How to Plan a PLC Retrofit Around a Shutdown Window
How do you replace a PLC on a plant that cannot stop?
You do almost all of the work before the shutdown. The new system is designed, built and tested away from the plant while the old one keeps running. The shutdown window is used only for what has to happen on site: disconnecting, connecting, checking and starting up. Every step has a rehearsed way back to the old system. If you plan to keep the existing enclosure, see whether a control panel can be reused in a retrofit.
The window should be the shortest part of the project, not the longest.
Why retrofit at all?
Controllers and drives reach end of support. Spares become hard to find, the supplier no longer offers repair, and the programming software may not run on current computers. A plant can often run on obsolete equipment for years, until a failure turns a manageable risk into a long stoppage. A retrofit is about choosing when to take that risk on your own terms.
A retrofit is not always the right answer. If a line is due to be replaced soon, spending on its controls may not pay back. That decision should be made before anything is quoted.
What happens before the shutdown?
- Survey the installed equipment. List every controller, drive, panel and network on the line, with its age, support status and spares position. This shows what is urgent and what can wait.
- Recover the documentation. On older plant the drawings and code are often out of date or missing. Record what is actually installed by tracing wiring and reading the existing program. Do not assume the drawings are right.
- Write down how the plant behaves. Capture the sequences, interlocks and alarms of the old system in a functional design specification. This is the reference the new system is built and tested against. If the old system has behaviour nobody can explain, decide with operations whether to keep it.
- Build and test off-plant. Program the new controller and test it against the specification, using simulation or a test rig. Factory acceptance testing should find most problems while nothing is at stake.
- Prepare the site. Pre-wire panels where possible, label every cable, and configure drives and instruments in advance so site work is mostly connecting what has already been tested.
- Stage the cutover. If the plant has several sections, decide whether they can be changed over one at a time, so each section can be reverted on its own.
What does the rollback plan look like?
A rollback plan is a procedure, not a good intention. It says how to restore the old system, who decides, and by when.
- Keep the old system intact. Do not remove or modify the old controller until the new one has run satisfactorily. Where possible, wire the new system in parallel.
- Set a go or no-go point. Agree beforehand the time by which the new system must be running, and what happens if it is not. Deciding this under pressure, during the window, leads to poor choices.
- Rehearse it. Run the cutover and the reversal in a test, at least in part, so the people involved have done each step before.
- Back up the old program and settings, and check the backup. A backup that has never been restored is an assumption.
What happens during the window?
The sequence is usually: confirm the plant is safe to work on, isolate and disconnect, install, check each signal point to point before any power-up, load the new program, test the functions that can be tested without running product, and then start up in stages. Operations witnesses the site acceptance tests against the specification.
Here is a short example. It is illustrative, does not describe a real plant, and contains no measured results. A single packaging line has an out-of-support controller and one weekend of planned downtime. The new controller is built and tested off-plant beforehand. On site, the first day is for disconnection, installation and point-to-point checks. The second day is for testing, start-up and a go or no-go decision. If the decision is no-go, the old controller is reconnected and the line starts on Monday as before.
What should operators and maintenance receive?
After startup, a run-in period with support available is worth having, because some problems only appear in normal production. Operators and maintenance staff should be trained on the running plant. You should also receive the as-built drawings, the program source, the specification and the test records. These are yours to keep, whoever supports the system later.
Where do safety functions fit?
If the controller being replaced performs a safety function, such as an emergency stop circuit or a guard interlock, changing it is a safety matter. The design and validation of those functions should be done with a qualified functional-safety assessor, and the work should be recorded as such. Using a standard as a reference does not mean the system is certified against it.
Related reading: how to prioritise obsolete plant control equipment, backing up and restoring PLC programs, what a factory acceptance test covers and the site acceptance test for plant control systems. Before booking installation crews, see how to check site readiness for a controls installation.
Facing an obsolete controller? Our retrofit and modernisation service explains what each stage delivers, and our process shows how a fixed-fee assessment decides whether a retrofit is worth doing. Our guide to what a functional design specification should contain covers the document the new system is built against, and our FAQ answers questions about plants that cannot stop.