11.4 Interpretation of vibration data

Vibration data interpretation is crucial for understanding mechanical system behavior. It involves analyzing natural frequencies, mode shapes, and damping ratios to characterize system dynamics. Frequency and time domain analyses, along with modal analysis techniques, provide deeper insights into vibration patterns and system properties.

Fault identification in machinery relies on spectral analysis, advanced signal processing, and trending techniques. These methods help detect common faults like imbalance, misalignment, and bearing defects. Correlating vibration data with other parameters and using visual representation tools enhances diagnostic capabilities and communication of findings.

Meaningful Information from Vibration

Natural Frequencies and Mode Shapes

Top images from around the web for Natural Frequencies and Mode Shapes

• 

  Damped and Driven Oscillations | Boundless Physics View original

  Is this image relevant?

• 

  Extracting Vibration Characteristics and Performing Sound Synthesis of Acoustic Guitar to ... View original

  Is this image relevant?

• 

  frequency - What are the first, second etc modes of vibration? - Physics Stack Exchange View original

  Is this image relevant?

• 

  Damped and Driven Oscillations | Boundless Physics View original

  Is this image relevant?

• 

  Extracting Vibration Characteristics and Performing Sound Synthesis of Acoustic Guitar to ... View original

  Is this image relevant?

1 of 3

Top images from around the web for Natural Frequencies and Mode Shapes

• 

  Damped and Driven Oscillations | Boundless Physics View original

  Is this image relevant?

• 

  Extracting Vibration Characteristics and Performing Sound Synthesis of Acoustic Guitar to ... View original

  Is this image relevant?

• 

  frequency - What are the first, second etc modes of vibration? - Physics Stack Exchange View original

  Is this image relevant?

• 

  Damped and Driven Oscillations | Boundless Physics View original

  Is this image relevant?

• 

  Extracting Vibration Characteristics and Performing Sound Synthesis of Acoustic Guitar to ... View original

  Is this image relevant?

1 of 3

• Natural frequencies represent inherent vibration tendencies of mechanical systems when disturbed
  • Determined by system mass and stiffness
  • Typically measured in Hertz (Hz)
  • Example: A guitar string's fundamental frequency
• Mode shapes describe deformation patterns at specific natural frequencies
  • Unique for each natural frequency
  • Characterized by nodal points (stationary) and anti-nodes (maximum displacement)
  • Example: Vibration patterns of a circular drum head
• Damping ratio quantifies oscillation decay after disturbance
  • Dimensionless measure between 0 and 1
  • Lower values indicate longer vibration persistence
  • Example: Shock absorber in a car (high damping) vs. tuning fork (low damping)

Frequency and Time Domain Analysis

• Fast Fourier Transform (FFT) extracts frequency content from time-domain signals
  • Converts time-based data to frequency-based representation
  • Reveals dominant frequencies and their amplitudes
  • Example: Identifying gear mesh frequencies in a gearbox
• Autocorrelation function identifies periodic components within a signal
  • Measures similarity of a signal with a delayed copy of itself
  • Useful for detecting hidden periodicities
  • Example: Detecting cylinder firing frequency in an engine
• Cross-correlation function analyzes relationships between different vibration signals
  • Measures similarity between two signals as a function of time-lag
  • Helps identify cause-and-effect relationships
  • Example: Correlating vibration between coupled shafts

Modal Analysis Techniques

• Experimental Modal Analysis (EMA) extracts modal parameters using controlled excitation
  • Requires known input force and measured output response
  • Yields frequency response functions (FRFs)
  • Example: Hammer impact testing on a bridge structure
• Operational Modal Analysis (OMA) extracts modal parameters during normal operation
  • Uses ambient excitation without measuring input forces
  • Suitable for large structures or continuous processes
  • Example: Analyzing wind turbine blade vibrations during operation

Fault Identification in Machinery

Spectral Analysis for Fault Detection

• Power Spectral Density (PSD) reveals energy distribution across frequencies
  • Identifies dominant frequency components
  • Useful for detecting periodic faults
  • Example: Identifying bearing fault frequencies in a motor
• Order analysis correlates vibration frequencies with machine speed
  • Tracks harmonic components related to rotational speed
  • Useful for variable speed machinery
  • Example: Detecting misalignment in a pump-motor system
• Characteristic frequencies associated with common faults
  • Imbalance: 1x running speed
  • Misalignment: 1x, 2x, and sometimes 3x running speed
  • Looseness: Multiple harmonics of running speed
  • Bearing defects: Specific frequencies based on bearing geometry
  • Gear damage: Gear mesh frequency and sidebands

