Your Chlorine Sensor
Is Lying to You —
Here's 65 Days of Proof
How a UK Water Utility Exposed Systematic Over-Dosing Caused by Membrane Sensor Fouling
A Field Evaluation at Compton Durville Water Treatment Works, Wessex Water. October 2025 – February 2026. 73 Operational Cycles | 65 Days of Continuous Data.
Key Findings: 65 Days of Membrane Fouling
Executive Summary
The operators at Compton Durville thought their dosing was under control. Siemens Depolox showed residuals at setpoint. When Wessex Water installed an MP5 alongside, it revealed the actual residual was rising steadily — climbing approximately 10% per cycle.
Iron deposits fouled the Depolox membrane, causing progressive under-reading. The PID system responded by dosing more chlorine. Not a dramatic failure but slow, invisible drift. Weekday calibrations partially corrected it; weekends accumulated fouling 20% faster.
The MP5 required zero intervention for 5 months — providing a baseline of what actual residual control should look like in high-iron borehole water.
Site Background: Compton Durville WTW
Compton Durville Water Treatment Works operates with significant operational constraints and water quality challenges typical of deep borehole installations in the UK.
Borehole Water Quality
Elevated iron (Fe) concentration creates progressive fouling challenges for membrane-based sensors.
Sensor History
Previously tried Kuntze and ATi probes — both experienced similar fouling issues in high Fe/Mn waters.
Sensor Deployment
Depolox provides PID feedback control; MP5 was installed as independent monitor only.
Operational Cycle
System shuts down 09:00–15:00; operates 15:00–09:00 daily (16-hour cycle).
Evidence of Progressive Fouling
Over 65 days and 73 operational cycles, consistent data revealed a systematic pattern of membrane degradation and increasing over-dosing.
+10.67%
Start of cycle: 0.6038 ppm → End: 0.6714 ppm
0.53%/hour
Consistent iron deposit buildup
5.3%
Mid-cycle residual vs. setpoint
+13.95%
Sunday (no calibration)
+8.32%
Tuesday (post-calibration)
Day-of-Week Pattern: Calibration Effect
Fouling accumulation shows a clear weekly cycle. Weekday calibrations partially reset the fouling signal, while weekends show uninterrupted accumulation.
| Day of Week | Mean Rise (%) | Notes |
|---|---|---|
| Sunday | 13.95% | Highest accumulation |
| Monday | 10.82% | Post-weekend fouling |
| Tuesday | 8.32% | Post-calibration effect |
Operator Calibration Effect
Weekday Accumulation: +0.0639 ppm average residual rise per cycle when operators perform daily calibrations.
Weekend Accumulation: +0.0767 ppm average residual rise per cycle without calibration intervention.
Net Weekday Effect: Nearly zero net change (operator calibrations correct fouling signal back to setpoint), but the underlying fouling continues undetected.
Critical Finding: Even daily calibration cannot keep pace with cumulative fouling. The membrane degradation is progressive and inevitable without physical cleaning.
Root Cause: Membrane Fouling in High-Iron Waters
The failure mechanism is well-understood but not adequately addressed in membrane sensor design for high-iron, high-manganese borehole waters.
Iron Deposits on Membrane
Borehole water with elevated iron (Fe) creates diffusion barrier, progressively degrading sensor signal.
Signal Degradation Feedback Loop
As sensor under-reads due to fouling, PID system increases dosing to reach setpoint — causing over-chlorination.
History of Membrane Issues
Previous probes (Kuntze, ATi) had identical fouling problems in high Fe/Mn waters.
Membrane Sensors Not Ideal for High Iron/Manganese
Diffusion membrane technology susceptible to particulate and mineral fouling in source water with elevated metals.
Conclusions
MP5 Accuracy Over 5 Months
Zero maintenance required. Consistent, drift-free readings — the control baseline.
Depolox Fouling Progression
Under-reading of 10.67% per cycle. Iron deposits create diffusion barrier with predictable accumulation rate.
PID Over-Compensation
Control system increases dosing to reach setpoint, resulting in 5–6% average over-dosing of chlorine.
Calibration Limitations
Daily calibration corrects the displayed reading but does not reverse underlying membrane fouling. Effects resume on every cycle.
Widespread Problem
Kuntze and ATi probes showed identical fouling patterns in the same environment. Membrane sensor vulnerability is systemic.