Vision Helps Perform Predictive Maintenance

A novel vision-based system enables engineers to perform predictive maintenance on assets in a plant quickly and accurately.

MaintWorld_ARG_Vision-helps-perform-predictive-maintenance_Fig1

Figure 1: A novel vision-based system called OptiVibe enables engineers to identify vibrational issues with their equipment with a greater degree of simplicity and in a much shorter period of time.

Vibration analysis is a widely used predictive maintenance technique that enables engineers to evaluate the condition of equipment. In doing so, maintenance personnel are able to schedule repairs on the equipment during maintenance shutdowns to avoid failures during production runs. 

One popular approach that has been deployed to visualize the vibration pattern of a machine involves the use of Operating Deflection Shape (ODS) analysis. In the process, vibration measurements are taken at a number of different locations on a machine using transducers. The data from the transducers is then analysed by software to create a geometric stick figure model of the structure of the machine and display the vibrational patterns associated with it.

However, the ODS analysis technique is not without its drawbacks. ODS requires a skilled end user to set up the mechanical transducers at specific locations on the machine before the data can be acquired. In many cases, the sheer number of collection points can prove problematic. 

Having deployed such a system, engineers then need to set up a database to acquire the data from all the collection points and post-process the data before a stick figure animation can be rendered. Although this can be achieved through off-the-shelf software packages, it can take many hours to complete such an analysis. To add to the complexity, the data may then need to be interpolated further to determine what is happening physically between each of the points from where the data was acquired. 

Vision Vibrations

Allied Reliability Group offers a service solution to its customers that utilizes a novel vision-based system called OptiVibe™ to identify vibrational issues with their equipment with a greater degree of simplicity and in a much shorter period of time (see Figure 1). 

The OptiVibe system itself typically captures video images of an asset in a plant using Allied Vision Technologies’ cameras, such as the Prosilica GE680. This camera uses OnSemi’s KAI-0340 CCD sensor, which delivers VGA resolution at a frame rate of 205 frames/sec. Since the resolution and the frame rate of the camera determines the resolution of the displacement and the maximum frequency of the vibration that needs to be identified, cameras with higher resolution imagers and greater frame rates can be deployed if necessary. Typically, however, around 15 seconds of data are captured by the camera and transferred over a GigE interface to a PC where the images are stored prior to analysis.

Proprietary algorithms running in a National Instruments’ LabVIEW application on a PC take the video input and process each pixel in the set of images to detect fluctuations in pixel intensity. By using a Fourier transformation, the time-dependent intensity data is then converted into frequency-dependent intensity data. The image processing software also generates Vibragrams, colour-coded representations that look exactly like thermograms, but which show the movement of the asset, instead of its temperature, providing an intuitive means by which the vibration of the asset can be interpreted (Figure 2.).

MaintWorld_ARG_Vision-helps-perform-predictive-maintenance_Fig2

Figure 2: Software running on the PC takes stored video data previously captured by a camera and processes each pixel in the set of images to detect fluctuations in pixel intensity. By using a Fourier transformation to convert the time-dependent intensity data into frequency-dependent intensity data, a range of frequencies can be displayed on screen that highlight where the fluctuations in intensity occur. The image processing software also generates Vibragrams, colour-coded representations that show movement.

Having captured several seconds of video from a specific asset, the OptiVibe system can display the frequencies and their relative amplitude in those parts of the image where vibration is occurring. An engineer can then select a specific frequency of interest and study the Vibragram to determine which parts of the asset were vibrating at that frequency and the relative amplitude of the vibration.

Another option supported by the software enables a complete time waveform or an FFT (fast Fourier transform) of every pixel within the image to be displayed. By simply moving a cursor around the image of the asset, a user can visualize all the frequencies and their relative amplitude at any point, rather than simply focus on frequencies of a specific wavelength. The result is the equivalent of placing an accelerometer on the equipment at different locations and viewing the different frequencies that are present.

Visualizing the Calender

To demonstrate the effectiveness of the system, engineers at Allied Reliability Group recently used the OptiVibe system to perform an analysis of a machine known as a calender. Widely used in the manufacture of paper, the calender comprises a series of heavy rollers driven by a series of motors through which the paper passes to ensure that it is smooth, glossy and of a uniform thickness.

The complexity of the calender and the many moving parts within it complicates the process of determining the cause and location of sources of any vibration using conventional techniques. However, by using the OptiVibe system, it is possible to simultaneously collect vibration measurement data across entire video images to help localize vibrations and determine what sources are causing individual components to vibrate.

One benefit of the OptiVibe system is its ability to isolate vibrations spatially and relate them to one another across a scene. To demonstrate this capability, images were captured by the camera of the structure around one of the motors on the calender (Figure 3). A Vibragram created from an analysis of the images clearly shows two bright objects -- the rotating shaft and a hand railing that are both vibrating at nearly the same frequency. From this, it can be deduced that it is most likely that the railing is being excited by the turning speed frequency of the rotating shaft. 

MaintWorld_ARG_Vision-helps-perform-predictive-maintenance_Fig3

Figure 3: A Vibragram created from an analysis of the images shows two bright objects -- a rotating shaft and a hand railing that are both vibrating at nearly the same frequency. From this, it can be deduced that it is most likely that the railing is being excited by the turning speed frequency of the rotating shaft.

Images were also captured of the paper itself as it moved though the production process. On the calender, the paper moves through several rolls as it dries. The process uniformly dries the paper as well as keeping the density of the paper as uniform as possible. Contacting methods are obviously impossible as the slightest pressure or disturbance on the paper can cause it to tear. Laser vibrometry would be possible but such systems are cost prohibitive and only measure one location at a time.

