Water systems in factories: matching supply, quality and continuity to what the process actually needs

One incoming pipe, several unrelated purchases

Most sites draw a single supply and then split it into uses that have almost nothing in common: cooling make-up, boiler feed, washdown, product contact, welfare and the fire reserve. Costing them as one utility line hides the fact that a small stream, usually the one closest to the product, carries most of the treatment expense. Break demand down by use before anyone sizes equipment. A plant that knows how much of its draw genuinely needs polishing, and how much runs perfectly well on incoming quality, often finds its treatment plant was scoped for a volume it never had to treat.

The quality specification comes from the failure mode

Ask what each process fails on rather than what grade sounds prudent. Hardness ruins one operation, dissolved solids another, microbiological load matters only where product contact or aerosol exposure exists, and some steps care about nothing beyond temperature and pressure. Have the process owners write the parameter down with the named failure beside it. Municipal supply varies seasonally and between sources, so a single sample proves little — ask the utility for its range, not its average. Where output is regulated, the acceptable grade is fixed by the product licence and the regulator, and internal preference does not enter it.

Interruption is a production risk wearing plumbing clothes

Supply loss is the failure planned for least. A cut lasting part of a shift is survivable at one plant and immediately terminal at another, and the difference is nearly always storage that nobody sized deliberately. Establish what the site can keep running on stored volume, how long a refill takes, and whether that storage sits upstream or downstream of treatment. Pressure deserves separate attention: a supply that behaves overnight can sag when an entire shift starts machines together. The fire reserve carries its own obligations, and quietly borrowing from it for process use is a decision that surfaces during an insurance survey.

Reuse arguments are usually won on discharge, not intake

Recovery projects get pitched on the water bill, which is generally the smaller half of the case. The larger amounts sit in trade effluent charges, treatment chemistry, the energy spent heating water that then goes down a drain, and the ceiling that effluent volume places on production growth. Cascading cooling blowdown, final rinse or recovered condensate into a lower-grade duty often pays back through avoided discharge. The catch is complexity: a recovery loop adds monitoring, failure modes and a contamination path back into a clean stream. Decide honestly whether the site's maintenance capability can run one, because an abandoned loop is a stranded asset that still needs attention.

What belongs to the engineer, and what you owe them

Pipe sizing, material selection, backflow prevention, storage design and anything connected to the fire supply are engineering decisions reserved for a qualified engineer working to the water and building regulations in force where the site sits. Nothing here is a design or a specification, and it should not be handed to a contractor as one. What the business owes that engineer is a demand profile by use and by hour, the parameters each use fails on, the continuity the operation needs, the growth expected within the building's remaining life, and the discharge route with its stated capacity. Arrive with those and the design conversation is short.

Frequently asked questions

Should we treat the whole incoming supply to the highest grade any process needs?

Rarely. Treating everything to the strictest requirement means paying capital and running cost to polish water destined for washdown and cooling towers. The usual pattern is a common incoming stage that protects the site, then point-of-use treatment sized for the small stream that actually needs it. The exception is where the risk of a wrong connection is serious enough that separating grades creates a bigger hazard than treating everything, which is a judgement the process owner and the designer make together.

How do we find out whether the supply can support an expansion before we commit?

Ask the utility for a written capacity assessment against the peak demand you expect, not the average, and get the answer before equipment is ordered. Utilities distinguish between what the connection can physically deliver and what the network can support at your busiest hour, and reinforcement can take longer than the plant build itself. Landlords rarely hold this information, so if you lease, ask for the meter details and go to the supplier directly rather than accepting a general reassurance.

Who inside the plant should own water quality day to day?

Give it to one named person with the authority to stop a process, usually within engineering or quality depending on whether the risk is equipment damage or product failure. Shared ownership between maintenance and production reliably produces a situation where sampling happens when someone remembers. The owner should hold the sampling schedule, the trend record, the escalation route when a parameter drifts, and the standing relationship with whoever maintains the treatment plant.