of integral Q and coefficient K3, is subtracted from the sum of level of integral Q and coefficient K3. Then sequential operating record for measuring of time constants of parameter sensor is performed, algorithm of which is shown on fig. 8 in the following way: Start 5 - End 5, determinant of level of signal Y (52), and if it is equal to zero, level of zero is assigned to reading of the counter ch1 (53); if level of signal Y is equal to one, counter ch1 (54) increases its reading by one. Level of counter ch1 (55) is calculated, if reading of counter ch1 is more than or equal to coefficient K8, then difference of DB (56) between level of the current calculable time T and level of its previous value BN is calculated. The current value of time level T is assigned to signal BN (56) and level of zero is assigned to counter ch1 (56). Level of signal DB (57) is defined, if absolute value of signal DB is less than or equal to accuracy of calculation Ch, record of the result of time T calculation is performed: value of coefficients a1 and a2 (58) are defined: if coefficients a1 and a2 are equal to zero, record of time Td1 is not performed. Value of coefficient a1 (62) is defined: if coefficient a1 is equal to zero, record of the first time Td1 (63) is performed: level of one is assigned to coefficients a1 and Tk1; the first calculated time T is assigned to time of the first correcting lag element Td1; value of summarized current time Si is assigned to time Ti1, which registers total time from the beginning of measurement till the record of the first measured time Td1; coefficient of actuating K1 of parameter sensor - low-level signal is assigned to command of actuating KO, if before signal of higher level K2 was entered, coefficient K9 is assigned to initial level of integral Q; coefficient K24 is assigned to accuracy of calculation Ch; zero is assigned to counter of power n; coefficient K10 is assigned to time of inquiry of parameter sensor TO; coefficient K11 is assigned to minimal boundary of dead zone Tz1; coefficient K12 is assigned to maximal boundary of dead zone Tz2. Difference of DTd 1 (63) is calculated and it shows possibility of re-recording of a new value of Td 1 in the first correcting element of the first correcting circuit DOO, if by absolute value difference of DTd 1 (64) between newly calculated time Td 1 and time TdO1 previously recorded is less than or equal to parameter d, then newly calculated time T is assigned to time TdO1 (65) and level of zero is assigned to counter ch3 (66). After next calculation of time T is performed, value of coefficients a1 and a2 (58) are defined again: if coefficient a1 is equal to one and at the same time coefficient a2 is equal to zero, time Td2 (59) is assigned to newly calculated time T, and so, wherefore level of one is assigned to coefficients a2 and Tk2; the second calculated time T is assigned to time of the second correcting element Td2; time Si is assigned to time Ti2, which registers total time from the beginning of measurement till the record of the second measured time Td2. Difference of DTd2 (59) is calculated, and it shows possibility of re-recording of a new value of Td2 in the second element of the first correcting circuit DOO, which is equal to difference between newly recorded time Td2 and time TdO2 previously recorded. Difference signal level DTd2 (60) is calculated, if absolute value of signal of difference DTd2 is less than or equal to parameter d, newly calculated time T is assigned to time TdO2 (61), and level of zero is assigned to counter ch3 (66).

Algorithm of measuring of numerical value of the counter of time of inquiry of parameter sensor and minimal and maximal boundaries of dead zone is shown on fig. 9 in the following way: Start 6 - End 6, determinant of level of signal W (67), and if signal W is equal to zero, counter ch2 (68) increases its reading by one. Level of counter ch2 (69) is calculated, if reading of counter ch2 is more than or equal to coefficient K13, then level of the calculable time T (70) is calculated, if level T is less than minimal boundary of dead zone Tz1 and at that derivative DM is less than zero, then one is subtracted from counter n (71) from numerical value of power of K. Level of counter n (72) is calculated, if number n on counter is less than or equal to zero, level of zero is assigned to counter n (73). Time of inquiry TO (74) of parameter sensor signal is calculated and it is equal to value of number K raised to power n and multiplied by coefficient K10. Minimal boundary of dead zone Tz1 (74) and it is equal to time of inquiry TO of parameter sensor signal multiplied by coefficient K14. Level of calculable time T (75) is calculated, if level of time T is more than maximal boundary of dead zone Tz2 and at that derivative DM is more than zero, then counter of number n (76) adds one. Level of counter n (77) is defined, if number n on the counter is more than or equal to coefficient K15, level of coefficient K15 is assigned to counter n (78). Time of inquiry TO (79) of parameter sensor is calculated and it is equal to value of number K raised to power n and multiplied by coefficient K10. Maximal boundary of dead zone Tz2 (79) is calculated and it is equal to time of inquiry TO of parameter sensor multiplied by coefficient K16. Level zero is assigned to counter ch2 (80). Then boundary of feasible accuracy of measuring Ch of parameter sensor element time T depending on counter numerical value in power n is determined, algorithm of which is shown on fig. 10 in the following way: Start 7 - End 7, determinant of number level n 81, 83, 85, 87, 89, 91, 93, if number n is less than or equal to numerical number of value of coefficient K17 ÷ K23, then numerical numbers of coefficients of accuracy K24 ÷ K30 are assigned accordingly in each particular case of feasible accuracy Ch 82, 84, 86, 88, 90, 92, 94. Level of calculable time T (95) is defined: if level of calculable time T is more than or equal to numerical number of coefficient K31, then value of coefficient K32 is assigned to feasible accuracy Ch (96).

