Complexity reduction using AI-based correlation analysis using the example of operator actions in a thermal waste treatment plant

Abstract: This Paper examines the use of artificial intelligence (AI) to analyse and optimise operating messages in a waste-to-energy plant. The aim of this thesis is to identify potential for improvement in automation technology. For this purpose, the 13,9 TB of data collected over the years from the message archive of the Wuppertal waste-to-plant are analysed, in particular the operating messages. The Siemens® Simatic PCS7 Process control system, which has been in use since 2016, is used as the data source.The lack of documentation on the database structure of the Siemens® reporting archive is addressed by re-documenting the structure of MSSQL database files (.mdf and .ldf). This provides a detailed insight into the data organisation and structure of the reporting archive for the first time. An MSSQL server is set up on a Linux system to import these files and make them accessible. With MATLAB as a client and using the Statistics and Machine Learning Toolbox, AI algorithms are used to analyse correlations between the operator messages. The aim is to recognise patterns and connections that indicate inefficient processes or unnecessary messages. The results of this analysis will be used to optimise the automation technology. This will reduce the workload for operators at the control centre and at the same time increase the efficiency and safety of power plant operation.The combination of back-documented data structures, AI-based data analysis and application to real operating data makes an important contribution to improving operational management in waste-to-energy plants. It shows how modern technologies and methods can be used to effectively address the challenges of thermal waste treatment and optimise operational processes.The development and structure of the paper is based on the V-model, which is commonly used in software development projects. The V-model was chosen for the structure of this paper because it offers a clear and systematic approach to the development and testing of software development projects. The development process starts at the top left, then goes down, then up again and finally ends at the top right. The descending branch contains constructive activities, while the ascending branch contains work steps for quality assurance. It begins with an “introduction”, which sets out the background, motivation and objectives of the research, followed by the “problem statement”, which highlights the current challenges and the current situation. The “target situation” defines the desired end state and derives requirements from this. These are further specified in the “Requirements” chapter. The “Theoretical foundations” provides the necessary scientific context, while the “Methodology” describes the research and analysis methods used. “Data analysis” and ‘Optimization proposals’ present the investigations carried out and the resulting suggestions for improvement. The “Discussion” reflects on the results and compares them with existing approaches. The “Conclusion” summarizes the findings and provides an outlook on future research opportunities.

Keywords: AI, Production, Process Management

DOI: 10.54941/ahfe1006446

Cite this paper:

​