Systems Integration / SCADA / Fieldbus / Wireless

Chips Are Up, Part 2

Find Out How More Powerful CPUs Have Improved Process Controllers

Jim MontagueBy Jim Montague, Executive Editor

Way back in November, I stated that microprocessors were at the root of most, if not all, recent advances in process control devices and systems. However, I realize now that I didn't provide a decent example to back up my claim, and without this real-world support, my opinion would be full of even more hot air than usual. Luckily, there's plenty of  evidence that faster and more powerful CPUs have greatly improved process controllers and the applications they serve. One recent example is from Dave Smith, Houston business unit manager for Wunderlich-Malec ( in Minnetonka, Minn. The 30-year-old system integrator and EPC firm and its almost 200 professional staffers  design and build everything from data centers to wastewater plants, and they've seen how rapidly evolving microprocessors make their controllers more intelligent and affect their projects.

"In the past, it was difficult for some large EPCs to get out into a more data-processing-centered world because it was hard to go from simply calculating the price per I/O point of a project to looking at less-tangible, longer-term strategies," says Smith. "When 4-20-mA networking and embedded protocols like HART began to emerge, everyone had to become more familiar with communications protocols, what measurements they were supposed to be delivering, and were they being done accurately. That's when talk began about adding diagnostic capabilities and possibly doing local or remote device maintenance and even process control on the wire via the network."   

For instance, Wunderlich-Malec just finished a project with Oak Ridge National Laboratory ( and its managing contractor UT-Battelle to deploy and install wireless sensors networks in steel plants, specifically for more sophisticated energy monitoring at CMC Steel's plant in Cayce, S.C., which is part of Commercial Metals Co. ( In August 2011, Smith and his colleagues designed and installed a network of WirelessHART devices, including about 1000 I/O points at distances up to 850 feet. 

CPU-enabled platforms are even causing DCSs and SCADA systems to combine some functions and provide them via the Internet

Despite the high levels of electrical noise that occur in virtually all steelmaking applications, CMC Steel uses its WirelessHART network to collect process values, such as stack gas and cooling water temperatures and flows from  the plant's melt shops and arc furnaces. Their data is archived and remotely monitored by CMC and UT-Battelle ORNL for use in their new energy-performance calculations. They can view the plant's operating status online, see health and history of instrumentation signals and retrieve stored data for calculations.

"We realized this system provided a new way of seeing into the plant and enabled much better understanding of its energy use," says Smith.

While the ingenuity and sweat of the project's many engineers and their software tools  deserve much of the credit for bringing wireless into CMC's plant, Smith acknowledges that much of their equipment relies just as heavily on embedded microprocessors. In fact, he adds these CPU-enabled, PC-based platforms are even causing typically separate DCSs and SCADA systems to combine some functions, and deliver them via Internet-based services.

"As microprocessors grow more powerful, smaller and less costly, more control devices can have their own operating systems," explains Smith. "Likewise, instead of having to read and write so much processor code, common CPU platforms are making it easier to run many functions. For example, to get temperature and mass flow data together for the AGA 357 calculations of gas through an orifice plate in a custody transfer application, it used to take many handwritten algorithms and manual programming. Now, we just grab the right software function blocks and schedule and enable them to run in a controller or RTU. So there's no reason why an AGA function for compressibility can't run on an instrument and have its algorithms closer to the actual measurements." How's that for capable?