1、 WORLDWIDE ENGINEERING STANDARDS Test Procedure GMW15243 Structural Durability Time - History Data Validation Guidelines Copyright 2013 General Motors Company All Rights Reserved November 2013 Page 1 of 37 1 Scope Note: Nothing in this standard supercedes applicable laws and regulations. Note: In th
2、e event of conflict between the English and domestic language, the English language shall take precedence. This standard provides a list of checks and processes to perform on time history data files to ensure data integrity and a common format for data used in laboratory structural durability simula
3、tions. 1.1 Purpose. The purpose of this standard is to verify the validity of data sent to the structural durability laboratory and ensure the data conform to the standards defined for the structural durability laboratories. This data is time history data from either physically measured Road Load Da
4、ta Acquisitions (RLDA) or analytical “virtual RLDA” (vRLDA). Standards described in this document: Standard Channel Measurement Conventions (Channel Nomenclature, Coordinate System, Positive Direction, Static Offsets), Data File Format and Data Quality Checks. Note: Throughout this procedure, “RLDA”
5、 means both RLDA and vRLDA unless noted otherwise. 1.2 Foreword. Not applicable. 1.3 Applicability. This procedure is applicable to all structural durability related time history data. It will not guarantee perfect data every time, but using these suggested data quality checks will eliminate the mos
6、t significant errors. 2 References Note: Only the latest approved standards are applicable unless otherwise specified. 2.1 External Standards/Specifications. None 2.2 GM Standards/Specifications. GMW15219 GMW15268 GMW15242 GMW15445 2.3 Additional References. Global Structural Durability Lab Expert T
7、eam (GSDLET) web site: https:/ 3 Resources 3.1 Facilities. 3.1.1 Calibration. The test facilities and equipment shall be in good working order and shall have a valid calibration label. 3.1.2 Alternatives. Alternative test facilities and equipment may also be used. However, all measuring variables as
8、 specified in this standard shall be determined correctly with respect to their physical definition. 3.2 Equipment. Not applicable. 3.3 Test Vehicle/Test Piece. Not applicable. 3.4 Test Time. Not applicable. 3.5 Test Required Information. Not applicable. 3.6 Personnel/Skills. Not applicable. Copyrig
9、ht General Motors Company Provided by IHS under license with General Motors CompanyNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-GM WORLDWIDE ENGINEERING STANDARDS GMW15243 Copyright 2013 General Motors Company All Rights Reserved November 2013 Page 2 of 37 4 Pro
10、cedure 4.1 Preparation. Data that will be used in the structural durability laboratories is expected to incorporate the following standards. These standards include: Remote Parameter Control (RPC) format files. The raw RLDA data should be received with a sample rate exceeding 1000 Hz to allow for ad
11、equate peak resolution for analyses. The lab will need to resample the data to either 409.6 or 512 points/s sample rate for compatibility with the controller being used. Frame size of 2048 points. Group size of 2048 or 4096 points. Adherence to channel naming standards (see Appendix A), file naming
12、standards (see GMW15219), and engineering units (Appendix C) with correct coordinate systems and polarity (Appendix D) is strongly encouraged but is not required. Correctly driven (RLDA) or calculated (vRLDA). 4.1.1 These data should be verified to adhere to these guidelines using the following proc
13、edure both before delivery to the laboratories (RLDA engineer or Computer Aided Engineering (CAE) engineer) and after delivery, the laboratory engineer should use the same procedure as a means of becoming familiar with the data. 4.2 Conditions. Not applicable. 4.2.1 Environmental Conditions. Not app
14、licable. 4.2.2 Test Conditions. Deviations from the requirements of this standard shall have been agreed upon. Such requirements shall be specified on component drawings, test certificates, reports, etc. 4.3 Instructions. 4.3.1 Data Integrity Check. Data issues fall into one of two categories: corre
15、ctable and non-correctable. When using software programs to correct the data, take care to ensure that the engineering values are not accidentally changed in the time history. The user must consult an experienced RLDA or laboratory engineer if they are not familiar with the suggested corrective acti
16、on. The following is a list of potential data issues, and suggested corrective actions. Some of the suggested actions are to rerecord (recalculate) the data. This may not be practical if the RLDA vehicle is already torn down, in which case the laboratory engineer must find the best repair of the dat
17、a possible, using good engineering judgment and possibly with the aid of a CAE engineer. Several of these checks require visual inspection. Examples are included in Appendix B. 4.3.2 File Naming Convention. For data shareability, it is recommended that the schedule names adhere to the standard namin
