1 November 2005
No more migration headaches
Plug in to better DCS system replacement.
By Betty Naylor-McDevitt
Upgrading legacy control systems to next generation systems poses a major problem for vendors and end users. The differences between the old and the new systems are dramatic. Every 10 to 15 years, vendors tend to introduce a new product line, one that meets the ever-changing market needs and leverages the best technology advances available on the market. Rather than compromising the new design to accommodate the past, migration considerations tend to be an afterthought. When it becomes impossible to continue to support the older systems due to obsolescence of parts and lack of support people knowledgeable in the older system, vendors consider developing migration solutions.
End users want to keep maintenance costs down, ensure top-level support, and take advantage of the features and capabilities only the new systems offer to improve performance and profitability. They also want a migration solution that is cost effective, low-risk, easy to install, and designed to minimize the process downtime.
In the past, cost and risk associated with a system replacement has been due to the need for rewiring and retesting the new system while the process is down. Every vendor's input/output (I/O) equipment has a different signature size, shape, and I/O capacity and cannot plug in to the legacy termination assemblies and leave the wiring intact.
Popular migration fixes
The bulldoze: In the past, vendors simply required their customers to rip out the old system and install new wiring, cabinets, termination assemblies, and equipment. This typically took weeks or months of process downtime to install and test the new system before bringing the process back online. Furthermore, it is the most costly and risky solution. Everything is new (wiring, equipment, and database), and if there are problems, the process must stay down until they resolve. In today's world, it is unrealistic to expect customers to take on this risk and excessive process downtime.
Cables: Here you can reuse the original cabinets, nests, wiring, and termination assemblies and remove the original I/O modules and control processor modules. Simply install new cabinets with new equipment beside the original setup. Then run cables between the two. If the capacities of the old and new I/O modules don't line up, move field wires to spread out the load. Whenever you need to cut wires and redistribute them, you run the risk of additional cost for engineering, construction, and testing, plus significant downtime. It also increases the footprint of the system. In the worst case, it means doubling the number of cabinets.
Transitions: This involves replacing parts of the control system with newer equipment. It usually starts with replacing the original HMI console with the next generation console and modifying it to interface to the legacy controllers. Some vendors also offer a console and controller replacement solution, leaving the original I/O modules intact. Eventually, you'll need to replace the I/O modules. But don't expect guarantees for long-term support for this equipment. Here is where the real cost, risk, and downtime come in. The choices available to customers are a bulldoze solution (rewire to new I/O modules) or a cable solution.
Gateways: A gateway is really an integration fix, not a migration answer. Installing new equipment in a plant area and gateway interface integrate the new and old systems. But it's a temporary solution. While this can be inexpensive and quick to implement, it's at best a Band-Aid. The legacy system is still obsolete and will continue to age with escalating maintenance costs.
The challenge is to develop a better migration answer that eliminates the high cost of construction, risk of rewiring, and cost of process downtime. The plug-in solution meets all these criteria. Manufactured standard I/O modules are one-for-one, form-fit replacements for other vendors' legacy I/O modules. They fit directly into the original I/O module slot and line up exactly with the existing I/O points.
The plug-in approach reuses the original legacy cabinets, termination assemblies, power supplies, and racks; it simply substitutes new I/O modules for existing legacy ones. This happens without changing the wiring or termination assemblies. The design and manufacture of each new I/O module should meet or exceed the capacity of the I/O module it replaces.
New controllers and workstations replace the original legacy equipment, located within the original cabinets, along with new installation of network connectivity. The result is a 100% new control system. The replacement I/O modules use the same components as traditional I/O modules. No custom software is involved.
Plug-in migration plans are available for the following legacy systems:
- Honeywell TDC2000 & TDC3000
- Emerson (Fisher) PROVOX Series 10 and Series 20
- Westinghouse WDPF
- Bailey Net90 & Infi90
- Foxboro SPECTRUM & SPEC 200
- Moore APACS
- Taylor Mod300 (early 2006)
A plug-in migration solution has proven to provide savings of between 30-40% over a bulldoze solution. System equipment costs are virtually the same: You still need to purchase new I/O modules, controllers, and workstations, but there are some savings in not having to purchase new cabinets, racks, termination assemblies, or power supplies. The biggest savings is eliminating construction costs (rewiring). None of the wiring diagrams have to change, which drastically reduces instrument engineering.
One cost we often overlook is downtime. Plug-in installations minimize the process downtime to less than one day. A bulldoze or cable solution will require much longer outages. Customers should identify the cost in terms of per diem, lost profitability due to a plant outage. When comparing a one-day outage (plug-in solution) to several days, weeks, or months, it could be enough to pay for the new system several times over.
The first step in a migration involving the retention of the legacy system infrastructure starts with a site audit. The auditors look for excessive corrosion on wiring or equipment, indicating the infrastructure may need to be replaced. With legacy systems, this is rare. Older systems had exposed electronics and had to live in very clean environments. Most installations are in pristine condition. In addition, the legacy racks and termination assemblies have no electronic components to break or wear out.
To simplify and speed the process of replacing a legacy system, extensive planning, testing, and preparation are required. Experts at converting legacy system control and display databases perform the engineering. You can use automated database conversion tools to capture and reuse database information without having to manually reenter it. Where appropriate, identify and implement opportunities for optimization.
Perform full factory acceptance testing (FAT) ahead of time to ensure proper database conversion work meets the customer's satisfaction. You can fully stage and test the system using process simulators. Staging occurs using legacy cabinets, power supplies, I/O racks, termination assemblies, and cabling stocked for that purpose.
While the plant remains operational, you can put in place new I/A Series workstations and controllers along with the legacy equipment and ready it for hookup. On the day of the planned switchover to the new system, shut down the process, replace the legacy I/O modules with plug-in I/O modules (in the original I/O nests), and attach new network connectors for communication to the new system. Operators can now turn to their new consoles and begin loop checkout on their system to ensure all is working properly. You only need a cursory check to ensure the signals are read into the system. (In-depth logic testing occurred during FAT.) Then, just restart the process, and the new system is in control.
The solution is non-destructive. Because none of the field wiring or nests/termination assemblies change, the original legacy system can easily go back in place in case there's a problem. Simply shut down the process, reinsert the original I/O modules in the nests, reattach the original network connection, and restart the old system and process. The problem and its resolution can go offline without impacting end users' production schedule. It is a built-in contingency plan. With other migration solutions, once you cut or move wires, you are committed to resolve the issue with the process down, no matter how long it takes.
Once the end user is disconnected, permanently remove the legacy controllers and workstations, and mount the new controllers and workstations in their place. The new controllers and their base plate are significantly smaller than the legacy equipment. They easily fit into the original controller space. There is no increase in system footprint. With careful planning, the switchover can take hours instead of days or weeks. The majority of installations have occurred in less than a day. Some have even done it online without shutting down the process. IC
Behind the byline
Betty Naylor-McDevitt is director of migration at Invensys Foxboro in Foxboro, Mass.