River Rapids Conference Proceedings of the Society for Experimental Mechanics Series Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry, Volume 8 Dario Di Maio Paolo Castellini Proceedings of the 35th IMAC, A Conference and Exposition on Structural Dynamics 2017 River Publishers
Conference Proceedings of the Society for Experimental Mechanics Series Series Editor Kristin B. Zimmerman, Ph.D. Society for Experimental Mechanics, Inc., Bethel, CT, USA
River Publishers Dario Di Maio • Paolo Castellini Editors Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry, Volume 8 Proceedings of the 35th IMAC, A Conference and Exposition on Structural Dynamics 2017
Published, sold and distributed by: River Publishers Broagervej 10 9260 Gistrup Denmark www.riverpublishers.com ISBN 978-87-7004-953-5 (eBook) Conference Proceedings of the Society for Experimental Mechanics An imprint of River Publishers © The Society for Experimental Mechanics, Inc. 2017 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, or reproduction in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Preface Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry represent one of ten volumes of technical papers presented at the 35th IMAC, A Conference and Exposition on Structural Dynamics, organized by the Society for Experimental Mechanics, and held in Garden Grove, California, January 30–February 2, 2017. The full proceedings also include volumes on Nonlinear Dynamics; Dynamics of Civil Structures; Model Validation and Uncertainty Quantification; Dynamics of Coupled Structures; Sensors and Instrumentation; Special Topics in Structural Dynamics; Structural Health Monitoring & Damage Detection; Shock & Vibration, Aircraft/Aerospace and Energy Harvesting, and Topics in Modal Analysis & Testing. Each collection presents early findings from experimental and computational investigations on an important area within Structural Dynamics. Topics represent papers on enabling technologies such as: rotating machinery, vibro-acoustics and laser Vibrometry, advances in wind energy and hybrid testing methods. The organizers would like to thank the authors, presenters, session organizers, and session chairs for their participation in this track. Bristol, UK Dario Di Maio Ancona, Italy Paolo Castellini v
Contents 1 Strategies for Testing Large Aerospace Structures with 3D SLDV................................................ 1 Daniel P. Rohe 2 Modal Model Validation Using 3D SLDV, Geometry Scanning and FEM of a Multi-Purpose Drone Propeller Blade .................................................................................................. 13 Daniel J. Alarcón, Karthik Raja Sampathkumar, Kamenzky Paeschke, Tarun Teja Mallareddy, Sven Angermann, Andreas Frahm, Wolfgang Rüther-Kindel, and Peter Blaschke 3 Effect of Dry Friction Damping on the Dynamic Response of Helicopter Tail Shaft ............................ 23 Onur Ozaydin and Ender Cigeroglu 4 Nonlinear Dynamic Analysis of a Spiral Bevel Geared System.................................................... 31 Siar Deniz Yavuz, Zihni Burcay Saribay, and Ender Cigeroglu 5 Estimating Material Wavespeed Using the Wavenumber Transform of Rectangular Plate Mode Shapes....................................................................................................... 41 Micah R. Shepherd 6 In-Operation Wind Turbine Modal Analysis via LPV-VAR Modeling ............................................ 47 L.D. Avendaño-Valencia, E.N. Chatzi, and S.D. Fassois 7 Structural Damage Identification Using Free Response Measured by a Continuously Scanning Laser Doppler Vibrometer System.................................................................................... 59 Y.F. Xu, Da-Ming Chen, and W.D. Zhu 8 Mitigation of Structural-Acoustic Mode Coupling in a Modal Test of a Hollow Structure ..................... 71 Ryan Schultz and Ben Pacini 9 Applications of 3D Scanning Laser Doppler Vibrometry to an Article with Internal Features ................ 85 Bryan Witt, Brandon Zwink, and Ron Hopkins 10 The Measurement of a Nonlinear Resonant Decay Using Continuous-Scan Laser Doppler Vibrometry ............................................................................................................... 97 David A. Ehrhardt, Matthew S. Allen, and Timothy J. Beberniss 11 Vibro-Acoustic Modulation of a Spinning Apparatus for Nondestructive Evaluation........................... 105 Martin J. Ward, Wesley E. Scott, Nicholas M. Diskerud, Alessandro Cattaneo, John D. Heit, and John D. Bernardin 12 Nonlinear Phase Separation Testing of an Experimental Wing-Engine Structure............................... 115 L. Renson, J.P. Noël, D.A.W. Barton, S.A. Neild, and G. Kerschen 13 Wind Turbine Health Monitoring: Current and Future Trends with an Active Learning Twist ............... 119 N. Dervilis, A.E. Maguire, E. Papatheou, and K. Worden vii
viii Contents 14 Nonlinear 3D Dynamic Model of an Automotive Dual Mass Flywheel ............................................ 131 G. Quattromani, A. Palermo, F. Pulvirenti, E. Sabbioni, and F. Cheli 15 Investigation of Notch-Type Damage Identification by Using a Continuously Scanning Laser Doppler Vibrometer System............................................................................................ 143 Da-Ming Chen, Y.F. Xu, and W.D. Zhu
Chapter 1 Strategies for Testing Large Aerospace Structures with 3D SLDV Daniel P. Rohe Abstract The 3D Scanning Laser Doppler Vibrometer (3D SLDV) has the ability to scan a large number of points with high accuracy compared to traditional roving hammer or accelerometer tests. The 3D SLDV has disadvantages, however, in that it requires line-of-sight from three scanning laser heads to the point being measured. This means that multiple scans can become necessary to measure large or complex parts, and internal components cannot typically be measured. In the past, large aerospace structures tested at Sandia National Laboratories typically have used a handful of accelerometer stations and instrumented internal components to characterize these test articles. This work describes two case studies that explore the advantages and difficulties in using a 3D SLDV to measure the same test articles with a much higher resolution scan of the exterior. This work proposes strategies for combining a large number of accelerometer channels with a high resolution laser scan. It explores the use of mirrors and laser head mounts to enable efficient re-alignment of the lasers with the test article when many scans are necessary, and it discusses the difficulties and pitfalls inherent with performing such a test. Keywords 3D • SLDV • Aerospace • Modal 1.1 Introduction Scanning Laser Doppler Vibrometry has been shown to have advantages over traditional mounted sensors [1]. For example, it allows for non-contact measurements which do not mass-load the test article. Additionally, a finer measurement point resolution can often be achieved through the use of scanning mirrors in