- Email: [email protected]
Managing the microvibration impact on satellite performances
Acta Astronautica 162 (2019) 461–468
Contents lists available at ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
Managing the microvibration impact on satellite performances a,∗
a
a
Frank Steier , Torben Runte , Anneke Monsky , Timo Klock a b
a,1
b
, Gregory Laduree
T
OHB System AG Bremen, Universitätsallee 27-29, D-28359 Bremen, Germany ESA/ESTEC, Keplerlaan 1, 2200 Noordwijk, the Netherlands
ARTICLE INFO
ABSTRACT
Keywords: Microvibration Control Isolator Disturbance Superposition Overlap MicroMap
Due to the increasing needs on the performance of satellites, in particular optical satellites, microvibrations have become more and more important over the last years. Microvibrations affect a large number of satellites for both Earth observation and space observation missions. Further, microvibrations are also a concern for other missions with sensitive instruments like Fourier spectrometers, high-precision accelerometers or reference oscillators. In classical fields like structural dynamic engineering, management approaches have been established and implemented in all space projects. This paper presents a corresponding approach for the area of microvibrations. Focus is on the definition of all key elements needed to establish a microvibration control approach. This includes the establishment of microvibration budgets and associated summation rules. A systematic way for microvibration disturbance minimization is presented and rules for the application are defined. Specific emphasis is given to the definition of microvibration interface requirements and the associated control process. The concepts provided in this paper have been verified by tests performed in the frame of various satellite programs and studies. Tests related to the superposition of different noise sources are presented in here.
1. Introduction In literature many publications on microvibrations can be found. These are mostly related to specific aspects, while only a few address system level aspects and the underlying engineering process. Certain insight on the microvibration engineering process is provided for Bepicolomobo [1], GOCE [2], Solar-B [3], an optical satellite system [4] and a satellite platform [5]. Others system level publications are related to simulation [6] and testing [7,8], and to the definition of microvibration interface requirements [9]. In addition, numerous publications can be found that focus on specific aspects of the microvibration engineering management process. These are related to the modelling and testing of microvibration sources [10–12], structural modelling [13] and design considerations for appendages [14,15]. Various recent publications in the area of microvibrations are related to isolations systems [16–21]. This paper provides a general microvibration engineering process which is not linked to a specific satellite. It complements the existing approaches and provides engineering processes and guidelines for the important elements of the engineering process. In addition to existing articles, it provides generalized budgeting rules and an interface requirement definition approach aiming for specifications with
minimized margins. Section 2 deals with the microvibration performance budget and provides guidelines on how to consider noise sources and the associated summation rules. Section 3 provides a systematic approach on the disturbance minimization process to be applied during the satellite design phase. Section 4 defines a methodology to establish microvibration interface requirements towards susceptible payloads. Section 5 provides the overall control logic that is applied during the satellite development. Emphasis is given to requirement definition and verification at different levels. Section 6 provides test results that provide information on the superposition of noise sources and the statistical behaviour of the disturbances. 2. Microvibration performance budget and noise sources In many satellites, microvibrations have to be considered as one contributor to the payload key performance parameters. Most commonly, this performance is linked to the pointing performance. However, there are other sensitive payloads such as high precision accelerometers, spectrometers or atomic reference oscillators, which could be impaired by microvibrations. One major activity of early satellite studies is the establishment of
Corresponding author. E-mail address: [email protected] (F. Steier). 1 Current address: Simula Research Laboratory, Martin Linges vei 25, 1364 Norway. ∗
https://doi.org/10.1016/j.actaastro.2019.06.027 Received 28 January 2019; Received in revised form 26 April 2019; Accepted 20 June 2019 Available online 22 June 2019 0094-5765/ © 2019 IAA. Published by Elsevier Ltd. All rights reserved.
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
the satellite performance budget, which is needed to define the overall satellite architecture. Here we explain, how the budget can be established when microvibrations have to be considered. Budgeting rules are defined for satellite pointing, however the method can be applied correspondingly for other performance parameters as well. There are guidelines provided by ECSS2 [22–24] on how to evaluate the pointing performance of a satellite. Microvibrations typically affect the error budgets related to the following performance parameters:
• APE (Absolute Pointing Error) • RPE (Relative Pointing Error) and • PDE (Pointing Drift Error). These errors are illustrated in Fig. 1, where ε is the deviation from the nominal satellite pointing and Δt the observation period. The RPE is the maximum deviation from the Mean Point Error during the time Δt. The Figure also shows the Mean Pointing Error (MPE) for which the microvibration contribution is usually negligible. Before establishing the budget, a consideration for all sources of microvibrations aboard the satellite has to be made in terms of their statistical behaviour in order to apply the appropriate summation rules. Typical sources on satellites are either periodic or transient sources. Random sources are usually less important and are therefore not further considered in the following discussion. Depending on the way the sources are operated during satellite operation, their statistical behaviour can be determined. For transient sources the statistical behaviour can be: - Overlapping, - Non-overlapping, or - Randomly appearing;
Fig. 1. Pointing errors affected by microvibrations.
while periodic sources can either be
generic budget for microvibrations is shown for short term disturbances in Table 1. The expectation value of the satellite pointing, < E > , is the sum of the individual contributors defined as maximum error emax. The budget for intermediate term disturbances is shown in Table 2. It uses modified summation rules that account for the fact that the disturbance does not reach its full amplitude over the observation time. In that case the summation has to consider the maximum amplitude over the disturbance period, e∼ max, for transient sources; and a weighing function for periodic sources. The weighting function is defined by w(ω,Δt) = sin(ωΔt/2) for ω ≤ π/Δt, and w(ω,Δt) = 1 for ω > π/Δt.
- Correlated, or - Uncorrelated. It shall be noted that a source can show more than one statistical behaviour at the same time – e.g. Reaction Wheels generate disturbances by imbalance effects which are harmonic; and friction jumps which are random effects. These effects have to be treated as individual errors in the performance budget. For periodic errors such as micro vibrations, the impact of the disturbance on the spacecraft performance depends on the relation of the microvibration oscillation period (disturbance period) with respect to the observation period. If the disturbance period is short compared to the observation period, the average deviation from the Mean Pointing Error is zero. When establishing a microvibration budget, two cases have to be distinguished:
3. Disturbance Minimization There are different measures to minimize the impact of microvibrations on the spacecraft performance. Minimization measures are implemented starting in phase A of the project, but have to be followed iteratively throughout the design phases. Possible minimization measures are:
- Short disturbance period: Disturbance period is short compared to the observation period. - Intermediate disturbance period: Disturbance period is long or of the order of the observation period.
