ML051520510
| ML051520510 | |
| Person / Time | |
|---|---|
| Site: | Quad Cities  |
| Issue date: | 05/31/2005 |
| From: | General Electric Co |
| To: | Document Control Desk, Office of Nuclear Reactor Regulation |
| References | |
| DRF-0000-0039-5860 GENE-0000-0039-5860-01-NP, Rev 1 | |
| Download: ML051520510 (125) | |
Text
ENCLOSURE 2 ATTACHMENT 1 "Test and Analysis Report, Quad Cities New Design Steam Dryer, Dryer #1 Experimental Modal Analysis and Correlation with Finite Element Results," GENE-0000-0039-5860-01-NP, Revision 1, Non-Proprietary, dated May 2005
WY~
GE Nuclear Energy General Electric Company 1989 Little Orchard Street, San Jose, CA 95125 GENE-0000-0039-5860-0 1 -NP Class I Revision I DRF-0000-0039-5860 May 2005 TEST AND ANALYSIS REPORT QUAD CITIES NEW DESIGN STEAM DRYER DRYER #1 EXPERIMENTAL MODAL ANALYSIS AND CORRELATION WITH FINITE ELEMENT RESULTS
GENE-0000-0039-5860-01-NP, Revision 1 IMPORTANT NOTICE REGARDING CONTENTS OF THIS REPORT Please Read Carefully The only undertakings of the General Electric Company (GE) respecting information in this document are contained in the contract between the company receiving this document and GE. Nothing contained in this document shall be construed as changing the applicable contract. The use of this information by anyone other than a customer authorized by GE to have this document, or for any purpose other than that for which it is intended, is not authorized.
With respect to any unauthorized use, GE makes no representation or warranty, and assumes no liability as to the completeness, accuracy or usefulness of the information contained in this document, or that its use may not infringe privately owned rights.
ii
GENE-0000-0039-5860-01-NP, Revision 1 Test and Analysis Report Quad Cities New Design Steam Dryer Dryer #1 Experimental Modal Analysis and Correlation with Finite Element Results Page I of 122
GENE-0000-0039-5860-01-NP. Revision 1 Table of Contents List of Tables.................................................
3 List of Figures................................................
3 Acronyms................................................
6 1.0 Executive Summary.................................................
7 2.0 Scope.................................................
9 3.0 Background................................................
9 4.0 Purpose................................................
10 5.0 Experimental Setup.....
1.....................................
1 5.1 Test Configuration................................................
I 1 5.2 Data Acquisition System and Instrumentation................................................ 12 5.3 Sensor Locations................................................
13 5.4 Frequency Response Function Measurement Settings................................................
13 5.5 Test Documentation................................................
14 6.0 Test Results................................................
15 6.1 Skirt Damping Measurement Results................................
15 6.2 Hood Damping Measurement Results................................
17 6.4 Discussion of Damping Results................................
18 6.5 Dryer Top Experimental Results................................
20 6.6 0° Vertical Side Experimental Results................................
21 7.0 Correlation................................
23 7.1 Procedure................................
23 7.2 Geometric Correlation................................
23 7.3 Modal Correlation................................
24 7.4 FRF Correlation................................
26 7.5 Changes to As-received model....................................................................................................... 28 8.0 Correlation of Test Results and FE Results................................
28 8.1 Global Comparison................................
29 8.2 Skirt Component Comparison................................
29 8.3 900 Hood Component Comparison................................
33 8.4 2700 Hood Component Comparison................................
35 8.5 Additional Observations................................
36 9.0 Summary and Conclusions................................
38 10.0 References................................
41 Figures.................................
42 Attachment A: Steam Dryer Experimental Modal Analysis Test Plan................................
82 Attachment B: Skirt Sub-Model Study................................
92 Attachment C: Effect of Number of Averages................................
97 Attachment D: Acquisition Front-end channel assignment................................
104 Attachment E: Test Log Sheets................................
110 Page 2 of 122
GENE-0000-0039-5860-01-NP, Revision 1 List of Tables Table 1: Dryer #I Skirt Damping Results at Low Water Level...............................
....................... 16 Table 2: Dryer #I Hood Damping Results at Low Water Level.......................................................
8 Table 3: Dryer #I Top Experimental Frequencies......................................................
20 Table 4: Dryer #1 00 Vertical Side Experimental Frequencies......................................................
22 Table 5: Dryer #I Skirt Experimental Frequencies......................................................
30 Table 6: Frequency Change after addition of instrumentation to 90° skirt panel........................................ 33 Table 7: Dryer #1 900 Hood Experimental Frequencies......................................................
35 Table 8: Dryer #1 2700 Hood Experimental Frequencies......................................................
36 List of Figures Figure 1: Finite Element Model Representation of Dryer 43 Figure 2: Photograph ofDryer #1,00 Side 43 Figure 3: Photograph of Dryer #1, 180° Side 44 Figure 4: Tripod Supports before Installation 44 Figure 5: Tripod Support Connection to Dryer 45 Figure 6: Comparison of Measurements using Hammer Tips of different hardness, Radial Skirt Response to Radial Skirt Impact (Red curve is soft tip, Green curve is medium tip) 45 Figure 7: Comparison of Measurements using Hammer Tips of different hardness, Vertical Side Response to Skirt Impact (Red curve is soft tip, Green curve is medium tip) 46 Figure 8: Comparison of Measurements using Hammer Tips of different hardness, 90' Hood Response to Skirt Impact (Red curve is soft tip, Green curve is medium tip) 46 Figure 9: Full view of 90 degree hood of dryer #I during measurements.
47 Figure 10: Pattern of measurement points for the 90 degree hood of dryer # 1, where light red mesh shows the 90 degree hood response locations, dark red represents the responses on top & aqua represents the responses on the ring.
47 Figure I 1: Response locations on upper right corner of 90 degree hood of dryer # 1.
48 Figure 12: Response locations on lower right corner of 90 degree hood of dryer # 1.
48 Figure 13: Response locations on upper left corner of 90 degree hood of dryer #1.
49 Figure 14: Response locations on lower left corner of 90 degree hood of dryer #I with green circle indicating input location on 90 degree hood.
49 Figure 15: Response location on side panel to the right of main panel of 90 degree hood of dryer #1 (at angle w.r.t. main panel).
50 Figure 16: Response location on side panel to the right of main panel of 90 degree hood of dryer #1 (at angle w.r.t. main panel).
50 Figure 17: Response locations on side panel to the left of main panel of 90 degree hood of dryer #I (at angle w.r.t. main panel).
5 1 Figure 18: Full view of 270 degree hood of dryer #1 (before installation of temporary sensors).
51 Figure 19: Pattern of measurement points for the 270 degree hood of dryer #I where yellow mesh is 270 hood response locations, dark red is responses on top & aqua is responses on the ring.
52 Figure 20: Response locations on upper part of 270 degree hood of dryer # 1.
52 Figure 21: Response locations on right side of 270 degree hood of dryer # 1 53 Figure 22: Response locations on lower part of 270 degree hood of dryer #1 53 Figure 23: Response locations on lower right corner of 270 degree hood of dryer #1 with green circle indicating input location on 270 degree hood.
54 Figure 24: Response locations on side panel to the left of main panel of 270 degree hood of dryer # I (at angle w.r.t. main panel).
54 Figure 25: Response locations on side panel to the right of main panel of 270 degree hood of dryer #I (at angle w.r.t. main panel).
55 Page 3 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 26: Full view of 0 degree side of dryer# I (before installation of temporary sensors).
56 Figure 27: Pattern of measurement points for the 0 degree side of dryer #I where pink mesh is 0 degree side response locations, dark red is responses on top & aqua is responses on the ring.
56 Figure 28: Response locations on left side of 0 degree side of dryer #1 with green circle indicating input location on 0 degree side 57 Figure 29: Response locations on middle part of 0 degree side of dryer # 1.
57 Figure 30: Response locations on lower right part of 0 degree side of dryer # 1.
58 Figure 31: Response locations on upper right part of 0 degree side of dryer # 1.
58 Figure 32: Response locations on 0 degree side of dryer #1.
59 Figure 33: Full view of 180 degree side of dryer #1 (before installation of temporary sensors).
59 Figure 34: Pattern of measurement points for the 180 degree side of dryer #I where blue mesh is 180 hood response locations, dark red is responses on top & aqua is responses on the ring.
60 Figure 35: Pattern of measurement points for the skirt & ring of dryer #I where purple mesh is ring response locations, aqua mesh is the response locations on the ring, blue is responses on 180 degree side & pink is responses on 0 degree side.
60 Figure 36: Pattern of measurement points for the skirt & ring of dryer #I where the red mesh is response locations on top, green is the response locations on the panels inside the dryer, aqua is responses on the ring, blue is responses on 180 degree side & pink is responses on 0 degree side.
