Abstract:Quality prediction holds significant importance in monitoring the operational status of industrial systems. However, when analyzing a newly established production line, there are often numerous quality variables that are difficult to measure directly. Therefore, domain adaptation methods are increasingly employed for quality prediction in intelligent industrial monitoring by leveraging the existing production line knowledge. Nevertheless, these methods often discard domain-specific information resulting from variations in working conditions as noise, which limits their performance. On the contrary, this paper proposes a feature structured domain adaptation (FSDA) method to predict the quality variables of cross working conditions in industrial processes by using domain-specific information in the target domain for the first time. First, FSDA incorporates a novel feature representation that separates domain-invariant and domain-specific features from the feature set. Specifically, when adversarial game discovers the domain-invariant features, a feature structured strategy is proposed by defining a structured projection matrix to maintain the structured relationships between sample pairs, thereby enhancing specific information between target samples. Second, a dynamic weighted pseudo-labeling factor is designed to quantify the distribution differences between the two domains, allowing for the full utilization of knowledge from the source domain. The proposed feature representation method preserves domain-specific information while learning domain-invariant features in the target domain. Finally, the performance of FSDA is evaluated by conducting extensive experiments on the public C-MAPPS dataset and a real industrial dataset from the polyester esterification stage of the DuPont three-reactor plant. The experimental results demonstrate that the proposed method outperforms existing state-of-the-art methods and exhibits superior performance for quality prediction in industrial processes.

A multisource domain adaptation method for quality prediction in small-batch production systems

Feature selection and domain adaptation for cross-machine product quality prediction

Feature structured domain adaptation for quality prediction of cross working conditions in industrial processes

Production quality prediction of cross-specification products using dynamic deep transfer learning network

Multiscale Cost-Sensitive Learning-Based Assembly Quality Prediction Approach under Imbalanced Data

Multi‐source domains transfer learning strategy based on similarity measurement for batch process quality prediction

Adaptive Soft Sensors for Quality Prediction under the Framework of Bayesian Network

An Adversarial Bidirectional Serial–parallel LSTM-based QTD Framework for Product Quality Prediction

A Wide-Deep-Sequence Model-Based Quality Prediction Method in Industrial Process Analysis

Source-Free Domain Adaptation for Transferable Remaining Useful Life Prediction of Machine Considering Source Data Absence

Intelligent Prediction Method for Small-Batch Producing Quality Based on Fuzzy Least Square SVM

Cooperative Deep Dynamic Feature Extraction and Variable Time-Delay Estimation for Industrial Quality Prediction

Final Quality Prediction Method for New Batch Processes Based on Improved JYKPLS Process Transfer Model

Domain Adaptation Transfer Learning Soft Sensor for Product Quality Prediction

A Two-Stage Multi-Target Domain Adaptation Framework for Prediction of Key Performance Indicators Based on Adversarial Network

Deep Learning of Complex Batch Process Data and Its Application on Quality Prediction

Uncertainty Estimation Pseudo-Labels Guided Source-Free Domain Adaptation for Cross-Domain Remaining Useful Life Prediction in IIoT

A locally weighted multi-domain collaborative adaptation for failure prediction in SSDs

A novel DL-BiGRU multi-feature fusion and deep transfer learning based modeling approach for quality prediction of injection molded products using small-sample datasets

Quality Prediction and Quality-Relevant Monitoring with Multilinear PLS for Batch Processes

Privacy-Preserving Adaptive Remaining Useful Life Prediction Via Source-Free Domain Adaption.