the laser heads which are much more precise than a roving hammer or roving accelerometer test. It can also be quicker to set up a scan of, for example, 100 measurement points than to mount 100 accelerometers on the test article or perform 100 roving hammer impacts. This last result suggests that the 3D SLDV may be well-suited for performing tests with large numbers of measurement points, for example a large aerospace structure. There are, however, some disadvantages to SLDV that might impede such a test. 3D SLDV requires line-of-sight from each of the three laser heads to the measurement point of interest. This often precludes the measurement of the motions of internal components of the test article, which can often times be the measurements that designers are most interested in. Additionally, scans may need to be performed from multiple directions to measure all sides of the test article, and these extra alignment steps may significantly increase the time required to perform a test. This work investigates the effectiveness of using the 3D SLDV system to perform a high-resolution scan of the exterior of a large aerospace structure using two case studies. It discusses strategies for incorporating a large number of accelerometer channels, as well as using mirrors and laser head mounts to enable efficient re-alignment of the lasers with the test article when many scans are necessary. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. D.P. Rohe ( ) Structural Dynamics Department, Sandia National Laboratories, P.O. Box 5800 - MS0557, Albuquerque, NM, 87185, USA e-mail: dprohe@sandia.gov © The Society for Experimental Mechanics, Inc. 2017 D. Di Maio, P. Castellini (eds.), Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry, Volume 8, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-54648-3_1 1
2 D.P. Rohe 1.2 Challenges Associated with Large Aerospace Test Articles Aerospace test articles are generally large, complex structures that have high programmatic value. As a result of their size, a large number of measurement points may be necessary to achieve a fine resolution scan of the exterior of the structure. Due to the sequential nature of the SLDV measurement, this can result in long test times, especially considering that the laser may need to be repeatedly aligned with the structure to scan different sections of the test article. Therefore, time saving strategies should be implemented whenever possible. For example, performing a multiple-input test with three shakers attached would reduce the amount of measurement time by a factor of three compared to performing three single-input tests sequentially. Also, because of the potentially large number of alignments, it is important to develop an efficient way of repositioning the laser heads, because the alignment procedure is the most labor intensive portion of the SLDV test. Additionally, the SLDV measurement may need to be supplemented with a large number of accelerometer channels measuring the internal components of the aerospace structure. Because large aerospace structures are often very valuable and costly to replace, it is standard procedure to minimize damage to the test article in the event that the test boundary conditions (e.g. bungee cords) should fail. This typically involves not lifting the test article higher than necessary from its supported position at rest. A requirement such as this can significantly limit the placement of the laser heads as they try to scan the bottom of the test article; there may not be enough room underneath the test article, so an alternative strategy must be devised, potentially using a different safety catch. Even for a top or side scan of a test article, there may not be adequate space to place the laser heads when considering the placement of shakers and the footprints of the shaker stands or laser tripods. Setting up for such a scan may not be trivial. Bungee cords, shakers, and accelerometer cabling may all occlude the path of the laser beams resulting in portions of the test article that are unmeasurable. 1.3 Case Study Tests Two test articles were used as case studies for performing tests on large aerospace structures. The first test article, shown schematically in Fig. 1.1, was conical with fewer external features and contained 58 channels of internal accelerometers. This test article was used to refine the approach before moving to the second, larger test article. The second test article is shown schematically in Fig. 1.2. 1.3.1 Conical Structure Modal Test Because shaker testing is more often used on large aerospace structures, shaker testing was investigated in this work. To excite all of the modes in the bandwidth of interest, three shakers were attached simultaneously to the aft end of the test article in the three principal directions. Initially MB-50 shakers were used, but they occluded much of the aft end of the test article, and the stands required to support these shakers had a very large footprint which would interfere significantly with the laser head tripods. Instead smaller, lighter Labworks MT-161 shakers were used, and these could be supported by Fig. 1.1 Conical test article
1 Strategies for Testing Large Aerospace Structures with 3D SLDV 3 Fig. 1.2 Empty bomb case test article Fig. 1.3 Shaker support strategies using MB-50 (left) and MT-161 (right) shakers a single frame made from 95 mm optical rail. As one can observe in Fig. 1.3, the smaller shakers supported on the optical rail system provided much better line-of-sight for the SLDV system. Bungee cords were used to support the test article, and retro-reflective tape was applied to the surface at each measurement point. 1.3.1.1 Data Acquisition Setup Because the 3D SLDV data acquisition system (DAS) only supported eight additional measurement channels, the 58 channels of internal accelerometer measurements had to be measured by a separate DAS. A 64-channel VXI DAS was used for this purpose. The signals from the three drive point accelerometers and three force gauges were split to both systems so each could compute frequency response functions (FRFs). Similar data acquisition parameters were used between the two DASs. Current to power all of the integrated electronic piezoelectric (IEPE) accelerometers was supplied by the VXI DAS, so this output was disabled in the SLDV DAS. Random white noise was applied to the shakers from the 3D SLDV sources. This eased the triggering and synchronization requirements between the two data acquisition systems. The SLDV system was set to measure 30 averages, but due to the
