Choosing an Estimating Method for a Conceptual Design
A chemical company is planning a new processing unit. The project is at the conceptual design stage (Class 5 estimate per AACE International Recommended Practice 18R‑97). The only available data are a capacity of 150,000 tons per year and a similar plant built 5 years ago that cost $12M for 100,000 tons per year. The older plant’s cost index was 350, and the current index is 420. The capacity exponent is typically 0.6 for this type of plant. The project team needs a cost estimate within ±30% accuracy to support a go/no‑go decision. The company has historical data for similar plants but no detailed engineering. The estimator proposes using the unit‑cost method based on $80 per ton of annual capacity from a recent trade publication.
1. Calculate the estimate using the capacity‑factored method and the unit‑cost method. Which method is more appropriate for this stage, and why?
Read the worked answer
Capacity‑factored: Cost = $12M * (150,000/100,000)^0.6 * (420/350) = $12M * (1.5)^0.6 * 1.2. (1.5)^0.6 ≈ 1.275, so cost ≈ $12M * 1.275 * 1.2 = $18.36M. Unit‑cost: $80/ton * 150,000 tons = $12M. The capacity‑factored method is more appropriate because it accounts for economies of scale and cost index changes, and it uses actual historical data from a similar plant. The unit‑cost method is less accurate because it relies on a single factor from a publication that may not reflect the specific process. For a Class 5 estimate, the capacity‑factored method is standard and provides a basis for the ±30% accuracy. The unit‑cost method is better for preliminary screening but lacks the adjustment for scale and time.
2. What are the key risks and limitations of using the capacity‑factored method in this scenario? How would you improve the estimate’s reliability?
Read the worked answer
Key risks: the exponent (0.6) may not be accurate for this specific plant; the older plant’s cost may include different scope, location, or technology; the cost index may not reflect the specific equipment or labor market; and the estimate does not account for site conditions, regulatory changes, or inflation beyond the index. To improve reliability: validate the exponent with additional data points, adjust for location and scope differences, perform a sensitivity analysis on the exponent and index, and use a range estimate (e.g., low/high) rather than a single point. Also, cross‑check with other methods (e.g., unit‑cost) and document assumptions. Since it’s a Class 5 estimate, the expected accuracy is -30% to +50%, so the estimator should present a range and highlight the largest uncertainties.