NOTE: To download this document, please scroll down until the bottom of the page.
The file will be attached at the bottom.
PROBLEM
During normal operation, a device may report a consistent tank level over time. This value may be accurate if there is genuinely no-usage, but it may also indicate a potential malfunction. For example, certain device and sensor types report a consistent value of 50% when the sensor is unplugged from the device or dial. Because Unify does not currently flag this condition, users have no reliable way to distinguish potential sensor malfunctions from genuine no-usage, leading to delayed detection and increased run-out risk.
SOLUTION - HOW ARE SENSOR ISSUES DETECTED?
The sensor issue detection model looks at the last 7 days of level data and flags changes of less or equal than 0.75%. This model does not account for seasonality, unlike the Stuck Gauge model. The following use cases are monitored by sensor type:
Level constant at 0% - Any sensor
Level constant at 48% - Rochester R3D Vertical variants
Level constant at 50% - Rochester R3D Horizontal variants
Level constant at 62% - Magnetel variants
At a high level, the algorithm to identify possible sensor issues works as follows:
OUTCOME
Possible sensor malfunctions are now detected earlier and more consistently by monitoring sustained, low variance tank level readings across multiple sensor types. This reduces ambiguity between true no-usage scenarios and device or sensor issues, enabling teams to proactively identify and address data quality problems and improving overall confidence in reported tank levels.
WORKFLOW - CASE CREATION
When a Possible Sensor Issue is flagged, a new Maintenance Case is created. The case is assigned a Medium priority and labeled with the case type 'Possible Sensor Issue'. The priority is increased to High if the case remains open for five days and the identified trend persists. The case is automatically closed once the issue is resolved or when a subsequent reading reports a normal value.
REPORTS
Built in reports list all open cases for a tenant. Users can also create custom reports filtered to Possible Sensor Issues only. Reports can be downloaded or scheduled for distribution.
TANK LEVEL ICON
When a tank is allocated to a device with an open Possible Sensor Issue, the tank level icon is displayed with an "!" indicator. This visual cue conveys that the tank telemetry may require further review.
DEVICE DETAILS
When a device has an open Possible Sensor Issue, the tank level chart displays subsequent readings with a yellow triangle indicator. This highlights that the reported reading may be inaccurate or impacted by a sensor issue.
ANOVA GO
This new case type is also available in Anova Go and functions in the same way as existing case types within the application. When a tank is allocated to a device with an open Possible Sensor Issue, the tank level icon is displayed with an "!" indicator. This visual cue conveys that the tank telemetry may require further review.
HOW TO ENABLE
Please reach out to Customer Success Team:
- Contact support@anova.com
NEED FURTHER INFORMATION OR HAVE FEEDBACK:
Please reach out to Customer Success Team to discuss at: support@anova.com
FAQ:
How does a Possible Sensor Issue differ from a Stuck Gauge?
The Stuck Gauge detection model flags any consistently reported value. The Possible Sensor Issue model specifically detects recurring values of 0%, 48%, 50%, and 62%.
The Possible Sensor Issue model runs before Stuck Gauge. If it flags a device with a Possible Sensor Issue, that device is excluded from the Stuck Gauge analysis.
Read more about the Stuck Gauge topic in our Knowledge Base Article: Stuck Gauge Detection with Maintenance Cases Creation
EXAMPLES OF POSSIBLE SENSOR ISSUES IN UNIFY: (*SAMPLE DATA - IN TEST ENVIRONMENT*)
EXAMPLE 1:
EXAMPLE 2:
EXAMPLE 3: A device that might have been flagged as Stuck Gauge is now flagged as a Sensor Issue, based on the sensor type and reported value
Comments
0 comments
Please sign in to leave a comment.