1. Mission concept:
ECSS [23] define the terms short disturbance period and long disturbance period, while the long disturbance period means that variations are much longer than the observation period. For microvibration the new term “intermediate disturbance period” is introduced. This is needed since for many typical spacecraft, the disturbance period is of the same order as the observation period which requires the definition of a weighting function as defined at the end of this paragraph. The budgets are established differently for these two cases. The 2
Based on the criticality of the satellite to microvibrations, an initial concept has to be established. One could choose e.g. standard attitude control concepts based on Reaction Wheels or lay out more comprehensive control concepts based on microNewton thrusters. 2. Frequency Control: In order to be able to control the satellite performance throughout the development cycle, it is mandatory to keep control over the frequencies relevant for microvibrations. These are the excitation frequencies of sources, susceptible frequencies of receivers and structural
ECSS: European Cooperation for Space Standardization. 462
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Table 1 Pointing Budget rules in case of short disturbance period. Source Type Statistical behaviour
Summation
Examples
Transient
Linear Worst case Linear Linear RSS
Σ emax emax Σ emax Σ emax σ assuming Gaussian distribution
Mechanisms operated at the same time Calibration wheel, scanner start-up Reaction Wheel friction jumps, sudden stress release Solar Array Drive stepper motors, resonances within one Reaction Wheel Disturbances within one stepper motor, disturbances from different Reaction Wheels
Periodic
Overlapping Non-overlapping Randomly appearing Correlated Uncorrelated
Table 2 Pointing Budget rules in case of intermediate disturbance period. Source Type Statistical behaviour Transient Periodic
Overlapping Non-overlapping Randomly appearing Correlated Uncorrelated
Summation
Examples
Linear Worst case Linear Linear Linear
Σ e∼ max e∼ max Σ e∼ max
Mechanisms operated at the same time Calibration wheel, scanner start-up Reaction Wheel friction jumps, sudden stress release Solar Array Drive stepper motors, resonances within one Reaction Wheel Disturbances within one stepper motor, Reaction Wheel imbalances, disturbances from different Reaction Wheels
Σ w(f)emax Σ w(f)emax
resonances. The resonance frequencies can be suppressed by design measures of these elements, but it is normally more important to separate them, in order to avoid amplifications.
and for the structural design, since it affects the harmonic disturbances and resonance frequencies. The above measures for minimization can be applied in a sequential process during the design phase as shown in Fig. 2. The process starts in Phase A, in which a dedicated microvibration requirement at system level is derived based on the mission requirements. Depending on the stringency of these requirements, the relevant sources and potentially sensitive devices are identified. Based on this rudimental information, a preliminary assessment of the achievable performance can be carried out, in order to indicate the measures to be implemented in Phase B. In Phase B, a detailed analysis on all possible measures can be performed taking into account more mature information of the microvibration critical elements. The flow shown in Fig. 2 aims for minimum iterations in the design process. However, some of the control measures could also be applied in parallel. The key element is the Frequency Control which should be analysed first, since it impacts the other control measures. After an initial iteration, a first detailed analysis can be performed which is indicated in Fig. 2 as “Microvibration Satellite Level Nominal Analysis”. This analysis provides the relevant data to decide whether or not to implement an isolation system and to define the location and cutoff frequencies of the isolator(s), which usually cannot be done a priori. Depending on the remaining margins, this process may be applied iteratively until a robust satellite design is found. For missions that are sensitive to microvibrations, a sensitivity analysis is performed after the minimization loop is finished. The objective of the sensitivity analysis is to define available margins and acceptable frequency ranges of critical harmonics that have to be known during the unit level development and implementation activities in later project stages. Similar to other engineering areas such as structure or EMC engineering, critcal frequencies, allowed ranges and amplitudfes of resonances are summarized in the Microvibration Frequency Control Plan, which is intended to support tracking of evolutions throughout the project life cycle.
3. Unit Level Coordination: At equipment level, there is some tuning capability of the disturbances. E.g. the frequency of a cryogenic cooler can be chosen for minimum impact or mechanisms can be actuated with specific actuation profiles that minimize the microvibration disturbances in the critical frequencies. In case of off-the-shelf components, such tuning capability is normally not given. 4. Accommodation & Structural Design: The satellite accommodation can be chosen such that emitters of microvibrations are located far away from the sensitive receivers. Further, the microvibration load path can be optimized for the transmission of critical frequencies. In order to design the load path effectively, the coupling of the source with the structure has to be modelled correctly and optimized for separation of the critical frequencies. Managing the coupling effect is typically most challenging, since its correct modelling requires correlation with test data. 5. Operational Concepts: In many cases it is possible to schedule the activation of mechanisms aboard a satellite such to minimize the impact of microvibrations. E.g. calibration devices can be activated outside the measurement time or an antenna can be pointed when the pointing performance is not critical. 6. Isolation Systems: In case the above measures do not provide sufficiently low microvibration environment, an isolation systems can be implemented. Microvibration sources and/or sensitive units may be isolated from each other by means of e.g. elastomer material or active damping systems. The design trade-off for such system needs to be performed with a certain frequency control plan, which depends on sensitive frequencies as well as source emission frequencies. The isolation system also needs to be considered for the specification of source and sensitive elements
4. Payload interface requirements A key question in the management of microvibrations is the definition of interface requirements towards a sensitive receiver. The most typical example is a large optical instrument mounted on top of a 463
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 3. Payload microvibration interface specification approach. Fig. 2. Microvibration disturbance minimization process.
The payload supplier is then supposed to evaluate the requirement by performing a sine-sweep of the peak-level in each frequency band and determine the Line of Sight impact at each frequency. This process is illustrated in Fig. 3. By this approach, the worst case LoS impact per frequency is estimated for the case that the satellite platform resonance in that frequency band hits the payload resonance in the same frequency band. The evaluated LoS impact provides an artificial worst case, since the frequencies will not match exactly and when coupling the payload with the satellite platform, the resonances split into more than one peak while the split resonances usually show a reduced amplitude. This split of resonances has been observed on a number of tests performed on different systems. However, the effect can only be quantified for a specific system, since it depends on the local design, e.g. the support structure of a source and the involved modes that are coupling. However, based on our heritage we found that a reasonable assumption for the reduction factor is
satellite platform. During the development of such an instrument, its performance with respect to microvibrations has to be analysed and/or tested. Thus microvibration interface requirements have to be defined against which the instrument performance is assessed. The situation for microvibrations is very similar to the case of mechanical launch environments, where the design load for instruments or other equipment are significantly over-specified. For mechanical design loads, relaxations are treated by notching. This practice is well established and suppliers of equipment know which relaxation of loads is likely to be accepted. In the case of microvibrations, such process is not established throughout space industry. There is also no common heritage of acceptable disturbances levels, such that a notching approach would not be adequate. In order to define microvibration interface requirements towards sensitive receivers we have defined an approach, which aims on one hand for the definition of interface levels that are always achievable in the system, and on the other hand for minimizing the margins in the specified interface levels in order to allow the design of the receiver against the specified levels. The specification approach is performed in the following steps: The frequency range, over which the LoS3 angle, φLoS, is defined, is split into several smaller frequency bands, based on a-priori knowledge about the system. These bands could for example be related to Reaction Wheel harmonic frequencies and satellite platform primary modes. The frequency bands are chosen such that they are dominated by a single peak or a very small number of peaks. The spectrum of interface acceleration is computed and split into the frequency bands. On each band, the dominating peak acceleration, apk, is identified and an average acceleration level, aavg, is derived from the residual ‘signal power’ in the band. 3
cred =
max (
LoS , coupled (f ))
max( apk HPL )
which is the ratio between the maximum Line of Sight error in the coupled analysis, φLoS,coupled, and the satellite bus interface peak accelerations apk, multiplied by the payload frequency response, HPL. The interface accelerations are injected commonly at the Payload-toPlatform mechanical interface points in hard-mounted configuration. This method has been applied on a large scale satellite model. It was shown that the calculated interface parameters derived from FE Models were representative also for the measured interface levels. The requirement and the measured interface acceleration spectrum are shown in Fig. 4. 5. Microvibration control logic In the previous sections, activities have been outlined that are needed in order to control the microvibration performance of a satellite.
LoS: Line of Sight. 464
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 4. Payload interface requirement and measured interface accelerations.
