Hello, I hope you find this post helpful. I loved working through this case because pi blows it out of the water and many folks may not be aware of that. The white paper summarizes the key steps. We'll look at a group of ANSI centrifugal pumps. The goal is to take you beyond real-time alarming. Your SCADA/DCS already does that. We want analyses that dynamically look at a pump's history (weeks, months, or years) and publish forecasts of Life Expectancy. The better answer is in the history. I think world class teams forecast. They constantly seek to know the future and they publish their forecasts in order to improve them. In this post we will show you how to calculate and publish forecast life expectancy with pi using email messages and web displays, .
A cool concept emerged from this work. What good is insight that can't be understood by everybody? The beauty of forecasting and where this case surges further is when it becomes simple beyond complex. Let me explain. When you do the complex math to forecast you can communicate in simple dates and time units with all users. Life expectancy, How many more months will this machine last? This means that operators, mechanics, supervisors, executives, or anyone will understand the results and gain insight.
Special thanks to all of these folks for contributing to this work: Ted Birky, David Gray, Joe Van Roosendaal, Carroll Sparks, Chris Felts, Heath Howland, Richard Beeson, Rachel Hemmer, Norton Green, Stephen Kwan, Jeremy Korman, Alexander Fiset
- White paper (Summary excerpt below): Using PI Server Analytics to See Rotating Equipment Life Expectancy: Best Practices for Consuming Wireless IIoT Smart Vibration Data to Reduce Downtime
- Samples of Asset Framework referred to in the white paper. (AF EXAMPLE KIT-RotEquipHealthForecasting.ZIP)
Using PI servers to analyze overall vibration readings is a breakthrough. Smart EDGE/IoT wireless vibration sensors are opening new doors. The low costs of connecting these data sources to PI servers allow you to stream and store life expectancy on machines that have never been monitored. This will create a tremendous boost in reliability and safety performance for many companies.
In this post we present best practices for consuming overall vibration data with PI servers. We demonstrate how to forecast machine health, alert users and capture feedback with PI Analysis Service. We provide sample templates for PI Asset Framework. We demonstrate effective dashboards and event visualization with web displays in PI Vision™ . We present several best practices that leverage local operations staff as key data consumers. We show you how to configure a standard PI system to stream forecasted machine failure into future data. We demonstrate “plug-n-play” flexibility techniques needed when sensors are often moved between machines in the field. We give you guidelines for selecting smart vibration sensors. We present how to build vertical asset health rollups. We present best practices for notification of machine health issues using eMail and SMS messages.
PI servers provide visibility into rotating equipment operation. Boosted tag counts streaming from production operations transform real-time operational data into fast, actionable insights. Barriers between Operational Technologies (OT) and IT are removed with PI Asset Framework. PI web-based visualization tools accelerate discovery through mobility.
Plant Engineering designers and managers
Plant Operations managers
Plant Maintenance managers, Engineers, Mechanics
Plant Process engineers and Process technologists
Plant Reliability Engineers
Unplanned rotating machine failure drives up costs and risks injury in nearly all plants. If you could only see the failures coming for more rotating machines in your plant…
More online sensors. Fewer inspections. Fewer spare parts. Improved safety. These are powerful moves toward world class. The only questions are… how to do it well and avoid overspending.
A PI System and EDGE/IIoT smart wireless vibration sensors can make it happen.
Use PI servers to log overall vibration data, analyze, forecast and inform your decision-makers. PI does the work behind the scenes. What you experience is simple beyond complex.
Witnessing the early stages of rotating equipment failure is essential to finding latent root causes. Response times are measured in minutes. Online PI analytics near the asset give local operations team’s immediate focus.
Traditional vibration monitoring uses contracted experts, hard-wired sensors, dedicated processors, and dense data storage. This “deep dive” approach is well suited for ultra-critical assets with high replacement cost. But what about the rest of the plant?
In power generation there is a class of assets called Balance of plant (BoP). These are the large quantities of pumps and fans that support the site. Often these are not monitored for health due to cost of traditional hard-wired sensors. These now are prime targets for improvement.
Don’t overspend on complex layers of contracted service-provider relationships that ultimately are too distant to sustain. With the PI system, keep a larger portion of machine health decisions within the reach of your operations staff.
We will show you how in this post.
For rotating equipment, vibration data from accelerometers is the fundamental visibility into machine health. For BoP assets you often don’t need very much detail to be effective. Replacement costs are low making the decision-making thresholds low. Early detection of machine failure can boost repairs and reduce replacements.
The goals of this post are to help you answer questions like these for all of your machines…
- I know the machine is sick but is it getting worse?
- How much longer can this machine run and how confident are we?
- If we know life expectancy do we need a spare?
- Did we get a good repair?
- Has this machine or any others like it ever had these symptoms before?
- How can we teach a new employee the history of each machine?
- Who saw the machine the last time it failed and what did they say about it?
- Did we capture the root cause from the last time it failed?
Using PI systems to store and analyze overall vibration readings is the breakthrough method presented here.
For BoP assets, the operations teams simply seek “clean” machines. You need less about why the machine is not healthy because the corrective action is often to swap for a refurbished unit. You push the heavier diagnostics work to the offline repair teams. Back in the plant the mission is simpler. You confirm a machine is “clean” on entry, monitor online, detect “unclean”, track progression and plan for maintenance accordingly.
There are three pieces that must be done well to make this method work.
The human element is still today the most important consideration. In plants it is the people that make the operating decisions. I read recently an article in Scientific American by Mark Fischette where IBM concluded our computing capability equates to the brain power of a house cat. I always knew cats were smart. IBM estimates that it will take 880,000 processors to demonstrate the capability of a human brain. We are best at creative thinking and our computers are best at repetitive operations. It is important to understand this and design for it. We will do that here.
The second piece is IIoT hardware. It’s what we techies all want to talk about. The Industrial Internet of Things (IIoT) is truly moving the needle. Wireless sensors with on-board processors are scrubbing and delivering data at ultra-low cost. It is important to choose the right sensors.
Finally, a PI system must be in place to store data, shape it, slice it and notify the right people. It must be intelligent so as to not add labor. Streaming analytics are the repetitive tasks done for us by PI. The PI system can store, aggregate, scrub, forecast and alert you. PI is ready to go to work. Web-based visualization, SMS text messages and email messages pull users in the loop on predictive maintenance.
Check out the white paper below for the FULL STORY.