A Saw-Tooth Learning Curve for the Economic Assessment of Small Modular Reactor Deployment

Small modular reactors (SMRs) are expected to benefit economically from serial production and learning by doing, yet their early deployment is likely to involve regulatory feedback, design changes, production interruptions, and discontinuous construction. Conventional experience curves assume smooth and monotonic cost reduction and therefore cannot represent these disruption-driven cost increases. This study introduces a saw-tooth learning curve that extends the classical experience curve with temporary cost shocks, deployment cadence effects, and organizational forgetting. Deterministic and Monte Carlo simulations are used to examine alternative learning rates, disruption frequencies and magnitudes, and deployment structures. The results show that disruption frequency has a stronger effect on early-stage cost bias than the assumed learning rate. Under batch deployment, mean cost bias over the first ten units increases to 18.1% for frequent medium disruptions and to 34.9% for high-magnitude disruptions, while the P90 unit-10 cost under stop-start deployment exceeds the classical estimate by approximately 36%. A historical comparison with the French pressurized water reactor program further shows that the saw-tooth formulation reproduces observed cost transition directions and reduces mean absolute percentage error from 13.75% for the classical model to 2.06%. The proposed formulation provides a transparent method for incorporating nonmonotonic learning and deployment instability into prospective SMR economic assessments.