4 D.P. Rohe sequential nature of that measurement, the accelerometer DAS could be set to measure many times that number of averages. 200 averages were selected for the accelerometer data acquisition system, which corresponded to measuring approximately seven points using the 3D SLDV. To acquire data, the VXI was first started in preview mode so the gauges would be powered. Then, the SLDV data acquisition process was initiated, which started the shakers. During the SLDV measurement, the VXI would be set to acquire so those data would be measured. After 200 averages, the VXI would stop recording, and some time later (depending on the number of measurement points being scanned) the SLDV system would stop. 1.3.1.2 Laser Alignment Perhaps the most difficult aspect of the test was the laser alignment. Though the conical test article was not an incredibly costly or programmatically important piece of hardware, throughout this work it was treated as such. It was therefore kept fairly close (about 15 cm) to the chocks from which it was lifted. The chocks were approximately 30 cm (1 ft) tall, which left approximately 45 cm (1.5 ft) between the ground and the test article. To scan the bottom of the test article, it would be necessary to position each of the laser heads such that it could see the bottom of the test article and still maintain the physical separation necessary to get a good 3D measurement. Initially it was thought that the conical test article could be scanned in four quadrants; however, it soon became apparent that six scans would be necessary due to the limited distance between the ground and the test article and the limited field of view of the laser heads. Similarly, it was found that the bungee cords blocked a portion of the test article, which required splitting the top scan into two portions. Since the bottom scan was judged to be the most difficult, it was attempted first. It became obvious that mirrors would be necessary, so a variety of laser head configurations were attempted (see Fig. 1.4). By using mirrors, one can effectively increase the distance between the test article and the laser head. This is advantageous because it can increase the field of view of the lasers. One limitation was that the mirrors provided in the mirror kit marketed by the SLDV vendor were quite small. This required the laser heads to be very close to the mirrors in order to maintain their field of view, which required the laser heads to be very close to the ground and prohibited the use of the standard tripods to support the laser heads. Laser heads were instead mounted on optical rails which could be set directly on the ground. Aligning the lasers to the test article required a two-step procedure for each scan. Due to the conical nature of the part, there were not any sharp features on the test article that could be used for alignment. Instead, when space was available, the laser heads were aligned using a separate reference object with known 3D coordinate locations. If there was no space for the reference object, an automatic 3D alignment procedure was used where the operator positions all three laser spots at points defining an origin, an axis, and a plane and the laser system computes the coordinate system from the mirror deflections at each point and the distance to each point from the top laser. This second approach was not as accurate as using the reference object, but it was necessary for some of the tighter setups. Both alignment procedures aligned the lasers to an arbitrary coordinate system, so some coordinate transformation was required to bring the laser heads into the part coordinate system and allow the measurements to be stitched together. Sixteen alignment points were drawn on the test article in two axial stations. The first station was approximately 4 cm (1.6 in) aft of the joint in the body, and the second station at the aft end of the test article. These points were drawn before the test article was set up, and therefore some oversight was made in their positioning. The aft alignment points were blocked by the shakers and the chocks, and were therefore only visible from one of the scan directions. The forward alignment points were sufficient for the alignment, but because they were all positioned in the same plane, small errors in their positioning would result in relatively large rotational alignment errors. This can be seen in a plot of the computed measurement points, shown in Fig. 1.4 Various attempts to measure the bottom quadrant of the conical test article
1 Strategies for Testing Large Aerospace Structures with 3D SLDV 5 Fig. 1.5 Measurement point positions computed by the SLDV software. Points on the scan boundaries should be nominally coincident, but are seen to be rotated due to sensitivity of the geometry to alignment errors Fig. 1.5. Points on the boundaries of each of the scan areas (differentiated by color in Fig. 1.5) should nominally be coincident, but because only the front alignment points were used for the majority of the scan positions, the gap between the nominally coincident points grows larger towards the aft end of the structure. After the coordinate transformation, the effective positions of the laser heads (i.e. the positions of the laser heads reflected through the mirrors) could then be extracted from the SLDV software for each scan, and these are shown in Fig. 1.6. 1.3.1.3 Data Analysis In this test, data came from two separate DASs: SLDV and VXI. Additionally, the laser data consisted of six separate scans. All of this data had to be combined into a suitable format for modal analysis using the chosen Synthesize Modes and Correlate (SMAC) [2] algorithm. The laser data were combined using the SLDV software. A combined file was created from all of the separate scan files and exported to MATLAB for analysis. The data from all internal gauges were measured each time a laser scan took place. Since the same channels were measured each time, the data should be nominally identical. This observation was used to gauge data consistency between laser scans; because a large amount of time might elapse between laser scans due to alignment difficulties, the bungee cord boundary conditions could sag and the shaker stingers could potentially be cross-loaded causing shifts in the natural frequencies. This was seen early in the testing before the bungee cords had significantly relaxed, and is shown in Fig. 1.7. After the bungee sag issue was discovered, the stingers were re-aligned after each scan to ensure consistent data. The internal data from each of the separate scans were then averaged into one set of FRFs and combined with the laser data to perform mode fitting. Fifty-three modes were extracted below 4000 Hz, many of which were ovaling modes of the case. However, these could be easily identified due to the relatively high resolution of measurement points on the case. The MAC matrix for this test was very good, with only two off-diagonal entries greater than 0.3, and the vast majority of the off-diagonals below 0.1. 1.3.1.4 Lessons Learned from the Conical Test Article The empty bomb case is significantly longer than the conical structure, so portions of the test that were tedious or laborintensive would be even more so when applied to that structure. The testing strategies applied to the conical article were therefore evaluated before the empty bomb case was tested. The optical rail shaker supports provided a very small footprint and did not significantly block line-of-sight to the test article, so this strategy was thought to be adequate for the empty bomb case. In the conical test, the 3D SLDV system was able to scan about 90ı of the test article at a time. This will be the goal of the empty bomb case test as well. However, there were some issues with the choice of quadrants used for the conical test article. For that test, the quadrants were chosen primarily due to the difficulty in setting up the bottom scan as described in Sect. 1.3.1.2. This required the alignment of three mirrors, one for each laser head, to enable each laser to scan as much
