Beyond the flowsheet — why plant simulation needs a brain

For twenty years, the promise of aggregate plant simulation has been simple and powerful: see the impact of every change before you make it. Build the flowsheet on screen, drop in your equipment, feed in a gradation, and read out the mass balance — tonnages, gradations, circulating loads — for every stream in the circuit.

That promise still holds. What has changed is everything around it.

The flowsheet was never the hard part

Drawing a circuit is not where engineers lose time. The hard part is the loop that comes after:

• Which deck or setting is actually capping production?
• If I open the closed-side setting 5mm, what happens to my 0–5mm yield and my margin per ton?
• How do I turn a solved balance into something a customer will sign off on?

Classic desktop simulators answer these one manual what-if at a time. You are the optimizer. You are the report writer. The software holds the physics; you hold everything else.

Adding a brain to the balance

The mass and water balance is still the foundation — there is no shortcut around the engineering. But once the balance is solved, a model can reason about it:

1. Find the bottleneck automatically. Read the balance, rank the constraints, point at the stage that matters.
2. Optimize toward a goal you state in plain language. "More manufactured sand without losing throughput" is a target a machine can search for.
3. Model profit per ton from the first run, not as an afterthought — because the decision was never really about tonnage.

What this means for how you work

The engineer stops being the search algorithm and starts being the judge. You describe the goal; the tool proposes the circuit; you decide whether it is right. The rigor is unchanged. The tedium is gone.

That is the idea behind AggMind: keep the mass balance every good simulator has always computed, and put an AI engine on top of it.

Want to see it run on your circuit? Book a demo.