18、g convention. See GMW15219 for the file naming convention. Corrective Action: Rename to standard naming convention. 4.3.3 Uniform Channels. All the schedules must have the same channel arrangement (number and order). Corrective Action: Discuss with the RLDA Engineer to understand, correct and/or rec
19、ollect the data as required. 4.3.4 Channel Naming Convention. Ensure that the standard channel naming convention is used. See Appendix A, Figure A1 and Table A1 for the channel naming convention. Corrective Action: Rename to standard naming convention. For example, use RPC Pro Setup to create and co
20、rrect the appropriate header and apply to all files using the RPC Pro Header Modification tool. Warning: If a given channel has a non-common full scale, this action will change the engineering data. 4.3.5 Engineering Units. Ensure that the standard engineering units are used. Special characters (e.g
21、., m/s2, , ) may not be correctly recognized by RPC software. See Figure B1 for an example. Appendix C, Table C1 contains a table with the preferred units for most common measurements. Corrective Action: Correct the engineering units. For example, use RPC Pro Setup to create and correct the appropri
22、ate header and apply to all files using RPC Pro Header Modification tool. Warning: If a given channel has a non-common full scale, this action will change the engineering data. 4.3.6 Full Scales. Ensure that the individual channel full scales are the same across all the schedules. Corrective Action:
23、 If the full scale is not the same across all schedules, a common full scale must be chosen and applied. For example, use the RPC Pro Header Modification Tool. Copyright General Motors Company Provided by IHS under license with General Motors CompanyNot for ResaleNo reproduction or networking permit
24、ted without license from IHS-,-,-GM WORLDWIDE ENGINEERING STANDARDS GMW15243 Copyright 2013 General Motors Company All Rights Reserved November 2013 Page 3 of 37 Warning: The value chosen for a given channel must be greater than the largest absolute peak observed across all schedules for that channe
25、l, or data will be clipped. 4.3.7 Correct Channels in Dataset. Ensure that all needed channels are included in data. GMW15268 provides the MINIMUM channels that should be included. Corrective Action: Either re-extract data or rerecord so that correct, required channels are included. 4.3.8 Sample Rat
26、e. Check the sample rate. All GM RLDA groups typically record data with a sample rate of at least 1000 Hz (1 kHz). The standard sample rate for RPC laboratory data is either exactly 409.6 points/s or 512 points/s so that the physical test controller can use the data. RLDA data will need to be resamp
27、led to a laboratory-acceptable data rate. The sample rate must be the same for all files. Sample rates which are slightly off (i.e., 409.599 points/s) indicate a data processing issue and must be corrected before using the data. Note: Some data acquisitions may require a higher sample rate to adequa
28、tely characterize the content of the data. Corrective Action: Correct the data. For example, this may require the use of RPC Pro Resample tool to filter the data (to prevent aliasing), and adjust the sample rate. 4.3.9 Frame Size. Check the frame size. Data received with a sample rate of 409.6 point
29、s/s should have a frame size of 2048 points, providing a frequency resolution of 0.2 Hz. For adequate frequency resolution, the laboratory typically needs at least 0.25 Hz or smaller resolution. Data with inconsistent frame sizes needs to be corrected. Corrective Action: Adjust the data to the corre
30、ct (2048) frame size. For example, use RPC Pro Setup to create and correct the appropriate header and apply to all files using RPC Pro Header Modification tool. Warning: If a given channel has a non-common full scale, this action will change the engineering data. 4.3.10 Points per Group. The points/
31、group must be an integer value when divided by the frame size. The points/group must be 2048 or 4096, when the frame size is 2048. Corrective Action: Correct to the correct points per group. For example, use RPC Pro Setup to create and correct the appropriate header and apply to all files using RPC
32、Pro Header Modification tool. Warning: If a given channel has a non-common full scale, this action will change the engineering data. 4.3.11 Recorded Maneuvers. Ensure that the modules were driven per test procedure instructions. Check for the proper time length (usually documented in the driving pro
33、cedure) and internal events. For example, in the Full Brakes Snubs durability event (GMW15556) there should be a total of five (5) forward/reverse Full Brake Snub maneuvers (15 forward snubs and 5 reverse snubs). If the user is not familiar with the proper number of maneuvers that should have been r