61 Figure 37: 90° Skirt Panel Input and Response Locations 61 Figure 38: 2700 Skirt Panel Input and Response Locations 62 Figure 39: Summation FRF for Top: Red - Test 62 Figure 40: Summation FRF for 00 Side: Red - Test 63 Figure 41: Summation FRF for Skirt: Red - FE, Green - Test 63 Figure 42: FRF on 900 Skirt Panel, Upper Point: Red - FE, Green - Test, Blue - Damping Test (different input point from Green) 64 Figure 43: FRF on 900 Skirt Panel, Lower Point: Red - FE, Green - Test, Blue - Damping Test (different input point from Green) 65 Figure 44: Additional FRFs on 90° Skirt Panel, Test Measurements: Red - Left of 707 as seen in Figure 31, Green - Right of Centerline, between 707 and 708 Vertically, Blue - Upper Middle, Pink - Lower Middle 65 Figure 45: FRF on 2700 Skirt Panel, Upper Point: Red - FE, Green - Test 66 Figure 46: FRF on 2700 Skirt Panel, Lower Point: Red - FE, Green - Test 66 Figure 47: FRF on 2700 Skirt Panel, from Damping Test: Red - Upper Point, Green - Lower Point 67 Figure 48: FRFs on 900 and 2700 Skirt Panel from Damping Test: Red -Test -2700 skirt, upper point, Green -
Test -2700 skirt, lower point, Blue -Test - 90° skirt, upper point, Pink -Test - 900 skirt, lower point 67 Figure 49: Mode Shapes for Skirt, Global View: Left - Test, ((
68 Figure 50: Mode Shapes for Skirt, Local View: Left - Test, [
69 Figure 51: Mode Shapes for Skirt, Local View: Left - Test, [
70 Figure 52: Mode Shapes for Skirt, Local View: Left - Test, ((
))
71 Figure 53: Summation FRF for 90'Hood: Red - FE, Green - Test 72 Figure 54: Summation FRF for 90'Hood, FE in X-direction, Test normal to surface: Red - FE, Green - Test 72 Figure 55: MAC Matrix for 90° Hood, ((
1]
73 Figure 56: Mode Shapes for 90° Hood, Global View: Left - Test, ((
))
73 Figure 57: Mode Shapes for 900 Hood, Global View: Left - Test, ((
))
74 Figure 58: MAC Matrix for 90° Hood, ((
74 Figure 59: Summation FRF for 2700 Hood: Red - FE, Green - Test 75 Figure 60: Summation FRF for 2700 Hood, FE -X-direction, Test - normal to surface (12.50 Difference): Red
- FE, Green - Test 75 Figure 61: Individual FRFs for 2700 Hood, FE - Top, X-direction; Middle, Y-direction; Bottom, Z-direction:
Red - FE, Green - Test 76 Figure 62: Individual FRF for 2700 Hood, X-direction: Red - FE, Green - Test 76 Figure 63: Mode Shapes for 2700 Hood, Local View: Top - Test, ((
1]
77 Figure 64: Mode Shapes for 2700 Hood, Global View: Top - Test, ((
))
78 Figure 65: MAC Matrix for 270° Hood, ((
79 Figure 66: Mode Shapes for 2700 Hood, Local View: Top - Test, ((
80 Figure 67: Mode Shapes for 2700 Hood, Global View: Top - Test, ((
))
81 Page 4 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure A-I: 900 Side Response and Input Locations 84 Figure A-2: 2700 Side Response and Input Locations 84 Figure A-3: 1800 Side Response and Input Locations 85 Figure A-4: 00 Side Response and Input Locations 85 Figure A-5: Top Side Response and Input Locations 86 Figure A-6: 90° Hood (hda), Ring and Skirt 89 Figure A-7: 1800 Hood (hdb), Ring and Skirt 89
' Figure A-8: 1800 Side (sidc), Ring and Skirt 90 Figure A-9: 00 Side (sidd), Ring and Skirt 90 Figure A-10: Top (top) and top view of 180° Side (sidc) 91 Figure B-I: Location of Conduit and Instrumentation Pads on Skirt Panel 93 Figure B-2 Left - Baseline, ((
))
93 Figure B-3: Left - Baseline, (( f]
94 Figure B-4: Left - Baseline, ((
94 Figure B-5: Left - Baseline, ((
1]
95 Figure B-6: Left - Baseline, ((
))
95 Figure B-7: Left - Baseline, ((
))
96 Figure E-l: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:785 on the 90 degree skirt panel of Dryer #2 98 Figure E-2: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:753 on the 90 degree skirt panel of Dryer #2 99 Figure E-3: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:707 on the 90 degree skirt panel of Dryer #2 100 Figure E-4: Comparison ofthe FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:722 on the 270 degree skirt panel of Dryer #2 101 Figure E-5: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:762 on the 270 degree skirt panel of dryer 2 102 Figure E-6: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and response at strain gage channel C at point skir:796 on the 90 degree skirt panel of Dryer #2 103 Page S of 122
GENE-0000-0039-5860-01-NP, Revision 1 Acronyms BWR DAS EMA EPU FE FEA FRF GE GENE MAC MPT NPT ODS OLTP QCI QC2 RPV Boiling Water Reactor Data Acquisition System Experimental Modal Analysis Extended Power Uprate Finite Element Finite Element Analysis Frequency Response Function General Electric General Electric Nuclear Energy Modal Assurance Criterion Mode Pair Table Node Pair Table Operational Deflection Shape Original Licensed Thermal Power Quad Cities Unit I Quad Cities Unit 2 Reactor Pressure Vessel Page 6 of 122
GENE-0000-0039-5860-01-NP, Revision 1 1.0 Executive Summary An experimental modal analysis was performed on new design Dryer #1, the instrumented dryer intended for Quad Cities Unit #2 (QC2), and the results were compared to finite element analysis results on a frequency basis and on a mode shape basis. The finite element analysis included a modal analysis and, using those modes from the modal analysis, a mode superposition to obtain FRFs that match the input and response points of the test data. The test results showed many similarities to the finite element results. In terms of frequencies, the trend is that the finite element model frequencies are generally in good agreement with the test frequencies. For example, the first dominant mode of the 90° skirt panel is at ((
)) in the finite element analysis and at ((
)) in the test results, a difference of 2.4%. The 2700 skirt panel, the 90° hood and the 2700 hood all had frequency differences between experimental and analytical results of less than 10% for their lowest frequency dominant modes. In terms of FRF comparisons, the various components examined showed generally good agreement in trends and levels between summation FRFs for test and FE for that specific component.
As expected, the test frequency versus finite element frequency agreement decreased as frequency increased and the modes increased in complexity. The largest potential discrepancy seen between the model and the test is the 90° skirt panel with a significant test frequency being at ((
)) and the finite element model frequencies at ((
)) for a vertical 2nd panel bending mode and near ((
)) for a lateral 2nd panel bending mode.
There are several explanations for this discrepancy; however, with the number of sensors on this panel and their locations, a conclusive explanation cannot be provided. The 2700 skirt panel and the 90° and 270° hoods showed better agreement for similar frequencies.
The decrease in test versus finite element agreement as frequency increases is expected because, as the mode shapes become more complex with higher frequency, the finite element model results become more sensitive to several parameters such as boundary condition details, element type and number of elements. The boundary conditions of the individual plates, the welds to other plates and beams, are not explicitly modeled in the finite element Page 7 of 122
GENE-0000-0039-5860-01-NP, Revision 1 model. The wavelengths of the higher modes are shorter so more elements are needed to accurately represent the shapes; however, increasing the number of elements prohibitively increases computation error and adds to round-off error accumulation. The lower skirt super element, which includes hydrodynamic effects of the water on the skirt up to Low Water Level (LWL), may have some effect on the higher mode frequencies and test versus finite element differences as well.
Also, some of the perceived difference for higher frequency modes may be that too few transducers were used in the experimental modal analysis to characterize the mode shape well. The 90° skirt panel is an example where the number of transducers and their location do not appear to allow adequate resolution of some modes whose frequencies appear significantly in the frequency response measurements. Time constraints imposed upon the Dryer # I experimental modal analysis prevented more locations from being measured. As of the writing of this revision, changes were implemented in the Dryer #2 test plan to add locations to specific components and to reduce the number of conditions (water levels) for testing so that both the time constraint condition and the need for better spatial resolution of the measurement locations are satisfied.
In addition to determining the natural frequencies and mode shapes, the hammer test responses are used to experimentally determine damping values on the skirt and hood at low water level. The purpose of the experimentally determined damping values was to validate the damping values used in the stress prediction analyses. The damping measurement results showed a range of damping values which form a technical basis for arriving at appropriate damping values to be used for structural response analyses.
From the above discussion on the good agreement of the first mode frequency comparisons and the reasonable agreement in the higher mode frequencies, it is concluded that the impact hammer test results verify that the finite element model used for dryer design calculations is sufficiently dynamically similar to the as-built dryer for engineering purposes. In addition the impact hammer test results also show that the ((
))
value used in the stress prediction analyses are realistic and adequately conservative.
Page 8 of 122
GENE-0000-0039-5860-01-NP, Revision 1 2.0 Scope This document summarizes the experimental modal analysis and correlation with finite element results performed to compare the finite element model of the new design steam dryer with the actual production Dryer #1. The experimental modal analysis and the finite element analysis were conducted during early April 2005. The following items are included in this document:
- 1. Description of the testing performed
- 2. Presentation of experimental data
- 3. Description of the finite element analysis
- 4. Comparison of experimental results with finite element results
3.0 Background
This section provides background information intended to help the reader understand the events that precipitated this program.
The original design BWR steam dryers functioned acceptably at Original Licensed Thermal Power (OLTP) for many years. In response to some cracking of original design and modified original design BWR steam dryers when operation shifted to Extended Power Uprate (EPU),
GENE has initiated a program to develop a new design of BWR steam dryer with the design intent of being able to survive loads imposed by EPU. As this design was substantially different from the original design, experimental correlation measurements were considered necessary to determine if the finite element analysis used to predict stresses at EPU with its higher loads correlated to actual hardware fabricated according to the new design. Figure 1 is a depiction of the finite element model for the new dryer design, and Figures 2 and 3 are pictures of the actual unit undergoing final assembly.
Page 9 of 122
GENE-0000-0039-5860-01-NP, Revision 1 4.0 Purpose The testing and analysis described in this document were defined to accomplish the following main objective:
- Determine if the lowest or first dominant frequencies of major components of the new design steam dryer are within 10% of the frequencies predicted by finite element analysis for the dryer configuration at low water level The specific purpose of the testing, the experimental modal analysis, is as stated in Reference 1, Steam Dryer Hammer Test Specification, GE 26A6380, Revision 1:
- To identify the as-built frequencies and mode shapes of the dryer's key components at ambient conditions.
These as-built frequencies and mode shapes are to be compared with mode shapes, frequencies and FRFs generated from the finite element model of the dryer.
The program has several side objectives as well:
- 1. To experimentally measure damping values on the skirt and hood at-low water level to validate assumptions used in the stress prediction analyses
- 2. To recommend areas for investigation of differences between model and as-built hardware
- 3. To recommend areas of improvement of the finite element model to more closely match the as-built hardware Page 10 of 122
GENE-0000-0039-5860-01-NP, Revision 1 5.0 Experimental Setup This section describes the test configuration and environment, identifies the instrumentation and data acquisition equipment used, and identifies the sensor locations. The setup and testing follows the requirements outlined in Reference 1, Steam Dryer Hammer Test Specification, GE 26A6380, Revision 1.
5.1 Test Configuration Steam Dryer #1, the first dryer with the new design, was tested at J.T. Cullen in Fulton, Illinois, a fabrication facility that served as the location for final assembly, installation of the final modifications and installation of the permanent sensors. For the experimental modal test, the steam dryer was supported in a water tank by 4 tripods with extensions that fit into its main support lug sockets. These tripods were welded to metal plates which were bolted to the concrete floor of the fabrication shop. One item to note is that, partway through the testing, the dryer was removed from its supports and transferred to another location for laser dimensional measurements. When it was removed, 2 of the supports had to be detached where they were welded at the floor. When it was returned to its supports, a third support had to be detached from its weld at the floor to align the dryer properly. Data taken before and after this move indicates that some of the low frequency (below 10 Hz) suspension modes were not at exactly the same frequency before and after this move, making the effort to match the actual support stiffness in the FE model more difficult. Figure 4 is a picture of one of the support tripods, and Figure 5 is a close-in picture of the support/dryer connection.
A circular tank with a liner was used to hold water for the testing with water. The tank's inner diameter replicated the inner diameter of the reactor pressure vessel (RPV) at the plant to attempt to match the hydrostatic loading at the plant. Testing was performed at 4 different water levels:
- 1. Dry - no water Page 11 of 122
GENE-0000-0039-5860-01-NP, Revision 1
- 2. LWL - Low Water Level - water up to 32.5 inches above the bottom of the bottom flange of the dryer.
- 3. NWL - Normal Water Level - water up to 36.5 inches above the bottom of the bottom flange of the dryer.
- 4. HWL - High Water Level - water up to 40.5 inches above the bottom of the bottom flange of the dryer.
All testing was performed at ambient conditions at the test site, with the temperature ranging from 600 F to 750 F 5.2 Data Acquisition System and Instrumentation The following instrumentation was used to perform the experimental modal analysis and the static test:
- 1. PCB Model 356A22, 356B08 and 356A15 triaxial accelerometers, and Model 333B30, 333A32, 333B32 and 352C43 single-axis accelerometers
- 2. PCB Model 086D50 Impact Hammer - after preliminary measurements, the softest hammer tip was selected. Figures 6, 7 and 8 are comparisons of the softest tip with the medium tip. The softest tip provides better low frequency results than the medium tip. The low frequency portion of the FRF is cleaner as seen in the middle plot in Figures 6, 7 and 8, and the low frequency portion of the coherence, the lower plot in Figure 6, 7 and 8, has a higher value with the softest tip. At the start of testing on Dryer #1, it was believed that obtaining very good low frequency data was more important than extending the frequency range of the results; however, subsequent review and discussion of the results produced a decision to use the medium tip for Dryer #2.
