Losses & Productivity¶
Billet recommendation¶
PyExtrusion calculates the billet length required by the selected supported process geometry and press constraints.
The public output can distinguish between:
- a calculated billet length with full numerical precision;
- a practical recommended billet length rounded up with
ceilto a whole millimetre.
Display rounding to the nearest mm is a different operation. Feasibility limits use the full-precision calculated length, not either displayed value.
The recommended whole-millimetre value is intended as an engineering convenience. Users should still apply plant-specific billet-cutting tolerances and operating practice.
Finished cut length is interpreted as the net finished-bar length. Physical downstream saw kerfs therefore require additional extruded metal beyond the finished bars. Puller and final-saw kerf allowances are included in the physical process length used for billet geometry, table occupancy and technical extrusion time.
For positive-trim multi-billet continuous pulls, only the puller kerf is shared once per pull; each billet commercial block requires its own cuts + 1 final-saw boundaries. The process-level billet recommendation represents the selected complete-pull geometry. If an order finishes with a partial multi-billet pull, exact order-level saw events are used for timing and losses; the final partial pull can require plant-specific billet-length handling. See trim topology for zero-trim behavior and actual-group accounting.
The exact internal decision logic used to select among valid candidates is implementation detail and is not part of the public Technical Reference.
Process-loss model¶
PyExtrusion separates process losses into understandable engineering categories.
Fixed process losses¶
The fixed-loss group can include:
- butt / discard;
- total trim/reject reserve (
front_scrap_kgis the retained compatibility result name); - billet saw kerf;
- puller saw kerf;
- final saw kerf.
The public fixed-loss total is:
and its percentage relative to manufactured good output is:
Total trim mass¶
For canonical trim, order rejection mass is N_B * trim_total_per_billet_m * combined_linear_weight. Do not multiply by p again. The compatibility result name scrap.front_scrap_kg denotes this total, not exclusively a leading crop.
Additional modeled allowances¶
The physical butt loss is calculated separately from these planning allowances. See the physical definition below.
PyExtrusion can also represent additional planning allowances such as:
- startup allowance;
- process-complexity allowance.
The total modeled scrap is therefore:
and:
Startup and complexity are modeled planning allowances, not additional physical length inserted into the billet and not additional extrusion time. They should therefore not be interpreted as an exact physical mass balance or as measured plant scrap unless a user has deliberately calibrated the corresponding assumptions to plant data.
These categories are calculation aids. They are not a universal scrap standard for every extrusion plant.
Physical butt mass¶
butt_mm is the residual thickness in the container. With A_c in m², butt mass per billet is rho * A_c * butt_mm / 1000, even when an incoming-billet kg/mm override is supplied. The incoming-billet equivalent is this mass divided by the effective billet kg/mm coefficient. Never multiply the physical thickness directly by the incoming-billet coefficient when the two sections differ.
Order-level butt scrap is the number of billets times this single-billet mass. The number of sequential pulls does not multiply the number of butts. Billet-saw scrap still uses the incoming-billet coefficient; total trim, puller and final saw losses use the through-die basis.
Changing butt thickness can change billet feasibility, selected cuts/family, losses and derived scores. If the family and events remain unchanged, through-die mass, good mass, extrusion time and net kg/h remain unchanged. A loss-based score can change without a change in physical net productivity.
Saw kerf convention¶
A saw kerf represents the physical thickness of material lost by a cutting operation.
PyExtrusion distinguishes billet, puller and final saw kerfs because they occur at different process boundaries:
- billet saw kerf is an upstream billet-stock loss and does not add extrusion length;
- puller saw kerf is downstream extruded metal and is reserved in the physical extruded length;
- final saw kerf is downstream extruded metal and is reserved in the physical extruded length for the represented cut events.
Accordingly, puller and final-saw kerfs contribute to billet material requirement, table occupancy where applicable, and technical extrusion time.
A configured kerf of:
0 mm
means:
that operation contributes no material loss through saw thickness.
It does not remove the declared event topology. Additional final kerf changes physical extrusion length, mass and extrusion time, but no separate saw-operation time or new press dead-time event is added.
Good output¶
For a finished bar:
where:
- \(L_{cut}\) = net finished bar length, m;
- \(w\) = linear weight of one finished profile, kg/m.
For a production quantity of \(N\) finished bars:
Saw kerf is additional material loss and does not reduce the declared finished-bar length.
The number of exits affects how many bars are produced simultaneously, but the mass of one finished bar is still based on the linear weight of one profile.
Nominal gross productivity¶
The nominal extrusion mass flow is based on total linear weight leaving the die and extrusion speed.
If:
then:
where:
- \(Q_{gross,nominal}\) = kg/h;
- \(w_{total}\) = kg/m across all exits;
- \(V_e\) = m/min.
This is an idealised extrusion-rate quantity and does not by itself represent finished good kg/h.
Technical extrusion time¶
The fundamental extrusion-time relationship is:
where the physical extruded length includes finished product, total trim/reject reserve, and downstream puller/final-saw kerf allowances.
When expressed in metres and metres per minute, the result is in minutes.
PyExtrusion then adds the technical dead-time events represented by the current process model:
This is technical press time, not complete plant elapsed time.
It excludes unmodeled waits such as die availability, die correction, caustic cleaning, shift changes, maintenance or material logistics. Startup and complexity allowances also do not add time unless a future or external model explicitly represents such time.
Real gross and net productivity¶
PyExtrusion distinguishes between gross and net production performance.
Real gross productivity¶
Real gross productivity relates the physically represented extruded material to the calculated technical time.
Conceptually:
The physically represented extruded mass includes good product, total trim and downstream saw kerfs. Modeled startup and complexity allowances remain outside this physical extrusion boundary.
Net productivity¶
Net productivity uses good manufactured output:
Net kg/h is one of the most useful results when comparing the practical production performance of the same profile on different presses, provided the same time boundary and process assumptions are used.
Productivity and geometric indices¶
PyExtrusion can expose composite indices intended to help interpret how favourably a case uses the represented press and process.
These indices are not physical laws and should never replace the underlying engineering results.
The public interpretation should focus on the factors they summarize, including:
- net productivity;
- relation to an applicable productivity reference;
- process losses;
- cut utilisation;
- table utilisation;
- billet utilisation.
The exact internal weighting and scoring rules are intentionally not part of the public Technical Reference.
When comparing presses, users should inspect the actual physical outputs — kg/h, billet length, RE, losses, table use and warnings — rather than choosing a press only from one composite score.