Vibration Analysis

What Is Vibration Analysis?

Vibration analysis is the measurement, recording, and diagnostic interpretation of mechanical vibration signals produced by operating equipment. Every rotating machine generates vibration — the question is whether that vibration represents normal operating behavior or the signature of a developing fault. Vibration analysis provides the answer by converting raw mechanical motion into frequency-domain and time-domain data that trained analysts can interpret to identify specific fault conditions, assess their severity, and estimate remaining useful life.

The science behind vibration analysis rests on a fundamental principle: every mechanical fault generates vibration at predictable frequencies determined by the machine’s geometry, operating speed, and component dimensions. A shaft imbalance produces vibration at the running speed. A misaligned coupling produces vibration at multiples of running speed with characteristic phase relationships. A defective rolling element bearing produces vibration at specific frequencies calculated from the bearing geometry — and each bearing component (outer race, inner race, rolling elements, cage) produces its own distinct frequency signature. These relationships are repeatable and physics-based, which is what makes vibration analysis a diagnostic tool rather than just a screening test.

The technology has matured significantly over the past two decades. Modern vibration data collectors capture high-resolution spectral data across frequency ranges that span from below 1 Hz to above 20 kHz, with dynamic ranges exceeding 90 dB. This resolution allows analysts to distinguish between fault signatures that are separated by fractions of a Hertz in the frequency domain — a capability that is essential when diagnosing complex machines with multiple bearings, gear meshes, and driven equipment components all generating vibration simultaneously. Advanced analysis techniques including envelope (demodulation) analysis, time waveform analysis, orbit plots for proximity probe data, and cepstrum analysis for gearbox diagnostics extend the diagnostic reach well beyond what basic spectral analysis alone can provide.

Route-Based vs. Online Monitoring: Choosing the Right Architecture

One of the most consequential decisions in building a vibration analysis program is determining which assets receive permanent online monitoring sensors and which are covered through periodic route-based data collection with portable instruments. This is not an either-or decision for most facilities — the optimal architecture is almost always a blend of both approaches, calibrated to each asset’s criticality, failure consequences, and P-F interval characteristics.

Route-based monitoring uses a portable data collector and manually placed or permanently mounted accelerometers to collect vibration data at defined measurement points on a scheduled interval — typically monthly for critical equipment and quarterly for general-purpose assets. A trained technician walks a predefined route through the facility, collecting data at each point with consistent sensor placement, instrument settings, and collection parameters. This discipline is critical: inconsistent data collection introduces variability that obscures real trends and generates false alarms. Route-based programs work exceptionally well when the collection interval is short relative to the P-F interval of the dominant failure modes. For most industrial rotating equipment operating above 600 RPM, monthly data collection provides adequate detection margin for bearing, balance, alignment, and looseness faults.

Online monitoring systems use permanently installed sensors — accelerometers, velocity transducers, or proximity probes — connected to continuous or periodic data acquisition hardware that collects and stores data automatically. These systems are justified when one or more of the following conditions apply: the equipment is safety-critical and a failure could result in personnel injury or environmental release; the production consequence of an unplanned failure exceeds the monitoring system cost by a large margin; the P-F interval is short enough that monthly route-based collection may miss rapidly developing faults; or the equipment is physically inaccessible for routine manual data collection due to hazardous environments, confined spaces, or remote location. In most industrial facilities, online monitoring is deployed on 5-15% of the rotating equipment population — the assets where the economic and safety justification is clear.

Severity Classification and the P-F Curve

Detecting a fault is the first step. Classifying its severity — determining where it falls on the progression from initial detection to functional failure — is what transforms detection into a planning tool. Our vibration analysis program uses a four-stage severity classification system that maps directly to maintenance planning actions.

