Home  ›  Blog  ›  Bus Duct Replacement

How Do You Replace an Old Bus Duct Without a Plant-Wide Shutdown?

By DHYN Engineering Team  ·  October 7, 2026

Quick answer: You replace an old bus duct in stages, not in one shutdown. Survey and measure it while it is still live, order the new made-to-measure duct weeks ahead, keep critical loads on temporary bypass cables, and swap the duct section by section in planned outage windows. The final cutover is a hard outage of hours, so everything before it is preparation.

Nobody budgets for a bus duct replacement. The duct runs for decades, then one summer the joints start showing up hot on the thermal camera and suddenly it is a project. Here is how we see these jobs go well, and where they go wrong.

Replace or repair: how do you decide?

Repair when the problem is local: one hot joint, one damaged tap-off, one section that took a forklift hit. Replace when the decay is systemic, meaning repeated hot joints along the run, insulation resistance trending down year over year, corroded housings, or a dead brand with no spare parts.

A well built duct lasts 30 to 50 years, but environment shortens that fast. The honest test is the trend, not the snapshot. If your annual thermal survey found two hot joints last year and six this year, the duct is telling you something. Megger readings falling toward 1 megohm across several sections say the same thing in numbers.

How do you plan a replacement around production?

Do everything that can be done live, live. Thermography under load shows which joints are worst. Physical measurement for the new duct happens with the old one still running, because busway is made to measure and the factory needs your numbers, not the old drawings.

Then order early. An engineered replacement duct takes 6 to 12 weeks to build, and it will not fit if the survey was lazy. Old buildings lie: beams move, ceilings fill up with forty years of other services, and the 1985 drawings are fiction. Budget a full day for the site survey, with a tape, a ladder and a camera. Order spares in the same batch, because a replacement duct deserves its own tap-off boxes on the shelf.

Can the loads stay live during the swap?

Mostly, yes, with temporary bypass. Critical feeders get jumper cables or a rental generator for the changeover window, and the old duct is removed section by section. The one honest exception is the main feeder cutover itself, which is a hard outage. Plan for 8 to 12 hours, usually a weekend night.

For a plug-in duct feeding many small loads, the work is gentler: move each load's feed to a temporary board or the neighboring circuit, swap that section, move on. For a 3000 A feeder between transformer and main switchboard, there is no parallel path unless you build one. That cutover window is the whole project in miniature, and it should be rehearsed on paper with the electricians before anyone touches a wrench.

What happens on cutover day?

Lockout and prove dead, remove the old duct, hang and align the new one, torque every joint with a calibrated wrench to the manufacturer's figure, check phase rotation, then megger the whole run before re-energizing.

Two things get skipped under time pressure. Joint torque, which is the single biggest cause of future hot joints, and the phase check, which is embarrassing when a motor spins backwards. After energizing, put the duct under real load for a day and thermal scan every joint. A joint that runs 10 K hotter than its neighbors goes on the snag list immediately, not at the next annual survey. The full test sequence is in our commissioning guide.

What do people forget?

Tap-off compatibility, mostly. Your old plug-in boxes will not fit the new duct, and discovering that during the outage is painful. Update the single-line diagram and the maintenance schedule while you are at it, because the new duct resets the clock and the paperwork should show it. And do not scrap the old copper casually. A few hundred meters of 2500 A copper duct is real money, enough to noticeably offset the project cost.