We used GPR to investigate construction details and identify building defects in a London school. Specifically, we determined if precast hollow core units had reinforcement placed within them and had been infilled with concrete as specified.

Building defects investigation

Cracking had been observed in the roof slab of a London school adjacent to a wall. Poor construction standards and building defects were suspected to have caused the problem.

The roof construction comprised hollow-core precast units, each containing six cores. The as-built drawings showed the tops of three alternate cores broken out for 500 mm, a reinforcement bar inserted, and the core filled with concrete to that point. Every other core should, therefore, be filled with concrete for a distance of about 500 mm from the supporting wall.

Survey objectives

The client wished to establish, by non-destructive means, the extent to which the cores had, in fact, been filled with concrete in the vicinity of the cracks.

GPR survey method

Sandberg used Ground Penetrating Radar (GPR) to establish the presence and extent of infill concrete in the cores of nominated hollow-core precast units. A GSSI SIR 3000 Ground Penetrating Radar system was used with a 2.0 GHz palm antenna. The antenna is compact and ideal for scanning within 45 mm of an obstruction.

Due to the roof construction detail, including a layer of insulation, all GPR scans were undertaken on the slab soffit. Single scan lines were collected within the nominated area up to 0.8 – 1.0 m from the adjacent internal wall.

Survey findings

GPR detected the following variations within the hollow cores:

  • Completely filled in with concrete
  • Partially filled in with concrete
  • Unfilled, no concrete

The results of the building defects were presented in CAD drawings; a different-coloured hatch indicated the state of the infill. The distribution of infilled cores was very haphazard, and the extent of the infill was not in accordance with the specification, indicating a poor standard of construction. No reliable correlation between the infill concrete and crack positions could be established.

GPR detected occasional linear features within the hollow cores, likely to be the specified reinforcement. It must be noted, however, that GPR will not reliably detect reinforcement within an empty or partially filled hollow core because the strong reflection from the void is likely to mask the smaller, weaker reflection from the reinforcement. Consequently, all rebar may not have been detected.