Stage 1 represents an initial detection — a subtle change from baseline that may indicate the very early onset of a fault. At this stage, the recommended action is to increase monitoring frequency and confirm the trend at the next collection. Stage 2 represents a confirmed developing fault with a clear and repeatable signature. The fault is real, but severity is moderate — typically months of remaining useful life remain. The recommended action is to generate a work order for planned repair during the next available maintenance window. Stage 3 indicates an advanced fault condition where vibration levels have risen significantly and the signature shows progression. Remaining useful life may be weeks to a few months depending on equipment type and load conditions. The recommended action is to schedule repair as soon as practical and consider operational adjustments — load reduction, speed reduction, or standby equipment activation — to extend remaining life until repair can be executed. Stage 4 is a critical condition where failure is imminent. The machine should be removed from service as soon as possible to prevent secondary damage, safety incidents, or catastrophic failure.

This staging system gives your planning and scheduling team clear, prioritized information. Not every finding is an emergency, and not every finding can wait until the next turnaround. The severity classification system ensures that the urgency of the diagnosis matches the urgency of the response.

What Are the Signs Your Facility Needs Vibration Analysis Services?

Vibration analysis is relevant to any facility that depends on rotating equipment. The following indicators suggest that your facility would benefit from professional vibration analysis services — either as a new program or as an upgrade to an existing one that is underperforming.

• Bearings are being replaced on fixed time intervals (annually, every two years) rather than based on actual condition — this approach replaces some bearings too early (wasting parts and labor) and misses others that deteriorate faster than the schedule assumes
• Operators report that equipment “sounds different” or “vibrates more than it used to,” but there is no objective measurement to confirm or quantify the observation
• Your maintenance team has portable vibration instruments but lacks the training or time to analyze the data beyond checking whether overall vibration levels exceed a threshold
• You are experiencing recurring bearing failures on specific equipment and have not been able to identify whether the root cause is misalignment, imbalance, lubrication, installation practice, or an operating condition
• Critical equipment has been operating for extended periods without baseline vibration measurements, leaving no reference point for evaluating current conditions
• Your facility has variable-speed equipment (VFDs, turbine-driven machines) that your current monitoring approach does not adequately cover because speed changes complicate spectral analysis
• Post-maintenance verification is not being performed — equipment is reassembled and returned to service without vibration checks to confirm that the repair achieved acceptable mechanical condition
• You have experienced secondary damage from equipment failures — a bearing failure that damaged a shaft, a coupling failure that damaged both the driver and driven machine — that could have been prevented with earlier fault detection
• Your insurance carrier has asked about your equipment monitoring practices or has recommended vibration monitoring as a risk mitigation measure
• Your facility is planning a transition from reactive or time-based maintenance to a predictive maintenance strategy and needs a foundational monitoring technology to build the program around

Our Vibration Analysis Approach

Our vibration analysis services are built on the principle that monitoring without diagnosis is just data collection, and diagnosis without communication is just an academic exercise. The value is delivered when a specific fault condition is identified, its severity is classified, the remaining useful life is estimated, and that information reaches the right person with enough lead time to plan an effective response. Every element of our process is designed to ensure that chain is unbroken.

Sensor and Instrument Selection

We match sensor technology to the measurement application. General-purpose industrial accelerometers with 100 mV/g sensitivity cover the majority of route-based monitoring requirements — pumps, motors, fans, and compressors operating above 600 RPM. For low-speed applications below 300 RPM, we deploy low-frequency accelerometers with extended low-end frequency response and higher sensitivity (500 mV/g or 1000 mV/g) to capture the small-amplitude vibration signals that standard accelerometers may not resolve. For equipment requiring shaft-relative vibration measurements — large turbomachinery, journal bearing machines, or equipment where casing vibration does not reliably represent shaft behavior — we specify non-contact proximity probes installed per API 670 guidelines. High-frequency enveloping analysis using accelerometers with resonant frequencies above 30 kHz provides early-stage bearing defect detection by isolating the high-frequency stress waves produced by defect impacts from the lower-frequency machine vibration.

Database Design and Trending Philosophy

A vibration analysis program is only as good as its database. We build measurement point databases with standardized naming conventions, consistent measurement parameters (frequency range, resolution, averaging, units), and sensor orientation documentation for every point. This standardization ensures that data collected over months and years is directly comparable — that a spectral trend showing a bearing defect frequency increasing over six months reflects a real change in equipment condition, not a change in how the data was collected.

