PSI - Issue 84

A.A. Khalil et al. / Procedia Structural Integrity 84 (2026) 983–990

989

5.2. Method 2 - Piece‑by‑Piece Mechanical Demolition without Intermediate Supports

Method 2 relies on the deck’s redistribution capacity of load through diaphragms and slabs while removing ~4 m girder segments sequentially (Figure 12). To model tendon’s development length after cuts, prestressing was conservatively set to zero over 4 m longitudinally and 3.7 m transversely adjacent to each removal.

1.00

Stage 2

Stage 6

Stage 8

Stage 10

Stage 14

Deflection [cm]

-25.

Stage 24

Stage 32

Stage 36

Stage 62

Stage 151

Fig. 12. Stage analysis: subsequent removal of 4mt portion of the deck, vertical deflection at selected stages.

Retaining portions of the diaphragms before moving to the next girder proved effective in limiting deflections during demolition, as these elements continued to redistribute across the remaining deck (Figure 13). This finding not only improved local stability but also reduced the risk of progressive failure, enhancing overall safety of demolition operations.

70.0

Deflection [cm]

(a)

0.00

(b)

Fig. 13. Deflection comparison between removing (a) and retaining (b) intermediate diaphragms.

The most critical condition was found after four girders had been removed, leaving only two to carry redistributed loads during the first cut in the fifth girder. Here, analyses incorporating damage states (inactive tendons) and material safety factors demonstrated that the deck safely sustained demolition with a global steel safety factor of ≈ 1.5. Additional observations included longitudinal seat behavior: maximum outward translation approached ≈ 300 mm during late‑stage diaphragm removals in the final girder, highlighting the need to confirm actual seat length (~730 mm) and inter‑span gap (~250 mm) in the field prior to execution and, where necessary, adjust sequencing to limit horizontal motion near expansion gaps. 5.3. Method 3 - Mechanical Demolition with Pier and Ground Interaction Method 3 was designed as an induced simultaneous drop of three girders, which required extending the analysis scope to include full pier models (caps, shafts, and foundations), along with a soil representation capable of capturing impact‑induced lateral lo ads, shear forces, and bending moments during the collapse stage. Across all scenarios, no pier failure was observed (Figure 14), confirming adequate reserve capacity under the studied conditions.