 This paper proposes a novel approach to diagnosing faults in wind turbines using graph neural networks, GNN, and one-shot learning, OSL. The GNN is able to extract features from vibrations captured by a simulated wind turbine rig while OSL allows for accurate classification of these features despite having only a few examples available. This combination of techniques has been shown to be more effective than other methods such as siamese, matching, and prototypical networks.