Condition monitoring is exploding. All manufacturers of any type of equipment are putting sensors in their machines,” said Michael Fry, a director with CIMdata.
This growth of sensors is not in isolation, he is quick to point out. While the sensors are getting cheaper and more powerful, they would be useless if there was not a way to use that data in a coherent manner.
Their growing existence and use is driven by an expanding information ecosystem.
“We can call it Industrial Internet of Things or Industry 4.0, or something else,” explained Fry. “The point is that there is a structure to this emerging system, a stack if you will, where at the bottom of the stack are sensors, feeding data for something like predictive analytics and other enterprise applications at the top of the stack.”
Sensors and condition monitoring is not new, of course. Operators of machinery have been keenly aware that a few fundamental measurements can help them keep critical assets running smoothly.
“Even a 30-year-old CNC machine will have some sensors and will have its control data fed to it through its MES system,” he said.
What is different today is the magnitude of the data available, and the ability to make sense of the data deluge.
“A new CNC machine will have sensors all over it for spindle speed, loads of various types, vibration, or timing,” he said.
This will quickly surpass the ability of a human or simple processing tools like electronic spreadsheets. to make sense of it.
What to do with it all?
Answering that question is what his consulting practice at CIMdata is all about. Fry believes the key is the marriage of advanced data processing and distribution.
“Data modeling tools dissect that data and send it back, typically through a comprehensive PLM system, to where it can be used such as in quality control, engineering, sales, whatever,” he said.
A place for everyone
The emerging infrastructure for this Industrial Internet of Things, or IIoT, in general and for condition monitoring in particular is huge—beyond the capability of any single organization to provide it. That is according to Brett Burger of National Instruments. “No one company can stand up and say we can do it all,” he said. “It is really important to talk about ecosystems and the ability to connect to adjacent technologies.”
In that ecosystem, offerings from National Instruments play a central role in collecting and transforming data from operational sensors into information systems used in monitoring and decision making. That is how they position their LabVIEW software, as a platform for customizing any number of such transformations. He points to a technology ‘edge’ between operational technology and information technology. “We pride ourselves on being an open platform, and that means reading and converting any kind of sensor data, typically in analog, converting it to a digital format and then extracting the features as needed,” he explained. “In LabVIEW we have over 1000 Intellectual Property, or IP, blocks for feature extraction.”
While LabVIEW is a development environment, National Instruments also provides application layer end-use software tools built using that very environment. “We did this because of customers using these tools in specific vertical areas, and condition monitoring is an important one,” he explained.
Just as important is the ability to provide data to downstream IT users, such as ERP, MMS or PLM software. “LabVIEW is an open platform so we have APIs or can provide data in any file format that those in the IT world might want to use,” he said.
Wireless sensors reduce cost
Burger also sees the importance of wireless sensors as critical to the growth of condition monitoring. Wired sensors require conduit, power supplies and often skilled labor to install. In some instances, this could make the installation more expensive than the sensor itself.
Fluke Corporation also thinks wireless is important. It has been investing in bringing a series of wireless sensors to the market, along with an interoperable platform it terms Fluke Connect. “Fixed condition monitoring has existed for some time,” explained Pronitha Shankarananda, business unit manager for Fluke. This is especially true for highly critical assets, those that can quickly shut a plant down. “What is changing is there are simple screening systems, systems which will let [organizations] monitor non-critical or semi-critical assets,” she said.
Shankarananda also sees sensor costs coming down, sensors that are both more sensitive and with better processing capabilities. “But what might be even more important is integration and interoperability capabilities and making sure there are no data silos in these organizations,” she explained. “Our idea is to democratize condition monitoring. Make it simple and easy to install.”
Fluke announced late in 2016 that its Fluke Connect reliability platform now offers cloud-based condition monitoring. This includes a system of voltage, current, temperature, and power sensors that can be moved from asset to asset or left in place for continuous monitoring.
The Fluke Connect reliability platform employs wireless sensors and a gateway that receives signals from the sensors from up to 30′ (9.1 m) away and stream to mobile app or desktop via secure cloud servers. “The cloud connection is either through an ethernet connection or via WiFi,” explained Shankarananda. “You can view the data live via a laptop or any remote visualization process” such as on a smartphone.
Maintenance technicians can set the system up and begin monitoring in a matter of minutes, with the sensors transmitting measurements to the cloud as frequently as one measurement per second, according to the company. “Predictive maintenance has been out of reach for most assets except for the most critical ones, now with this easier system, condition monitoring can be pushed down the criticality matrix with an adequate return-on-investment,” she said.
