The low water levels in 2006 made it feasible to dig that out and install a proper facility for the pump.The first step was to relocate the pump temporarily so that I could keep water levels as low as possible. Then I dug a large pit in the clay floor. Not very easy - the clay is quite sandy there and it was still flooding despite the low levels.
I lined the pit with a permeable geotextile (the sort usually used for weed-suppression) and part-filled it with gravel. I drilled 2cm diameter holes in a sawn-off plastic barrel and covered it with the geotextile, and immersed that in the pit, adjusting the gravel layer to get the top of the barrel flush with the floor surface. Then I backfilled with gravel to the surface and installed the pump in its new home.
I haven't quite worked out how to protect the well from clay falling into it while at the same time allowing easy recovery of failed pumps - experience has shown that it is necessary to be able to haul the pump up from below a couple of metres of water.
The more recent excavations needed a floor before I could start any more extensive shoring up. The old section has a floor made of wooden floorboards on a gravel-filled frame. This works quite well but I was keen to start on something rather more permanent and I decided on a concrete raft, particularly as the new pump well allowed me to keep water levels low.I cut out the mud down to a depth of between 5 and 10cm (the base was sloping slightly) and installed a sub-floor drainage channel. This was a length of perforated drainpipe, with the perforations covered in metal gauze. It was laid along the length of the area close to the centre-line, in a trench lined with geotextile, backfilled with gravel and, eventually, covered in concrete (shown on the diagram below). The lower end of the pipe led to a gap in the wall of the pump well.
I lined the floor area with a light plastic sheet and dumped several bags of pre-mixed concrete on it, mixing in the required amount of water.
The concrete began to set quickly, which was fortunate because the pump started playing up. I spent the rest of the day and much of the night keeping it going - eventually I went to bed and by morning the concrete was under 10cm of water.
As it turned out the concrete had set enough by then not to be harmed by its immersion. I had to cut a drainage channel in it (the surface was slightly concave) and that took a lot of chiselling, so I'm reasonably confident that the concrete floor is sound.
Some time in Autumn 2007 there was a substantial cave-in. A hefty chunk of roof over the concrete floor fell down, and the sides collapsed inwards so that the wooden floor was also covered.It's going to take a while to clear that lot up, and I can't begin until water levels start to drop in Spring. Even though the pump has enough umph to empty the hole, the drains flood after a day or so of pumping - sooner if it's raining at the time.
So in the meantime I will concentrate on building a controller for the pump. The float-switch gets stuck on the sides of the pump well so I need something rather more sophisticated. Watch this space.