Validation of On-Orbit Thermal Deformation and Finite Element Model Prediction in X-Ray Astronomical Satellite Hitomi

Trans. JSASS Aerospace Tech. Japan Vol. 16, No. 3, pp. 242-247, 2018

DOI: 10.2322/tastj.16.242

By Taro KAWANO,1) Kosei ISHIMURA,1) Ryo IIZUKA,1) Kuniyuki OMAGARI,2) and Akira KITO3)

1)Japan Aerospace Exploration Agency, Sagamihara, Japan 2)NEC Corporation, Fuchu, Japan 3)NIPPI Corporation, Yokohama, Japan

(Received June 15th, 2017)

The validation of deformation caused by the thermal expansion of structures is an important issue for the success of a mission that requires the large-scale and precise positioning of instruments. Therefore, an accurate alignment monitor system was installed in the X-ray astronomy satellite Hitomi to correct observation data. We compared the analysis results of a finite element model with the flight telemetry data of the alignment monitor installed on Hitomi to validate the finite element model. As a result, it was shown that the finite element analysis was able to reproduce the global and slow trends of thermal deformation on orbit. However, the analysis could not reproduce the rapid changes in thermal deformation. To realize more precise and large structures in the future, we need to identify these rapid changes and correctly improve the FE model of thermal deformation through ground tests prior to launch.

Key Words: Thermal Deformation, Large Scale Structure, Hitomi

Nomenclature validate the mathematical model with the satellite structure during the development phase. Most satellite projects with a : relative position of the center of corner large-scale structure such as the James Webb Space cube mirror against the laser spot Telescope require abundant resources to validate thermal Ԧ Dܥ : distance deformation prior to launch.1) However, few studies have : displacement of laser spot on corner confirmed the validity of the analysis model compared with cube mirror. actual on-orbit thermal deformation. The reason is that only Ԧ௖௖௠ ݀ : displacement of laser spot CMOS a few satellite systems have a measurement system to : laser spot position on the surface of measure on-orbit deformation due to limited resources. We Ԧ௖௠௢௦ ݀ corner cube mirror obtained on-orbit actual thermal deformation of the X-ray ሬԦ ܮ : laser direction vector astronomy satellite Hitomi on-orbit by using alignment : position vector monitoring system. In this study, we compared actual ݈Ԧ : center position of corner cube mirror on-orbit deformation with the analysis results of the finite ݎሬሬԦప element (FE) model, and then discussed the differences. ܶሬԦ 1. Introduction 2. ASTRO-H (Hitomi) Modern satellites require increasingly large and precise structures for the development of long focal length X-ray 2.1. Overview of the satellite Hitomi telescopes, large-diameter telescopes or high-speed The astronomy satellite ASTRO-H (“Hitomi”) was broadband communication antennas. The major causes of launched from the Tanegashima Space Center in February 2) degraded positioning accuracy and shape stability of 2016. Unfortunately, its operation was terminated in April structures are mechanical assembly errors, deswelling 2016 due to communication trouble. The details of those 3) deformation of CFRP, thermal deformation, micro vibration circumstances have been reported by JAXA. caused by rotating parts, and deformation due to differences In order to observe broadband X-rays simultaneously at a between 0 G and 1 G. When a satellite experiences drastic high-energy resolution emitted from such high-energy fluctuations in thermal conditions such as a full shadow and astronomical objects as a black hole, four kinds of sunlight, thermal deformation becomes the most serious observation systems consisting of two soft gamma-ray problem. detectors, two hard X-ray imaging systems, a soft X-ray To deal with this problem, we use thermal and structural imaging system, and a soft X-ray spectrometer system were analysis model of the satellite structure in structural design. installed facing the same direction. The mathematical model is also used to predict on-orbit Since X-rays have high transparency and are difficult to thermal deformation. Therefore, it is very important to bend, the main structure of Hitomi was extremely long at 14 m when the Extensible Optical Bench (EOB) was extended

