Vibration analysis (or vibration monitoring) is a powerful diagnostic tool for identifying, monitoring and preventing mechanical failures in rotating and reciprocating machinery. It’s an essential component of predictive maintenance programs, allowing technicians to detect developing problems in equipment before they lead to costly unplanned downtime or catastrophic failures.
Vibration analysis—a component of condition monitoring systems—utilizes vibration sensors to measure frequencies in an asset and detect abnormalities that may indicate a problem. At its core, vibration analysis is the study of the oscillatory motions of machines and their components around an established equilibrium point. These oscillations can result from a number of issues, including imbalances, misalignments, looseness, bent shafts and bearing defects, among others.
By measuring and analyzing machine vibrations, technicians can better assess the health and performance of equipment, enabling them to pinpoint issues and implement corrective actions. And since rotating and reciprocating components (e.g., motors, compressors, pumps, etc.) tend to vibrate more intensely and more loudly as they age, analyzing changes in vibration can help maintenance managers monitor machine conditions in real-time and proactively identify asset wear and damage.
Understanding vibrations
Vibrations are multidimensional, so vibration testing requires an understanding of various parameters. The primary parameters are amplitude, frequency and phase:
- Amplitude refers to the magnitude of the vibration, typically measured in units like displacement (mils or micrometers), velocity (inches per second or millimeters per second) or acceleration (g’s).
- Frequency measures the number of oscillations per unit of time, usually expressed in Hertz (Hz).
- Phase refers to the relative timing of the vibration waveform, typically measured in degrees.
Vibration analyzers can use these factors—along with others like waveform shape and harmonic content—to identify specific machine faults and their severity.
There are also two major types of vibrations:
- Forced vibration occurs when an external force is applied to a system, causing it to vibrate. Examples include the vibrations produced by an operating motor or pump.
- Natural vibration, also known as resonance, occurs when a system vibrates at its natural frequency. This type of vibration can be problematic if the machine’s natural frequency coincides with the operating frequency, as it can lead to excessive vibrations.
Conducting a vibration analysis
In today’s competitive industrial landscape, ensuring the smooth operation and longevity of your machinery is paramount. Following a step-by-step vibration analysis process can help you leverage machine vibrations to better understand equipment and system health.
Step 1. Gather preliminary data
To conduct a vibration analysis, start by gathering data about the equipment and/or structure you will analyze. Be sure to identify the location and type of sensors you will need for vibration measurements, including any amplitude and frequency ranges, as the sensors will serve as your data collectors throughout the process. It is also important to gather information about the operating conditions of the equipment, such as speed, load and temperature.
Step 2. Measure vibrations
The next step in the process is measuring vibrations from the chosen machinery. Vibration analysis typically involves measuring the vibration amplitude of an asset or system.
In this case, you will rely on an accelerometer attached to key vibration sites to take measurements. Accelerometers (e.g., Piezoelectric accelerometers, proximity probes, Laser Doppler vibrometers, etc.) use high-frequency responses to measure the acceleration of a vibrating object in three directions: up-and-down, side-to-side and front-to-back.
You can also measure vibrations using velocity sensors and displacement sensors. Velocity sensors measure the velocity of a vibrating object, offering a mid-frequency response ideal for monitoring overall machine health. Displacement sensors, on the other hand, measure the displacement of a vibrating object, providing a low-frequency response suitable for detecting structural issues and imbalances.
Step 3. Analyze the data
Once you’ve collected all the relevant data, it’s time to analyze it. The analysis stage will require you to use computer algorithms and software to identify vibration patterns and data trends, and you’ll need to compare the data to established standards and guidelines.
Maintenance teams can utilize a variety of techniques to analyze vibration data, including the following:
- Time domain analysis: To conduct a time domain analysis, you will analyze the raw vibration signal over time. Technicians can extract and assess data (e.g., peak amplitude, crest factor, skewness, root mean square (RMS), etc.) of the signal directly from the time waveform. This technique is useful for detecting transient phenomena like impacts or shocks.
- Frequency domain analysis: Using a mathematical frequency analysis operation called the Fast Fourier Transform (FFT), frequency domain analysis involves converting the time-domain signal into a frequency-domain signal. The resulting frequency spectrum facilitates easier identification of the specific frequencies associated with mechanical faults.
- Envelope analysis: Also known as demodulation, envelope analysis is primarily used to detect early-stage bearing defects. It involves extracting the high-frequency impact signals generated by damaged bearings from the overall vibration signal, allowing technicians to identify the presence and severity of bearing faults.
- Modal analysis: Modal analysis is a more advanced technique that focuses on identifying a machine’s natural frequencies, mode shapes and damping characteristics. Studying these properties helps analysts assess the dynamic behavior of a machine and identify potential structural problems and/or resonance conditions.
