AI-based equipment monitoring provides benefits beyond streamlining maintenance processes, so why do we bucket every use case as predictive maintenance?
Predictive maintenance can be defined as a method of using data analysis tools and techniques to detect anomalies in your operation and possible defects in equipment and processes so you can fix them before they result in failure.
The Value of Predictive Maintenance
The term predictive maintenance has come of age in parallel with terms such as Industry 4.0.
Both got their roots in manufacturing and improving operations on the factory floor. One primary reason for this is that factory floors were already collecting vast amounts of data on SCADAs and PLCs that were now becoming digitized. As a result, monitoring of assets typically centered around factory equipment. Equipment that is prone to failure while also abiding by strict maintenance schedules.
The value proposition of AI is that it can analyze the machines to predict issues in advance providing a notification of an issue that could be addressed with a specific maintenance task. Hence the term Predictive Maintenance.
Predictive Maintenance as a Subset of Intelligent Asset Monitoring
As IoT has emerged, helping to connect numerous types of assets, the use cases have grown but the term that has become synonymous with monitoring operations has not. As the use cases change so does the value proposition.
An asset performing outside of normal operations could easily need to be lubricated, require a filter change, or another routine maintenance task that will return the asset to normal. However, with the new types of data being collected from sensors, and the ability for AI to analyze that data, the alert could be telling us so much more.
Cavitation is the silent killer of water pumps and piping. Cavitation slowly wears down the materials such as the metal of a pump impeller resulting in a decrease in waterflow and eventually full degradation of the pump itself.
Through AI and IoT sensors measuring attributes such as flow, temperature, and vibration, we have the ability to detect cavitation as it occurs. The question becomes, what do you do with the insights generated from the AI?
In this example, if we detect cavitation, the goal may not be to perform maintenance on the pump to extend it’s useful life but instead to actually stop the current cavitation in order to extend the useful life. In other words, an operator may look to optimize the settings on the pump, or on another machine in the system, to reduce or stop cavitation as it occurs.
Oil & Gas Lines
On a pipeline, such as those for water, oil, or gas, IoT sensors can be deployed to monitor the current status of the system.
With AI based monitoring, we can monitor these sensors to learn normal operation. If the series of real-time measurements deviate from normal operations that could indicate issues such as a potential leak or even discover a current leak. The end result could be viewed as maintenance to the line but it is not necessarily routine maintenance but instead a fix or repair to a current problem.
AI offers predictive insights, but those insights are not always in advance and they are not always tied to a maintenance schedule. Often times, the insights can notify you of a problem happening in real-time that could otherwise go unseen for a length of time. Real-time monitoring of a fleet of equipment is very difficult and manual. AI makes it possible to automate this, helping to cut costs, limit catastrophic failures, and increase uptime not only of manufacturing equipment but all equipment.
For industrial equipment that does follow a maintenance schedule, such as HVAC systems, intelligent monitoring does offer the benefits of predictive maintenance to help lower costs and streamline maintenance processes.
However, there are also indirect effects of an asset performing abnormally.
For example, when an HVAC system, or any powered asset for that matter, is operating abnormally it has an impact on the power usage and load of that asset. AI based systems can therefore monitor for abnormal activity to help optimize settings for peak performance.
Think of a car struggling to get up to highway speed. Through the lens of predictive maintenance, the problem could be that the car requires an oil change and routine maintenance. Through the lens of Intelligent Asset Monitoring, the problem could be maintenance related but it could also mean that the vehicle settings are not optimized, such as running in too low of a gear.
The same AI based monitoring techniques deployed for predictive maintenance on the factory floor can be utilized to produce more actionable insights.
Remote asset monitoring for real-time incident discovery and equipment optimization are just two examples of leveraging AI to monitor equipment.
Many in the industrial space shy away from predictive maintenance because it’s framed in a way that makes it sound as if it is not for them. Intelligent Asset Monitoring on the other hand is for everyone.
Predictive maintenance using AI provides tremendous value but it is just a subset of the valuable insights that AI provides and is just a subset of Intelligent Asset Monitoring.
So, let’s help people understand the broad implications of AI and IoT and it starts by broadening the terminology to Intelligent Asset Monitoring.