Case studies
Every engagement starts with a measured baseline and ends with an audited delta. The studies below describe the wedge workflow, the metric and the mechanism. Named references publish at GA.
Outcomes shown are modelled from design-partner baselines. [ASPIRATIONAL — to be replaced with audited, named results.]
Measured on the press
Every Extruon engagement starts with a baseline and ends with an audited delta. These are the target bands we underwrite in a paid pilot.
+4.2 pts
Press recovery uplift, saleable vs charged metal
−38%
Surface-defect and dimensional rejects
−61%
Die-trial pushes before a die is signed off
−12%
kWh per tonne across heating, quench and aging
Target outcome bands modelled from design-partner baselines. [ASPIRATIONAL — to be replaced with audited pilot results.]
Studies
Wedge: adaptive ram-speed and exit-temperature control. Mechanism: taper set per billet so exit temperature stays flat instead of drifting, allowing higher sustained speed without die lines. Metric: recovery points per shift.
Wedge: quench control plus per-rack temper prediction. Mechanism: quench rate held against predicted T6 margin rather than a fixed recipe. Metric: T6 first-pass yield and audit evidence per rack.
Wedge: cut nesting plus aging batch consolidation. Mechanism: cut-to-length nested across the order book and aging batches consolidated into cheaper energy windows. Metric: kWh per tonne and offcut loss.
Wedge: twin sweeps and die-flow correction. Mechanism: flow imbalance corrected in simulation before press time is committed. Metric: proving pushes per new die.
Wedge: press-exit surface inspection. Mechanism: die lines, pickup and scoring classified at line speed and attributed to the die, so causes are fixed rather than symptoms sorted. Metric: escape rate to packing.
Wedge: fleet baseline and standardised heat-and-press control. Mechanism: the best plant's effective recipe encoded and transferred as priors. Metric: recovery spread between best and worst site.
Anatomy of a study
A case study that cannot survive a finance review is marketing. Ours are built to survive one.
01Baseline
Recovery, reject rate, die trials, scrap, butt loss and kWh per tonne measured from your own MES and quality records over a representative period, agreed in writing.
02Control
Where possible we compare like alloys, like dies and like profile families, and we report the confounders rather than hiding them.
03Shadow
Prediction accuracy against operators and gauges is measured before any control claim is made.
04Delta
The result is reconciled against the same MES and quality records that produced the baseline, using the metric agreed at week zero.
05Publish
Studies publish with customer approval, including the caveats. Where we cannot publish a name, we publish the mechanism.
Outcome bands
| Wedge workflow | Primary metric | Target band |
|---|---|---|
| Adaptive press & taper control | Press recovery (saleable vs charged) | +2.5 to +5.0 points |
| Profile & surface-defect sensing | Reject and rework rate | −25% to −45% |
| Die & flow optimisation | Proving pushes per new die | −40% to −70% |
| Quench & stretch control | T5/T6 first-pass yield | +3 to +8 points |
| Cut & age optimisation | Energy per tonne | −8% to −16% |
| Robotic handling | Handling damage and packing labour | −20% to −40% |
The mechanism
Recovery is not won in a report. It is won by holding exit temperature flat, catching a die line before it repeats, and not scrapping the front of the profile because the taper was wrong.
Every study traces its delta to a specific mechanism on a specific push, not to a general improvement.
Design partners
The baseline exercise alone was worth it. We found out our recovery number had three different definitions inside our own company.
What convinced engineering was shadow mode. Three months of predictions scored against our gauges before anything wrote back.
We negotiated the success metric before the contract. That is the only reason the pilot ended in a decision instead of a debate.
Design-partner quotes are illustrative of the target outcome. [PLACEHOLDER — to be replaced with named references at GA.]
Common findings
The gap between the best operator and the average shift on the same press is routinely larger than the gap between two presses. That gap is the fastest ROI in the plant.
Surface-defect rates concentrate in a minority of dies. Attribution at the press exit turns a diffuse quality problem into a short, actionable die list.
Under-loaded aging batches and conservative soak times account for a surprising share of kWh per tonne — invisible until batching is optimised across the order book.
Most plants budget proving pushes as a fixed cost of a new die. Measuring them per die makes the twin's value immediately legible.
Seven agents, one orchestrator
Each agent owns a physical stage of the line, senses it directly, and acts inside an approved envelope. A plant orchestrator arbitrates between them so nobody optimises recovery at the expense of temper.
Controls billet temperature taper, ram speed, breakthrough pressure, container and die temperature, and exit temperature to keep metal flow balanced and the profile on dimension.
Senses and predicts cross-section dimensions, wall thickness, straightness and surface defects — die lines, pickup, scoring, blisters — from fused vision and laser gauge.
Optimises die temperature, correction and flow balance to cut die trials and the twist and bow that send profiles back to the stretcher.
Controls press-quench cooling rate and stretch-straightening so temper and flatness land inside spec on the first pass.
Controls cut-to-length, batching and aging ovens to reach T5 and T6 mechanical properties without over-soaking the oven.
Drives robotic pulling, stacking, racking and packing of long, hot, delicate profiles — the handling nobody wants to staff at 3am.
Optimises press recovery, scrap, butt and offcut loss, and energy per tonne across the whole line rather than one station at a time.
Answers metallurgy and die-design questions with citations into your die books, profile drawings, press recipes and alloy specs.
Simulates billet heating, metal flow through the die, quench and profile properties — hitting the target profile and yield before the push, not after the scrap.
Assurance-grade by default
Extruon never asks for control it has not earned. Shadow mode proves accuracy against your operators and gauges; advisory mode proves ROI; bounded autonomy only follows once the gates are met.
Data, models, retrieval indexes and memory are scoped per tenant and per entity. On-prem deployment is available for sensitive producers.
Die designs, customer drawings and recipes never leave your boundary. Fleet learning shares defect signatures and process priors, never geometry.
Every perception, recommendation, approval and setpoint write is recorded with its evidence, ready for quality and customer audits.
SSO and role-based access, encryption in transit and at rest, SOC 2 programme in progress, and fail-safe interlocks on every write path.
Proof
The plant assessment produces the baseline. The pilot produces the delta. Both are measured from your own systems of record.
Land on one press. Expand press by press, module by module, site by site.