Thermal-hydraulic network engineering

Model your fluid network before you touch a valve

ArchiSist AI turns pipe layouts, pump curves, and boundary conditions into pressure-drop and thermal-load estimates — built from the field by an engineer who runs a live thermal and water network, not a lab.

StatusIn development
Core engineDarcy-Weisbach
NextPINN acceleration
SOURCE HEAT EXCHANGER 18.4°C SENSOR S-04 2.1 bar RETURN
Live calc · Re 63,408
Δp 2.2 kPa / 50m run
System status

What's actually running under the hood

No claims here get ahead of the build. This is the honest state of the computational core today, and where it's headed.

TODAY

Classical hydraulic core

Pressure drop and flow calculations run on the Darcy-Weisbach equation with a Swamee-Jain friction-factor approximation — live and interactive. See it work in the demo below.

ROADMAP

Physics-informed neural network layer

A PINN core trained against Navier-Stokes and Darcy-Weisbach constraints is in development, extending single-pipe math to full branched networks.

DESIGN

Field-first interface

Built to be read on a phone in a mechanical room, not just a desktop in an office — large touch targets, high-contrast readouts.

Demo mode — single-pipe model, illustrative only

Estimate pressure drop for one pipe run

Adjust the inputs below. The output updates live using the Darcy-Weisbach equation — this is a genuine calculation, not a mock number, but it's simplified to a single straight run of pipe.

Flow velocity m/s
Reynolds number
Head loss m
Pressure drop kPa