How IIoT Will Drive the Next Generation of APM Predictive Analytics

There is no doubt the Industrial Internet of Things (IIoT) is reshaping the manufacturing industry. Indeed, Asset Performance Management (APM) is already being affected. in a number of ways. The ever more data being available improves the quality and the timeliness of the predictive analytics used to drive reliability centered and condition based maintenance (RCM/CBM). Recently, Matt Littlefield speculated on the future of Data Historians, reflecting one of the shifts we are likely to see due to the IIoT. LNS Research’s Smart Connected Operations vision is dependent on having Smart Connected Assets. We have also written about both Smart Connected Operations and Smart Connected Assets, and the changes they will have, particularly in driving Operational Excellence (OpEx). However all this data we'll have at our fingertips creates another problem, one that could be the next wave off innovation in the APM space.

The Data Conundrum

At a recent conference I heard someone refer to the “Internet of Anything.” The implication being that we are getting to the stage where connectivity isn’t limited to people and smart devices, but virtually anything in existence. The point was made that in the last year or two mankind has created more data than was created cumulatively in all of the preceding years. Even more amazing is that it will take only six months to match that level of data creation in the future, and then three months, six weeks, and so forth. Even though asset centric data reflects only a fraction of the total data pool, it is a significant portion. So, the amount of asset data available is growing exponentially.

As a user some 30 years ago, I was involved with the implementation of the first Vax based data historian from OSIsoft at a Weyerhaeuser paper plant. One thing that drove us to look at data historians and the Vax 32 bit architecture was the realization that we would need to store large amounts of data. Being able to access it rapidly to gain the benefits we were seeking was also key.

Dr. Pat Kennedy of OSIsoft made the point “you need to collect everything because you don’t always know what data you will need until you need it.” If you don’t have the data to diagnose an issue when it occurs, you might be put in the difficult position of explaining to management why you spent so much to implement a system that doesn’t help you solve your problem. That’s because you didn’t collect the right information. The answer becomes “collect everything just in case.” Hence the conundrum. The idea that you need to capture everything, yet the exponentially increasing volume of data to capture leads to seemingly insurmountable problem.

The Next Level of Predictive Analytics

The winner in the predictive analytics race will be the one that not only creates the most powerful tools to analyze the data to predict asset health and impending failures, but also adds a new level of predictive capability to their solution. Going from asset health predictive capabilities to data relevance predictive capabilities will help separate the winners from the “also-rans” in the APM space.

Data relevance predictive capability is the ability to identify how long asset data needs to be stored to accurately forecast machine health and risks. Being able to store data only as long as it has relevance will be the power that unleashes APM and makes the movement to Cloud and hybrid environments not only possible but desirable. Once APM providers cannot use any and all of the data available, but determine when it is no longer of value, they will have achieved the next level of performance and utility.

Understand the capabilities of twenty of the leading vendors in the APM space by downloading our APM Solutions Section Guide. The guide contains comparison charts for the factors listed above and the detailed profiles of the twenty vendors ranging from automation companies, to enterprise software providers and includes many specialized APM solutions as well.