34、ecorded, read the applicable test procedure. See Figure B2 for an example. If available, reference a “speed template” for the road surface in question. Also ensure that the filenames accurately represent the module recorded. Corrective Action: Resolve discrepancies by adjusting the repeat counts (if
35、 possible) or recollecting the data. An alternative method is to duplicate the existing laps to create a file with the proper number of laps. 4.3.12 Polarity. Visually check the data polarity for consistency with the RLDA Workbook. During known test maneuvers, the time history plots must be consiste
36、nt with the polarity stated in the RLDA Workbook. See Figure B3 for an example. Appendix D contains the coordinate system standards. Corrective Action: Correct the RLDA Workbook or invert the data. For example, use RPC Pro Time History Math tool. If the data are corrected by inverting, correct the s
37、ource of the inversion in the lab, or the same issue will re-emerge during laboratory testing if the same instrumentation is used on the road and in the laboratory. 4.3.13 Symmetry. Check for symmetry between similar channels. For example, the left front Fore/Aft peak load should be similar to the r
38、ight front Fore/Aft peak load. This check should be done between all channels where similar magnitudes would be expected. See Figure B4 for an example. Corrective Action: Discuss with the RLDA Engineer to understand, correct and/or recollect the data as required. 4.3.14 Saturation. Ensure that the d
39、ata do not reach the full scale value. Look for flat tops in the data (due to reaching a previously used full scale value). Figure B5 gives an example. Corrective Action: Discuss with the RLDA Engineer to understand, correct and/or recollect the data as required. Copyright General Motors Company Pro
40、vided by IHS under license with General Motors CompanyNot for ResaleNo reproduction or networking permitted without license from IHS-,-,-GM WORLDWIDE ENGINEERING STANDARDS GMW15243 Copyright 2013 General Motors Company All Rights Reserved November 2013 Page 4 of 37 4.3.15 Auto-Spectral Density (ASD)
41、 Plots. Look for unusual spikes or frequency roll-offs in the ASD plots. Symmetrical channels should have similar ASDs. Also, integrally-related channels, such as acceleration, displacement and velocity should have the appropriate relation, e.g., velocity is the integral of acceleration so the veloc
42、ity ASD should approximately roll off as the inverse of frequency from the acceleration ASD. See Figure B6 for an example of symmetric channels. Corrective Action: Discuss with the RLDA Engineer to understand, correct and/or recollect the data as required. 4.3.16 Signal Resolution. Confirm that ther
43、e is an acceptable signal resolution. When closely viewing the data, the bit resolution (stair stepping) should not be visible and the peak loads should be a reasonable percentage of the full scale value. Ideally the peak loads should be about 80% of the full scale value. Using less than 10% of the
44、full scale value is unacceptable. See Figure B7 for an example. Corrective Action: Discuss with the RLDA Engineer to understand, correct and/or recollect the data as required. 4.3.17 Signal to Noise Ratio. Ideally, the RLDA engineer will have run a “noise test” where the individual transducers are u
45、n-powered but data are recorded. This test will allow review of data for Electro-Motive Force (EMF)-type noise issues. A less effective option is to review the before-run zeroes before the vehicle is moving. See Figure B8 for an example. Corrective Action: Discuss with the RLDA Engineer to understan
46、d, correct and/or recollect the data as required. 4.3.18 Wheel Force Transducer (WFT) Ripple. Although not often seen on newer WFTs, verify that the ripple on the WFT channels is acceptable (Fore/Aft and vertical) in all files. The maximum acceptable value is 600N peak-peak. Ripple is best viewed du
47、ring the smoothest section of each module. See Figure B9 for an example. Corrective Action: Discuss with the RLDA Engineer and correct as required. Options include: recollect the data, replace a bad WFT and recollect the data, or post-process the data with appropriate software tools. These options s
48、hould be executed by the RLDA Engineer. 4.3.19 Signal Loss. Look for any signal loss. For example, when looking at a file where there is no expected change in magnitude, ensure that the signal doesnt suddenly get larger or smaller. See Figure B10 for an example. Corrective Action: Discuss with the R
49、LDA Engineer to understand, correct and/or recollect the data as required. 4.3.20 Switched Channels. Check for switched channels. For example, a lateral load channel should not look like a Fore/Aft load channel, or vice-versa. See Figure B11 for an example. Corrective Action: Discuss with the RLDA Engineer to understand, correct and/or recollect the data as required. Correction may include modifying header to correct channe