242 Trans. JSASS Aerospace Tech. Japan Vol. 16, No. 3 (2018)

on-orbit. Maintaining the co-alignment of each observation was provided by CSA, Neptec, and St. Mary’s University. instrument was important for achieving the simultaneous CAMS consists of a Laser Detector (LD) unit and a Corner observation of broadband X-rays. Therefore, the highly Cube Mirror (CCM). Two LD units were mounted on the accurate prediction of thermal deformation prior to launch top plate of the FOB near the Hard X-ray Telescope (HXT), was needed to ensure sufficient shape stability on-orbit. In addition, two CCMs were installed on the HXI plate at the end of the EOB. Figure 2 shows the mechanism of CAMS measurement. For the alignment measurement, a laser was emitted from a LD unit to the end of the EOB, reflected by the CCM mounted on the HXI plate back to the LD unit, and finally enters a CMOS element on a detective plane in the LD unit. CCM displacement relative to the LD unit can be calculated by using (the vector of movement of the �� beam spot on a CMOS surface) in Eq. (1) below.� �� . (1)

�� The laser axes� of �� both CAMS/2 units were mounted mirror-symmetrically about the Y-Z plane (Fig. 2). Based on the principle of measurement, CCM rotation does not affect the measurement results. Conversely, rotation of the laser direction vector of the LD unit affects the measurement Fig. 1. Transparent view of X-ray astronomy satellite Hitomi.4) results. Any LD unit rotation or deviation in laser direction Figure 1 shows a transparent view of the X-ray astronomy against the optical axis of the telescope will cause erroneous satellite Hitomi. The structure of Hitomi consists of the base measurement in telescope alignment. The major cause of plate, Fixed Optical Bench (FOB), and Extensible Optical such rotation was expected to be thermal deformation of the Bench (EOB). All telescopes were mounted on the FOB. LD unit itself. However, LD unit temperature remained The HXI plate was attached to the bottom end of the EOB. stable within 2 °C during the entire on-orbit operation. Two Hard X-ray Imagers (HXI) were installed on the HXI According to the thermal stability test of the LD unit prior to plate. Bus equipment was mounted on the side panels launch, the influence of fluctuations in temperature within 5) attached to the periphery of the base plate. The central area 2 °C was negligible even with a 12 m working distance. enclosed by eight side panels was left vacant and used for CAMS units were turned on before the EOB was optical paths of X-rays collected by telescopes. The main extended during on-orbit operation. After that, CAMS units requirement of regarding alignment in the Hitomi mission measured alignment continuously and obtained was the precise alignment of the mutual optical axes of all telescopes, and the mechanical axes that were defined by the center points of each detector and telescope. 2.2. Canadian ASTRO-H metrology system The measurement instruments collectively known as the 5) Canadian ASTRO-H Metrology System (CAMS) were installed for correction of the hard X-ray imaging system that was expected to have difficulty in maintaining alignment due to its long focal length. CAMS measures the amount of misalignment between a telescope and a detector to correct the observation data. It

Fig. 2. CAMS measurement mechanism and mounting position. Fig. 3. CAMS flight data in all periods of the Hitomi mission.

243 Trans. JSASS Aerospace Tech. Japan Vol. 16, No. 3 (2018)

measurement data up to the anomaly event just prior to the In the FE model of Hitomi, the LD units of CAMS were end of operation. placed on the top plate of the FOB, and CCMs were installed on Figure 3 shows the flight data of both CAMS units as the HXI plate at the end of the EOB. With the position vectors obtained in all periods of the Hitomi mission. The vertical axis of nodes on the interface plane of a LD unit defined as , ,

