Keropos Lokal sebagai Cacat Penampang Kritis: Bukti Eksperimental Kapasitas Aksial dan Lokalisasi Keruntuhan Kolom Beton Bertulang

This study evaluates axial capacity and failure localization in reinforced concrete columns with a controlled local cross-sectional reduction used as a model of honeycombing. Eight 125 × 125 × 750 mm columns were loaded in nominally concentric compression at 28 days: two intact controls and two specimens each with a defect at the support (K0), quarter height (K1/4), and midheight (K1/2). The 40 mm-high defect removed 387.84 cm³, equivalent to 3.31% of the gross column volume, and left a 77 × 77 mm section, corresponding to a 62.05% local area loss. Mean maximum loads for N, K0, K1/4, and K1/2 were 389.44, 250.14, 217.37, and 220.87 kN, respectively. The reductions relative to N were 35.77%, 44.19%, and 43.29%. N and K0 shared the first batch, whereas K1/4 and K1/2 shared the second batch, whose mean cylinder strength was lower (19.15 versus 24.41 MPa). Thus, the latter two reductions combine defect and batch differences. Within the second batch, K1/4 and K1/2 differed by 1.61% in mean load. Failure localized at the reduced section in all defective specimens. The contribution is a descriptive comparison of controlled defect locations and failure patterns; two replicates per group, batch differences, and regular defect geometry limit causal inference and application to field honeycombing.