If you're responsible for an industrial component that millions of people depend on, the last thing you want is a middle-of-the-night alert that the part has failed. The Internet of Things could be your ticket to sweet dreams.
IBM and National Instruments have built a testbed to find out how sensors in industrial gear, and analytics software in the cloud, can prevent failures in places like power plants. It's part of a broader effort by the Industrial Internet Consortium (IIC) to bring together various technologies in industrial IoT so companies can feel confident in adopting it.
NI and IBM set up the Condition Monitoring and Predictive Maintenance Testbed to prove that energy utilities can send data from their monitoring systems to a public cloud for analysis. Specifically, they are connecting NI's measurement and edge computing technology with IBM's BlueMix PaaS (platform as a service). They've shown that the systems can work together, and they're now identifying potential security holes and closing them, said Jamie Smith, director of embedded product systems at National Instruments. They are also talking with other companies about joining the testbed.
The testbed will explore one of the most promising applications of IoT in the industrial world. The idea is that sensors can monitor how well every generator, turbine and machine is running in real time, while big-data software can crunch all the numbers they collect to draw conclusions about where and when things will start to go wrong. This allows companies to keep using each system for as long as possible and replace it without unplanned downtime. Even the failure of a non-critical component of a power plant can cause a catastrophic shutdown, Smith said.
Technology like this could cut costs and improve operations across many industries, but there's no recipe for implementing it, especially when the equipment that needs monitoring has been there for years. Monitoring and maintenance systems have mostly been proprietary, but that will have to change if industries want to take advantage of Internet scale, Smith said.
Utilities are interested in cloud computing for the same reasons as other companies, he said: Buying, deploying and maintaining big data centers inside an enterprise is expensive and difficult, and public clouds can be a simpler alternative. But there are a host of other advances coming with IoT that could make industrial monitoring and maintenance much better. For one thing, today many kinds of equipment get tested only as often as a specialized engineer can come around to check them out, Smith said. Built-in IoT sensors collect data all the time.
The testbed is still in a National Instruments lab in Austin, Texas, where the companies are currently monitoring a motor and two fans. Fans are ideal for testing condition monitoring because it's relatively cheap to introduce faults by breaking or weighing down one of the blades, Smith said. Later, more components will be added. In the next phase of the plan, a testbed will be set up in a utility's own facilities, probably beginning next year.
Other industries, including rail, heavy equipment and mining, are likely to be interested in the results. The third phase of the project calls for testbeds in other industries. The project's findings will probably influence companies' decisions about the new monitoring and maintenance technology before all the testbeds are done, Smith said.
IIC was formed in March 2014 with a mission to foster interoperability among industrial IoT systems. IBM, General Electric, AT&T, Intel and Cisco Systems were the founding members of the group, which now includes more than 100 companies. IIC has also started testbeds for energy microgrids, tracking and tracing of tools in factories, and software-defined infrastructure.
The new cooperation among companies in this field represents a big change, Smith said.
"Historically, industrial systems have been very siloed," Smith said.