Predicting Strength of FRP‑Wrapped Concrete Columns with a Smart AI Model

The safety of buildings that use fiber‑reinforced polymer (FRP) to strengthen concrete depends on knowing how much weight the wrapped columns can hold. A new approach uses an adaptive neuro‑fuzzy inference system (ANFIS) to estimate that capacity. The model learns from a large set of 812 real‑world tests, covering many different column sizes and FRP types. It takes four key inputs: the plain concrete strength, the diameter of the column, how thick the FRP sheet is, and the stiffness of that FRP. With these numbers it predicts the maximum compressive load the wrapped column can resist.

To train ANFIS, researchers applied a Levenberg‑Marquardt algorithm that fine‑tunes the model while stopping early to avoid overfitting. The result is a tool that generalises well to new columns not seen during training. When the ANFIS predictions were compared with five popular analytical formulas, the AI model showed less spread in its errors. That means it gives more consistent results across a wide range of strengths and designs. The study concludes that this ANFIS framework is reliable for early design checks and quick assessments of FRP‑strengthened columns. It offers engineers a practical, data‑driven alternative to traditional equations.

Actions