Gauge R&R (GR&R), as defined through Measurement Systems Analysis (MSA), has long been an accepted technique to validate the consistency of a measurement system (gauge), whether it is in the lab or used in end of line (EOL) applications. The analysis process, defined by Automotive Industry Action Group (AIAG), is typically applied to an attribute that is inherently stable and repeatable, and a measurement process that is just as robust and single valued, e.g. measuring the diameter of a part with a caliper. In the world of NVH, the parts are stable and the measurement system is robust; however, the measured results are not as repeatable as desired due to the nature of the system. In the NVH realm, the results are typically multi-value, e.g. an order slice vs. RPM. How is a traditionally single value process applied to a multi-value result, pick 1 point? An additional challenge to the traditional MSA analysis is that it requires a linear amplitude relationship of the measured result.This paper will highlight the application of an alternate AIAG referenced process to perform a GR&R of NVH results. Benefits noted of the alternate method will show how it can be used with multi-value results and use statistical math that is compatible with NVH data and analysis. Also covered is a review of the amplitude relationship of NVH data and a description of the various results available for interpreting the validation of the EOL stand. At the conclusion, a short case study will be reviewed highlighting the topics covered.
Validating NVH EOL Measurements using GR&R
Sae Technical Papers
7th International Styrian Noise, Vibration & Harshness Congress: The European Automotive Noise Conference ; 2012
2012-06-13
Conference paper
English
Field Test: Validating the User Study Measurements
Springer Verlag | 2016
|VALIDATING PIV MEASUREMENTS IN SUPERSONIC JETS USING SHADOWGRAPH OPTICAL IMAGING
British Library Conference Proceedings | 2011
|Validating the Adaptability of Travel Time Reliability Measurements using Probe Data
Transportation Research Record | 2019
|