STANDARDS October 2017 CHECK AND CALIBRATION PROCEDURES FOR FATIGUE TEST BENCHES OF WHEEL E S 3.29 Page 1/13 PROCÉDURES DE CONTRÔLE ET CALIBRAGE DE FATIGUE BANCS D'ESSAIS DE ROUE PRÜFUNG UND KALIBRIERUNG VERFAHREN FATIGUE PRÜFSTÄNDE RADER Changes: 10/2017 First release No part of this standard may be reproduced in any form without the prior written permission of EUWA References: ISO 7500-1 ES3.23 ASTM E 1049-85 Content: 1 - SCOPE... 2 2 - FIELD OF APPLICATION... 2 3 - REFERENCES... 2 4 - TEST BENCH DEFINITION AND LOAD CHAIN... 3 5 - METHODS AND CALIBRATION PROCEDURES... 3 6 - GENERAL DESCRIPTION AND PRELIMINARY SETUP... 4 7 - EQUIPMENTS... 9 8 - TEST BENCH MEASURE DESCRIPTION... 9 9 - TEST BENCHES ACCEPTABILITY LIMITS... 10 10 - CALIBRATION... 11 11 - RECCOMENDATIONS... 11 ANNEX1... 13 Main changes compared to the last issue: EUWA - Association of European Wheel Manufacturers ES - 3.29
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 2 1 - SCOPE This EUWA Standard specifies methods to check and calibrate test benches used for the fatigue wheel life evaluation. 2 - FIELD OF APPLICATION The wheel fatigue performance is a necessary requirement to be fulfilled to grant a safe working condition. Thus, the fatigue test benches must be maintained always in the best efficiency and shall work correctly. Test benches maintenance, periodic checking and calibration are important to assure their best performances. The following standard shows four procedures describing methods and rules to check and calibrate fatigue test benches. The described procedures, in this standard, can be used to check all test bench types, but some of them better fit to specific test bench. Other procedures may be used but accuracy level specified in this norm shall be respected. This norm is intended to be used for the following test benches. The test bench types are subdivided by load application and secondly by bench structure: Dynamic Cornering fatigue test benches - rotating eccentric mass - rotating wheel Dynamic Rolling fatigue test benches - radial load - radial and axial load Biaxial Road Simulator 3 - REFERENCES The following methods are based on the ISO 7500-1 force-proving instruments used for the verification of uniaxial testing machine even if the application field of this norm has been extended to the wheel test benches. The generic guidelines about test execution, measures analysis and the CLASS definition can be also adopted for this ES norm s purpose.
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 3 4 - TEST BENCH DEFINITION AND LOAD CHAIN The fatigue test bench is a device, with a: - user interface (PC or generic I/O interface) - controller - active loading devices (electric or hydraulic system,...) - loading measuring system (load cells,...) - load frame - wheel - load feedback dynamic control (PID,...) All the test bench elements shall be taken into consideration during the machine calibration process but the most important are for sure the loading measuring system. Fig.1 Load chain schema 5 - METHODS AND CALIBRATION PROCEDURES Two are the methods about checking the test benches: Check: stand alone methods able to define the proper working conditions of the test bench (procedure as per ISO 7500-1) Comparison: method to compare two test benches or the same test bench in different period The following procedures must be carried out only on bench in good functional conditions in terms of mechanical, hydraulic and electric systems, tools and cleanliness. The test benches checking can be carried out follow different procedures. Everyone is useful to calibrate or check the complete test bench or part of it. In some case, more than one of following procedures must be carried out for a complete checking of the test bench. The different procedures are: 1. Static offline Load Cell calibration 2. Static Test Bench calibration 3. Dynamic bench calibration by dynamometric wheel 4. Dynamic bench checking by stress analyses comparison
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 4 In the table1 are shown the main characteristics which identify general performances and main limitations of different procedures. Procedure ID #1 #2 #3 #4 Check Type Static Static Dynamic Dynamic Load Chain Checked (figg.1-3-5-7) Partial Complete Complete Complete Complexity of Execution LOW LOW HIGH HIGH Execution Costs LOW LOW HIGH LOW Required skill LOW LOW HIGH HIGH Correct error cause by Dynamic Controller Device stability Correct error cause by dynamic effects (inertia, friction, clearances,...) Correct error cause by mechanical set-up, misalignment Correct error cause by sensor mass and stiffness NO YES YES YES NO NO YES YES YES NO YES High risk YES YES YES High risk NO NO test bench calibration YES YES YES NO tab.1 main performances and limitations of test bench checking procedures 6 - GENERAL DESCRIPTION AND PRELIMINARY SETUP The selection of the procedure is mainly dependent on the test bench. It is advisable to orient the choice on dynamic calibration procedure where the all loading chain is considered 6.1 Static offline Load Cell calibration In figure 2 the dash line delimits the part of load chain checked. Keeping the load cell electronic system connected to the test bench, dismount the cell and assemble it aligned (series connection) with the Master Load Cell in an external structure. The figure 3 shows an example of series connection of both load cells.
