scholarly journals Semi-supervised Approach to Soft Sensor Modeling for Fault Detection in Industrial Systems with Multiple Operation Modes

2017 ◽  
pp. 683-693
Author(s):  
Shun Takeuchi ◽  
Takuya Nishino ◽  
Takahiro Saito ◽  
Isamu Watanabe
Keyword(s):  
Fault Detection ◽  
Soft Sensor ◽  
Industrial Systems ◽  
Operation Modes ◽  
Multiple Operation ◽  
Sensor Modeling

Soft sensor design for melt index estimation of a polypropylene process with multiple operation modes

2001 ◽  
Vol 34 (25) ◽  
pp. 359-364
Author(s):  
Minjin Kim ◽  
Chonghun Han ◽  
Byungwoo Lee ◽  
Jinsuk Lee ◽  
Woo Kyoung Kim ◽  
...  
Keyword(s):  
Soft Sensor ◽  
Sensor Design ◽  
Melt Index ◽  
Operation Modes ◽  
Multiple Operation ◽  
Index Estimation

Kernel PCA Performance in Processes with Multiple Operation Modes

2009 ◽  
Vol 4 (5) ◽  
Author(s):  
Mauricio L Maestri ◽  
Miryan C. Cassanello ◽  
Gabriel I Horowitz
Keyword(s):  
Fault Detection ◽  
Bench Mark ◽  
Kernel Pca ◽  
Statistical Process ◽  
Tennessee Eastman Process ◽  
Industrial Installation ◽  
Linear Relations ◽  
Operation Modes ◽  
Non Linear ◽  
Multiple Operation

Kernel PCA, as a multivariate statistical process monitoring (MSPM) tool, is a powerful technique capable of coping with non linear relations between variables, thus outperforming classical linear techniques when non linearities are present in data. In real industrial chemical processes, multiple plant operating modes often lead to multiple nominal operation regions, and MSPM techniques that do not take account of this fact show increased false alarm and missing alarm rates. The existence of multiple operation modes is often more frequent than clearly expressed strong non linear relations between the variables involved. Non linear relations do exist, but the small variability allowed in key variables during normal plant operation prevents these non linear relations from being expressed in the data. In this work, a fault detection tool based on Kernel PCA is tested in such multiple operation modes environments, with final objective of implementing the tool in a real industrial installation. Robustness of the tool for coping with a certain percentage of outliers is particularly examined. The tool is applied to three case studies: (i) a two dimensional toy example, (ii) a realistic simulation usually used as a bench-mark example, known as the Tennessee Eastman Process, (iii) real data from a methanol plant. Hence, the feasibility of implementing fault detection tools based on this technique in the field of chemical industrial processes is discussed.

Hierarchical Quality-Relevant Feature Representation for Soft Sensor Modeling: A Novel Deep Learning Strategy

2020 ◽  
Vol 16 (6) ◽  
pp. 3721-3730 ◽  
Author(s):  
Xiaofeng Yuan ◽  
Jiao Zhou ◽  
Biao Huang ◽  
Yalin Wang ◽  
Chunhua Yang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Learning Strategy ◽  
Soft Sensor ◽  
Feature Representation ◽  
Relevant Feature ◽  
Sensor Modeling

scholarly journals Industrial Semi-Supervised Dynamic Soft-Sensor Modeling Approach Based on Deep Relevant Representation Learning

Sensors ◽  
2021 ◽  
Vol 21 (10) ◽  
pp. 3430
Author(s):  
Jean Mário Moreira de Lima ◽  
Fábio Meneghetti Ugulino de Araújo
Keyword(s):  
Short Term Memory ◽  
Soft Sensor ◽  
Fine Tuning ◽  
Current Layer ◽  
Layer By Layer ◽  
Linear Processes ◽  
Method Performance ◽  
Modeling Approach ◽  
Ensemble Strategy ◽  
Sensor Modeling

Soft sensors based on deep learning have been growing in industrial process applications, inferring hard-to-measure but crucial quality-related variables. However, applications may present strong non-linearity, dynamicity, and a lack of labeled data. To deal with the above-cited problems, the extraction of relevant features is becoming a field of interest in soft-sensing. A novel deep representative learning soft-sensor modeling approach is proposed based on stacked autoencoder (SAE), mutual information (MI), and long-short term memory (LSTM). SAE is trained layer by layer with MI evaluation performed between extracted features and targeted output to evaluate the relevance of learned representation in each layer. This approach highlights relevant information and eliminates irrelevant information from the current layer. Thus, deep output-related representative features are retrieved. In the supervised fine-tuning stage, an LSTM is coupled to the tail of the SAE to address system inherent dynamic behavior. Also, a k-fold cross-validation ensemble strategy is applied to enhance the soft-sensor reliability. Two real-world industrial non-linear processes are employed to evaluate the proposed method performance. The obtained results show improved prediction performance in comparison to other traditional and state-of-art methods. Compared to the other methods, the proposed model can generate more than 38.6% and 39.4% improvement of RMSE for the two analyzed industrial cases.

Sign in / Sign up

Export Citation Format

Share Document