Selected Use Case Scenarios

Process Mining in Automated Production Systems

Commissioning of automated production systems usually follows standardized, stability-oriented procedures. Hence, the resulting performance is limited and the potential to increase productivity often exceeds 10%. However, relevant factors (e. g. system parameters, environmental conditions) influencing the process are often unknown.

Problem

Relevant key figures for the quantitative assessment of system performance are not available.

Selection and prioritization of appropriate optimization measures require extensive time and expertise.

With increasing system complexity these tasks become even more challenging, even for experts.

Our Solution

Complexity reduction through centralization of system information, process characteristics as well as control and sensor data streams

Generic algorithms for automated training of hybrid process models based on discrete control events and continuous data

Analysis under holistic view on machines, products and processes reveals additional relevant information and system characteristics

Structured data orchestration providing a solid basis for targeted application of specific analytics services from our toolboxes or third parties

Benefits & Advantages

Gain deep insights into actual system behavior under interpretable visualization of discrete process flows and sensor data features through recurring process cycles

Simplified and reliable quantification of Key Performance Indicators (KPI) for profound evaluation of Process Integrity or Overall Equipment Effectiveness (OEE), Proofs of conformity for system and process certification

Comparative assessment of real-world system behavior based on the initial system specifications (re-engineering of the PLC code)

Detection of critical process steps, system bottlenecks and unstable sequences to prioritize optimization measure

Benchmarking of Systems and Processes

Production processes of identical or similar lines, cells or machines often run in parallel, at the same or at a different location. Alternatively, equal processes are operated using similar systems from different manufacturers.

Problem

Root causes for deviations and occurring discrepancies in performance are often unclear

The given problem complexity is not manageable by conventional approaches

Our Solution

Data driven training of system and process models for direct and quantifiable comparison

Comparison of statistical key figures and data features to identify causes for occurring deviations

Localization of relevant sub-process steps and defective technical components

Benefits & Advantages

Minimized effort and time to identify deviations and detect causalities

Incentives for plant operators through measurable performance indication and competition

Increased overall performance through focused modification of underperforming systems

Transparency about proven potentials, validated by best performing counterpart

Online Process Monitoring & Anomaly Detection

Maintenance measures take place preventively or after fixed intervals. Furthermore, critical deviations of process and Products quality are detected only with delay, resulting in unexpected system downtime.

Problem

The majority of critical states is not detectable by local monitoring of subsystems (plus additional sensors are expensive)

Even slight deviations reduce overall performance excessively but are usually not detectable even by experienced personnel

Often system are too complex to centralize essential information to combine individual state information of subsystems

Our Solution

Structurization and centralization of accessible process and sensor data via heterogeneous interfaces

Software services to implement performant data stream pipelines based on IIoT infrastructures

Implementation of digital process twins integrating machines, product and process characteristics

Anomaly detection in physical and statistical behaviour of machine components and process sequences

APIs for standardized access to process and machine data for additional analytics and monitoring services

Benefits & Advantages

Messaging system for detected deviations lead maintenance teams to critical components with extended information

Traceability of product individuals with evaluation of process integrity

High-Sensitivity and Real-Time Monitoring of relevant key performance indices and individual data features

Web-Interfaces for Visualization and direct Integration to MES/ERP

Reduce downtime by up to 50%