Condition monitoring, Industry 4.0, analytics, predictive maintenance, diagnostics, prognostics…
With an increasing use of technology to drive businesses forward in an agile, efficient, compliant and - crucially - competitive manner, does the skilled manufacturing sector worker need to go back to school to study computer science? What do these terms mean? Does it even matter? Despite claims that “every business is a technology company”, every business has its own core skillset and can utilise appropriate specialist outsourcing solutions for additional expertise, utilising software for advanced data analytics to better understand machinery health.
Thankfully, quality and maintenance teams don’t need a degree in computer science, but it is helpful to know the principles of some key data processing and interpretation terms. Here, we look at three key words used in the world of condition monitoring – and we won’t test you on it!
Why diagnostics, detection and prognostics?
Condition monitoring and prognostics look to make factory maintenance and management easier, more efficient and more cost effective – they enable Predictive Maintenance. The outcomes of effective machinery management are enhanced product quality, confident delivery timescales, and reduced overheads through reduced downtime, spares inventory, emergency repairs and over-servicing costs, and under-estimating remaining useful life. Diagnostics, detection and prognostics are key tools in proactive machinery maintenance.
Detection: The action or process of identifying the presence of something concealed.
Detection is the foundation of condition monitoring. Detection software monitors sensor readings, checking them against predefined thresholds. If the readings fall outside these thresholds then an alert is sent to highlight the defect. These alerts identify problems early, flagging the issue to the user before the machine fails. This information is often passed to a diagnosis tool to identify the most likely cause and a recommended course of action.
Some software providers, such as Senseye, use more advanced analytics and machine learning to take detection a step further; developing data monitoring techniques that automatically learn and tune themselves. This machine learning model can identify unknown failures – those not identified and categorised within the diagnosis stage.
Diagnostics: A program or routine that helps a user to identify errors.
Moving away from its medical roots, in the context of condition monitoring, diagnostics takes known data, such as a problematic component, and tells users what the possible failure mode is.
Diagnostics helps users by taking them beyond the detection stage’s simple threshold breach alert; it provides insight into possible causes for the problem, helping users to make decisions on the correct course of action. Further information can be found on our blog: Reduce machine downtime, embrace prognostics.
The manufacturing sector has a bias towards diagnostics, but it is reliant on hardware (good condition-indicating sensor data) and access to internal expertise from analysts, maintenance and quality teams who have the knowledge from which to build the diagnosis.
Prognostics: Relating to or serving to predict the likely course of a given condition.
Prognostics is the science of predicting when machines will stop being able to perform work; the remaining useful life. Used with the other outputs from the detection and diagnosis processes, prognostics tools analyse data and use modelling techniques to form a case of potential failure timelines to pass to the user.
This information will help the user decide whether a machine can continue running until the next scheduled stop, or whether it needs correcting sooner to avoid bigger problems such as machine downtime affecting business performance.
The benefits of prognostics are huge; the biggest being that it can prevent machine downtime and increase machine lifespan due to a better understanding of machine health, rather than using a benchmark lifetime calculation. A free prognostics white paper is available here.
The ideal solution
Diagnostics, detection and prognostics; we have looked at what they can each bring to the smart factory, but what is the ideal solution – or is there no one-size-fits-all solution?
Typically, all three methods are not required, but a combination is used to suit the priorities and specific nature of the site. And often, the higher cost of machine downtime, the higher the chance of employing more than one technique.
When it comes to condition monitoring, the end user is often a maintainer, not an analytics expert and should not be expected to understand the processes and terms used throughout the data collection, processing and analytics cycle. However, it is useful to have an appreciation of the different approaches used and what each can do for your specific requirements.
To summarise each: detection is a measured change, a diagnosis is derived from measured change assessed against given knowledge, and prognostics is a forecast based on information and given knowledge with varying levels of confidence.
There no one-size-fits-all solution, so a grasp of these concepts can help users to get the best solution for their requirements.
Senseye® is the leading cloud-based software for Predictive Maintenance. It helps manufacturers avoid downtime and save money by automatically forecasting machine failure without the need for expert manual analysis. Its intelligent machine-learning algorithms allow it to be used on any machine from any manufacturer, taking information from existing Industrial IoT sensors and platforms to automatically diagnose failures and provide the remaining useful life of machinery.