shows the CAMS output (indicating CCM movement); the , the laser direction vector is expressed in Eq. (2) below.� � horizontal axis shows the date and time. The results shown are �� in order of CAMS1 X, CAMS1 Y, CAMS2 X and CAMS2 Y � �� from the top. The direction of each axis is as defined in Fig. 2. . (2) Large fluctuations at the start of CAMS measurement were �� 6) � � caused by extending the EOB. After that, the initial distortion |�� | was gradually released over a long period. As depicted in Fig. 3, In addition, the position vector of the laser spot on a CCM the rapid shifts represent an attitude change due to a changed surface is expressed as follows: observation target. It is difficult to identify details at this scale, �� but periodic fluctuations accompanying the orbital round of 96 � , (3) minutes are observed. The breadth of fluctuation of the line �� represents the magnitude of change in one orbital period. The where is the position� ��vector�� of the origin of laser emission

magnitude of deformation during the whole period was within and D is� the length between the LD unit and CCM. CCM 1.5 mm, taking into account the mechanical deformation caused displacement�� relative to the laser spot position on the HXI plate by the initial latch. The deformation caused by orbiting the is defined by the following equation. Earth was less than 500 μm. In the case of NuSTAR, which has a similar focal length, the motion of the optical axis on the , (4) detector was about 3 mm in one orbit period.7) In view of that � �� finding, the structure of Hitomi was extremely stable. where is a position � vector�� of a CCM center. The vector obtained by Eq. (4) corresponds to the flight data of CAMS. To �� 3. Finite Element Model and Analysis show validation� of the FE analysis, the flight data of CAMS is compared with the displacement of (Eq. (4)) obtained by FE In the following section, we discuss the FE model and analysis. � analytical conditions that were used for the prediction of At first, we conducted thermal analysis� by using the same on-orbit thermal deformation. The FE analysis used here was a attitude actually taken on-orbit to prepare the load set of linear static analysis of Nastran. Figure 4 shows an overview of temperature distribution for thermal deformation analysis.10,11) the Hitomi FE model. Although the whole model had about one To obtain highly accurate temperature distribution, we selected million nodes, the structure and components to be included two kinds of periods when the same attitude continued more were selected based on the purpose of analysis. For example, than two days and fluctuations in temperature had been the solar array paddle (SAP) is not included in the analysis of converged. The period of an anomaly event near the end of the thermal deformation because the joint between the satellite operation was also selected due to a very large change in body and the SAP is so flexible that there is no significant temperature that was expected to be suitable for validation of influence on the thermal deformation of optical benches. Prior the analysis model. Then, we conducted thermal deformation to launch, the FE model was modified based on ground tests analysis. For each normal attitude, thermal deformation analysis using a mechanical test model.8,9) was conducted at six timings in one orbital period. The first timing is set to the culmination time. In case of the anomaly, the analysis was conducted at four timings when telemetry data was obtained.

Fig. 4. Overview of the Hitomi FE model.

244 Trans. JSASS Aerospace Tech. Japan Vol. 16, No. 3 (2018)

Fig. 5. The flight data of CAMS x, CAMS y, SAP temperature and FE Fig. 6. The flight data of CAMS x, CAMS y, SAP temperature and analysis results with equal boundary condition. FE analysis results at different attitude from Fig. 5.

4. Validation of FE Analysis

Figures 5 and 6 show the relative displacement between the top plate and the end of the EOB as measured by the CAMS units at two attitudes. The solid line denotes the two orbital periods of flight data superimposed 10 times. The horizontal axis represents time; the vertical axis shows displacement in the x and y directions. The coordinate systems of both CAMS1 and CAMS2 flight data were converted to the satellite coordinates. Fig. 7. General characteristics of the CAMS flight data. In normal operation, the satellite coordinate +Y was directed toward the Sun, and the satellite coordinate +Z was directed toward the telescope targeting direction. CAMS1 is installed on the +X side of the satellite; CAMS2 is installed on the -X side. CAMS1 and CAMS2 are symmetrically placed with respect to the Y-Z plane. To represent the thermal conditions of the

