EE - Forum Styles
fo

Up digital-transformation profits

Paul Studebaker

It’s becoming clear that taking profitable advantage of digital transformation depends on the ability to draw on the right set of a vast and evolving array of technologies. No one facility needs them all, at least not right away, but the odds of success will rise with wise selection.

Numerous presentations at Emerson Global Users Exchange in Nashville, Tennessee, made it clear that Emerson has been working to make available a full range of “digital transformation technology enablers,” according to Peter Zornio, chief technology officer, by extending the company’s portfolio of “operational analytics,” the technologies that bring results where the greatest profit potentials are commonly identified: reliability, performance and energy efficiency.

Zornio announced four new Plantweb Insight applications, providing templates for monitoring, troubleshooting and optimizing cooling towers, pressure relief valves, networks and power modules. Plantweb Optics v1.5 asset performance management platform expands connectivity and increases collaboration, and new Digital Twin software-as-a-service for well exploration and production has added a model for subterranean systems.

Advancing the operational analytics portfolio

“Customers want to use analytics to look at data, convert it to information and use it to improve plant performance,” said Manasi Menon, product manager, analytics and machine learning with Emerson’s Automation Solutions business.

Operational analytics often use principles-driven analysis on individual units and data-driven analysis on the plant—a combination of units. Principles-driven analytics are for equipment or systems that follow known physical principles, such as gearboxes or steam traps. Data-driven analytics use statistics and tools such as artificial intelligence (AI) or machine learning to derive correlations and probable causes where interactions are not as well-defined.

Plantweb Insight principles-driven analytics are available for common and some less common assets, from pumps and heat exchangers to corrosion and cooling towers, including the recent additions of power modules and network management.

“Get answers from your hidden data, and make them available through PlantWeb Optics.” Manasi Menon, product manager, analytics and machine learning with Emerson’s Automation Solutions business.

Integrating KNet software, acquired in April, brings data-driven analytics and first-principles failure mode and effects analysis (FMEA) templates for 492 asset classes. Data and information acquired through standard connectors such as OPC, control systems (DCS or PLC), process flow diagrams, FMEA and historians can be processed by KNet process and asset models, machine learning and online analytics to provide actionable failure mode analyses, workflows, predictive insights and key performance indicators (KPIs).

For example, on a debutanizer, Reid vapor pressure (RVP) is controlled to maintain efficiency. “KNet can calculate and predict RVP, using root cause analysis embedded in decision and fault trees,” Menon said. “The 492 FMEA templates can be embedded in the fault trees for faster analysis.”

Going forward, “KNet will be integrated in Plantweb Optics so customers using Optics can embed more predictive maintenance,” Menon said.

As a result, “former silos are integrated,” Menon said. For example, information from AMS Device Manager (instrument and valve health), DeltaV (control loop performance), KNet (first-principles and data-driven analytics) and AMS Machinery Manager (machinery protection and health) can be presented to mobile devices and augmented-reality (AR) systems, integrated with computerized maintenance management system (CMMS) workflows, and rolled into Plantweb to provide persona-based content delivery to operators, engineers and technicians.

“Augmented reality, external data and enhanced Optics for faster views are coming, and services are available to develop these systems for you,” Menon said. “Get answers from your hidden data, and make them available through PlantWeb Optics.”

1 Reply

  • Grab sampling product, sending it to the lab, and waiting for results the traditional way take time before you can take corrective action such as making setpoint adjustment. Using predictive analytics you can predict product properties such as Reid vapor pressure (RVP) much faster. Thus you can make any adjustments to setpoints etc. much faster to make sure product is within spec. In the same way you can predict other process problems and prescribe corrective actions, which process operators otherwise would not have seen coming, or may not be sufficiently experienced to tackle. Analytics alone may not be able to do this reliably and timely based on the data coming in from existing sensors and systems. Very often you need additional data to reliably predict changes and upsets, and prescribe the corrective action. Wireless add-on sensors are ideal for this. Additional pressure, flow, temperature, and other data that feeds into the analytics to provide stronger correlation for robust and trustworthy results. Software is not enough. Sensors is not enough. You need both sensors and software. It is works best when they are all built as part of the same digital ecosystem. Learn what other plants are doing from this essay: www.linkedin.com/.../