Overview of Elevator Door Machine Fault Diagnosis Methods

As a critical component in elevator operation, the health status of the door operator system directly impacts elevator safety and operational efficiency. Currently, there is limited research literature on diagnosing faults or failures in elevator door operator systems, necessitating further studies on diagnosing common door operator malfunctions. This paper reviews the application of Extreme Learning Machines (ELM), Multiscale Convolutional Neural Networks (MSCNN), and Nonlinear Vector Autoregressive Convolutional Neural Networks (NVAR-CNN) in elevator door operator fault diagnosis. A multidimensional unified analysis framework is constructed to compare these three methods through a combination of quantitative and qualitative evaluations. Concurrently, this study addresses data-related challenges and model cross-scenario generalization issues in door operator fault diagnosis, proposing a strategy for combined method application. This provides a comprehensive and practical reference for subsequent elevator door operator fault diagnosis research.