- 3. Vishay Micro-Measurements CEA-06-125UR-350 Strain Gages Page 12 of 122
GENE-0000-0039-5860-01-NP, Revision 1
- 4. Omega LC304-5K Load Cell The following equipment was used to record and analyze the test data:
- 1. A 116 channel LMS SCADAS 111 dynamic signal analyzer (2 SCADAS I11 Model 316 front ends in a Master-Slave configuration) with PQA and PQFA modules was used to provide ICP power to and receive the signal from all of the ICP sensors. For the strain gage measurements, PQBA modules were used in the SCADAS 316 to provide bridge completion and signal conditioning. The system was controlled by a personal computer using LMS Test.Lab 5A softwvare, specifically the Modal Impact, Modal Analysis and Spectral Acquisition modules of software.
5.3 Sensor Locations The following items detail the contents of Figures 9 through 39 which are pictures that identify the input and response locations used for the experimental modal analysis and static load test.
- Figures 9 through 17 are the 900 hood
- Figures 18 through 25 are the 2700 hood
- The 0° vertical side locations are shown in Figures 26 through 32
- TheI 80° vertical side locations are shown in Figures 33 through 34
- Figures 35 and 36 contain the test wireframe for the 00 vertical side locations and theI80' vertical side locations
- Figures 37 and 38 depict the impact locations and the accelerometer and strain response locations for the 90° skirt panel and the 2700 skirt panel 5.4 Frequency Response Function Measurement Settings The signal processing parameters used for data acquisition were the following:
- 5 to 10 averages (1 average for the time domain damping measurements)
(Attachment C is a comparison of results from 5 averages and 20 averages for one measurement location on Dryer #2 to show that using 5 to 10 averages is Page 13 of 122
GENE-0000-0039-5860-01-NP, Revision 1 adequate, and the GE Hammer Test Specification, Reference 2, specifies 3 or more)
- Force window of 5% to 20% on the input (Uniform window used for damping measurements)
- Exponential window of 3% to 20% on the responses (Uniform window used for damping measurements)
- Effective Frequency bandwidth of 400 Hz (actual bandwidth setting 512 Hz, sampling frequency of 1024 Hz)
- 4096 Spectral lines (0.125 Hz resolution/8.0 second time length) 0.1 second pretrigger on the hammer input For the measurements, the following results were saved:
- Frequency Response Function
- Coherence
- Input Autopower
- Response Autopowers
- Time Record (for time domain damping measurements only) 5.5 Test Documentation Attachment D contains the channel assignment data sheets as referred to in the GE Hammer Test Specification. Attachment E contains the test log sheets. The test log sheets have been rewritten from the original LMS test log sheets to match the GE format and to clean them up.
Page 14 of 122
GENE-0000-0039-5860-01-NP, Revision 1 6.0 Test Results This section presents a subset of the test results. It covers the damping results based on experimental measurements and other experimental results on components that are not compared to FE results on a component basis. Additional test results that are correlated to FE results are presented in Section 8.0, Correlation of Test Results and Finite Element Results.
6.1 Skirt Damping Measurement Results Specific measurements were performed to measure the damping on the 900 skirt panel and on the 2700 skirt panel at low water level. Acceleration and strain responses were acquired on these panels in response to impacts. Figures 38 and 39 show the impact and response locations for the 90° skirt panel and on the 2700 skirt panel, respectively, for the damping measurements.
Damping was calculated in the frequency domain using modal curve-fitting methods on individual FRF measurements and on all of the FRF measurements for the specific component and using the half power bandwidth method or equivalent methods with different "dB down" values and in the time domain using the log decrement method. Reference 2, SAE Recommended Practice J 1637, contains the calculations for the equivalent methods to the half power bandwidth method, and Reference 3 provides a discussion of the modal curve-fitting method used for the individual FRFs and for the whole component analysis.
The damping results on the skirt are presented in Table 1.
Page 15 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Table 1: Dryer #1 Skirt Damping Results at Low Water Level 1!
))
Page 16 of 122
GENE-0000-0039-5860-01-NP, Revision 1 The damping in terms of percent critical damping ranged from ((
)) on the 90° skirt panel, with the higher frequencies generally showing less damping. The 2700 skirt panel showed a similar trend, with a range of ((
)) In general, the strain gages showed slightly higher damping levels for the same mode than the accelerometers.
6.2 Hood Damping Measurement Results Specific measurements were performed to measure the damping on the 90° hood and on the 270° hood at low water level. Acceleration was acquired on these panels in response to impacts. Figures I I through 15 and Figures 20 through 25 show the impact and response locations for damping measurements on the 90° hood and on the 270° hood, respectively.
Damping was calculated in the frequency domain using modal curve-fitting methods on individual FRF measurements and on all of the FRF measurements for the specific component and using the half power bandwidth method or equivalent methods with different "dB down" values. Table 2 contains the damping values for the hoods.
Page 17 of 122
GENE-0000-0039-5860-01-NP. Revision 1 Table 2: Dryer #1 Hood Damping Results at Low Water Level 11
))
The hood panels generally showed less damping than the skirt panels. Percent of critical damping values ranged from ((
)) on the 90° hood and from
((
))on the 2700 hood. Again, damping decreased as frequency increased.
6.4 Discussion of Damping Results Several questions arise in review of damping results. One question is, "Why does the damping decrease with frequency?"
- 1. The input force decreased as frequency increased. Quite often, there is a relationship between input force and measured damping, with damping increasing as input force Page I8 of 122
GENE-0000-0039-5860-01-NP, Revision 1 increases, so possibly in this case the decrease in force versus frequency is also affecting the decrease in damping versus frequency
- 2. The loading of the water on the lower skirt panel has a greater damping effect at low frequency than high frequency A brief literature review produced no results that appear to be immediately relevant to a large welded structure such as the dryer.
Another question concerned the increase in damping from accelerometer to strain results.
Again, 2 possibilities are present but no conclusive determination has been made:
- 1. The strain gages were at the panel edges, greater than 1 inch from the weld but closer than 3 inches, while the accelerometers were in the middle of the panels. It is believed that these locations could show more damping than the accelerometer locations
- 2. The strain gage is a displacement-based device while the accelerometer is acceleration-based. There is some belief that this difference is causing a slight increase in damping estimates from strain versus those from acceleration.
Further review of both damping versus frequency and damping from strain versus damping from acceleration will be performed using Dryer #2 data when it is available.
Page 19 of 122
GENE-0000-0039-5860-01-NP, Revision 1 6.5 Dryer Top Experimental Results This section contains experimental results from the dryer top as well as a very limited number of the points on the inner hood and perforated plates. No specific correlation exercise was performed for the dryer top so only the experimental results for modes from curve-fitting up to 100 Hz are presented here.
Table 3: Dryer-#I Top Experimental Frequencies Percent Mode Test Data Critical Number Frequency Damping Mode 1 Mode 2 Mode 3 Mode 4 Mode 5 Mode 6 Mode 7 Mode 8 Mode 9 Mode 10 Mode 1 1 Mode 12 Mode 13 Mode 14 Mode 15 Mode 16 Mode 17 Mode 18 Mode 19 Mode 21 Mode 22 Mode 25 Mode 26 Mode 28 Mode 29 Mode 30 Mode 32 Fr 1]
Table 3 contains significant frequencies for the dryer top below 100 Hz, and Figure 40 contains the summation FRF for the dryer top measurements. The summation FRF is an average of the FRFs in all 3 directions for the included points unless noted. Some of the Page 20 of 122
GENE-0000-0039-5860-01-NP, Revision 1 summation FRFs presented in this report will only be the measurements perpendicular to the surface and will be noted as such 1))
6.6 00 Vertical Side Experimental Results As with the dryer top, no specific correlation to finite element results was performed for the 0° vertical side. Figure 41 is the summation FRF for the 00 vertical side, and Table 4 on the next page contains significant frequencies for this component below 94 Hz. The 1800 vertical side is omitted because there was not a specific impact point on that side due to lack of access.
Page 21 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Table 4: Dryer #1 0° Vertical Side Experimental Frequencies Percent Critical Mode Test Data Damping Number Frequency Mode 5 Mode 6 Mode 7 Mode 8 Mode 9 Mode 10 Mode 11 Mode 12 Mode 13 Mode 15 Mode 17 Mode 18 Mode 20 Mode 21 Mode 22 Mode 23 Mode 24 Mode 25 Mode 26 Mode 27 Mode 28 Mode 29 Mode 30 Mode 31 Mode 32 Mode 33 Mode 34 Mode 35 Mode 36 Mode 38 Mode 39 Mode 40 Mode 41 Mode 42 Mode 46 Mode 49 Mode 50 Mode 51 Mode 52 Mode 53 Mode 54 Mode 55 Mode 56 Page 22 of 122
GENE-0000-0039-5860-01-NP, Revision 1 7.0 Correlation This section discusses the process of using finite element model results to correlate with experimental test results.
7.1 Procedure The goal of the correlation is to determine the differences between the FE model and the test object and to determine the sensitive spots of the FE model. After vards, one can improve the correlation, focusing on those hot spots, and finally obtain a model that is more realistic than the previous one.
The correlation procedure contains the following steps:
Geometric correlation: Definition of a relationship between the units, the coordinate systems and the measuring points of the TEST model on one hand, and the units, the coordinate systems and the nodes of the FE model on the other hand.
Modal correlation: Comparison of the experimental mode shapes and the FE mode shapes, based on wireframe animations and MAC (modal assurance criterion) calculations.
Correlation of the transfer functions: Comparison of the measured FRFs and the ones calculated directly or synthesized from the FE mode shapes.
The calculations of the FE mode shapes and FRFs have been done here with ANSYS 8.1.
The correlation has mainly been done in LMS/LINK and Test.Lab.
7.2 Geometric Correlation This phase ensures the compatibility between the measured and the calculated data. It includes the following steps:
- 1. Definition of rational entities/groups in the model, grouping nodes that have common properties or are part of the same panel for example. Those entities are called components.
Page 23 of 122
GENE-0000-0039-5860-01-NP, Revision 1
- 2. Definition of pairs between TEST measuring points and FE nodes. These pairs are stored in a node pair table (NPT). This table allows the automatic projection of the TEST geometry on the FE geometry.
- 3. Compatibility between the global and local coordinate systems. This information is also contained in the node pair table and ensures a proper projection of the TEST geometry on the FE geometry.
- 4. Transformation of the TEST data (modes, FRFs, wireframes) to the verification system (FE model), in order to make the comparison easier.
All these steps were done in LMS/LINK.
7.3 Modal Correlation The dynamic behavior of a structure, at least in the low frequency range, is best described by its normal modes. From experimental normal analysis, the modes of the structural components are known with their shapes, frequencies, and modal damping.
For the FE model, shapes and frequencies of normal modes are calculated in ANSYS 8.1 (analysis ANTYPE, 2). For the test results, the general process has been to calculate the complex modes first, review those modes, and then use the same poles to calculate the normal modes for comparison with FE results.
Experimental modes can be directly compared to FE modes by using the wireframe animations from both Test and FE. Using this technique, important conclusions can be drawn regarding parts that are not well modeled in the FE.
After the first mode extraction for example, four grounded springs were added to the model as boundary condition. These grounded springs (COMBIN14) were fixed to the dryer ring in the 4 support points. This was intended to model as simply as possible the support beams used in the test rig that support the dryer. This was necessary to be able to extract the suspension modes and model the influence that the supports have on the rest of the structure as there was concern that the support stiffness influenced some of the lower frequency flexible modes of the dryer.