By capturing images of one of the paper rolls and analysing those images with the OptiVibe software, the frequency at which the paper is vibrating on the roll can immediately be visualized (Figure 4). The use of the OptiVibe software in such an application clearly opens up new opportunities for engineers to monitor such assets in many industries where material is produced on rolls. 

MaintWorld_ARG_Vision-helps-perform-predictive-maintenance_Fig4

Figure 4: By capturing images of one of the paper rolls and analysing those images with the OptiVibe software, the frequency at which the paper is vibrating on the roll can immediately be visualized.

Video image data was also collected under the rolls of the calender. The goal of collecting the data in this case was to demonstrate how the OptiVibe system could collect data on all parts of the machine simultaneously. Although the number of independently vibrating elements is enormous, the OptiVibe system can simply measure the vibration on all locations of the machine in a matter of minutes. Here, the belt vibration frequencies were measured, another impossible task for contacting measurement methods. A belt vibration can be clearly seen, as the belt vibration is highlighted red in the Vibragram (Figure 5).

MaintWorld_ARG_Vision-helps-perform-predictive-maintenance_Fig5

Figure 5: Here, the belt vibration frequencies were measured, an impossible task for contacting measurement methods. A belt vibration can be clearly seen, as the belt vibration is highlighted red in the Vibragram.

In another case, the OptiVibe system was deployed to show major structural vibrations in the machine itself. Here, it is apparent from the Vibragram produced from an analysis of the video data that a structural vibration is present as the entire structure is vibrating while other elements in the scene are not (Figure 6). With single point contact methods, it would be difficult to separate local vibrations from global vibrations. But because this is feasible with the OptiVibe system, engineers are now able to determine what structural vibrations are present and what the cause of those vibrations might be.

MaintWorld_ARG_Vision-helps-perform-predictive-maintenance_Fig6

Figure 6: The OptiVibe system can definitively show structural vibrations in the machine as the entire structure is vibrating while other elements in the scene are not.

Useful Images

The use of the vision-based system to perform condition monitoring clearly opens up some new opportunities for engineers to accurately measure the vibration of machinery, especially at locations in the machine that are moving where the use of traditional fixed transducers would prove impossible. 

Since the OptiVibe system collects data simultaneously across all the aspects of a machine, engineers also now have the ability to visualize how various components are interacting with one another, or to look at the asset as a whole. Hence, it is no longer necessary to choose which parts of a machine should be analysed. 

In addition, those aspects of the assets not specifically targeted during the collection process can be analysed at a later date to enable patterns and correlations in the data to be visualized. The visualization process can also be used to help engineers easily communicate where and how troublesome issues are occurring. 

Clearly, the OptiVibe system could be a good fit for monitoring a complex asset as it can capture the overall view of the asset’s vibration and may help determine unknown vibration sources by painting a larger picture of the motion of the asset. 

ChrisKlosterman

Chris Klosterman,
Vice President Products and Technology
Allied Reliability Group, Charleston, SC, USA
chris.klosterman(at)alliedreliability.com

Cmms | 22.11.2016

Latest articles

It’s About “Time”

At Bentley Systems, we have been talking to our users, especially owners, and it has become clear that despite our collective best efforts, there is a gap in the market.  Owners need easy access to real-time, accurate, engineering information and they need tools that make it simple to keep engineering data up-to-date, especially those engaged in brownfield and revamp projects.  It takes too much time and effort to gather and verify the information they need to make decisions effectively. And for older plants that do not have 3D models, there is simply no visual way to verify and check information easily.

Partner Articles | 10.12.2018

Energy savings 4.0

We have all read about it: leak detection should be a top priority since, if no leak detection program has been implemented, leaks can account for 30 to 40% of consumed volume. So, why is this issue still on the table? Why is it difficult to change things in the field?

Asset Management | 10.12.2018

How to Choose an Air Compressor, According to Science – 8 Factors to Consider

Buying an air compressor, for the first time, can be challenging in many ways. There is a lot going on behind this power tool that offers faster and more efficient performance. Before we get to know about the details of an air compressor then, let’s understand how it works. How do air compressors work?

Applications | 10.12.2018

Maintenance 4.0 hype or game changer?

Last September Antwerp (Belgium) hosted the Euromaintenance 4.0 conference. With 1187 participants from 64 different countries, the European maintenance federation EFNMS and the organisers BEMAS (Belgian Maintenance Association) and Reliabilityweb.com can look back on a very successful initiative. The massive interest and very positive vibe that prevailed confirm that maintenance 4.0 is currently at the centre of attention. But is this really justified? Isn't the whole buzz surrounding the Industrial Internet of Things (IIoT) just a hype that will have faded in a year's time?

Cmms | 10.12.2018

8 Ways to Identify and Eliminate Noise Hazards in the Workplace

Occupational noise hazards are one of the most common workplace safety concerns. Our hearing is sensitive and it does not take much to temporarily impair or permanently damage it. 

HSE | 10.12.2018

Bacterial Bandages, Natural Dyes and Recycled Fibres: Aalto Brings Materials Revolution to Slush

New materials play an important role in sustainable development and combatting climate change. New uses for old materials can also be a major industrial opportunity: for example, the value of biomass from forests in Finland can be doubled if used for manufacturing products of higher added value.

R&D | 4.12.2018

Maintenance: A Necessary and Important Function in the Future

Euromaintenance 2016 will take place in Athens at the end of May. It is the ideal moment to reflect on maintenance in a European context. Euromaintenance is known as the summit for all involved in maintenance across Europe, it’s the place to be. The conference, with the support of the EFNMS, is the only commercially independent conference covering the topics we deal with in the maintenance world.

EFNMS | 20.5.2016