PyExtrusion

Engineering calculation toolkit for aluminium extrusion

Version: 0.19.0 — 2026-10-03.

0.19.0 adds trim_total_per_billet_m, explicit legacy migration, and corrected positive-trim multibillet final-saw events, including partial pulls. Read the trim contract and migration guide before migrating sequential-pull data. Examples below retaining front_scrap_m deliberately use the legacy contract; do not mechanically rename it in p=2 data.

Version 0.18.0 corrects butt_mm as physical residual thickness inside the container: butt mass uses the container section, then converts once to incoming-billet equivalent length. Existing physical mm inputs and the measured incoming-billet coefficient override remain supported. Calculated lengths, losses and feasibility may differ from 0.17.0; see the changelog.

PyExtrusion is a deterministic Python toolkit for evaluating aluminium profiles on direct extrusion presses. The same calculation engine is available through Python, CLI and JSON workflows.

PyExtrusion includes a source-traced engineering layer, basic deterministic economics and physical downstream saw-kerf accounting. The engineering models keep their scope explicit: the implemented pressure/force result is an equivalent-axisymmetric baseline, not a final porthole-die force prediction, and the thermal models are analytical/source-term baselines rather than a production-grade exit-temperature predictor.

Three working areas

• Production & Planning — billet sizing, extrusion geometry, productivity, multi-billet processes, operational planning, sequences and multi-press comparison.
• Engineering — AA6063/AA6060 constitutive models, modified Feltham mean strain rate, Zener-Hollomon, flow stress, equivalent-axisymmetric mechanical baselines and bounded thermal tools.
• Economics — recurring production cost, material/scrap economics, tooling/development cost and sales-margin calculations.

What changed in 0.18.0

• Corrects physical butt mass and equivalent incoming-billet length without changing the physical input values or redefining upsetting.
• Adds AA6060 and AA6063 as identical-object aliases of the existing source-traced constitutive models, preserving scientific names and provenance.
• Integrates the public Engineering section and separate Engineering/Economics API examples into the documentation.
• Retains the existing pressure, rheology, thermal and economics equations; no new alloy dataset or registry is introduced.
• Keeps shaped-section, bridge/porthole pressure corrections, complete transient thermal reconstruction and production-grade exit-temperature prediction outside the current model.

Installation

From PyPI:

python -m pip install pyextrusion==0.19.0

From a locally built release wheel:

python -m pip install pyextrusion-0.19.0-py3-none-any.whl

Quick calculation

from pyextrusion import Press, Profile, Production, StudyCase, calculate_case

press = Press(
    name="Example 8-inch press",
    nominal_size_in=8,
    table_length_m=54,
)

case = StudyCase(
    profile=Profile(1.35, exits=2, profile_type="solid"),
    production=Production(
        exit_speed_m_min=24,
        cut_length_mm=7000,
        bars_requested=1000,
        front_scrap_m=2,
    ),
)

result = calculate_case(press, case)

print(result.recommended_configuration)
print(result.billet.recommended_length_mm)
print(result.production.billets_per_pull)
print(result.productivity.real_net_kg_h)

table_length_m is mandatory. Billet/container geometry, billet limits, dead time, saw kerfs and a reference productivity target can be inferred from nominal_size_in; real plant values supplied by the user always take priority.

Quantity-free process definition for planning

from pyextrusion import (
    PlanningCase,
    PlanningRequest,
    Press,
    Process,
    Profile,
    calculate_planning,
    calculate_process,
)

planning_case = PlanningCase(
    profile=Profile(1.35, exits=2, profile_type="solid"),
    process=Process(
        exit_speed_m_min=24,
        cut_length_mm=7000,
        front_scrap_m=2,
        complexity="normal",
    ),
)

process = calculate_process(press, planning_case)
print(process.billet_length_mm, process.bars_per_billet)

plan = calculate_planning(press, planning_case, PlanningRequest.bars(300))
capacity = calculate_planning(press, planning_case, PlanningRequest.hours(2))

Planning works with complete billets. A request may be expressed in bars, kg, metres, billets, minutes or hours. The annual optional 10% supplement is not applied implicitly to an operational planning request.