245 Trans. JSASS Aerospace Tech. Japan Vol. 16, No. 3 (2018) satellite at each moment, the bottom figure shows the SAP of the primary structure having high thermal capacity and temperature with the same phase. In Figs. 5 and 6, the FE consisting of the FOB, EOB, thrust tube, and rocket coupling analysis results are shown as circled dotted lines. Because the ring. However, FE analyses could not reproduce the rapid 200 displacement of CCM should be evaluated instead of its μm rise at the timing of entering the penumbra in 2,000 seconds. absolute position in discussing the validation of thermal These phenomena accompany a fast convergence. One deformation analysis, the result of FE analysis at the starting suspected cause must be the structural members having low point was shifted to the averaged value of the CAMS flight data. thermal capacity such as the face sheets of honeycomb panels, Figure 7 shows the general characteristics of the CAMS flight thin plates or bolts. Further discussion of the reason would data discussed below. We can observe three trends in the figure: require a more detailed and comprehensive analysis. rapid, slow and short periodic changes. According to the solid In the Y direction, the flight data of CAMS2 is in good lines in Figs. 5 and 6, a large peak appeared at the timing of agreement with FE analysis in terms of global tendency. entering the penumbra in every orbital period. One orbit period However, the flight data of CAMS1 has a large difference from of Hitomi was 96 minutes (5760 seconds). The part of the the result of FE analysis. In addition, the notable point of the Y changes over time shorter than the orbital period is clearly direction is that CAMS1 and CAMS2 have different tendencies. synchronized with the timing of heater switching. There are two suspected causes of this difference. The first one The slow changes in the X direction of CAMS flight data are is that the HXI plate was rotated around a point with an offset, in good agreement with the FE analysis results as shown in Fig. instead of the centroid. The second one is that the laser 5. The slow changes are considered to come from deformation directions of a LD unit were shifted asymmetrically by each other due to asymmetric thermal conditions. We will investigate

which suspected cause is more likely by using additional data,

such as the observation data of scientific instruments.

3000 Figure 6 shows the flight data of CAMS and FE analysis

results for Hitomi at another attitude. Also in this case, the slow

2500 changes in the X direction are in good agreement with FE

analysis and the rapid change of roughly 100 μm at the timing

2000 of entering the penumbra is not reproduced. As well as in Fig.

03/25 18:00 03/25 21:00 03/26 00:00 5, FE analysis is in good agreement with the flight data in the Y

direction of CAMS2.

Figure 8 shows the flight data and FE analysis results at the

last four timings during the end of Hitomi operation. The 1000 horizontal axis in Fig. 8 represents the date and time; the

vertical axis shows displacement. Hitomi had rotated at 21.7 500 degrees/hour about the Z axis just before communication was

lost due to an error in the attitude control system.3) Such a 0 drastic change in thermal conditions has never occurred in

03/25 18:00 03/25 21:00 03/26 00:00 normal operation. This data is extremely effective as a

validation of FE analysis because we were able to estimate the

bias component of thermal deformation caused by the change in 2500 the satellite’s thermal potential. Unfortunately, continuous

flight data has not been obtained due to trouble. In Fig. 8, the 2000 downloaded flight data of CAMS is plotted with a blue solid

line, and the period without data is indicated with a blue dotted 1500 line. Orange asterisks denote the FE analysis results. The

orange dotted lines show the period without data. It should be 1000 03/25 18:00 03/25 21:00 03/26 00:00 noted that the temperature used here was a transient one and the

estimated temperature was not as accurate compared with the

periodic results in Figs. 5 and 6. To limit discussion to the

500 deformation caused in this period, the starting point of analysis

was shifted to the point of the flight data. The flight data of both

0 CAMS1 and CAMS2 in the X direction are in good agreement

with the FE analysis results. In the Y direction, the flight data is

-500 in agreement with the analysis data globally. Therefore, it is

confirmed that the FE analysis of Hitomi was able to reproduce

03/25 18:00 03/25 21:00 03/26 00:00 thermal deformation including the bias component globally. Day/time (-) However, the FE analysis results showed excessive displacement in the Y direction between 3/25 18:12 and 3/25 Fig. 8. Comparison of the flight data and FE analysis results in 21:02. Furthermore, it is interesting that there are few the period of attitude anomaly. asymmetric displacement components of the Y direction as seen