Advanced Signal Processing Techniques

• Envelope analysis detects modulation in vibration signals
  • Extracts amplitude modulation caused by impacts
  • Effective for early-stage bearing and gear faults
  • Example: Detecting a localized fault on an inner race of a rolling element bearing
• Cepstrum analysis identifies periodic structures in frequency spectrum
  • Useful for detecting harmonics and sidebands
  • Helps diagnose gear faults and echo effects
  • Example: Identifying gear tooth wear in a gearbox
• Machine learning algorithms automate fault detection and classification
  • Supervised learning for known fault patterns
  • Unsupervised learning for anomaly detection
  • Example: Neural network classifying vibration patterns in a wind turbine

Trending and Comparative Analysis

• Vibration data trending tracks changes over time
  • Establishes baseline vibration levels
  • Detects gradual deterioration
  • Example: Monitoring gradual increase in overall vibration levels of a pump
• Comparison with historical data and industry standards
  • Identifies deviations from normal operation
  • Utilizes severity charts (ISO standards)
  • Example: Comparing measured vibration velocity to ISO 10816 alarm levels

Vibration Data Correlation

Multi-parameter Integration

• Cross-correlation techniques identify relationships between vibration and other parameters
  • Reveals cause-and-effect relationships
  • Helps in root cause analysis
  • Example: Correlating pressure pulsations with vibration in a hydraulic system
• Temperature correlation with vibration data
  • Indicates friction or lubrication issues
  • Helps diagnose bearing problems
  • Example: Detecting a failing bearing by correlating increasing temperature with vibration
• Pressure fluctuations in fluid systems
  • Identifies issues like cavitation or surge
  • Correlates with flow-induced vibrations
  • Example: Detecting pump cavitation by correlating suction pressure drops with vibration spikes

Speed-related Analysis

• Vibration analysis at different operating speeds
  • Identifies speed-dependent faults
  • Helps diagnose resonance conditions
  • Example: Creating a Bode plot to identify critical speeds in a rotor system
• Time synchronous averaging (TSA) correlates vibration with rotational speed
  • Separates deterministic and random signal components
  • Enhances periodic signals related to shaft rotation
  • Example: Isolating gear mesh vibrations in a complex gearbox

Data Fusion Techniques

• Sensor fusion algorithms combine information from multiple sensors
  • Improves fault detection accuracy
  • Provides more comprehensive system health assessment
  • Example: Combining vibration, acoustic emission, and oil analysis data for bearing health monitoring
• Multivariate statistical analysis for complex systems
  • Handles large datasets with multiple variables
  • Identifies patterns and correlations
  • Example: Principal Component Analysis (PCA) for fault detection in a gas turbine

Communication of Vibration Findings

Visual Representation Techniques

• Waterfall plots display vibration spectra over time or speed
  • Shows frequency content changes
  • Useful for analyzing variable speed machinery
  • Example: Visualizing resonance passages during machine startup
• Orbit plots represent shaft centerline motion
  • Diagnoses rotor-related issues
  • Reveals information about system dynamics
  • Example: Identifying oil whirl in a journal bearing
• Polar plots show vibration amplitude and phase angle
  • Useful for balancing operations
  • Helps visualize vibration vector changes
  • Example: Displaying rotor unbalance before and after correction

Performance Indicators and Severity Assessment

• Key Performance Indicators (KPIs) quantify machinery health
  • Overall vibration levels
  • Trend parameters (e.g., peak values, RMS)
  • Alarm and alert thresholds
  • Example: Tracking overall velocity RMS for a pump over time
• Severity charts communicate urgency of identified issues
  • Based on industry standards (ISO, API)
  • Color-coded for easy interpretation
  • Example: ISO 10816 severity chart for vibration velocity

Actionable Insights and Recommendations

• Root cause analysis methodologies applied to vibration findings
  • Ishikawa (fishbone) diagrams
  • 5-Why analysis
  • Fault tree analysis
  • Example: Tracing high vibration in a fan to misalignment caused by a loose foundation
• Cost-benefit analysis of proposed interventions
  • Quantifies potential savings from preventive actions
  • Justifies maintenance decisions
  • Example: Comparing cost of scheduled bearing replacement vs. potential downtime cost
• Tailored reporting for different stakeholders
  • Technical details for maintenance teams
  • Summary reports for management
  • Visual dashboards for operators
  • Example: Executive summary highlighting critical machine health issues and recommended actions