Transcript
Experience with Force Limited Vibration Applied to the OLA Sensor Head Optical Instrument Yvan Soucy
Peter P. Krimbalis
Canadian Space Agency Saint-Hubert, QC, Canada
MDA Robotics and Automation Brampton, ON, Canada
Spacecraft and Launch Vehicle Dynamic Environments Workshop June 21-23, 2016
Introduction The NASA OSIRIS-REx* mission will visit the asteroid Bennu with the major objective of returning a sample: • Principal Investigator, Science Processing and Operations Center are at the University of Arizona • NASA GSFC provides overall mission management & systems engineering • Lockheed Martin is building the spacecraft
• Collaborators from the US and other countries (including Canada) • http://www.asteroidmission.org (Credit: Lockheed Martin) 2
*Origins Spectral Interpretation Resource Identification Security Regolith Explorer
Introduction (Cont’d) • The OLA* instrument is the Canadian Space Agency’s contribution to the OSIRIS-REx mission • MDA was contracted to design, build and test OLA • OLA is a scanning laser altimeter (LIDAR) that will be used to characterize the geometry of Bennu (surface craters, slopes, total volume) as well as providing enhanced navigational estimates and surface roughness estimates for the purposes of sample collection. • The OLA instrument consists of the • Sensor Head (SH) to transmit and receive the laser(s) • Main Electronics (ME) which is an avionics box
• The Sensor Head is a vibration-sensitive unit since it contains optics and mechanisms
• The Semi-Empirical method of Force-limited Vibration (FLV) was used for testing the Sensor Head because: • Presence of vibration-sensitive hardware • Project on a very tight and even tighter (more critical) schedule • FLV used to mitigate hardware failure caused by over-testing 3
*OSIRIS-REx Laser Altimeter
Content of Presentation The main sections of the presentation are: • Overview of Sensor Head • Quick review of FLV semi-empirical method • Vibration test sequence
• Test configuration and instrumentation • Test parameters and results • Sensitivity study of comparison between peak CG acceleration and Limit Load Factor (MAC) for different • Peak factors for CG acceleration (between 3 to 5) • Different bandwidths of force PSD’s (full test bandwidth or only up to first anti-resonance)
• Concluding remarks
4
Overview of OLA Sensor Head • The OLA SH design and system architecture are proprietary and subject to CGRP/ITAR restrictions
• Contains low and high energy lasers for a large effective range of measurement • Complex array of optics to manipulate outgoing and incoming laser light • Contains a bi-axial, precision controlled scanning mirror (azimuth & elevation) that enables OLA to scan complex patterns on Bennu’s surface • Laser sub-assemblies, optics and in particular the scanning mirror are all vibration sensitive components which served as the impetus for implementing FLV, in addition to manual notching 5
Review of FLV Semi-Empirical method • Basic relations for random vibration [1] :
Sff(f) = C2 M2 Saa(f)
f < fb
Sff(f) = C2 M2 Saa(f) / (f/fb)2n
f ≥ fb
where Sff : Sought force limits PSD Saa : Input acceleration specification PSD M : Physical mass of test item fb : Break frequency (usually fundamental frequency of test item) n : Constant to account for decrease in frequency of residual mass and asymptotic mass C2 : Configuration-dependent dimensionless constant 6
Selection of C2 Parameter • C2 is the only parameter that cannot be measured or derived directly from low-level runs • Several criteria can be used for selecting C2 - See for example [1, 2] • These criteria include • Extrapolation of interface force or notching of input acceleration from similar mounting structures and test items • Comparison with force limits from more analytical TDFS methods • Comparison with coupled-system interface force
• Comparison with quasi-static limit load
7
Selection of C2 Parameter (Cont'd) • Comparison with limit load factor or MAC • Peak CG acceleration obtained from multiplying rms of notched force PSD with peak factor & dividing by M • Peak CG acceleration compared with and limited to limit load factor or MAC [3, 4] • Provide an upper bound for force limits and thus C2 • However, it is important to keep in mind the warning of Ref. [1] regarding the possible inappropriate use of force limiting to compensate for an acceleration specification that may be too high
