Once you have Live Data feeding your Qatium model, sensor icons will populate the assets assigned to them.
Qatium will display a model accuracy score and a model deviation panel for each sensor type implemented.
Both model deviation and model accuracy are indicators of the difference between the data received from the sensors and the one calculated by the model.
In networks without AMI data but with inlets and outlets informed in zones, Qatium will automatically update the network demand with the flow balance improving the network accuracy levels.
The model accuracy is an hourly percentage indicator calculated from every model deviation value of the hour. The goal is to have a model accuracy as high as possible with the maximum value being 100%. Although we can consider a good model accuracy score when it’s above 75%.
The accuracy score is calculated from the different model deviation values, the lower the value the more accurate the model will be.
The model deviation is calculated for every time step (1 hour at a time) and for tank level, flow, and pressure. It averages the difference between the sensor value and the model value, assigned to the same asset for each hour step. The goal is to have a model deviation as small as possible with the minimum value being 0%. The lower the model deviation, the closer are live data and simulated values.
To visualize the model deviation values, you can click on the double arrow headed button on the right-hand side of the window.
By turning the toggle ON, you will be able to visualize the locations where there’s a higher model deviation, or something unplanned is happening in the network (like a pipe burst), hence where you might need to focus your efforts.
The deviation values will be shown in:
- Green, for errors below 10%
- Orange, if the error is between 10% and 20%
- Red, if the error is above 20%
If there is no sensor value for a specific timestep or variable, the model deviation panel will not display that deviation.
Once you’ve set thresholds for your assets, Qatium will analyze and process your live data to find relevant events and display warnings when:
- Tank levels are unusually low
- Junction pressures are unusually low
- Pipe flows are unusually high
If you want to see how the network behaves with different thresholds, you can edit them directly from the asset tooltip. In that moment you’ll switch to scenario mode, and you will see the changes inside the asset detail information.
To persist the change you’ll need to edit the threshold through the import process. Learn how to do it by reading this article.
If you have live data added to your network and delimited zones, and you need to evaluate if there is an abnormally high amount of non-revenue water (NRW) in a specific sector, we’ll give you some insights to find the cause and improve water balance in your network.
The zone tooltip will show extra information for:
- Net flow: total flow consumed in a zone (by customers, leaks, etc.)
- Efficiency: percentage of water that actually gets to customers
In addition, with the Zones panel you will be able to quickly locate the inlet and outlets in each sector. This will help you to understand how the zones are supplied in case you notice that any value is out of the expected. This way you will quickly find out if the error is in the flow meters, in the AMI data or if there really is an anomalous consumption in the network.
To activate it you just have to:
- Click on the "Zones" icon
- Switch the toggle ON
- Hover over each zone
The icons for inlets and outlets will appear on the map:
- Alone, if there are no readings available.
- With a sensor icon in gray: when in the current time there are no readings available.
- With a sensor icon in blue: when there are readings for the current period.
The "current time" or “current period” it’s the hour that you see in the timeline but also near dates (e.g. a reading at 10:57 is valid and used when the timeline is pointed at 11:00).
To have a clearer view of this information you can activate the Overview work mode.
To link your historic or real-time SCADA, sensor, or AMI data to Qatium, please visit our Ingest API article.