WaterSight - Active Events

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

Overview

The Active Events page lists all events that are automatically generated by the software (based on some user defined alerts configuration in the administration) or manually added by the user. These events should represent anomalies that are happening or that occurred in the network, and should be managed by the user. 

The generated alerts can be triggered using the sensor pattern as reference, or can be based on absolute sensor values defined by the user. All trigger configurations are defined by the user at the alerts administration page. For some tips how to configure bursts alerts, also see this article

Adding Manual Events, Editing and Deleting

It is also possible to manually add an event, by clicking  in the "Add" button just above the alerts list. This is particular useful to track events that are happening or occurred in the network, but were not automatically generated by the software. The "Edit" and "Remove"  buttons located above the table also allow the user to edit existent alerts as well as delete a specific alert (in this case only administrators have permission to delete alerts). To edit or delete a specific alert, select first the row that contains the alert, and then click on the edit or delete button. 

Events Filter Options

Status

Click on the symbol   located on the right of the Status to see all filter options.

By default the list of alerts is showing all alerts that are active and that require user attention. This includes the alerts with the following status: New, Under Investigation, Field Work and Open (all status except Closed). Once an alert is closed, it will be hidden from the default view list, and the user will need to check the “Closed” status from the status table available and located on the right side. The user can also search for specific alerts in the list based on one or more selected status.

Category

Click on the symbol    located on the right of the Category to see all filter options.

The user can filter the alerts that are being displayed in the list based on the selected category: Unknown, Meter Failure, Burst, Leak, Flow Increase, Maintenance Work, Zone Change or show all categories.

Search

The search field allows the user to search for a specific alert or groups of alerts based on their name or based on the name of the sensor from which the alerts were generated.

CSV Export

The user can download the list of alerts and all the respective details to CSV.

Events Details

Name

Name of the alert. If the alert was automatically generated by the software, this field is blocked for editing, and the default name will follow the rule “Alert + ID”. If the alert was manually added by the user, the default name will be “User-defined”, but the user can change it.

Origin

Sensor or zone which the time series that generated the alert belongs to. Click in the sensor or zone name to quickly see the generated alert in the time series graph. For more information about the difference betwee triggering alerts for zones vs sensors, please click here

Status

Status of the event. The user should update status according with the respective events workflow: 

  1. New: by default all events are generated with this status;
  2. Open: change to this status to acknowledge this alert. Also a preliminary analysis should already have been made to filter possible false positives;
  3. Under Investigation: change to this status if it is required further investigation, before taking any action;
  4. Field work: change to this status if the user is confident that the alert can represent any anomaly in the network and there are already some people in the field looking for the anomaly;
  5. Closed: the event should only be close whenever the conditions that triggered the alarm went back to normal, and when a cause was identified. If it was not possible to identify a cause, please close the event and categorize it as unknow. If the alert is also considered a false positive, the user can directly move from step 1 to 5.

The user has the option to filter the events that appear on the list. By default the list is always showing all alerts/events that are in progress (this means all alerts that are not closed).

Category

Whenever an alert is generated, the category is always set to unknown, expect if the alert type (configured in the Alerts Administration page) was set to "No Data" or "Flat Reading". In these two last cases, the default category of the event generated is Meter Failure. 

After the cause of the event was identified, the user must select the most appropriate field defined in WaterSigth: Unknown, Meter Failure, Burst, Leak, Flow Increase, Maintenance Work, Zone Change. Maintenance work includes all operational works developed in the network (flushing, opening/closing valves, etc). Zone Change refers to a temporary or permanent change of the zone layout (by opening/closing valves)

The user can filter the events that appear in the list by clicking in the category filter. 

Symptom

This easily identifies the reason why the alert was generated. The symptom is identified based on the user admin configurations defined in the alerts configuration page:

For alert types "Pattern" and "Absolute":

  • If an alert was triggered based on a flow time series, high: symptom will be flow increase, 
  • If an alert was triggered based on a flow time series, low: symptom will be flow decrease, 
  • If an alert was triggered based on a level time series, high: symptom will be level increase, 
  • If an alert was triggered based on a level time series, low: symptom will be level decrease, 
  • If an alert was triggered based on a pressure time series, high: symptom will be pressure increase, 
  • If an alert was triggered based on a pressure time series, low: symptom will be pressure decrease, 

For alert types "No Data" and "Flat Reading":

  • If an alert was triggered based on the type "No Data": symptom will be No Data,
  • If an alert was triggered based on the type "Flat Reading": symptom will be Flat Reading

Start Date

The date the event started (and not the date the alert was triggered). 

