Techno-economic analysis as a method

Techno-economic analysis is not a forecast and not a market study; it is an accounting identity applied to a plant that does not yet exist. The central output is the minimum selling price: the price at which the discounted cash flows of building and running the process return exactly the required rate of return — net present value of zero at the chosen discount rate. Everything else in a TEA, from equipment sizing to labour lines, exists to serve that one computation, and the quality of the answer is inherited entirely from the quality of the flows it starts from.

The mass balance is the substrate

Every cost line in the model traces back to a material or energy flow, and those flows come from the mass-and-energy balance of the process flowsheet: feed consumed per kilogram of product, heating and cooling duties, volumes through each unit operation, solvent and resin inventories. The accounting then multiplies flows by prices and adds capital recovery. This is why TEA is unforgiving about process assumptions. An optimistic titer propagates everywhere at once — larger vessels, fewer batches per year, less labour and depreciation per kilogram — and nothing else in the model can absorb the error. For dilute products the recovery train dominates: a low titer inflates the volume of water that must be moved, filtered and worked up per kilogram of product, and the cost of that movement appears as capital and utilities rather than as a single visible line.

Sensitivity analysis names the owner of the answer

The base case is the least informative output, because no input is known to better than its plausible interval. The useful result is the ranking: vary each assumption across that interval and record how the minimum selling price responds. The tornado chart that emerges typically has a short head and a long tail — for dilute fermentation products usually titer and downstream step yields; for commodity products, feedstock price and conversion yield — meaning a handful of assumptions owns most of the variance. Monte Carlo extends the same logic to a distribution: draw all inputs from their ranges simultaneously and the output is a probability curve over prices, not a point. The decision value lies in that curve’s width and in identifying which single experiment would narrow it most.

Why an early-stage TEA is a range

The biological inputs are measured at bench scale, and bench data are systematically optimistic: small vessels behave differently, feeds are clean and expensive, recycle loops are absent, and strains have been selected for plates as much as for productivity. Downstream performance is frequently assumed rather than measured, because purification is developed last. Scale-up then changes mixing, heat transfer and sterility assurance in ways no small-scale experiment encodes. An honest early-stage TEA therefore reports the price as a distribution and names the assumption it is most sensitive to; a single-figure result presented without its uncertainty is a rhetorical artefact. In early development the method’s real function is directional: it tells the developer where the next experiment buys the largest reduction in uncertainty — and, just as often, that no plausible titer will rescue a route whose recovery economics fail.