6 D.P. Rohe Fig. 1.6 Laser head positions for the various scans of the conical test article. The transparent laser heads in the figure show the effective position after being reflected across the mirror of the bottom side of the test article as possible. It was originally thought for the conical test article that performing just one “bottom” scan would be the best approach. However, it was found as testing proceeded that bungee cords, dangling accelerometer cables, and shakers were located directly in the middle of the quadrants, and it was difficult to scan around these obstructions. An alternative approach would be to shift the measured quadrants by 45ı, positioning the bungee cords, dangling accelerometer cables, and shakers at the boundaries of the quadrants rather than in the middle. These approaches were investigated during the conical test (the first two frames of Fig. 1.4 show two such setups). The disadvantage to this approach (and the reason it was not used during the conical structure test) is that the test setup still required aligning three mirrors, one for each laser head, only now there were twice as many “bottom” quadrants to scan, so the effort for this approach is effectively doubled. Therefore, a strategy for efficiently orienting the laser heads and mirrors would be necessary in order to make this approach worthwhile.
1 Strategies for Testing Large Aerospace Structures with 3D SLDV 7 Fig. 1.7 Shifting peaks in the drive point FRFs between two scans Fig. 1.8 Photograph (left) and schematic (right) of the laser cart used to scan the bottom quadrants of the empty bomb case. The schematic shows the effective position of the mirrored laser head Performing the alignment, especially through mirrors, can be a labor-intensive part of any 3D SLDV measurement. However, if the test article moves with respect to the laser heads and the effective positions of the laser heads remain constant with respect to one another, then the laser head positions can be updated in the software using a simple coordinate transformation. This can save a significant amount of time when a large number of scans are needed. The goal then is to devise a system where the laser heads and any mirrors can be moved and positioned as one rigid body. The SLDV vendor markets a wheeled stand that holds all three laser heads just for this purpose, but this stand is quite large and does not incorporate mirrors. Instead, a cart was made from 95 mm optical rails. Several design iterations were evaluated: the key flaws discovered in early designs were a cart that was too big so it interfered with shaker supports as well as a cart with four “feet” so an uneven floor would warp the cart and misalign the lasers. The final design is shown in Fig. 1.8. The laser cart was immensely useful for measuring the bottom of the empty bomb case; a single 3D alignment could be performed, then all the bottom sections could be measured. Only a coordinate system transformation was required between each scan. The coordinate system transformation requires alignment points at which the (x,y,z) coordinates were known, and since the conical structure test revealed that an insufficient number of these alignment points could lead to significant errors in geometry, a grid of alignment points was drawn every 15 cm (6 in) and 45ı. This ensured that approximately 20 alignment points would be visible for each scan. Finally, it was recognized that there was not an efficient way to produce documentation for large laser scans. This documentation should provide information about the alignment accuracy, measurement point locations, and other data acquisition parameters that would be useful in evaluating a test, and collecting all of this data manually would be very tedious, especially for a large number of scans. Thankfully, the SLDV software provides a programming interface from
8 D.P. Rohe Fig. 1.9 Empty bomb case which much of this information can be extracted, so a script was written to traverse the output from the laser scan and produce a report that contained data, images, and test information. 1.3.2 Empty Bomb Case Modal Test With the improvements from the conical structure test implemented, the empty bomb case testing was initiated. The test article was suspended from an overhead crane using bungee cords to provide a soft support approximating a free boundary condition. The test article is shown in Fig. 1.9. Because the laser cart required more space below the test article, the existing chocks could not be used as a safety catch, though they were kept below the test article regardless. Instead, two slings were attached loosely to the test article to serve as a safety catch in case the bungee cords were to fail. Due to the large number of measurement points anticipated in this test, no surface preparation was applied to the test article. 1.3.2.1 Data Acquisition Setup The setup for the SLDV DAS was similar to that of the conical structure modal test. One important distinction is that the frequency content of the white noise signal that was applied to the shakers was cut off below 10 Hz. The low frequency content below 10 Hz induced rigid body motion in the structure, and while this motion was small it significantly increased the amount of noise in the laser measurements. 1.3.2.2 Laser Alignment For this test, 17 scans were used to measure the entire test article. Each scan was able to view approximately 60 cm (2 ft) axially and 90ı circumferentially of the surface of the test article. The surface of the structure visible to the laser vibrometer in each of the scans is shown in Fig. 1.10. The effective laser head positions for all of the scans are shown in Fig. 1.11. For the bottom scans, the lasers were mounted in the cart shown in Fig. 1.8, and an alignment was performed using a reference object. The laser cart was then moved into position beside the test article, and measurement points were placed at
1 Strategies for Testing Large Aerospace Structures with 3D SLDV 9 Fig. 1.10 Portion of the test article visible for each scan Fig. 1.11 All effective laser head positions for measuring the empty bomb case
