Hydrogen leakage pressure detecting method for hydrogen rail vehicles

Hydrogen safety monitoring is a crucial element in ensuring the safe operation of hydrogen-powered rail vehicles. To address the limitations of existing hydrogen leakage pressure detection methods in terms of real-time performance and accuracy, this paper proposes an intelligent detection method that combines laser visualization technology with machine learning. By establishing a quantitative relationship between hydrogen leakage pressure and deformation in the laser beam profile, geometric features were extracted from laser spots using image processing techniques. These parameters were then used to construct a prediction model based on neural networks. Experimental results demonstrate the excellent predictive performance of the optimized neural network model (with a 2-16-1 topology), achieving a coefficient of determination (<italic>R</italic>²) of 0.988 and a root mean square error (RMSE) of 0.004 6. This method enables non-contact, high-precision detection of hydrogen leakage pressure and offers a novel intelligent monitoring solution for hydrogen-powered rail vehicles, holding significant engineering application value in enhancing hydrogen safety in rail transit.