Page 24 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Of course, FE modes can also be visualized with the entire model, and not only the wireframe geometry. This way, one can easily understand the real nature of the modes, which is not always obvious with wireframe animations.
Another tool to judge modal correlation is the Modal Assurance Criterion (MAC). It expresses the nature of the relationship between two sets of vectors. For each pair of modes compared, it is calculated from the vectors of each degree of freedom considered in the correlation model. Mathematically, the MAC is defined as
}te
{
,,,Iest} JiFE}{YFE })
where (T,'e5)is a modal vector from test and (W.
F'E)is a modal vector from the finite element solution. These modal vectors each represent a single frequency and contain the common degrees of freedom between the test and the FE, in this case either on a global basis or on a component basis. For identical modes, the MAC is 1. For linearly independent mode shapes, the MAC is 0. As the MAC formulation includes a quadratic term, small or local deviations between Test and FE will result in considerably lowered MAC values.
Unfortunately, the FE modal density at low frequency is already very high for the dryer. And too many local effects are already present at the very low frequencies (this is a direct result of the low stiffness of the perforated inner panels of the dryer). As a result, the global modes are mixed/lost among many local modes. For example here, in the 200 first FE modes, only 10 had global characteristics; all the others were 1st, 2nd and 3rd bending modes of the inner perforated panels.
On the other hand, the low number of measuring points on a region makes it impossible to capture those local modes in the experimental TEST modal analysis.
Page 25 of 122
GENE-0000-0039-5860-01-NP, Revision 1 As result, one TEST mode corresponds to many FE modes, global, local or mixed. A one-to-one TEST-FE mode comparison is thus very difficult, and makes it almost impossible to draw any conclusion of the MAC analysis.
The MAC analysis was still used, but generally only after other methods discussed in the next sections had narrowed down the frequency range to be examined.
7.4 FRF Correlation Normal modes are an efficient representation of the dynamic behavior of a structure, but only useful for correlation as long as the modes are not combining too many local effects. As mentioned above, even at low frequency, there are many local modes, making it virtually impossible to combine the same local effects in one mode of the FE model, as can be observed in the test model. In reality, different local effects get shifted to different frequencies, making it impossible to find the same combination of the same local effects in test and FE. As long as the same local effects occur at approximately the same frequency, a good dynamic correlation is obtained, even if this is not visible in a MAC matrix.
Another way thus to compare dynamic behavior of two models is looking at frequency response functions (FRFs). They can be evaluated over the whole frequency range of interest.
In an FRF, the (acceleration) response of one point is plotted as function of a unit (force) input in another point. Modes can be found as peaks in an FRF, and the higher the modal damping, the lower and broader the peak will be.
For checking correlation, the synthesized FE FRFs are compared with the corresponding measured test FRFs. More interesting than the exact amplitude is the general shape of a FRF.
A good correlation means that important peaks from the test should be found in the synthesized FRFs at a similar frequency, and the general amplitude of the FE and test FRFs are similar.
A FRF can be calculated in ANSYS with the harmonic response analysis (ANTYPE, HARMIC). The idea is to impose a unitary force in one DOF of the hitting points and Page 26 of 122
GENE-0000-0039-5860-01-NP, Revision 1 calculate the frequency response (displacements or accelerations) in all the DOFs of all the other points. The obtained response functions can then effectively been seen as displacements or accelerations per unit input force.
Three methods of solution are available in ANSYS to calculate FRFs with the harmonic response analysis:
- 1. Full (HROPT, FULL)
This method solves the general equation of motion of a structural system transposed in frequency domain directly.
- 2. Reduced (HROPT, REDUC)
The reduced solution method uses reduced structure matrices to solve the equation of motion. This method runs faster than the full harmonic response by several orders of magnitudes, because the technique of matrix reduction is used so that the matrix used to represent the system is reduced to the essential DOFs required to characterize the response of the system.
- 3. Mode superposition (HROPT, MSUP)
The mode superposition method uses the natural frequencies and mode shapes from the modal analysis (ANTYPE, MODAL) to compute the dynamic response to steady harmonic excitation. It converts the equation of motion in its modal form. The advantage of this method is that the computationally expensive matrix algebra can be calculated inexpensively in modal coordinates. The individual modal responses are then superimposed to obtain the actual response.
The Full method is very memory and time-consuming and was not used in this project. The Mode Superposition technique was preferred here. This method has as an advantage that the mode shapes only have to be calculated once and can be re-used to calculate FRFs for different input points. On the other hand, the reduced and full harmonic methods have to be repeated for each hitting point.
Page 27 of 122
GENE-0000-0039-5860-01-NP, Revision 1 7.5 Changes to As-received model Several changes were made to the model after it was received from GE to more closely replicate the test conditions.
- Elastic modulus and damping changed to ambient conditions from reactor conditions (the lower skirt super element with its loading by water was left unchanged)
- Removal of pressure loading (a pressure loading from reactor condtions)
- Addition of grounded springs instead of perfectly rigid constraints at the support locations (the dryer as tested was supported on stands that exhibited flexibility as compared to the completely rigid FE constraints. A complete iteration of these springs to match the test results was not performed, and the dryer move for laser dimensional measurements caused the final sets of data to have slightly different support stiffnesses.)
8.0 Correlation of Test Results and FE Results This section presents comparisons between test results and finite element analysis results on several levels:
- Global
- Skirt Component 900 Hood Component
- 270° Hood Component The comparisons are presented as comparisons of frequencies, comparison of test FRFs to FE FRFs, comparison of local mode shapes and comparison of global mode shapes. Comparison started on a global level and moved to a component level due to the many closely spaced modes of some of the dryer components in the finite element model. Because of the modal density, many of the test modes appear similar to numerous FE modes, making it very difficult to find a single corresponding mode. Correlation on a component basis was performed by observing significant peaks in the summation FRF and some individual FRFs Page 28 of 122
GENE-0000-0039-5860-01-NP, Revision 1 of the component to narrow down the frequency range for searching the FE modes and then reviewing the mode shape animations and MAC matrices for the test and FE components.
Also, for the skirt and both hoods, 2 relatively widely spaced peaks in the frequency domain were examined. After that review, the other components were added to the mode shape to observe the global mode that the component mode was related to.
8.1 Global Comparison The results from the experimental modal and the finite element analyses have been compared on a global level. There are similar modes under 10 Hz that are directly controlled by the modeling of the supports. The stiffness of the support springs is still too low as the suspension modes of the FE model are at a lower frequency than those of the test results for similar modes under 10 Hz. For this study, it was decided to stop the iterative adjustment of the support springs as its effect on the component modes was expected to be minimal, but this adjustment is still an area for potential improvement.
The first significant global flexible mode is found at ((
)) a mode in which the hoods appear to slide laterally out of phase with each other with what appears to be a twist or rotation of the skirt and ring. The finite element counterpart to this mode ended up at ((
)) with the grounded springs used at the support locations. This frequency difference could probably be reduced with further adjustment of the grounded spring supports; however, as mentioned above, it was decided to stop iterating the FE grounded springs to match the test.
After the initial attempt at the global comparison, the focus shifted to a component basis and sometimes expanded to the global modes related to those specific component modes.
8.2 Skirt Component Comparison The skirt was the first component to be analyzed on a component basis. Table 5 contains frequencies from the test data analysis and a limited number of frequencies from the finite element results.
Page 29 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Table 5: Dryer #I Skirt Experimental Frequencies Figure 42 contains the skirt summation FRF, a comparison between test and FE results.
Figures 43 and 44 contain individual FRF comparisons between test and FE for the points measured on the 90° skirt panel, while Figure 45 contains additional results for points located off of the panel centerline. Figures 46 through 48 are similar results for the 2700 panel.
Figure 49 compares the 900 skirt panel and the 2700 skirt panel test results.
The first mode of the 900 skirt panels appears to be within 10% of the finite element prediction (first panel mode at ((
)) for FE). Figure 50 is the mode shape comparison of these modes.
The first mode of the 2700 skirt panel appears to be in the ((
)) range when the damping measurement FRFs are reviewed along with the modal information. For the FE Page 30 of 122
GENE-0000-0039-5860-01-NP, Revision 1 results, the range of frequencies is ((
)).
There are several more complex skirt modes in the ((
)) frequency range as well. In terms of visual comparison of modes based on a peak in the FE summation FRF, the test mode at ((
)) as shown in Figure 51.
For the panel modes, the 900 skirt panel appears to have an intermediate mode (not clearly 2nd bending, more of a cantilever) in the test results in the ((
)) range where the lower point is active normal to the panel surface, but the upper point is stationary.
In FE there is clearly a 2nd panel mode of the 900 skirt panel at ((
)) This 2nd panel mode is most clearly similar to test frequencies of ((
ii visually but most clearly similar to the ((
I] test mode in the FRF comparison as shown in Figures 43 and 44. Figures 52 and 53 contain the mode shape comparisons for both test frequencies. Objections have been raised to this comparison because the FE results show a lateral bending mode of this panel at approximately ((
)) and the test sensor locations would be at or near the node line for this mode; however the FRFs from both the modal and the damping measurements in Figures 43 and 44 show the clearest 2nd peak of the panel FRFs at ((
)) Figure 45 contains additional FRFs from points not on the centerline of the 90° skirt panel compared to the points on the centerline of the 90° skirt panel. In summary, there are several explanations for the appearance of the FRF and mode shapes but none of the explanations are conclusive with the data available for the 90° skirt panel:
I. The sensors are picking up the 2nd lateral panel bending mode even though the centerline sensors should be insensitive to this mode
- 2. The sensors are observing the 2nd vertical panel bending mode, and this mode is 20%
higher in frequency than predicted by FE
- 3. The 2nd lateral panel bending mode is geometrically rotated. The 900 skirt panel is the last panel welded into the structure. It is believed that it is very sensitive to the Page 31 of 122
GENE-0000-0039-5860-01-NP, Revision 1 weld sequence. Preliminary results from Dryer #2 with a larger number of sensors on this panel indicate that the 2nd lateral panel mode is not strictly side-to-side but rather upper-left to lower-right when observed by an observer facing the panel exterior.
This explanation coincides to some degree with explanation 1. as well.
Also, a comparison of the FRFs from the damping measurements with those from the 2700 skirt panel in Figure 49 shows the 2nd major mode of the 2700 skirt panel near 11 Based on the apparent discrepancy in frequency between test and FE results for the 900 skirt panel, a skirt sub-model study was performed to determine if the instrumentation pads, conduit and clamps affected this mode on the 90° skirt panel. The 900 skirt panel had this instrumentation while the 2700 skirt panel did not. The next section describes this study and its results.
8.2.1 Skirt Sub-Model Study A study of the skirt was performed to determine the effect of the instrumentation pads, clamps and conduit on the frequencies of the modes of the 90° skirt panel. These components were found to have no significant effect on the frequency or shapes of the panel modes of the 90° skirt panel. The frequency difference between base model and model with instrumentation pads, clamps and conduit was ((
))or less for all modes examined.
Table 6 contains the percentage change for the first 6 modes of the panel.
