Data Acquisition: Smart Factory, Automation, Digital Twin & Traceability

The integration of data acquisition systems into networked production lines requires 3 technical properties: deterministic latency, structured output formats, and configurable data interfaces. AT Sensors 3D laser profile sensors deliver measurement results in less than 1 ms per scan line, making them compatible with high-speed production lines running at conveyor speeds of up to 3 m/s. Infrared cameras provide calibrated temperature matrices at frame rates of 25 Hz to 100 Hz, enabling real-time thermal monitoring of heat-sensitive production processes such as welding, casting, and battery module assembly.

Smart Factory data acquisition systems produce 2 types of measurement output: primary measurement data (geometric coordinates, temperature values, intensity images) and derived evaluation results (defect classifications, dimensional deviations, temperature gradients). Both output types are available via digital interfaces, enabling automated control loops in which the measurement result directly triggers a process action — for example, a reject signal for a non-conforming part.

Industrial data acquisition generates high data volumes. A single AT Sensors 3D laser profile sensor produces up to 4,096 measured points per scan line at scan rates of 4 kHz, resulting in a data throughput of 16 million measured points per second. Infrared cameras with 640 × 512 pixel resolution at 50 Hz deliver 16.4 million temperature values per second. Smart Factory architectures process this data using 2 approaches: edge processing directly at the sensor (on-sensor processing) and cloud-based processing for statistical evaluation and long-term trend analysis.

Inline data acquisition integrates the measurement system directly into the production line. The part moves through the measurement station during normal production flow, and the sensor acquires all necessary data within one machine cycle. Offline measurement, by contrast, removes the part from the production flow for measurement in a separate station. Inline 100% inspection eliminates the time and handling cost of offline measurement and provides real-time feedback on production quality.

AT Sensors 3D sensors support 2 inline measurement configurations: static measurement (part stops briefly under the sensor) and dynamic measurement (part moves continuously past the sensor). In dynamic configurations, the sensor synchronizes data acquisition with the conveyor speed via an encoder signal, ensuring that each scan line corresponds to a defined physical distance on the part surface. Infrared cameras in inline configurations capture the full thermal image of a part or weld seam within a single frame, enabling temperature-based quality decisions at 100% throughput.

100% inspection at full production speed generates large data volumes that require structured storage and evaluation architectures. A production line running at 60 parts per minute, each inspected by a 3D sensor producing a 2-million-point cloud, generates 120 million measured points per minute. Efficient data acquisition systems address this volume through 3 strategies: on-sensor preprocessing (reducing raw data to relevant features), selective storage (storing only non-conforming parts or statistical summaries), and real-time streaming to edge evaluation systems. AT Sensors sensors provide configurable output modes that allow engineers to balance data completeness against system bandwidth.

Infrared cameras provide thermal measurement data that the digital twin uses to model heat behavior in production processes. A thermal digital twin of a welding process captures the temperature distribution at the weld pool, the heat-affected zone, and the surrounding base material for every single weld produced. AT Sensors infrared cameras measure temperature values in the range from −20 °C to 1,500 °C with a measurement uncertainty of less than 2 °C, providing the precision needed for reliable thermal twin models in applications such as automotive body welding, battery module manufacturing, and semiconductor processing.

The digital twin uses thermal measurement data for 2 primary functions: real-time process control (adjusting process parameters based on measured temperature deviations) and predictive analysis (identifying patterns that precede process failures). Both functions require data acquisition systems with low latency. AT Sensors infrared cameras deliver calibrated temperature data at 50 Hz with a latency of less than 20 ms, enabling closed-loop thermal control in real-time production environments.

The fidelity of a digital twin depends directly on the measurement frequency and the spatial resolution of the data acquisition system. A digital twin updated with every part measurement at full sensor resolution mirrors the physical production process with high accuracy. A twin updated only with statistical summaries at hourly intervals reflects the production average but misses individual part variations. AT Sensors sensors support both operating modes, allowing engineers to configure the update frequency and data resolution according to the requirements of each digital twin application.

Effective traceability requires linking the measurement data of each part to a unique part identifier. AT Sensors data acquisition systems support 3 identification methods: external trigger signals with serial number input from the production control system, integrated barcode or DataMatrix code reading (via connected vision system), and manual part identifier entry via the sensor API. When the part identifier is linked to the measurement data at the time of acquisition, the complete measurement history of every part is retrievable from the quality database using the part number alone.

Product liability law in the automotive sector (IATF 16949) requires retention of quality records for a minimum of 15 years after the end of production. Data acquisition systems generate large data volumes over this period. A production line inspecting 1 million parts per year, each with a 3D measurement dataset of 5 MB, accumulates 5 TB of measurement data annually. Effective traceability architectures address long-term retention through 3 strategies: data compression (reducing raw point clouds to deviation maps and tolerance results), hierarchical storage (hot storage for recent data, cold storage for archived data), and cryptographic data integrity verification (ensuring that stored records have not been altered). AT Sensors measurement data includes hash-based integrity signatures that allow auditors to verify the authenticity of stored measurement records.

Measurement traceability in the metrological sense requires that each measurement result is linked to a national or international measurement standard through an unbroken chain of calibrations. AT Sensors sensors are calibrated against traceable reference standards, and the calibration certificates include the measurement uncertainty values that define the reliability of each traceable measurement result. Calibration documentation and recalibration intervals are part of the measurement record, ensuring that traceability extends not only to the measured part but also to the measurement instrument itself.