CONDITION monitoring is perhaps the most misused and misunderstood of all industrial plant improvement programs.
It is typically defined as a means of preventing catastrophic failure in critical rotating machinery - such as power generation plant, larger pumps and main arterial conveying systems - and providing the data needed to accurately determine the optimal schedule for maintenance activities on this plant.
It is largely viewed as a “maintenance tool”, with little relevance to plant management and operations.
In truth, condition monitoring can and should play a much broader role in the modern industrial operations: it is a tool that helps effectively manage site plant assets, logistics and labour requirements.
If we take, for example, a typical Australian mine site, located away from any major infrastructure and immediate transportation, labour is usually finite, plant on site is specialised and availability requirements are usually “24/7”. Faced with these challenges, we can see that careful management of plant assets is critical to ensuring overall plant performance.
There are a variety of technologies that are used as part of a complete plant condition monitoring program, the more common being:
• Vibration analysis.
• Oil analysis.
• Infrared thermography.
• Motor current analysis.
Typically, we see a range of analysis technologies applied to any given piece of equipment, which allows the analyst to make the most informed decision.
Offline and online analysis
OF all these technologies, vibration analysis is the one most commonly used and the one that provides the most amount of information from the data acquired.
Vibration monitoring basically encompasses two different techniques: offline and online. Before looking at the condition monitoring program itself, we need to fully understand these two techniques.
Offline vibration monitoring (sometimes know as the “walk around system”) is based on the collection of vibration data by means of a handheld data logger, and then subsequently downloading this information to a database for further analysis. Specialist software is used to view the data and report findings.
Online monitoring, by contrast, provides some level of permanent connection to the monitored plant.
Online monitoring can be further broken down into two techniques: surveillance and continuous monitoring.
Surveillance monitoring sits between offline monitoring and continuous online monitoring. This intermediate technique is akin to having a permanently connected data logger complete with a number of inputs, which is in turn permanently connected to a PC.
Transfer of data and alarm log information is performed automatically at a set interval, with the data collection performed by multiplexing from one input to the next.
The advantage of surveillance monitoring over conventional offline monitoring is that the monitoring intervals can be increased without incurring additional labour cost. This allows for better fault detection, and permits collection from dangerous or hazardous areas without incurring risks to engineers. Remote site monitoring is also made more economical, as data can be transferred over great distances using this technique.
Continuous monitoring is the preferred technique where full-protection of an asset is required. It is similar to the surveillance technique, except that there is no lag time between data acquisition and processing.
Continuous monitoring systems collect all the input channels simultaneously and process the data immediately. There have been great developments in continuous monitoring protection systems over recent times. In the past, the traditional approach has been to use “shut down protection” systems, which only monitored overall values.
By contrast, the latest generation systems collect a mass of data that includes phase data, and even captures “start up” and “coast down” data. While not all this data would be used for protection alarming, it does provide operations and maintenance managers with the ability to access high-end data without incurring the additional cost associated with advanced offline systems.
Overall and spectral band
EACH of these techniques can be used with either overall or spectral band alarming of the machine vibration.
Overall, as its name implies, measures the overall energy within a large vibration frequency range. This is typically between 5 and 2000 Hertz, or 5 and 5000 Hertz.
Overall monitoring has its limitations, and is typically used as shutdown protection for failures that might result in large increases in vibration amplitude. An example of such an event is an impeller losing a blade, which would result in a large increase in the vibration amplitude in the lower frequency range.
By contrast, spectral band alarming (or narrow band alarming, as sometimes it is known) divides a set frequency range into bands of varying sizes. These are based on the vibration frequencies that would manifest as a result of specific machinery faults.
For example, a bearing fault in its early stages of degradation will normally appear in the higher frequency range at amplitudes that would not normally trigger an overall alarm.
Using band alarms we can band this particular frequency range and set an alarm level that is appropriate for this particular type of fault. So in summary, spectral band alarming permits the detection of machinery faults at an earlier stage of degradation than that provided by overall alarming.
Both overall and spectral alarming can be applied to the offline and online monitoring techniques.
Spectral alarming used with continuous monitoring systems is the latest development - it allows engineers to protect against more than a catastrophic event and program for mechanical fault detection. This would normally have only been available in the offline or online surveillance programs.
BEFORE determining which technology should be applied to an equipment item, it’s best to register and “rank” the equipment accordingly.
Ranking is itself a complex issue that takes into account many factors. These factors are weighted differently depending on what role the plant plays in the facility’s operations, and the actual plant design. The following are some such factors:
The effect the equipment failure might have on the facility’s production or output.
The failure modes of the equipment.
The cost of replacement (whole or part) of the machine.
The health and safety consequences of equipment failure.
The environmental consequences of equipment failure.
This list is not exhaustive, but it does give an overview of what factors are taken into account when setting the appropriate ranking. We can quickly see a number of positives that can be made in terms of dollar savings and corporate governance.
If we look at the first point - the effect the equipment failure might have on the facility’s production - we see that a number of questions need to be asked first:
How do stoppages in this area affect overall production output? An example could be a crusher at a mine site, where downtime will cause stoppages downstream, and downstream production is already at 100 percent.
How long will repairs take? Both repair and replacement should be considered in the event of catastrophic failure.
What is the lead time in obtaining spares and/or replacement equipment, if not on site?
What product wastage due to incomplete processing might occur?
Most of these points relate to potential production time loss, with the exception of the last. The last includes the additional cost of materials and prior processing time as a cost of failure.
Alone, these points make a good case for implementing a condition monitoring program and realising the gains. But we can also add the cost of carrying spares, as a capital cost, labour cost to manage logistics, and the cost of scrapping due to obsolescence if spares are no longer useable.
Modes of failure
TO review the second point, failure modes, we need to put on the technical hat.
Plant equipment typically have usual modes of failure. For example, an electric motor rated running at 1500 rpm and fitted with rolling element bearings could possibly incur a bearing failure and run to failure in approximately three months.
By contrast, a turbine with sleeve bearings may only run for two hours with a bearing fault. These extremes demonstrate the failure differences between equipment. Knowing what the possible failure modes are allows us to select the appropriate monitoring technique, and to select the appropriate testing interval (for example, weekly, month or bi-monthly).
The third point, replacement cost in case of failure, is similar to point two in regards to what type of monitoring technique is applied. Plant equipment that incurs a high repair/replacement cost in the event of failure would typically warrant an online protection system or more frequent offline testing.
The last two points explore what the health and safety and environmental consequences would be of such a failure. Typically, a risk assessment is carried out on the specific production area to determine the possible failure scenarios, and what may be required to reduce or eliminate risk. As described earlier, the end result will be a recommendation for either online or offline and an appropriate testing interval.
The accompanying figure depicts the various elements that impact on the final make-up of any facility’s condition monitoring program.
Clearly, it is much more than a means of scheduling the maintenance activities. Rather, condition monitoring should be seen as an asset management tool that ensures the optimum level of plant availability, and a performance level that meets business objectives.
Viewed in this light, and with appropriate support from senior management and cooperation across the entire operation, the potential impact of a well-implemented condition monitoring system can be dramatic.
* Commentary by Mark Liebler, condition-based monitoring coordinator, Rockwell Automation Australia .