Then end of cycle of calculation of element time constants of parameter sensor T is determined, algorithm of which is shown on fig. 11 in the following way: Start 8 - End 8, determinant of level of coefficient Y (97) is formed, and if it is equal to one, then levels of coefficients a1 and a2 (98) are defined, if coefficient a1 or a2 is equal to one, then counter ch3 (99) adds one. Reading of counter ch3 (100) is defined, if reading of counter ch3 is more than or equal to coefficient K33 and at that absolute value of the first derivative DEO (101) from signal of the second correcting circuit DOK is less than or equal to accuracy of calculation of K24, then current value of control variable i is assigned to condition of cycle continuation Tl (102), which aborts calculation cycle.

If necessary, the moment for current parameters record of the counter is calculated additionally, when parameter sensor time T is being defined, and its algorithm is shown on fig. 12 in the following way: Start 9 - End 9, determinant of level of coefficient W (103) is formed, if it is equal to one, then zero is assigned to coefficient W (104) and if it is equal to zero, one is assigned to coefficient W (105). Moment for calculated current parameters record is defined, if coefficient W (106) is equal to zero and at that signal Si is more than zero, then the following signals for print or registration (107) are displayed: Si - summarized current time of calculations, DO - input signal of parameter sensor, KO - output actuating signal, which goes onto parameter sensor, DOK - output corrected signal of parameter sensor, T - current calculable time of the parameter sensor element, TO - time of inquiry of parameter sensor and coefficient Y, which allows/disallows to calculate time constant T of parameter sensor element Condition of cycle continuation Tl (108) is calculated, if it is equal to the current value of control variable i, then following parameters (109) are recorded only once' Ti1, which is time before record of the first calculated time Td1 and time Td1 itself; Ti2 is time before record of the second calculated time Td2 and time Td2 itself or temperature reading, for example, signal DOO on medical thermometer. Parameter sensor equivalent is formed for scheduled inspection of the counter of time constant T of parameter sensor element, algorithm of which is shown on fig. 13 in the following way: Start 10 - End 10, first time t1 (110) is calculated in accordance to set time of the first circuit T1 , which is equal to the sum: quotient of time T1 divided into time of inquiry of parameter sensor TO and coefficients K3 and K4, coefficient K4 is multiplied by quotient of time of inquiry of parameter sensor TO divided into set time T1, which should not be equal to zero. Output signal of the first lag element (111) is equal to sum of signal D1 and difference of actuating signal KO and signal of the first element D1 divided into time tl Second time t2 (112) is calculated in accordance to set time of the second circuit T2, which is equal to the sum: quotient of time T2 divided into time of inquiry of parameter sensor TO and coefficients K3 and K4, coefficient K4 is multiplied by quotient of time of inquiry of parameter sensor TO divided into set time T2, which should not be equal to zero. Output signal of the second lag element D2 (113) is equal to sum of signal D2 and difference of signal of the first element D1 and signal of the second element D2 divided into time t2. Output signal of parameter sensor equivalent DO (113) is equal to the sum of the compositions: signal of the first element D1 multiplied by coefficient d1 - output of the 1st-order lag element and signal of the second element D2 multiplied by coefficient d2 - output of the 2nd-order lag element. Whether one should be assigned to d1 or zero should be assigned to d2, or one should be assigned to d2, or zero should be assigned to d1 is defined by output signal of parameter sensor equivalent DO.

