Re-checking the client's financial model

The brief asks for an independent re-check. We rebuilt the client's workbook from its own formulas and cell references and reproduced all eight of its scenarios to within 0.19 percentage points of internal rate of return, so what follows is criticism of the model as built rather than of a misreading of it.

Six findings, in order of how much they matter.

The eight scenarios are ex-plant, and the case is not. Row 6 of the Simulation sheet reads "Ex-plant" in all eight columns. Rental transportation service, round-trip distance, trips per day and fuel consumption are all zero. The line the workbook labels "Transportation" is a compressor and storage tanks at the plant gate. There is also no cost anywhere for collecting biogas from the mills to the hub, which is the concept's central mechanism.

The headline case prices a domestic product. The compressed case sells at $14.42/MMBtu, which multiplied by the model's exchange rate of 33 gives 476 THB/MMBtu, identical to the LPG price the model carries one row below it. The liquefied case sells at $18.00. Neither is the $20 Singapore price the brief names as a hard assumption. The 40 tonne-per-day compressed case returning 36.3% is a Thai plant selling at domestic parity from the fence.

Working capital is charged in full every year and never released. The discounted cash flow charges 30 days of revenue less cost in each of years zero through fourteen, which is the level of working capital rather than the change in it, and nothing is returned at the end. Correcting it raises the model's own returns:

The workbook also carries three different definitions of the same working capital across three sheets.

The small-scale cases fail for a spreadsheet reason. The upgrading electricity line is computed from throughput only in the 40 tonne-per-day columns; the 30 tonne columns take 75% of that figure and the 20 and 10 tonne columns all take 70% of it, so both 10 tonne cases carry exactly the same absolute electricity bill as the 20 tonne cases while producing half the gas. Implied specific energy runs 0.24, 0.24, 0.336 and 0.779 kWh per Nm3 of biomethane across the four scales. Computed from each case's own throughput, the 10 tonne compressed case returns 18.26% rather than 2.17%. The client's stated conclusions that 10 tonne cases struggle on economies of scale and that the sweet spot is 30 to 40 tonnes rest on that line.

The desulfurization line is built on a media capacity that is far too low, and on an inlet sulfur concentration that is too low as well. The workbook sets iron-oxide media capacity at 8.1 mg of hydrogen sulfide per gram, which turns 11.12 tonnes of sulfur a year into 1,715 tonnes of media, or 154 kg of media per kg of sulfur removed, costing 51.5M THB a year and 21% of total operating cost. Stoichiometry gives 640 mg per gram at complete conversion and a working capacity of about 150 is the figure this analysis uses, 18.5 times the workbook's. Against that, the workbook's inlet concentration of 200 ppm is 3.1 times below the 943 mg/Nm3 measured at an operating POME plant. The two errors point in opposite directions and the net effect is to overstate the cost.

There is no construction period, no ramp, no escalation and no terminal value. Year one runs a full 346 operating days at constant rate, against a brief that requires payback within 8 years including about six months of construction, and against a Thai palm sector that averaged 37.1% capacity utilization. Neither revenue nor cost escalates across fifteen years while a 3% inflation assumption sits unused and the discount rate is a nominal 8%.

Two smaller points. The model's tonne-per-day row measures contained methane rather than product, so its "40 TPD" is 45.81 tonnes of gas on the basis a shipper would load. And the financing block is broken throughout, so the 75% leverage and 3.43% interest rate it carries feed nothing; the reported return is unlevered, which is what the brief's hurdle asks for.