The purpose of this study was to investigate how to detect abnormalities in electric submersible pumps (ESPs) in advance and how to classify the faults by monitoring the production data before pumps break down. Additionally. a new method based on the denoising autoencoder (DAE) and support vector machine (SVM) is proposed. Firstly. https://www.drcristinosuarez.com/