Numerical values of coefficients used in calculable algorithm are defined by empirical approach and represented in table 1.

Results of calculations of time constants of parameter sensor are represented in tables 2, 3, 4, in which the following current parameters are recorded in progress: Si, DO, KO, DOK, T, TO and Y when calculating time constant of parameter sensor element T.

Table 1.

Table 2

Tl=180 Kl =0.25 K2=0.366

Til=14.680060 s. TdI=I 80.0000000125 s.

Table 3

Tl=36000 Kl=0.25 K2=0.366

Ti 1=80.740348 s. TdI =36000.0050790888 s.

Table 4

Til=441.450492 s. Ti2=445.186155 s.

Tdl=480.0074563413 s. Td2=29.9999668771 s.

So, table 2 shows calculations of time constant T of one lag element of parameter sensor with the following original input parameters: set time of one lag element T1 is equal to 180 sec. (3 minutes); original temperature 25 ⁰C is represented by coefficient K1, which is equal to 0.25; then temperature jumped up to 36.6⁰C and applied to parameter sensor input, is represented by coefficient K2, which is equal to 0.366. Final output parameters of the counter are the following: summarized time of the calculation at the first Ti1 switch being active is 14.680060 sec. and calculated time Td1 is 180.0000000125 sec. After TM time and Td1 time are calculated, input temperature is changed from 36.6⁰C down to 25⁰C and then again up to 36.6⁰C together with time of inquiry of parameter sensor TO being changed. As it can be seen on the table, output temperature parameter follows input temperature changes during 1-2 microseconds rapidly. As a result, input lag element with set time T1 equal to 180 sec. (3 minutes) is reduced and in such a way, higher speed and better accuracy of parameter sensor prediction are achieved. Table 3 shows calculations of time constant T of one lag element of parameter sensor with the following original input parameters: set time of one 1^st order lag element T1 is equal to 36000 sec. (10 hours); original input parameters are the following: original temperature 25 ⁰C is represented by coefficient K1, which is equal to 0.25; then temperature jumped up to 36.6 ⁰C and applied to parameter sensor input, is represented by coefficient K2, which is equal to 0.366. Final output parameters of the counter are the following: summarized time of the calculation at the first TM switch being active is 80.740348 sec. and calculated time Td 1 is 36000.0050790888 sec. After TM time and Td1 time are calculated, input temperature is changed from 36.6 ⁰C down to 25 ⁰C and then again up to 36.6 ⁰C together with time of inquiry of parameter sensor TO being changed. When input signal is changed, output parameter follows input signal during 1-2 microseconds rapidly. This also confirms, that input 1^st order lag element with set time equal to 36000 sec. (10 hours) is reduced and in such a way, higher speed and better accuracy of parameter sensor prediction are achieved.

Table 4 shows calculations of time constant T of the 2^nd lag element with the following original input parameters: set times of elements T1 are equal to 480 sec. (8 minutes) and T2 is equal to 30 seconds, original temperature 25⁰C is represented by coefficient K1, which is equal to 0.25; then temperature of the 2^nd lag element jumped up to 36.6 ⁰C and applied to parameter sensor input is represented by coefficient K2, which is equal to 0.366. Final output parameters of the counter are the following: summarized time of the calculation of the first time TM is 441.450492 sec. at the first switch, and calculated first time Td1 is 480.0074563413 sec, but summarized time of the calculation of second time Ti2 is 445.186155 sec. at the first switch and time Td2 is 29.9999668771 sec. After TM time, Td1 time, Ti2 time and Td2 are calculated, input temperature is changed from 36.6⁰C down to 25⁰C, and output parameter temperature follows reading of input temperature 25⁰C during 0.0655 sec. rapidly. As it can be seen on the table, input 2^nd order lag element with set time T1 equal to 480 sec. (8 minutes) and time T2 equal to 30 sec. are reduced and that means that highei speed and better accuracy are achieved.

So, suggested process for prediction of parameter sensor response allows to calculate time constant of the 1^st and 2^nd order lag elements, included into transfer function of parameter sensor and prediction of its response, as a result of which higher speed and better accuracy of parameters measuring are achieved, which ensures well-timed regulation and control over parameters in different production facilities in service.

