Browsing by Author "Tanaka, Hiroto"

Now showing 1 - 6 of 6

Data Assimilation Applied Thermal Analysis of Mars Airplane for High-altitude Flight Test (MABE2)
(2020 International Conference on Environmental Systems, 2020-07-31) Tanaka, Hiroto; Misaka, Takashi; Nagai, Hiroki
As a new means of Mars exploration method, Mars airplane has been studied for a long time. The Mars airplane enables to bridge the scale and resolution measurement gaps between rovers and orbiters. However, it is not realized yet due to the severe flight conditions. Now, JAXA advances the development of the Mars airplane and plans to make a high altitude flight test called MABE2 (Mars Airplane Balloon Experiment 2). In order to demonstrate a severe condition: atmospheric temperature and pressure on Mars, the flight model will be launched to 35 km in altitude on Earth by High-Altitude Balloon. There, the Mars airplane is released from the balloon, and the flight test is conducted under the Martian-like atmospheric condition. For the high altitude flight test, a thermal mathematical model of the Mars airplane was built on Thermal Desktop©, and thermal analysis was conducted. In the analysis, we introduced a new thermal analysis method using data assimilation to build a reliable thermal mathematical model. The data assimilation is a statistical method to combine physical simulation and observation data obtained from an actual system, and has been used to solve dynamic inverse problems. We employed the Ensemble Kalman Filter, which is one of the data assimilation techniques, to the thermal mathematical model and estimated uncertain parameters, such as thermal conduct conductance. This paper describes the thermal modeling of the Mars airplane for a high-altitude flight test, firstly. Then, the results of the thermal vacuum test and model-test correlation are described. Finally, the result of the thermal conductance estimation and the availability of the new thermal analysis method using data assimilation is discussed.
Development of an advanced thermal mathematical model construction method for spacecraft using artificial neural networks
(50th International Conference on Environmental Systems, 2021-07-12) Tanaka, Hiroto; Nagai, Hiroki; Fujita, Koji
Thermal analysis of spacecraft is one of the most critical processes for the flight model, and the number of missions that need severe thermal requirements is increasing these days. Hence, constructing an accurate thermal mathematical model (TMM) is indispensable to enable the operation to be safe and stable. However, the model has many uncertain parameters, such as thermal contact conductance. In general, these unknown parameters are tuned up by model-test correlation spending much cost with engineer s know-how. This is because the TMM is too complex to tune up by hand. Then, the automatic correlation method for TMM of spacecraft is required. Here, we propose a technique to estimate thermal conductance based on machine learning for the TMM of spacecraft systems. We especially focused on a deep neural network that has inputs of temperature data and outputs of thermal conductance. In this study, a fully-connected feed-forward neural network and a TMM of a spacecraft were installed, and numerical experiments were conducted to evaluate the new method. Then, their network architecture, calculation condition, and estimation results are shown. Also, we discuss the benefits and problems of the model-test correlation based on a deep neural network.
Evaluation of temperature estimation accuracy using Physics-Informed Neural Network for small satellite model
(51st International Conference on Environmental Systems, 2022-07-10) Tanaka, Hiroto; Fujita, Koji; Nagai, Hiroki
Thermal analysis of spacecraft is one of the most critical processes for the flight model, and the number of missions that need severe thermal requirements is increasing these days. However, the thermal mathematical model has many uncertain parameters, such as thermal contact conductance; hence, it is impossible to predict the true value of temperature distribution. On the other hand, the number of temperature sensors on the small satellites is limited, and it is difficult to predict the temperature distribution accurately. In this study, we propose a method to estimate the temperature distribution of the entire spacecraft system based on a small amount of temperature data. To realize the temperature estimation, we use the Physics-Informed Neural Network, which is a neural network that uses the physical conservation law and the observation error as evaluation functions. Specifically, the actual value of the temperature distribution is estimated using the conservation law of the thermal mathematical model, the difference between the operational temperature data and the estimated value, and the boundary conditions as the evaluation function of the neural network. As a result, the temperature distribution of the system can be reproduced from a small amount of temperature data. In this presentation, the temperature estimation accuracy of the proposed method will be shown by numerical experiments using a thermal mathematical model of a pseudo small satellite.
