The Squeaky Wheel
Measurement Systems for Noise, Vibration, and Harshness
Measurement Systems for Noise, Vibration, and Harshness
In the spring of 2003, my father traded up to a twin-engine motor yacht. As with most used vehicles, engine trouble usually comes with the territory. Hopefully, the seemingly never-ending minor repairs cost less than the stratospheric price tag of brand new. By mid-summer, both engines were purring, and I was fortunate to pilot a few excursions around the Chesapeake Bay. Centered on the instrument cluster between the duplicate gauges for the port and starboard engines is a curious needle gauge titled “synchronization.” When both engines are running at identical rpm, the gauge reads a value of zero and a white status light is illuminated. As one engine increases or decreases in rpm relative to the other, the needle sways toward the faster engine and the status light is extinguished. I can appreciate the fact that the boat travels straightest when both props are spinning at the same rate. However, the resulting resonate vibrations powered by the beat frequencies between the engines are intolerable. We soon decided the white synchronization light is a safety feature warning the captain of imminent damage to any dental work unable to abide violent chattering.
Photo by Fidel Fernando on Unsplash
Individually, each engine is tuned to factory specifications and the hull mounts are in good condition. An analysis of the individual components reveals all is in order. The unwanted hull vibrations are an emergent property appearing only when the entire system is functioning. In this case, a “looking in” analytical view does not reveal the problems so obvious from the “looking out” systems view. Engineers can spend any amount of time minimizing the vibration signature of individual components that nevertheless generate unexpected and unacceptable vibrations when operating in concert.
This situation is not novel. Researchers in the life sciences can gain fundamental understanding of structure through dissection; however, a full appreciation of function can only be gleaned by observing the entire system while operating in its environment. Thankfully, psychiatrists do not maintain a collection of scalpels behind their couch. The equipment, methods, and analysis routines used for the study of systemic vibrations fall under the general topic of noise, vibration and harshness (NVH). In the transportation industry, NVH testing is not only prevalent for locating defective components, but also for minimizing occupant discomfort and complying with environmental noise legislation.
Noise levels can be acquired using sensitive microphones while vibration is frequently measured by accelerometers, optical deflection or optical interferometry. Multiple sensors are used simultaneously to capture an accurate representation of the system in operation. Analysis of the results often falls back to the analytical “separation of variables” view. As all of the noise and vibration signals are obtained simultaneously, they must be separated in frequency space, a process facilitated by the discrete Fast Fourier Transform (FFT). By observing the power contained at each frequency, the performance of individual components and their interactions with the entire system can be studied. Unfortunately, the FFT generates a frequency spectrum that is averaged over all time and important intermittent noises and vibrations are not apparent. One way to approach this weakness is to divide the entire data set into short, partially overlapping segments in time and to perform an FFT on these segments. This “short-time” Fourier Transform (STFT) permits the discovery of transient frequencies before they are averaged into the entire recorded data set. A second related method is to filter the data into different frequency bands before segmenting them in time and then performing a wavelet transform (WT) on the segments. The STFT analyzes transient signals with constant time and frequency resolution while the WT provides high frequency resolution at low frequencies and high time resolution at high frequencies. Both analysis methods are part of the ever-growing Joint Time-Frequency Analysis (JTFA) algorithms available to the NVH community.
Increasingly, the enabling technologies in data acquisition are moving beyond correct transducer selection and calibration and into innovative methods of data analysis. In acknowledgement of the evolving systems view, this column has expanded its title to “Data Acquisition and Analysis” as of January 2004. If you have a clever analysis method you would like to share please let me know.
This material originally appeared as a Contributed Editorial in Scientific Computing and Instrumentation 20:12 November 2003, pg. 16.
William L. Weaver is an Associate Professor in the Department of Integrated Science, Business, and Technology at La Salle University in Philadelphia, PA USA. He holds a B.S. Degree with Double Majors in Chemistry and Physics and earned his Ph.D. in Analytical Chemistry with expertise in Ultrafast LASER Spectroscopy. He teaches, writes, and speaks on the application of Systems Thinking to the development of New Products and Innovation.