OpenFlows | Water Infrastructure - WaterSight - Zone Details

Product(s):	WaterSight
Version(s):	10.00.
Area:	Documentation

Overview

The zones information is only referred to flow. The chart displays initially a time-window of 2.5 days back into past to 12 hours into future with the range-slider below the main graph turned on. The range-slider below shows by default a 3 month signal history. The is automatically obtained by making the flow balances of the zone (sum of all inflows minus all outflows plus the difference in tank storage).

Shows detailed information of the flow for a specific zone. Zones should correspond to isolated areas of the system, surrounded by closed valves and where all inflows and outflows (if exist) must be measured. Zones can include for example DMA (District Metered Areas), pressure zones or operational zones. The configuration of the inflow and outflow sensors as well as tanks levels that may exist inside the zone are configured in the Zones configuration page (at the administration level) by the user.

The user can understand if "patterned" variables lie within the expected range of values. Particularly useful to quickly spot any anomalies in a specific zone (that can represent bursts, leaks, tank overflows, pressure problems, maintenance works, etc) in real time. The pattern confidence bands are represented by a grey band, while real (re-sampled) values are represented as a black, red or orange circles. Whenever real data (re-sampled) goes outside the pattern confidence bands (above percentile 95 or below percentile 5 of the pattern), an outlier is identified and the real value is color coded with a red circle. The user has also the option to navigate in the time history series to understand past behaviors or have access to future expected behavior (short-term forecasts), by moving the range-slider below the main graph.

Zone Graph

Pattern: the grey band on each chart represents the pattern confidence bands computed from one month of historical sensor readings within the zone. A pattern corresponds to an expected behavior for the zone. In OpenFlows WaterSight pattern bands are computed for each zone using advanced data analytics (statistic equations together with machine learning techniques to remove outliers or bad data). For these calculations several considerations such as the data history for each specific sensor as well as the distinction between week-day and weekend are taking into consideration.

A wide grey band indicates that there were large variations in flow during the time period for which the bands were developed. A narrow band indicates that the flow has been relatively constant at that time of day for the time period (hour and day of week) for which the bands were developed. If there is not enough historical data to compute the pattern confidence bands an informative message is displayed, and only real (re-sampled) values are shown. Clicking on this message removes the message.

Patterns can be customized by the user in the Administration page, under Settings > Pattern

Real time data: Real (re-sampled) values are represented as a black, red or orange circles. Those real (re-sampled) values are calculated based on the final balance between all inflows and outflows (plus accounting for storage inside the zone if exists). Whenever real data (re-sampled) goes outside the pattern confidence bands (above percentile 95 or below percentile 5 of the pattern), an outlier is identified and the value is color coded with a red circle.

Estimated data: If estimating data option is set to yes, every time one of the sensors fails or every time a meter failure event is registered at the sensor level, the software will automatically estimate the final zone flow. Estimated data will be represented in the graph with empty circles. Please note that the gap filling option only exists at the zone level, this means that in the sensor graph this problem (missing data or gaps) will always be noted so the user can be always aware of the real issues at the sensor level. Please also note that zone events/alerts are never generated based on estimated data. The option to use gap filling (estimating data) or not is done in the zone administration page. More information here - Zone Gap Filing: Estimating missing or wrong data.

Zone Name

Select from the list of available zones the one to create the chart for.

Interval

Specifies on the one hand the rate at which flow raw signal readings are re-sampled (15min, 1hour, daily) and on the other hand how the zone information is charted.

15 minutes: raw signal data is re-sampled in every 15 minutes interval.
1 hour: raw signal data is re-sampled in every 1 hour interval.
Daily minimum: The minimum value for each day (one value per day) is computed using statistical analysis (based on the 5th percentile of the time series). The daily minimum for the zones represents the minimum nightly flow.
Daily average: The average value for each day (one value per day) is computed using statistical analysis (based on the 50th percentile of the time series)
Daily maximum: The maximum value for each day (one value per day) is computed using statistical analysis (based on the 95th percentile of the time series)

All 15 minutes values and hourly values are evaluated looking backwards. For example if the value at 9h15 is 10 l/s, this represents the re-sampled value from 9h00 to 9h15. If the hourly value is 20 l/s at 10 AM, this represents the re-sampled value from 9 AM until 10 AM. All daily values are evaluated for the specific day that is being displayed. For example if the value for day 10th is 20 l/s, this value is calculated for the time period between 00h:00mm and 23h:59mm of day 10th.

