JPH0516542B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention controls the thermal behavior of a satellite by feeding back heat input to the satellite and measured temperature data in a space chamber and controlling the heat input to the satellite. The present invention relates to a satellite thermal testing device that enables higher accuracy of the mathematical model (satellite thermal model) that represents the satellite, as well as shorter thermal testing periods and lower costs.

[Background of the Invention] FIG. 1 is a schematic diagram of a conventional thermal testing apparatus. It is difficult to obtain the data necessary to create a thermal model with sufficient accuracy, and it is therefore impossible to reproduce the thermal behavior of a satellite in space using computer simulation. , the heat input device 1 simulates each characteristic thermal environment in the orbit and transmits heat input information 11 to the satellite in each state.
The heat input 21 converted into a physical quantity in 01 is transferred to the satellite 102.
The temperature distribution 3 of the satellite obtained by giving .

The temperature distribution of the satellite in orbit was predicted by applying corrections to the measured temperature distribution data based on the differences in the thermal environment between the space chamber and the orbit, so space chamber tests under many thermal environment conditions were possible. I had to do it. this is,
This is one of the major reasons why Japan is only able to launch about two satellites a year.

[Purpose of the invention]

SUMMARY OF THE INVENTION An object of the present invention is to provide a thermal testing device for achieving high accuracy of a thermal model of a satellite, with the aim of shortening the period of expensive thermal testing using a space chamber.

[Summary of the invention]

In order to achieve the above object, the present invention solves the problem of how to obtain enough heat input data and temperature measurement data to a satellite in a short time to meaningfully determine a thermal model. (the ability to uniquely define a thermal model) is sequentially verified.
The feature is that the heat input is controlled so that the remaining uncertain model parts can be determined.

According to this method, only enough thermal tests in a thermal environment are performed to determine the thermal model, so it is possible to estimate a highly reliable thermal model with a minimum amount of data.

[Embodiments of the invention]

Embodiments of the present invention will be described below with reference to FIG.

The thermal model of the satellite is given by the thermal balance equation at each node (area of the satellite that can be considered isothermal) as follows.

C _i dT _i /dt=Q _i + _o 〓 ^j=i K _ij (T _j −T _i )+σ _o+1 〓 ^j=i R _ij (T ⁴ _j −T ⁴ _i ) i=1,...,n ...(1) where, n: Number of nodes on the satellite C _i : Heat capacity of the i-node T _i : Absolute temperature of the i-node Q _i : Sum of internal calorific value of the i-node and input heat amount to the i-node from sources other than the satellite K _ij : Heat conduction coefficient between nodes i and j R _ij : Radiation heat exchange coefficient between nodes i and j t : Time σ : Stefan-Boltzmann constant T _o+1 : Temperature of space chamber wall (temperature of space chamber wall (sometimes it's absolute zero).

In Fig. 2, the heat input 22 to the satellite 102 is based on the results optimally calculated for thermal model identification, and is different from that in Fig. 1, which simulates the heat input to the satellite in the thermal environment in orbit. It is different from the heat input 21.

The coefficients of this model (heat capacity C _i , heat conduction coefficient K _ij ,
Since the accuracy of the model that uses estimates from a priori information (basic data on the satellite's shape, physical laws, and physical properties) as the radiant heat exchange coefficient (R _ij ) is not good, the heat input data to the satellite12 and the temperature measurement data are 4, the coefficients must be identified. If the coefficient vector to be identified is represented by p, then from equation (1), g = A × p ... (2) where g: vector A consisting of heat input data (Q _i , i = 1, ..., n) : A matrix consisting of the measured temperature time series data of all nodes and σ. Therefore, whether p can be determined significantly can be determined by examining the regularity of the data matrix A ^T A (A ^T is the transposed matrix of the A matrix). When A ^T A is regular, it is estimated that p = (A ^T A) ^-1 A ^T g.

However, since the temperature of each node depends on the coefficients to be identified, the heat input cannot be determined in advance so that the data matrix ^AT A is regular. Therefore, the regularity of the sequential data matrix is determined, and the heat input is controlled so that the part of the coefficient that cannot be determined from the data up to that time can be determined. This heat input calculation procedure is shown in FIG. 3 and will be described in detail below.

