Projects

Neural Networks and T48 Profiles

EQUIPMENT
PRODUCT

SISO collected T48 probe data from an aeroderivative engine across multiple runs to develop an anomaly detection algorithm. The process involved data collection, cleaning, and training a neural network. The trained model was then validated against a separate dataset containing both normal and faulty conditions to assess error detection performance. If the algorithm met accuracy requirements, it was deemed suitable for future monitoring; otherwise, it was retrained with adjusted inputs.