Process-Anomaly Detection

How advanced AI algorithms enable the discovery of previously unknown process anomalies.

Process anomalies represent a risk for manufacturing companies that should not be underestimated. However, their efficient identification and control is often difficult – especially in the chemical industry, where wide-ranging processes and constant change are part of daily business. Rule-based risk-detection methods are of limited help: When an algorithm applies knowledge learned from past data to new cases, it finds only those anomalies acquired from previous experience.

4

different types of anomaly can be identified using a single application

Solution

ONE LOGIC developed a centralized “Anomaly Detection App” divided into the components “Audit”, “Process Analysis”, “Generic Clustering”, “Text Mining” and “Time-Series Analysis”, which provides the solution for the customer. Based on advanced AI algorithms, the application identifies unknown anomalies almost in real time. “Audit”, for example, detects instances of non-compliance and mismatch, while “Process Analysis” alerts the user to high costs, process deviations and long lead times. “Generic Clustering” and text analysis uncover errors and anomalous behavior. “Time-Series Analysis” shows KPI fluctuations over multiple periods.

The application can be quickly connected to existing SAP systems via the Data Product Platform ONE DATA and works with standard tables. A conversion layer automatically converts data into standard functions and feeds them into the algorithm. This can be efficiently adapted through extensions and the scalable architecture and expanded to new application areas.

Result

Through advanced algorithms and central management of all use cases, the application enables continuous monitoring of risks – at scale and while significantly reducing manual effort. Automated pipeline and quality- assurance measures make monitoring, ad-hoc analysis, and auditing effortless.


Download ,Process-Anomaly Detection’ for free

Process anomalies are a high risk for production and often difficult to identify. ONE LOGIC developed an “Anomaly Detection App” for a manufacturing company. The app detects process deviations via AI algorithms in real time.

  • In this use case, you learn how the Data Product Platform ONE DATA can be easily connected to SAP systems.
  • We show you how the tool reduces manual effort and easily enables monitoring, ad-hoc analysis and auditing using automated processes.