10 D.P. Rohe Fig. 1.12 Geometry showing the 1075 measurement point locations colored per scan. White elements denote the space between scan patches 5–10 alignment points. A coordinate system transformation was then computed by the SLDV software using the locations of the alignment points in the current coordinate system and the locations in the part coordinate system. This process allowed efficient re-alignment of the laser heads and mirrors with the test article for each scan. For the top scans, the lasers were arranged on tripods; therefore, the two step alignment-and-transform procedure described in Sect. 1.3.1.2 was followed for those scans. In hindsight it would have been prudent to build a second stand from the 95-mm optical rails in order to bypass the alignment portion as was done for the bottom scans. A large number of measurement points were to be scanned for this test, so the most efficient way to define them in the laser software was to import the measurement points from an external geometry file. These points then needed to be reduced for each scan to those points within the lasers’ field of view and on the side facing the laser heads. Points were also removed if the angle of incidence between any of the laser beams and the surface were too high. Figure 1.12 shows the measurement geometry color-coded by scan. Even with the efficient bottom alignment strategy, this test was long running due to the sequential scanning of over 1000 points as well as labor-intensive setting up and aligning the lasers for all of the scans. Testing was performed over 5 days, so it would be very difficult to go back and retake data if some data acquisition parameter needed to be changed (e.g. forgetting to apply a window, bandwidth too low, etc.). This is contrary to testing with a large number of accelerometers where once the gauges are adhered to the surface data can be taken and retaken rather easily if there are enough channels available. 1.3.2.3 Data Analysis Similarly to the conical structure test, the laser data was collected into a single scan file and exported to analyze in MATLAB. Over 6000 FRFs were measured by the laser system, accounting for the over 1000 scan points, two shaker inputs, and three measurement directions per point. It became clear when performing this analysis that the SMAC algorithm began to struggle with a data set of this size. When all 6000CFRFs were included in the SMAC analysis, the correlation coefficient—a key parameter indicating the presence of a mode in the SMAC algorithm—was approximately unity for the entire bandwidth. Essentially, SMAC was identifying a mode at every frequency line. When the set of FRFs was reduced by only keeping one of every 5 or 25 measurement points, the correlation coefficient began to look more reasonable, as shown in Fig. 1.13. The frequency and damping parameters extracted from the reduced analysis could then be imported into a full analysis to extract the full mode shapes. Extracted mode shapes were somewhat noisy which is attributed to the large angle of incidence and relatively poor surface properties of the test article, especially on the darker surface near the tip of the nose. Two example mode shapes are shown in Fig. 1.14. 1.4 Lessons Learned and Lingering Deficiencies A number of lessons were learned from the testing of the two test articles described in this report. Large, complex test articles will often require multiple scans to measure all parts of the test article, so developing an efficient method of repositioning the laser head is likely the most important step to making SLDV measurements of large structures practical. Because multiple
1 Strategies for Testing Large Aerospace Structures with 3D SLDV 11 Fig. 1.13 Complex Mode Indicator Function (CMIF) and the SMAC correlation coefficient based on number of points kept in the analysis scans will need to be stitched together into one coherent model, a sufficient number of alignment points need to be available; in many cases it may be advisable to define and mark these as a preliminary step in test setup. Simultaneous excitation from multiple input locations is also an excellent way to decrease the amount of time spent scanning a test article. Utilizing mirrors proved to be an excellent way to allow the laser heads to be positioned farther from the test article. However, because the test articles in this effort were quite large, the mirrors marketed by the SLDV vendor were found to be too small to gain the full benefit; the field of view is artificially limited unless the laser heads were quite close to the mirrors. This reduced the measurable surface to approximately 60 cm (2 ft) along the length of the test article and drove the number of scans required. Since this work was performed, two larger first-surface mirrors have been acquired which should allow larger scan areas. The laser cart allowed the laser heads and mirrors to be moved as one rigid body, and this greatly simplified the alignment procedure for this test. However, even with the refined alignment procedure, this test still required a large amount of time to perform. Because of the sequential nature of the SLDV it is imperative when performing such a test that all test setup and data acquisition parameters be correct the first time, because repeating such a test due to some simple oversight in test setup (for example if the shaker stinger were misaligned) would be particularly painful. Accelerometer testing can be labor intensive when a large number of gauges must be adhered to the test article, but that labor is not wasted if the test needs to be repeated for any reason. With SLDV testing, the labor of aligning the laser heads multiple times would need to be repeated if the test needed to be redone. Surface preparation using retroreflective tape can be used to increase the signal returning to the laser head, which can reduce the measurement noise. For the conical structure, retroreflective tape was applied at all measurement points. However, for the 1000Cmeasurement points on the empty bomb case, it was thought that the application of that much retroreflective tape would require a prohibitive amount of time. Spray-on retroreflective beads were attempted as an easierto-apply alternative to tape, but this technique was not met with much success. The beads would drip down the test article before the solution could dry, leaving clumps of beads in some locations with very few beads over the majority of the area. Many curve fitters and modal software were designed to handle a “reasonable” number of FRFs. As techniques such as DIC and SLDV become more standard in the modal testing community, it may be necessary to revisit these tools to make sure they are sufficiently capable in dealing with the large amount of data that can arise from these techniques. The SMAC algorithm, for example, struggled to identify modes when too many FRFs were used. This was easy enough to work around by reducing the number of functions supplied to the curve fitter, extracting the natural frequencies and damping ratios extracted from that analysis, and importing those modal parameters into an analysis of all of the FRFs.