• Peak factor traditionally selected as 3 • To avoid flight hardware damage, larger peak factors (up to 5) have been proposed [5] and is recommended (especially for brittle structures) [1] • Larger peak factor results in lower force limits and C2 for same limit load factor
8
Vibration Test Sequence The Sensor Head Protoflight Model (PFM) was subjected to the following test campaigns: • Low-level flat random characterization testing - 4 August 2015 • •
•
0.5 grms applied to SH in Z axis In-situ measurement of scanning mirror frequency response via two single axis accelerometers (elevation & normal axes) provided frequency and transmissibility data for design of manual input notches Additional instrumentation applied internally and externally
• Vibration Test Campaign - 17 to 19 August 2015 • • •
9
FLV + manual notching random runs along 3 axes C2 = 4 as agreed with GSFC prior test campaign Test level and sequence: • Reduced PFM level with applied manual notching along Z axis • Modified workmanship level with manual notching along X and Y axes
Vibration Test Campaign
10
Test Configuration and Instrumentation • Random vibration testing performed in all 3 axes • •
Vertical Z axis first Lateral X and Y axes next
• 11 triaxial force sensors measuring I/F force •
Sensors rotated and flipped over to allow easy access to connectors
• 2 triaxial control accelerometers
• 2 triaxial monitor accelerometers measuring response at key locations
11
FLV Parameters Used for Sensor Head • C2 = 4 for all three axes as recommended by GSFC • In line with GSFC experience that C2 of 4 appropriate with instruments similar to the Sensor Head • In line with CSA experience and R&D activities • Position reasonable since quite low input specification and additional manual (pre-defined) notching
• ‘n’ constant derived from curve-fitting of apparent mass at low-level runs • Minimal values defined (0 or 0.25) • Not enough peaks beyond fb for defining larger slopes • Imply conservative (higher) force limits beyond fundamental mode
12
0 dB Force & Acceleration PSD’s for Z Axis •
Input spec modified from original specification to include manual pre-defined notching
•
Two pre-defined input notches around 875 & 1175 Hz at scanning mirror resonant frequencies (for all three axes)
•
Force limits computed with input spec without pre-defined notches (for all three axes)
•
20 % reduction of rms force from FLV (based on -6 dB run)
•
Maximum FLV notch of 9 dB at 616 Hz
•
FLV notches right on top of pre-defined input notches (i.e. aligned well with mirror resonant frequencies)
13
-9 dB @ 616
-6.5 dB @ 1104
-2.5 dB @ 1196
0 dB Force & Acceleration PSD’s for X Axis • Modified GEVS workmanship input specification to include manual notches • 17 % reduction of rms force from FLV (based on -6 dB run)
• Maximum FLV notch of 5 dB at 608 Hz • FLV notches away from pre-defined input notches
-5 dB @ 608
14
0 dB Force & Acceleration PSD’s for Y Axis • Modified GEVS workmanship input specification to include manual notches
• Axis with most significant reduction of rms force from FLV notching • 36 % reduction of rms force from FLV (based on -6 dB run) • Maximum FLV notch of 10 dB at 556 Hz
• Large FLV notch right on top of second pre-defined input notch (corresponds well with characterized mirror response)
-10 dB @ 556
-5.5 dB @ 620
15
-9 dB @ 1164
Sensitivity Study Comparison between peak CG acceleration and MAC
16
Introductory Remarks to Study • Study based on data from vibration test campaign • Study performed after completion of the tests • The context of this test campaign implies minimal conservatism in acceleration specification beyond the amount required to ensure no hardware nor alignment issues will be caused by launch vibration: • Vibration-sensitive hardware • Very tight schedule for OLA delivery to spacecraft provider
This led to • Probably lowest acceptable acceleration spec in main axis of test (Z direction) • Minimum workmanship level for the other two axes
• This is an ideal situation for investigating the C2 selection criterion of comparison of peak CG acceleration with limit load (such as MAC), without the risk of excessive notching caused by compensating for overly conservative acceleration spec as mentioned in the NASA Handbook [1] 17
Introductory Remarks to Study (Cont’d) • Data from -6 dB runs used simulate real-life situation • Highest run before going to full-level testing • In cases where value of C2 finalized from low-level runs, peak CG acceleration could influence C2 and thus force limits and input notching