Details of the skirt sub-model for the study are as follows:
- Use only skirt section for which elements are available (do not use lower skirt super element)
- Constrain so that no displacements are possible at top and bottom of skirt section (full constraint - 3 directions/3 rotations)
- The sub-model frequencies were slightly different than those of the full model but alwvays within 10%
Page 32 of 122
GENE-0000-0039-5860-01-NP, Revision 1
- Instrumentation pads - stitched to skirt Conduit modeled as beam elements
- Clamps modeled as rigid connections (no mass addition, only stiffness - conservative assumption - will not reduce frequency)
- Plate thickness remains same from whole model
- Material Properties (reactor properties allowed frequency match to whole model modes more closely)
- 2.56e6 psi elastic modulus
- 7.4e-4 slugs/inA3 density
- Also checked using ambient properties - frequencies shifted up by approximately 4% to 4.5% but did not change effect of instrumentation pads, clamps and conduit Table 6: Frequency Change after addition of instrumentation to 90° skirt panel Appendix B contains finite element mode shapes from the Skirt Sub-Model analysis that was used to determine the effect of the instrumentation pads, clamps and conduit.
8.3 900 lHood Component Comparison The 90° hood test results were compared to finite element predictions. Figure 53 is the summation FRF for all 3 directions for this component - a comparison of test and FE, while Figure 54 is the summation FRF for this component in the X-direction for FE and normal to the surface for test (12.50 Euler angle between the FE and test results). A comparison of Figures 53 and 54 shows that the direction normal to the surface dominates the summation Page 33 of 122
GENE-0000-0039-5860-01-NP, Revision 1 FRF. As with the other components examined, the ff
)) clearly appears. Table 7 contains the test mode frequencies and comparisons with several FE mode frequencies.
The first flexible mode of the 900 hood appears to be in the ((
))
Visually these appear similar, but the component summation FRF more readily supports a
((
].
For the hood upper frequency range around ((
)) similar modes emerged in both the visual comparison and the MAC calculation. Figure 57 compares mode shapes at ((
)) solution. Figure 58 is a MAC matrix of this frequency range that shows a cluster of test modes being very similar to a larger cluster of FE modes.
Page 34 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Table 7: Dryer #1 900 Hood Experimental Frequencies
((I 8.4 2700 Hood Component Comparison The 270° hood test results were compared to finite element predictions at frequencies similar to the frequency regions used for the 900 hood. Table 8 contains the significant test modes and frequencies of similar FE modes. As with the other components evaluated the global mode at ((
flwas present. For the 270° hood, the first flexible mode, although one related to significant skirt activity, appears at ((
)). Figure 59 compares test and FE summation FRFs, and Figure 60 compares test and summation FRFs for the direction normal to the hood surface for test and in the X-direction for FE. Note that there is an angular difference of 12.50 between the test and FE directions in this case. Individual FRF comparisons between Test and FE are contained in Figures 61 and 62. Figure 63 shows a comparison of component-level mode shapes at ((
]Jwhile Figure 64 is the global comparison a those frequencies. This difference in frequency is approximately 10% although there is a cluster of
((
)) is similar to.
Page 35 of 122
GENE-0000-0039-5860-01-NP, Revision 1 As for the 90° hood, the other frequency range of interest for the 270° hood is the ((
)) Figure 65 contains the MAC matrix display around ((
Ilresults. The ((
)). It is most similar to a ((
component level and Figure 67 on a global level.
)), shown in Figure 66 on a Table 8: Dryer #1 2700 Hood Experimental Frequencies 1II A general observation of the summation FRFs for the 2700 hood is that the level of the FRF predicted by FE is lowver in the ((
)) This difference can be seen in Figures 59 and 60. The hood in FE may be too stiff or its connections may be too stiff as the general trend of the FE versus test comparison for the hood is that the FE levels are lower than the test levels.
8.5 Additional Observations This section contains additional observations from reviewing the test results in comparison with the analytical results.
Page 36 of 122
GENE-0000-0039-5860-01-NP, Revision 1
- 1. When comparing predicted FRFs to experimentally measured FRFs, responses on the ring or skirt from inputs to the hood show the FE levels ranging from similar to too high, indicating that the skirt may not be stiff enough in FE.
- 2. In many of the animations reviewed, the amplitude of the skirt compared to the amplitude of the hoods in mode shapes in FE is higher than the relative amplitudes for a similar test mode shape. The possibilities for this occurrence are the following:
- a. in FE, the relative stiffness of the hoods is too high with respect to the skirt.
- b. in FE, the mass of the hoods is too high with respect to the skirt.
- c. in FE, the skirt and the hood are too closely coupled and adjustments need to be made to the connections between the hoods, the support ring and the skirt
- 3. In terms of being a potential area for improvement, the skirt and its connection to the ring ranks as the highest areas because the 90° skirt panel 2nd vertical and lateral panel modes predicted by FE do not match the test results because there is either a frequency difference or a shape/orientation difference and because the FRF and animation trends (cross-FRFs with other components and relative amplitude in FE animations of skirt to other components compared with test animations) show that the skirt may generally not be modeled as stiff enough in the finite element analysis, or its connection to the ring may require review or there may be some aspect of the lower skirt super element affecting it more than desired for this comparison.
Page 37 of 122
GENE-0000-0039-5860-01-NP, Revision 1 9.0 Summary and Conclusions An experimental modal analysis was performed on new design Dryer # 1, the instrumented dryer intended for Quad Cities Unit #2 (QC2), and the results were compared to finite element analysis results on a frequency basis and on a mode shape basis. The finite element analysis included a modal analysis and, using those modes from the modal analysis, a mode superposition to obtain FRFs that match the input and response points of the test data. The test results showed many similarities to the finite element results. In terms of frequencies, the trend is that the ((
)) dominant modes.
As expected, the test frequency versus finite element frequency agreement decreased as frequency increased and the modes increased in complexity. The largest potential discrepancy seen between the model and the test is the 90° skirt panel with a significant test frequency being at ((
)) for a vertical 2nd panel bending mode and near 70 Hz for a lateral 2nd panel bending mode. There are several explanations for this discrepancy; however, with the number of sensors on this panel and their locations, a conclusive explanation cannot be provided. The 2700 skirt panel and the 900 and 2700 hoods showed better agreement for similar frequencies. In terms of FRF comparisons, the various components examined showed generally good agreement in trends and levels between summation FRFs for test and FE for that specific component.
The decrease in test versus finite element agreement as frequency increases is expected because, as the mode shapes become more complex with higher frequency, the finite element model results become more sensitive to several parameters such as boundary condition Page 38 of 122
GENE-0000-0039-5860-01-NP, Revision 1 details, element type and number of elements. The boundary conditions of the individual plates, the welds to other plates and beams, are not explicitly modeled in the finite element model. The wavelengths of the higher modes are shorter so more elements are needed to accurately represent the shapes; however, increasing the number of elements prohibitively increases computation error and adds to round-off error accumulation. The lower skirt super element, which includes hydrodynamic effects of the water on the skirt up to Low Water Level (LWL), may have some effect on the higher mode frequencies and test versus finite element differences as well.
Also, some of the perceived difference for higher frequency modes may be that too few transducers were used in the experimental modal analysis to characterize the mode shape well. The 90° skirt panel is an example where the number of transducers and their location do not appear to allow adequate resolution of some modes whose frequencies appear significantly in the frequency response measurements. Time constraints imposed upon the Dryer # I experimental modal analysis prevented more locations from being measured. As of the writing of this revision, changes were implemented in the Dryer #2 test plan to add locations to specific components and to reduce the number of conditions (water levels) for testing so that both the time constraint condition and the need for better spatial resolution of the measurement locations are satisfied.
In addition to determining the natural frequencies and mode shapes, the hammer test responses are used to experimentally determine damping values on the skirt and hood at low water level. The purpose of the experimentally determined damping values was to validate the damping values used in the stress prediction analyses. The damping measurement results showed a range of damping values which form a technical basis for arriving at appropriate damping values to be used for structural response analyses.
Page 39 of 122
GENE-0000-0039-5860-01-NP, Revision 1 To summarize:
- 1. The FE results contain many modes so generally a specific test mode will be similar to a group of FE modes in a specific frequency range
- 2. Specific lower frequency dominant modes of the skirt and the hoods appear to be similar in frequency and shape between Test and FE
- 3. FRF comparisons of component summation FRFs showed generally good agreement of levels and trends, with the only major exception being the ((
)) of the 270° hood
- 4. Individual FRFs often showed good agreement of levels and trends. An exception is the 90° skirt panel. It is discussed further in the next item.
- 5. The 90° skirt panel as installed appears to show the most difference from FE results in its 2nd panel modes. Further work is necessary to determine the source of the differences as initial checks of the model produced no insight. Additional analysis of Dryer #2 results is expected to provide information for this panel.
- 6. Plans for similar testing on the next dryer will include more detailed measurements/additional locations on the 90° skirt panel and the 2700 skirt panel and on both of the hoods.
From the above discussion on the good agreement of the first mode frequency comparisons and the reasonable agreement in the higher mode frequencies, it is concluded that the impact hammer test results verify that the finite element model used for dryer design calculations is sufficiently dynamically similar to the as-built dryer for engineering purposes. In addition the impact hammer test results also show that the ((
)) used in the stress prediction analyses are realistic and adequately conservative.
Page 40 of 122
GENE-0000-0039-5860-01-NP, Revision 1 10.0 References
- 1. Ramani, Venkat "Steam Dryer Hammer Test Specification." GE 26A6380. GENE.
San Jose, CA. April 2005.
- 2. "Laboratory Measurement of the Composite Vibration Damping Properties of Materials on a Supporting Steel Bar." Recommended Practice J 1637, Society of Automotive Engineers. Warrendale, PA. February, 1993.
- 3. Peeters, Bart; Guillame, Patrick; Van der Auweraer, Herman; Cauberghe, Bart; Verboven, Peter; Leuridan, Jan. "Automotive and Aerospace Applications of the PolyMAX Modal Parameter Estimation Method." Proceedings of International Modal Analysis Conference (IMAC) 22.
Dearborn,
Ml. January, 2004.
Page 41 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figures Page 42 of 122
GENE-0000-0039-5860-01-NP, Revision 1 I
I I
Fir 1
Figure 1: Finite Element Model Representation of Dryer l
is 1
11 2__
LE I
l l_
A_
I Figure 2: Photograph of Dryer # I, 0° Side Page 43 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 3: Photograph of Dryer #1, 1800 Side cv)
.-- do
-W I
Figure 4: Tripod Supports before Installation I
Page 44 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 5: Tripod Support Connection to Dryer I
I I
I I
Figure 6: Comparison of Measurements using Hammer Tips of different hardness, Radial Skirt Response to Radial Skirt Impact (Red curve is soft tip, Green curve is medium tip)
Page 45 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
I I
I I
I Figure 7: Comparison of Measurements using Hammer Tips of different hardness, Vertical Side Response to Skirt Impact (Red curve is soft tip, Green curve is medium tip)
I I
I I
Figure 8: Comparison of Measurements using Hammer Tips of different hardness, 900 Hood Response to Skirt Impact (Red curve is soft tip, Green curve is medium tip)
Page 46 of 122
GENE-OOOO-0039-5860-O1-NP, Revision 1 I
l
~I
-~I An I
- mn polints forf h 90 degree hood of dryer #I wheren lighturedmen sh.
hw te9 I
Figure 9:
Figure 10: Pattern ot measure degree hood response I Page 47 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure I 1: Response locations on upper right corner of 90 degree hood of dryer # 1.
gure I z. Kesponse jocations on lower right corner oi vu degree nood ot dryer Th1.
Page 48 of 122
I GENE-0000-0039-5860-01-NP, Revision 1 Figure 14: Response locations on lower left corner of 90 degree hood of dryer #1 with green circle indicating input location on 90 degree hood.
Page 49 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 15: Response location on side panel to tie right ot main panel o w.r.t. main panel).
he right ot main panel o w.r.t. main panel).
