May 31, 2021
By: Lyat Avidor Peleg
How to Leverage Existing Data to Monitor ALL your Plant’s Valves
Digital transformation is spreading across process plants, but there are so many different areas to address and so many different solutions to use. One of the approaches that brings the most immediate value is predictive maintenance or PdM.
PdM can make an enormous difference to the profitability of a process plant. No failure arises in a vacuum; there are always small, early signs that something minor isn’t going the way it should. Gradually those signs grow, becoming more serious and more obvious as time goes by, until there’s an unmissable crisis. The sooner you catch and repair the issue, the lower the cost in time, money, and lost production.
All these signs are present in the process data that each plant already captures as a matter of course. If a process engineer, maintenance expert, or any other domain expert were to look at the data, they would spot far more nascent failures long before they become serious. But of course, each process plant has hundreds of sensors. There’s no way that any human can keep track of so much data, 24/7.
But AI can, which is where PdM is born. Problems can and do take place every day, somewhere in the plant, leading to increased costs, wasted resources, flare events, safety incidents, environmental damage, unplanned downtime, low quality product, and more. PdM solutions can monitor all the thousands of tags to spot the earliest signs of failure and enable process engineers to correct them before the failure takes place.
Valves responsible for 35% of plant failures
Most companies already have predictive monitoring solutions in place for their most expensive equipment, like turbines, compressors, and other unique non-redundant machinery. But plant issues can come from any direction, not just from the important and expensive items, and valves make up a significant proportion of them. In fact, up to 35% of all plant failures derive from valve issues.
This means that ongoing valve monitoring is the perfect place to start with a PdM solution. SAMSON, the parent company of SAM GUARD, has been manufacturing valves and carrying out valve diagnostics for 114 years. We run testing facilities and innovation centers where we test valves constantly, so it made sense for us to bring our expertise to a total valve monitoring solution.
We felt that in addition to our complete PdM solution, which is running effectively in many process plants, there is a need for a simpler, faster, dedicated solution for rapid valve predictive monitoring. We offer it as a service, to make it easy and stress-free for process plants to implement and run, covering every valve with more efficiency than a control room solution and better results than software that checks only certain valves.
How SAM Valve Monitoring can monitor all your valves
Our PdM valve monitoring service is a combination of advanced technology plus decades of knowhow and experience at making and maintaining valves.
We know that a valve monitoring solution can’t measure valve data alone. If you examine valves without the context of the rest of the plant, you’ll mostly draw the wrong conclusions. You need to understand what affects each valve, otherwise you’ll receive a number of false alarms.
That’s why our PdM software goes beyond using machine learning to analyze data and also applies plant context. We take data from the P&IDs, convert it into a plant model which describes all the dependencies and casualties within the plant, and then convert that knowledge map into a graph. With this context data, we can train our model to know what to expect when valve X is opened or valve Y is closed, and how any pump, compressor, or combination of items of equipment affect each valve.
Valve monitoring as a service means that we filter the alerts and help you choose how to respond. This added layer of expertise helps prevent the typical false alarms that noisy process data can generate if context is ignored.
The SAM GUARD software is based on proprietary algorithms that analyze your data and detect anomalies that merit an alert. It sends alerts to a SAM GUARD engineer, based in our analytical monitoring center (AMC) in Frankfurt, who examines the alert, applies their expert knowledge, and then chooses whether to share it with you, or continues to monitor it for a while to see how it develops, or dismisses it as simply noise.
What you need in order to implement SAM Valve Monitoring
Our algorithms are ready to go, and have been proven effective in several plants. You can learn more in this webinar, where we discuss these steps towards implementation in more detail:
- Draw up a list of the valves you want us to monitor
- Share the process data you’re collecting anyway, ideally, around 12 months of historical data
- Send some additional information like process book diagrams and schemas
- Decide whether to use the SAM GUARD servers or your own servers
SAM GUARD’s valve monitoring solution is completely vendor agnostic so it can work with every valve and every data storage device. Our server centers, network, and valve knowledge experts are all ready for you. It’s easy to get connected and protected.
When valve monitoring is within reach, the value is clear
Early detection of impending failures saves repair time, cost, and productivity within process plants, and that’s possible when you use PdM. In the webinar, we discuss three specific use cases where the valve monitoring solution prevented material waste, safeguarded end product quality, prevented a potential safety and environmental incident, and stopped damage from occurring to the plant.
The SAM Valve Monitoring PdM solution can monitor all the valves in your plant, no matter where they are located or which vendor they are from, to deliver another layer of protection in your plant.
It takes just a few days to set up the SAM Valve Monitoring solution. Watch the webinar to learn more about how to implement SAM GUARD remote valve monitoring in your plant.