Scan Time per Layer

Area divided by the rate the beam covers area. The beam sweeps a strip of width $h$ at speed $v$, so it covers $v h$ of area per second, and a cross-section of area $A$ takes $A/(v h)$ to hatch. It is the same shape of calculation as machining time on a lathe, and it comes with the same warning: this is the laser-on time for the hatch, and it is a floor rather than an estimate.

What it leaves out is substantial and it is all in the same direction. The contours — the outline passes that set the surface finish — are usually run at a different and lower speed, and on a thin-walled part they can dominate the layer entirely. The jumps between separate islands of cross-section happen with the beam off but the galvanometers still moving, and a layer with thirty small features spends real time flying between them. The acceleration and deceleration at every vector end mean the commanded speed is a peak rather than an average, and on short vectors inside a thin section it may never be reached at all — a 3 mm hatch vector on a machine that needs 2 mm to get up to speed is a vector that averages well under the number you typed. Add any deliberate inter-layer dwell for cooling and the real laser-on time on a fiddly part can be several times this calculation.

The second thing to hold onto is that $A$ is the area at this height, and it changes layer by layer as the part's cross-section changes. Using a single average area over a whole build is perfectly reasonable for a quotation and useless for understanding where the build slows down. It is also why a solid block and a lattice of the same envelope have completely different build times in opposite directions from what people expect: the lattice has far less area to hatch but far more contour to trace, and on a fine lattice the contours win.

Multi-laser machines change the arithmetic but not the equation. Four lasers working four zones of a plate divide this time by close to four in the ideal case, and by rather less in practice, because the zones have to overlap somewhere and the stitching regions need care — and because the plume from one laser drifting across another's path is a real defect mechanism that gas-flow design exists to prevent.