Dynamic AI

FlowReporter's Dynamic AI harnesses the power of machine learning to analyse past water consumption patterns detected by your flow sensor. This innovative technology calculates alarm thresholds tailored to your specific usage, ensuring optimal leak detection and water management.

While training mode provides a single alert threshold, Dynamic AI offers a nuanced approach by setting different thresholds depending on the time of day and day of the week. This adjustment mirrors your actual water usage, establishing higher thresholds during peak usage hours and stricter thresholds during periods of inactivity.

Dynamic AI generates two sets of thresholds based on learning data:

• Weekday: Active from Monday to Friday.
• Weekend: Active on Saturday and Sunday.

To initiate learning, simply navigate to the "Dynamic AI" section under the alarms settings of your device page.

Tailored Learning Options

Use default values for most cases or customise learning parameters for advanced scenarios:

1. Set minimum and maximum threshold values using the alert slider. Note that the maximum might be overridden by the device owner.
2. Specify the learning timeframe, considering data from the previous week, month, last 3 months, or all-time.
3. Enable "Drip Detection" for heightened sensitivity to minimal flow events, such as dripping taps. If you are worried about these types of leaks, enable this setting.
4. Initiate learning by clicking the "Learn" button.

Graphical Representation and Fine-Tuning:

• Visualise learned thresholds on a graph, allowing for adjustments or relearning with different time frames if necessary.
• Easily refine values by clicking "Edit" and entering precise numbers as needed.

FlowReporter's Dynamic AI empowers users with adaptive, data-driven alarm thresholds, enhancing water consumption monitoring and leak detection capabilities with unparalleled precision.