This section puts these activities into a larger scheme, which is shown in Fig. 5. The diagram considers a split of the satellite into a payload and a platform on the so-called Module level. That accounts for the fact that many satellites are split in terms of responsibility for platform and payload. The upper part of the flow deals with the preliminary satellite design and analysis that is performed in early project phases. In later project phases, the microvibration performance is ensured by constraining each critical element with microvibration requirements. 5.1. Requirement definition As described in the previous section, payload microvibration requirements can be defined. In the same manner the platform requirements can be specified in terms of maximum interface accelerations at the platform to payload interface. Lower level microvibration requirements are derived at module level. This derivation can be in fact quite challenging and may lead to large additional margin. Alternatively, the unit level requirements could be derived from satellite level analysis. However, this requires to make various detailed design assumptions at satellite level, which would need to be ensured throughout the satellite development process by imposing additional requirements on platform and instrument providers. In order to ensure that margins are kept reasonably small, it is foreseen to derive lower level microvibration requirements at module level (instead of satellite level) and to confirm these by a satellite level analysis.
Fig. 5. Generic microvibration control logic.
terms of microvibrations, this is either by heritage or analysis data, or – for medium of highly sensitive missions – by test using Kistler table setups. 5.2.2. Verification of structural transfer functions The structural transfer behaviour has to be verified. Representative results cannot be obtained at structure level, since a representative behaviour is only present in the fully equipped satellite. Thus transfer function verification has to be performed at module and satellite level. 5.2.3. Verification of susceptible receivers Verification of large optical instruments are combined with measurements at module level, since they cannot be treated separately from the structure. There are other more compact susceptible receivers such as Fourier Spectrometers, atomic reference clocks or gyroscopes that can be verified at assembly level. Many companies that provide such susceptible receivers have developed test setups that allow disturbance injection at the interface of the receiver. This allows verification of the receiver's performance stand alone.
5.2. Requirement verification Microvibration requirement verification is performed at different levels. How they are verified (by analysis and/or test) depends on the criticality of the microvibration impact on the satellite performance. From a pure technical perspective, structure requirements would be defined for the satellite as a whole. However, in many projects, they are defined at module level, since the payload structure and the platform structure are under the responsibility of different entities. Therefore, a pre-verification is done at module level (dashed boxes in Fig. 5). However, a full verification of the structure design is only achieved at satellite level, when both modules a coupled.
5.2.4. Module level verification Verification at module (platform and instrument) level is sometimes foreseen. However, the results that can be obtained at module level have limited representativeness, since any coupling between the platform and the payload cannot be characterized and verification is performed against rather artificial interface levels. Nevertheless, a preverification at module level often makes sense, since it can be
5.2.1. Verification at source level The verification of sources is done on unit/sub-system level. A standard approach can be used. Depending on the mission criticality in 465
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 6. Satellite free-free suspension implemented by springs attached to the hoisting points.
Fig. 9. Overlap of disturbances of two Reaction Wheels. Comparision between measurement and prediction using a linear sum and a Root Sum Square. Fig. 7. Microvibration Test Setup used in the MicroMap Study.
Fig. 8. Mini shakers installed in the test setup. Left: combined axial and radial force exciter installed at Reaction Wheel 1 location; right: axial exciter installed at Reaction Wheel 2 location.
Fig. 10. Superposition of disturbances injected in different Degrees of Freedom of a single Reaction Wheel.
elements are sampled and used as input for optical modelling.
performed rather early in the program and some important aspects such as coupling of sources with the structure can be characterized.
6. Satellite microvibration testing
5.2.5. Satellite level verification Full verification is achieved at satellite level, where all sources, the entire satellite structure and the receivers are present. A common approach is to characterize the overall transfer function from relevant sources to the susceptible receiver. The coupling of sources with the structure can be characterized using a representative noise source. In case of optical telescopes, the performance measurement, i.e. Line of Sight, is barely measureable directly. Thus the accelerations at optical
In the past decade OHB has gained significant experience in microvibration testing due to activities performed in the frame of various satellite programs and studies. In this section some results are presented that validate the methods described in this paper. Microvibration tests at satellite level are performed on a satellite that is configured to be as representative as possible for its in-flight behaviour. This is achieved by spring suspension that decouples the 466
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 11. Disturbance at receiver composed of a number of resonances.
Fig. 12. Disturbance at receiver when one resonance is dominant.
satellite from the mechanical environment with low cut-off frequency and thus providing a near free-free configuration as shown in Fig. 6. Commonly, microvibration tests are performed on structural models of satellites. Compared to Flight Models, these allow much more flexibility in terms of test instrumentation, while providing fully flight representative structural behaviour in the low frequency range, which is in many cases sufficient to demonstrate the performance compatibility with respect to microvibrations. In case the mid and high frequencies are of concern, an additional test on the Flight Model can be performed which should focus on determining the different behaviour compared to the Structural Model. An example of a test setup is shown in Fig. 7. It was used in the frame of the MicroMap Study performed at OHB in the frame of an ESA4 development activity. It made use of a Structural Model of a typical geo-stationary satellite. A telescope breadboard was mounted on top. This breadboard was built and characterized in the frame of another study performed by OHB Munich. The MicroMap study included various investigations aiming at defining methods for the treatment of microvibrations in the satellite development process. In this paper, we present results related to the behaviour of different noise sources and their superposition. For this study the Reaction Wheels were replaced by mini shakers for injection
4
of forces. For both test runs presented hereafter, a combined axial and radial force exciter was used at one of the Reaction Wheel locations. The other Reaction Wheel location was equipped with an axial exciter as shown in Fig. 8. One frequent topic appearing during the development of satellites is related to summation rules to be considered for different microvibration noise source channels. This aspect was addressed in the test setup shown in Fig. 8 and results are provided in the following. 6.1. Superposition of two separate sources In order to understand how the noise of two different Reaction Wheels overlap, we have introduced disturbances at their interface bracket (random noise at Reaction Wheel 1 and sine sweep at Reaction Wheel 2) using the exciters shown in Fig. 7. The superposition at the receiver was measured and evaluated. The results in Fig. 9 show that a Root Sum Square has to be considered when superimposing two different Reaction Wheels. 6.2. Superposition of disturbance contributors of a single source In a second step the superposition of noise injected into different Degrees of Freedom of a single Reaction Wheel were investigated. The test was performed using the exciter installed in Reaction Wheel 1 location. Two data sets are compared:
ESA: European Space Agency. 467
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
- correlated noise simultaneously injected into 2 radial and 1 axial Reaction Wheel channel - Input injected separately into 2 radial and 1 axial Reaction Wheel channels and multiplied with the Frequency Response that was measured in a separate test run
Reaction Wheels and a Microvibration Isolator. J Virtanen and V. Nieminen from VTT Finland have supported the activities with FE modelling. A. Grillenbeck and T. Lechelmayr from IABG have supported system level test with test instrumentation and execution. References
The result is shown in Fig. 10 comparing the direct measurement with the summation of the different channels using a Linear Sum and a Quadratic Sum. Also for the different Degrees Of Freedom of a single Reaction Wheel, the results indicate that a quadratic Root Sum Square can be assumed. However, deviations have been identified for a small number of resonances, where a linear sum seems more suitable.
