Predictive Maintenance / Prescriptive Maintenance / Asset Monitoring

Platforms for predictive maintenance and prescriptive maintenance progress

Sheila Kennedy says tools from pattern recognition to self-powered sensors power advanced maintenance.

By Sheila Kennedy, CMRP, contributing editor

Asset reliability is a maintenance organizations’ Holy Grail – highly desired yet difficult to consistently achieve. Fortunately, new software and technologies for predictive maintenance (PdM) and prescriptive maintenance (RxM) are bringing the goal closer to reality.

Insightful solutions

Artificial-intelligence-driven predictive and prescriptive analytics platforms let factory workers exchange data with remote maintenance teams and equipment manufacturers, receive predictive alerts, and prescribe solutions. One example is MachineMetrics. The ability of this platform to connect easily to any machine has allowed it to quickly aggregate a powerful global data set of machine performance, says Bill Bither, CEO of MachineMetrics.

“That data is empowering MachineMetrics’ AI and machine-learning algorithms to predict almost any shop-floor problem, whether it be when a machine will need maintenance or a spindle about to crash or a tool that’s worn down, and prescribe solutions via alerts through fully automated workflows,” adds Bither.

SmartCBM, Allied Reliability’s new IIoT-based condition monitoring and PdM solution, is capable of prescribing corrective actions. It combines IIoT architecture, including wireless sensors, with Allied’s proprietary failure-mode library to monitor asset health in real time. Its “enhanced” deployment option supports virtual and augmented reality capabilities and includes integration with the PTC ThingWorx platform for improved visualization of condition and process data.

“SmartCBM leverages Allied’s expertise in reliability engineering and data analytics to deliver early diagnosis of defects, identify the root cause of failures, and prescribe corrective actions that enable proactive decision-making,” says Robert DeLaGarza, product manager of CM services/SmartCBM at Allied Reliability.

OEM-specific solutions also are available. For Sullair compressor customers, the company’s AirLinx compressor remote monitoring software provides preventive system warnings and alerts via a dashboard on the user’s smartphone or tablet or any device with an internet connection. Users can view details, trend data, and remotely troubleshoot a condition before they send out a technician.

Manhar Grewal, R&D product manager at Sullair, says that what differentiates AirLinx is that it “(allows) the customer to set any preventive parameters they want to be alerted on – anything from the line pressure hitting a certain psi to the machine’s ambient temperature.” He adds, “This allows for greater predictive maintenance, where the customer can receive customized, advanced warnings and alerts via text or email.”

Diagnostics and prognostics

A monitoring system should not only warn about problems, but also it should provide an accurate diagnosis with the specific component identification, location, and indication of the extent of damage, suggests Jost Anderhub, product manager of software at PROGNOST Systems.

PROGNOST-Predictor is an online condition monitoring and component diagnostics solution for rotating equipment. The company’s patented Confidence Factor technology, applied on each analysis, leverages pattern recognition to indicate the degree of confidence in whether a measurement represents a real fault or damage. “To calculate a reliable projection, it is mandatory to take all the experiences (i.e., historical data, real-time signals, and other analysis results) in parallel into account,” adds Anderhub.

Senseye applies automated prognostic analysis, dynamic time warping of data, AI, and fleet clustering to its scalable PdM solution. It can automatically account for operational differences of various machine measurements, present the output of prognostics for a given asset, and propagate the prognostic risk up through the hierarchy of asset and organizational levels.

“Proprietary machine-agnostic algorithms characterize and learn from machine failures to provide the remaining useful life for any asset type, automatically and without any bespoke model development, says Rob Russell, CTO of Senseye. “This eliminates the need for costly data analysts to build and interpret custom data models and constant modification as new machines are added or operating conditions change.”

Uninterrupted condition data

Battery replacements for wireless sensors may soon be a thing of the past. PsiKick’s ESP platform uses extremely small amounts of harvested energy to power its suite of completely self-sustaining industrial sensors. For example, PsiKick’s first product continuously monitors steam traps by leveraging the energy of steam passing through the steam trap to power the sensors that monitor the unit’s condition.

Temperature differences of 1°C or 100 lux of light (less than most dimly lit facilities) can generate all the power required for PsiKick’s devices to measure, process, and wirelessly transmit data, says Shadi Hawawini, VP of product management at PsiKick. “We’re applying this technology to condition-based monitoring of assets such as steam traps and pressure relief valves – and this year to rotating equipment – throughout a range of industrial environments,” he adds.