Method for identifying the model of a thermal enclosure and application to regulation of thermal enclosure

Claim 2

Method for identifying the model of an enclosure comprising the following steps: applying a power rectangular wave front to the enclosure, measuring and recording the temperature response of the enclosure to the said power rectangular wave front and computing from the said temperature response, the response time of the enclosure, generating an identification signal formed of a multiplicity of square waves having a randomly variable height, said square waves further having a randomly variable width comprised between lower and upper limits which are related to the said response time of the enclosure, applying the said identification signal to the enclosure, and measuring by sampling to obtain a plurality of successive samples each having a corresponding rank number, calculating and storing the pulsed response of the enclosure to each of the said square waves, calculating the model representing the enclosure from the stored pulsed responses.

Claim 3

Method according to claim 1 in which the calculation of the model representing the enclosure from the stored pulsed responses comprises the following steps: sampling at constant time intervals the response to the identification signal, thus obtaining a plurality of successive samples each having a corresponding rank number, and calculating from said samples a first multiplicity of first values representative of the pulsed response of the enclosure as a function of time, storing said first values in a first series of memories at addresses corresponding to the rank number of the corresponding sample, measuring a second multiplicity of values representative of the variation in the gain of pulsed responses as a function of the operating temperature and storing in a second series of memories said second values at addresses corresponding to the different operating temperatures, storing in a third series of memories a third multiplicity of values representative of the variation of the sampling period as a function of the different operating temperatures, at addresses corresponding to the temperatures.

Claim 4

Method according to claim 1 which further comprises a validation and control stage of the model including the following steps: applying action signals to the enclosures, calculating by means of the model the response to the signals of the enclosure, and comparing the calculated response with the measured response.

Claim 5

Method according to claim 1, in which the calculation of the pulsed response of the model is effected according to the following identification algorithm: ##EQU8##

Claim 6

SO, SM being respectively the object and the model values, λ being a constant term comprised between 0 and 2, E n being the input vector formed by the values of the input at the sampling times n-1, n-2, . . . n-N, e n-1 , e n-2 , e n-i , e n-N being the input values at the sampling times n-1, n-2, n-i, n-N, h i M (n) represents the i th component of the mathematical representation of the model at sampling time n.

Claim 7

SO, SM being respectively the object and the model values, λ being a constant term comprised between 0 and 2, E n being the input vector formed by the values of the input at the sampling times n-1, n-2, . . . n-N, e n-1 , e n-2 , e n-i , e n-N being the input values at the sampling times n-1, n-2, n-i, n-N, h i M (n) represents the i th component of the mathematical representation of the model at sampling time n.

Claim 8

Regulation method of a thermal enclosure by means of a regulation system having a first closed loop and including means for measuring the temperature to be regulated and a regulator element for generating from the values of the measured temperature and a reference point, a control signal and heating means acting on the temperature of the enclosure in response to the control signal, said method comprising the following first set of steps: measuring the temperature of the enclosure, calculating the future desired behaviour, as a function of the desired response time of the system functioning in closed loop mode, estimating, by means of a mathematical model the future development of the temperature of the enclosure and adjusting a future control signal, applying to the heating means a first adjusted control signal, measuring the temperature of the enclosure and comparing the measured temperature with the estimated values, applying to the heating means a following control signal in the case of a mismatch or effecting a new calculation, a new estimation and a new adjustment before applying a new first control signal to the heating means; said mathematical model being obtained by means of an identification procedure which comprises the following stages: applying a power rectangular wave front to the enclosure, measuring and recording the temperature response of the enclosure to the said power rectangular wave front and computing from the said temperature response, the response time of the enclosure, generating an identification signal formed of a multiplicity of square waves having a randomly variable height, said square waves further having a randomly variable width comprised between lower and upper limits which are related to the said response time of the enclosure, applying the said identification signal to the enclosure and measuring by sampling, thus obtaining a plurality of successive samples each having corresponding rank number, calculating and storing the pulsed response of the enclosure to each of the said square waves, calculating the model representing the enclosure from the stored pulsed responses.

Claim 9

Regulation method of a thermal enclosure by means of a regulation system having a first closed loop and including means for measuring the temperature to be regulated and a regulator element for generating from the values of the measured temperature and a reference point, a control signal and heating means acting on the temperature of the enclosure in response to the control signal, said method comprising the following first set of steps: measuring the temperature of the enclosure, calculating the future desired behaviour, as a function of the desired response time of the system functioning in closed loop mode, estimating, by means of a mathematical model the future development of the temperature of the enclosure and adjusting a future control signal, applying to the heating means a first adjusted control signal, measuring the temperature of the enclosure and comparing the measured temperature with the estimated values, applying to the heating means a following control signal in the case of a mismatch or effecting a new calculation, a new estimation and a new adjustment before applying a new first control signal to the heating means; said mathematical model being obtained by means of an identification procedure which comprises the following stages: applying a power rectangular wave front to the enclosure, measuring and recording the temperature response of the enclosure to the said power rectangular wave front and computing from the said temperature response, the response time of the enclosure, generating an identification signal formed of a multiplicity of square waves having a randomly variable height, said square waves further having a randomly variable width comprised between lower and upper limits which are related to the said response time of the enclosure, applying the said identification signal to the enclosure and measuring by sampling, thus obtaining a plurality of successive samples each having corresponding rank number, calculating and storing the pulsed response of the enclosure to each of the said square waves, calculating the model representing the enclosure from the stored pulsed responses.

Claim 10

Regulation method according to claim 5, in which said control signal is calculated according to the following algorithm: SMR(n+1)=αSO+(1-α)CONS SMR(n+h)=SMR(n+h-1)+(1-α)CONS with α=e(-T/τ r ) and h≧2 T being the sampling period, τ r being the desired response time of the system, SMR(n) being the value of the reference model at the time (n), SO being the value of the measured object, CONS being the value of the reference.

Claim 11

Regulation method according to claim 5, in which said future control signals are adjusted iteratively according to the following algorithm: ##EQU9## with j varying from 1 to h-i+1 i varying from 1 to h λ being a relaxation coefficient ej being the value of the j th future control obtained either at the preceding iteration, or by shifting for the first iteration, ej* taking the place of ej at the following iteration, h J M being a coefficient of the mathematical model, ε(n+h-i+1) being the predicted difference at time n and at (h-i+1) point of coincidence.

Claim 12

Regulation method according to claim 5 further comprising the following second set of steps: measuring a parameter of the heating means, calculating and generating a control signal to be applied to the heating means as a function of the measured parameter and of the control signals calculated in the first set of steps.

Claim 13

Regulation method according to claim 8, in which the calculation of the control W to be applied to the heating means as a function of control determined in the first set of steps, and of the parameter measured of said heating means, is effected according to the following algorithm: W.sub.(n) =W.sub.(n+1) +GP[ε.sub.(n) -ε.sub.(n+1) ]+GI.T.ε.sub.(n-1) with ε.sub.(n) =e.sub.(n) -TRA.sub.(n) W n being the control value to be applied to the heating means, e.sub.(n) being the control value calculated by the first set of steps, GP and GI correspond respectively to gains of a proportional and integral regulator, T is the sampling period, TRA is the value of said measured parameter of the heating means.