Figure 1: Typical equipment needed to perform specialized acceptance testing includes a laptop computer with data analysis software, a data acquisition analyzer with the appropriate channel count, associated cables, the required number of tri-axial accelerometers, uni-axial accelerometers, and an instrumented impact hammer. Note that impact testing is not required for standards such as ANSI/Hydraulic Institute 9.6.4, but it is often specified by the Engineer of Record (EOR), especially if a separation margin has been included as part of the analysis effort.
Two-Part Test Plan – Operational and Non-Operational
Operational Testing
Once temporary sensors, including triaxial accelerometers, are mounted in the correct locations (typically on each bearing housing), vibration data is measured and recorded from a single operating machine. In accordance with ANSI/Hydraulic Institute Standard 9.6.4, vibration measurements should also be taken on adjacent non-operational machines to confirm that any sympathetic vibration levels remain within acceptable limits.
For variable-speed machinery, data is recorded at several speed points as specified. Testing individual machines independently is standard practice; however, in many cases, data is also collected with multiple machines operating simultaneously to assess system interactions.
Key terms such as natural frequency, resonance, amplification factor, and separation margin are explained in this one-page document.
The recorded vibration data is analyzed off-site, and a report is prepared. Figure 2 shows an example of a triaxial accelerometer mounted on a mixer, while Figure 3 illustrates how the maximum overall vibration is summarized in a table against specified acceptance limits. This data serves as the basis for determining whether the vibration acceptance test passes or fails.
Vibration orientation is reported as parallel, perpendicular, and vertical (or axial), with the reference frame defined relative to the draft tube inlet for mixers or the discharge piping for pumps, blowers, compressors, and fans.
Figure 2: Sensor locations for VAT operational testing.
Figure 3. Overall vibration at specified locations reported. Overall data is used to determine if a machine passes or fails a vibration acceptance test.
Figure 4 shows the overall vibration at a key location (the motor outboard bearing in this example) plotted over time in the vertical, perpendicular, and axial (inline) directions. The maximum amplitudes presented in Figure 3 were taken from this type of plot, where the highest vibration occurred at 9:37:30.
Figure 5 presents the corresponding Fast Fourier Transform (FFT) of the maximum amplitude data. The FFT is a computational method that converts time-domain data into the frequency domain. Both time-domain and frequency-domain plots are valuable tools, especially when troubleshooting. For this reason, FFT plots are often specified in acceptance criteria to support diagnosis in case overall vibration exceeds allowable limits.
In this case, the natural frequency mode shapes observed near 10 Hz will be discussed later in the blog.
Figure 4: A plot of overall vibration at the motor outboard bearing (in this example) in the vertical, perpendicular, and axial (or inline) directions over time. The maximum amplitudes shown in Figure 3 and the FFT plot in Figure 5 occur at 9:37:30.
Figure 5: An FFT plot of the motor outboard bearing's maximum amplitude data, which occurred at 9:37:30. The FFT plot, vibration spectra, or filtered readings are useful if troubleshooting is required. Therefore, FFT plots are generally specified in case the machinery exceeds the maximum allowable overall vibration and some diagnosis is needed.
Non-Operational (with Limited Operational) Testing
Experimental Modal Analysis (EMA), or impact testing, is performed to identify the natural frequencies of the mixer system up to approximately 2x the vane pass frequency. When conducted properly, EMA also provides valuable information on damping and amplification factors for each natural frequency. The procedure involves impacting the system with an instrumented hammer at a limited number of accessible locations on the motor, mixer, and nearby support structure, while measuring the response with accelerometers and a vibration analyzer. Most of the testing is done while the machinery is not operational, with some supplemental data collected during operation to assess how operating conditions influence the natural frequencies.
A table summarizing the natural frequencies and their separation margins from key excitation frequencies is a key deliverable. Mixers can be more complex because the excitation frequencies include both the motor running speed (1750 rpm or 29.17 Hz in this example) and the mixer running speed (285 rpm or 4.75 Hz), in addition to the Mixer Vane Pass Frequency (VPF) of 14.25 Hz and the Belt Drive Frequency (BDF).
Another important deliverable is the set of Frequency Response Function (FRF) plots. The FRF shows the degree of separation between natural frequencies and excitation sources. The amount of damping or the Amplification Factor can also be derived from the FRF. Any natural frequencies with an Amplification Factor below 3.33 are considered critically damped and should not become resonant, even if they are near a primary excitation frequency, such as one times the running speed. In practice, the Amplification Factor can be thought of as the steepness of the mode shape curve (illustrated by the red lines in Figure 6). Since FRF data can be complex, interpreting it accurately requires both technical expertise and field experience. Analysts must distinguish between genuine resonance concerns and patterns that may appear problematic but are, in fact, benign.
Figure 6: Example FRF from an impact test on a motor driving a mixer. The FRF clearly shows a natural frequency of the mixer motor's outboard bearing housing, located about 43 Hz (20.5 Hz / 14.25 Hz) above the mixer vane pass frequency, while also maintaining an acceptable separation margin of 42% below the motor running speed. Experience also tells us that the mixer vane pass frequency excitation will have minimal or no impact on motor modes, even if the separation margin criteria are not met.
Deliverables
A properly specified Vibration Acceptance Test (VAT), which includes Experimental Modal Analysis (impact testing), should result in a clear report outlining the condition of the machinery and offering specific recommendations if the vibration acceptance criteria are not met.
The report should also include:
- Overall vibration data (in velocity, inches per second RMS) at each measurement point, including notes on the region of operation.
- FFT spectrum analysis for each measurement point, interpreted by an experienced engineer, with commentary on issues such as imbalance, misalignment, bearing concerns, or resonance.
- Impact test results, including:
- A summary table of natural frequencies up to two times the vane-pass frequency, noting their separation margin from the specified excitation frequencies (such as running speed).
- Frequency Response Function (FRF) plots for each identified natural frequency.
- Animations (videos) illustrating significant mode shapes of interest.
This three-part blog series has outlined the complete vibration risk reduction process: Part 1 covered the importance of clear specification wording, Part 2 described the analysis phase and expected deliverables, and Part 3 demonstrated how a VAT with impact testing confirms separation margins near critical natural frequencies.
Please contact MSI if you would like more information or assistance in reducing the risk of post-installation rotating machinery system vibration problems.
