Case studies

Proof in the numbers.

We gave ORCANET three human-designed networks, from the most naive to the most expert. Every time, the algorithm uncovers savings invisible to the naked eye — without ever violating the service pressure or the allowable velocities.

In short — Three networks sized by engineers, taken over by ORCANET's optimisation under the same pressure and velocity constraints. Each case gives the cost before and after, and the pipes that changed.

01

The "safety" reflex — oversized by hand

54 nodes · 82 pipes · gravity-fed supply

5 141 550 €782 452 €−85 %

  • Diameters set "for caution": DN 250 to 600 mm everywhere.
  • Ridiculous velocities (~0,1 m/s), largely excessive pressures.
  • The optimizer resizes the 82 pipes — cost divided by 6,5.
02

The engineer — by the book

57 nodes · 95 pipes · pumping + reservoir + PRV

919 443 €593 691 €−35 %

  • Cleanly sized: target velocity 1,2 m/s, DN 100 minimum, service-pressure margin.
  • A feasible network, beyond reproach… and yet improvable.
  • ORCANET optimizes 85 pipes while meeting every constraint (P ≥ 20 m, V ≤ 2,5 m/s).
The engineer’s network (57 nodes) loaded in ORCANET.
919 443 € → 593 691 € · 85 pipes optimized · feasible solution.
03

The large metropolitan network — at city scale

161 nodes · 266 pipes · 4 districts · 2 pumps + reservoir + 2 PRV

3 981 719 €2 341 565 €−41 %

  • Multi-district engineering design, with pressure zones and storage.
  • 236 pipes resized automatically in a single optimization.
  • The gain holds — and grows — at large scale, where humans reach their limits.
The metropolitan network (161 nodes, 266 pipes) in ORCANET.
3 981 719 € → 2 341 565 € · 236 pipes optimized · P ≥ 20 m, V ≤ 2,5 m/s.

Real screenshots of the application. Results obtained with the "Sizing" optimization.

Load your own network

Import an existing .inp file and compare: the gap takes a minute to measure.