Reactor Period–Reactivity Modelling and Cross-Validation of In-Hour and PRCT Methods for the LEU Nigeria Research Reactor-1

Accurate conversion of reactor period to reactivity is essential for control-rod calibration and reactor-kinetics interpretation in low-power research reactors. This study cross-validates the In-hour equation and Period Reactivity Conversion Table (PRCT) using secondary low-power calibration data from the low-enriched uranium Nigeria Research Reactor-1 (LEU NIRR-1), and develops a reproducible reactor-specific period–reactivity model. Six reactor periods (45–180 s) were analysed. In-hour estimates ranged from 5.64 to 1.28 mk, whereas PRCT estimates ranged from 5.50 to 1.20 mk. The paired methods showed close agreement (mean absolute error = 0.120 mk; RMSE = 0.122 mk; mean absolute percentage deviation = 4.38%). Linear agreement analysis gave R^2=0.99994, with a slope of 1.0140 (95% CI 1.0032–1.0247). Exploratory Bland–Altman analysis showed a positive bias of 0.120 mk and 95% limits of agreement of 0.069 to 0.171 mk. To make the predictive model fully reproducible, the arithmetic mean of each In-hour/PRCT pair was fitted to a power law, reproducing the dissertation model as ρ=337.48T^(-1.070). The fit was strong (R^2=0.9963; adjusted R^2=0.9954), and leave-one-out cross-validation yielded RMSE = 0.159 mk and MAPE = 5.00%. The model exponent was -1.070(95% CI -1.161to -0.980), supporting an approximately inverse period–reactivity relationship over the investigated interval. The two conventional methods are therefore mutually consistent for the analysed LEU NIRR-1 period range, while the fitted model is best regarded as a compact interpolation and cross-checking tool. It is recommended that independent validation with new reactor measurements is required before operational use.