I Page 50 of 122 I
GENE-0000-0039-5860-01-NP, Revision 1 Figure 17: Response locations on side panel to the left of main panel of 90 degree hood of dryer #I (at angle w.r.t. main panel).
gure Ix: Full view ot 2/u degree hood of dryer t#i (before installation ot temporary sensors).
Page 51 of 122
Pattern o0 measui response Ic GEN E-0000-0039-5860-01-N P, Revision 1 i
I I
'I
~I respons pocaionts fon uppe rato 270 degree hood of dryer #1.weeylo es s20ho I
Page 52 of 122
GENE-0000-0039-5860-01-NP, Revision 1 I
r igure 22: Response locations on lower par OT Z/U 0degree hood o0 oryer ThI Page 53 of 122 I
GENE-0000-0039-5860-01-NP, Revision 1 Figure 23: Response locations on lower right corner of 270 degree hood of dryer #1 with green circle indicating input location on 270 degree hood.
Figure 24: Response locations on side panel to the left of main panel of 2 w.r.t. main panel).
degree hood of dryer# I (at angle Page 54 of 122
GENE-0000-0039-5860-01-NP, Revision 1 IC 1 Ignlt 01 llal pIIacl U1 L w.r.t. main panel).
degree hood of dryer # I (at angle Page 55 of 122
I GENE-0000-0039-5860-01-NP, Revision 1 I
I I
I Figure 27: Pattern of measurement points for the 0 degree side of dryer #1 where pink mesh is O degree side response locations, dark red is responses on top & aqua is responses on the ring.
Page 56 of 122
GENE-0000-0039-5860-01-NP, Revision 1 tions on left side of 0 degree side of dryer #1 with green circle indicating input Figure 28: Response local location on 0 degree side Figure 29: Kesponse locations on n Page 57 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 31: Response locations on upper right part Page 58 of 122 I
GENE-0000-0039-5860-01-NP, Revision 1 Figure 32: Response'e locations on 0 degree side of dryer #1.1 I
i d o d y r 1 b fo r n tl l a i o o f t e p r r y s n o r ).
_lI P a g 5 9 o f 1 2 degree s
GENE-0000-0039-5860-01-NP, Revision 1 Figure 34: Pattern of measurement points for the 180 degree side of dryer #1 where blue mesh is 180 hood response locations, dark red is responses on top & aqua is responses on the ring.
I I
I I
I I
I I
I qI Figure 35: Pattern of measurement points for the skirt & ring of dryer #1 where purple mesh is ring response locations, aqua mesh is the response locations on the ring, blue is responses on 180 degree side & pink is responses on 0 degree side.
Page 60 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 36: Pattern of measurement points for the skirt & ring of dryer #1 where the red mesh is response locations on top, green is the response locations on the panels inside the dryer, aqua is responses on the ring, blue is responses on 180 degree side & pink is responses on 0 degree side.
Figure 37: 900 Skirt Panel Input and Response Locations Page 61 of 122 l
I I
I
GENE-0000-0039-5860-01-NP, Revision 1 Figure 38: 2700 Skirt Panel Input and Response Locations I
I I
I I
Figure 39: Summation FRF for Top: Red - Test Page 62 of 122 I
(co
GENE-0000-0039-5860-01-NP, Revision 1
))
Figure 40: Summation FRF for 0° Side: Red - Test Figure 41: Summation FRF for Skirt: Red - FE, Green - Test 11 Page 63 of 122
GENE-0000-0039-5860-01-NP. Revision 1
((
Figure 42: FRF on 900 Skirt Panel, Upper Point: Red - FE, Green - Test, Blue - Damping Test (different input point from Green)
Page 64 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
11 Figure 43: FRF on 90° Skirt Panel, Lower Point: Red - FE, Green - Test, Blue - Damping Test (different input point from Green)
((
Figure 44: Additional FRFs on 900 Skirt Panel, Test Measurements: Red - Left of 707 as seen in Figure 31, Green - Right of Centerline, between 707 and 708 Vertically, Blue - Upper Middle, Pink - Lower Middle Page 65 of 122
GENE-0000-0039-5860-01-NP, Revision 1 1]
Figure 45: FRF on 2700 Skirt Panel, Upper Point: Red - FE, Green - Test Figure 46: FRF on 2700 Skirt Panel, Lower Point: Red - FE, Green - Test Page 66 of 122
GENE-0000-0039-5860-01-NP, Revision 1 11 Figure 47: FRF on 2700 Skirt Panel, from Damping Test: Red - Upper Point, Green - Lower Point
((t
))
Figure 48: FRFs on 900 and 2700 Skirt Panel from Damping Test: Red -Test - 2700 skirt, upper point, Green -
Test - 2700 skirt, lower point, Blue -Test - 900 skirt, upper point, Pink -Test - 900 skirt, lower point Page 67 of 122
GENE-0000-0039-5860-01-NP, Revision 1 t]
Figure 49: Mode Shapes for Skirt, Global View: Left - Test, ((
1 Page 68 of 122
GENE-0000-0039-5860-01-NP, Revision 1 11 Figure 50: Mode Shapes for Skirt, Local View: Left - Test, ((
1 Page 69 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
r M))
Figure 51: Mode Shapes for Skirt, Local View: Left - Test, ((
))
Page 70 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 52: Mode Shapes for Skirt, Local View: Left - Test, ((
))
Page 71 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
))
Figure 53: Summation FRF for 90'Hood: Red - FE, Green - Test
[1 Fo Figure 54: Summation FRF for 90'Hood, FE in X-direction, Test normal to surface: Red - FE, Green - Test Page 72 of 122
GENE-0000-0039-5860-01-NP, Revision 1
[I
))
Figure 55: MAC Matrix for 90° Hood, f[
Figure 56: Mode Shapes for 90° Hood, Global View: Left - Test, ((
))
Page 73 of 122
GENE-0000-0039-5860-01-NP, Revision 1 11 Figure 57: Mode Shapes for 900 Hood, Global View: Left - Test, ((
))
Figure 58: MAC Matrix for 900 Hood, ((
))
Page 74 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 59: Summation FRF for 2700 Hood: Red - FE, Green - Test
[I Figure 60: Summation FRF for 2700 Hood, FE - X-direction, Test - normal to surface (12.50 Difference): Red
- FE, Green - Test Page 75 of 122
GENE-0000-0039-5860-01-NP, Revision 1 I[I Figure 61: Individual FRFs for 2700 Hood, FE - Top, X-direction; Middle, Y-direction; Bottom, Z-direction:
Red - FE, Green - Test
((
Figure 62: Individual FRF for 2700 Hood, X-direction: Red - FE, Green - Test Page 76 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
))
Figure 63: Mode Shapes for 2700 Hood, Local View: Top - Test, j[
))
Page 77 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
Figure 64: Mode Shapes for 270° Hood, Global View: Top - Test, ((
Page 78 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure 65: MAC Matrix for 2700 Hood, f[
1]
Page 79 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
Figure 66: Mode Shapes for 270 Hood, Local View: Top - Test, ((
))
Page 80 of 122
GENE-0000-0039-5860-01-NP, Revision 1
))
Figure 67: Mode Shapes for 2700 Hood, Global View: Top - Test, ((
))
Page 81 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Attachment A: Steam Dryer Experimental Modal Analysis Test Plan Page 82 of 122
GENE-0000-0039-5860-01-NP, Revision 1 General Electric Steam Dryer Experimental Modal and Static Load Test
- 1. Equipment setup (PC, front end and cables not on dryer itself) 2 people can be done while other work is done on dryer
- 2. Equipment Setup (cables on dryer) 1 person (simultaneous with Task 3) can be done while other work is done on dryer as long as other work is not interfered with by cable routing
- 3. Geometry - node selection and preparation 1 person then 2 people after Task 2 is finished)
Marking of measurement locations for accelerometers and strain gages Cleaning of accelerometer and strain gage locations can be done while some other work is done on dryer (work that is not interfered with by cable routing or accelerometer position)
Accelerometer locations determined by PreTest and information/requests from Structural Analysis Group GE supposed to supply markers or paint pens allowed to be used on dryer
- 4. Attachment of initial accelerometer set and of strain gages 1 person or 2 people focus on accelerometers 1 person focuses on strain gages (if strain gage attachment and wiring is not finished during this specific task it will continue while accelerometers are moved later) can be done while some other work is done on dryer (work that is not interfered with by cable routing or accelerometer position)
End-to-end measurement chain checks Initial set of locations will focus on one outer hood but include locations spread out over dryer so that quality of excitation locations can be verified Adhesive to be used is M-Bond 200 (a super glue with an accelerator) as it has been approved for use on the dryer Solvent will be acetone as it is approved for use on the dryer Sandpaper is Silicon Carbide paper in strain gage kit
- 5. Frequency Response Function (FRF) Measurements/Hammer Impact Testing Perform driving point measurements at support locations Re-verify that all transducers are operational Verify acceptability of excitation locations (4 to 8 excitation locations, starting with 11)
Perform initial set of measurements dry (Current plan is to have 35 to 38 triaxial accelerometers, 1 impact hammer, and 1 single-axis driving point accelerometer in each set.
Perform all sets of measurements wet at high water levels Current plan is for the first 2 sets to be spread out over dryer and the 3Id set to focus on skirt. Also, first set contains single axis accelerometers for driving point reciprocity checks.
Perform the skirt focus set and a set with selected points from the other 2 sets at medium and low water levels Page 83 of 122
GENE-0000-0039-5860-01-NP, Revision 1 The following pictures show the proposed excitation locations as green circles or ovals and the response locations as yellow dots.
Figure A-I: 900 Side Response and Input Locations 00LL-
[
Figure A-2: 270° Side Response and Input Locations Page 84 of 122
GENE-0000-0039-5860-01-NP, Revision 1 l
l Figure A-4: O0 Side Response and Input Locations Page 85 of 122 co5
GENE-0000-0039-5860-01-NP, Revision 1 I
Figure A-5: Top Side Response and Input Locations I
Proposed plan for sets is included as Attachment A Wireframe model with node identification is included as Attachment B I
Data checks include but are not limited to:
- a. Review of input time records for input force and acceleration response during data acquisition
- b.
Review of Input Autopower
- c.
Review of FRF and coherence for each location (specific review of driving point)
- d.
Reciprocity between driving point locations
- e. Repeatability of driving point measurements among sets
- f.
After checks for each set are done, frequency content will be available for that set
- g.
As sets are performed, preliminary curve-fitting will be performed to check data
- h. Specific order and composition of sets still being finalized with input from structural analysis
- 6. Static load testing: 2 people, no other work done on dryer
- a.
2 people
- b. Apply 500 lb. load with hydraulic jack (assistance from JT Cullen or other required for providing backstop for jack), measure strain using 6 rosette strain gages
- c. Tentative Load application location - Outer hood near MSL nozzle location -
still being reviewed as are strain gage locations Page 86 of 122 coA
GENE-0000-0039-5860-01-NP, Revision 1
- 7. Final hammer impact measurements if necessary
- a. 2 or 3 people
- b. Repeat any points where data was inadequate for any reason or add additional points as determined by review during data acquisition
- c. Would be determined by analysis and processing done along way and while static load testing is performed
- 8. Removal of accelerometers and cables on dryer
- a. 2 or 3 people
- b. can be done while some other work is done on dryer (work that is not interfered with by cable routing or accelerometer position)
- c. Discussion will occur with GE about timing. If it makes sense, 1 person will continue processing data while 1 or 2 others remove accelerometers based on need to make dryer available.
