Mining, mineral beneficiation and metal production are done for business reasons.  The processes used must therefore be financially competitive.  Since no two ore bodies are exactly alike, the design of the plant for processing material from a new ore-body will differ in some way from that of existing operations.  The detailed design of a new plant always requires engineering design, which requires process information.  Usually, this process information cannot be generated without testwork.  The direct costs associated with the testwork are not normally a significant fraction of the overall capital expenditure of the project, particularly when taking into account exploration costs and certainly not if a commercial plant is built.  However, the pace at which the testwork proceeds can be important in a business context.  Also, the consequential costs associated with inappropriate testwork can be extremely severe.  (For example, correcting design faults during start-up of the commercial plant can entail major extra expenditure and lost revenue.)  If the project does not proceed to a commercial operation, expenditure on testwork is effectively lost.  For these reasons, there is a need for a sound methodology to follow in testwork programs.  First, the design concept should be evaluated.  Typically, appropriate proof-of-concept tests would be done and the results evaluated before substantial amounts of testwork are commissioned.  (Unless, of course, the situation is one of applying known technology to a new feed that is similar to an existing one.)  Thereafter, laboratory-scale tests would be done to optimize the various unit operations envisaged, and the results would be used to evaluate the viability of the project.  If the project appears to be viable, pilot-plant work would be commissioned to refine or verify the process design and generate any outstanding process information.

Testwork can consume appreciable amounts of time and money.  Anything that can expedite it is of interest.  Process modeling, correctly combined with relevant experience and theory, can be used very effectively to enhance the productivity of testwork. Any flowsheet can be described in terms of a number of interconnected unit operations, inputs and outputs.  Process modeling entails the calculation of the quantities and compositions of all input, output and internal streams in the flowsheet.  The inner ‘workings’ of the various unit operations are described mathematically (e.g. reaction stoichiometry, conversion, fractional recoveries, etc.) and the connections between unit operations are specified.  The input streams are usually specified (the feed streams in terms of composition and magnitude and any reagent/utility streams in terms of composition) and the other streams are calculated.  The numbers so generated constitute a mathematically rigorous mass/energy balance over the flowsheet.  This balance is useful at all stages of a project, viz.

Conceptual stage
It is often possible to use preliminary information on the feed material and existing knowledge about the performance of individual unit operations to model one or more flowsheets.  At this stage, one makes assumptions about the nature and extent of chemical reactions, recoveries, etc.  It is rare for the results of testwork to substantially exceed early predictions.  The ‘developed’ flowsheet (after extensive testwork) is highly unlikely to outperform the one envisaged before testwork is done.  The mass/energy balance generated enables one to calculate the sensitivity of potential operating costs to the assumptions made about the performance of individual unit operations, and/or to compare different conceptual flowsheets.  The water balance, build-up of impurities and the magnitude of recycle streams are calculated rigorously (on the basis of the assumptions made).  This allows a much better evaluation of the flowsheet/alternatives at a much earlier stage than would otherwise be possible.  The exercise is normally an iterative process.  The assumptions are set up, the model is developed and solved numerically, the numbers are examined, changes are made and the model is re-run until it satisfies the people involved.  This exercise greatly refines one’s understanding of the flowsheet/alternatives.  Unrealistic expectations are exposed and the important process variables are identified significantly earlier, with much less effort, than in the absence of such a modeling exercise.

The results of a correctly-executed modeling exercise can enable one to avoid the development of a flowsheet that is not theoretically viable (technically or in terms of its potential operating costs – this is often not initially apparent) and direct the resources saved towards other projects.   When the modeling exercise gives numbers that indicate the flowsheet to be potentially viable, it enables one to design a significantly more productive program of testwork.

Flowsheet definition
When the results of laboratory-scale testwork are available, they can be ‘plugged into’ a flowsheet model and the model re-run.  (The model could have been developed previously, or it could be developed at that stage.  The former is more efficient, but the latter is quite workable.)  This gives a rigorous mass/energy balance that embodies the results of the testwork.  Any anomalies in the data (e.g. an acid consumption that disagrees with the dissolution of valuable metals and other elements) are fairly easily exposed, therefore the modeling exercise adds significantly to the credibility of the results of the testwork.  At this stage, the mass/energy balance would be a suitable basis for a pre-feasibility study by an engineering contractor.  The resulting improved estimates of the operating costs and estimate of the capital costs associated with the flowsheet would enable one to evaluate the project financially, with a high degree of confidence.  The output of a well-executed modeling exercise incorporating the results of laboratory-scale testwork better enables one to decide rationally on whether or not to proceed to more costly pilot-plant work and a full feasibility study.

Piloting
Once the laboratory-scale testwork has been completed and evaluated (by of means pre-feasibility study or some other way), the mass/energy balance generated from the results of the laboratory-scale work can be used to design a significantly more productive pilot-plant campaign than would otherwise be possible.  The sizing of the various items of pilot-plant equipment can be done much more accurately.  Recycle streams can be synthesized at their predicted compositions and magnitudes, and the pilot plant started up using these.  This enables the pilot plant to reach ‘steady state’ much sooner, thereby shortening its running time and/or enabling it to produce more ‘good’ data.  As before, the results of the pilot-plant campaign can be incorporated into the process model, thereby generating a rigorous mass/energy balance that accurately reflects the measured process data.  This balance is an excellent starting point for the detailed engineering design that constitutes the bulk of the feasibility study.

Environmental
The process model gives quantitative predictions of all effluents at each stage of the development sequence (conceptual modeling to pilot-plant data processing) .  This is useful for environmental impact studies, which are becoming mandatory (or at least prudent) at increasingly earlier stages in the development of projects.

Simple process modeling can be done using spreadsheets.  However, changing the process configuration, accounting for recycles and adding control actions is difficult with spreadsheets.  There are software packages that are designed for flowsheet modeling.  The better ones include powerful convergence algorithms, control loops, background chemical equilibrium, aqueous thermodynamics and extensive databases containing information on the physical properties of hundreds of components.  In these packages, drawing the flowsheet establishes the connections between the unit operations.  Once the inputs have been entered, the package does the arithmetic, freeing the user to concentrate on the flowsheet itself.  Changing the process configuration is relatively easy, as is varying process parameters to study the impact on the circuit.  In this way, one can produce rigorous mass/energy balances for as many variations of the model as one desires, with minimal extra effort.  This makes for enhanced analysis of the flowsheet, and therefore for better business decisions.