# Lpbf Parameter Development > Use when developing LPBF process parameters, mapping energy density to melt pool regime, screening keyhole mode risk, or sizing the parameter development matrix for a powder bed fusion build. Develop the laser powder bed fusion (LPBF) parameter window: compute the volumetric energy density from laser power, scan speed, hatch spacing, and layer thickness, check hatch overlap between melt tracks, classify the process window as conduction mode, transition, or keyhole mode with porosity expectations, and build the parameter development matrix across power, speed, and hatch grids plus the qualification test matrix of coupon builds, density, and mechanical testing per the additive manufacturing qualification framework. Trigger: LPBF, laser powder bed fusion, volumetric energy density, process window, keyhole mode, conduction mode, melt pool, parameter development matrix. - Skill: `ashfordeou/lpbf-parameter-development` (Agent Skill, multi-file: 3 files) - Install (CLI): `npx skillmds@latest add ashfordeou/lpbf-parameter-development` - Raw SKILL.md: https://api.skillmd.com/api/skills/ashfordeou/lpbf-parameter-development/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - License: Apache-2.0 - Author: ashfordeOU (https://skillmd.com/u/ashfordeou) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/ashfordeou/lpbf-parameter-development --- # LPBF Parameter Development (manufacturing-quality/additive/lpbf-parameter-development) Use when developing the laser powder bed fusion (LPBF) process parameter window for a metal powder build: computing the volumetric energy density from the four build parameters, checking hatch overlap, classifying the process window as conduction mode, transition, or keyhole mode, and building the parameter development matrix plus the qualification test matrix. ## Domain quick reference - LPBF parameter set: laser power (W), scan speed (mm/s), hatch spacing (mm), and layer thickness (mm). These four values pin the energy input to the melt pool of a powder bed fusion build. - Volumetric energy density: VED = laser power / (scan speed x hatch spacing x layer thickness), in J/mm^3. High VED means a hot, deep melt pool; low VED means a shallow, cool pool. - Melt pool regimes: conduction mode (shallow, wide, stable pool), transition (mixed, intermittent keyholing), and keyhole mode (deep, narrow pool with a vapor depression). Keyhole mode carries porosity risk from trapped vapor and keyhole collapse. - Hatch overlap: overlap = (melt pool width - hatch spacing) / melt pool width. Positive overlap means adjacent melt tracks merge; negative overlap means un-melted gaps between tracks, which flags incomplete fusion risk. - Melt pool penetration: melt pool depth / layer thickness. A ratio well above 1 is the deep penetration signature of keyhole mode. - Qualification test matrix: each candidate parameter set is proven by coupon builds: density coupons (Archimedes density), tensile, fatigue, and hardness coupons, per the additive manufacturing qualification framework. - AS9100 link: parameter development feeds the additive manufacturing qualification program under production control and quality management; AS9100 is referenced, not reproduced. ## Workflow 1. Compute the volumetric energy density from the four build parameters: volumetric_energy_density(laser_power, scan_speed, hatch_spacing, layer_thickness) returns VED in J/mm^3. 2. Check hatch overlap between melt tracks: hatch_overlap_fraction(melt_pool_width, hatch_spacing) returns the overlap fraction, negative when the hatch spacing leaves gaps. 3. Check melt pool penetration: melt_pool_penetration(melt_pool_depth, layer_thickness) returns the depth to layer ratio, a keyhole signature when well above 1. 4. Classify the process window: classify_process_window(ved) maps the energy density to conduction, transition, or keyhole mode with the porosity expectation. The window bounds (conduction_ved, keyhole_ved) are material dependent and may be passed explicitly; the defaults are 60 and 100 J/mm^3. 5. Build the parameter development matrix: build_parameter_matrix(power_values, speed_values, hatch_values, layer_thickness) builds every power x speed x hatch combination at the fixed layer thickness, with the VED and regime per row, sorted deterministically by power, scan speed, then hatch spacing. 6. Apply the process window to the matrix: process_window_verdict(matrix) counts conduction, transition, and keyhole rows, flags any keyhole exposure, and returns a one-line verdict for the screen. 7. Derive the qualification test matrix: build_qualification_test_matrix(parameter_sets) assigns density, tensile, fatigue, and hardness coupons to each candidate parameter set per the additive manufacturing qualification framework. 8. Validate inputs first: non-numeric or non-positive parameters, empty grids, unknown regimes, and malformed parameter sets raise ValueError instead of returning a silent result. ## Worked example A powder bed fusion parameter screen on a 0.03 mm layer thickness: - Power grid: 200 W and 350 W. Speed grid: 800 and 1200 mm/s. Hatch grid: 0.08 and 0.12 mm. - build_parameter_matrix([200, 350], [800, 1200], [0.08, 0.12], 0.03) returns 2 x 2 x 2 = 8 rows. The corner points: - 200 W, 800 mm/s, 0.08 mm: VED = 200 / (800 x 0.08 x 0.03) = 104.2 J/mm^3, keyhole mode. - 200 W, 1200 mm/s, 0.12 mm: VED = 200 / (1200 x 0.12 x 0.03) = 46.3 J/mm^3, conduction mode. - 350 W, 800 mm/s, 0.08 mm: VED = 350 / (800 x 0.08 x 0.03) = 182.3 J/mm^3, keyhole mode, highest porosity risk. - Hatch overlap check: a 0.12 mm melt pool at 0.10 mm hatch gives (0.12 - 0.10) / 0.12 = 0.167, a 17% track overlap. At 0.15 mm hatch the same pool gives -0.25, un-melted gaps between tracks. - Melt pool penetration: a 0.12 mm deep pool over a 0.03 mm layer gives a ratio of 4.0, deep penetration consistent with keyhole mode. - process_window_verdict() over the 8 rows reports the keyhole count; the 350 W, 800 mm/s corners are screened out and the 200 W, 1200 mm/s corners move to coupon builds. - build_qualification_test_matrix() then assigns the density, tensile, fatigue, and hardness coupon builds to the surviving parameter sets. ## Pitfalls - Confusion with additive-manufacturing-qualification: the qualification leaf owns the qualification record, witness coupon sample planning, material property verification, and first article checks. This leaf owns the parameter window development itself: VED, hatch overlap, melt pool regime, and the parameter development matrix. Develop the window here, then feed the survivors into the qualification program. - Unit mixing: VED in J/mm^3 needs W, mm/s, mm, mm. Converting the scan speed to m/s or the hatch spacing to cm changes the result by orders of magnitude. - Treating the window bounds as material independent: the conduction and keyhole thresholds are alloy specific. Use per-alloy bounds instead of the defaults for production screening. - Hatch overlap as a single number: overlap needs the melt pool width, which itself changes with power and speed. Re-check overlap per matrix row, not once for the whole build. - Keyhole mode at the corners: high power with low scan speed and tight hatch spacing is exactly where keyhole porosity appears; the verdict count exists to catch corner combinations, not just the center of the grid. - Qualification test matrix vs production part: coupon builds prove the parameter set; they do not replace the production first article checks owned by the qualification framework. ## Behavior contract (gate 3) The parameter development logic is exercised by the gate 3 contract test: scripts/test_lpbf_parameter_development.py against scripts/lpbf_parameter_development_logic.py (stdlib unittest, offline). Run: python3 scripts/test_lpbf_parameter_development.py ## Compliance - Standards referenced, not reproduced: AS9100 frames additive manufacturing parameter development within production control and quality management; summary-only per standards-map.yaml. - compliance: STANDARDS-REF, gated: false.