Most water treatment plants have not changed how they operate in decades. Operators still walk rounds checking meters that SCADA already reads, chemical dosing gets adjusted manually from lab results that arrive hours after the process has already drifted, and a coagulant dose that was correct at eight in the morning is often wrong by noon once raw water turbidity shifts. The plants breaking this cycle are not ripping out SCADA and starting over, they are layering AI directly on top of the systems already running, and the operators who've made the switch describe it as finally getting to see the plant the way the data has always shown it, just hours sooner.
Chemical Dosing That Adjusts Before Water Quality Does
iFactory connects to your existing SCADA over OPC-UA, calibrates AI baselines from live turbidity and pH data, and adjusts coagulant and disinfectant dosing continuously, cutting chemical costs by up to 15% while improving treated water quality.
Six Process Stages, One Continuous View
Raw Water Intake
Turbidity, alkalinity, and flow are baselined continuously so coagulant dosing starts from real conditions, not yesterday's jar test.
Coagulation & Flocculation
Dose optimization models trained on historical jar tests and SCADA dose-response data adjust coagulant continuously as raw water shifts.
Sedimentation
Basin efficiency is tracked against turbidity removal targets so underperforming basins are flagged before they affect filtration load.
Filtration
Filter headloss and turbidity breakthrough are modeled per filter cell, timing backwash before efficiency drops instead of on a fixed clock.
Disinfection
Chlorine residual and CT compliance are tracked continuously against SDWA thresholds and your state's primacy permit limits.
Treated Water Quality
Every stage feeds a single quality dashboard, so a drift at intake can be traced through to its effect on final treated water instantly.
What Continuous Dosing Optimization Delivers
Typical reduction in chemical costs from continuous dose optimization
Reported cut in chemical usage where AI-driven dosing has fully replaced manual adjustment
Water quality compliance rate achieved by plants running integrated SCADA and AI monitoring
Reduction in unplanned downtime from proactive pump and filtration maintenance
See Your SCADA Tags Mapped to a Live Dosing Model
Bring a recent turbidity and dosing trend from your plant and we'll show how continuous optimization would have adjusted the dose in real time.
Manual Rounds vs. Continuous AI-Driven Operation
| Operational Task | Manual, Lab-Driven Process | iFactory Continuous Optimization |
|---|---|---|
| Coagulant dosing | Adjusted from jar tests, hours delayed | Continuous, from live turbidity and pH |
| Filter backwash timing | Fixed schedule | Triggered by headloss and breakthrough model |
| CT compliance tracking | Calculated periodically by operator | Continuous, alerted before threshold breach |
| Equipment maintenance | Calendar-based inspection rounds | Predictive, from live SCADA condition data |
| Compliance documentation | Compiled manually for state reporting | Auto-generated, compliance-ready templates |
Frequently Asked Questions
Will this disrupt our existing SCADA or PLC infrastructure?
No, the connection to your plant is established as a read-only OPC-UA link into your existing SCADA system, meaning your current control logic, PLC programming, and operator interfaces continue running exactly as they do today. All process tags including turbidity, flow, chemical dose, CT, filter headloss, and UV intensity are mapped into the treatment stage model without requiring any changes to your control system architecture. All AI processing runs on-premise, so plant data does not need to leave your network for the optimization models to function. You can review the integration approach for your specific SCADA platform through this booking link.
How does the dosing model learn what the correct dose should be?
The coagulant dose optimization model is trained on historical jar test results paired with SCADA dose-response data from your own plant, so the baseline reflects how your specific raw water and treatment train actually behave rather than a generic industry curve. Once live, the model continuously refines its recommendations as new turbidity, pH, and dose-response data comes in, adjusting for seasonal raw water changes that a fixed dosing table would miss entirely. Operators retain full override control throughout, the model recommends and alerts, it does not remove human judgment from the loop.
Can this help specifically with CT compliance and state primacy reporting?
Yes, CT compliance thresholds are configured directly against Safe Drinking Water Act requirements and your specific state primacy permit limits, and the system tracks disinfection contact time continuously rather than relying on periodic manual calculation. When a value approaches a threshold, the system generates a pre-alert rather than waiting for a violation to occur, and every automated detection generates a compliance-ready record with supporting evidence already attached. This is designed to plug directly into the reporting format your state agency already expects.
What kind of chemical cost savings should we realistically expect?
Plants moving from manual, lab-delayed dosing to continuous optimization commonly see chemical cost reductions in the range of 15%, with some facilities reporting usage cuts closer to 23% once the model is fully calibrated to plant-specific conditions. The savings come primarily from eliminating over-dosing, which is the default response operators use to stay safely above compliance thresholds when they can't see real-time water quality. For a savings estimate based on your plant's current chemical spend, reach out through our support page.
How long does it take to get from initial connection to active dosing recommendations?
The OPC-UA connection to SCADA is typically established within the first days of a deployment, after which historical data is loaded and stage-specific baselines begin calculating from about a week of live operating data. Dose optimization models generally reach reliable recommendations within a few weeks as they calibrate against your plant's actual jar test history and dose-response patterns. Most plants see compliance pre-alerts active early in the process, with full dosing optimization following shortly after baseline calibration completes.
Let Your SCADA Data Do More Than Get Read on a Round
Book a free automation readiness assessment and see what continuous dosing optimization looks like on your own plant's data.
