On one 0.8-ton jacketed unit, this came out.
Material rose 3.7 kg, yet the total cost went down.
Look only at material cost and it's a loss. But the total is a gain. How does that happen?
The money in one pressure vessel splits into two lumps of different character.
In accounting terms the first is closer to acquisition (CAPEX), the second to operating input (OPEX). More important than the terms is that these two often move in opposite directions.
A common method is to find the material cost and then add fabrication as a fixed ratio — 6:4, 7:3 and so on. It's fast, and not bad as a quick reference.
The problem is that the ratio is fixed. If material drops, fabrication is calculated to drop automatically too. But on the floor, trying to cut material often adds one more cut, one more seam, one more inspection line. At that moment the ratio is wrong.
The shell length comes out at 1075 mm, but the plate in stock is 1100 mm. Going up to 1100 raises material, but one cutting step and the weld seam at that spot disappear.
| Item | Change |
|---|---|
| Material +3.7 kg | money spent goes up |
| One cut · one weld seam gone | money to make goes down (by more) |
| Total | total cost goes down |
By material alone it's a choice you should reject; counting fabrication, it's a choice you should take.
A nozzle reinforcing pad is not just the price of the steel. It is cut separately, aligned to the hole, welded all the way around, and inspected. If raising the base metal one step lets you drop that pad entirely, you have to compare the added plate cost against the removed steps.
This is when, keeping capacity, a small size change drops the thickness one step. On a 1-ton jacket the bottom head went from 10T to 8T, cutting material by 14.5 kg. Here the thinner plate also reduces the weld bevel area and pass count, so both go down together.
In ① and ② material and fabrication move opposite ways; in ③ they move together. To measure all three by one yardstick, material cost alone isn't enough.
AUTONEST VC calculates the material change and the fabrication change separately and adds them up. Standard-plate matching is adopted only when the total is a gain; if it's a loss, the original size is kept.
That said, whether to remove a reinforcing pad is not decided by the program. How much the plate cost rises and how much welding/inspection labor falls differs by company, and that comparison belongs to that company's cost structure. So we just give you a switch to turn it on or off.
That covers the cost of one unit. But there is one more kind of money that never shows on the balance sheet: the time from design to order.
Change the thickness once and the drawing, material list, weld map, and plate layout all have to be recomputed. By hand, that chain costs hours. So most shops never even examine ① and ② — there's no time to work out whether they pay off.
Saving is not buying less material. It is measuring the money spent and the money to make it by the same yardstick. When that calculation finishes in minutes, the options that used to be skipped finally become options worth examining.
Related — Design is done, so why can't you order? — automating the drawing and BOM too · Why we didn't automate everything — the places we chose not to put into code