[1] M. Vitelli, B. Specht, F. Boquet, A Process to Verify the Microvibration and Pointing Stability Requirements for the Bepicolombo Misson, International Workshop on Instrumentation for Planetary Missions, 2012, p. 1023 PI Contribution No. 1683. [2] T. Wacker, L. Weimer, K. Eckert, GOCE platform micro-vibration verification by test and analysis, Proc. 13th European Conf. On Spacecraft Structures, Materials & Environmental Testing (ESA SP-727), 2005. [3] N. Yoshida, O. Takahara, T. Kosugi, K. Ninomiya, T. Hashimoto, K. Minesugi, S. Tsuneta, K. Ichimoto, S. Shimada, Systematic approach to achieve fine pointing requirement of SOLAR-B, IFAC Proc. Vol. 37 (2004) 101–106, https://doi.org/10. 1016/S1474-6670(17)32157-2. [4] S. Li, H. Zhang, S. Liu, Y. Wang, Multilevel microvibration test for performance predictions of a space optical load platform, Rev. Sci. Instrum. 89 (2018) 054901, , https://doi.org/10.1063/1.5005590. [5] M. Remedia, G. Aglietti, Z. Zhang, B.L. Page, G. Richardson, A general methodology to study the transmission of micro-vibrations in satellites, Proc. Int. Astronaut. Congr. IAC 8 (2012) 6386–6393. [6] Y. Gu, L. Wang, Q. Wu, A micro-vibration generated method for testing the imaging quality on ground of space remote sensing, Proc. SPIE 10710, Young Scientists Forum, vol. 2017, 2018, p. 107101D. [7] S. Rapp, A. Kuisl, C. Zauner, G. Reinicke, H. Baier, Experimental verification of a micro-vibration analysis software for pointing error prediction of space optical instruments, Proc. 13th European Conf. On Spacecraft Structures, Materials & Environmental Testing (ESA SP-727), 2014. [8] A.R. Prashant, M. Madheswaran, V. Kartik, V.R. Naidu, P. Govindan, P. Aravindakshan, System development for micro vibration measurements on spacecrafts, International Conference on Control, Instrumentation, Communication and Computational Technologies (ICCICCT), 2016, pp. 98–103, , https://doi.org/ 10.1109/ICCICCT.2016.7987925 Kumaracoil, 2016. [9] J. Franck, S. Denis, Specification of sensitive instruments under micro-vibrations orbital environment. Case of the MTG-S sounder interferometer's corner cube, Proc. 13th European Conf. On Spacecraft Structures, Materials & Environmental Testing (ESA SP-727), 2014. [10] G. Smet, G. Richardson, S. McLaren, A. Haslehurst, Managing reaction Wheel microvibration on a high resolution EO small spacecraft, 15th European space mechanisms and tribology symposium, Proceedings of the Conference Held 25-27 September, 2013 in Noordwijk, the Netherlands, vol. 718, L. Ouwehand. ESA-SP, 2013#978-92-9221-282-7. [11] M. Remedia, G. Aglietti, G. Richardson, M. Sweeting, Integrated semiempirical methodology for microvibration prediction, AIAA J. 53 (2015) 1236–1250, https:// doi.org/10.2514/1.J053339. [12] N. Jedrich, D. Zimbelman, M. Turczyn, J. Sills, C. Voorhees, B. Clapp, Brian, Cryo Cooler Induced MicroVibration Disturbances to the Hubble Space Telescope, NASA STI/Recon Technical Report N (2002). [13] K. Komatsu, H. Uchida, Microvibration in spacecraft, Mech. Eng. Rev. (2014), https://doi.org/10.1299/mer.2014se0010 1. SE0010-SE0010. [14] F. Angeletti, P. Gasbarri, Uncertainty Impact on Micro-Vibration Control of an Orbiting Large Adaptive Space Structure, EasyChair Preprint, 2018, p. 645, https:// doi.org/10.29007/332j. [15] J.R. Maly, S.C. Pendleton, J. Salmanoff, G.J. Blount, K. Mathews, Hubble Space Telescope solar array damper, Proc. SPIE 3672 (1999) 186–197 Smart Structures and Materials 1999: Passive Damping and Isolation. [16] P. Valentin, Robust Microvibration Control and Worst-Case Analysis for High Pointing Stability Space Missions (Thesis), (2017). [17] Q. Luo, D. Li, J. Jiang, Analysis and optimization of microvibration isolation for multiple flywheel systems of spacecraft, AIAA J. 54 (2016) 1–13, https://doi.org/ 10.2514/1.J053885. [18] L. Li, L. Tan, K. Lin, D. Wang, Modeling and test of flywheel vibration isolation system for space telescope, J. Vibroeng. 19 (2017) 1613–1623, https://doi.org/10. 21595/jve.2016.17520. [19] Y. Zhenguo, K. Xianren, X. Sirui, Exploration of Magnetic Suspension Technology used for Spacecraft Micro-Vibration Isolation, International Conference on Mechatronics and Control (ICMC), Jinzhou, 2014, pp. 1545–1549, , https://doi. org/10.1109/ICMC.2014.7231818 2014. [20] H. Xing, Y. Feng, L. Liu, Passive double-layer micro-vibration control of optical spacecrafts, 2nd International Workshop on Materials Engineering and Computer Sciences, Atlantis Press, 20152352-538X. [21] C. Xu, Z.-D. Xu, X.-H. Huang, Y.-S. Xu, T. Ge, Modelling and analysis of a viscoelastic micro-vibration isolation and mitigation platform for spacecraft, J. Vib. Control 24 (18) (2018) 4337–4352. [22] ESSB-HB-E-003: ESA Pointing Error Engineering Handbook. [23] ECSS-E-ST-60-10C: Space Engineering - Control Performance. [24] ECSS-E-HB-60-10A: Space Engineering - Control Performance Guidelines.
6.3. Statistical behaviour of disturbances The test data was analysed for the statistical behaviour of the disturbance at the receiving end in order to identify what kind of statistical distribution the signal shows. In most cases the disturbance is composed of a number of resonances as shown in Fig. 11. The resulting distribution of the signal shows a behaviour close to a Normal distribution. However, in some cases, in particular if a resonance of the system is excited, the disturbance at the receiver is dominated by a single resonance as shown in Fig. 12. In that case the statistical distribution shows a bi-modal behaviour. The appearance of this behaviour is very limited and shows up only in a subset of measured channels and for a small number of excitation frequencies. We have concluded that it is not mandatory to consider this behaviour in the budget. However, their occurrences should be identified in test data and assessed in system level analysis case by case. 7. Conclusions This paper provides guidelines on how to establish a performance budget for satellites for which microvibrations are critical. The guidelines are generic and can be applied to a large number of budgets. For application, the disturbance source characteristic has to be identified and the system disturbance period has to be defined. We have identified potential microvibration minimization measures and provided a scheme on how to apply them during the satellite development phase. This scheme is embedded into a larger scheme that defines all levels at which requirements have to be defined and how they have to be verified. A possible way on specifying microvibration interface requirements towards sensitive receivers has been outlined. Various different approaches are applied in industry. Our approach focuses specifically on the definition of interface levels with minimized margin to allow the providers of the receivers to design against these interface loads. We have presented test results related to the superposition of noise sources, which confirm our methods defined in this paper. Acknowledgements This paper is largely based on results obtained in the frame of the MicroMap study performed within the study AO/1-7706/13/NL/SFe “Methodology for Microvibration Management at System Level” funded by the European Space Agency / European Spaceh Research an Technology Centre (ESA/ESTEC), The Netherlands. Special thank go to E. Yoo and A. Kuisl from OHB Munich for the provision of the optical telescope breadboard and the support of the test activities. B. Paijmans and C. Lauwerys from QuinetiQ Space Belgium have contributed to the activities with the modelling and testing of
468
Contents lists available at ScienceDirect
Acta Astronautica journal homepage: www.elsevier.com/locate/actaastro
Managing the microvibration impact on satellite performances a,∗
a
a
Frank Steier , Torben Runte , Anneke Monsky , Timo Klock a b
a,1
b
, Gregory Laduree
T
OHB System AG Bremen, Universitätsallee 27-29, D-28359 Bremen, Germany ESA/ESTEC, Keplerlaan 1, 2200 Noordwijk, the Netherlands
ARTICLE INFO
ABSTRACT
Keywords: Microvibration Control Isolator Disturbance Superposition Overlap MicroMap
Due to the increasing needs on the performance of satellites, in particular optical satellites, microvibrations have become more and more important over the last years. Microvibrations affect a large number of satellites for both Earth observation and space observation missions. Further, microvibrations are also a concern for other missions with sensitive instruments like Fourier spectrometers, high-precision accelerometers or reference oscillators. In classical fields like structural dynamic engineering, management approaches have been established and implemented in all space projects. This paper presents a corresponding approach for the area of microvibrations. Focus is on the definition of all key elements needed to establish a microvibration control approach. This includes the establishment of microvibration budgets and associated summation rules. A systematic way for microvibration disturbance minimization is presented and rules for the application are defined. Specific emphasis is given to the definition of microvibration interface requirements and the associated control process. The concepts provided in this paper have been verified by tests performed in the frame of various satellite programs and studies. Tests related to the superposition of different noise sources are presented in here.
