Year: 2026
Type: article
Abstract: This paper addresses the problem of identifying the optimal iron content in Zr-1%Nb alloy that minimises the corrosion rate of WWER-1000 fuel rod cladding under primary-circuit conditions (350 °С, 16.5 МPа). Based on seven interval-valued corrosion-rate measurements of specimens with Fe=0.012...0.192 wt.%, we show that ordinary leastsquares (OLS) regression is structurally inadequate: no polynomial of degree 𝑚 ≤ 4 is compatible with all measurement intervals, and OLS estimates of the optimal iron content (0.108...0.117 wt.%) carry a systematic bias of ΔFe=0.006...0.015 wt.%. A set-membership (bounded-error) identification framework is applied instead. The feasible parameter set P* – all 5th-degree polynomial coefficients simultaneously compatible with every measurement interval – is characterised via linear programming. The resulting LP envelope localises the corrosion-rate minimum at Fe ∈[0.101;0.102] wt.% with, V = 0.02975 mg/(dm²·h), an order of magnitude more precisely than OLS. The identified systematic OLS error corresponds to a corrosion-rate increase of up to 22% above the achievable minimum – an unacceptable margin for cladding with a 60-year design service life.
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W7204193486