A novel semi-supervised pre-training strategy for deep networks and its application for quality variable prediction in industrial processes

•A semi-supervised autoencoder (SS-AE) is first developed as the basic network to extract quality-related features.•By hierarchically stacking multiple SS-AEs, a novel semi-supervised strategy is proposed for pretraining of deep networks.•SS-SAE can learn deep hierarchical quality-relevant features...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Chemical engineering science Jg. 217; S. 115509
Hauptverfasser: Yuan, Xiaofeng, Ou, Chen, Wang, Yalin, Yang, Chunhua, Gui, Weihua
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 18.05.2020
Schlagworte:
Deep learning
Quality prediction
Semi-supervised SAE (SS-SAE)
Soft sensor
Stacked autoencoder (SAE)
Deep learning
Stacked autoencoder (SAE)
Soft sensor
Semi-supervised SAE (SS-SAE)
Quality prediction
ISSN:0009-2509, 1873-4405
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:•A semi-supervised autoencoder (SS-AE) is first developed as the basic network to extract quality-related features.•By hierarchically stacking multiple SS-AEs, a novel semi-supervised strategy is proposed for pretraining of deep networks.•SS-SAE can learn deep hierarchical quality-relevant features from process data for quality prediction.•High prediction performance and fast convergence ability of SS-SAE are validated on two refining industries. Deep learning-based soft sensor has been a hot topic for quality variable prediction in modern industrial processes. Feature representation with deep learning is the key step to build an accurate and reliable soft sensor model from massive process data. To deal with the limited labeled data and abundant unlabeled data, a semi-supervised pre-training strategy is proposed for deep learning network in this paper, which is based on semi-supervised stacked autoencoder (SS-SAE). For traditional deep networks like SAE, the pre-training procedure is unsupervised and may discard important information in the labeled data. Different from them, SS-SAE automatically adjusts the training strategy according to the given data type. For unlabeled data, it learns the shape of the input distribution layer by layer. While for labeled data, it additionally learns quality-related features with the guidance of quality information. The proposed method is validated on two refining industries of a debutanizer column and a hydrocracking process. The results show that SS-SAE can utilize both labeled and unlabeled data to extract quality-relevant features for soft sensor modeling, which is superior to multi-layer neural network, traditional SAE and DBN.
ISSN:0009-2509
1873-4405
DOI:10.1016/j.ces.2020.115509