1. Introduction In literature many publications on microvibrations can be found. These are mostly related to specific aspects, while only a few address system level aspects and the underlying engineering process. Certain insight on the microvibration engineering process is provided for Bepicolomobo [1], GOCE [2], Solar-B [3], an optical satellite system [4] and a satellite platform [5]. Others system level publications are related to simulation [6] and testing [7,8], and to the definition of microvibration interface requirements [9]. In addition, numerous publications can be found that focus on specific aspects of the microvibration engineering management process. These are related to the modelling and testing of microvibration sources [10–12], structural modelling [13] and design considerations for appendages [14,15]. Various recent publications in the area of microvibrations are related to isolations systems [16–21]. This paper provides a general microvibration engineering process which is not linked to a specific satellite. It complements the existing approaches and provides engineering processes and guidelines for the important elements of the engineering process. In addition to existing articles, it provides generalized budgeting rules and an interface requirement definition approach aiming for specifications with
minimized margins. Section 2 deals with the microvibration performance budget and provides guidelines on how to consider noise sources and the associated summation rules. Section 3 provides a systematic approach on the disturbance minimization process to be applied during the satellite design phase. Section 4 defines a methodology to establish microvibration interface requirements towards susceptible payloads. Section 5 provides the overall control logic that is applied during the satellite development. Emphasis is given to requirement definition and verification at different levels. Section 6 provides test results that provide information on the superposition of noise sources and the statistical behaviour of the disturbances. 2. Microvibration performance budget and noise sources In many satellites, microvibrations have to be considered as one contributor to the payload key performance parameters. Most commonly, this performance is linked to the pointing performance. However, there are other sensitive payloads such as high precision accelerometers, spectrometers or atomic reference oscillators, which could be impaired by microvibrations. One major activity of early satellite studies is the establishment of
Corresponding author. E-mail address: [email protected] (F. Steier). 1 Current address: Simula Research Laboratory, Martin Linges vei 25, 1364 Norway. ∗
https://doi.org/10.1016/j.actaastro.2019.06.027 Received 28 January 2019; Received in revised form 26 April 2019; Accepted 20 June 2019 Available online 22 June 2019 0094-5765/ © 2019 IAA. Published by Elsevier Ltd. All rights reserved.
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
the satellite performance budget, which is needed to define the overall satellite architecture. Here we explain, how the budget can be established when microvibrations have to be considered. Budgeting rules are defined for satellite pointing, however the method can be applied correspondingly for other performance parameters as well. There are guidelines provided by ECSS2 [22–24] on how to evaluate the pointing performance of a satellite. Microvibrations typically affect the error budgets related to the following performance parameters:
• APE (Absolute Pointing Error) • RPE (Relative Pointing Error) and • PDE (Pointing Drift Error). These errors are illustrated in Fig. 1, where ε is the deviation from the nominal satellite pointing and Δt the observation period. The RPE is the maximum deviation from the Mean Point Error during the time Δt. The Figure also shows the Mean Pointing Error (MPE) for which the microvibration contribution is usually negligible. Before establishing the budget, a consideration for all sources of microvibrations aboard the satellite has to be made in terms of their statistical behaviour in order to apply the appropriate summation rules. Typical sources on satellites are either periodic or transient sources. Random sources are usually less important and are therefore not further considered in the following discussion. Depending on the way the sources are operated during satellite operation, their statistical behaviour can be determined. For transient sources the statistical behaviour can be: - Overlapping, - Non-overlapping, or - Randomly appearing;
Fig. 1. Pointing errors affected by microvibrations.
while periodic sources can either be
generic budget for microvibrations is shown for short term disturbances in Table 1. The expectation value of the satellite pointing, < E > , is the sum of the individual contributors defined as maximum error emax. The budget for intermediate term disturbances is shown in Table 2. It uses modified summation rules that account for the fact that the disturbance does not reach its full amplitude over the observation time. In that case the summation has to consider the maximum amplitude over the disturbance period, e∼ max, for transient sources; and a weighing function for periodic sources. The weighting function is defined by w(ω,Δt) = sin(ωΔt/2) for ω ≤ π/Δt, and w(ω,Δt) = 1 for ω > π/Δt.
- Correlated, or - Uncorrelated. It shall be noted that a source can show more than one statistical behaviour at the same time – e.g. Reaction Wheels generate disturbances by imbalance effects which are harmonic; and friction jumps which are random effects. These effects have to be treated as individual errors in the performance budget. For periodic errors such as micro vibrations, the impact of the disturbance on the spacecraft performance depends on the relation of the microvibration oscillation period (disturbance period) with respect to the observation period. If the disturbance period is short compared to the observation period, the average deviation from the Mean Pointing Error is zero. When establishing a microvibration budget, two cases have to be distinguished:
3. Disturbance Minimization There are different measures to minimize the impact of microvibrations on the spacecraft performance. Minimization measures are implemented starting in phase A of the project, but have to be followed iteratively throughout the design phases. Possible minimization measures are:
- Short disturbance period: Disturbance period is short compared to the observation period. - Intermediate disturbance period: Disturbance period is long or of the order of the observation period.