References cited 1. International Application WO/2000/070316, publ. 23.11.2000, MfIK- G01 K7/42.

Claims

C l a i m s

1. Process for prediction of parameter sensor response, comprising the steps of: the first derivative and the second derivative of the parameter measured are determined; time constants of elements of parameter sensor (DO) are determined; at that in advance the first correcting circuit (DOO), comprising at least two elements, is built; output signals of each circuit element are defined; then repetitive structure (for) is set and time from the beginning of calculation (Si) is defined; afterwards, the second correcting circuit (DOK), comprising at least two elements, is built; output signals of each circuit element are defined; and then the third correcting circuit (P), comprising one element, is built; after that, the first derivative (DM) of the output signal of the third correcting circuit (P) is determined and constant time of the element (T) of sensor parameter (DO) is calculated; then the calculated time (T) is recorded and it is assigned to the first time (Td1), which is entered into the first element of the second correcting circuit (DOK); then the next time parameter (T) is defined and it is assigned to the second time (Td2), which is afterwards entered into the second element of the second correcting circuit (DOK), herewith, the current level of the calculated time (T) is being determined without interruption, and if it is lower than the minimal boundary of dead zone (Tz1) and the derivative (DM) is lower than zero, then degree counter (n) subtracts one; but if (T) is more than the maximal boundary of dead zone (Tz2) and the derivative (DM) is more than zero, then degree counter (n) sums up one, at that new values of time of inquiry of parameter sensor (TO) are calculated by the current time (T) and the current minimal boundary of dead zone (Tz1) is defined by its value, when degree counter (n) subtracts one and the current maximal boundary of dead zone (Tz2), when degree counter (n) sums up one; the accuracy boundaries of the calculations (Ch) of the calculated time (T) are defined by reading showed on degree counter (n) without interruptions; the current parameters of the counter are constantly recorded; and the time of the closing of the measuring cycle is defined after the calculated times are recorded; at that absolute value of the first derivative (DBO) is defined from the signal of the second correcting circuit (DOK), and if time is less than the value of the set coefficient (K24), then the current value of control variable ( i ) is assigned to cycle continuation (Tl); and after that, output signal of parameter sensor equivalent (DO) is created for the purpose of checking the counter of parameter sensor time constant.

2. The process according to claim 1, wherein said output signal of the first correcting element (P01) of the first correcting circuit (DOO) is calculated by formula below:

P01 = D0 + C01 * KtO1, where: P01 - output signal of the first correcting element; DO - output signal of parameter sensor;

C01 - derivative of parameter sensor signal; KtO1 - first coefficient of amplification from the first time TdO1; C01 = DO - E01; E01 = E01 + C01;

E01 - integral from derivative C01 ; KtO1 = TkO1 * ((TdOI / TO) - K3 + K4 * (TO / TdOI)); TkO1 - coefficient of protection; Td01 - first time of the first correcting element; TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.5; K4 - coefficient, which is equal to 0.083333194445.

3. The process according to claim 1, wherein said output signal of the second correcting element of the first correcting circuit (DOO) is calculated by formula below:

DOO = P01 + C02 * Kt02, where: DOO - output signal of the first correcting circuit, which corresponds to output signal of the second correcting element;

P01 - output signal of the first correcting element; C02 - derivative of the first correcting element signal;

KtO2 - second coefficient of amplification from the time TdO2;

C02 = P01 - E02;

E02 = E02 + C02;

E02 - integral from derivative C02; KtO2 = TkO2 *((TdO2 / TO) -K3 + K4 ^* (TO / TdO2));

TkO2 - coefficient of protection;

TdO2 - second time of the second correcting element;

TO - time of inquiry of parameter sensor;

4. The process according to claim 1, wherein said time from the beginning of calculation (Si) up to time for record (T) is calculated by formula below:

Si = Si + TO. 5. The process according to claim 1, wherein said output signal of the first correcting element (P1) of the second correcting circuit (DOK) is calculated by formula below:

P1 = DO + C1 * Kt1 , where: P1 - output signal of the first correcting element; DO - output signal of parameter sensor; C1 - derivative of parameter sensor signal; Kt1 - first coefficient of amplification from the first time Td 1 ; C1 = D0 - E1; E1 = E1 + C1;

E1 - integral from derivative C1 ; Kt1 = Tk1 *((Td1 / TO) - K3 + K4 * (TO / TdI)); Tk1 - coefficient of protection; Td 1 - first time of the first correcting element; TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.

