About the project

In a deep drilling process, sufficiently assessing the stability of the process is no easy task. This project examined the possibility of determining the stability using data from accelerometers on the machine and AI models and mathematical methods. The aim was to optimize the drilling process so as to reduce scrap and material wear.

The research found that autoencoders, a special form of artificial neural network, can learn and reconstruct the drill speed. In the event of an anomaly, the reconstruction error increases, which can be used as an indicator of process stability.

In addition to artificial neural networks, the results obtained from applying mathematical methods, such as principal component analysis (PCA), also allow for the drilling process to be described and abstracted. Using these methods, further knowledge about the drilling process can be gained, which can then be used to optimize the process.