1. Mission concept:
ECSS [23] define the terms short disturbance period and long disturbance period, while the long disturbance period means that variations are much longer than the observation period. For microvibration the new term “intermediate disturbance period” is introduced. This is needed since for many typical spacecraft, the disturbance period is of the same order as the observation period which requires the definition of a weighting function as defined at the end of this paragraph. The budgets are established differently for these two cases. The 2
Based on the criticality of the satellite to microvibrations, an initial concept has to be established. One could choose e.g. standard attitude control concepts based on Reaction Wheels or lay out more comprehensive control concepts based on microNewton thrusters. 2. Frequency Control: In order to be able to control the satellite performance throughout the development cycle, it is mandatory to keep control over the frequencies relevant for microvibrations. These are the excitation frequencies of sources, susceptible frequencies of receivers and structural
ECSS: European Cooperation for Space Standardization. 462
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Table 1 Pointing Budget rules in case of short disturbance period. Source Type Statistical behaviour
Summation
Examples
Transient
Linear Worst case Linear Linear RSS
Σ emax emax Σ emax Σ emax σ assuming Gaussian distribution
Mechanisms operated at the same time Calibration wheel, scanner start-up Reaction Wheel friction jumps, sudden stress release Solar Array Drive stepper motors, resonances within one Reaction Wheel Disturbances within one stepper motor, disturbances from different Reaction Wheels
Periodic
Overlapping Non-overlapping Randomly appearing Correlated Uncorrelated
Table 2 Pointing Budget rules in case of intermediate disturbance period. Source Type Statistical behaviour Transient Periodic
Overlapping Non-overlapping Randomly appearing Correlated Uncorrelated
Summation
Examples
Linear Worst case Linear Linear Linear
Σ e∼ max e∼ max Σ e∼ max
Mechanisms operated at the same time Calibration wheel, scanner start-up Reaction Wheel friction jumps, sudden stress release Solar Array Drive stepper motors, resonances within one Reaction Wheel Disturbances within one stepper motor, Reaction Wheel imbalances, disturbances from different Reaction Wheels
Σ w(f)emax Σ w(f)emax
resonances. The resonance frequencies can be suppressed by design measures of these elements, but it is normally more important to separate them, in order to avoid amplifications.
and for the structural design, since it affects the harmonic disturbances and resonance frequencies. The above measures for minimization can be applied in a sequential process during the design phase as shown in Fig. 2. The process starts in Phase A, in which a dedicated microvibration requirement at system level is derived based on the mission requirements. Depending on the stringency of these requirements, the relevant sources and potentially sensitive devices are identified. Based on this rudimental information, a preliminary assessment of the achievable performance can be carried out, in order to indicate the measures to be implemented in Phase B. In Phase B, a detailed analysis on all possible measures can be performed taking into account more mature information of the microvibration critical elements. The flow shown in Fig. 2 aims for minimum iterations in the design process. However, some of the control measures could also be applied in parallel. The key element is the Frequency Control which should be analysed first, since it impacts the other control measures. After an initial iteration, a first detailed analysis can be performed which is indicated in Fig. 2 as “Microvibration Satellite Level Nominal Analysis”. This analysis provides the relevant data to decide whether or not to implement an isolation system and to define the location and cutoff frequencies of the isolator(s), which usually cannot be done a priori. Depending on the remaining margins, this process may be applied iteratively until a robust satellite design is found. For missions that are sensitive to microvibrations, a sensitivity analysis is performed after the minimization loop is finished. The objective of the sensitivity analysis is to define available margins and acceptable frequency ranges of critical harmonics that have to be known during the unit level development and implementation activities in later project stages. Similar to other engineering areas such as structure or EMC engineering, critcal frequencies, allowed ranges and amplitudfes of resonances are summarized in the Microvibration Frequency Control Plan, which is intended to support tracking of evolutions throughout the project life cycle.
3. Unit Level Coordination: At equipment level, there is some tuning capability of the disturbances. E.g. the frequency of a cryogenic cooler can be chosen for minimum impact or mechanisms can be actuated with specific actuation profiles that minimize the microvibration disturbances in the critical frequencies. In case of off-the-shelf components, such tuning capability is normally not given. 4. Accommodation & Structural Design: The satellite accommodation can be chosen such that emitters of microvibrations are located far away from the sensitive receivers. Further, the microvibration load path can be optimized for the transmission of critical frequencies. In order to design the load path effectively, the coupling of the source with the structure has to be modelled correctly and optimized for separation of the critical frequencies. Managing the coupling effect is typically most challenging, since its correct modelling requires correlation with test data. 5. Operational Concepts: In many cases it is possible to schedule the activation of mechanisms aboard a satellite such to minimize the impact of microvibrations. E.g. calibration devices can be activated outside the measurement time or an antenna can be pointed when the pointing performance is not critical. 6. Isolation Systems: In case the above measures do not provide sufficiently low microvibration environment, an isolation systems can be implemented. Microvibration sources and/or sensitive units may be isolated from each other by means of e.g. elastomer material or active damping systems. The design trade-off for such system needs to be performed with a certain frequency control plan, which depends on sensitive frequencies as well as source emission frequencies. The isolation system also needs to be considered for the specification of source and sensitive elements
4. Payload interface requirements A key question in the management of microvibrations is the definition of interface requirements towards a sensitive receiver. The most typical example is a large optical instrument mounted on top of a 463
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 3. Payload microvibration interface specification approach. Fig. 2. Microvibration disturbance minimization process.
The payload supplier is then supposed to evaluate the requirement by performing a sine-sweep of the peak-level in each frequency band and determine the Line of Sight impact at each frequency. This process is illustrated in Fig. 3. By this approach, the worst case LoS impact per frequency is estimated for the case that the satellite platform resonance in that frequency band hits the payload resonance in the same frequency band. The evaluated LoS impact provides an artificial worst case, since the frequencies will not match exactly and when coupling the payload with the satellite platform, the resonances split into more than one peak while the split resonances usually show a reduced amplitude. This split of resonances has been observed on a number of tests performed on different systems. However, the effect can only be quantified for a specific system, since it depends on the local design, e.g. the support structure of a source and the involved modes that are coupling. However, based on our heritage we found that a reasonable assumption for the reduction factor is
satellite platform. During the development of such an instrument, its performance with respect to microvibrations has to be analysed and/or tested. Thus microvibration interface requirements have to be defined against which the instrument performance is assessed. The situation for microvibrations is very similar to the case of mechanical launch environments, where the design load for instruments or other equipment are significantly over-specified. For mechanical design loads, relaxations are treated by notching. This practice is well established and suppliers of equipment know which relaxation of loads is likely to be accepted. In the case of microvibrations, such process is not established throughout space industry. There is also no common heritage of acceptable disturbances levels, such that a notching approach would not be adequate. In order to define microvibration interface requirements towards sensitive receivers we have defined an approach, which aims on one hand for the definition of interface levels that are always achievable in the system, and on the other hand for minimizing the margins in the specified interface levels in order to allow the design of the receiver against the specified levels. The specification approach is performed in the following steps: The frequency range, over which the LoS3 angle, φLoS, is defined, is split into several smaller frequency bands, based on a-priori knowledge about the system. These bands could for example be related to Reaction Wheel harmonic frequencies and satellite platform primary modes. The frequency bands are chosen such that they are dominated by a single peak or a very small number of peaks. The spectrum of interface acceleration is computed and split into the frequency bands. On each band, the dominating peak acceleration, apk, is identified and an average acceleration level, aavg, is derived from the residual ‘signal power’ in the band. 3
cred =
max (
LoS , coupled (f ))
max( apk HPL )
which is the ratio between the maximum Line of Sight error in the coupled analysis, φLoS,coupled, and the satellite bus interface peak accelerations apk, multiplied by the payload frequency response, HPL. The interface accelerations are injected commonly at the Payload-toPlatform mechanical interface points in hard-mounted configuration. This method has been applied on a large scale satellite model. It was shown that the calculated interface parameters derived from FE Models were representative also for the measured interface levels. The requirement and the measured interface acceleration spectrum are shown in Fig. 4. 5. Microvibration control logic In the previous sections, activities have been outlined that are needed in order to control the microvibration performance of a satellite.
LoS: Line of Sight. 464
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 4. Payload interface requirement and measured interface accelerations.
