Fault diagnosis of tin oxide gas sensor using energy barrier and art-2 neural network

We propose a method of fault diagnosis for tin oxide gas sensors using energy barriers at the contacting surfaces of the particles of tin oxide film and ART-2 NN (adaptive resonance theory 2 neural network) with uneven vigilance parameters. We diagnosed tin oxide gas sensors upon exposure to oil vapor, silicon vapor, and high humidity. The sensor feature for diagnosis was an energy barrier between particles extracted by temperature-simulated conductance measurement. The feature was manipulated by an ART-2 neural network and the performance was finally evaluated with real n-C₄H₁₀ gas. This method proves to be helpful to diagnose a fault that was typically generated by oil vapor, silicon vapor, and high humidity.