A comprehensive perspective is provided by J P Provencher, VP for Connected Operations Solutions at PTC, the PLM company that also owns the ThingWorx IIoT platform. ThingWorx is a complete platform designed to enable the Industrial IoT, according to Provencher. These modules include connecting machines and sensors, collecting and transmitting the data, and analyzing it to review it and make predictions. There is even an augmented reality ‘studio’ module to enhance the user experience in interpreting the data.
The ThingWorx user community is broad and includes machine providers and their customers. “We have both customers who are monitoring assets from within the four walls of the factory and customers who are monitoring assets from an OEM perspective, remotely connecting, diagnosing, and servicing those assets,” Provencher said. In this case, think of a CNC machine provider who wants to monitor their installed base and use data collected from the entire fleet of machines for predictive maintenance, quality reviews and influence future machine designs.
Provencher’s philosophy is that IIoT is a team sport. Partnerships are particularly important, such as with National Instruments for access to sensor data and with OSI Soft to transform machine time series data into new business value. “PTC works with 26,000 different manufacturers which gives us a unique opportunity for discussions around needs and use cases inside and outside the factory,” he said.
Another interesting gage is a recent survey conducted by PTC. “Our survey indicates that 50% of manufacturers will run an IIoT pilot or are running one this year,” Provencher said. “Whether you are a plant manager or an OEM selling equipment, people are recognizing that IIoT is a significant business opportunity.” He believes the future is coming up fast. Technologies today, such as wireless sensors and software platforms like ThingWorx mean that new applications can start fast—in as little as six weeks according to him.
There is also the simple question of what to do with that data. While predictive analytics and big data analysis will continue to grow, there may be simpler ways to achieve value.
“Simple” is the keyword for companies like Leading2Lean with its CloudDISPATCH for real-time data aimed at visibility, improvement and operations management. CloudDISPATCH gets its data from manually entered information or from sensors. Some customers are “integrating machine and sensor data into CloudDISPATCH and doing innovative things,” said Bob Argyle, CCO for Leading2Lean.
How does it work? “Once a machine reaches a certain threshold, or condition where there’s maintenance or a fix that needs to be addressed, CloudDISPATCH will automatically initiate the dispatch of a technician to resolve the issue. Otherwise, someone has to be constantly looking at the data and ensuring it gets fixed,” Argyle said. Machine alarms are often ignored. It could just stop running until it’s fixed. “If the machine stops running, you then have downtime that comes at a cost,” he explained.
“Essentially, the machine condition is generating a human response,” he said. The machine creates the event and is bringing in a person that can resolve the issue before it becomes a much bigger issue and causes the machine to go down.
Condition monitoring in process industries, from oil and gas to pulp and paper, is especially important. “We work with customers where losing a critical asset can cost millions of dollars per day,” said Chris McMillen, a product manager for the Bently Nevada division of General Electric. They provide a host of sensors, such as vibration and proximity sensors, as well as condition monitoring and diagnostic software. System 1 is their condition monitoring software platform for real-time optimization of equipment and selected processes, condition monitoring, and event diagnostics. McMillen also said that GE’s Predix IIoT software can also be part of a comprehensive monitoring solution.
An interesting point he made is that the variety of sensors is also contributing to both an increased ability to know what is happening and the chance to get overwhelmed with data.
“There are sensors for temperature, vibration, proximity, acoustic, ultrasonic,” McMillen said.
As an example of how sensitive these cost-effective sensors are getting, McMillen noted that one of their noncontact proximity sensors measures displacments down to 0.0005″ (0.013 mm).
“Some vibrations that might be an early indicator of failure in a critical pump or turbine cannot be detected with the human hand,” he said, explaining why such sensitive measurements are needed.
Data can go into the history database as well as into real-time control room. There are different alarm bells raised based on ISO standards or best practices, from a simple notification to a critical shut down response that can happen in milliseconds.
“The future is bright for condition monitoring of all assets,” he said. He noted that as the industry grows and embraces new technology, new methods for running manufacturing and process plants are emerging. “An emerging challenge is an ever increasing amount of technology and the amount of data that is being generated. There is an ever-decreasing amount of expertise in the field due to an aging workforce as well new individuals who are tech savvy and work in different ways,” he explained. “It is all about embracing this new technology and staying sharp to take advantage of it.”