This section puts these activities into a larger scheme, which is shown in Fig. 5. The diagram considers a split of the satellite into a payload and a platform on the so-called Module level. That accounts for the fact that many satellites are split in terms of responsibility for platform and payload. The upper part of the flow deals with the preliminary satellite design and analysis that is performed in early project phases. In later project phases, the microvibration performance is ensured by constraining each critical element with microvibration requirements. 5.1. Requirement definition As described in the previous section, payload microvibration requirements can be defined. In the same manner the platform requirements can be specified in terms of maximum interface accelerations at the platform to payload interface. Lower level microvibration requirements are derived at module level. This derivation can be in fact quite challenging and may lead to large additional margin. Alternatively, the unit level requirements could be derived from satellite level analysis. However, this requires to make various detailed design assumptions at satellite level, which would need to be ensured throughout the satellite development process by imposing additional requirements on platform and instrument providers. In order to ensure that margins are kept reasonably small, it is foreseen to derive lower level microvibration requirements at module level (instead of satellite level) and to confirm these by a satellite level analysis.
Fig. 5. Generic microvibration control logic.
terms of microvibrations, this is either by heritage or analysis data, or – for medium of highly sensitive missions – by test using Kistler table setups. 5.2.2. Verification of structural transfer functions The structural transfer behaviour has to be verified. Representative results cannot be obtained at structure level, since a representative behaviour is only present in the fully equipped satellite. Thus transfer function verification has to be performed at module and satellite level. 5.2.3. Verification of susceptible receivers Verification of large optical instruments are combined with measurements at module level, since they cannot be treated separately from the structure. There are other more compact susceptible receivers such as Fourier Spectrometers, atomic reference clocks or gyroscopes that can be verified at assembly level. Many companies that provide such susceptible receivers have developed test setups that allow disturbance injection at the interface of the receiver. This allows verification of the receiver's performance stand alone.
5.2. Requirement verification Microvibration requirement verification is performed at different levels. How they are verified (by analysis and/or test) depends on the criticality of the microvibration impact on the satellite performance. From a pure technical perspective, structure requirements would be defined for the satellite as a whole. However, in many projects, they are defined at module level, since the payload structure and the platform structure are under the responsibility of different entities. Therefore, a pre-verification is done at module level (dashed boxes in Fig. 5). However, a full verification of the structure design is only achieved at satellite level, when both modules a coupled.
5.2.4. Module level verification Verification at module (platform and instrument) level is sometimes foreseen. However, the results that can be obtained at module level have limited representativeness, since any coupling between the platform and the payload cannot be characterized and verification is performed against rather artificial interface levels. Nevertheless, a preverification at module level often makes sense, since it can be
5.2.1. Verification at source level The verification of sources is done on unit/sub-system level. A standard approach can be used. Depending on the mission criticality in 465
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 6. Satellite free-free suspension implemented by springs attached to the hoisting points.
Fig. 9. Overlap of disturbances of two Reaction Wheels. Comparision between measurement and prediction using a linear sum and a Root Sum Square. Fig. 7. Microvibration Test Setup used in the MicroMap Study.
Fig. 8. Mini shakers installed in the test setup. Left: combined axial and radial force exciter installed at Reaction Wheel 1 location; right: axial exciter installed at Reaction Wheel 2 location.
Fig. 10. Superposition of disturbances injected in different Degrees of Freedom of a single Reaction Wheel.
elements are sampled and used as input for optical modelling.
performed rather early in the program and some important aspects such as coupling of sources with the structure can be characterized.
6. Satellite microvibration testing
5.2.5. Satellite level verification Full verification is achieved at satellite level, where all sources, the entire satellite structure and the receivers are present. A common approach is to characterize the overall transfer function from relevant sources to the susceptible receiver. The coupling of sources with the structure can be characterized using a representative noise source. In case of optical telescopes, the performance measurement, i.e. Line of Sight, is barely measureable directly. Thus the accelerations at optical
In the past decade OHB has gained significant experience in microvibration testing due to activities performed in the frame of various satellite programs and studies. In this section some results are presented that validate the methods described in this paper. Microvibration tests at satellite level are performed on a satellite that is configured to be as representative as possible for its in-flight behaviour. This is achieved by spring suspension that decouples the 466
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
Fig. 11. Disturbance at receiver composed of a number of resonances.
Fig. 12. Disturbance at receiver when one resonance is dominant.
satellite from the mechanical environment with low cut-off frequency and thus providing a near free-free configuration as shown in Fig. 6. Commonly, microvibration tests are performed on structural models of satellites. Compared to Flight Models, these allow much more flexibility in terms of test instrumentation, while providing fully flight representative structural behaviour in the low frequency range, which is in many cases sufficient to demonstrate the performance compatibility with respect to microvibrations. In case the mid and high frequencies are of concern, an additional test on the Flight Model can be performed which should focus on determining the different behaviour compared to the Structural Model. An example of a test setup is shown in Fig. 7. It was used in the frame of the MicroMap Study performed at OHB in the frame of an ESA4 development activity. It made use of a Structural Model of a typical geo-stationary satellite. A telescope breadboard was mounted on top. This breadboard was built and characterized in the frame of another study performed by OHB Munich. The MicroMap study included various investigations aiming at defining methods for the treatment of microvibrations in the satellite development process. In this paper, we present results related to the behaviour of different noise sources and their superposition. For this study the Reaction Wheels were replaced by mini shakers for injection
4
of forces. For both test runs presented hereafter, a combined axial and radial force exciter was used at one of the Reaction Wheel locations. The other Reaction Wheel location was equipped with an axial exciter as shown in Fig. 8. One frequent topic appearing during the development of satellites is related to summation rules to be considered for different microvibration noise source channels. This aspect was addressed in the test setup shown in Fig. 8 and results are provided in the following. 6.1. Superposition of two separate sources In order to understand how the noise of two different Reaction Wheels overlap, we have introduced disturbances at their interface bracket (random noise at Reaction Wheel 1 and sine sweep at Reaction Wheel 2) using the exciters shown in Fig. 7. The superposition at the receiver was measured and evaluated. The results in Fig. 9 show that a Root Sum Square has to be considered when superimposing two different Reaction Wheels. 6.2. Superposition of disturbance contributors of a single source In a second step the superposition of noise injected into different Degrees of Freedom of a single Reaction Wheel were investigated. The test was performed using the exciter installed in Reaction Wheel 1 location. Two data sets are compared:
ESA: European Space Agency. 467
Acta Astronautica 162 (2019) 461–468
F. Steier, et al.
- correlated noise simultaneously injected into 2 radial and 1 axial Reaction Wheel channel - Input injected separately into 2 radial and 1 axial Reaction Wheel channels and multiplied with the Frequency Response that was measured in a separate test run
Reaction Wheels and a Microvibration Isolator. J Virtanen and V. Nieminen from VTT Finland have supported the activities with FE modelling. A. Grillenbeck and T. Lechelmayr from IABG have supported system level test with test instrumentation and execution. References
The result is shown in Fig. 10 comparing the direct measurement with the summation of the different channels using a Linear Sum and a Quadratic Sum. Also for the different Degrees Of Freedom of a single Reaction Wheel, the results indicate that a quadratic Root Sum Square can be assumed. However, deviations have been identified for a small number of resonances, where a linear sum seems more suitable.
