Where test work turns into a flowsheet
The flowsheet is the point where laboratory results harden into an engineering commitment. Here's how the derivation works — from liberation data to grind size to stage counts — and why the diagram is cheap compared with the concrete poured on top of it.
Discuss your project
Overview
A mineral processing flowsheet looks harmless on paper. Boxes for crushing, grinding and separation, arrows for slurry, concentrate and tailings. What the paper doesn't show is that every box becomes a purchase order and every arrow becomes a pipe, a pump and a line in the mass balance. Once concrete cures around that diagram, changing it means demolition rather than revision — which is why flowsheet development sits at the exact midpoint between test work and engineering design, and why it deserves more scrutiny than it usually gets.
This page walks through where the diagram comes from: the chain of reasoning that leads from laboratory data to a process flow diagram, the recurring flowsheet shapes and the ores they suit, and what it costs when the shape is wrong.
What the flowsheet commits you to
Strictly speaking, the flowsheet is the conceptual sequence of unit operations, and the process flow diagram is its formal engineering expression, complete with stream numbers and equipment tags. In practice the two travel together. Between them they fix the grind size targets, the separation method, the number of cleaning and scavenging stages, and the recirculating streams that tie the circuit into a loop.
Everything downstream inherits those choices. The mass balance comes from the flowsheet. So do the water balance, the equipment list, the electrical loads, the plant layout and the civil design underneath all of it. Xinhai's Mine Design Institute coordinates 17 design disciplines — geology, mining, mineral processing, civil, structural, electrical, automation, water, tailings and more — and every one of them takes its input, directly or indirectly, from the same diagram. A flowsheet error doesn't stay a process problem. It propagates into foundations, cable trays and pipe racks.
The derivation chain: from test data to a diagram
A flowsheet you can defend is derived, not picked from a catalogue. The derivation runs in a fairly stable order.
- Mineralogy and liberation set the grind. Microscopy and liberation analysis tell you at what particle size the valuable mineral actually parts company with the gangue. That size becomes the grind target, and the grind target sizes the mills, which draw a large share of plant power.
- Separability curves set the method. Sink-float data from heavy-liquid work shows what a density separation can achieve at a given crush size. Grade-recovery curves from batch flotation show what the chemistry can do at a given grind. Magnetic and gravity response tests do the same for their methods. You're comparing curves, and the curve with the acceptable trade-off wins.
- Open- and closed-circuit tests set the stage count. Open-circuit work indicates how many cleaner and scavenger stages the ore needs. Locked-cycle tests then confirm what happens when middlings recirculate — the point where many optimistic flowsheets quietly fall apart.
A worked example from our engineering archive: for a copper project in Kazakhstan feeding 1.5 million tonnes a year, the flotation test program settled on a grind of 50% passing 200 mesh and a one-rougher, two-scavenger, three-cleaner flotation configuration. The confirmed result was a copper concentrate at 3.50% yield, 16.23% Cu grade and 66.23% recovery. Each figure maps onto a flowsheet commitment. The grind fixes the milling circuit, the stage counts fix the flotation banks, and the yield fixes how much concentrate handling the back end needs. None of those numbers came from an assumption — and if the vocabulary here is unfamiliar, our page on metallurgical testwork covers the test hierarchy from exploratory work through locked-cycle and pilot runs.
Three flowsheet shapes — and the ores they suit
Most concentrator flowsheets are variations on a few recurring shapes. Which one fits depends on where the ore offers a cheap rejection opportunity.
Gravity first, flotation after
Where a dense valuable mineral liberates coarse — placer gold, cassiterite, zircon-titanium sands — a gravity circuit ahead of, or instead of, flotation recovers it early, before further grinding smears it into slimes. The broader logic of matching method to mineral is covered in our ore dressing process guide.
Stage grinding, stage separation
Finely disseminated ores punish a grind-everything-at-once approach with power cost and slime losses. Grinding in stages and separating between them pulls out liberated values early and regrinds only the middlings that genuinely need it.
Heavy media separation up front
When there's a workable density contrast at coarse sizes, dense media separation can reject a large share of barren mass before it ever reaches the mill. You crush, you sink-float, and the light fraction leaves the plant at a fraction of the grinding cost.
The Zimbabwe lithium case in our project brochures shows how a flowsheet shape gets chosen under real constraints. For a 2-million-tonne-per-year spodumene operation — and Zimbabwe hosts five operating lithium mineral operations, per the current USGS Mineral Commodity Summaries — the test program found that petalite, one of the lithium-bearing minerals in the ore, wouldn't separate cleanly by flotation. Rather than force the chemistry, the program pivoted to density: heavy-liquid tests with tetrabromoethane at laboratory scale, then pilot verification with ferrosilicon media. The recommended flowsheet became crushing and screening, heavy media separation to recover petalite, grinding and classification, a gravity circuit for tantalum-niobium byproducts, desliming, mica flotation and finally spodumene flotation. With batteries taking an estimated 88% of global lithium end use according to the same USGS summary, the concentrate had a market waiting — and the flowsheet could serve it because it followed the mineralogy instead of the first method tried.
The product spec pulls from the other end
Test data pushes the flowsheet from the feed side. The product specification pulls it from the market side, and it can add or delete whole stages. Fluorspar is the textbook case: acid grade means more than 97% CaF2, metallurgical grade means 97% or less, per USGS commodity definitions. That single threshold decides how many cleaner stages the flotation circuit carries and how tightly regrind is controlled.
Our engineering archive includes a 1,200 t/d fluorite concentrator delivered in Italy under an EPC scope — Xinhai's first full-scale EPC project in Europe, executed with CE certification throughout. On a project like that, the flowsheet conversation starts with the buyer's specification sheet as much as with the drill core.
What a wrong flowsheet costs
Here's the uncomfortable arithmetic. Flowsheet development — the test program plus the engineering hours to interpret it — is a small fraction of total project cost. A flowsheet retrofit after construction is a different animal: changing a separation method or adding a stage in an operating plant means new foundations, rerouted piping and launders, rework across the electrical and control scope, and lost production for every day of tie-in. The study fee you saved by shortcutting test work gets repaid with interest, in concrete.
Recovery points carry the same weight in the other direction. At the same Zimbabwe spodumene operation, run under a cost-plus-profit-sharing EPC+M+O contract, recovery rose from 59.5% to 69% with concentrate grade held at 5.5% or better, and annual concentrate output increased by roughly 70,000 tonnes, as recorded in our project brochures. The same project moved from contract signature to production in 364 days. You can see how we document this kind of work on our project record.
Commissioning flowsheet development: what to ask for
If you're at the stage of buying this work, a few practical filters help.
- Insist on a derivation, not just a diagram. Ask which test result justifies each grind target, each stage, each recirculating stream. A supplier who can't trace the diagram back to data is selling you a template.
- Send a representative sample, and say what you don't know. No laboratory can make an unrepresentative sample representative. Geological variability belongs in the conversation from day one.
- Check the laboratory's breadth. Xinhai operates a CNAS-accredited laboratory and an industrial-scale pilot base, and reports approximately 200 test studies a year across more than 70 ore types, according to the company's published figures. Breadth matters because odd mineralogy — like petalite refusing flotation — is only odd if you haven't seen it before.
- Keep testers, designers and commissioning engineers in one loop. The flowsheet sits its final exam at plant commissioning, and the exam goes better when the people who wrote the diagram are still in the room.
Our mine design services page describes how test work, flowsheet development and detailed design connect inside one team. If you'd rather start with a conversation, send the mineral, the tonnage and whatever reports you already have. A flowsheet is only as good as the data behind it, and knowing what's missing is half the job.