Skill Guide

Domain knowledge in rotating machinery, vibration signatures, and failure-mode taxonomy

The specialized knowledge required to interpret the mechanical condition, operational anomalies, and root causes of failure in rotating equipment (e.g., turbines, pumps, motors, compressors) by analyzing their vibration data against a structured taxonomy of failure modes.

This skill is critical for maximizing asset uptime and reliability, directly impacting production yield and reducing costly unplanned downtime. It enables predictive maintenance strategies, shifting organizations from reactive fixes to proactive, data-driven asset management, thereby extending equipment life and optimizing capital expenditure.

1 Careers

1 Categories

8.5 Avg Demand

20% Avg AI Risk

How to Learn Domain knowledge in rotating machinery, vibration signatures, and failure-mode taxonomy

Focus on: 1. Foundational mechanical concepts (mass unbalance, misalignment, bearing defect frequencies, resonance). 2. Standard vibration terminology (amplitude, frequency, phase, waveform, FFT spectrum). 3. Key industry standards (ISO 10816 for vibration severity, ISO 13373 for condition monitoring).

Practice: Correlate vibration spectra with physical components using actual machine train schematics. Analyze time waveform and enveloped spectra for early bearing/gear fault detection. Common mistake: Misinterpreting spectral peaks without considering machine operating context (speed, load, temperature) and ancillary sensor data (process, acoustic emission).

Master: Complex fault interactions (e.g., looseness-induced misalignment, foundation resonance with unbalance), advanced signal processing (wavelet analysis, order tracking), and building vibration-based failure mode and effects analysis (FMEA) trees. Strategic focus: Developing predictive models for remaining useful life (RUL) and aligning monitoring programs with business risk and criticality ranking.

Practice Projects

Beginner

Project

Vibration Signature Analysis of a Single-Stage Centrifugal Pump

Scenario

You are provided with baseline vibration data (overall level, FFT spectra from motor and pump bearings) for a pump operating at 1780 RPM. The overall vibration level has increased by 40% over 3 months.

How to Execute

1. Download the data file (CSV) and plot the FFT spectrum. 2. Identify the 1X (1780 CPM) and 2X (3560 CPM) peaks and their harmonic amplitudes. 3. Compare the peak amplitudes and patterns against ISO 10816 severity charts for a Class III machine. 4. Draft a preliminary diagnostic report attributing the increase to probable misalignment (high 2X) with supporting severity assessment.

Intermediate

Case Study/Exercise

Root Cause Analysis of a Gas Turbine Bearing Failure

Scenario

A 25MW gas turbine experienced a catastrophic #2 bearing failure. Pre-failure vibration data shows a classic 'bathtub' curve of increasing broadband vibration, with emerging sidebands around inner race defect frequency (BPFI) peaks in the high-frequency envelope. Oil analysis shows high particle counts.

How to Execute

1. Map the vibration timeline to operational events (starts, load changes). 2. Analyze the sideband spacing (shaft speed) to confirm inner race defect localization. 3. Correlate the envelope spectrum trend with the oil debris trend to build a unified degradation story. 4. Construct a fault tree diagram leading to root cause (e.g., insufficient lube flow, resulting in micro-pitting and spalling on the inner race).

Advanced

Project

Develop a Predictive Model for Centrifugal Compressor Fouling

Scenario

You are tasked with building a model to predict fouling in a multi-stage centrifugal compressor, which causes a shift in vibration signature from a pure unbalance pattern to a complex pattern involving sub-synchronous instability (oil whirl/whip) as clearances degrade.

How to Execute

1. Integrate vibration (orbit, spectrum cascade), process (flow, pressure), and temperature data into a time-series database. 2. Use order tracking and wavelet transforms to isolate the sub-synchronous energy from the synchronous component. 3. Develop a multivariate model (e.g., using a physics-informed neural network) where fouling is a latent variable affecting both vibration and efficiency. 4. Define alarm thresholds based on model residuals and validate against a historical failure database.

Tools & Frameworks

Software & Analysis Platforms

Vibration Analysis Software (e.g., SKF @ptitude Analyst, Emerson AMS Machinery Health, Bently Nevada System 1)Data Science Tools (Python with SciPy, NumPy, and libraries like 'vibes' or 'pyro' for signal processing)Statistical Process Control (SPC) and Machine Learning Platforms (e.g., Python's scikit-learn, MATLAB)

Use dedicated vibration software for daily monitoring and diagnosis. Use Python/MATLAB for advanced signal processing, custom algorithm development (e.g., envelope detection, demodulation), and building predictive models when commercial software limits are reached.

Diagnostic Frameworks & Standards

ISO 20816 / 10816 (Vibration Severity)ISO 13373 (Condition Monitoring)Failure Mode Identification Charts (e.g., by Technical Associates of Charlotte)Cause-Effect (Ishikawa) and Fault Tree Analysis (FTA)

Apply ISO standards for benchmarking and severity grading. Use published charts (e.g., 'The Vibration Analysis Handbook') as reference for spectral pattern identification. Employ FTA and Ishikawa for structured root cause analysis of failures, ensuring all possible contributing factors (machine, process, maintenance) are considered.

Interview Questions

Answer Strategy

The candidate must demonstrate they understand the difference between a static (amplitude-only) and a dynamic (phase-varying) problem. Sample answer: 'This indicates a resonant condition, not simple unbalance. The stable phase at the motor and unstable phase at the gearbox input suggest the gearbox shaft or its support structure is in resonance at the running speed. The immediate action is to perform a bump test or run-up/coast-down test to confirm the natural frequency, then investigate structural stiffness issues or consider a tuned-mass damper.'

Answer Strategy

Tests ability to translate technical risk into business language. Sample answer: 'I would present the data trend showing the bearing degradation curve, alongside the machine's criticality rating in our asset register. I would frame the cost of the planned, scheduled replacement (during a planned turnaround in 3 weeks) versus the risk of unplanned failure, which based on our MTBF data, would likely cause a 48-hour production loss and incur 3x the cost in parts, labor, and lost revenue. I would emphasize the action is to monitor closely and prepare parts, not to shut down immediately, thus giving management control over the risk.'