103. Winding Bridge, 163 on the Grand Junction Canal

My thanks to Rob Gilbert (@robertgilbert86 via twitter) for sending me along for this one. It involved a nice bike ride on a sunny afternoon. What attracted Rob’s attention when he passed on a boat was the cast iron rib beneath the arch.

As usual a pdf version is available here.


There is lots to see here, though and lots of speculation to make. We are on the River Gade, which here forms part of the Grand Junction Canal and this is bridge 163. It is clearly a winding bridge, designed to allow horses to change sides when pulling barges. The horse would come under the bridge from the right on the far side, wind up and over and exit behind the camera in this view.


That movement is clearer in the plan view. Most of the remaining pictures come from a 3D model done very hastily from less than an hour’s worth pf photography. Normally for this work I would use a 12m pole and get the camera close to every part of the bridge. Cycling here on my Brompton I was limited to the camera I wear on my belt and a 1.2m monopod for reach. At the range imposed, the white painted areas offered very little visible texture and the black even less. Despite that, Hamish managed to conjure a model of real value. You can explore a lower resolution version here.


This square elevation and all subsequent images come from a 3D photographic model made from around 1000 photographs taken with a Sony RX100VI, sometimes lifted or lowered on a light monopod.

Even in this distant shot, the arch does not seem an entirely happy shape.


An upward view of the soffit shows some of the difficulties that can arise in modelling. The white painted surface is very bland. Where I got the camera reasonably close, the brick pattern makes the modelling work but there is a large area where the surface is made up of large triangles and part of the rib is just missing altogether. The high contrast between white and black means that getting definition on the black is more difficult.


A closer elevation shows more detail, but again, the rib is missing in places. Just below the 163 sign is an apparent ripple in the brickwork where the software didn’t quite get everything to fit. In photogrammetry that tends to show quite starkly and it is often masked (but present) in point cloud surveys.


There are a few minor blemishes in this image too. If I were doing this commercially, I would use a 10m window washing pole to get the camera into more and better places. I would also use a laser level, as I did at Rutters Bridge: BoM99, to put levelling dots on the surface.



Striking a horizontal line across the face shows the disturbance to the bedding of the bricks that tells us why that rib went in. There can be little doubt that the arch has sagged substantially. The mortar beds would not normally slope down to mid span.

The span measures at 30ft6in which is quite unlikely. A little juggling with possible shapes says the original span was 30ft, the side radii 6ft and the main radius 20ft producing a rise of 9ft. The present rise is around 8ft and the end radii are unchanged but most of the crown is at about 29ft radius.


Looking closer at the left hand of the arch, the white line is horizontal. Note that there is a high point, a little to the right of the springing but the beds go down to the left and more steeply down to the right.


The right-hand end movement is more obvious. Here, both the sag into the span and the tilt of the abutment is more pronounced. I think this is the side that moved first. The tilted beds in the abutments indicate that the rising thrust as the span sagged tipped the abutments backwards so the effective increase in span is rather greater than the 6in measured at ground level.


With that extra movement, the change from 20ft to 29ft radius at the centre is viable.

It is not surprising, therefore, that the engineer felt the need to stiffen the bridge up. A quick response was needed, and with the canal open, closing the bridge to rebuild it would have been difficult. So, an iron rib to hold it up. Which raises the question “would that work?” and if so how?

If the arch had continued to spread the cast iron would surely have fractured with relatively small movement, at the connections if nowhere else. There would have been no way of stressing the rib into place beyond packing behind it with mortar so the initial load would have been negligible. If the central section of the arch continued to sag by creep, it would transfer load to the rib, which might then generate some thrust and provide support. I suspect the truth is that the rib is doing very little. The arch found a natural settlement and stopped. A similar effect can be more easily seen at Over Bridge, BoM 4.

It is, perhaps, worth noting that this is all normal behaviour for an arch, only the scale here is larger. The very soft ground would have provided weak foundations and there would have been a tendency for the abutments to rotate backwards as well as spread. The thrust in the arch would increase considerably if the sag was as big as it seems so there would be considerable movement before the resistance grew.


Since the rib doesn’t render properly in the model, we need to go back to the original photos to get something worthwhile.


By pushing the photos to bleach out the white, it becomes possible to see detail in the rib. It is T shaped, Cast iron. It has end plates cast on at the ends and the joints and is in three parts. The springing plate sits on a stone pad.

The T shape says it was not intended to take bending in the sagging sense because cast iron needs an inverted T to work like that. I think the engineer saw it just as an additional layer to the arch. The connections tell a similar story, two bolts in a cast flange do not constitute a moment connection, they are merely locators.


Notice the mortar packing above the rib which shows how far the shape differs from that of the bridge.


And that is month 103, Next month another look at “relieving arches”.