Reliability-based mix design for concrete compressive strength using a physics-prior residual-learning surrogate with calibrated uncertainty

Pengfei Qu, Lei Song, Sihan Wang · 2026

Read Reliability-based mix design for concrete compressive strength using a physics-prior residual-learning surrogate with calibrated uncertainty by Pengfei Qu, Lei Song, Sihan Wang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0350575.txt.

Description

Pengfei Qu, Lei Song, and Sihan Wang blend physics with learning to predict concrete strength and design mixes that meet strength targets reliably, even with little data. A physics based starting formula uses water, binder, and age, then a small neural net learns only what physics misses, keeping results sensible and monotone where they should be.

On a public set of over one thousand mixes, the hybrid scored high accuracy and held up when trained on only forty percent of data. Checks confirmed age and cement help strength while water hurts, as physics expects. Honest uncertainty bands covered the promised share of cases. Builders can design greener mixes with less cement and known risk.

Catalog record

CREATORPengfei Qu, Lei Song, Sihan Wang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62315
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0350575

Browse the full library · About the Tower of Babel