- 9. Surface cleanup
- a. 2 or 3 people
- b. can be done while other work is done on dryer or while dryer is at plant (proposed in late January by GE)
- c. Discussion will occur with GE about timing. If it makes sense, 1 person will continue processing data while I or 2 others remove accelerometers based on need to make dryer available.
- 10. Equipment breakdown (PC, front end and cables not on dryer)
- a. 2 or 3 people
- b. can be done while other work is done on dryer Page 87 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Attachment A of Report Aft. A-Proposed Sets of Measurement Locations/Nodes Set Set Set A
B C
Impact Hammer Impact Hammer Impact Hammer Driving Point Driving Point Driving Point 301 302 810 814 811 813 809 303 812 804 806 807 606 808 805 801 802 803 601 110 604 603 704 404 605 734 608 607 731 610 609 206 207 208 205 508 215 507 506 505 713 504 503 714 114 119 117 111 412/121 115 118 103 107 104 101 109 108 106 401 705 105 102 707 402/120 403 708 112 113 711 406/218 725 726 405 728 729 702 735 732 502 205 210 202 209 204 203 202 216 214 217 211 212 201 213 407 722 408/219 7---
719 716 720 717 410 411 Repeats or Extra Repeats or Extra Repeats or Extra Repeats or Extra Triaxes replaced by single axis accels at driving points for reciprocity Page 88 of 122 CO-7
GENE-0000-0039-5860-01-NP, Revision 1 Attachment B of Report Att. A - Wireframe with Measurement Points/Node Numbers B
I coQ
Page 89 of 122
GENE-0000-0039-5860-01-NP, Revision 1 I
l 1
_lgue A-:'Sd _sd) Rin an Sit I
Page 90 of 122 C fi
I GENE-0000-0039-5860-01-NP, Revision 1 Figure A-1O: Top (top) and top view of 180° Side (sidc)
I I
Page 91 of 122 C I 0
GENE-0000-0039-5860-01-NP, Revision 1 Attachment B: Skirt Sub-Model Study Page 92 of 122
GENE-0000-0039-5860-01-NP, Revision 1 I
I Figure B-1: Location of Conduit and Instrumentation Pads on Skirt Panel I
Figure B-2 Left - Baseline, ((
))
Page 93 of 122 C11 i
GENE-0000-0039-5860-01-NP, Revision 1 Figure B-3: Left - Baseline, ((
Figure B-4: Left - Baseline, ((
]1
))
((
))
Page 94 of 122
GENE-0000-0039-5860-01-NP, Revision 1
[1 11 Figure B-5: Left - Baseline, ((
Figure B-6: Left - Baseline, ((
))
))
Page 95 of 122
GENE-0000-0039-5860-01-NP, Revision 1
))
Figure B-7: Left - Baseline, ((
))
Page 96 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Attachment C: Effect of Number of Averages Page 97 of 122
GENE-0000-0039-5860-01-NP, Revision 1 This attachment illustrates how the FRF measurement and coherence are affected when a different number of hammer impacts is used. Two measurements have been taken with the same excitation and response locations, one measurement with twenty hammer impacts and one measurement with five hammer impacts. The data presented in the following figures was measured on the 90 degree and 270 degree skirt panels of Dryer #2 (the dryer without permanent sensors intended for Quad Cities Unit # I (QC I)).
Specific measurements show some difference between 20 averages and 5 averages, particularly near ((
)). Theses differences are attributed to inconsistency in the hammer impact. In general, the frequency, shape, and phase of the FRF are consistent between 20 averages and 5 averages. The phase of the FRF is generally "cleaner" with 20 averages as well. The coherence shows some differences between 20 averages and 5 averages, but the differences appear to be specific to individual measurements and usually at low frequency - below ((
)). The other coherence differences are near
((
)) and correspond to those measurements with some difference in FRF amplitude that has been attributed to inconsistency in the hammer impact.
))
Figure E-l: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with S averages (green) for normal excitation at point skir:785 and normal response at point skir:785 on the 90 degree skirt panel of Dryer #2 Page 98 of 122
GENE-0000-0039-5860-01-NP, Revision 1
))
Figure E-2: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:753 on the 90 degree skirt panel of Dryer #2 Page 99 of 122
GENE-0000-0039-5860-01-NP, Revision 1
((
Figure E-3: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:707 on the 90 degree skirt panel of Dryer #2 Page 100 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure E-4: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:722 on the 270 degree skirt panel of Dryer #2 Page 101 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure E-5: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and normal response at point skir:762 on the 270 degree skirt panel of dryer 2
))
Page 102 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Figure E-6: Comparison of the FRF amplitude, FRF phase and coherence between a hammer impact measurement with 20 averages (red) and one with 5 averages (green) for normal excitation at point skir:785 and response at strain gage channel C at point skir:796 on the 90 degree skirt panel of Dryer #2 Page 103 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Attachment D: Acquisition Front-end channel assignment Page 104 of 122
GENE-0000-0039-5860-01-NP, Revision 1 AID CHANNEL ASSIGNMENT DRY SET A lAID CHASSIS:
2 LMS Scadas III frontends in master-slave setup AID AID A/D A/D CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID 1
21233 33 37277 X 65 16224 Z 97 16684 Y 2
34 37277 Y 66 15625_X 98 16684 Z 3
36614 X 35 37277 Z 67 15625 Y 99 16955 X 4
36614 Y 36 37278 X 68 15625 Z 100 16955 Y 5
36614 Z 37 37278 Y 69 17051 X 101 16955 Z 6
36616 X 38 37278 Z 70 17051 Y 102 20267 X 7
36616 Y 39 37279 X 71 17051 Z 103 20267 Y 8
36616 Z 40 37279 Y 72 17333_X 104 20267 Z 9
21836 41 37279 Z 73 17333 Y 105 220 X 10 21860 42 37280 X 74 17333 Z 106 264 Y 11 21852 43 37280 Y 75 27796 X 107 8751 Z 12 21855 44 37280 Z 76 27796 Y 108 8754 X 13 21857 45 37281 X 77 27796 Z 109 8777 Y 14 NA 46 37281 Y 78 27800 X 110 8778 Z 15 37126 X 47 37281 Z 79 27800 Y 111 8766 X 16 37126 Y 48 21869 80 27800 Z 112 8767 Y 17 37126 Z 49 22143 81 31423 X 113 8768 Z 18 37129 X 50 82 31423 Y 114 11156 X 19 37129 Y 51 37283 X 83 31423 Z 115 18026 Y 20 37129 Z 52 37283 Y 84 31425 X 116 19586 Z 21 37130 X 53 37283 Z 85 31425 Y 22 37130 Y 54 15514 X 86 31425 Z 23 37130 Z 55 15514 Y 87 31426 X 24 37131 X 56 15514 Z 88 31426 Y 25 37131 Y 57 15990 X 89 31426 Z 26 37131 Z 58 15990 Y 90 15314 X 27 37132 X 59 15990 Z 91 15314 Y 28 37132 Y 60 15591 X 92 15314 Z 29 37132 Z 61 15591 Y 93 15333 X 30 37275 X 62 15591 Z 94 15333 Y 31 37275 Y 63 16224 X 95 15333 Z 32 37275 Z 64 16224 Y 96 16684 X Data Collected By:
Signature:
Verified By:
Signature:
Date Performed:
Test Equip., Serial No. & Calibration Due Date:
LMS Scadas IlIl frontend (master)
LMS Scadas IlIl frontend (slave) 41011904 41001503 09-Mar-06 15-Feb-06 Page 105 of 122
GENE-0000-0039-5860-01-NP, Revision 1 lAID CHANNEL ASSIGNMENT DRY SET B LWL SET A HWL SET A LWL SET C lA/D CHASSIS:
2 LMS Scadas III frontends in master-slave setup A/D A/D A/D A/D CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID 1
21233 33 37277 X 65 16224 Z 97 16684 Y 2
37322 34 37277 Y 66 15625 X 98 16684 Z 3
36614 X 35 37277 Z 67 15625 Y 99 16955 X 4
36614 Y 36 37278 X 68 15625 Z 100 16955 Y 5
36614 Z 37 37278 Y 69 17051 X 101 16955 Z 6
36616 X 38 37278 Z 70 17051 Y 102 20267 X 7
36616 Y 39 37279 X 71 17051 Z 103 20267 Y 8
36616 Z 40 37279 Y 72 17333 X 104 20267 Z 9
37124 X 41 37279 Z 73 17333 Y 105 220 X 10 37124 Y 42 37280 X 74 17333 Z 106 264 Y 11 37124 Z 43 37280 Y 75 27796 X 107 8751 Z 12 37125 X 44 37280 Z 76 27796 Y 108 8754 X 13 37125 Y 45 37281 X 77 27796 Z 109 8777 Y 14 37125 Z 46 37281 Y 78 27800 X 110 8778 Z 15 37126 X 47 37281 Z 79 27800 Y 111 8766 X 16 37126 Y 48 37282 X 80 27800 Z 112 8767 Y 17 37126 Z 49 37282 Y 81 31423 X 113 8768 Z 18 37129 X 50 37282 Z 82 31423 Y 114 11156 X 19 37129 Y 51 37283 X 83 31423 Z 115 18026 Y 20 37129 Z 52 37283 Y 84 31425 X 116 19586 Z 21 37130 X 53 37283 Z 85 31425 Y 22 37130 Y 54 15514 X 86 31425 Z 23 37130 Z 55 15514 Y 87 31426 X 24 37131 X 56 15514 Z 88 31426 Y 25 37131 Y 57 15990 X 89 31426 Z 26 37131 Z 58 15990 Y 90 15314 X 27 37132 X 59 15990 Z 91 15314 Y 28 37132 Y 60 15591 X 92 15314 Z 29 37132 Z 61 15591 Y 93 15333 X 30 37275 X 62 15591 Z 94 15333 Y 31 37275 Y 63 16224 X 95 15333 Z 32 37275_Z 64 16224_Y 96 16684_X Data Collected By:
Signature:
Verified By:
Signature:
Date Performed:
Test Equip., Serial No. & Calibration Due Date:
LMS Scadas Ill frontend (master)
LMS Scadas Ill frontend (slave) 41011904 41001503 09-Mar-06 15-Feb-06 Page 106 of 122
GENE-0000-0039-5860-01-NP, Revision 1 AID CHANNEL ASSIGNMENT DRY SET C DRY SET D LWL SET D MWL SET D HWL SET D HWL SET B MWL SET B LWL SET B AID CHASSIS:
2 LMS Scadas IlIl frontends in master-slave setup AVD A/D AID A/D CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID 1
21233 33 37277 X 65 16224 Z 97 16684 Y 2
37322 34 37277 Y 66 15625 X 98 16684 Z 3
36614 X 35 37277 Z 67 15625 Y 99 16955 X 4
36614 Y 36 37278 X 68 15625 Z 100 16955 Y 5
36614 Z 37 37278 Y 69 17051 X 101 16955 Z 6
36616 X 38 37278 Z 70 17051 Y 102 20267 X 7
36616 Y 39 37279 X 71 17051 Z 103 20267 Y 8
36616 Z 40 37279 Y 72 17333 X 104 20267 Z 9
17998 41 37279 Z 73 17333 Y 105 220 X 10 18004 42 37280 X 74 17333 Z 106 264 Y 11 18011 43 37280 Y 75 27796 X 107 8751 Z 12 21870 44 37280 Z 76 27796 Y 108 8754 X 13 NA 45 37281 X 77 27796 Z 109 8777 Y 14 NA 46 37281 Y 78 27800 X 110 8778 Z 15 37126 X 47 37281 Z 79 27800 Y 111 8766 X 16 37126 Y 48 37282 X 80 27800 Z 112 8767 Y 17 37126 Z 49 37282 Y 81 31423 X 113 8768 Z 18 37129 X 50 37282 Z 82 31423 Y 114 11156 X 19 37129 Y 51 37283 X 83 31423 Z 115 18026 Y 20 37129 Z 52 37283 Y 84 31425 X 116 19586 Z 21 37130 X 53 37283 Z 85 31425 Y 22 37130 Y 54 15514 X 86 31425 Z 23 37130 Z 55 15514 Y 87 31426 X 24 37131 X 56 15514 Z 88 31426 Y 25 37131 Y 57 15990 X 89 31426 Z 26 37131 Z 58 15990 Y 90 15314 X 27 37132 X 59 15990 Z 91 15314 Y 28 37132 Y 60 15591 X 92 15314_Z 29 37132 Z 61 15591 Y 93 15333 X 30 37275 X 62 15591 Z 94 15333 Y 31 37275 Y 63 16224 X 95 15333 Z 32 37275 Z 64 16224 Y 96 16684_X Data Collected By:
Signature:
Verified By:
Signature:
Date Performed:
Test Equip., Serial No. & Calibration Due Date:
LMS Scadas Ill frontend (master)
LMS Scadas Ill frontend (slave) 41011904 41001503 09-Mar-06 15-Feb-06 Page 107 of 122
GENE-0000-0039-5860-01-NP, Revision 1
[AID CHANNEL ASSIGNMENT I
DRY SET E LWL SET E MWL SET E HWL SET E
[A/D CHASSIS:
2 LMS Scadas IiI frontends in master-slave setup A/D AID AID A/D CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID 1
Strain 33 37280 Y 2
Strain 34 37280 Z 3
Strain 35 NA 4
Strain 36 NA 5
Strain 6
Strain 7
Strain 8
Strain 9
Strain 10 Strain 11 Strain 12 Strain 13 NA 14 NA 15 NA 16 NA 17 NA 18 NA 19 NA 20 NA 21 21233 22 37322 23 27796 X 24 27796 Y 25 27796 Z 26 37132 X 27 37132 Y 28 37132 Z 29 37279 X 30 37279 Y 31 37279 Z.