12 D.P. Rohe Fig. 1.14 Example mode shapes of the empty bomb case. Some points are noisy, particularly near the tip of the nose where the angles of incidence of the laser beams to the part were the highest and where the surface was darkest 1.5 Conclusions Two case studies were performed to determine the feasibility of using a 3D SDLV to measure a large aerospace structure with fine measurement point resolution. The alignment procedure to measure a large number of points on a long cylindrical test article was very labor-intensive. If there is no explicit reason to measure a large number of measurement points for the test, it is feasible that more traditional methods using accelerometers could be more appropriate. However, if a large number of data points are required, it may be difficult to acquire a data acquisition system that could measure that many accelerometer channels simultaneously. In such scenarios, SLDV may be an attractive option. Compared to roving hammer testing, the 3D SLDV system is quicker, more precise, and perhaps most importantly it can additionally measure directions tangent to the surface rather than just perpendicular. References 1. Castellini, P., Martarelli, M., Tomasini, E.P.: Laser Doppler vibrometry: development of advanced solutions answering to technology’s needs. Mech. Syst. Sig. Process. 20(6), 1265–1285 (2006) 2. Hensley, D.P., Mayes, R.L.: Extending SMAC to Multiple References. In: Proceedings of the 24th International Modal Analysis Conference, pp. 220–230 (February 2006)
Chapter 2 Modal Model Validation Using 3D SLDV, Geometry Scanning and FEM of a Multi-Purpose Drone Propeller Blade Daniel J. Alarcón, Karthik Raja Sampathkumar, Kamenzky Paeschke, Tarun Teja Mallareddy, Sven Angermann, Andreas Frahm, Wolfgang Rüther-Kindel, and Peter Blaschke Abstract The ATISS (Autonomous Flying Testbed for Integrated Sensor Systems) is a measurement unmanned aircraft vehicle—commonly known simply as “drone”. The ATISS is a multi-purpose sensor carrier drone used for professional aerial photography, aerial surveying, air pollution measurements, etc., propelled by two carbon fiber propellers. The kinetic energy of these propellers can excite undesired vibration modes on these blades and excite the drone structure unless a careful design is chosen. Generally, a detailed modelling of the vibrational characteristics is needed for any aerospace component; for the successful correlation and validation of these models through experimental vibration testing. However, lightweight, innovative composite components such as carbon fiber propellers pose challenges on this validation cycle. Material non-linearities and non-proportional damping responses are inherent to composite materials and therefore, a further degree of precision on the testing and simulation is necessary for the successful validation of these components. The experimental modal analysis is performed by the combination of a 3D Scanning Laser Doppler Vibrometer and a scalable automatic modal hammer. The geometry used for the FE simulation is obtained by means of a 3D geometry scanner, a high resolution-digitalization is carried out on the tested rotor blade surface. The obtained points are reverse-engineered in a CAD model and imported into the FEA software. A further validation of the FE modal model would prove the successful implementation of these techniques on the study of the vibrational modes of composite lightweight structures. Keywords Experimental modal analysis • 3D SLDV • Composite material • FEA validation • Reverse engineering Abbreviations ATISS Autonomous Flying Testbed for Integrated Sensor Systems CAD Computed-Aided Design DOF Degree(s) of Freedom EMA Experimental Modal Analysis FEA Finite Elements Analysis FRF Frequency Response Function IGES Initial Graphics Exchange Specification LDV Laser Doppler Vibrometry NVH Noise, Vibration, Harshness SAM Scalable Automatic Modal Hammer 2.1 Introduction and Motivation The ATISS (Autonomous Flying Testbed for Integrated Sensor Systems) is an unmanned aircraft vehicle, also known as drone, based on a modular concept; designed, engineered, manufactured and tested at the Technical University for Applied Sciences Wildau (Fig. 2.1). The ATISS is designed as a motor glider and intended as a multi-purpose sensor carrier, which