Our trending philosophy emphasizes the rate of change over absolute values — a vibration level actively increasing warrants investigation even when still below guideline thresholds.

Our trending philosophy emphasizes the rate of change over absolute values. A vibration level that has been stable at 0.25 inches per second for two years is not a concern, even if it is above a generic guideline threshold. A vibration level that has increased from 0.08 to 0.18 inches per second over three months — while still below the same guideline threshold — is actively deteriorating and warrants investigation. We configure alarm structures that capture both absolute exceedances and rate-of-change anomalies, ensuring that developing faults are not masked by thresholds that are insensitive to trends.

Multi-Technology Integration

Vibration analysis is the backbone of most predictive maintenance programs, but it does not operate in isolation. We integrate vibration findings with data from oil analysis, thermography, ultrasonic testing, motor current analysis, and process parameters to build a complete diagnostic picture. A vibration signature indicating a bearing defect is more diagnostically powerful when correlated with oil analysis showing elevated iron and chromium wear metals. A thermal anomaly on a motor junction box gains context when vibration data confirms the motor is operating under normal mechanical conditions — pointing to an electrical rather than mechanical root cause. This cross-technology correlation increases diagnostic confidence, reduces false positives, and provides a more complete understanding of equipment health than any single technology can deliver alone.

Reporting and Communication

Every vibration analysis report we deliver includes a plain-language diagnosis, a severity classification, a recommended action with a timeline, and the supporting spectral data for analysts who want to review the technical detail. We do not deliver reports filled with unexplained spectra and generic recommendations. If we identify a bearing defect, the report will specify which bearing, which component (inner race, outer race, cage, rolling element), the current severity stage, the estimated time to failure at current operating conditions, and the specific corrective action. If the diagnosis is uncertain, we say so — and we specify what additional data or testing would resolve the uncertainty.

What Equipment Is Typically Covered?

Centrifugal Pumps

Centrifugal pumps are the most common rotating equipment type in most industrial facilities and are a core component of any vibration monitoring program. Monitoring typically covers the motor drive end, motor non-drive end, pump inboard, and pump outboard bearing positions. Common fault conditions detected include mechanical imbalance from impeller erosion or buildup, misalignment between motor and pump (angular, offset, or combined), bearing defects at all stages of development, structural looseness at the base or coupling, and cavitation — which produces a distinctive broadband high-frequency vibration signature that indicates the pump is operating outside its preferred operating range. Vertical pumps present additional monitoring challenges due to structural resonance modes and require specialized analysis approaches.

Electric Motors (AC Induction and DC)

Vibration monitoring on electric motors detects both mechanical and electrically induced faults. Mechanical faults — bearing defects, imbalance, looseness — produce the same signatures as on any rotating machine. Electrically induced vibration appears at twice the line frequency (120 Hz in 60 Hz systems) and its harmonics, and can indicate rotor eccentricity, broken rotor bars, stator winding faults, or supply voltage imbalance. Distinguishing between mechanical and electrical vibration sources is critical for correct diagnosis — a motor vibrating at 120 Hz due to a broken rotor bar will not be fixed by balancing or alignment, and misdiagnosis leads to repeated failed repairs.

Fans and Blowers

Industrial fans — particularly large induced-draft and forced-draft fans in combustion processes — are high-value monitoring targets because of their size, the production impact of their failure, and their susceptibility to imbalance from material buildup, erosion, and corrosion on the impeller. Fans operating in dirty gas streams may require monthly or bi-weekly monitoring to track balance changes between cleaning cycles. Overhung fan designs are particularly sensitive to imbalance and misalignment due to their cantilevered rotor geometry. Axial vibration monitoring on fans often reveals thrust-related conditions — including misalignment and aerodynamic instability — that are not apparent in radial measurements alone.

Gearboxes and Speed Increasers/Reducers

Gearboxes contain some of the most complex vibration signatures of any industrial equipment due to the simultaneous generation of gear mesh frequencies, bearing defect frequencies on multiple shafts, sideband patterns from gear tooth wear or damage, and resonance interactions. High-resolution spectral analysis is essential — a minimum of 3,200 lines of resolution is typically required for parallel shaft gearboxes, and 6,400 or more lines may be needed for planetary gear sets where the carrier, sun, ring, and planet components generate closely spaced frequencies. Gear defect diagnosis requires expertise in interpreting sideband spacing and amplitude patterns, and in distinguishing normal gear mesh characteristics from developing defects.