32 37280 X Data Collected By:
Signature:
Verified By:
Signature:
Date Performed:
Test Equip., Serial No. & Calibration Due Date:
Page 108 of 122
GENE-0000-0039-5860-01-NP, Revision 1 AID CHANNEL ASSIGNMENT Static load test jAID CHASSIS:
2 LMS Scadas IlIl frontends in master-slave setup A/D A/D A/D A/D CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID CHANNEL SENSOR ID 1
Strain 2
Strain 3
Strain 4
Strain 5
Strain 6
Strain 7
Strain 8
Strain 9
NA 10 NA 11 NA 12 NA 13 NA 14 NA 15 NA 16 N A 17 NA 18 NA 19 NA 20 NA Data Collected By:
Signature:
Verified By:
Signature:
Date Performed:
Test Equip., Serial No. & Calibration Due Date:
Page 109 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Attachment E: Test Log Sheets Page 110 of 122
GENE-0000-0039-5860-01-NP, Revision 1 s
F 11 1.
tI i
I Dat Log Ip7 -,, A" rite of Tet Impact Ha mmer Location D
Date and Time of Acquisdtil m QmL%6ettngb
,3_ 2.
3 a
t V4.
.4-(RoYtT A7 c
':¢-a:
A7 01/
YT Ir If JV
?,,' 2SA2 4... 7 4
4 -
- 4.
':
A=..
r/# M
.*~
t u
e t
3.
.7i.
L
- v. -YZ.
0_Vol R 35 Ir
-r-..
__t I
A4,
/,
r - _
f f
{
w
.1
- -I U t -
L
- 7.WF t-t*
,. Le lil/.;Y3 I
Ie
{#~
a
-j
.7 7-' :W-:W-If I qrIe ct'/V.
m ws
.7 I.
- w.
-}
i. :.
- f I m
- r I
I wz%.
iLoe Daft Coected Br.
S
/
s 1
VerDfied B' Ale Signature-at P e one 9 0.-.
Test Euiup 'Serial No. & calibmsaion Du Dte -:
_C -
Page I I I of 122
GENE-0000-0039-5860-01-NP, Revision 1 Dab~o
- brye, /f' 4 L/
- tD :
T Ieof eefUnac Hannmer Locatton Data PMenarie Date and Time of Aolutsltbo 4d V8:X 4
fZ-ic l 4f yOfylZ 05-vt~yO-v S.~t2-~
£'
aif
(
MilijM r
3ZZZ _
i'r.
r -re lee 3
1 01 1
I.,'
11 II I
IL I
I 0I I.
i Data coulected ar. S, erA: Aka iaLe..ki Signatum jJ--
Verffied Br.
c r4 aur Test Wquip, SerWa No. & Cailbratlo Due Dae-Page 112 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Tidle of Test Impact Hammer Location DaaF____,Daea__imf cuito
- k.. y z
f1 s#I t I
.-4a:-£9'Weyy
-p,~-l
-__-XI
~.
0~
/031
, ~ V 'f 2 l Data colectedw.
e.es nt a
Verified 0y fSI-atre:C- -
Date Perdofnne
£i5-oqO:3 Test Equip, Serial No. &Calibation Due Date:
Page 113 of 122
GENE-0000-0039-5860-01-NP, Revision 1 I~
br,[4 r, Ad, :
. ;,
A,
AM Tite oi Test Impact Hammer Location l___Fite_
lDate and Time of Acquisition
__h i
p Sel 6-
/
A*~~~-f.-
As.
_~O AL.4y-t i
1
-7 V t tI 4
-
w1
Xllt
,. :::
- rS72,-V
--~
k~- ~"S: -X 4-
-4 lq dl-Irl I-4 21
.Data Co.led El:
B C
C tse(
ernfied By:
sty f
i S-i natu:e:.,5a.
Date Performed:
0 5> Vo 3 Test Eguip, Seril No. & CaIbiation Due Datar I
Page 114of 122
GENE-0000-0039-5860-01-NP, Revision 1 I
tSe, I -:- :-r
._r_ of Tes tnpact Hamner Locadon Dat Flimme Date and Tie of Acquo W l
.%e L.t s
A-1 I
.. quip, S iN.
D
=
c-I-
_ _ _ __~
t
~
i:
0
.a
-at Cofete Br L@z p&R
/h'u°s 5rE
- S
... fied. B y.:
Akr:f Z t. -X.
- ea E i..Seia No.,,airto u
at.
.~.
=
.1
- i. -
t S.
E Page 115 of 122
GENE-0000-0039-5860-01-NP, Revision 1 Dataog ~
rye-M J
______eg trmpact Hammer Location Daa
~
Date and Time, of Acquisition AiA r ht
. y i.
r r
Zma %
DY____3______
.T-
$)'c.,
r
-10 ler-Veffied By:
dy4 Sigmature" DatPerfomroe&
-0571',- /M's-oryo leit Equip, SerWa Mo.'& Calibration Due Date:L.I I
I I
Page 116 of 122
GENE-0000-0039-5860-01-NP, Revision 1
- Ago, Z>^','
.a-e
.ct MataColeid
-CSigmure Verifed By:
A/:C Sign ture: 7:
0 Date P
,..o.
ed:
Test Equip Serial No. & Calibration De Dat:
Page 117 of 122
GENE-0000-0039-5860-01-NP, Revision 1 rAt ImpaCt Hrmer Location naeDate and Iimne of Acquisition Gah2 Somi~sj 4~#
Z
____pro~o~1
_ _ _ _ _ _ _ _ _ _ _ _ _ _fy 2/0 z i l
-pop:
4O-Irl 0/0 M__________________________
A.
4 4 ~ -_
- I2~X D-a ouct_
y:Se_
_6i1.
Verfied Byv 0.'-Ilz --
sianatufa:
I in Date Pformed:
-y lest mgulp.;enal n'o- & LCalibrmjn uue Lame Page 118 of 122
GENE-0000-0039-5860-01-NP, Revision 1
_______9!
Impact Hammer Location =
a,~Date and Tenamo Acquisition Gerwettins)
V
.4 0$DV1O
.. /f7d
-22.JI 7.9S.
.frpt :
-Y
_ __ _ _ __ _ _ __ _ ___In_ _ __ _ _ __ _ _ __ _
Daft Collected By. -
Ase /~L1~(Signature:
Verified By-A'Ii
~
- -~--
Signaturie:
Data Perormed:
9t 2 0
'Test Equip., Serial No. & Calbibrtion Du iae it
.. :-: :.. . -
. :1 -
c +
I. a L
i I - -. I . I, -
Page 119 of 122
GENE-0000-0039-5860-01-NP, Revision 1 IDXX~
~~0 L0 1b L-M
-~
- XKO r P%-
.- 'I)
.4 T
h__l__ bnipact Hammer Location d
iui Date and Time of Acquisition CSci.,g_
D
-Jr ro ~
~
t 7-S
~;r 0
S i;
Wa.}YV, t
)S fi:-,a r
fi:
- }
.j
)
3
-4 I i
rk.
^!7 3:-
r_
_ft_
f l/
t 3 Z w
t
- et/
.at C.lce B.
OD:
/4 5/4 IEt af.Sign-a-ture:D Vie B
W
/
S ur47
~-~: ~
Vi 3 e
o~aS=
f/
z Y /S7^
g_
.aeefond
-.5 qc C
-. -t!ts......... s8& _...j Tes- 't Equip. Se -ria Co. & Cabrtio Due
. ~ ~
~
~ ~ ~ ~ ~ ~
- Y
.y
. e-t:f
~-;
- es l
.at
.e
.ne.
s
)q c
1 Tes Eqip da No
.airto'u tg ax Page 120 of 122
GENE-0000-0039-5860-01-NP, Revision 1 roatLo.-l ~)rZv- ^
4e(
-,--r S wmpfecfll-wunrLocation.
et-lie afo DCate and TimveofAcquisition - a~~t Ij
_ I C.
C.
_L.*Co_
G-6,0 j",o"-
,:Ip!
-7' Y9 I,-
-~
I I
V.,
DaaC.
ee B.-
b t 3-
.t 4 ii-ntre
________________.d
_ft
_ed
__7i..-,
Page 121 of 122
1 GENE-0000-0039-5860-01-NP, Revision 1 IData LOg i~/r e
1o(X14
_ i pac Ha mr ef etig T__11__sT ImpaCtHammer LocatiOn lam taFly
.Date and Tame of Acquisition -
,M 107Vf
-ZZ.
_e 4
-IR.~z qI;7"5z:-wt o
D r
z 0
.ll e
t:
a 3
lb
-3..r zr
~~~~~_.
~~~~~t :..
_,r iI,
.I
.I I
Da~t. Cotlected %
ki Ar.6.
'/h 4,e
..; a t~igrature:,
Z a
a
-r
=
.r VerliB..y- -
I -
Slanature:
I eto Ep.,p t No.
C...on Due D. t..:
'Test Equip. Serlat No.
Calltxtlon DueOie:
i J.-
.I i' '
r.
f I
I.
'e.,. : :. I A........-....
t
. r ww_
.v Page 122 of 122