D.J. Alarcón ( ) • K.R. Sampathkumar • K. Paeschke • T.T. Mallareddy • P. Blaschke Laboratory for Machine Dynamics and NVH, Technical University of Applied Sciences Wildau, Hochschulring 1, 15745, Wildau, Germany e-mail: daniel.alarcon@th-wildau.de S. Angermann • A. Frahm • W. Rüther-Kindel Institute for Aeronautical Engineering, Technical University of Applied Sciences Wildau, Hochschulring 1, 15745, Wildau, Germany © The Society for Experimental Mechanics, Inc. 2017 D. Di Maio, P. Castellini (eds.), Rotating Machinery, Hybrid Test Methods, Vibro-Acoustics & Laser Vibrometry, Volume 8, Conference Proceedings of the Society for Experimental Mechanics Series, DOI 10.1007/978-3-319-54648-3_2 13
14 D.J. Alarcón et al. Fig. 2.1 The ATISS drone with the red painted payload container between the rotors. It has a wingspan of about 5 m and an empty weight of 15 kg Fig. 2.2 Manufacturing process of the propeller blade used on this study allows its deployment in a variety of cases: professional aerial photography, aerial surveying, the generation of 3D terrain models and air pollution measurements. In the current project SAPODS (Smart Airborne Pollutants Detection System) the ATISS is intended to fly through volcanic ash clouds in order to measure the concentration and composition of air pollutants [1]. With a wingspan of about 5 m and a basic or empty weight of 15 kg, the ATISS drone is capable of carrying payloads of more than 10 kg. Driven by two brushless electric motors with 2.5 kW power each, two especially designed folding carbon propellers and two Lithium-Polymer batteries, the ATISS can fly for more than 60 min without gliding. The propeller analyzed on this paper is part of a bi-propeller rotor. The other propeller and the joint in between are not analyzed on this study for simplicity reasons. The ATISS propellers were designed to be folded if necessary, improving the gliding properties of the drone. Only one twin-shell negative mold was used for the manufacturing of both propellers. These were manufactured in the negative mold shown in Fig. 2.2 using carbon-fiber and epoxy resin. The propeller blades of the ATISS work therefore at very high rotational speeds for long periods of time. The kinetic energy of this rotation can excite undesired vibration modes on these blades and excite in turn the drone structure unless a careful design is chosen. Consequently, the data gathered by the integrated sensors could be distorted by excessive resonance on the drone structure. One clear example of resonance-induced problems is the wobbling in aerial videography recordings, also known as “jello effect”, a well-known problem among hobbyists and aerial videography aficionados. The resonance of the propellers at certain rotation regimes is transferred to the drone body and to the camera as a result, rendering the recordings unusable [2]. Several home-made techniques have been developed in order to balance the drone propellers. Nevertheless these procedures are not as immediate or as straightforward in a fast-growing large-scale drone industry and in applications where excessive vibration at undesired regimes/frequencies can impact the data recorded by very sensitive transducers and the integrity of the propeller blades. Generally, these vibrational problems, both at small scale (for drones) and large scale (i.e. for gas turbine blades) require for their analysis a detailed Finite Element Analysis (FEA) modelling of their vibrational characteristics. The aerospace and automotive NVH industries have developed and improved cycles of correlation and validation during the last years by updating the parameters of the FEA models with experimental vibrational data usually derived from experimental modal analysis (EMA). This procedure is widely known as “modal updating”, and it is still hindered by several challenges nowadays. For example, the use of lightweight innovative composite materials is constantly growing in the aerospace industry [3] but components made of this kind of materials present difficulties on this correlation and validation cycle [4]. Material non-linearities, anisotropy, non-proportional force/response curves, etc., are inherent to composite materials—such as the composite of carbon fiber and epoxy resin used on the propeller analyzed on this study.
