Kevin J. Farrell, Principal Engineer, Computational Simulation & Validation

NVH (noise, vibration, and harshness) is an increasingly important design metric for many commercial products—from vacuum cleaners and dishwashers, to lawnmowers and chain saws, to cars, trucks, and aircraft. Consumers demand a product that performs well but with low "observables."

The first two qualities are measureable. A sound power meter can measure the radiated sound from a source, and a motion pickup or accelerometer on the product housing can quantify the vibration amplitude.

But how can we determine how harsh a sound is?

In psychoacoustic studies, a group of people (i.e., a jury) is asked to describe their perception of a sound. That perception can vary among each person in that group. What sounds "good" to my ear does not necessarily sound the same to yours.

Many years ago, I was in a work group that was tasked with quieting a handheld vacuum cleaner. We made several modifications to the rotor and flow path in the base unit that reduced the noise and vibrations significantly. We then asked a jury of vacuum users to choose their preferred model among the base unit and several modified versions.

Interestingly, the jury did not pick the quietest vacuum, because the diminished sound seemed to indicate to them a diminished ability or capacity for picking up dirt. Similar perceptions can apply to other household equipment. For example, a quiet lawn mower may not cut high grass without stalling, and a quiet dishwasher will not remove baked-on food residues like a louder one. More sound “means” better performance.

So what would a jury of operators in the process industry consider harsh?

Operators are pleased if their exchangers transfer heat effectively and limit noise and vibration.

Since our founding, HTRI has often served as a clearinghouse for information about vibration, actively increasing our understanding of flow-induced vibration phenomena in shell-and-tube heat exchangers, developing evaluation techniques for new designs, and coming up with remediation methods for field problems. We have collaborated in important research activities, chaired relevant symposia in engineering societies, and hosted technical meetings on the subject. For many years, HTRI has worked with many of the important industrial and academic researchers in the field.

The fruits of our research and ongoing review of the subject manifest in our vibration screening tool in Xist® and our finite element tube analysis in Xvib®. Xist checks for excessive tube vibration as well as for the potential for acoustic resonance.

Acoustic vibration results from the complex interaction among the tubes, the cross flow, and the volume of gas enveloped by the shell. The bundle geometry affects the magnitude and frequency of the pressure fluctuations from the wakes of the cross flow that can excite a standing acoustic wave—typically across the diameter of the shell as shown in Figure 1. The speed of sound in the fluid media (a thermodynamic quantity) determines the natural frequency of the gas volume. If the two frequencies match, the exchanger will let you know—sometimes.

Figure 1. Most common acoustic resonance in a shell-and-tube heat exchanger

Acoustic vibration remains enigmatic due to its apparent sensitivity to nuances in the vapor flow pattern and profile. These observations have been used to spark remediation techniques like surface roughening, tube removal, and variations in tube spacing, but only the deresonating baffle has shown the requisite reliability in silencing an acoustic resonance.

Unfortunately, the many years of wind tunnel research have not produced a reliable damping model or drain for acoustic vibration energy in the exchanger. Consequently, prediction methods like the Eisinger-Sullivan [1] map predict our data well, as shown in Figure 2, but they also tend to be overly conservative and predict many false positives, sometimes leading the heat exchanger designer to make some changes to the design that are not warranted. The coordinates on the map are functions of the crossflow pressure drop, Mach number, and acoustic natural frequency.

Figure 2. Distinct separation of vibration (resonant) cases from no vibration (non-resonant) cases using map of dimensionless acoustic particle velocity vs. dimensionless input energy

In addition to the wakes of flows over bluff structures like the tubes, a process plant itself offers many other sources of acoustic vibration:

  • Fans and pumps. These generate pressure pulses at the shaft rotation rate as well as the blade passing frequency (shaft rotation rate times the number of rotor blades) and harmonics.
  • Valves, pipes, and various flow components. These sources can contribute to turbulent flow, cavitation events, and shock waves, and can be discrete or continuous.
  • Furnace-duct systems. These can excite standing acoustic waves during the addition or removal of heat.

The jury in the process industry is unanimous: they desire minimal sound and vibration. No psychoacoustic analysis is necessary. The not-so-harsh reality is that few shell-and-tube heat exchangers experience failure due to flow-induced vibration!


  1. F. L. Eisinger and R. E. Sullivan, Further evidence for acoustic resonance in full size steam generator and tubular heat exchanger tube banks, J. Pressure Vessel Tech. 132(4), 044501-1 – 044501-4 (2010).