
AI-supported shop floor management

Objective
Extension of a classic shop floor management to include "smart error management" with AI algorithms and "cognitive analysis"
Initial situation

Lack of transparency in the processing of internal errors on the shop floor
Lack of comprehensive error knowledge base in conventional shop floor management
Need for action on data analysis from error databases
Lack of technology infrastructure and limited practicality of modern data mining methods
High latency in error detection

Solution strategy
Sustainable error prevention through transparent real-time analysis of product and process error events
Building an error knowledge base to store acquired cognitive skills in the systematic problem-solving process
Development of an intelligent AI system for smart problem solving that supports employees in troubleshooting
Improving the intelligent AI system with extensive methods and tools of quality science and data mining

Building an error knowledge base
1. Determination of knowledge sources and methods
- Identification and analysis of error-relevant information and problem-solving methods
- Development of an intelligent AI system for smart problem solving that supports employees in troubleshooting
2. Expert regulation for knowledge generation
Computer-aided extraction, classification, systematic implementation and interpretation of problem-solving methods from the error-relevant data
3. Building a knowledge base & applying smart methods
Smart collection and analysis of historical data from the pilot manufacturing companies to build the knowledge base

Continuous improvement
4. Analysis and prediction of error patterns
Processing the knowledge base to detect unknown error relationships, determine and define error patterns and predict future error events
5. Derivation of error causes / measures
Derivation of error causes and measures from the error knowledge base using inference models
6. Design and verification of a functional prototype
Verification of the algorithms developed so far by building a functional model
7. Cognitive modeling
Development of cognitive algorithms for unknown problems

Advantages
Supported problem-solving processes on the shop floor
Transparency in troubleshooting
Cost and resource savings
Positive error management culture in the company
Increased competitiveness
In collaboration with:
Technical University of Berlin
Department of Quality Science
