Being able to predict failure of assets could be a game-changer for businesses that rely on their plants being fully operational.
With Internet of Things (IoT) and AI, maintenance programs are more data-driven, more predictive, and “turning data into insights and into action,” according to Simon Lee, Product Marketing Manager for SAP Asset Performance Management.
In a recent ASUG webcast, “From Data to Action: Leveraging IoT and AI for Predictive Maintenance,” Lee discussed how SAP Asset Performance Management brings connected devices together onto one platform to monitor assets and make fast, informed decisions regarding both immediate maintenance and the larger trajectory of the customer's business.
The SAP APM Landscape
In the webcast, Lee spoke first about the wider product landscape as it relates to SAP Asset Performance Management.
SAP's supply chain management portfolio, along with its intelligent asset management portfolio (alternately referred to as asset service management), sits on top of SAP Business Technology Platform (BTP) and keeps SAP S/4HANA at its core. This allows for full asset management and service capabilities, enterprise asset management (EAM) functionality, and extensions or add-ons to work within customers' SAP implementations.
SAP Asset Performance Management, specifically, extends SAP S/4HANA asset management with more capabilities around asset risk assessment, asset reliability, and AI-enabled predictive maintenance.
Lee says that the best way to close the loop with SAP S/4HANA asset management between reliability and maintenance is with SAP Asset Performance Management, because the solution takes advantage of all maintenance records and data collected via measurement points, with its capabilities extending through the enterprise’s maintenance organization.
Embedded IoT is a new technology that SAP is using, powered by Cumulocity, a leader in connected devices. It uses Gartner’s Magic Quadrant for IoT, Industrial Internet of Things platforms, and is fully embedded into SAP Asset Performance Management, meaning that there’s no direct action needed by the customer to license these capabilities. Sensors connected through IoT perform continuous monitoring, detecting signals of potential failure early. The subject matter expert is then able to see the likelihood of failure and the consequence of failure based on whatever dimensions the organization identifies.
This allows time for planning and scheduling of the actual maintenance. Assets then have more uptime, bringing greater productivity to the business. Less overall maintenance is also needed, reducing the total cost of ownership.
Choosing the Right Data and Assets
The key to unlocking the value of a business’s connected assets is choosing the right ones, Lee said.
The assets that should be monitored are ones that constitute substantial safety risks or production risks, that have bad records, or that attract bad actors. Starting at these points is recommended before gradually expanding the program. “The ability to do these things systematically beyond just using spreadsheets or tribal knowledge is a really good thing from our point of view [at SAP],” Lee said.
It’s not quite as simple as choosing which assets to monitor; targeting the right failure modes is also critical. The most successful customers, Lee said, identify the top failure modes using industry standards, typically recommended by SAP partners, who outline those industry standards as aligned with master data and catalog profiles. That way, data can be more comprehensive and leveraged for more actionable insights.
Now that devices are connected and data is being collected, it should go to a central place. “That’s really, really important for analysis, for asset health monitoring, so that you can be able to monitor these receiving alerts in a consistent fashion,” Lee said.
Then, an enterprise can compile historical data and new data coming in — from sensors that collect time series data, image data from cameras or drones, and more — to see a holistic picture of asset health.
But by starting with the same master data that is speaking all the same language is key to success. This means having the same technical objects, the same functional location, the same classes characteristics, hierarchy, failure modes, cost codes, catalog profiles, notifications, and work orders.
Then, an enterprise can take advantage of AI capabilities, powered by that centralized data. “Let AI actually do all the hard work,” Lee said.
The Data-Driven Process of An AI-Powered Business
SAP follows an AI-first strategy in general, and especially with SAP Asset Performance Management, Lee said. The company first emphasizes intelligent rules, which monitor equipment consistently, in batches, or in near-real-time. The rules will continuously run in the background, monitoring for any potential failures of equipment, supplemented or even replaced by AI, catching the patterns that humans might miss.
Reliability engineers can set these rules up, specific to the enterprise’s unique data set, historical data, equipment, and operating conditions, to train AI to detect anomalies. This can be configured in the company’s GUI, and it’s as simple as setting the model, targeting specific settings, and selecting specific technical objects. This training of AI can be done easily, even without IT or data scientists.
Then, the technology can quantify how abnormal the incoming data is, based on the historical pattern, providing an early signal of a potential failure. And with a new capability called Visual Asset Health Monitoring, AI assists and streamlines the visual inspection process. SAP Asset Performance Management works with the AI models to bring in image data, and the AI analyzes if the asset is performing correctly.
Data alone doesn’t perform the work. Turning the data into actions, by connecting the reliability process with maintenance, is what it’s all about, Lee added.
He advised making sure that maintenance organizations trust the insights that are coming their way. Otherwise, the predictive maintenance isn’t going to be trusted or used.
Ultimately, SAP Asset Performance Management aims to help reliability and maintenance teams identify and target the right assets, and work together to streamline inspections and keep assets running smoothly. “The key here, of course, is to make the process as data-driven as possible, creating confidence and trust between reliability and maintenance,” Lee said.