On-orbit demonstration of Advanced Thermal Control Devices using JAXA Rapid Innovative payload demonstration SatellitE-2 (RAISE-2)
(50th International Conference on Environmental Systems, 2021-07-12) Nagai, Hiroki; Tanaka, Hiroto; Kajiyama, Satoshi; Mizutani, Takuji; Nagano, Hosei; Sawada, Kenichiro; Matsumoto, Kan
In recent years, advances in thermal control technology have become essential for deep space exploration to achieve exploration goals. For missions that explore an outer planet, the limited power resources available from solar panels must be used to maintain the temperature of the spacecraft. Therefore, there is an urgent need to develop lightweight thermal control technology that does not use power resources. ?We have conducted research and development of original thermal control devices such as flexible deployable radiators, thermal straps, self-excited oscillating heat pipes, and heat storage devices, and their effectiveness has been confirmed in ground tests, but there has been no opportunity for on-orbit technical demonstrations, and there has been no path to practical application. However, we were selected in 2018 to participate in the Innovative Satellite Technology Demonstration Program proposed by JAXA, and we have the opportunity to conduct on-orbit experiments with the RApid Innovative payload demonstration SatellitE-2 (RAISE-2) in 2021. However, we were selected in 2018 to participate in the Innovative Satellite Technology Demonstration Program proposed by JAXA, and we have the opportunity to conduct on-orbit experiments with the Innovative Satellite Technology Demonstration Satellite 2 in 2021. This program provides opportunities for private companies and universities to acquire and accumulate new knowledge using nano-satellites, to create future mission projects, and to conduct on-orbit demonstrations of key components and new element technologies for space systems.
Thermal Vacuum Testing of Advanced Thermal Control Devices for Flight Demonstration
(51st International Conference on Environmental Systems, 2022-07-10) Kajiyama, Satoshi; Mizutani, Takuji; Ishizaki, Takuya; Tomioka, Kota; Tanaka, Hiroto; Nagai, Hiroki; Matsumoto, Kan; Sawada, Kenichiro; Machida, Yoshihiro; Matsumoto, Kazuaki; Nagano, Hosei
In Japan, several unique thermal control technologies have been developed. However, there are no opportunity to demonstrate in orbit. Therefore, we have proposed to apply our thermal control devices named advanced thermal control devices (ATCD) to the Innovative Satellite Technology Demonstration Program conducted by JAXA, and accepted to apply to the Rapid Innovative payload demonstration SatellitE-2. In this paper, the test results of the thermal vacuum testing of ATCD are presented. ATCD consists of two types of flexible thermal straps: one is made of high-thermal-conductive material, and the other is made of a fluid-loop, and a re-deployable radiator. The conductive-type thermal-strap (CTS) is made of high-thermal-conductive graphite-sheets and aluminum blocks. The fluid-type thermal-strap (FTS) is made of a ultrathin loop-heat-pipe. The re-deployable radiator named reversible-thermal-panel (RTP) is made of high-thermal-conductive graphite-sheets as a flexible fin, and a shape-memory-alloy as a passive re-deployable actuator. As a result, it was confirmed that the thermal conductance between the two ends of CTS was 0.50-0.55 W/K. As for FTS, it was confirmed that it could operate even after recovering from the freezing condition of the working fluid, and that there was no leakage of the working fluid and no performance degradation under vacuum environment. As the heat load increased, the thermal conductance between the evaporator and condenser increased, and finally a thermal conductance value of 4.1 W/K (at 5 W heat load) was confirmed. For RTP, it was confirmed that the radiator fins were fully expanded to 130 when the SMA actuator reached 30 ? during heating. On the other hand, during cooling, the temperature of the SMA actuator dropped only to -15?, and the fins retracted only to 40 . Furthermore, the temperature hysteresis of the SMA actuator was estimated to be about 40? based on the experimental results.
Transient Surrogate Model using Recurrent Neural Networks for Spacecraft Thermal Analysis
(2024 International Conference on Environmental Systems, 2024-07-21) Yamashita, Daichi; Tanaka, Hiroto; Ikami, Tsubasa; Nagai, Hiroki
Thermal analysis is critical to spacecraft mission success. However, uncertain parameters such as contact conductance and effective emissivity make it difficult to predict temperatures accurately. Therefore, uncertainty quantification (UQ) is important in the thermal analysis of spacecraft. In recent years, Monte Carlo simulation (MCS) has been used as one of the most useful methods for UQ. On the other hand, Monte Carlo simulations are computationally expensive because they require a large number of physical simulations. To significantly reduce computational cost, a surrogate model with data-driven has been proposed. However, many have limited temperature prediction points and do not predict system-wide temperatures. If system-wide temperatures could be predicted at low cost, a different approach to thermal analysis could be taken. This study proposes a transient surrogate model using proper orthogonal decomposition (POD) and Long Short-Term Memory (LSTM), a type of recurrent neural network that can predict the entire system and sequence at a low cost. The spatial mode U is extracted from the prior data set by POD, and the coefficients ? of the spatial mode U are predicted each time by LSTM. The use of POD allows for low-cost system-wide prediction, while the use of LSTM allows for prediction against time-varying uncertainty. In this study, the method was applied to the satellite thermal model. Numerical experiments were conducted on a satellite thermal model to compare the accuracy and analysis time of surrogate and thermal mathematical models. The numerical experiments show that the proposed method can significantly reduce the analysis time without losing accuracy.