“How long do I have until this machine fails?”

That’s a key question always on the mind of a plant operator, in the water sector and elsewhere. Preventive maintenance “by the book” is fine, but how can an operator know when something big is about to go wrong with a motor, pump or engine so it can be fixed before a major failure and the downtime and lost production that go with it?

GE Intelligent Platforms now offers the Proficy Monitoring & Analysis software suite that can help provide answers and in the bargain can help operators optimize process efficiency and throughput. The platform combines four existing GE offerings with a Proficy Knowledge Center browser-based visualization tool, and Proficy Historian HD, a Hadoop-based Industrial Big Data product.

Brian Courtney, general manager of the GE Intelligent Platforms Industrial Data Intelligence Software Group, talked about the offering in an interview with Water System Operator.

wso: What overall need in the drinking water sector does this offering address?

Courtney: In water treatment you have large assets like pumps, motors and generators. How do you optimize the system? A lot of system software today is about alarm and event — how to react when something goes wrong. Analytics are about how to prevent things from going wrong in the first place. We provide tools to help figure out what could go wrong.

Our Proficy SmartSignal system ‘learns’ the behavior of each piece of equipment by monitoring the equipment itself for about two weeks. From that we’re able to identify what normal behavior looks like for each item. Then as we continue monitoring, 24 hours a day, we’re able to identify outputs that are not behaving correctly.

wso: What kinds of anomalies are detected and what is done with the information?

Courtney: We have various analytic tools to help us identify issues based on known failure modes for large pieces of equipment. An abnormal vibration could mean a bolt is loose. Rising temperature could mean a fan has shut off. Bad data could indicate that a sensor has failed. Sometimes it takes collaboration with the customer: We can tell that the equipment isn’t behaving correctly, but it takes a person on site to do diagnostics to understand the problem.

wso: What is the ultimate benefit of this technology?

Courtney: We detect minor issues that will lead to big problems if they go unchecked. If you take asset reliability a little bit higher, you get additive value: Every day you can treat more water. The way to measure success is: How much unplanned downtime did you have to start with, versus how much you have today? And as a result, how much have you saved, both in the cost of repairs and in recovering lost revenue from downtime and lost production?

wso: How does this offering support process optimization?

Courtney: The Proficy CSense toolset has advanced analytics that identify process variables that can lead to suboptimal performance. It allows for what-if analysis to find solutions that optimize processes and so increase yield. If you can get the variability out of a process, you can run it closer to the optimum. We have another set of analytic tools for doing closed-loop process optimization — similar to advanced process control.

wso: How exactly is the data acquired for each piece of equipment?

Courtney: Data is collected and stored with GE’s Proficy Historian data collection software. We work off existing sensor data. We assume there is a SCADA system or a historian application on site that we can pull data from. We tell customers what kinds of analytics we can do with the data they have available. And we let them know what additional information we could obtain if they were to acquire additional sensors.

wso: Can this offering be used for benchmarking? Could customers use the data to see how their equipment performs versus other similar equipment industrywide?

Courtney: We’re building that capability now. We would use a comparative analytic to show how a given customer’s equipment behaves versus an industry average, as defined by all the other similar assets we monitor. This depends on customers’ willingness to share their data anonymously. It is something customers would need to opt into.

wso: Do the analytics apply to all equipment and not just devices made by GE?

Courtney: It is designed for any asset that has rotating or vibrating components. We don’t do electrical boards or electrical circuits today, although we are working on solutions in that space.

wso: Is there a size threshold at which this offering becomes cost-effective?

Courtney: On the asset health and process health side, any organization can get value, and obviously the bigger you are the more value you get. I think anyone doing water treatment is in the size range of organizations we could help in a cost-effective way.

wso: How is the cost of this offering structured?

Courtney: We offer it in two ways: As a license deal where users buy the software and operate it themselves, and as a service where we do the monitoring, setup and analytics and the user pays a monthly or quarterly fee with pricing set by duration of contract. When we do the monitoring, we typically don’t go into the field and fix equipment. We call customers weekly and identify things we see as potentially failing in the next several weeks. If we detect something at imminent risk, we contact the customer right away.

wso: What do customers actually see when using this offering to help them visualize and understand issues?

Courtney: A part of this offering is Proficy Knowledge Center, a model-driven, browser-based visualization application. Users can view results of data analyzed by Proficy CSense or Proficy SmartSignal in ways that make it easy to navigate and understand. They can see trend lines, key performance indicators and advisories. There are also simulation and modeling capabilities. For example, users can explore how an action they might have taken would have affected a given problem.

wso: Does this solution require on-site IT infrastructure? Or are cloud-based applications available?

Courtney: We offer Proficy Historian HD, which can extract data from Proficy Historian and store it into Hadoop clusters so that it is available in a cloud format. That is a very cost-effective way of storing large sets of data.