LOAD TEST TO COLLAPSE ON A FULL SCALE MODEL SIX METRE SPAN BRICK ARCH BRIDGE

The Transport and Road Research Laboratory is undertaking a research programme to re-examine the present, commonly used, method of assessing the traffic load carrying capacity of brick and stone masonry arch bridges (referred to as the MEXE method). The programme consists of the development of analytical models of their structural behaviour under load, and load tests at full and model scale to calibrate the analyses. This report describes the construction and testing of a full scale model bridge built in the laboratory at Bolton Institute of Higher Education. The bridge had a span of 6m, a rise at midspan of 1m, and an arch thickness of 220mm. The arch was built of two rings of solid concrete engineering bricks. A 1:2:9 cement:lime:sand mortar was used throughout, with a strength of 2.33 N/sq mm. The bridge was 6m wide and the spandrel walls were 436mm thick at road surface level, increasing in two steps to 768mm. The parapets were 216mm thick. The fill was a 50mm graded limestone. A dense bitumen macadam road surface was applied. The total length of the spandrel and wing walls was 14m and the fill was retained by brick end walls which were structurally separate from the wing walls. Loading to failure was applied using hydraulic jacks reacting against prestressing tendons cast into the slab on which the bridge was built. In addition macalloy bar and coupler anchorages were provided along the span of the bridge to enable point loads to be applied across the span and width of the bridge. Instrumentation included load cells, linear variable differential transformers (LVDT), displacement transducers, vibrating wire earth pressure cells, electrical resistance strain gauges, embedded and surface mounted vibrating wire strain gauges, and temperature gauges. The first load test was to apply a 100kN load to a 340mm diameter patch over a grid of points between one springing and midspan. However cracking occurred between the arch ring and the spandrel wall at a load of about 70kN during the first load application so subsequent loads were restricted to 50kN. The final load test was to apply a line load the full width of the bridge between parapets to the road surface at a quarter span point. The line was actually 750mm wide to distribute the load enough to avoid a premature fill failure. Collapse occurred as a four hinge mechanism and the failure mode of the arch ring and the spandrel wall at the west face of the bridge, is illustrated. Comparisons are made between the loads achieved and the loads permitted by a MEXE assessment, and also the collapse loads predicted by a computer program which assumes that failure occurs due to the formation of a four hinge mechanism.