A task facing the newly formed Railtrack in 1994 was ensuring that its overbridges could carry the new 40t and 44t gvw trucks introduced onto the UK's roads.
It was a task that was simply too large for Railtrack to manage, and so it left local authorities to make their own investigations. They did, however, need to make sure that the conclusions arrived at by a number of different agencies were consistent.
"Our job was to act as technical advisors to Railtrack – later Network Rail," says Tony Small, Pell Frischmann's head of rail.
"The project was due to complete in 1999, but went on much longer. Some of the work was to prepare for privatisation in valuing major assets like the Forth Rail Bridge, but most of the work was in checking," he says.
Bridgeguard 3, as the project was known, required 2,500 bridges across the country to be assessed, some by local authorities, some by consultants. It is still ongoing in some parts of the country, although Pell Frischman's involvement has come to an end.
"We checked the reports bridge-by-bridge, highlighting any deficiencies, to ensure they were up to standard," he says.
The checking process was designed to discover whether the data and its analysis were sound. The reports varied a great deal. Generally, the data collected in inspections were sound, the analysis less so.
"We took the inspections at face value, but we had to ensure their conclusions were right,"
But, assuming the data was sound, Small found that much of the analysis was not, and many bridges were being failed – not suggesting they were dangerous, but that they couldn't take the new, heavier loads.
"We took the inspections at face value, but we had to ensure their conclusions were right. To measure corrosion, we looked at the detailed construction, making a moπdel, and then making a structural appreciation which took a varying amount of work.
"By and large, we took the measurements on, and re-did the assessment. More often than not, reports gave a lower capacity for the bridge, when it was able to take a higher capacity, but this did sometimes work the other way around."
Small and his team used eigen values to model beams, and if this was inconclusive, then ran a full finite element analysis. This gave his team a virtual model of the bridge beams, to see where strengths lay.
"These were usually classic deep-edge half-through girder bridges," says Small.
"We were looking for lateral torsional buckling. The top flange is in compression, and will tend to buckle.
"You might improve the bridge's rating, the loading limit – that was the exercise. What is the capacity of the bridge – do we recommend strengthening or placing a load limitation? There were limits on the prescribed methods, leading to technical failures which were not realistic."
The result was that bridges were found – generally – to be sturdier than first thought.
Although estimates are difficult, Small believes the work saved Network Rail some £10M in bridge remediation or even replacement. It covered many historic bridges that would otherwise have been lost. The project has been so successful that Pell Frischman's team won the 2006 Historic Bridge and Infrastructure Award.