Why commissioning is where systems fail
Almost every component arrives working. What fails is the installation's assumptions — pressures never checked, settings left at factory defaults, protective valves omitted as "extras". These are the recurring offenders, in roughly the order they bite.
Dosing and chemical
- Syphoning through an unvalved injection point. The tank drains itself; the process gets unmetered chemical. Prevention: injection/back-pressure valve at the quill, and the switched-off syphon test before handover.
- FKM seals on hypochlorite. Mystery leaks and lost prime within months. Prevention: EPDM for hypochlorite, stated on the order.
- Doser sized to run at 3% of capacity. Erratic dosing that measurement keeps "confirming and unconfirming". Prevention: the 20 – 80% rule; dilute the chemical.
Vessels and valves
- Backwash hydraulics never verified. The bed silts, mudballs, channels; the media gets blamed. Prevention: medium's backwash rate × bed area, checked against valve, drain and source before the media order.
- No vacuum breaker on a drainable FRP vessel. First proper drain-down or siphon event crumples the liner. Prevention: air admittance at the high point, fitted at installation.
- Factory-default valve settings. Regeneration for a resin volume nobody has, backwash for a different filter. Prevention: programme resin volume, salt dose, hardness and cycle times per the actual build, and write them on the vessel.
- Salt-dose economics never set — the softener regenerates daily on a bed exhausted 30%. Prevention: size the interval per the sizing article; volumetric control where use varies.
- Filter valve on a softener (or a brine cycle left active on a carbon vessel). Hard water on schedule / water wasted on schedule. Prevention: check the brine port and the code's first digit.
Membranes
- Chlorinated water reaching the elements — spent carbon, missing SMBS, a "temporary" bypass. Permanent rejection loss, measured in weeks. Prevention: free-chlorine test after the carbon at commissioning and every service.
- Preservative never flushed / first permeate used. New elements ship in preservative; the first production water is not product. Prevention: flush per the datasheet before anything downstream sees permeate.
- Reused vessel closure seals. Salty permeate immediately after a service visit. Prevention: new seals every opening — and when quality drops post-service, suspect the seal before the membrane.
- Recovery "adjusted" on site. The concentrate valve gets throttled for more product; the tail scales within the quarter. Prevention: recovery is a design number; lock it, log it, and re-project before changing it.
Disinfection and general
- UV before the tank, or without its prefilter. The lamp treats water that gets recontaminated, or shadows past half its dose. Prevention: UV last, 5 µm ahead, lamp on the calendar.
- No gauges, no sample points. The system that cannot be diagnosed — every fault above hides longer without them. Prevention: gauges either side of every filtration stage; a sample tap between every major stage. It is the cheapest instrumentation that exists and the first thing value-engineering deletes.
Rule of thumb: commission with three tests and a pen — syphon test on every doser, free-chlorine test after every carbon stage, drain-down test with the vacuum breaker watched — and write every setting on the equipment it belongs to. An hour of commissioning discipline buys years of not meeting this list again.