ResizedImage168213-Sean-Riley-02368

Memory-edge analytics for predictive maintenance

July 13, 2016
Even the possibilities of IoT and digital transformation are not yet fully understood.

Dell recently hosted its Connect What Matters contest, which asked companies to submit their innovative IoT solutions using the Dell Edge Gateway device. Software AG was a gold-medal winner for its predictive-maintenance solution that includes memory-edge analytics for acting on machine data in real-time.

“The digital transformation is changing enterprise IT landscapes from inflexible
application silos to modern software-platform-driven IT architectures, which deliver the openness, speed and agility needed to enable the digital real-time enterprise,” said Sean T. Riley, global industry director (manufacturing, supply chain & logistics) with Software AG. “[The Software AG] modular platform allows users to develop the next generation of application systems to build their digital future today.”

Smart Industry: What is special about your predictive-maintenance solution? 

Sean: The ability to combine real-time condition monitoring with dynamic prediction. This allows users to understand not only what is happening with their equipment right now and have weeks of lead time to detect issues, but also to aggregate that data on predictive-model-specific time frames and automatically have that data sent to the models utilized. This provides months of lead time when an issue is predicted. While time is a critical commodity, understanding where the issue resides and the severity of the issue is another critical function. Because of this technology’s ability to handle massive scales of data at high speeds, the Software AG solution is able to understand the true issue at hand. For example, when monitoring a compressor for lubricant level, temperature, pressure-in, pressure-out, RPM, vibration and coolant level if the vibration sensor drops to zero…that’s a strong indicator that a bearing has just bored a hole through the casing and a catastrophic failure has occurred. By being able to measure and understand these variables both uniquely and in a multi-variant manner, we are able to quickly diagnose that a sensor failure has occurred because vibration is at zero but all other indicators show normal. This ensures that a high-priority issue is correctly classified as an issue of a much lower priority.

Smart Industry: What are applications for this solution?

Sean: Currently the solution is being used to monitor and predict maintenance requirements on both field-based/remote and in-plant-operated equipment. This can also be extended to the monitoring of transportation-asset performance, facility monitoring, smart-meter monitoring, continuous optimization of production processes, and production-quality management.

Smart Industry: How critical is the real-time element of your solution? 

Sean: It ensures that both immediate and future actions can be taken and that the model used to predict issues has a safety net. Unfortunately, a crystal ball with perfect future vision has not yet been created. Real-time, continuous monitoring ensures that all issues are caught with time to act, even if they aren’t diagnosed with the predictive model. When dealing with high-speed production processes, the real-time element ensures that when a quality issue occurs, the out-of-spec product is minimized, thereby limiting the cost exposure of the user.

Smart Industry: How is predictive maintenance changing the world of manufacturing? 

Sean: At first examination, predictive maintenance is viewed as increasing overall equipment effectiveness by ensuring maintenance activities only occur during planned downtimes, virtually guaranteeing equipment uptime and ensuring maintenance is completed only when it is really needed, resulting in decreased costs. While these benefits are useful, predictive maintenance is truly changing the world of manufacturing by enabling a path to outcome-based sales models, which enables purchasers to have a cost outlay that exactly matches their uses and needs for the equipment.

99% uptime and availability? Not a problem to provide, but the cost is going to be different than a requirement of 90% uptime and availability. A provider will be able to deliver these services efficiently and effectively, which will ensure the risk of this type of business model will be minimized. Equipment manufacturers will be able to strategically align with their customers at a very deep level and have conversations about present/future business requirements that they were previously not able to.

Smart Industry: What will you do with the winnings from the Connect What Matters contest?

Sean: The monetary funds will be used to support the continued development of the edge solution with Dell by adding additional use-cases to the current list. The IoT lab engagement and business and engineering consulting will also be leveraged to support the continued development of current and new use-cases on the (already robust) functional capabilities.

Smart Industry: What most excites you about the digital transformation of industry? 

Sean: Even the possibilities of IoT and digital transformation are not yet fully understood. Many companies have an understanding of what they want to do but have not fully conceived of the possibilities of how this technology is going to fundamentally alter business models, business processes and market offerings. It allows for manufacturers to almost become childlike and let their imaginations run wild with the possibilities. To me, not understanding exactly what will be offered in the future is exciting because being a part of this process will be much more than just a little bit of fun.