One of the biggest challenges for newer, more sophisticated data analytics projects and software is securing the information they require from legacy process control devices and systems—and delivering it quickly. To keep on serving up all the signals and results that analytical teeth need to chew on, several suppliers have developed software and support products to keep the all-you-can-eat data portions coming.
ClearBlade and Phoenix Contact
For instance, after partnering for two years, Phoenix Contact integrated ClearBlade Inc.'s Intelligent Assets turnkey software with its PLCNext scalable controllers so users can look at data streams from sensors and other devices that can’t do their own data processing. No programming or data scientists are needed to establish these links, stream their information, upload it to an analytics platform, and get it into a useful format such as a rules alert or an artificial intelligence (AI) plug-in.
“This gives us an open platform on PLCNext that we can attach to dumb assets, and turn them into intelligent assets by using ClearBlade’s edge, run, compute functions, which can be uploaded via IIoT to on-premises analytics of the cloud,” says Dave Eifert, senior business development for IIoT at Phoenix Contact.
Eric Simone, ClearBlade’s CEO, adds: “We can automatically stream tags data into our Intelligent Assets software and produce human-readable values. This lets users quickly read compressor or pump data for example, and add alerts or assign actions, all without complex coding. We don’t replace existing supervisory control and data acquisition (SCADA) systems, we just let operators simply analyze what they want, and allow them to streamline their operations.”
Besides handling a few or hundreds of I/O, Eifert and Simone report that PLCNext and ClearBlade can also handle very remote locations that can’t send all their data back to a central location by instead performing some analytics and AI functions out at the edge, such as developing and reviewing models to direct and optimize equipment at the edge. They add that AI and models can perform mundane tasks; augment what humans have been doing in the field; reduce trial-and-error “science experiments” by IIoT; reduce the need for system integrators to design and build time-consuming, customized, and often-riskier solutions; and deliver return on investments (ROI) more quickly.
“Oil and Gas drilling data streams may include information about ingredients, mixing, penetration rates and viscosity, as well as what to do based on present conditions or those observed in recent weeks or months,” explains Simone. “Operators can use these results, automatically apply them to AI models to automate processes, such as reducing RPMs when feasible to increase equipment longevity.”
Eifert adds, “This is a convergence of technologies that have been available in process automation for years, but are finally easier due to edge computing. This is why Phoenix Contact and ClearBlade work together.”
Control Station
To provide production staff with actionable and state-specific control loop KPIs and analyses, Control Station Inc. reports its software distinguishes between various process conditions and states, and provides advanced, state-based insights using a facility’s existing historized process data.
“Process manufacturers are increasingly looking to control loop monitoring solutions for intelligence on how to improve a plant’s PID controller performance, but data from different product runs, different batches, different load levels are all included in the analysis. This comingled information is rarely helpful when the goal is to optimize each of the many and unique production states in a process,” says Bob Rice, VP of engineering at Control Station. “Simple averages tend to mask extremes in controller performance and obscure the production states that undermine efficiency and throughput. The state-based analytics capability in PlantESP lets users apply loop analytics to distinct operating states, and pinpoint opportunities for performance improvement.”
In addition to this new state-based analytics capability, PlantESP now complies with Representational state transfer (REST) application program interface (API) requirements for uniform interfaces. This lets other analytics and reporting tools like TrendMiner, Wedge or Microsoft’s Power BI software query PlantESP’s KPI results and to incorporate controller performance information in advanced data analytic investigations.
Rice reports that large site and multi-site manufacturers typically have multiple silos of data, whether it comes from loop monitoring, process historians, manufacturing execution systems (MES) or other self-contained databases. This results in islands of automation data that unnecessarily limit a manufacturer’s enterprise analytics effort. With its REST-compliant architecture, PlantESP helps manufacturers to bridge across those islands so that PID controller data can be used as part of more strategic data mining initiatives.
Control Station also conducts 30- and 90-day evaluations of its PlantESP solution at customer sites during which their engineers benchmark existing PID loop performance, identify controller issues and associated corrective actions, and estimate a return on investment (ROI). This limited use of PlantESP allows the company to demonstrate how improved controller performance can have a significant and sustainable impact on a plant’s long-term financial health.
Fluke Reliability Systems
After about 70 years of making instruments that generated data that usually went onto clipboards, Fluke Corp. logically wanted to get that information to cloud-computing services, and develop software that could help users take advantage of it. Consequently, the company launched Fluke Connect cloud-based data repository for its instruments. More recently, it acquired eMaint, a cloud-based e-maintenance software platform in 2017, which lets users produce work orders in response to condition-based data and multi-variable, predictive analytics.
“In addition to being leaders in the CMMS space we’re investing heavily to innovate and build industry leading software that can be used for analytics automated analysis leveraging AI,” says Mitch Kruse, global software product marketing manager at Fluke Reliability. “Linking this to our CMMS software will give maintenance teams a full end-to-end solution for their reliability program. The hurdle everyone’s facing now is needing a lot more information to inform and train software algorithms. This is where Fluke has an advantage with decades of data generated from handheld tools. Training the software to recognize patterns in sensor data to find potential trouble before it happens is how AI will revolutionize maintenance. The root cause of any machine failure often involves many variables - some that may not be obvious using traditional troubleshooting methods. Truly ‘listening’ to a machine involves much more than occasionally checking it while on an inspection route. It needs constant monitoring with Industrial Internet of Things (IIoT) sensors and adaptive software. For example, even though people would change the oil in their vehicles and equipment, but they’d still break down due to events like vibrations they weren’t monitoring before.”
Kruse reports that Fluke software will incorporate multiple parameters of users’ assets in addition to vibration data into its rule-based algorithms and compare with known library of vibration signatures. “This lets all of a user’s vibration data feed into an algorithm that can be compared with the profiles and prior wear and failures by other machines and equipment. Users can employ this library to set up new systems and speed, RPM and amperage ranges for all kinds of rotating equipment, such as pumps, compressors, fan motors, belt or shaft drives and shaft couplings.”
These capabilities are being incorporated into Fluke's new Live-Asset products, where users can input data feeds, such as present machine status, and with optional modules get back predicted conditions. LIVE-Asset automated analysis, coming in 2023 is built on AI-based, analytics software from SymphonyAI that provides predictive insights.
Most recently, Fluke acquired Pruftechnik and its alignment technology, vibration sensors, handheld tools and analysis software in 2019. Its vibration analysis software products are being rebuilt for hybrid on-premises and cloud-based deployment, so technicians and other users can automate their analytics if they don’t want to do calculations themselves, and just want to get alerts when items like bearings need to be changed.
“Previously, maintenance was just a necessary evil. Now, maintenance teams can flip it, and show substantial bottom line savings by predicting and preventing costly downtime on critical assets,” adds Kruse. “When users perform a FMECA (Failure Modes Effects and Criticality Analysis) assessment and understand the financial impact of downtime maintenance can be prioritized and assigned according to asset importance, cost of stops, downtime, likely failures and needed replacements. Next, this can be coupled with predictive maintenance based on condition monitoring and AI that can handle sensor data.
"This can make maintenance more effective and more cost-positive because we’re tying analytics directly to plant reliability, and we know how to get more uptime at less cost," Kruse continues. "Knowing what devices aren’t going to fail and confirming what’s important is a more intelligent way to manage a maintenance and reliability program because we’ve learned to trust our software tools and we’re not guessing anymore.”