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 5 This is the simplest and cheapest bench check. All mechanical and dynamic effects depending on bench structure, passive masses, control system, etc. are not considered. Figure 2: Static offline Load Cell Calibration Checked Load Chain Figure 3: Static offline Load Cell Calibration Example of loading frame 6.2 Static Test Bench calibration This procedure checks statically all the load chain. In figure 4 the dashed line delimits the part of load chain checked. The test bench calibration foresees a Master Load Cell installation where the load is applied during the test execution (see figure 5).
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 6 This is a simple and cheap bench check. During the Master Load Cell installation paying attention to the precise alignment between load application axle and Master Load Cell axle. The dynamic effects and inertia masses are not considered in this procedure. The effects of test bench geometry and active load devices are considered. Figure 4: Static Test Bench Calibration Checked Load Chain Figure 5: Static Test Bench Calibration Example of loading equipment 6.3 Dynamic bench calibration by dynamometric wheel This procedure checks dynamically all the load chain. In figure 6 the dashed line delimits the part of load chain checked. The figure 7 shows two examples of Dynamometric Wheel application by multiaxial force sensor. The sensor assembled in the dynamometric wheel must replace the wheel in dimensions and allow to be fixed to the test bench with standard fixing system. The test bench must work in normal way without any limitation when the dynamometric wheel is installed. The equipment used for this procedure is expensive and high skilled technicians are needed to carry out the bench check.
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 7 The bench checking is performed in real loading and working conditions, then all dynamic effects are taking into account even if the sensor mass and stiffness may influence the measures. Usually this procedure is coupled with a preliminary static calibration as above described Figure 6: Dynamic Calibration by Dynamometric Sensor Figure 7: Dynamic Check: Dynamometric Sensor example of dummy wheels 6.4 Dynamic bench checking by stress analyses comparison This procedure is usually used for checking the biaxial road simulators. All the load chain is checked dynamically. In figure 8 the dashed line delimits the part of load chain checked. Differently from the others procedures, the calibration is not applicable because there isn t a direct and absolute relationship between applied load and
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 8 measured stress, furthermore the wheel cannot be calibrated itself. This procedure is carried out to compare two test benches keeping one as an Average Behaving Test Bench. This method consists in execute two stress analyses, one on the Average Behaving Bench and one on the bench to be checked. It is important that wheel, electronic devices (DAQ, signal transmission cables/telemetry system, amplifier) and the postprocessing software are the same for both measurements. The acquisition can be carried out using complex load history (see standard ES3.23); in this case the stress parameter to be compare shall be the RFS (Required Fatigue Strength) (see figure 9 and Annex1). In case of constant loading condition the comparison can be done evaluating directly the stress amplitude. The equipment used for this procedure is cheaper but high skilled technicians are needed to carry out the stress analysis. The bench checking is performed in real loading and working conditions, then all dynamic effects are considered. Usually this procedure is coupled with a preliminary static calibration as above described. Figure 8: Dynamic Check Stress Analysis Comparison Figure 9: Dynamic Check Stress Analysis Comparison
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 9 7 - EQUIPMENTS In table 2 are summarised for each procedure the main sensor needed to carry out the test bench checking. PROCEDURE ID SENSOR CLASS OTHER EQUIPMENTS #1 Master Load Cell 0.5 External loading structure #2 Master Load Cell 0.5 Fixing system to the bench precise aligned to the load axes #3 Load Sensor 1 Adapters to connect the dyn. wheel to the machine #4 Average Behaving Test Wheel with Strain - Gauges (*) Bench or Stress Analysis as reference Tab.2 Procedure sensor (*) it is recommended to use minimum 4 strain gauges. The strain gauges shall be installed in different wheel areas (rim and disc). Every sensor must be adequate to check the test bench in all the working load range. 8 - TEST BENCH MEASURE DESCRIPTION Here follows the summary of main steps which shall be performed for the test bench checking. 8.1 Thermal compensation: Thermal compensation is requested to avoid measure thermal drifting due to electronic, mechanical components and tire (if present) heating. Recommendation is to switch on the electronic system 30 minutes before to carry out the measure, and for dynamic checks, run the test with a generic loading conditions for the same time. 8.2 Hysteresis compensation: In case of static checks also the Load Cells hysteresis must be removed by three loading cycles from zero to maximum Load Cell capacity before to carry out the measures. 8.3 ZERO Systems Setting Load sensors shall be set to zero while completed unloaded.
