First-floor taps starved in a three-storey house
A three-storeyed house in Bhubaneswar has a 1000 litre HDPE tank on a 1.2 m stand on the roof terrace, with the terrace about 9.5 m above ground level. The service connection is an old 20 mm (3/4 inch) galvanised iron pipe, about 18 m long, with a stopcock at the boundary and a 15 mm gate valve near the tank. With the tank full, a pressure gauge on the ground-floor tap reads 1.1 bar when nothing is drawing water. The owner complains that the first-floor tap gives only a trickle whenever the ground-floor or kitchen tap is opened, that flow almost stops after 8 pm, and that the problem became much worse after two new 15 mm taps and a shower rose were added on the first floor last month. There is no booster pump. When a section of the GI pipe was opened, the bore was heavily roughened with rust tubercles.
1. Explain, using the numbers given, why the first floor is starved while the ground floor is acceptable.
Read the worked answer
Start with what is available. Static pressure at the ground tap is 1.1 bar, which corresponds to roughly 11 m of water column - consistent with a tank water level about 11 m above that tap (9.5 m terrace plus the 1.2 m stand, less the tap height). So the supply head itself is not generous; it is a low-rise gravity system, not a pressurised one.
Subtract the height first. A first-floor tap is about 4 m above ground, so it loses about 0.4 bar of head before any water moves: static pressure there is only about 0.7 bar in the best case.
Then subtract friction, which is where the real loss is. Flow through a pipe loses head roughly with the square of the velocity, and the velocity for a given draw-off is set by the bore. An 18 m run of 20 mm GI that is tuberculated may behave like a 12–15 mm pipe; its roughness also rises several-fold. When a ground-floor tap opens, the whole draw passes through that same lonely 20 mm line, and the friction loss over 18 m plus bends and two valves can easily exceed 0.5–0.6 bar. The residual at the first floor therefore falls to near zero - exactly the 'trickle when another tap is open' symptom.
The 8 pm effect is peak-hour low pressure in the municipal main feeding the tank, so the tank fills more slowly and the standing head at the time of use is lower still.
Why last month made it worse: two extra 15 mm taps and a shower rose raised the simultaneous design flow. The pipe was already at its limit; extra outlets did not create water, they only divided a fixed, loss-limited supply - and each new fitting added its own minor losses.
2. Recommend remedial measures in order of cost and effect, and justify the order.
Read the worked answer
Order the fixes by benefit per rupee, and by what the diagnosis says is limiting.
- Remove the bottleneck before adding pressure. Replace the tuberculated 20 mm GI service line with 25 mm (or preferably 32 mm) CPVC or HDPE from the boundary to the tank, and ream/deburr and use full-bore valves. This attacks the actual cause: high friction in a choked small bore. It is the cheapest single fix and reduces loss at every outlet at once.
- Reduce simultaneous demand and balance the system. Fit a 15 mm stop valve or a flow-restricting screw on the ground-floor and kitchen taps so that one tap cannot steal the whole line, and replace the shower rose with a low-flow head. Cheap, immediate, but a management fix, not an engineering cure.
- Raise the head. Increase the tank stand from 1.2 m to 2.5–3 m, or shift the tank to the highest point of the terrace. Every extra metre of stand gives about 0.1 bar at every tap for a modest, one-time cost; check the structure can carry it and that the terrace waterproofing is not disturbed.
- Only if 1–3 are insufficient, add a booster. A small 0.5 HP pressure-boosting pump with a pressure switch and a small diaphragm tank on the down-take from the tank, set to cut in at about 0.3 bar and cut out at about 1.5 bar. This is the costliest and the one that adds running cost and a failure point, so it comes last.
- Housekeeping: clean the tank, check that the float valve is not throttling the fill, and check the gate valve spindle is fully open - a half-closed gate valve is a common hidden loss.
Say clearly to the owner that a booster on top of a choked 20 mm GI line would only pressurise a pipe that still cannot deliver flow, and would raise the leak risk in old joints.
3. How would you verify, after the work, that the complaint has actually been resolved?
Read the worked answer
Measure, do not rely on 'it feels better'. Use two tests the owner can see.
1. Static and residual pressure. Fit a pressure gauge with a 15 mm adaptor at the first-floor tap. Record static pressure with all taps closed, then residual pressure with the ground-floor tap fully open and with two taps open. Accept the job only if the residual at the first floor stays at roughly 0.3 bar or more with the ground tap running - enough for a shower and a cistern to work together. Record the before and after figures; the 'before' number is your proof of improvement.
2. Flow test. Time the filling of a 10 litre bucket at the first-floor tap with the ground tap closed, then with it open. A 15 mm tap at 0.5 bar should give roughly 6–8 litres per minute; the 'trickle' complaint corresponds to 1–2 litres per minute. Repeat at peak hour (about 8 pm) so the test covers the worst case, not the best.
3. Leak and settling check. After the pipe replacement, do a hydrostatic test on the new line at 1.5 times working pressure (here about 1.6 bar, held for an hour) before boxing up or concealing, and check every new joint for weeping. Then run all outlets together for ten minutes and look for water hammer, vibration at the booster (if fitted), and any drop in the tank level faster than the float valve can refill.
4. Written handover. Note the pressure and flow readings, the pipe size and route, the booster settings and the date in the owner's record. That converts your work into documented performance and protects you if the complaint returns.