Multi-sensor data fusion platform for reagent-free total chlorine monitoring in water

Overview

This solution offers a chemical-free approach to continuous total chlorine monitoring in drinking water by combining several simple, low-cost sensors through advanced data fusion. Rather than relying on wet-chemical or membrane-based chlorine probes, the platform interprets signals from optical spectroscopy (LEDs and photodiodes), pH, temperature, and conductivity sensors using process-model and machine-learning algorithms. The result is a compact, reagent-free measurement system capable of estimating free chlorine and combined chloramine species in real time.

The technology addresses a long-standing pain point for water utilities and industrial water operators: conventional chlorine sensors require frequent calibration, consume reagents, and are sensitive to interfering substances such as humic acids and turbidity. By selecting wavelengths that carry information on both the target analytes and the interfering parameters, the platform can self-correct for common matrix effects, reducing or eliminating the need for manual recalibration.

Technical specifications

Core approach:

• Sensor data fusion of optical, pH, temperature, and conductivity measurements
• Algorithms grounded in known process models combined with machine learning to infer parameters not directly measured
• Wavelength selection based on absorption profiles of free chlorine, monochloramine, dichloramine, and trichloramine
• Built-in compensation for turbidity, humic acids, and other interferences through multi-wavelength optical design

Design priorities:

• Reagent-free operation, eliminating ongoing chemical costs and waste
• Miniaturization as an explicit design criterion, enabling portable and distributed deployments
• Simple, robust sensor suite suited to long-term continuous monitoring

Target performance:

• Detection of total chlorine below 0.1 mg/L in tap water
• Measurement accuracy of approximately ±20%

Technology readiness level

The concept is supported by prior modeling work from the research group on multi-sensor oxychloride monitoring and by partner experience demonstrating that smart wavelength selection can correct for interference in water-quality measurements. Laboratory validation experiments have been initiated but were paused due to external disruptions, and experimental results are currently pending. The immediate next step is to build a first-generation multi-sensor prototype and develop an initial data-fusion algorithm to demonstrate proof of concept. The platform is therefore at an early to mid-stage technology readiness level, with strong theoretical and modeling foundations and prototype validation as the next milestone.