Quickstart: Explore the Azure Time Series Insights Preview demo environment
This quickstart gets you started with the Azure Time Series Insights Preview environment. In the free demo, you tour key features that have been added to Time Series Insights Preview.
The Time Series Insights Preview demo environment contains a scenario company, Contoso, that operates two wind turbine farms. Each farm has 10 turbines. Each turbine has 20 sensors that report data every minute to Azure IoT Hub. The sensors gather information about weather conditions, blade pitch, and yaw position. Information about generator performance, gearbox behavior, and safety monitors also is recorded.
In this quickstart, you learn how to use Time Series Insights to find actionable insights in Contoso data. You also conduct a short root cause analysis to better predict critical failures and to perform maintenance.
Explore the Time Series Insights explorer in a demo environment
The Time Series Insights Preview explorer demonstrates historical data and root causes analysis. To get started:
Create a free Azure account if you don't have one.
Go to the Contoso Wind Farm demo environment.
If you're prompted, sign in to the Time Series Insights explorer by using your Azure account credentials.
Work with historical data
In Contoso Plant 1, look at wind turbine W7.
Recently, Contoso found a fire in wind turbine W7. Opinions vary about what caused the fire. In Time Series Insights, we can see that the fire alert sensor was activated during the fire.
Review other events around the time of the fire to understand what occurred. Oil pressure and active warnings spiked just before the fire.
The oil pressure and active warning sensors spiked right before the fire. Expand the displayed time series to see other signs that were evident leading up to the fire. Both sensors fluctuated consistently over time. The fluctuations indicate a persistent and worrisome pattern.
Examining two years of historical data reveals another fire event that had the same sensor fluctuations.
Using Time Series Insights and our sensor telemetry, we've discovered a long-term and problematic trend hidden in the historical data. With these new insights, we can:
- Explain what actually occurred.
- Correct the problem.
- Put superior alert notification systems into place.
Root cause analysis
Some scenarios require sophisticated analysis to uncover subtle clues in data. Select the windmill W6 on date 6/25.
The warning indicates an issue with the voltage being output by the generator. The overall power output of the generator is operating within normal parameters in the current interval. By increasing our interval, another pattern emerges. A definite drop-off is evident.
By expanding the time range, we can determine whether the issue has stopped or whether it continues.
Other sensor data points can be added to provide greater context. The more sensors we view, the fuller our understanding of the problem is. Let’s drop a marker to see the actual values.
Select Generator System, and then select three sensors: GridVoltagePhase1, GridVoltagePhase2, and GridVoltagePhase3.
Drop a marker on the last data point in the visible area.
Two of the voltage sensors are operating comparably and within normal parameters. It looks like the GridVoltagePhase3 sensor is the culprit.
With highly contextual data added, the phase 3 drop-off appears even more to be the problem. Now, we have a good lead on the cause of the warning. We’re ready to refer the issue to our maintenance team.
You're ready to create your own Time Series Insights Preview environment. To start:
Learn to navigate the demo and its features: