Atlas Core

Decision intelligence
for the plant floor.

Atlas Core installs beside your existing control system, reads the data you already collect, and tells your operators the cheapest safe way to run the plant — every few minutes, all year. No new equipment. No cloud. No control handover.

2htypical site bring-up
0new sensors required
<1stypical solve time
Two ways to work with Atlas

Decide what to build.
Then run it well.

One optimisation engine, two different jobs — a small service that runs your plant, and a modelling studio that designs it.

Atlas Core · Runs your plant

Installs next to your SCADA in an afternoon.

Reads live tags, solves your site’s model every few minutes, and recommends setpoints to your operators through the HMI they already use. Advisory only — Atlas never writes to a PLC.

For manufacturers, water utilities, and the system integrators who serve them.See a real solve
Atlas Platform · Designs your system

A modelling studio for the questions before operations.

Import your EPANET, SWMM, Raven or HEC-RAS model, design networks, set goals and constraints, run scenarios, and compare outcomes.

For engineers and consultancies doing capital planning and system design.Explore the platform
BUILT TO WORK WITHSWMM/EPANETFEWS PIWISKI / KiWISRTC-TOOLSRAVENHEC-RAS/HMSMODELICACADGISNetCDFZarrREST APISCADA
Before you ask

How it connects to your plant.

The three questions every engineer asks first — answered before you have to ask them.

Atlas is advisory. It cannot move your equipment.

There is no control path from Atlas to your PLC. Recommendations reach operators through your HMI, subject to your existing interlocks. The worst failure mode is no recommendation — which leaves the plant exactly as it runs today.

Your data never leaves your network.

Atlas Core runs on-site in a single container. No cloud calls, no telemetry, no licence check-in. It reads your historian and answers your HMI. Nothing else.

It works with what you have.

OPC-UA out of the box — Ignition, Kepware, or any compliant server. Modbus, SQL historians, MQTT and CSV are configuration, not custom development.

Decision intelligence, applied

Wherever equipment
runs on a schedule.

Any process with equipment or storage that can shift in time, and a price that moves with it. New sectors are a document, not a rewrite.

Pumping stations

Multi-pump scheduling against tank storage, minimum flows, and a demand charge.

Chiller & HVAC plant

Pre-cooling against a comfort band — timed to when outdoor air makes the chiller efficient, not just when power is cheap.

Compressed air

Multi-compressor scheduling against a receiver, accounting for how a fuller receiver costs more to push into.

Steam & boiler plant

Multi-boiler scheduling against a steam accumulator — including which boiler to run first when they don’t turn down the same way.

Battery storage

Charge and discharge scheduling against a time-of-use tariff and a demand charge, with round-trip efficiency.

New

EV fleet charging

Depot charger scheduling to hit a departure deadline without spiking the demand charge — and accounting for how cold weather derates the chargers.

Refrigeration, wastewater aeration, district heating and cooling, and CHP fit the same pattern. If your plant has equipment that runs on a schedule and a storage buffer of any kind, we can very likely map it.

Beyond water

Any engineered system.
One decision layer.

Any engineered system with decisions to make, constraints to satisfy, and data to work with is an Atlas fit — the platform doesn't impose a domain ceiling.

01

Water networks

Pressure, pumps, storage, and demand — coordinated as one optimised system.

EPANET hydraulic modelling
02

Hydropower

Dispatch generation against inflows, constraints, and live NEM market value.

Multi-unit RTC + pumped storage arbitrage
03

Flood operations

Turn forecasts into defensible gate, reservoir, and response decisions hours ahead.

Raven HBV + HEC-RAS 2D routing
04

Mine water

Keep site water balance and ANCOLD freeboard inside operational limits, automatically.

CasADi OCP + ANCOLD compliance
05

Canals & polders

Automate pumps and gates while holding levels within tight bands — the Dutch way.

GoalProgramming OCP + Modelica
06

Treatment plants

Shift blowers and digesters to lower-cost energy windows, maximise biogas value.

Time-of-use scheduling · 910 MLD scale
07

Drought management

Plan ahead for dry years — know when and how much to cut back before crisis hits.

Robust scenario-based scheduling
08

Reservoir systems

Balance releases across connected storages so every community gets a reliable share.

LP yield optimisation
09

Regional allocation

Share limited water fairly across towns, farms, and industry when regulators call for cuts.

Cascade LP + staged goal programming
10

Manufacturing

Shift energy-intensive machining into off-peak windows without missing a ship date.

Job-shop MILP · TOU energy shifting
11

Logistics & fleet

Route the fleet the way the road network actually looks, not the way a spreadsheet guesses.

Vehicle-routing MILP + OSRM road network
12

Agriculture & greenhouse

Pre-heat on cheap gas, coast on thermal mass — never chase a flat setpoint again.

Thermal-mass LP · 48-hour forecast horizon
13

District energy

Charge thermal storage overnight; shave the peak that actually sets your demand charge.

Chiller dispatch LP + TES round-trip dynamics
14

Gas & pipeline

Compress off-peak, draw linepack at peak, and never miss a city-gate delivery.

Compressor scheduling · linepack storage
15

Vehicle mfg & R&D

Time battery formation, cure ovens and dyno tests around the tariff — without slowing the line.

Process MILP · Modelica battery & thermal models
See it live

The Atlas console, in your browser.

Atlas ships as a self-contained browser interface with secure authentication for registered clients. We'll walk you through a live call tailored to your infrastructure — no account needed to start that conversation.

  • Domain-aware KPI dashboard across 13 built-in sectors
  • Live network map with Leaflet tiles and node-level drill-down
  • Import SWMM/EPANET, HEC-RAS/HMS, CAD, GIS Shapefile, NetCDF & Zarr
  • Live time-series over FEWS PI, WISKI/KiWIS, SCADA & REST APIs
  • 23 worked scenarios — from pump scheduling to gas-compressor dispatch
  • Every result labelled with its provenance — real solve, or demo simulation
Request a platform walkthrough

Client sign-in is by registration — see Atlas Core above for a demo you can try right now, no account required.

How Atlas works

From system signal
to operator action.

A transparent optimisation workflow that strengthens operator judgment instead of hiding it behind a black box.

01

Connect

Bring forecasts, telemetry, hydraulic models, and operating rules into a shared decision layer.

02

Optimise

Atlas evaluates thousands of feasible actions against cost, risk, and service objectives in under 30 seconds.

03

Act

Operators receive a clear plan, its trade-offs, and the evidence needed to approve and execute it.

Engineered for reality

Optimal on paper.
Practical in the field.

Forbay closes the gap between optimisation research and daily infrastructure operations. Every Atlas recommendation respects real equipment limits, service obligations, and the way your team actually works.

  • Explainable recommendations, not black-box outputs
  • Integrates with models and data you already use
  • Scenario testing before any action is taken
15Infrastructure domains
500+Automated test cases
<30sTypical solve time
FA

F. Awol, P.Eng., Ph.D.

Founder & Principal Engineer

Hydraulic modelling, mathematical optimisation, and real-time control systems across water, energy, and industrial infrastructure. The same optimisation core that schedules a pump station also schedules a job shop or a gas compressor.

Your system, clearer

Bring us the constraint
that keeps you up.

We'll map an Atlas walkthrough to your infrastructure, models, and operating challenge — no generic demo, no sales deck.

info@forbaysolutions.com
Ready when you are

Let's talk about your infrastructure.

We'll map an Atlas walkthrough to your specific models, data, and operating challenges — no generic demo.

Request a tailored demo