As illustrated in figure below, the alert is generated when the trigger configurations defined by the user in the alerts configuration page are met, however the real event can already have started some hours earlier. 

Figure 1. Difference between event start date and date the alert was triggered, for a pattern based alert. 

End Date

End date of the event, for which the alert was triggered.

For pattern based alerts (please see alerts configuration page for more detail) it is when the event returns to the pattern. For high type alerts this means below percentile 80th (values are color coded with a black circle). For low type alerts this means above percentile 20th (values are color coded with a black circle). 

For absolute based alerts (please see alerts configuration page for more detail) it is when real values are below (for high type) or above (for low type) the absolute numerical value defined by the user.

Figure 2. Start date and End Date for a pattern based alarm. 

Duration

The duration of the event is automatically calculated by the software using the Start Date and End Date (for automatically generated alerts). For manually added alerts, this field must be manually completed.

Figure 3. Duration of the event. 

Max.Value

Máx value registered by the sensor. This value is automatically calculated by the software for automatically generated alerts. For manually added alerts, this field must be manually completed. For pattern based alarms (please see alerts configuration page for more detail), this corresponds to the burst (or event) flow, and it is given by the maximum difference obtained between the sensor real value and the pattern. For absolute based alarms (please see alerts configuration page for more detail), this corresponds to the maximum absolute value registered by the sensor (within the period of the event). 

Figure 4. Máx value registered by the event, for a pattern based alert.  Figure 5. Máx value registered by the event, for a absolute based alert

Lost Volume

Volume lost in anomalies and other events (like bursts, leaks, flushing, fire usage, etc). This volume is automatically calculated by the software for automatically generated alerts, and only for alerts that were generated based on a flow time-series using pattern high as a criteria (please see alerts configuration page for more detail).

The volume is automatically calculated by integrating the difference between real flow and expected flow (the pattern) along time, for the all duration of the event. 

Figure 5. Identification of the lost volume.

Remove From Patterns

The user has the ability to control if an event should be considered for the pattern and forecasts calculations of the sensor or if should be ignored. The pattern is represented in the graph of each sensor by a grey band and corresponds to the expected behavior of the sensor. For more information about patterns, go for example to the Sensor Overview help page.

Generally all alerts that represent anomalous or extraordinary events (such as bursts, leaks, fires, tank overflows, temporary zone changes) should not be considered for the pattern calculations and forecasts of the respective sensor. In these cases the option "Remove From Patterns" should be checked  . Otherwise, if unchecked, the values related to the anomalous events that have occurred will be considered for the patterns calculations and forecasts, and with this "polluting" the patterns. 

On the other hand, in case an alert is generated and represents a new operational condition of the network, or something  that will tend to repeat in the future at similar hours, the option "Remove From Patterns" should not be checked  . This way WaterSight will acknowledge this event as something normal, and it will be automatically considered for the pattern and forecasts calculation, avoiding that in the future new alerts are generated due to the same and already identified reason (and avoiding generation of false positives). Some examples of this can include for example: i) permanent zone changes, ii) when entering in the summer period - where flows can increase significantly - or in the winter period - where flows can decrease significantly , iii) any event in the network that will tend to repeat in the future at similar hours.

By default all alerts that are automatically generated have the option "Remove From Patterns" disabled. The user should manually enable this option, whenever applied as explained above. Also to mention that when an event is categorized as meter failure, burst, leak or maintenance work the option "Remove From Pattern" is automatically enabled/Checked. 

Comments

The user can update the event with some relevant comments (for example the type of burst, location, and other relevant information for the utility).

"Cleaning" Patterns

One way to assure "cleaned" patterns is to set alerts for anomalies and meter failures, and make sure that once they are generated that data is removed from the pattern calculation (see explanation above). 

However, it is important to refer that the user does not need to wait for an alert to be generated in order to get "cleaner" patterns. The user can  manually "Add" alerts directly form the Active Events table (clicking the button   ), define the period in which something unusual as occurred in the network, and then selecting the option  "Remove From Patterns" from the events list.  By doing this patterns do not become "polluted" by anomalous events or data failures, even if no alerts are triggered/configured. 

Flat Reading event. Remove from Pattern option NOT checked.  Flat Reading event. Remove from Pattern option checked. 

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

Recommended
Related