Cut a hole in the shell and you lose strength where the metal used to be. That has to be made up somewhere, and there is more than one way to do it — weld on a reinforcing pad, thicken the nozzle neck, or step the parent plate up a size. Which one is cheaper depends on that shop's cost structure. So the program does the calculation and leaves the choice as a switch.
Cut a 150 diameter hole in the shell and the material that was there is gone. What is gone has to be made up from around the hole. The check runs like this:
That "what the surroundings can supply" comes from the excess thickness of the parent plate. If the actual plate is thicker than the calculation requires, the difference counts as reinforcement. That is the basis of the third option below.
This is the default. The nozzle wall is raised first, and if that is still not enough,
a reinforcing pad is added automatically. Its thickness and diameter appear in the report
and on the drawing — for example 6T x D150.
It is the safest and by far the most common answer. The cost is that one more part enters the build. It gets cut, rolled to the curve, welded and inspected. The plate itself is cheap; the hours are not.
Reinforcement can come from the neck as well. A thicker neck adds available area. So the nozzle wall is raised before any pad is considered, and if that alone passes, no pad is fitted.
It has a ceiling, though. Nozzle wall thickness can only move within pipe schedules and cannot exceed the maximum thickness setting, and a corrosion allowance eats into that further. When even the top of that range is not enough, the next step comes in.
Turn on pad avoidance and, instead of fitting a pad, the plate that nozzle sits on is raised one step up the standard thickness list. The excess thickness grows, the required area is met, and no pad is needed.
A real case — one tonne jacketed vessel, one nozzle on the top head.
| Avoidance OFF | Avoidance ON | |
|---|---|---|
| Top head | 3T | 6T |
| Nozzle N1 | pad required | no pad |
| Part count | one extra plate | nothing added |
The plate costs more. In exchange, all the cutting, welding and inspection hours for that pad disappear. Which side wins is decided by that shop's fabrication rates. Where welders are expensive and inspection is strict, thicker plate wins. Where plate cost dominates, the pad wins.
So the program does not weigh the two and decide for you. It gives you the switch.
| 1 | Raise the nozzle wall, up to its limit |
| 2 | Avoidance OFF: fit a pad. ON: step the parent plate up |
| 3 | If the thicker plate still is not enough: revert it and fit a pad |
| 4 | If a pad is not enough either: mark it NOT OK |
Step 3 matters. If the plate went up and a pad is needed anyway, the plate is put back. Otherwise you pay for thick plate and a pad both.
Step 4 matters too. When it does not work, it says so. Quietly passing it would be far more dangerous.
When avoidance actually fires, it shows up in three places.
[PAD-AVOID] 3T -> 6T(PAD) marker in the report filenameIf you flip a switch and cannot tell why a thickness went up, that is not automation, it is an incident.
All three pass the code check. What differs is the kind of cost — plate or hours. And that is known better by someone who knows the shop than by someone drawing the vessel. Which is why the calculation is the program's job and the judgement stays a person's.
There is a video running the same case with avoidance on and off, side by side.
Watch: avoiding nozzle pads →