Automatic conversion of explorative algorithms into edge-apps for machine tools
The advancements in technology have led to the collection and storage of substantial amounts of data in real-time monitoring of machining processes. However, the intelligent analysis of this data, particularly through standardized solutions that would empower process experts, has received little attention. Production-grade edge apps in manufacturing tools companies like Siemens are created following a two-phase procedure: First, process experts create explorative algorithms that operate on static data using platforms such as Jupyter notebooks. Then, these explorative algorithms need to be converted into factory-floor apps that are executable on edge platforms such as the Siemens Analyze MyWorkpiece/Monitor (AMW-M). This manual procedure is time-consuming and error-prone, especially for inexperienced workers, as it involves complex configuration of parameters.
The objective of this thesis is to automate this conversion of explorative algorithms into edge-apps and demonstrate the solution in the context of Siemens edge infrastructure. The following challenges need to be addressed:
the resulting apps should be able to handle live streaming data efficiently, provide event notification mechanisms, support configuration and deployment, and inform users about missing input data without causing the app to malfunction. The solution created in this thesis needs to also focus on improving process experts' convenience.
The objective of this thesis is the creation of a solution or the combination of existing techniques to solve the problem of automatically converting explorative algorithms for machine monitoring into factory-floor apps executable on edge platforms through a suitable conversion technique and supporting software tool as described above. This comprises the analysis of the state of the art of software conversion/transformation techniques as well as the demonstration of the solution by prototypical implementation in the context of a real-world software environment by Siemens and a suitable experimental evaluation in a pilot study as outlined above.