Step 4. Interpret the results
A successful analysis will require vibration analysts to understand the specific frequencies associated with various mechanical issues. Some common fault frequencies include the following:
- Fundamental frequency: The frequency at which a rotating component completes one full rotation. Deviations from the baseline fundamental frequency can help you identify specific faults, such as unbalance, misalignment and looseness.
- Harmonics: The integer multiples of the fundamental frequency often associated with gear-meshing issues or other cyclical events.
- Sidebands: The result of the modulation of the fundamental frequency by another frequency, as in the interaction between a bearing fault and the frequency of a rotating component. Sidebands often appear as symmetrical peaks around the fundamental frequency or its harmonics and can be indicative of bearing issues and/or broken gear teeth.
Step 5. Take corrective action
Once you identify potential issues in the interpretation stage, set alarm thresholds that trigger a response when vibrations exceed them. These thresholds can be based on industry standards, historical data or manufacturer recommendations. The most commonly used types of alarm thresholds are absolute thresholds, trending thresholds and statistical thresholds.
- Absolute thresholds are fixed values based on acceptable levels of vibration for specific types of equipment.
- Trending thresholds are dynamic values that change over time based on the historical performance of the equipment. When vibration levels increase suddenly, it may indicate a developing issue.
- Statistical thresholds are derived from statistical analyses of historical data, taking into account the mean and standard deviation of vibration levels.
When a system or asset exceeds the alarm threshold, the maintenance manager should take appropriate action to address the issue. This may involve scheduling maintenance, adjusting operating parameters or conducting further analysis to pinpoint the cause of the problem.
Benefits of vibration analysis
Implementing vibration analysis as part of a predictive maintenance program offers numerous benefits:
- Early fault detection: Vibration analysis can detect developing faults in machinery long before they become visible or audible to human senses. These early detection capabilities help maintenance teams schedule repairs or replacements before a failure occurs, reducing downtime and improving overall productivity.
- Improved maintenance scheduling: By identifying the severity of machine faults, vibration analysis allows maintenance teams to prioritize their efforts and allocate resources more effectively. This targeted approach can result in significant cost savings and improved equipment reliability.
- Reduced maintenance costs: Finding and fixing machine faults before they lead to catastrophic failures can help organizations avoid costly repairs and equipment replacements. Furthermore, predictive maintenance based on vibration analysis can extend the lifespan of machinery, further reducing long-term capital expenditures.
- Enhanced safety: Vibration analysis can play a crucial role in maintaining a safe working environment for employees. Proactively detecting and addressing potential mechanical failures significantly reduces the risk of accidents and injuries in the workplace.
- Improved energy efficiency: Faulty machinery often consumes more energy than it should, leading to increased operating costs. Rectifying issues using vibration analysis helps organizations optimize asset efficiency and reduce energy consumption.
Applications of vibration analysis
The versatility of vibration analysis makes it a useful methodology across a wide range of industries and equipment types.
Manufacturing plants use vibration analysis to monitor the condition of motors (including electric motors), gearboxes, conveyors and machine tools. The vibration data can be used to optimize production processes, reduce the risk of equipment failure and improve overall plant efficiency.
In the automotive industry, vibration analysis plays a significant role in designing, developing and testing components. Analyzing the vibration characteristics of engines, transmissions and suspension systems can help engineers optimize their designs for improved real-world performance and reliability and increased passenger comfort.
In the aerospace industry, vibration analysis enables engineers to identify and address issues like excessive vibration, resonance or material fatigue to enhance the reliability and longevity of aircraft systems. And in the wind power sector, vibration analysis helps turbine operators monitor turbine health in order to identify blade imbalances, gearbox failures and/or bearing defects.
IBM Maximo can help you leverage vibration analysis for your business
Vibration analysis is an invaluable tool in the world of predictive maintenance, offering organizations the ability to detect machine faults early on. Understanding the principles of vibration analysis and using advanced software like IBM Maximo Application Suite can help maintenance teams significantly improve the reliability, efficiency and safety of their equipment.
IBM Maximo is an integrated platform that helps manufacturing companies optimize asset performance and streamline day-to-day operations. With Maximo, teams can set up meters for observing and measuring asset health and connect those meters to vibration sensors installed on key assets, helping them more easily store and access machine data.
Using an integrated AI-powered, cloud-based platform, Maximo offers CMMS, EAM and APM capabilities that produce advanced vibration data analytics and help managers make smarter, more data-driven maintenance decisions.
As industries continue to evolve and adopt new technologies, vibration analysis will remain at the forefront of maintaining machine health and ensuring optimal performance. By leveraging IBM Maximo’s advanced features and capabilities, companies can transform maintenance operations to keep pace with the ever-shifting manufacturing landscape.
Find out how by booking a live demo with an IBM expert
The post What is vibration analysis and how can it help optimize predictive maintenance? appeared first on IBM Blog.