in the flight data of CAMS in typical orbital periods at this

246 Trans. JSASS Aerospace Tech. Japan Vol. 16, No. 3 (2018) time. Structure, AIAA Paper 2011-2161, 2011. 2) Takahashi, T. et. al.: The ASTRO-H (Hitomi) X-ray Astronomy Satellite, Proc. SPIE, 9905 (2016), pp.99050U-1-99050U-16. 5. Conclusion 3) JAXA, Investigation of Anomalies Affecting the X-ray Astronomy Satellite “Hitomi” (ASTRO-H), In this paper, we discussed the actual on-orbit thermal http://global.jaxa.jp/projects/sat/astro_h/files/topics_20160608.pdf (accessed June 08, 2016). deformation of the X-ray astronomy satellite Hitomi and the 4) ASTRO-H Press kit, validation of the FE model. According to CAMS, the http://fanfun.jaxa.jp/countdown/astro_h/files/astro_h_presskit.pdf maximum magnitude of deformation during the whole (in Japanese)(accessed September 12, 2017). 5) Gallo, L., Lambert, C., Koujelev, A., Gagnon, S., and Guibert, M.: period was approximately 1.5 mm. During the period, the The Canadian Astro-H Metrology System, Proc. SPIE, 9144 optical bench of Hitomi was more stable as compared to (2014), pp. 914456-1-914456-10. NuSTAR with a similar focal length. Throughout this 6) Ishimura, K., Ishida, M., Kawano, T., Minesugi, K., Abe, K., Sasaki, T., Iizuka, R., and Bando, N.: Induced Vibration of validation, the FE model was confirmed as being High-Precision Extensible Optical Bench during Extension on sufficiently accurate to predict the behavior of on-orbit Orbit, 31st International Symposium on Space Technology and thermal deformation. However, several rapid displacements Science, Ehime, Japan, 2017-c-30, 2017, (in press). 7) NASA Goddard Spaceflight Center: NuSTAR Observatory Guide, originating from structural members having low thermal https://heasarc.gsfc.nasa.gov/docs/nustar/nustar_obsguide.pdf, capacity in the flight data were not reproduced. To realize (accessed April 11, 2017). more precise and large structures in the future, we need to 8) Kawano, T., Ishimura, K., Minesugi, K., Omagari, K., and Tanaka, identify these rapid changes and correctly improve the FE K.: The evaluation technique of on-orbit thermal deformation for large precise structure in ASTRO-H, AIAA Paper 2015-0203, model of thermal deformation through ground tests prior to 2015. launch. 9) Ishimura, K., Minesugi, K., Kawano, T., Wada, A., Shoji, K., Ikeda, M., and Omagari K.: Novel Technique for Spacecraft's Thermal Deformation Test Based on Transient Phenomena, Trans. Acknowledgments JSASS Aerospace Tech. Japan, 12, ists29 (2014), pp. Pc_29-Pc_34. 10) Iwata, N., Usui, T., Miki, A., Ikeda, M., Yumoto, T., Ono, Y., Abe, The authors would like to acknowledge the contributions K., Ogawa, H., and Takahashi, T.: Thermal Control System of X-ray Astronomy Satellite ASTRO-H: Current Development Status of the entire ASTRO-H (Hitomi) project team, particularly and Prospects, 44th International Conference on Environmental the team members from Canada. Systems, Tucson, Arizona, ICES-2014-285, 2014. 11) Iwata, N., Usui, T., Ikeda, M., Takei, Y., Okamoto, A., Ogawa, H., Yumoto, T., Ono, Y., Kokubun, M., and Takahashi, T.: Evaluation References of In-Orbit Thermal Preformance of X-ray Astronomy Satellite "Hitomi", J. Spacecraft Rockets, 2017-03-A33903.R1 (in press). 1) Johnston, D. J. and Cofie, E.: An Overview of Thermal Distortion Modeling, Analysis, and Model Validation for the JWST ISIM

247