TSE and RO water treatment plant performance assessment

Overview

The facility treats and reuses water on site. Treated sewage effluent (TSE) is processed through multimedia filtration and a reverse osmosis stage before being distributed for irrigation and other non-potable uses. The plant is not a standalone utility — its output, its reject stream and its backwash demand all sit inside the building's overall water balance, and each affects the others.

The assessment set out to establish how the plant was actually operating against how it was intended to operate, and where the gap between the two was costing water, energy or plant life.

Basis of the assessment

Meeting an output target is not the same as performing well. A treatment train can produce its required volume while running at poor recovery, consuming more backwash water than it needs, or shortening membrane life through operating practice rather than equipment fault.

The assessment therefore began as a survey of what was installed, its configuration and its observed operating behaviour — not as a fault investigation.

Engineering challenge

Three constraints shape work of this kind.

A plant in service cannot readily be taken offline for assessment, so the investigation has to be non-intrusive and read from normal operation.

Where the water balance is not fully instrumented, plant performance cannot simply be read off a meter. It has to be inferred from the readings that do exist, with the assumptions stated rather than buried.

And the plant is one consumer among several. An improvement that raised RO recovery but increased backwash frequency, or reduced reject volume at the cost of membrane fouling, would move the problem rather than solve it.

Technical analysis

The approach was to establish the water balance first and judge the plant second.

Inflow, product, reject and backwash streams were traced through the plant, and the available meter readings and operating records were used to establish which quantities are measured, which are derived, and which are not known at all. Being explicit about that third category matters: a balance built on an unstated assumption looks rigorous and is not.

Against that, the operating behaviour was examined for recovery ratio, filtration performance and backwash demand, differential pressure trends across the filtration stages, dosing practice, and the cleaning-in-place (CIP) regime and its interval relative to observed performance decline.

Findings and recommendations

The assessment produced engineering observations and recommended actions covering:

• Water balance instrumentation — the metering points needed to make plant recovery measurable rather than inferred, so future assessment is arithmetic and not argument.
• Backwash and reject practice — reviewing initiation criteria and duration against filtration condition rather than elapsed time alone.
• CIP interval — considering cleaning frequency against observed performance decline rather than schedule alone.
• PPM scope — bringing the maintenance schedule into line with how the plant is actually loaded.
• Operating documentation — recording configuration, setpoints and operating intent so that plant behaviour is reviewable by the next engineer.

Results

Performance improvement opportunities were identified. Stating a saving would require recovery to be directly metered, so no percentage is published — a figure here would be a claim the available data does not support.

Lessons learned

The most useful outcome of an assessment like this is often a measurement gap rather than a fix. Where instrumentation can only answer whether enough water was produced, that becomes the only question asked — and recovery, reject fraction and backwash demand stay invisible, which means they are never managed.

The lesson transfers to most utility plant in facilities management: if a quantity is not metered, it is not managed, and no amount of inspection substitutes for the meter. The first engineering recommendation on a plant like this is often instrumentation, not equipment.

Second, cleaning and backwash regimes inherited as fixed schedules deserve scrutiny. A timer is a proxy for condition, and proxies drift.