【正文】 across a construction joint， The original design concept had been for in site concrete construction. Inspection and assessment Inspection Inspection work was undertaken in a number of phases and was linked with the testing required for the structure. The initial inspections recorded a number of visible problems including: Defective waterproofing on the exposed surface of the top flange. Water trapped in the internal space of the hollow box with depths up to 300 mm. Various drainage problems at joints and abutments. Longitudinal cracking of the exposed soffit of the central span. Longitudinal cracking on sides of the top flange of the prestressed sections. Widespread sapling on some in site concrete surfaces with exposed rusting reinforcement. Assessment The subject of an earlier paper, the objectives of the assessment were: Estimate the present loadcarrying capacity. Identify any structural deficiencies in the original design. Determine reasons for existing problems identified by the inspection. Conclusion to the inspection and assessment Following the inspection and the analytical assessment one major element of doubt still existed. This concerned the condition of the embedded prestressing wires, strands, cables or bars. For the purpose of structural analysis these elements、 had been assumed to be sound. However, due to the very high forces involved,、 a risk to the structure, caused by corrosion to these primary elements, was identified. The initial remendations which pleted the first phase of the assessment were: 1. Carry out detailed material testing to determine the condition of hidden structural elements, in particular the grouted posttensioned steel cables. 2. Conduct concrete durability tests. 3. Undertake repairs to defective waterproofing and surface defects in concrete. Testing procedures Nondestructive radar testing During the first phase investigation at a joint between precast deck segments the observation of a void in a posttensioned cable duct gave rise to serious concern about corrosion and the integrity of the prestress. However, the extent of this problem was extremely difficult to determine. The bridge contains 93 joints with an average of 24 cables passing through each joint, . there were approx. 2200 positions where investigations could be carried out. A typical section through such a joint is that the 24 draped tendons within the spine did not give rise to concern because these were protected by in site concrete poured without joints after the cables had been stressed. As it was clearly impractical to consider physically exposing all tendon/joint intersections, radar was used to investigate a large numbers of tendons and hence locate duct voids within a modest timescale. It was fortunate that the corrugated steel ducts around the tendons were discontinuous through the joints which allowed the radar to detect the tendons and voids. The problem, however, was still highly plex due to the high density of other steel elements which could interfere with the radar signals and the fact that the area of interest was at most 102 mm wide and embedded between 150 mm and 800 mm deep in thick concrete slabs. Trial radar investigations. Three panies were invited to visit the bridge and conduct a trial investigation. One pany decided not to proceed. The remaining two were given 2 weeks to mobilize, test and report. Their results were then pared with physical explorations. To make the parisons, observation holes were drilled vertically downwards into the ducts at a selection of 10 locations which included several where voids were predicted and several where the ducts were predicted to be fully grouted. A 25mm diameter hole was required in order to facilitate use of the chosen horoscope. The results from the University of Edinburgh yielded an accuracy of around 60%. Main radar survey, horoscope verification of voids. Having pleted a radar survey of the total structure, a