Pipeline Leakage Detection via Extreme Seeking Entropy

In this article, we present a case study of pipeline leakage detection via an adaptive novelty detection algorithm. The detection is based on an affine projection filter applied to a sound recording of the pipeline outlet. Adaptive weight increments of the filter are evaluated by the Extreme Seeking Entropy algorithm to estimate the amount of novelty in data. The suitability of the proposed approach is presented in three case studies where the leakage was simulated in the degassing process model by increasing the discharge flow.