Abstract
This work considers industrial process monitoring using a variational autoencoder (VAE). As a powerful deep generative model, the variational autoencoder and its variants have become popular for process monitoring. However, its monitoring ability, especially its fault diagnosis ability, has not been well investigated. In this paper, the process modeling and monitoring capabilities of several VAE variants are comprehensively studied. First, fault detection schemes are defined in three distinct ways, considering latent, residual, and the combined domains. Afterwards, to conduct the fault diagnosis, we first define the deep contribution plot, and then a deep reconstruction-based contribution diagram is proposed for deep domains under the fault propagation mechanism. In a case study, the performance of the process monitoring capability of four deep VAE models, namely, the static VAE model, the dynamic VAE model, and the recurrent VAE models (LSTM-VAE and GRU-VAE), has been comparatively evaluated on the industrial benchmark Tennessee Eastman process. Results show that recurrent VAEs with a deep reconstruction-based diagnosis mechanism are recommended for industrial process monitoring tasks.
Highlights
Statistical process monitoring (SPM) is an important decision-making module in modern manufacturing sectors, allowing them to achieve higher plant safety, product quality, and enterprise profitability [1]
In the literature, [17] proposed the dynamic stacked autoencoder model to extract discriminative features for fault classification. Variant autoencoders such as denoising autoencoders and contractive autoencoders have been evaluated in extracting nonlinear feature representations for the fault detection of industrial processes in [18]; the results show that both models can deliver simple and effective performance
We show how to combine the recurrent unit with variational autoencoder (VAE)
Summary
Statistical process monitoring (SPM) is an important decision-making module in modern manufacturing sectors, allowing them to achieve higher plant safety, product quality, and enterprise profitability [1]. For fault detection and diagnosis, a new deep neural network, the multichannel one-dimensional convolutional neural network (MC1-DCNN), is proposed to investigate feature learning from high-dimensional process signals from the literature [16]. A new fault detection method, a convolutional gated recurrent unit auto-encoder (CGRU-AE), for feature learning from process signals is proposed in [23]. All these works have recognized that deep generative models can often outperform shallow generative models in process monitoring tasks. The fault detection and diagnosis capabilities have been established for the VAE variants, and all deep monitoring paradigms are comparatively studied on the TE process.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
