Variables & time series
What variables and time series are and how ENLYZE records machine data.
A variable is a single measurement value that changes over time. When a variable is recorded, a time series is created: a sequence of values with timestamps that documents the progression of the measurement.
What is a variable?
Every controller contains data points that reflect the current state of the machine. Examples:
Temperatures of individual heating zones
Pressures at inlet and outlet
Speeds and RPMs
Counter values (piece counters, meter counters)
Status messages and error codes
In the controller, these data points often have technical names like DB1.DBD24 or GDB.Static.rTemperatureZone1. These names are meaningful for the controller but difficult for humans to read.
From variable to time series
When you select a variable for recording, ENLYZE begins regularly querying its value. Every 10 seconds, the current value is read from the controller and stored together with a timestamp. This creates a time series point by point.
10:00:00
245.3 °C
10:00:10
245.1 °C
10:00:20
245.4 °C
10:00:30
245.2 °C
The recorded data is permanently stored in the cloud and is available both in near-real-time and historically.
The 10-second recording interval is a compromise between data volume and level of detail. For most production processes, this is sufficient.
How recording works
The edge device queries the controller every 10 seconds. Industrial controllers respond in the microsecond range. The edge device reads all configured variables within each 10-second window. The responses are collected and assigned to the window's start timestamp.
No averaging or aggregation takes place. Each stored value is a direct readout from the controller. If a value changes between two queries, the next recording shows the new value.
Multiple variables, one complete picture
In practice, not just one variable is recorded but many simultaneously. Typically, 10 to 150 variables per machine are recorded, depending on the use case.
Each variable becomes its own time series. Together, the time series form a picture of the machine state: the temperature remains stable, the throughput rises during startup, the heating power decreases when the machine reaches operating temperature. Patterns and correlations between variables become visible.
Variable status
Every variable in ENLYZE goes through three statuses:
Inactive
Discovered in the data source, but not being read.
Exploration
Being read and recorded, but not yet available to platform services. Used for evaluation.
Active
Being read and available to all services (dashboards, alerts, OEE, exports).
Contextualization
To make a variable understandable for others, it is contextualized upon activation:
Display name: A readable name (e.g. "Extruder throughput actual" instead of
DB1.DBD24)Unit: The physical unit (e.g. kg/h, °C, m/min)
Scaling factor: If the raw value needs to be converted
ENLYZE always stores variables unscaled (as raw values). Scaling is only applied at query time. This means that if a scaling factor was set incorrectly, you can correct it at any time without affecting the stored data. Historical values will then automatically be displayed correctly.
How to select and contextualize variables is described under Variable Mapping.
What variables tell you (and what they don't)
Time series show patterns and events: startup phases, stable operation, interruptions, speed changes. The data shows you the how of production.
What time series alone don't show: what was produced, for whom, and with what target values. For that, you need the production context.
Computed variables
Not all relevant KPIs can be read directly from raw data. Sometimes a raw value needs to be converted, sometimes multiple variables need to be combined.
Simple conversions
When contextualizing a variable, a scaling factor can be specified. This allows raw values to be converted directly, e.g. from an internal unit to °C or kg/h. This is sufficient for many use cases.
More complex calculations
When simple scaling is not enough, ENLYZE creates computed variables. Typical examples:
Rate calculation: Calculate the current production rate from a running counter (e.g. pieces per minute)
Sums and differences: Combine multiple variables (e.g. total throughput from individual components)
Outlier removal: Mark sensor fault values (e.g. 9999) as invalid
Status mapping: Convert codes to readable text (e.g. code 3 = "Technical breakdown")
Smoothing: Rate calculation over a time window for more stable values
Configuration is done by ENLYZE. You define which KPIs are needed and provide the necessary information about the raw data. ENLYZE implements and validates the results.
Computed variables depend on the recorded raw data. If no data is available for a period (e.g. because the edge device was offline), no computed values can be generated either.
Related topics
Production context: How order data gives meaning to machine data.
Variable Mapping: Selecting and contextualizing variables (guide).
Last updated