Continuous production sequences

from datetime import datetime
from pyextrusion import ProductionOrder, calculate_production_sequence

orders = [
    ProductionOrder("OF-001", planning_case, PlanningRequest.bars(300)),
    ProductionOrder("OF-002", planning_case, PlanningRequest.billets(20)),
    ProductionOrder("OF-003", planning_case, PlanningRequest.kg(2500)),
]

sequence = calculate_production_sequence(
    press,
    orders,
    start_at=datetime(2026, 9, 7, 5, 0),
)

print(sequence.total_press_time_min)
print(sequence.orders[0].cumulative_end_min)

The returned timeline is continuous technical press time only. The supplied list is never reordered and PyExtrusion does not estimate setup or waiting time between orders.

Multi-press comparison

from pyextrusion import (
    compare_planning,
    compare_processes,
    compare_production_sequences,
)

process_cmp = compare_processes([press_a, press_b, press_c], planning_case)
plan_cmp = compare_planning(
    [press_a, press_b, press_c],
    planning_case,
    PlanningRequest.bars(300),
)
sequence_cmp = compare_production_sequences(
    [press_a, press_b, press_c],
    orders,
)

PyExtrusion returns results in the same order as the supplied presses. It does not rank them or declare a winner. A sequence comparison gives every press the complete same order list; it does not distribute work between presses.

Supported production geometry

The current model supports:

• k_billets_1_profile: one continuous pull formed from one or more billets; k is calculated dynamically after billet optimisation;
• 1_billet_2_profiles: one billet produces two complete sequential pulls.

A scenario requiring 3 or more complete sequential profiles from one billet is not supported and is reported explicitly through support-status fields.

Engineering model

Publicly documented calculations include:

• profile section from linear weight;
• extrusion ratio from container-bore area;
• ram speed from volume constancy;
• billet mass and kg/mm from actual billet geometry or a measured mass override;
• billet-first cut/billet optimisation;
• dynamic billet-on-billet continuous pulls;
• modeled startup, complexity, butt, front-scrap and saw losses;
• technical dead time and extrusion timing;
• nominal, real gross and real net productivity;
• geometric and productive utilisation indicators;
• annual-demand normalisation and an explicit optional 10% supplement;
• operational planning and time-window capacity calculations;
• continuous technical calculation of user-supplied order sequences;
• JSON persistence, validation and multi-press comparison;
• source-traced engineering baselines for hot-working rheology, equivalent-axisymmetric pressure/force and limited thermal analysis;
• basic deterministic production economics.

Model boundary

PyExtrusion 0.18.0 models direct aluminium extrusion. It does not perform full industrial scheduling or complete extrusion-force, thermal, metallurgical, die-life or plant-resource prediction.

The mechanical result is F_baseline, an equivalent-axisymmetric baseline. A comparison F_reserve = F_press - F_baseline is only a screening indicator; it is not guaranteed remaining capacity, a safety margin, force available for a porthole die or a prediction of the real die load.

The productivity index is orientative. It is not a physical quantity and must not be used alone as the final industrial selection criterion.

Documentation

The public documentation is organised into:

• User Guide — task-oriented usage;
• Engineering — public analytical models and explicit interpretation limits;
• Technical Reference — selected engineering basis and formulas;
• API Reference — production, Engineering, Economics, JSON, CLI, result fields and errors;
• Examples — practical workflows;
• About — project, citation, licence and changelog.

MkDocs source is included under docs/ and configured by mkdocs.yml. The documentation site is intended for https://pyextrusion.com.

• Documentation: https://pyextrusion.com
• Source repository: https://github.com/enrique-co/pyextrusion
• Issue tracker: https://github.com/enrique-co/pyextrusion/issues

Project identity

• Project: PyExtrusion
• Author: Enrique Calvo Ordonez
• Website: https://pyextrusion.com
• License: Apache-2.0