Perform a root-cause analysis in Cognite Data Fusion

Use Cognite Data Fusion (CDF) to do root-cause analysis (RCA) and eliminate the need for traditional paper-based data collection.

All the relevant information is readily available in CDF, and you can access and monitor real-time data, and collaborate with your coworkers on a Canvas. You can also save your work and refer to it later.

Step 1: Collect data

Use search to find the resources you want to include in the root-cause analysis. The search results include all directly linked items that match your query. For instance, if you enter the prefix for a pump ID, you'll also find the associated assets, documents, sensor data, and engineering diagrams.

1. When you've found the data you're looking for, select Open in canvas and add the data to a new or existing canvas.

2. On the canvas, select + Add data to add any additional resources you need for your analysis. You can also drag and drop local PDF files and images directly onto a canvas.

3. Use visual elements, such as shapes, lines, and sticky notes from the toolbar to organize the canvas.

4. Select any item on the canvas to see options for changing labels and finding related data.

5. For assets and events, such as work orders, use Change fields to select which properties to display.

Step 2: Identify root causes

To collaborate with your coworkers, share it with individual users, or make it public to everyone using the CDF project.

sfat

Organize your RCA canvas as a storyboard to build referenceable investigations and analyses.

These tips can help you analyze and identify the root cause of an issue:

• View live sensor data by pinning a time series connected to the asset to the canvas. Then, add rules to set threshold values and use colors to visualize any breaches. You can also open the time series as a separate item on the canvas.

• Create cause map diagrams with the Cause map () tool, and add stamps, sticky notes, images or data to each card. Select a card to set a status, bring up options for formatting, and set color-coded thresholds for monitoring.

• For time series, select a time frame to navigate between current and historical data, or set a specific date. Select Apply to all charts and time series to use the time frame for all time series on the canvas.

• Open a time series in Charts to trend and analyze the data further, for example to see current or historical performance data and visualize and assess data quality. In Charts, you can also:

  • Trend data by using the Slider (🎚) tool to see current values and compare values at different points in time.

  • Add calculations to make predictions.

  • Add events, such as a faulty shutdown, from data historian systems, maintenance systems, or control systems.

Step 3: Lock and save the canvas

When you've completed the RCA, select Lock canvas to prevent it from being changed. You can also select ⋯ > Download as PDF to save the Canvas or specific parts of the canvas as a PDF file.

Step 4: Monitor asset health

When you're done with the RCA, you can keep monitoring the issue to see if you really resolved the issue. Monitor health indicators and key performance indicators (KPIs) by setting up real-time monitoring of time series. The monitoring checks for updates every 5 minutes and sends you emails for every threshold breach.

Create thresholds for items on a canvas or in engineering diagrams by setting up color-coded rules. For instance, you can keep track of pump values, corrosion rates, or operating temperatures. Canvas checks for updates every 30 seconds and changes the color of an item when a set threshold is breached.