Understanding Transient Vibration

Definition: What is Transient Vibration?

Transient vibration refers to a temporary, short-duration vibration that occurs when a machine’s operating state is changing. It is a non-steady-state event. The most common examples of transient vibration events are machine startups and shutdowns (coast-downs).

Unlike steady-state vibration, which is measured while a machine is running at a constant speed and load, transient vibration analysis focuses on capturing and understanding the dynamic response of the machine as it passes through a range of speeds or conditions.

Why is Transient Vibration Analysis Important?

Analyzing transient vibration is crucial for understanding the fundamental dynamic properties of a rotor and its support structure. It is the primary method for identifying a machine’s critical speeds.

During a startup or shutdown, the machine’s speed sweeps through a wide range. As the rotational speed (1X) passes through any of the machine’s natural frequencies, a resonance condition is created. This causes a significant amplification of the vibration amplitude. By capturing the vibration data during this speed sweep, engineers can precisely identify the frequencies at which these resonances occur.

This information is vital for:

• Machine Design and Acceptance Testing: Verifying that the machine’s critical speeds are not too close to its normal operating speed.
• Diagnostics: A change in the location of a critical speed over time can indicate a developing structural problem, such as a cracked shaft or loose foundation.
• Flexible Rotor Balancing: Balancing flexible rotors requires understanding the rotor’s response at its critical speeds, and this data is acquired during transient runs.

Specialized Analysis Plots

Because the speed is constantly changing, a standard FFT spectrum is not sufficient for analyzing transient vibration. The data is typically displayed on specialized plots that show how the vibration changes with respect to speed (RPM):

• Bode Plot: This is the most common plot for transient analysis. It displays the 1X-filtered vibration amplitude and phase on two separate graphs, both plotted against the machine’s speed. Resonance is clearly identified by a peak in amplitude and an associated 180-degree shift in phase.
• Nyquist (Polar) Plot: This plot combines the 1X amplitude and phase into a single polar plot. Resonance is identified as a characteristic loop on the graph.
• Waterfall/Cascade Plot: This is a 3D plot that stacks multiple FFT spectra together as the speed changes, creating a “waterfall” effect. It is excellent for visualizing the response of all frequency components (not just 1X) during the transient event.

Data Acquisition Requirements

Capturing transient vibration data requires specific instrumentation and setup:

• Multi-Channel Analyzer: A data acquisition system capable of simultaneously sampling multiple channels of vibration and speed data is needed.
• Tachometer/Keyphasor: A once-per-revolution speed/phase reference signal is absolutely mandatory. The analyzer uses this signal to track the machine’s speed and to enable the phase measurements required for Bode and Nyquist plots.
• Sufficient Memory and Processing Speed: The analyzer must be able to record a continuous stream of data for the duration of the startup or shutdown, which can sometimes last for several minutes on very large machines.

← Back to Main Index