The last decade has seen maintenance management evolve immensely. From paper to spreadsheets to Computerized Maintenance Management Software (CMMS), it’s been quite the journey for CMMS Software. The obvious next step for CMMS Software is predictive maintenance. However, with ever evolving technologies, what is it going to take to stay on top of maintenance management? We’re hoping to provide some answers.
Did you know artificial intelligence was one of the most searched for keywords in Google this past year? Artificial intelligence is everywhere but do we understand what it really is? I’m assuming a bot comes to mind right away ? Or a science fiction movie? Or battlebots?
Imagine you are entering your office facility, and the security system addresses you and opens the door. The security system quickly identifies you, does a background check from public and secure profiles and offers you a visitor’s pass with your name written on it. Your office security systems are smart. Perhaps on their way of becoming intelligent. Artificial intelligence is a journey, with lots of auxiliary technologies involved. In this article we will try to explore the subject matter from maintenance management’s perspective.
Pre-AI Maintenance Management
Post industrial revolution, when we started getting a flavor of “bulk”, (whether in terms of production, information or processing) we started organizing things on an industrial scale. Hence came the concept of standardization. And then a completely different era of evolution started. For example, repairing-jobs post a breakdown, evolved into preventive maintenance, paper and spreadsheets to track work order evolved to cloud based work order management.
With the onset of the ever increasing maintenance operations demands comes the need for a more evolved technology that addresses intelligent monitoring, prioritizing and optimizing maintenance schedules, self-maintenance and more!
AI in Maintenance Management
A Boeing study suggests that 85% of equipment fails despite calendar-based maintenance and one-third of all maintenance investments are wasted through ineffective maintenance management methods. This need for a more thorough and accurate maintenance management system led to the birth of AI in maintenance management. Maintenance Management coupled with technologies like internet of things, big data and AI can revolutionize management. A study by Manufacturing Business Technology stated that predictive maintenance using AI can save companies over $630 billion in costs over the next 15 years.
All it needs is data. With each passing day, the rate at which data is captured and stored is increasing. A study from Gartner points out that 72 percent of manufacturing industry’s data is unused due to the complexities involved with different variables, such as pressure, temperature and time. It’s now getting humanly impossible to collate and process data using computers and processors. Augury, a New York- and Israel-based predictive maintenance technology specialist, has its HVAC (heating, ventilation and air-conditioning) maintenance systems installed in more than 2,000 facilities across USA and Canada3. Data gathered from one location goes into a “malfunction dictionary” that can be used for all.
This brings us to two important aspects of the use of AI in maintenance management:
Continuous-monitoring as the name suggests is a vigilante system which involves both the failure system and the anomaly system. Failure system reads from the perspective of data patterns which indicate and predict operation failure. This means, like a doctor, the system knows what are the symptoms and indications of a failure.
On the other hand, an anomaly system reads data as deviations from normal routine operations. Unlike failure system, it does not focus on pre-defined failure symptoms but picks up variations from normal patterns. Both combined, give us a fair and a seamless monitoring of operational processes. Continuous-monitoring becomes important when low turn-around-time and down-time are of prime importance. It becomes even more important when operations involve very vital and critical processes. At times, they can be too complex and difficult for human interventions like gas pipelines over long distances, and tough climatic conditions.
If maintenance is like friction, the necessary evil, self-maintenance is roller-skates. Someone wearing roller-skates is always in motion to avoid a fall. Self-maintenance combined with continuous-monitoring avoid the same thing, fall or breakdown. A breakdown or a maintenance task brings a set of activities with itself, like raising an alarm, initiating a work order, ordering a spare part etc. Many times the set of activities are pre-decided in form of an SOP, standard operating procedure. We rely on these SOPs as an alternative to human intelligence, mainly where a human intervention is not readily available.
Harnessing the AI wave
It’s important to use the available technology to its utmost potential. Look for some relevant signs like developments in the areas of open data sources, machine learning and embedded AI. There are two categories of players who are active in these areas, the big enterprises like Amazon and Google, who’ve already accumulated massive amount of data. And the other, the relatively smaller players who act as solution providers or vendors. It is expected that these vendors would work on machine learning and open source (APIs and data) more rigorously, which they will eventually borrow from the big enterprises. These small players will develop tools which can be customized with the help of embedded AI and pre-built algorithms. Nobody must reinvent the wheel both in terms of data gathering and developing solutions. Existing codes and tools will be re-jigged and engineered to provide a custom fit solution. Of course this will give price benefits to clients and better margins to vendors as well.
Conclusively, it might be a good idea for service providers and vendors to start collecting relevant raw materials (codes and data sources), which can be quickly cooked into solutions, and thus served to their clients. For companies and solution providers, the best way to ride the AI wave is to find a right vendor who can bring technology and functional expertise on one platter, perhaps through using machine learning as a service. We know that technology plays an important role today in our day to day life. And it should in future. But the truth is, nobody knows what exactly the future holds. At best, we can speculate with calculated risks, and not miss the first mover’s advantage.