WaterSight - Sensor Details

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

Overview

Sensor details shows a detailed chart for a specific sensor. By default the a 3 month signal history will be read.

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. This view is particularly useful to monitor network parameters and to easily spot any anomalies (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.

Note: The user can navigate to the sensor details from the Sensor Status or Sensor Overview pages, by clicking in the name of the sensor.

Sensor Name

The name of the sensor is displayed above the graph, in the upper left corner. The user can easily switch between different sensors by directly clicking on the down arrow   located on the right side of the sensor name.

Pattern

The pattern confidence bands are represented with a grey color. The Pattern drop down lets the user select the time period for which the confidence bands are based.

The user has the option to define the historic time range to be used to compute these pattern confidence bands: 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. Patterns can be customized by the user in the Administration page, under Settings > Pattern.

For more information about patterns or forecasts, please click here.

Interval

Specifies on the one hand the rate at which raw signal readings are re-sampled (15min, 1hour, daily) and on the other hand how the sensor 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).  
  • 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. 

Create Event

The user can directly create an event based on the graphs observations, by clicking in the "Create Event" button.  Created events will automatically be saved in the Active Events listplace where they should be managed. 

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 selected current measurement and sensor will be already displayed and populated in the Comparison page.

Adding Model Results and Raw Data

Click on the visibility button, under the graphic tools, to add model results and raw data in the graphs. More information below.

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

- Raw Data

- Model Results

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 Sensor

Field Description
Id Scada ID/Tag of the sensor
Type Kind of measurement
Status Status obtained from number of signal readings in the selected signal time range and the expected number of readings.
Latest Timestamp of most recent signal reading

Statistic

Field Description
Minimum (p05) Percentile 5 of the historical time period defined by the user (default is last 3 months), for the integration type time-series defined.  
Median (p50) Percentile 50 of the historical time period defined by the user (default is last 3 months), for the integration type time-series defined.
Maximum (p95) Percentile 95 of the of the historical time period defined by the user (default is last 3 months), for the integration type time-series defined.
Minimum/Median

Ratio of "Minimum (p05)"and "Median (p50)" of the of the historical time period defined by the user (default is last 3 months), for the integration type time-series defined.

Note: When a sensor is recovering from a long period without data, the first few weeks may have patterns that do not look intuitive. Factors to consider when interpreting patterns include:

  • Weekends and weekdays are calculated separately, so there may be regions where the weekdays have sufficient data to display normal patterns, while the adjacent weekends do not.
  • If no historical data is available in the specified pattern duration, the empty regions are displayed as a straight line connecting across the gap.
  • If only one weekday or only one weekend of historical data is available, all percentile values will be equivalent, so a pattern line will be visible but there will be no shaded region.

Figure 3 - Special case when very few historical data is available.

For more information about WaterSight, please go to OpenFlows WaterSight TechNotes and FAQ's.

Recommended
Related