Case study

UV Disinfection Low Transmittance Dose Shortfall Case Study

Case study on UV disinfection dose shortfall from low UVT, lamp aging, sleeve fouling, peak flow, dose permissives and release validation.

Branch: Environmental Engineering
Content: Case study
Updated: Jul 03, 2026
Revision: v1.1.0 · reviewed

UV disinfection can fail even when water is still flowing through the reactor and all lamps appear energized. The delivered dose depends on flow, reactor volume, UV transmittance, lamp output, sleeve fouling, sensor calibration, hydraulics, and the validated operating envelope. A plant that judges UV performance only from lamp-on status can overclaim disinfection credit.

This case study follows a wastewater reuse facility where indicator organisms rise after a storm period. Operators initially suspect sampling error because the UV system did not trip. The engineering review shows that low UV transmittance, aged lamps, fouled quartz sleeves, and peak flow reduced the delivered UV dose below the permit basis.

The purpose is to connect UV dose, contact time, UV transmittance, lamp condition, fouling, flow control, maintenance response, and compliance validation into one operating decision.

Operating Envelope and Evidence Boundary

The failure boundary is the full disinfection credit chain: upstream filtration, UVT, turbidity, reactor flow, reactor hydraulics, lamp output, sleeve condition, UV intensity sensors, UVT analyzer, flow meter, control algorithm, diversion logic, microbial sampling and permit record. A UV reactor can be electrically healthy and still outside its validated dose envelope.

The case should be treated as a release decision, not only a maintenance event. The plant can claim disinfection credit only when online evidence proves that the credited dose algorithm remains inside the validated envelope. If the dose is below the validated basis, the correct state is hold, divert, reduce flow or operate under a non-credit condition until corrective evidence is complete.

The evidence boundary should include:

current flow and peak-flow history;
UVT measurement and grab-sample confirmation;
UV intensity sensor status and calibration;
lamp age and output factor;
sleeve fouling and wiper performance;
upstream filtration/turbidity condition;
validated dose algorithm and alarm basis;
interlock or diversion test;
microbial confirmation after corrected operation.

This boundary prevents the common mistake of treating lamp-on status as disinfection proof.

Case Context

The plant disinfects filtered secondary effluent before non-potable reuse discharge. The UV system was validated for a minimum dose of:

D_{req}=40\ \text{mJ/cm}^2

at the credited operating envelope.

Item	Value
UV reactor effective volume	$4.0\ \text{m}^3$
Operating flow during event	$500\ \text{m}^3/\text{h}$
Peak wet-weather flow observed	$575\ \text{m}^3/\text{h}$
Reference average intensity at clean condition	$2.0\ \text{mW/cm}^2$
Reference UV transmittance basis	$75\%$
Measured UV transmittance during event	$62\%$
Lamp output factor from lamp age	$0.74$
Sleeve fouling factor from inspection	$0.82$
Minimum validated dose requirement	$40\ \text{mJ/cm}^2$

The UV intensity sensors report reduced intensity but remain above the alarm threshold. The event therefore becomes a question of dose margin, not whether the electrical system is powered.

Failure Signature

The failure signature is a dose-margin collapse:

all lamps can be energized;
UV intensity is reduced but not necessarily in hard alarm;
UVT drops after wet-weather or solids/turbidity deterioration;
sleeves show fouling or wiper underperformance;
lamp-hour records show output derating;
flow approaches or exceeds the validated contact-time basis;
indicator organisms rise after the hydraulic/water-quality event.

That pattern points to a validated-dose shortfall, not a simple power failure. The first response should be to protect credited discharge while the dose envelope is rebuilt.

Contact Time at Event Flow

Convert event flow to cubic metres per second:

\displaystyle Q=\frac{500}{3600}=0.139\ \text{m}^3/\text{s}

The hydraulic contact time is:

\displaystyle t=\frac{V}{Q}

With:

V=4.0\ \text{m}^3

the contact time is:

\displaystyle t=\frac{4.0}{0.139}=28.8\ \text{s}

This is only a screening contact time. Real UV validation also depends on reactor hydraulics, short-circuiting, lamp arrangement, sensor location, dose-response testing, and the validated dose algorithm.

Hydraulic Caution

The volume-over-flow contact time assumes idealized use of the reactor volume. Real UV reactors can have non-uniform velocity, short-circuiting, dead zones, lamp shadowing and nonuniform dose distribution. The calculation is useful for screening, but it does not replace validated reactor performance.

If peak flow rises during wet weather, the loss is double: contact time falls, and UVT often falls at the same time. That coupling makes peak wet-weather operation the critical release condition.

Effective UV Intensity

Use a first-pass delivered-intensity screen:

I_{eff}=I_{ref}f_{UVT}f_{lamp}f_{foul}

where:

$I_{ref}$ is clean reference average intensity;
$f_{UVT}$ accounts for UV transmittance loss;
$f_{lamp}$ accounts for lamp aging;
$f_{foul}$ accounts for sleeve fouling.

