Planning a PLC migration with minimal downtime

Before any design work, document the system as it really is, not as the drawings say. Old systems often carry years of undocumented changes.

• Hardware, I/O count and signal types, panel layouts and spare capacity.
• The application software, including interlocks, sequences, alarms and any workarounds.
• Communication links to HMIs, SCADA, drives and other controllers.
• Which functions are safety-related, and whether they belong in a separate safety system.

Conversion tools can translate much of the old code, but the result still needs review by engineers who understand the process. Use the migration to fix known problems, standardise alarm handling and improve diagnostics, while keeping the control philosophy agreed with operations.

Run the migration through the plant’s management of change process from the start, and design network security into the new, more connected system (for example to IEC 62443).

A factory acceptance test (FAT) with simulated I/O lets operators and engineers check every interlock, sequence and alarm before installation. A site acceptance test (SAT) then confirms wiring and communications in place. Each test is run to a written procedure and closed out against a punch list. Use the FAT to train operators and maintenance staff too, so they know the new system before it takes control.

How the new system takes over depends on how long the plant can stop:

• Full cutover: replacing the whole system during a planned shutdown, which is the simplest option when the window is long enough.
• Phased cutover: moving one area or I/O group at a time, often with adapters that reuse the existing field wiring, and with a temporary interface between the old and new controllers.
• Parallel running: the new system monitors alongside the old one before taking control, which lowers risk but needs more planning.

Update the drawings, back up the final software and agree spare parts and support arrangements. A short period of close support after start-up catches the issues that only appear in real operation.