Fusing convolutional generative adversarial encoders for 3D printer fault detection with only normal condition signals Chuan Li a,1 , Diego Cabrera a,b,⇑,1 , Fernando Sancho c , René-Vinicio Sánchez b , Mariela Cerrada b , Jianyu Long d , José Valente de Oliveira e a National Research Base of Intelligent Manufacturing Service, Chongqing Technology and Business University, Chongqing 400067, China b GIDTEC, Universidad Politécnica Salesiana, Ecuador c Dpt. of Computer Science and Artificial Intelligence, Universidad de Sevilla, Spain d School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou, China e Universidade do Algarve, Portugal Keywords: Fault detection 3D printers Condition-based maintenance Convolutional neural networks Adversarial learning abstract Collecting data from mechanical systems in abnormal conditions is expensive and time consuming. Consequently, fault detection approaches based on classical supervised learn- ing working with both normal and abnormal data are not applicable in some condition- based maintenance tasks. To address this problem, this paper proposes Fusing Convolutional Generative Adversarial Encoders (fCGAE) method to create fault detection models from only normal data. Firstly, to obtain an adequate deep feature space, encoder models based on 1D convolutional neural networks are created. Then, these encoders are optimized in an unsupervised way through Bidirectional Generative Adversarial Networks. Finally, the multi-channel features collected from the system are merged with One-Class Support Vector Machine. fCGAE is applied to fault detection in 3D printers, where experimental results in two fault detection cases show excellent generalization capabilities and better performance compared to peer methods. 1. Introduction 3D printing has been attracting much attention as an innovation technology in the industry [1], as well as in medical [2], food [3], and other areas [4]. This is due to the accuracy level achieved by this technology [5], which improves the quality of the obtained products. Of course, the quality that a 3D printer offers depends mainly on the transmission elements that con- stitute it [6]. Despite the fact that the quality of the components of a 3D printer is the best or not, these elements suffer from wear and tear due to continuous use in the printing process, as in any other machinery [7]. Additionally, the union of the components can be compromised by the loosening of the fastening elements such as nuts, bolts, and join bearings. Those wear and/or loosening cause unintended movements in the printer head along with unwanted vibrations of the machine [8], resulting into loss of print quality, an accelerated deterioration of other components of the printer [9], and possible risk of damage ⇑ Corresponding author at: National Research Base of Intelligent Manufacturing Service, Chongqing Technology and Business University, Chongqing 400067, China. E-mail address: dcabrera@ups.edu.ec (D. Cabrera). 1 The two authors contributed equally to this paper.