Using the heat input/temperature measurement data 31 to the satellite, the eigenvalues of the data matrix are calculated in step 32, and if there are s (>0) zero eigenvalues, the regularity of the data matrix is determined in step 33. It is determined that the eigenvalue is non-regular, and in the next step 36, the normalized eigenvector corresponding to the zero eigenvalue is calculated as follows. The normalized eigenvectors corresponding to the non-zero eigenvalues of the data matrix ^ATA at the present time are expressed as V ₁ , . . . , V _os . In the vector space consisting of the coefficients to be identified, {w _i , i=1, ..., s|w _i ⊥w _j , i≠j; ‖w _i ‖=1
;w _i ⊥v _l , l=1,..., n-s}...(3) The data matrix does not have direction information. The symbols ⊥ and ‖‖ in the above formula represent orthogonality and norm, respectively. w _i , (i=1,...,s) in the above equation is a matrix It is obtained by orthonormalizing s linearly independent column vectors of . This w _i (i=1, . . . , s) is the normalized eigenvector corresponding to the zero eigenvalue of the data matrix to be determined.

In the next step 37, the following calculations are performed. The data matrix obtained between the present and a unit time in the future is represented by A ^* . Data matrix after unit time (A ^T
In order for A+A ^*T A ^* ) to approach regularity, A ^* w _i ,
(i=1, . . . , s) needs to take a sufficiently large value. However, A ^* cannot be determined because it depends on the coefficients to be identified. Therefore, the estimated value A ^** of A ^* is substituted, and the heat input per unit time is determined so that A ^** w _i (i=1, . . . , s) takes a sufficiently large value. A ^**
is obtained by integrating equation (1) using the estimated value p^ of the coefficient of the thermal model at the current time instead of the coefficient in equation (1). The estimated value p^ of the coefficient at the current time is calculated as follows. Pseudo-inverse matrix of data matrix (A ^T A) ⁺
The roughness estimate of the coefficient p = (A ^T A) ⁺ A ^T g (5) is calculated as follows. p^ is from p〓, () If the elements of p〓 are significantly different from the estimated value of the coefficient accumulated from a priori information, the estimated value based on a priori information is adopted, () After (), a negative value For elements that take , zero is used as the value of the element.

Heat input Q _i up to unit time is increased so that the evaluation function, _s 〓 ^i=l w ^T _i (A ^** ) ^T A ^** w _i ……(6)
All you have to do is find (i=1,...,n), but
Making Q _i time-dependent requires a huge amount of calculation and is practically impossible, so the following is used instead.

Q _i (i=1,...,n) is a time constant for unit time, and the maximum physically possible heat input from zero
Q _i , limited to values up to _nax . Under this constraint, by numerical simulation, Q _i (i=1,...,n) that takes the extreme value of equation (6) is SUMT
(Sequential Unconstrained Minimization
Technique, Reference..."Nonlinear Optimization Calculation Method", written by LCW Dickson). However, during the calculation process using SUMT, if the estimated value of each node exceeds the permissible temperature range of that node (the temperature range in which the satellite equipment represented by that node operates reliably and normally), -∞ give a penalty of The above is the calculation in step 37, where the calculated heat input 22 is given to the satellite and temperature measurement data per unit time is obtained.

The above steps are repeated until the data matrix is determined to be regular. When the data matrix is determined to be regular in step 33, the coefficients of the satellite thermal model are determined as p=(A ^T A) ^-1 in the next step 34. A ^T g is calculated, the satellite thermal model identification is completed, and the thermal test is also completed.

〔Effect of the invention〕

According to the present invention, by introducing a device that calculates coefficients of a satellite thermal model that cannot be determined from past temperature measurements and heat input data, it is possible to control subsequent heat input to the satellite so that the coefficients can be determined. I can do it. Further, according to the present invention, it is possible to reduce the acquisition of meaningless data and to acquire enough data to identify coefficients. Therefore, the effect of shortening the thermal test period using the space chamber, thereby reducing costs, and increasing the accuracy of the mathematical model can be obtained.

[Brief explanation of the drawing]

Figure 1 is an explanatory diagram of the conventional satellite thermal test equipment, Figure 2
The figure is an explanatory diagram of the thermal test device of the present invention, and FIG. 3 is an explanatory diagram showing the calculation procedure in the heat input calculation and satellite thermal model identification device of the present invention. 104... Heat input calculation and satellite thermal model identification device.

Claims (1)

[Claims] 1. Determine the identifiability of the measured temperature data in a thermal test device for identifying a satellite thermal model using a space chamber consisting of a means for inputting heat to the satellite and a means for measuring the temperature of each satellite component. An artificial satellite characterized by comprising: a means for extracting an unidentified portion of a satellite thermal model based on the determined result; and a means for determining a heat input necessary to determine the extracted unidentified portion. thermal testing equipment.