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 10 8.4 Measure Execution: Perform three load cycles from zero to maximum load subdividing the load range in minimum five steps. For every step and from both master and checked system, the measures must be saved in the calibration report. After every cycle remove completely the load and save also the deviation from zero. Between two cycles wait at least 30 seconds. Repeat the measurement. (Ref. ISO 7500-1) For load systems with tensile and compression working condition, the two situations shall be verified independently. Only for procedure #4 (Test Benches Comparison), the load condition shall be representative of a complete and most commonly use Load History. The measure execution consists to carry out two stress analyses: one in the Average Behaving Bench and one in the Bench to be checked. The stress measured are synthesize in Stress Spectrum on which RFS is evaluated. During the dynamic data acquisition the sample rate must guarantee at least 36 points every wheel or load revolution. 8.5 Measure Analysis Evaluate Accuracy, Repeatability and percentage deviations from the Master System. In the dynamic measures, the data acquisition is also affected from the stability of the signal during the time. So, in stationary loading situations the load variation shall be measured. The load scatter shall be evaluated as the maximum measured load minus the minimum measured load in stationary loading situation. 9 - TEST BENCHES ACCEPTABILITY LIMITS In table 3 are summarised the acceptability limits in terms of accuracy, repeatability and scatter for each procedure. METHODS CHECK COMPARISON LIMITS PROCEDURE ID #1 #2 #3 #4 ACCURACY CLASS1 CLASS2 CLASS3 RFS% < 5% (*) FREQUENCY CHK 1 year (*) the comparison has been performed by RFS parameter Tab.3 Acceptability Limits
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 11 The CLASS definitions are in according with ISO 7500-1. If the % Accuracy deviation is bigger than defined limit as per the reference class (see tab.3), the calibration procedure must be carried out. If the % Accuracy deviation is equal or less than defined limit as per the reference class (see tab.3), no actions are requested. In the dynamic procedures, in all the data acquisition, the Load Scatter must be within the 1% of the full load scale of the test bench. For all the procedures, the frequency is fixed to 1 year. About the Check Methods, the same procedure not necessarily has to be carried out every year but at least one of the three available checks can be used. 10 - CALIBRATION The calibration depends of the type of test bench. Major steps of the calibration are the following: - Set up the test bench as per one of the procedures #1, #2 or #3. - Perform thermal and hysteresis compensation as per par. 8.1 and 8.2. - Set ZERO in both systems: checked and master system (par. 8.3). - Impose a loading condition to the test bench well representing the most standard working condition. - Adjust the test bench gain to match the master system indication. After calibration, the test bench checking shall be repeated. Calibration may be necessary in case of: 1. specified time period has elapsed 2. extraordinary event occurrence (e.g. shock, vibration, tire explosion, mechanical failures, software update, electronic failure,...) 3. anomalous test results. 4. instrument of the load chain has been repaired or modified 11 - RECOMMENDATIONS To be sure of the proper of the test benches functionality, the working conditions must be periodically checked following one of the above procedures described. A repeatability feedback about the working condition will be guarantee if one of the procedures will take as master. The test bench checking can be carried out following more than one procedure, but one of them can be considered the best one depending mainly from the bench structure.
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 12 In the table 4 are represented the recommended checking procedures for the main test benches. METHODS CHECK COMPARISON PROCEDURE ID TEST BENCH Dynamic cornering with fixed wheel Dynamic cornering with rotating wheel #1 STATIC OFF LINE LOAD CELL CHECK #2 STATIC TEST BENCH CHECK #3 DYNAMIC BENCH CHECK By DYNA WHEEL #4 DYNAMIC BENCH CHECK by STRESS ANALYSIS X (O) (O) (O) X (O) Dynamic Rolling X (O) Biaxial Road Simulator (O) X X X standard checking procedure (O) optional procedure Tab.4 Recommended checking procedures
ES - 3.29 Check and Calibration Procedures for Fatigue Test Benches of Wheel Page 13 ANNEX1 RFS Definition Given a structure subjected to a variable loading, each point of the structure will be subjected to a stress time history. By stress cycle counting, the stress time history is reduced to sequence of pairs number of cycles stress semi amplitude: ni - Sai. This representation is called the stress amplitude spectrum; The fatigue material behaviour is characterized by bilinear S-N curve as follow: Slope of the portion over the fatigue limit K = - Log(N)/ Log(Sa) Slope of the portion below the fatigue limit K1 = - Log(N)/ Log(Sa) Number of cycles at fatigue limit Nfl Fatigue limit stress Fl For each pair of stress amplitude spectrum, the damage di is calculated as: di=ni/ni. Where Ni is the number of cycles relevant to Sai on the S-N curve. The damage sum D is the sum of all di. The target damage DT represent the limit at which the material will fail. RFS is the local fatigue limit Fl of the material that leads to the target damage DT. How to proceed in RFS calculation Here follows the description of RFS calculation procedure fixing all parameter needed to have comparable results. RFS value can be calculated only in an iterative way. These conditions shall be used for homogeneous metallic materials only. Stress time history: the sampling rate shall be capable to describe each stress cycle with 36 points minimum. In case stress data are derived from strain measurements, use 210000 MPa for steel Young modulus for steel and 70000 MPa for aluminium alloy. Cycle counting: the cycle counting procedure shall comply to Rain Flow Counting ASTM E 1049-85. S-N curve parameters: use K=5.5, K1=10, Nfl= 2*10 6. Target damage: DT shall be 0.5 for steel and 1.0 for other metal s alloys. Stress amplitude: Sai shall be stored with precision of 0.1% of maximum stress cycle or better. RFS iteration convergence criteria: final iteration of RFS calculation shall show a variation less than 0.001% of RFS value.