Pattern

A pattern corresponds to an expected behavior for the zone. In OpenFlows WaterSight pattern confidence bands are computed for each sensor and for each zone using advanced statistic equations together with machine learning techniques to remove outliers or bad data. The pattern confidence bands are represented by a grey band.

The user has the option to define the historic time range to be used to compute these patterns: last month, last 2 months, last 3 months, last 6 months or last year of data. If there is not enough historical data to compute the pattern confidence bands an informative message is displayed, and only real (re-sampled) values are shown. Clicking on this message removes the message.

Method

Specifically for the zones the user can choose the type of pattern to visualize in the graph:

Percentile based approach - Statistical analysis is applied to the historical data in order to calculate the patterns
AI generated approach - Variables such as historical data, events generated, special periods (holidays or festive seasons), temperature and precipitation (weather) are used by the AI algorithm to calculate patterns that can respond better whenever there is a clear relation between zone demand and temperature or precipitation.

More information about Patterns and forecast calculation and methodology can be found here - Patterns and Demand Forecast Calculation.

Patterns can be customized by the user in the Administration page, under Settings > Pattern. For more information about customizing patterns, please click here - Patterns Customization.

Create Event

The user can directly create an event based on the graphs observations, by clicking in the "Create Event" button. Created events will be displayed in the graph and also automatically be saved in the Active Events list. Events created with category meter failure will be displayed in the graphs with a red cross icon while all other event categories will be displayed with a purple symbol. To define the right event duration make sure before clicking in the create event button, to maximize the main graph to the catch all the event period.

Editing the event

The event can also be edited after creation, such as the duration and other fields, by directly clicking in the event symbol, to access the event detail graph - workflow 1. In that case update the event by using the side panel, image below.

Another workflow is to go the active Events list and search for the specific event and manage from there. Please note that if the event was created with the status closed, you will need to specifically filter for the closed events in the active Event list.

Adding manual events like meter failure events might be important to have reliable water balance calculations and also a correct gap filling for the zones. More information here - Zone Gap Filing: Estimating missing or wrong data and Water Balance Calculation.

Compare

By clicking in the compare button the user will be redirected to the comparison page where he will be able to simultaneously compare in the same graph any user defined variable. By default the comparison page will open with all the inflow and outflow sensors as well as the zone flow enabled.

Graphic Tools

Use the graphic toolbar to control the display of the charts shown on the sensor card, with the possibility to add additional information:

Entry	Description
	Turn on/off: - Legend of the graphs - Signal data - Events highlighted in the graphs (with the additional flexibility to display only active events or all events) - Statistics (or patterns) - Current time marker
	Range slider - Turn on or off the time range slider below the main graph Full y-Range - Y axis will adjust automatically to include the maximum value registered Time range - Define the time period included in the range slider below (last 3 months, last 6 months, last year, last 2 years). User will be able to navigate in the time series history, for the period defined here.
	Download data that is being displayed on graphs (PNG or CSV format)
	More graph configurations: spike lines, data tooltip and font size.

Selected Zone

Statistic

Field	Description
Minimum (p05)	Percentile 5 of the last 3 months period, for the integration type time-series defined.
Median (p50)	Percentile 50 of the last 3 months period, for the integration type time-series defined.
Maximum (p95)	Percentile 95 of the last 3 months period, for the integration type time-series defined.
Minimum/Median	Ratio of "Minimum (p05)"and "Median (p50)" of the last 3 months period, for the integration type time-series defined. High ratio means that the minimum flow is closer to the average flow, and therefore there is a higher probability of the zone having high levels of real losses.

Inflows, Outflows, Storage

List of all inflows, outflows and tank level names that are used to compute the zone flow balance. Clicking directly in one of the sensors there will forward the user to the sensor details page.

WaterSight - Zone Details