Summary
A public water utility serving one of India's largest metropolitan areas needed to reduce pollution load beyond what conventional dosing could achieve, and to see plant health in real time. Plaxonic designed and deployed an AI-driven smart dosing system that brought BOD and TSS well below target and put live plant monitoring in operators' hands.
The Challenge
The utility's traditional dosing systems could not consistently bring BOD and TSS down to required levels, and there was no real-time way to monitor the health of the plant or its equipment. Operators were dosing without live feedback, reacting to lab results after the fact, with no early warning when equipment or treatment drifted.
They needed a system that could automate and optimize dosing, monitor plant and equipment health continuously, fold in predictive maintenance, and use AI to calculate and reduce pollution load.
What We Built
Plaxonic built an Intelligent Self-Administered and Self-Monitored Automatic Chemical dosing system (ISASMAC), engineered with AI and ML. Rather than dosing to a fixed schedule, the system adjusts coagulant dosing to the actual flow, predicts maintenance needs before equipment fails, and brings the dosing system and the health data of every electrical and mechanical device onto one platform. IoT devices stream real-time data from the dosing system and plant equipment into an intuitive dashboard that shows system health and utilization at a glance.
The platform included: real-time IoT monitoring of dosing data, equipment health, and AI-based analytics; a color-coded operations dashboard; historical plant-health and utilization charts; automated calculation of pollution-load reduction and the aeration oxygen needed to neutralize it; an energy-savings calculator; ingestion of lab parameters such as SVI and sludge age to recommend better treatment; AI-based dosing-pattern suggestions; and a preventive-maintenance plan with an advanced alarm system. It was integrated with plant and laboratory modules including LIMS.
Our Approach
We worked collaboratively with the utility to understand their plants and constraints, then designed an end-to-end solution across hardware, software, and IoT, delivered in agile iterations and deployed into live plants. The software was customized to their requirements and connected the dosing system with the health information of all electrical and mechanical devices on a single platform.
The Technology
IoT for device connectivity and real-time data. Machine Learning to analyze dosing data and predict both optimal dosing patterns and maintenance needs. Big Data and analytics to process high volumes of dosing data. AI for health scoring, report generation, and pollution-load calculations.
Tech stack: Next.js and Tailwind CSS on the web; Node.js APIs with MongoDB and Apache ZooKeeper; MQTT for IoT device control; MongoDB and InfluxDB for system and sensor data with Redis caching; Kafka and Python for data analytics; Kubernetes and GitLab CI/CD for orchestration and delivery.
The Result
After four months in operation, the system brought BOD and TSS well below the 10 PPM target, recording an average of 5 PPM BOD and 8 PPM TSS. Real-time monitoring and predictive maintenance improved overall plant health and efficiency, and the utility planned to extend the solution to more of its plants.