[1] M. Vitelli, B. Specht, F. Boquet, A Process to Verify the Microvibration and Pointing Stability Requirements for the Bepicolombo Misson, International Workshop on Instrumentation for Planetary Missions, 2012, p. 1023 PI Contribution No. 1683. [2] T. Wacker, L. Weimer, K. Eckert, GOCE platform micro-vibration verification by test and analysis, Proc. 13th European Conf. On Spacecraft Structures, Materials & Environmental Testing (ESA SP-727), 2005. [3] N. Yoshida, O. Takahara, T. Kosugi, K. Ninomiya, T. Hashimoto, K. Minesugi, S. Tsuneta, K. Ichimoto, S. Shimada, Systematic approach to achieve fine pointing requirement of SOLAR-B, IFAC Proc. Vol. 37 (2004) 101–106, https://doi.org/10. 1016/S1474-6670(17)32157-2. [4] S. Li, H. Zhang, S. Liu, Y. Wang, Multilevel microvibration test for performance predictions of a space optical load platform, Rev. Sci. Instrum. 89 (2018) 054901, , https://doi.org/10.1063/1.5005590. [5] M. Remedia, G. Aglietti, Z. Zhang, B.L. Page, G. Richardson, A general methodology to study the transmission of micro-vibrations in satellites, Proc. Int. Astronaut. Congr. IAC 8 (2012) 6386–6393. [6] Y. Gu, L. Wang, Q. Wu, A micro-vibration generated method for testing the imaging quality on ground of space remote sensing, Proc. SPIE 10710, Young Scientists Forum, vol. 2017, 2018, p. 107101D. [7] S. Rapp, A. Kuisl, C. Zauner, G. Reinicke, H. Baier, Experimental verification of a micro-vibration analysis software for pointing error prediction of space optical instruments, Proc. 13th European Conf. On Spacecraft Structures, Materials & Environmental Testing (ESA SP-727), 2014. [8] A.R. Prashant, M. Madheswaran, V. Kartik, V.R. Naidu, P. Govindan, P. Aravindakshan, System development for micro vibration measurements on spacecrafts, International Conference on Control, Instrumentation, Communication and Computational Technologies (ICCICCT), 2016, pp. 98–103, , https://doi.org/ 10.1109/ICCICCT.2016.7987925 Kumaracoil, 2016. [9] J. Franck, S. Denis, Specification of sensitive instruments under micro-vibrations orbital environment. Case of the MTG-S sounder interferometer's corner cube, Proc. 13th European Conf. On Spacecraft Structures, Materials & Environmental Testing (ESA SP-727), 2014. [10] G. Smet, G. Richardson, S. McLaren, A. Haslehurst, Managing reaction Wheel microvibration on a high resolution EO small spacecraft, 15th European space mechanisms and tribology symposium, Proceedings of the Conference Held 25-27 September, 2013 in Noordwijk, the Netherlands, vol. 718, L. Ouwehand. ESA-SP, 2013#978-92-9221-282-7. [11] M. Remedia, G. Aglietti, G. Richardson, M. Sweeting, Integrated semiempirical methodology for microvibration prediction, AIAA J. 53 (2015) 1236–1250, https:// doi.org/10.2514/1.J053339. [12] N. Jedrich, D. Zimbelman, M. Turczyn, J. Sills, C. Voorhees, B. Clapp, Brian, Cryo Cooler Induced MicroVibration Disturbances to the Hubble Space Telescope, NASA STI/Recon Technical Report N (2002). [13] K. Komatsu, H. Uchida, Microvibration in spacecraft, Mech. Eng. Rev. (2014), https://doi.org/10.1299/mer.2014se0010 1. SE0010-SE0010. [14] F. Angeletti, P. Gasbarri, Uncertainty Impact on Micro-Vibration Control of an Orbiting Large Adaptive Space Structure, EasyChair Preprint, 2018, p. 645, https:// doi.org/10.29007/332j. [15] J.R. Maly, S.C. Pendleton, J. Salmanoff, G.J. Blount, K. Mathews, Hubble Space Telescope solar array damper, Proc. SPIE 3672 (1999) 186–197 Smart Structures and Materials 1999: Passive Damping and Isolation. [16] P. Valentin, Robust Microvibration Control and Worst-Case Analysis for High Pointing Stability Space Missions (Thesis), (2017). [17] Q. Luo, D. Li, J. Jiang, Analysis and optimization of microvibration isolation for multiple flywheel systems of spacecraft, AIAA J. 54 (2016) 1–13, https://doi.org/ 10.2514/1.J053885. [18] L. Li, L. Tan, K. Lin, D. Wang, Modeling and test of flywheel vibration isolation system for space telescope, J. Vibroeng. 19 (2017) 1613–1623, https://doi.org/10. 21595/jve.2016.17520. [19] Y. Zhenguo, K. Xianren, X. Sirui, Exploration of Magnetic Suspension Technology used for Spacecraft Micro-Vibration Isolation, International Conference on Mechatronics and Control (ICMC), Jinzhou, 2014, pp. 1545–1549, , https://doi. org/10.1109/ICMC.2014.7231818 2014. [20] H. Xing, Y. Feng, L. Liu, Passive double-layer micro-vibration control of optical spacecrafts, 2nd International Workshop on Materials Engineering and Computer Sciences, Atlantis Press, 20152352-538X. [21] C. Xu, Z.-D. Xu, X.-H. Huang, Y.-S. Xu, T. Ge, Modelling and analysis of a viscoelastic micro-vibration isolation and mitigation platform for spacecraft, J. Vib. Control 24 (18) (2018) 4337–4352. [22] ESSB-HB-E-003: ESA Pointing Error Engineering Handbook. [23] ECSS-E-ST-60-10C: Space Engineering - Control Performance. [24] ECSS-E-HB-60-10A: Space Engineering - Control Performance Guidelines.
6.3. Statistical behaviour of disturbances The test data was analysed for the statistical behaviour of the disturbance at the receiving end in order to identify what kind of statistical distribution the signal shows. In most cases the disturbance is composed of a number of resonances as shown in Fig. 11. The resulting distribution of the signal shows a behaviour close to a Normal distribution. However, in some cases, in particular if a resonance of the system is excited, the disturbance at the receiver is dominated by a single resonance as shown in Fig. 12. In that case the statistical distribution shows a bi-modal behaviour. The appearance of this behaviour is very limited and shows up only in a subset of measured channels and for a small number of excitation frequencies. We have concluded that it is not mandatory to consider this behaviour in the budget. However, their occurrences should be identified in test data and assessed in system level analysis case by case. 7. Conclusions This paper provides guidelines on how to establish a performance budget for satellites for which microvibrations are critical. The guidelines are generic and can be applied to a large number of budgets. For application, the disturbance source characteristic has to be identified and the system disturbance period has to be defined. We have identified potential microvibration minimization measures and provided a scheme on how to apply them during the satellite development phase. This scheme is embedded into a larger scheme that defines all levels at which requirements have to be defined and how they have to be verified. A possible way on specifying microvibration interface requirements towards sensitive receivers has been outlined. Various different approaches are applied in industry. Our approach focuses specifically on the definition of interface levels with minimized margin to allow the providers of the receivers to design against these interface loads. We have presented test results related to the superposition of noise sources, which confirm our methods defined in this paper. Acknowledgements This paper is largely based on results obtained in the frame of the MicroMap study performed within the study AO/1-7706/13/NL/SFe “Methodology for Microvibration Management at System Level” funded by the European Space Agency / European Spaceh Research an Technology Centre (ESA/ESTEC), The Netherlands. Special thank go to E. Yoo and A. Kuisl from OHB Munich for the provision of the optical telescope breadboard and the support of the test activities. B. Paijmans and C. Lauwerys from QuinetiQ Space Belgium have contributed to the activities with the modelling and testing of
468