Feature Highlight: Transform Sensor Readings into Derived Metrics and Meaningful States
Collecting sensor data gives you visibility into your system. But making that data useful often involves another step: putting measurements on a consistent scale, calculating relationships between readings, and interpreting values in terms of operating conditions.
As your monitoring setup grows, these calculations become part of everyday work. Consistent units make readings easier to compare. Derived metrics reveal relationships between measurements. Clear status labels help people recognize what needs attention without interpreting every number themselves.
With Machinechat JEDI’s transforms, you can explore what your readings can tell you. Turn existing measurements into useful new metrics and give raw values meaningful labels—then put the results to work in your dashboards and monitoring rules.
Give Your Sensor Data More to Say with Transforms
A measurement doesn’t have to be the end of the story. You can use it as the starting point for another calculation, a comparison, or an operating status. JEDI gives you two tools for doing this. Compute transforms evaluate expressions to calculate values. Translate transforms assign a value when a condition is true.
Derive New Metrics with Compute Transforms
Compute transforms let you build on your measurements. A little arithmetic can give you another way to understand what’s happening.
New Units, Familiar Meaning: Bring measurements into the units your team uses. Convert temperatures or scale electrical readings into physical quantities such as pressure and level.
Different Sensors, Common Ground: Make comparisons easier by expressing readings on a consistent basis. Standardize units across devices, apply a known offset correction, or express values as a percentage of their measurement range. Percentages are particularly useful when equipment differs. You can compare how full two tanks are without first working out what their different liquid heights mean.
Another Measurement, Without Another Sensor: What could you learn by combining two readings? Calculate pressure drop across a filter or temperature difference across a cooling circuit. These derived metrics give you additional quantities to monitor using sensors you already have. You might also hear them called virtual sensors.
Put Performance into Perspective: Give a production count some context by comparing it with a target. Calculate a reject percentage to see the proportion of output that needs attention. Use inputs covering the same period or batch, and account for any denominator that could be zero.
Spot What Changed: Sometimes the interesting part is how much a reading moved. Calculate the difference between consecutive values and look for sudden shifts or counter increments.
Map Conditions to Meaning with Translate Transforms
With Translate, you can make your dashboard easier to read at a glance. Define a condition and choose the value to assign when it matches.
Give Machine Codes a Human Voice: Replace the mental lookup between a numeric equipment code and its meaning. Map documented codes to labels such as “Stopped,” “Running,” and “Fault,” so readers can recognize the operating state directly.
Turn Ranges into Readable States: Give measurements useful categories. Define what “Low,” “Normal,” and “High” mean for your application, then display the resulting status alongside the value. Display the number alongside the category when precision matters; let the label help readers interpret it.
Bring Conditions Together: Combine comparisons to describe a more specific situation. High temperature alongside low pressure may deserve a different label from high temperature alone. Use these combinations to express Boolean conditions that matter to your equipment and process.
Example: Put Scaling and State Mapping to Work with Tank Monitoring
For this example, take a cylindrical tank with a usable liquid height of 200 cm. Its sensor reports liquid height measured from the bottom.
Start with the Measurement: Suppose your device Tank reports a metric named Level_cm, currently reading 68 cm.
Compute How Full It Is: Create a Compute transform with the expression metric('Tank', 'Level_cm') / 200 * 100. Set the output device to Tank and name the output metric Fill_Percent.
Translate the Level into a Status: Give the level a label that someone can interpret at a glance. Create three Translate entries using the conditions below. For each entry, select String as the output type and write to the metric Level_Status on device Tank.
Condition | Assigned Value |
|---|---|
metric('Tank', 'Level_cm') < 40 | Low |
metric('Tank', 'Level_cm') >= 40 && metric('Tank', 'Level_cm') < 160 | Normal |
metric('Tank', 'Level_cm') >= 160 | High |
Manage Compute and Translate Transforms
One reading now tells you the measured height, the proportion filled, and its status against your chosen limits.
What Could Your Next Transform Reveal?
Use Cases
Watch Relationships Across Equipment: Compare inlet and outlet measurements to calculate pressure or temperature differences. Observe the calculated value alongside the original readings as operating conditions change.
Make Environmental Readings Easier to Compare: Bring sensors into common units, then classify readings into bands suited to the space you monitor. Give readers both the measurement and a quick way to interpret it.
Explore Sudden Changes: Use differences between consecutive readings to highlight shifts worth investigating.
Notifications
Take the experiment beyond the dashboard by using the resulting metrics in JEDI rules. Automatically send a notification when the tank’s level status becomes Low by configuring a condition in JEDI rules that triggers an email or SMS notification action.
Start Turning Readings into Answers
Transforms close the gap between what your sensors measure and what your team needs to know. Compute transforms give you derived metrics in engineering units, and Translate transforms give you states people can read at a glance.
Ready to try it? Download Machinechat JEDI and create your first transform.
Comments