Your Flowsheet Is a Hypothesis, Not a Deliverable
Why treating process flowsheets as finished documents — and ignoring feed variability — creates the risks your project can least afford.
In mineral processing and hydrometallurgical plant design, flowsheets play a central role in project development. They organise complexity, communicate intent, and guide testwork programmes. But there is a pervasive and costly misconception about what a flowsheet actually represents — one compounded by a second, equally dangerous gap: the near-universal undercharacterisation of feed material variability.
The Misconception
Flowsheets are treated as endpoints
Across the industry, flowsheets are frequently delivered and received as completed artefacts. The implicit assumption is that the underlying chemistry is known, testwork is sufficient, and the interactions between unit operations are well understood. In reality, this is rarely the case, particularly in the early and middle stages of project development.
This misconception is not born from negligence. It emerges from the natural pressures of project timelines, the desire to demonstrate progress, and a tendency to conflate a well-drawn flowsheet with a well-understood process. The document looks finished. The risks remain hidden.
A flowsheet is a hypothesis about how a system will behave — not a record of how it does.
The Reality
A flowsheet describes what we believe, not what we know
A more precise definition: a flowsheet is a structured hypothesis. It describes expected material movement, anticipated reactions, predicted impurity behaviour, and assumed recycle interactions — all based on the data available at the time of its creation.
The key phrase is “at the time of its creation.” As testwork matures, as pilot data becomes available, and as process interactions are better characterised, the underpinning assumptions change. A flowsheet that does not evolve in response to new data is not a sign of a robust design — it is a sign that the data is not being listened to.
The Most Underestimated Gap
Feed variability: the assumption nobody tests
Of all the uncertainties embedded in a flowsheet, feed grade variability is among the most consequential — and the most consistently undercharacterised. Most early-stage and even mid-stage studies are advanced using a small number of representative composite samples. These composites describe a central tendency. They rarely describe the range.
This matters enormously for any process that includes a concentration step — flotation, gravity, magnetic separation — before a hydrometallurgical refinery. The concentrate is not a fixed material. Its grade, its gangue mineralogy, its impurity profile, and its deportment of penalty elements will vary with ore type, mining domain, and blending strategy. A flowsheet developed against a single composite feed is, in effect, optimised for a material that may never exist in steady-state operation.
Industry gap: engineering advancing faster than ore knowledge
A persistent pattern in project development is the progression of engineering scope — flowsheet selection, equipment sizing, reagent suite definition — while ore variability studies remain incomplete. The result is a design envelope calibrated to a narrow slice of the actual feed population. When the plant encounters the full range of the orebody, the process was never designed to handle it.
The consequences are not abstract. Variability in feed grade translates directly into variability in leach kinetics, reagent demand, solution chemistry, and downstream unit operation performance. Impurity elements that appear benign at average concentrate grades can accumulate to problematic levels during periods of elevated feed concentration. Separation circuits sized for a nominal feed can fail when gangue mineralogy shifts. These are not edge cases — they are the normal operating condition of any long-life processing plant.
Risk Identification
Where the technical risk actually sits
Understanding that a flowsheet is a hypothesis — built on limited feed characterisation — reframes where risk lives in a project. The highest-consequence uncertainties are rarely found in the major unit operations. They sit in the gaps between them, and in the assumptions made about the material entering the circuit.
Feed grade variability
Concentrate composition shifts with ore domain, mining sequence, and blending. Designs built on composite averages are not designed for real operating conditions.
Impurity behaviour
Minor components have outsized effects on downstream operations. Their behaviour changes as feed composition varies — and is rarely characterised across the full feed range.
Recycle accumulation
Impurities that accumulate in recycle streams are sensitive to feed variability in ways batch testwork cannot reveal.
Operability challenges
Scaling, fouling, and separation difficulties emerge readily when feed composition moves off-nominal, but rarely appear in testwork on average composites.
Commercial Consequence
The cost of poor assumptions
When a flowsheet is treated as a fixed deliverable and feed variability is insufficiently characterised, assumptions compound. Pilot plants underperform against design, not because the pilot is wrong, but because the feed it received was unrepresentative of the operational range. Reagent consumption exceeds estimates. Equipment sized for nominal conditions struggles at the tails of the feed distribution.
The redesign and delay costs that follow are rarely traced back to the flowsheet or to the variability study. They are attributed to the ore, to operational issues, to the pilot. In most cases, the root cause is an assumption that was never identified as an assumption — and a feed characterisation programme that stopped at the average.
A Better Approach
Treating flowsheets as living risk-management tools
The shift required is not purely technical — it is philosophical. Flowsheets should be treated as the best current description of a system, subject to revision as understanding improves. Feed variability characterisation should be treated as a primary input to flowsheet development, not a parallel workstream that catches up later. In practice, this requires five changes:
Define the feed envelope early. Identify the expected range of concentrate grades, gangue mineralogy, and penalty element deportment across mining domains before committing to flowsheet selection.
Explicitly identify the assumptions embedded in the flowsheet at each stage of development. Make uncertainty visible — including uncertainty about feed composition — not implicit.
Design testwork programmes to probe the extremes of the feed distribution, not just the central composite. The process must be designed to handle the range, not the average.
Use process simulation to stress-test recycle behaviour, impurity accumulation, and sensitivity to feed variability before committing to pilot scale.
Update the flowsheet continuously as ore knowledge and testwork data improve. Version control and documented rationale for changes are not administrative overhead — they are engineering rigour.
The flowsheet as a decision-making instrument
A flowsheet that is treated as a completed deliverable stops being useful the moment it is issued. A flowsheet built on a narrow feed characterisation dataset carries risk that compounds through every subsequent stage of project development. Both problems share the same root: the conflation of what has been tested with what is known. The goal is not to produce a document — it is to progressively reduce uncertainty and improve the quality of decisions at every stage of project development. That means characterising the full range of feed materials the plant will actually see, designing for variability rather than for averages, and treating the flowsheet as a living instrument that reflects current understanding. That is where Vector2 Consulting focuses its process design practice.
Vector2 Consulting — Process Design & Metallurgical Engineering
At Vector2 Consulting, we treat flowsheet development as a process of risk reduction — ensuring that key assumptions are clearly defined, tested, and aligned with real operating conditions. We support clients by:
Linking testwork directly to design decisions and mass balance development.
Refining flowsheets as data evolves, focusing on areas of highest uncertainty.
Using simulation to understand system behaviour, including recycle effects and variability.
As projects progress toward FID, this approach helps ensure that flowsheets are not only technically sound, but also robust, operable, and decision-ready.