2 Modal Model Validation Using 3D SLDV, Geometry Scanning and FEM of a Multi-Purpose Drone Propeller Blade 15 EMA FEA FE-model FE-calculation model parameters measurement modal parameters matching calculation comparison Experimental Modal Analysis Finite Element Analysis update Fig. 2.3 Continuous correlation cycle for the validation of modal models proposed in [5] A solution for this problem is a continuous correlation cycle between FEA modal models and experimental modal data, proposed already by [5–7] among others (Fig. 2.3). Experimental data can be collected in a variety of means, being experimental modal analysis of wide application in the industry. During the recent years, the world of vibration testing has undergone many changes that allow nowadays a proper data collection that makes this continuous validation process feasible. Non-contacting response measurement techniques, such as Laser Doppler Vibrometry (LDV) are of common use on its single-point or scanning variants. LDV is nowadays the standard method for the acquisition of output signals on EMA due to its versatility, accuracy, reduction of the testing time, and the fact that the tested system is free of mass loading effects. The advantages of this method have been widely described in [8, 9] to name a few. In the case of this paper, this system offers an extra advantage: while only flap (out-of-plane) modes can be investigated with 1-point and SLDV techniques, the edgewise (in-plane), simultaneous vibrational components of these modes can be investigated only through the application of 3D SLDV. These edgewise modes need to be evaluated for a good correlation, as they appear on the FEA modal models. A high degree of adjustability and repeatability on the excitation is needed on the modal testing of non-linear materials, as their mechanical properties depend on non-proportional force/response ratios [10]. Several excitation techniques are applied in the industry, each one with its advantages and drawbacks: semi-automatic modal hammers, electrodynamic/piezoelectric shakers and shaker-based, non-contact magnetic exciters. The automatic alternatives currently available in the market fail to allow a fine-tuned and truly repeatable impacting force adjustment while making no changes in the tested structure. In most cases, the experimental repeatability will be an issue due to the fact that the human factor is not fully eliminated during the hammer positioning and force adjustment or re-adjustment throughout the measurement. Other techniques such as electrodynamic and piezoelectric shakers have a series of drawbacks described in [10], they not specified on this paper for brevity. For this reason, the Scalable Automatic Modal Hammer (SAM) is later introduced. Leaving the experimental techniques aside, computational models can be also a source of error. Models used for the correlation are typically exported from the CAD software used on the design phase of that component. These CAD models, if available, are ideal; and tend to not faithfully represent the real component in terms of surface and dimensional tolerances, or due to limitations on their manufacturing process. There might be cases where the CAD model is not available for the studied part, i.e., when reverse-engineering a product from a competitor. This paper proposes generating these FE models via 3D geometry scanning. The 3D geometry scanner used for this study, property of the Laboratory for Machine Dynamics and NVH at the TUAS Wildau, is capable of digitalizing any kind of surface with a tolerance of around ˙10 m. In theory, small manufacturing defects, holes, or wear marks can be incorporated to a faithful model, which can be later imported in the FEA software for its further processing, as described in [11]. The aim of this work is threefold: (1) demonstrating that a precise modal test, at a single point of the non-linear force/response curve can be achieved in complex structures, such a drone propeller blade, through the full automatization of a modal test by combining automatic modal excitation with a 3D SLDV system; (2) showing that, in the studied case, a FE surface model obtained by means of 3D geometry scanning can replace an unavailable CAD model on a FE simulation, and (3) proving that an acceptable degree of correlation on the modal model generated in FEA is possible by means of the joint work of the devices described in the next chapters. 2.2 Materials and Methods: Experimental Modal Analysis The experimental leg of this correlation cycle consisted on performing a modal test on the propeller blade. The experimental setup is shown in Fig. 2.4, built inside the hemi-anechoic chamber at the TUAS Wildau. A workbench clamp was used to constrain the propeller at its pinhole in order to imitate real constraints (Fig. 2.4, left). This fixture was, at its time, rigidly
16 D.J. Alarcón et al. Fig. 2.4 Left—Clamping of the propeller blade and SAM impact position. 101 DOFs were used for this modal analysis. Right—Experimental setup for this test. A large angle between scanning laser heads was needed in order to better acquire the in-plane vibration components fixed to the testing surface with help of a set of parallel clamps. The propeller blade was excited at its pressure surface [12], at the height of the degree of freedom (DOF) number 21 (Fig. 2.4, left) with a Scalable Automatic Modal Hammer (SAM) (NV-TECH, Steinheim a.d. Murr, Germany), presented in [13]. The response signal was measured by means of a 3D SLDV system model PSV-500-3D-H (Polytec GmbH, Waldbronn, Germany), property of the Laboratory for Machine Dynamics and NVH at the TUAS Wildau. The tested propeller was uniformly sprayed before the measurement with a nonaqueous white color developer (ARDROX 9D1B, Chemetall GmbH, Frankfurt a.M., Germany) to improve the laser beams reflectivity. A mesh with 101 DOFs (Fig. 2.4, left) was chosen to better display the mode shapes at higher frequency ranges and thus better to observe the local in-plane phenomena occurring in some out-of-plane modes; while keeping a relatively low testing time. A large number of these DOFs were placed directly at the trailing edge of the propeller, where a large amount of local motion occurs. The excitation force was provided in this modal test by the aforementioned SAM, as seen on Fig. 2.4. The sensor tip applied on this case was a model 086E80 (PCB Piezotronics, Inc., Depew, NY, USA), assembled on a small stepper motor. The rotation movement and its velocity are digitally controlled via the stepper motor software and transmitted to the motor via USB. This fact allows a very fine rotation tuning. For this experiment, a resolution of ˙0.45ı per step proved to be precise enough. Without the SAM, the testing time for a large number of DOFs would be of several hours and difficulties related with operator fatigue would arise. By using the PSV Scanning Vibrometer Software form Polytec GmbH for the data acquisition, the sampling frequency was set at 24.6 kHz, which resulted in an effective bandwidth of 10 kHz. With 12,800 FFT lines, the measurement resolution is of 0.78 Hz and thus, each measurement block was 1.28 s long. No function windows or signal overlap were applied to avoid distorting the time input and output signals in any way. The laser velocity range was fixed at 100 mm/s/V with a fast tracking filter. It was already known beforehand that some DOFs close to the propeller tip are very compliant, and they would present velocity peaks higher than 100 mm/s/V. For this reason, the software options regarding laser auto-ranging and the re-measurement of DOFs out of velocity range were set to on. For this experimental setup, the SAM was set with an inward acceleration of 1000 microsteps/s2 and 90 microsteps rotation (this is, approximately, 5ı between the hammer tip at its starting position and the impact point at the propeller). This
RkJQdWJsaXNoZXIy MTMzNzEzMQ==