5; K4 - coefficient, which is equal to 0.083333194445.

6. The process according to claim 1, wherein said output signal of the second correcting element of the second correcting circuit (DOK) is calculated by formula below:

Kt2 - second coefficient of amplification from the second time Td2;

C2 = P1 - E2;

E2 = E2 + C2;

E2 - integral from derivative C2; Kt2 = Tk2 * ((Td2 / TO) - K3 + K4 * (TO / Td2));

Tk2 - coefficient of protection;

Td2 - second time of the second correcting element;

TO - time of inquiry of parameter sensor;

7. The process according to claim 1, wherein said output signal of the third correcting circuit (P) is calculated by formula below:

P = D0K + DEO * Q, where: P - output signal of the third correcting circuit; DOK - output signal of the second correcting circuit; DEO - first derivative from output signal of the second correcting circuit DOK; Q - integral, which is proportional to calculated time T; DEO = DOK- F;

F = F + DEO; F - integral from derivative DEO;

Q = Q + A1 * Y*W; A1 - temporary variable, which manipulates integral;

W - coefficient of stability of integral calculation, which changes every tact from zero into one and vice versa;

A1 = DM*KU ^* Q² ; DM - first derivative of output signal of the third correcting circuit (P);

8. The process according to claim 1, wherein said element time constant (T) of parameter sensor (DO) is defined in accordance to time of inquiry of parameter sensor (TO) and is calculated by formula below:

T = T0 * ((Q + K3)- K7 / (Q + K3)), where: T - element time constant of parameter sensoi DO; TO - time of inquiry of parameter sensor; Q - integral; K3 - coefficient, which is equal to 0.5;

K7 - coefficient, which is equal to 0.0833338887.

9. The process according to claim 8, wherein said time of inquiry of parameter sensor (TO) is calculated by formula below: T0= pow (K, n)*K10, where: TO - time of inquiry of parameter sensor; pow (K, n) - number K raised to n power; K10 - coefficient, which is equal to 0.000001.

10. The process according to claim 1 , wherein said value of the minimal boundary of dead zone (Tz1) is calculated by formula below:

Tz1 = TO ^* K14, where: Tz1 - minimal boundary of dead zone; K14 - coefficient, which is equal to 250.

11. The process according to claim 1, wherein said value of the maximal boundary of dead zone (Tz2) is calculated by formula below:

Tz2 = TO * K16, where: Tz2 - maximal boundary of dead zone;

K16 - coefficient, which is equal to 750.

12. The process according to claim 1, wherein said value of the first output signal (D1) of parameter sensor equivalent (DO) is calculated by formula below:

D1 = D1 + (K0 - D1) /tl, where: D1 - first output signal of parameter sensor equivalent (DO) (output signal of the first lag element (LE);

KO - actuating signal of parameter sensor; t1 - time of the first lag element (LE) from the first time to be set (Tϊ); t1 = (T1 / T0) + K3 + K4 * (T0 / T1);

T1 - first time to be set of the first element;

TO - time of inquiry of parameter sensor;

K3 - coefficient, which is equal to 0.5;

K4 - coefficient, which is equal to 0.083333194445.

13. The process according to claim 1, wherein said value of the second output signal (D2) of parameter sensor equivalent (DO) is calculated by formula below:

D2 = D2 + (D1 - D2) / t2, where: D2 - second output signal of parameter sensor equivalent (DO) (output signal of the second lag element); D1 - output signal of the first lag element; t2 - time of the second lag element (LE) from the second time to be set (T2); t2 = (T2 / T0) + K3 + K4 * (T0 / T2);

T2 - second time to be set of the second element;

TO - time of inquiry of parameter sensor; K3 - coefficient, which is equal to 0.5;

K4 - coefficient, which is equal to 0.083333194445.

14. The process according to claim 1, wherein said checking of the counter is performed by output signal of parameter sensor equivalent (DO): D0 = D1 * d1 + D2 * d2, where: D1 - output signal of the first lag element; d1 - coefficient of choice, if it is equal to one, then coefficient of choice d2 is equal to zero and vise versa;

D2 - output signal of the second lag element; d2 - coefficient of choice. ... .