The UV transmittance factor is:

\displaystyle f_{UVT}=\frac{UVT_{event}}{UVT_{ref}}

\displaystyle f_{UVT}=\frac{62}{75}=0.827

Therefore:

I_{eff}=2.0(0.827)(0.74)(0.82)=1.00\ \text{mW/cm}^2

The lamp status display can therefore be misleading. The lamps are on, but the effective intensity is about half of the clean reference value.

Instrument Validity

Before using the dose calculation, verify the instruments that feed it. UV intensity sensors can drift or foul. UVT analyzers can be affected by sample cell fouling, air bubbles or calibration. Flow meters can be biased at high flow. Lamp-hour factors can be wrong after lamp replacement if the maintenance record is not updated.

The release record should state whether each input is measured, calculated, assumed or manually entered. A permissive based on invalid inputs is not a validated control.

Delivered Dose

A simplified UV dose screen is:

D=I_{eff}t

Using:

I_{eff}=1.00\ \text{mW/cm}^2,\quad t=28.8\ \text{s}

gives:

D=1.00(28.8)=28.8\ \text{mJ/cm}^2

The dose shortfall is:

G_D=40-28.8=11.2\ \text{mJ/cm}^2

The delivered dose is only:

\displaystyle \frac{28.8}{40}=0.72

or $72\%$ of the required value. The system should not claim the validated disinfection credit at this condition.

Hold Decision

At this point the case has enough evidence for a hold decision. The plant should not wait for the next microbial sample before protecting the discharge. The online dose screen already shows that credited operation is outside the validated basis.

The correct operating decision is to reduce flow, divert, increase UV capacity if available, or declare that disinfection credit is not being claimed for the affected flow. Any water already discharged during the shortfall should be reviewed under the site compliance procedure.

Flow Limit Without Maintenance

If the plant does not immediately restore UV transmittance, lamp output, or sleeve condition, the required contact time is:

\displaystyle t_{req}=\frac{D_{req}}{I_{eff}}

\displaystyle t_{req}=\frac{40}{1.00}=40\ \text{s}

The maximum allowable flow at the existing effective intensity is:

\displaystyle Q_{max}=\frac{V}{t_{req}}

\displaystyle Q_{max}=\frac{4.0}{40}=0.100\ \text{m}^3/\text{s}

Convert back to cubic metres per hour:

Q_{max}=0.100(3600)=360\ \text{m}^3/\text{h}

So the plant must either reduce credited flow to about $360\ \text{m}^3/\text{h}$ or restore dose capacity. Continuing at $500\ \text{m}^3/\text{h}$ is not a compliant operating state under this simplified dose basis.

Flow Permissive

The control system should translate this calculation into a permissive, not a spreadsheet-only conclusion. The permissive should compare calculated dose with the validated setpoint and act before the reactor leaves the credited envelope.

Possible states are:

State	Dose and flow condition	Action
normal credit	dose above setpoint with margin	discharge/reuse allowed
warning	dose above setpoint but below warning margin	investigate UVT, lamps, fouling and flow
hold	dose below setpoint but diversion available	divert or reduce flow
non-credit operation	dose below setpoint and flow continues under approved contingency	record no disinfection credit
invalid data	required dose inputs not valid	hold or divert per site rule

The exact actions depend on the permit and process design, but the key point is that the system needs a defined state before the event.

Corrective Action Screen

The reviewed corrective package is:

clean quartz sleeves and verify wiper operation;
replace aged lamps that are below the validated output factor;
improve upstream filtration control to recover UV transmittance;
calibrate UV intensity and UVT instruments against traceable checks;
add a flow permissive that diverts or reduces flow when calculated dose falls below the validated setpoint;
document the validated dose algorithm and alarm basis instead of relying on lamp-on status.

After maintenance and upstream correction, assume:

UVT_{corr}=70\%,\quad f_{lamp,corr}=0.92,\quad f_{foul,corr}=0.96

The corrected intensity is:

\displaystyle I_{eff,corr}=2.0\left(\frac{70}{75}\right)(0.92)(0.96)=1.65\ \text{mW/cm}^2

At the normal event flow of $500\ \text{m}^3/\text{h}$ , contact time remains:

t=28.8\ \text{s}

The corrected dose is:

D_{corr}=1.65(28.8)=47.5\ \text{mJ/cm}^2

The dose margin is:

\displaystyle M_D=\frac{47.5-40}{40}=0.188

or about $18.8\%$ . That is acceptable for the normal event flow, but the peak wet-weather flow must still be checked.

Maintenance Evidence

Corrective action is complete only when the maintenance action is tied to measured dose recovery. Cleaning sleeves without checking intensity, replacing lamps without updating lamp-hour factors, or improving filtration without confirming UVT does not close the case.

The maintenance evidence should include before/after UVT, before/after intensity, lamp replacement record, sleeve inspection, wiper function, flow confirmation and control-system alarm/permissive test.

