.png){kind=icon}
返回
Causality-driven sequence segmentation assisted soft sensing for multiphase industrial processes
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.129612.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
The dynamic characteristics of multiphase industrial processes present significant challenges in the field of industrial big data modeling. Traditional soft sensing models frequently neglect the process dynamics and have difficulty in capturing transient phenomena like phase transitions. To address this issue, this paper introduces a causality-driven sequence segmentation (CDSS) method. The CDSS method segments the industrial time series into difference phases, based on the causal mechanism shifts. A novel distance-based dual similarity metric is designed to evaluate the temporal consistency of causal mechanisms, including both the causal similarity and the stable similarity. Furthermore, a soft sensing model, called phase-specific temporal-causal graph convolutional network (PS-TC-GCN), is established for each phase, by using the segmented phase data and the discovered temporal causal graphs (TCGs). The numerical examples are utilized to validate the proposed CDSS method, and the segmentation results demonstrate that the CDSS has excellent performance on segmenting both the stable and the unstable multiphase series. Especially, the CDSS has higher accuracy in separating non-stationary time series compared to other methods. The effectiveness of the proposed CDSS method and the PS-TC-GCN model is also verified through the penicillin fermentation process. The experimental results indicate that the breakpoints identified by the CDSS are accurately aligned with the ground truth. Furthermore, the PS-TC-GCN achieves higher predictive accuracy, compared with other soft sensing models.
Keyword:
Time series segmentation
Multiphase industrial process
Causal discovery
Soft sensing
Graph convolutional network
期刊
IF:
6.5
论文数:
2.5W
被引数:
6.5W
机构
引用论文
Causal Generative Model for Root-Cause Diagnosis and Fault Propagation Analysis in Industrial Processes工业过程根本原因诊断和故障传播分析的因果生成模型
PREAI
分享收藏Soft sensor model development in multiphase/multimode processes based on Gaussian mixture regression基于高斯混合回归的多相/多模过程软测量模型开发
PREAI
分享收藏DFFIR-net: Infrared Dim Small Object Detection Network Constrained by Gray-level Distribution ModelDffir-net: 灰度分布模型约束下的红外弱小目标检测网络
PREAI
分享收藏Distributed event-triggered estimation for dynamic average consensus: A perturbation-injected privacy-preservation scheme.png)
INFORMATION FUSION
INFORMATION FUSIONIF15.5
PREAI
分享收藏The influence of comorbidities on mortality in bronchiectasis: A prospective, observational study
OAAI
OAAI分享收藏
查看更多内容