A steel processing plant in South Africa needed to upgrade its Wonderware InTouch 7.0 to version 9.5. Plant operators tried the initial upgrade and no errors were reported. But the implementation on the live system was a disaster. Work-in-process sheet metal ended up in the trash. Because of the plant’s dependency on the supervisory control and data acquisition (SCADA) system, especially the database-logging process of the steel coils’ work-in-process, they had to revert to the old SCADA system.
Plant operators brought in system integrator AMR Automation. “They originally tried to do it themselves and they failed,” says Andrew Rennie, president, AMR Automation, in Cape Town South Africa. “We realized they have a 24/7 plant. They don’t stop. The production of the sheet metal is critical. When they failed, a lot of product went down the drain.”
AMR turned to simulation tools from Software Toolbox Inc., Matthews, N.C., to validate the upgrade. “We simulated everything on the line, ensuring that everything was working correctly. We were actually able to run the plant and be in agreement,” says Rennie. “It’s absolutely imperative to do simulation. It solves unnecessary problems and ensures the client can sign off before we go live.”
As they worked through the simulation, technicians figured that the client simply forgot to install the Wonderware SQL access objects on the machines. “For the simulation, users created name spaces that match the real name spaces of the process control system,” says John Weber, president of Software Toolbox. “We mimic the same in the InTouch system.”
AMR managed to successfully upgrade and install the applications on the live plant. “The plant operators were so excited, they decided to upgrade to InTouch 10.0 while we were at it,” says AMR’s Rennie.
Simulation and modeling are gaining traction as tools to validate a plant’s automation system before turning it on. Simulation is particularly useful for upgrading and optimizing process plants that sometimes go years between shutdowns and start-ups. The technology is also being deployed as a training tool to help young engineers cope with the massive knowledge loss that’s occurring in North America, Europe and Japan as baby-boomer plant engineers retire and take their decades’ worth of knowledge with them.
Simulation lets plant managers create an exact replica of the plant operating system and use that simulation to collaborate with vendors. “People are taking what is traditionally used for video gaming and creating a realistic 3D model of the plant,” says Robert Shear, business line director of the process and power industries at software supplier Autodesk Inc., in San Rafael, Calif. “You can walk through the simulated plant and collaborate in real time.”
Simulate before launch
For large, greenfield process plants, the advanced control automation process modeling might begin 12 to 18 months before the automation system is actually started. “Before you commission an asset, you want to make sure the plant and control performance is understood well to minimize any surprises,” says Martyn Blanchard, senior director of professional services at automation software supplier Aspen Technology Inc., in Burlington, Mass. “This helps test and debug the process and control system designs in advance for a smoother start-up.”
Major refineries use simulation and modeling to test all aspects of plant operation prior to launching a new plant. “Most of the big oil-and-gas plants have groups that do the simulation work for the process design and process validation,” says Mark O’Rosky, engineering manager at automation supplier Emerson Process Management, in Eden Prairie, Minn. “They use it for initial line, process validation and sizing equipment.”
Unlike discrete manufacturing, some process plants are rarely shut down. That means tweaking the system during upgrades, and optimization needs to be done offline, while the plant keeps moving. The changes have to be fully validated before they’re implemented. Simulation is used to debug the new system before it is deployed. “Historically, simulation was used in the design phase, but now, in the last 10 or 15 years, it’s used in the daily operation for our customers,” says Michelle Manzulli, principal engineer in professional services at Aspen Technology.
As with the South African steel operation, plants are vulnerable when changes are made to improve the system. Those vulnerabilities can be mitigated if the system can be tested prior to implementation. “There is a great deal of risk in doing advanced control optimization. Will it affect the plant in a positive manner?” says Martin Berutti, president and chief executive officer of simulation supplier Mynah Technologies, in Chesterfield, Mo. “Offline, you can run through the scenarios and test it. You have documentation for regulators, and your unit is running all the while.”
*****************************************************************
The simulation can run simultaneously with the actual plant. This lets operators monitor and optimize the plant’s system continuously. “Today, companies are using simulation, steady state and dynamic, to run he daily operation of the plant, to do monitoring, to track performance of the unit, to optimize the production or the energy consumption, and to improve the plant controllability,” says Aspen Technology’s Manzulli.
Problem solving
When something goes wrong in a plant operation, a simulation can help operators track down the problem. “If they experience something they don’t understand on the night shift, the day shift—with greater expertise—can take the incident from the historian and play it in simulation until they see what happened,” says Pete Henderson, product manager for simulation business, at Phoenix-based automation supplier Honeywell Process Solutions. “Simulation is a great tool for cause analysis. We can look at the problem as it is becoming upset, look at the reactors, backtrack and jump ahead.”
The simulation can also give vendor representatives a chance to examine the system before they arrive at a plant to conduct an upgrade or track down a problem. “With simulation, we can be debugging a customer’s system on the way to visit the customer,” says Emerson’s O’Rosky. “Our customers can virtualize the system and put it in virtual software. We take that and we’re using the same control and graphics for export and import from the DCS (distributed control system).”
How can you train new operators on start-up, shutdown and emergencies when the plant has been running without crisis or shutdown for years? And who is around to train the new engineers when the aging engineers retire and take their expertise out the door? Plants are turning to simulation and modeling for the answer. For the young engineer who grew up on video games, the solution sounds great.
The knowledge of the aging engineer can be programmed into the simulated model that can later be accessed by a newer engineer with less direct knowledge. “The aging generation is leaving, and the new generation will only stay for a few years,” says Tobias Scheele, vice president of advanced applications at automation solutions supplier Invensys Operations Management, based in Plano, Texas. “You need a long time to become an expert, so we have to put the intelligence in the software. Then the engineer can call up the intelligence to focus on the problem that needs to be solved.”
Some very basic operations—start-ups and shutdowns—are rare occurrences at some process plants. Yet operators need the ability to start up and shut down operations. With the infrequency of those events, operators can only be trained on virtual systems. “Some people use simulation to train for a shutdown. An operator may only see a shutdown every one or two years,” says Emerson’s O’Rosky. “The operator can experience the training scenario on a pace that suits him. You can train on start-up, shutdown, safety procedures—everything that is unusual is a candidate for simulation.”
Simulation will take a greater role in designing and running plants as the technology improves and continues to come down in price. “Simulation is at the beginning of its life,” says Honeywell’s Henderson. “Now, we evolve the design six months before the start-up. We train the operators before the start-up. In the next chapter of simulation, we will improve the advanced control strategy by creating a dynamic picture of the plant and use it to enhance the decisions the process control operator needs to make.”
There will come a time when the simulation can be developed from a stock program. “Portable modeling is in the future. You can dumb it down, and then tailor it to the specific plant,” says Scheele from Invensys.
Simulation and modeling have come into their own for designing control processes in advance of a process plant start-up. Simulation tools are also being deployed for upgrades and optimization as well as troubleshooting and training. The goal is to make sure that the system runs right before you throw the switch.
Related Feature - DOE Using Simulation to Train Power Plant Operators. To read this feature, visit:
www.automationworld.com/feature-8234
