| name | cost-estimation |
| description | Engineering cost estimation — analogical, parametric (CERs), bottoms-up, learning curves, DCFM, ROM vs definitive estimates, AACE classes, manufacturing cost models. |
| metadata | {"priority":7,"promptSignals":{"phrases":["cost estimation","cost estimate","parametric cost","learning curve cost","ROM estimate","AACE estimate","manufacturing cost"],"minScore":3}} |
Engineering Cost Estimation — Complete Skill
AACE International Estimate Classes
| Class | Maturity | Accuracy | Effort | Use |
|---|
| Class 5 | 0–2% | -50% to +100% | Low | Concept screening (ROM) |
| Class 4 | 1–15% | -30% to +50% | Low-Med | Study/screening |
| Class 3 | 10–40% | -20% to +30% | Medium | Budget authorization |
| Class 2 | 30–75% | -15% to +20% | High | Control/bid |
| Class 1 | 65–100% | -10% to +15% | Highest | Check/bid/tender |
ROM (Rough Order of Magnitude): Class 5; quick assessment; single point or range
Estimation Methods
Analogical (Reference Based)
Use cost of similar completed project as baseline
Scale factor: C_new = C_base × (S_new/S_base)^x
x = scaling exponent:
- Capacity-based (plants): x ≈ 0.6 (six-tenths rule)
- Linear systems (pipelines): x ≈ 1.0
- Area-based: x ≈ 0.6–0.8
Scale factors for processing plants (AACE 18R-97):
Chemical plant: x = 0.5–0.7
Refinery unit: x = 0.6–0.8
Heat exchanger: x = 0.5
Parametric (Cost Estimating Relationships, CER)
C = a × X^b × adjustment factors
X = physical parameter (weight, power, area, complexity)
Developed from regression of historical data
Examples:
Structural steel: ~$2,000–4,000/tonne installed (US 2023)
Pressure vessel (CS): C = 3,000 + 350 × M^0.95 [$ USD; M = weight in kg; rough]
Piping: multiply equipment cost by 0.45–0.90 (factored estimate, Nelson)
Murphy and Pye (fabricated equipment):
C_f = C_CS × F_m × F_p
F_m = material cost factor (CS=1.0; SS 316=3.2; Ni alloy=6.0)
F_p = pressure factor (atmospheric=1.0; 10 bar=1.15; 50 bar=1.5; 100 bar=2.0)
Bottom-Up (Work Package)
Detailed WBS; estimate each work package:
Labor: hours × labor rate
Materials: quantity × unit cost
Subcontracts: from quotations
Overhead + G&A: % of direct cost
Contingency: % based on risk register
Factory cost breakdown (discrete manufacturing):
Total manufacturing cost = Direct materials + Direct labor + Manufacturing overhead
Direct materials: 40–60% typical
Direct labor: 10–25%
Manufacturing overhead: 30–50% (allocated on labor hours or machine hours)
Learning Curve
Cumulative average cost decreases as production experience accumulates
Wright's cumulative average model:
C_n = C_1 × n^b
b = log(r) / log(2) where r = learning rate (e.g., 80% → unit doubles → average cost × 0.80)
Unit cost model (Crawford):
C_unit,n = C_1 × n^(b-1)
Learning rates: aerospace complex = 75–80%; electronics = 80–85%; defense mfg = 85–90%; standard mfg = 90–95%
Effect of 80% learning:
From unit 1 to unit 2: cost × 0.80
From unit 1 to unit 4: cost × 0.80² = 0.64
From unit 1 to unit 8: cost × 0.80³ = 0.512
Discounted Cash Flow (DCF)
Net Present Value:
NPV = Σ CF_t / (1+r)^t - C_0
r = discount rate (WACC typically); C_0 = initial investment
IRR: discount rate where NPV = 0
Payback period: time for cumulative CF = C_0
Life Cycle Cost (LCC):
LCC = C_acquisition + C_operations + C_maintenance + C_disposal
Discount to present value; often 10–30 year horizon for capital equipment
Cost Indices (Escalation)
Chemical Engineering Plant Cost Index (CEPCI):
C_2 = C_1 × (CEPCI_2 / CEPCI_1)
CEPCI 2023 ≈ 800 (vs. 400 in 2000 → prices roughly doubled)
Marshall & Swift (M&S): similar purpose; less current
ENR Construction Cost Index: for civil/structural
Manufacturing Cost Model (Casting Example)
C_total = C_material + C_labor + C_overhead + C_tooling/N_parts
C_material = ρ × V × c_material [$/kg]
C_labor = t_cycle × n_workers × w_rate [$/cycle; w_rate in $/hr]
C_overhead = C_labor × f_overhead (f ≈ 1.0–2.0)
C_tooling: amortize die/mold cost over production run N
Break-even analysis:
N_BEP = (C_tooling_A - C_tooling_B) / (C_variable_B - C_variable_A) (between two processes)
Output
Provide: estimate class, method used (analogical/parametric/bottoms-up), total cost [$ and/or $/unit], cost breakdown (materials, labor, overhead, contingency), contingency % and basis, NPV or LCC if requested, learning curve adjustment, cost index used for escalation.
