Plumbing water leak

One morning the PowerPal power monitoring app on my wall-mounted iPad Mini showed some unexpected activity, which I recognized as the house supply water pump running continuously for a couple of hours. That's the block in the graph in image 1 between 8:40am and 10:30am with a baseline of just below 1kW.

Closing the main stop valve between the pump and the house stopped the pump, so there was definitely enough of a leak to cause the pump to run. The pump and flow control unit were relatively new and was not showing any error condition.

All the taps were off, toilet cisterns weren't leaking and there were no telltale damp areas in walls, floors , ceiling spaces or outside around the house or yard. Closing off the valve feeding into the hot water heater did not stop the pump, indicating that the leak was not in the hot water tank or lines.

There were some clues:

The water coming out of taps was muddy if the lines were left to drain out before reopening the main valve again. This indicated that the leak point was sitting in muddy water, that is, underground.
The muddiest water came from the verandah garden tap.
If the system was shut off most of the time and only opened briefly when absolutely needed, the amount of muddiness reduced significantly because the water that had pooled around the leak had time to dissipate.
Attaching a clear plastic tube to a connection at a low point near the ground would indicate the level of water in the lines as they drained through the leak. This indicated that the leak was below floor level.

1. Unexpected power usage

I called a plumber who didn't show up the following afternoon as I thought had been arranged, so I decided to continue investigating on my own, starting with digging between the pump and the house.

I discovered that the line from the pump branches into two separate lines underground, one supplying the kitchen and the other supplying the rest of the house. I was able to reconfigure and add isolation valves and found the leak was in the "rest of house" line. I could now at least run water to the kitchen without going outside to manually switch the pump on.

2. Pipes disconnected during a previous reworking

3. Two new pipes replaced a buried tee branch

4. New isolation valves

Initially, because muddy water drained out when the pipe was disconnected, I suspected the leak was between the point of entry under the house (image 3 top left of centre) and what I suspect is the take-off point for the wet areas, which are all on that side of the house. I bought a cheap endoscope and was able to see the joiner ('T' piece) connection that I measured as 3.3m from the entry point. This was a huge concern as this section is below a tiled concrete bathroom slab.

After making an inflatable pipe plug that I could use to block the pipe at various points, combined with the clear plastic tube as described earlier, I determined that the leak was not in this section afterall. That was a relief even though I'd still not located the leak.

Inserting a garden hose into the pipe indicated that the 1 inch poly pipe ran for at least 20 metres without obstructions such as right-angle bends etc. That is, right under the house and beyond. The next step was to try to locate an entry point at the other end and repeat the process with the inflatable plug. The only place to start that search was at the verandah garden tap.

The next day I hired a radiodetection pipe locator and attempted to trace the pipe from where it ran under the house. It was looking like some flooring would need to be opened up, however from previous investigation, this wing of the house has a floor build on top of a previous floor and it was not going to be easy if the leak was under all of that. To trace a non-metalic pipe requires feeding a conductor through the pipe and energising it using the transmitter that comes with the set. I was not able to find an unambiguous trace, although there were some indications I had no confidence in them.

Digging down beside the verandah garden tap (image 5) exposed an elbow joiner and a horizontal pipe section running back under the verandah slab. Initially it appeared to run at a 45 degree angle however this was just a bend around the footing of the verandah post above. Unscrewing the vertical pipe section allowed the endoscope to be run along the horizontal section where another joiner could be seen about 60cm from the opening. So the digging continued following the horizontal pipe. Taking a break and with the garden tap reattached to the elbow and the isolation valve opened I ran a washing machine cycle.

Resuming the dig there was now evidence of saturated soil in one wall of the hole but it was on the opposite side to where the pipe ran. This was a very lucky break.

Progressing with the dig into the softer wet side of the hole and repressurizing the line soon exposed a small blowhole in the soil (image 6) only 60cm under the slab. It may not be necessary to break or cut the slab to make a repair. The objective now was to fully expose the leak and enough of the pipe to cut out the damage (image 7 right of centre) and insert a joiner (image 8).

5. Digging beside the verandah tap

6. Blowhole created by water spraying from the leak

7. Fully exposed leaky section

8. Damage cut out, stones and brick removed and joiner inserted

When the pipe was cut to remove the damaged section, water flowed out from the left side (connected to the 'T' piece) suggesting that the rest of the house is upstream to the left. Where the right side goes to is so far undiscovered.

I suspect that the damage (the deformed wall shown in image 9) has been there since the pipe was installed because it appears to be an impact crease and the pipe has been buried since the verandah roof was constructed and the slab was put down, probably at the time the last major renovations were done in 1983, i.e. fourty years ago. The pipe appears to have been laid at that time as the pipe section leading to the garden tap riser bent around the concrete verandah post footing. Half bricks can be seen lying alongside and in contact with the pipe (image 7) and these would have been thrown or dropped into the trench. Although there is evidence the pipe was bedded in sand, the presence of gravel, stones and brick in direct contact indicates that sand was not covering it when the trench was filled in.

The wall has split along the crease, which is both longitudinal and latitudinal, and will have borne significantly higher stress concentration.

9. Impact damage

I have long suspected there was a very slow leak in the system because the previous pump would run periodically for no reason (it included an internal pressure bladder that would prevent the pump running continuously). The new pump delivers higher pressure (90 psi) and over time has caused the split to grow and open up further.

I have also ordered an adjustable pressure reduction valve to go after the pump to set the system to within the 80 to 60 psi range considered normal.

Costing

$300 for components to reconfigure the buried supply line branch with above ground isolation valves.

$60 for a WiFi Endoscope (10m)

$10 for a bicycle inner tube (used to make the DIY pipe plug).

$150 for the hire of the cable and pipe locator.

$18 for 30m of multistrand steel wire to use as a trace conductor.

$DIY for labour.

The joiner used to fix the damaged section came from the reconfigured branch.

Total cost was approximately $530. It's not clear is my insurance covers fixing leaks or just fixing damage from leaks but either way the minimum cost to me would have been $1000. The maximum could have been double that or higher.

Importantly to me, I was able to use the locator to also trace the power line between the house and garage distribution boxes and I now have some new useful tools, materials and system improvements including removing a redundant section of copper pipe and a tap that was part of the previous hot water tank that was removed in 2015 when the ground source heat pump was installed.