• Study investigated 2 different parameters • Peak factor (from rms to peak force values to get peak CG acceleration) • Primary investigation • Range from 3 (traditional) to 5 (recommended for brittle hardware) • Irrelevant for OLA since C2 defined and fixed prior testing
• Maximum frequency of force PSD used to compute rms force • Two cases investigated • Upper frequency of random vibration testing : 2000 Hz • First anti-resonance (since some analysts and organizations limit the PSD just past the fundamental mode to get the rms value) However, not proper approach according to [6], since this does not yield the true rms value nor limit the true peak CG acceleration 18
Vertical Z axis Factored MAC value for OLA SH: MACfact = 39 g Peak factor
Freq. bandwidth 20 to 2000 Hz rms force (lbf)
Freq. bandwidth 20 to 652 Hz
Peak CG rms force acc (g) (lbf)
Peak CG acc (g)
3
290
29
233
23
4
290
39
233
31
5
290
48
233
39
• For reduced frequency bandwidth • No peak CG acc exceeds MACfact
• For full frequency bandwidth • Peak factors of 3 and 4 result in peak CG acc not exceeding MACfact • If a peak factor of 5 was used, C2 not pre-set and already pretty low (4) and the input spec not already quite low for the SH, then a peak CG acc of 48 might have triggered discussion on test parameters (either input spec and/or C2 value) 19
Lateral X axis Factored MAC value for OLA SH: MACfact = 39 g Peak factor
Freq. bandwidth 20 to 2000 Hz
Freq. bandwidth 20 to 824 Hz
rms force (lbf)
Peak CG acc (g)
rms force Peak CG (lbf) acc (g)
3
264
26
230
23
4
264
35
230
31
5
264
44
230
38
• For reduced frequency bandwidth • All peak CG acc are lower than MACfact
• For full frequency bandwidth • Peak factors of 3 and 4 result in peak CG acc not exceeding MACfact • If a peak factor of 5 was used, C2 not pre-set and already pretty low (4), then a peak CG acc of 44 might have not triggered any discussion on test parameters since the input spec is already at its lowest level (workmanship) and the MACfact is basically halfway between between 35 and 44 20
Lateral Y axis Factored MAC value for OLA SH: MACfact = 39 g Peak factor
Freq. bandwidth 20 to 2000 Hz
Freq. bandwidth 20 to 824 Hz
rms force (lbf)
Peak CG acc (g)
rms force (lbf)
Peak CG acc (g)
3
254
25
230
23
4
254
34
230
31
5
254
42
230
38
• For reduced frequency bandwidth • All peak CG acc are lower than MACfact
• For full frequency bandwidth • Peak factors of 3 and 4 result in peak CG acc not exceeding MACfact • If a peak factor of 5 was used, C2 not pre-set and already pretty low (4), than a peak CG acc of 42 might have not triggered any discussion on test parameters since the input spec is already at its lowest level (workmanship) and the peak CG acc is marginally above the MACfact (i.e. less than 10 % above) 21
Concluding Remarks •
The combination of FLV, in-situ characterization of critical components, manually designed notches and specific tailored design changes based on EM test results arrived at a solution for OLA to meet the OSIRIS-REx launch environment
•
Since the OLA Sensor Head was subjected to quite low acceleration specifications and since force limits have been derived using a reasonable low C2 value of 4, a sensitivity study comparing peak CG acceleration with the MAC value was performed and shows that (using the full frequency bandwidth): •
•
22
Peak CG accelerations (with peak factors of 4 and 5) are fairly close to the factored MAC and could have thus been used to detect a C2 value that would have been set too high The comparison supports the warning of the NASA Handbook on FLV regarding the possible inappropriate use of force limiting to compensate for an acceleration specification that may be too high
References 1. Anon. 'Force Limited Vibration Testing', NASA-HDBK-7004C, Nov 2012 2. Soucy, Y. 'On Force Limited Vibration for Testing Space Hardware', Proc. of IMAC XXIX Conference, Jacksonville, Florida, Feb. 2011 3. Scharton, T.D. 'Force Limits Measured on a Space Shuttle Flight', Journal of the IEST, Vol. 45, pp. 144-148, 2002 4. Chang, K.Y. 'Deep Space 1 Spacecraft Vibration Qualification Testing', Sound and Vibration, March 2001 5. Scharton, T.D, Pankow, D. and Sholl, M. 'Extreme Peaks in Random Vibration Testing', The 2006 S/C and L/V Dynamic Environments Workshop, Hawthorne, CA, June 2006 6. Dunford, J. and Scharton, T.D. ‘Development of the Random Vibration Test Program for the ICON Instruments', The 2015 S/C and L/V Dynamic Environments Workshop, El Segundo, CA, June 2015 23