Reciprocating Compressors and Engines

Reciprocating equipment generates vibration that is fundamentally different from rotating equipment — dominated by impulsive forces from piston motion, valve events, and cylinder pressure fluctuations rather than continuous sinusoidal vibration. Analysis techniques for reciprocating machines focus on time-domain waveform analysis, peak-to-peak amplitude trending, and event-based analysis tied to crankshaft position. Fault conditions including loose crossheads, worn piston rings, valve leakage, and foundation deterioration produce identifiable waveform characteristics that experienced analysts can diagnose and stage.

Turbomachinery

Steam turbines, gas turbines, and centrifugal compressors operating at high speeds typically require shaft-relative vibration monitoring with non-contact proximity probes in addition to casing-mounted seismic measurements. Orbit analysis, shaft centerline position tracking, Bode and polar plots during startup and shutdown transients, and gap voltage monitoring provide diagnostic information that is critical for these high-speed, high-value machines. API 670 defines the instrumentation and monitoring system requirements for this equipment class, and our programs are designed to comply with and leverage those standards.

What Results Do Companies Typically See?

Vibration analysis is the most widely deployed and most extensively documented predictive maintenance technology in industrial use. The results it produces are well-established, and the outcome ranges below reflect what we consistently deliver across a range of industries and starting conditions.

Catching a $300 bearing defect before it becomes a $15,000 shaft replacement or a $50,000 gearbox rebuild is the single highest-value outcome of vibration monitoring.

• Bearing life extension of 30-50% by detecting and correcting installation defects (misalignment, soft foot, improper preload) and operating conditions (overlubrication, contamination, overload) that would otherwise shorten bearing service life below its design capability
• Unplanned rotating equipment downtime reduction of 40-60% as mechanical faults are detected and corrected during planned maintenance windows rather than through emergency response after failure
• Secondary damage avoidance — catching a $300 bearing defect before it becomes a $15,000 shaft replacement or a $50,000 gearbox rebuild is the single highest-value outcome of vibration monitoring, and most facilities experience multiple such saves per year
• Precision maintenance improvement — post-maintenance vibration verification identifies installation problems (misalignment, soft foot, balance issues) before equipment is returned to full-load service, reducing infant mortality failures by 25-40%
• Preventive maintenance optimization of 15-25% as time-based bearing replacements and coupling inspections are converted to condition-based triggers, eliminating unnecessary intrusive maintenance while maintaining or improving reliability
• Energy consumption reduction of 3-7% on rotating equipment where vibration analysis identifies and drives correction of misalignment, imbalance, and resonance conditions that increase mechanical losses
• Mean time to repair (MTTR) reduction of 20-30% because vibration-based diagnosis identifies the specific fault condition and affected component before the machine is opened, allowing the maintenance team to have the right parts, tools, and skills staged before the work begins

Over the first three to six months, a new vibration monitoring program typically identifies existing fault conditions — the backlog of developing problems that have been accumulating without detection.

The timeline for realizing these results is important to set expectations correctly. In the first three to six months, a new vibration monitoring program typically identifies existing fault conditions — the backlog of developing problems that have been accumulating without detection. This initial burst of findings often produces immediate high-value saves and builds credibility with operations and maintenance leadership. Over months six through eighteen, the program transitions from identifying existing faults to catching new faults as they develop, and the trending database begins to enable more accurate remaining-life estimates. Beyond eighteen months, the accumulated data supports statistical analysis of failure patterns across equipment classes, evaluation of maintenance practices, and optimization of monitoring intervals and alarm thresholds — benefits that compound as the database grows.

Vibration analysis is not a tool that produces results once and then plateaus. It is a capability that becomes more valuable over time as the depth of equipment-specific data increases and the analytical expertise sharpens. Our team brings that depth of expertise and the systematic discipline to build a vibration monitoring program that delivers sustained, measurable value to your operation.