Peak Flow Check

At:

Q_{peak}=575\ \text{m}^3/\text{h}

the flow is:

\displaystyle Q_{peak}=\frac{575}{3600}=0.160\ \text{m}^3/\text{s}

The contact time becomes:

\displaystyle t_{peak}=\frac{4.0}{0.160}=25.0\ \text{s}

Using the corrected intensity:

D_{peak}=1.65(25.0)=41.3\ \text{mJ/cm}^2

The peak-flow margin is:

\displaystyle M_{peak}=\frac{41.3-40}{40}=0.033

or about $3.3\%$ . This is too narrow to treat peak flow as routine. The operating envelope should include an interlock, a conservative alarm margin, or a lower permitted peak flow until more validation evidence is available.

The corrected maximum flow is:

\displaystyle Q_{max,corr}=\frac{V I_{eff,corr}}{D_{req}}

\displaystyle Q_{max,corr}=\frac{4.0(1.65)}{40}=0.165\ \text{m}^3/\text{s}

or:

Q_{max,corr}=0.165(3600)=594\ \text{m}^3/\text{h}

The measured peak of $575\ \text{m}^3/\text{h}$ is below this value, but with little margin. The release decision should therefore use a lower operational alarm limit than the theoretical maximum.

Peak-Flow Release Margin

A (3.3%) peak-flow margin is too small for routine release because UVT, lamp output, sensor cleanliness and flow measurement all have uncertainty. The plant should set an operational limit below the theoretical maximum unless site-specific validation supports the full range.

If peak flow must be accepted, the evidence should include repeated UVT under wet-weather solids conditions, verified flow-meter accuracy near peak, lamp output margin and microbial confirmation during representative high-flow operation.

Why Microbial Sampling Alone Is Too Slow

Post-disinfection microbial sampling is essential, but it is not fast enough to control the UV reactor in real time. A failed sample can arrive after water has already been discharged or reused. UV operation needs online evidence:

flow;
UV intensity;
UV transmittance;
lamp status and lamp age;
sleeve cleaning state;
validated dose calculation;
diversion or shutdown status.

Microbial samples validate the control strategy. They do not replace the dose permissive.

Compliance Evidence Timing

Online dose evidence is preventive. Microbial sampling is confirmatory. A failed microbial sample can reveal that the barrier failed, but it usually cannot prevent the failed water from leaving the system. The monitoring plan should therefore link online permissives to immediate action and microbial results to validation, reporting and corrective-action effectiveness.

Release and Validation

The corrected UV system should be released only after the plant proves that the validated dose is available across representative operating conditions.

Acceptance evidence should include:

UVT analyzer calibration and grab-sample comparison;
UV intensity sensor calibration and sensor-window cleaning records;
lamp-hour records and replacement thresholds;
sleeve fouling inspection and wiper function test;
flow-meter verification at normal and peak ranges;
validated dose calculation implemented in the control system;
diversion or flow-reduction interlock test below dose setpoint;
upstream filtration record showing solids and turbidity control during wet weather;
microbial sampling after corrected operation to confirm the process response;
permit record stating when disinfection credit may and may not be claimed.

Release criteria should include:

D_{calc}\geq40\ \text{mJ/cm}^2

with:

UVT\geq68\%

and:

Q\leq560\ \text{m}^3/\text{h}

unless site-specific validation supports a broader envelope. The system should divert, reduce flow, or alarm before the validated dose falls below the permitted limit.

Revalidation Triggers

Revalidate the UV envelope if:

lamps are changed to a different type or supplier;
UVT analyzer, intensity sensors or flow meter are replaced;
upstream filtration performance changes;
peak flow basis changes;
sleeves, wipers or reactor hydraulics are modified;
microbial results worsen without an obvious sampling explanation;
the control algorithm, alarm setpoint or permissive logic is changed.

These triggers keep the dose claim connected to the actual hardware, water quality and validated algorithm.

Operator Handover

The handover should state the credited flow limit, minimum UVT, dose setpoint, alarm margin, diversion rule, lamp replacement threshold, sleeve-cleaning response and microbial confirmation plan. Operators should know which signal creates a warning, which signal removes disinfection credit, and who can authorize return to normal operation.

Engineering Lessons

The first lesson is that UV disinfection is not proven by lamp-on status. Dose depends on intensity, transmittance, fouling, lamp age, flow, and validation basis.

The second lesson is that water quality is part of disinfection capacity. Low UV transmittance from upstream solids, color, iron, organics, or poor filtration can reduce dose even when the UV reactor is mechanically intact.

The third lesson is that compliance requires fast permissives. Laboratory microbial results are important, but the operating system must prevent uncredited flow from leaving the reactor before samples return.

Good environmental engineering therefore treats UV disinfection as an integrated treatment and control system. Flow, UVT, intensity, maintenance state, validated dose, diversion logic, and compliance records must agree before disinfection credit is claimed.

REF

Disciplines