Handed the task of overseeing the implementation of communication and control systems that would serve as the automation data lifeline for two new plants, Johan Claassen, engineering manager for control systems, instrumentation, and electrical, started small, but finished big. The plants, to be constructed by South Africa's Sasol Solvents and Olefins and Surfactants, housed some 10,000 I/O points. Claassen began with a vision to implement a control system and corresponding field instrumentation that had the intelligence to predict failures before downtime resulted. This initial vision ultimately led him to a major installation of HART-capable intelligent devices and big savings by using the technology in the company's chemical plants.

Claassen estimates Sasol's yearly savings as a result of the new system to be about 6.5 million South African Rand ($1 million). "We significantly exceeded my expectations. I was expecting two or three million Rand in savings," he says. However, Claassen believes this figure to be conservative, as a number of benefits are hard or impossible to quantify.

There's also another satisfying part to this tale for Sasol—this implementation won the company the HART Plant of the Year Award for 2005.

Claassen's and Sasol's results benchmark what's achievable, according to Ron Helson, executive director of the HART Communication Foundation. "They really stepped out of the box and set the example for what is possible when you start using the full potential of smart instrumentation," he says.

To finish this big, however, Claassen faced numerous obstacles that had him battling balky technology and overcoming objections from sometimes skeptical management.

In 2000, Claassen dusted off his crystal ball and peered into the future. He didn't do this because of the new millennium. Rather, he went into the forecasting business because Sasol, an integrated oil-and-gas company with substantial chemical interests, needed to expand.

In response to that need, Sasol would, over the next few years, build a butanol plant capable of producing 150,000 metric tons a year of solvent for the company's customers in the paint and ink industries. The company would also construct an acrylic acid and acrylates project that would eventually supply about 125,000 tons of products. Because of these projects, two plants would arise from empty fields near the company's complex in Sasolburg, a town midway between Cape Town and Johannesburg, where Sasol had gotten its start 50 years earlier.

Claassen, who'd joined the company in 1995 with a control-system engineering degree from the University of Pretoria, focused on the automation and control systems proposed for the new facilities. He had his own plans for these systems, although he didn't have much of a team to help him push his ideas. "The team consisted of two people—myself and a Sasol technology engineer—with the vision to implement a control system and field instrumentation that had the intelligence to predict failures before they caused plant downtime," he recalls.

The group started looking for solutions, investigating various technology options. It also looked at different distributed control systems (DCS) to provide needed functions.

But Claassen could not start from a blank slate. The automation and control systems for the new plants had to take into account Sasol's existing relationships and installed technology. On that basis, HART-based instruments and HART I/O had an advantage. Sasol already had a limited installation of HART instruments and systems in other sites within Sasol's Solvents group. In addition, Claassen notes that the communications protocol offered another benefit.

"HART had the advantages of normal 4-20mA where you have one pair of wires per field device. Sasol's design specifications for the new plants were built around this," he says.

Changing to another industrial communication protocol would have forced the company to update design specifications before building new facilities. But Claassen was prepared to go down that road because he felt that new technologies had to be part of the new facilities.

Others weren't quite as sure. The group's proposal for a control system involving HART, Foundation Fieldbus, and a new DCS was brought before Sasol's board of directors in 2001, but that approach was rejected. Disappointed, Claassen and his team continued planning. "We had to make the best of the situation. Some of the board's arguments made sense, and we had to honor it," he says.

Several of those arguments involved timing and risk. The board approved a $200 million butanol project in mid-2001 and fast tracked it for completion by the end of 2002. This approach, which involved taking on two new technologies—one for the DCS and the other for the communication protocol—seemed too difficult and presented too high a chance for failure.

Claassen's team reworked their ideas, concentrating on implementing HART along with a DCS from a vendor that Sasol had worked with before this project. When this idea was presented to the board, Claassen got approval for the combination.

A wrinkle at the time was that this particular vendor offered no integrated HART I/O solution, meaning that Sasol wouldn't be able to make use of the diagnostic capabilities of the technology. So the group decided to implement device management software from a rival vendor, using multiplexers to strip off the diagnostic information from the field devices. This solution required the use of OPC for communication and the development of a diagnostic bit-extraction tool to handle traffic between the hardware and software.

By the time the butanol plant was up and running in 2003, Claassen was hopeful that he had the desired automation and communication solution in place. Over time, though, his optimism faded. Part of the communication set-up in the plant involved signals being sent to indicate the existence of diagnostic errors in an instrument. An operator would then receive an alarm alerting him to the problem. Unfortunately, operators and engineers were getting 10 alerts a minute, and many weren't about anything an operator could affect. Consequently, those alarms were largely ignored.

Due to this alarm situation, the Sasol team switched its approach in 2004, implementing a solution that used the same vendor for both hardware and software. The company also embarked on an alarm management program and changed its philosophy on maintenance-related alarms, meaning that operators were only presented with those alarms they could do something about. Claassen reports that, today, the number of alarms seen by operators has been cut by more than a factor of 10, and that's only the beginning of the upgrades.

"As things improve, we will start training the operators to do first-line fault finding on our asset management software as well," he says.

The change in alarm philosophy coincided with the completion of the acrylic acid and acrylates facility and completion of current building projects. At present, Sasol uses some 4,000 HART instruments and about 400 FOUNDATION fieldbus devices. The mix includes multiplexers, shutdown and control valves, and pressure and temperature transmitters. Claassen uses the intelligence and diagnostic capabilities in these instruments for predictive maintenance strategies, root cause failure analysis, and fault tree analysis.

By using the diagnostic capabilities on the control valves, his group has determined which valves should be removed, targeting 40 out of 350 at a savings of over a million Rand. They've also detected faulty or poorly optimized valve positioners at a savings estimated to top two million Rand.

There are also advantages to the new systems that are harder to quantify. Claassen recalls that, after installing the latest asset-management portal-software, the acrylic acid plant went through a scheduled shutdown in September 2005. Using the asset management portal, Claassen and his team could give the production department an update on the health status of all field instrumentation and control valves before restarting the plant. This was not possible before because it was not easy to tell how well quality-assurance and quality-control tasks were done during the shutdown. Thanks to the new technology, Claassen was able to provide an answer much quicker than before.

"In less than five minutes I could give my production department the go-ahead to start up the plant. Before we had HART and the new software packages, it was very difficult to do this. You had to rely on the information that your people could give you," he says.

Like any new control or automation installation, the new technology and setup aren't yet perfect in the eyes of the engineer whose vision they are employed to fulfill. Claassen notes two drawbacks: not all instruments are created equal and, as such, the diagnostic information available from HART-compatible devices can vary widely; and, at 1,200 bps, the communication rate only allows about two transactions a second—much less than the number Claassen would like to see.

Claassen, however, is already developing solutions to these drawbacks. The first problem is being handled by careful work up front to ensure that instruments satisfy Sasol's needs. Those that don't will be removed from the list of permitted future purchases. The data rate problem, according to HART's Helson, is partly market driven. Specifications are in place for an eight-fold increase in the raw data rate, but vendors won't supply such devices unless the market demands it. For his part, Claassen believes that planned software changes at Sasol will boost communication speed.

Results obtained through the use of intelligent instruments have been so encouraging to this point that Sasol plans to deploy similar systems in the company's other plants over the next few years.

For further information about this project, contact Johan Claassen at johan.claasen2@sasol.com