Xiyue Yang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Xiyue Yang has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Xiaoyong Gao
Department of Automation, China University of Petroleum (Beijing), Beijing 102249, China
2 shared publications - Changfang Ji
CNOOC Energy Technology & Services Limited-Drilling & Production Co., Tianjin 300452, China
1 shared publication - Chaodong Tan
China University of Petroleum (Beijing)
1 shared publication - Gang Wei
CNOOC Energy Technology & Services Limited-Drilling & Production Co
1 shared publication - Guanlin Wang
Department of Automation, China University of Petroleum, Beijing 102249, China
1 shared publication - Mingming Li
CNOOC Energy Technology & Services Limited-Drilling & Production Co., Tianjin 300452, China
1 shared publication
Latest Publications
- Open Access
Fault Diagnosis Method and Application for GTs Based on Dynamic Quantile SPC and Prior Knowledge
This paper addresses the challenges of fault diagnosis in gas turbines (GTs) utilized in oil and gas pipeline systems by proposing a novel multiparameter analysis framework that integrates dynamic, quantile-based Statistical Process Control (SPC) with prior domain knowledge. The proposed approach initially employs a dynamic quantile SPC model to establish adaptive control limits, effectively handling the non-stationarity and non-normality of gas turbine operational data. By analyzing parameter variations under typical operating conditions and incorporating expert insights, a multiparameter fault analysis matrix and corresponding weighting factors are constructed to facilitate fault diagnosis with prior knowledge. Furthermore, a fault probability model based on parameter change rates and weighting factors is developed to quantify the likelihood of different fault modes. An operating condition clustering and correction mechanism enables the dynamic adjustment of control limits, thereby preventing misdiagnoses caused by varying operational states. The validity of the proposed method is demonstrated using real data from a domestic pipeline gas turbine, validated by real domestic pipeline GT data, outperforming existing models, with a fault accuracy up to 10%. The approach efficiently estimates fault probabilities and accurately detects both sudden and gradual faults, significantly enhancing intelligent fault diagnosis capabilities for gas turbines.
Open AccessFault Diagnosis Method and Application for GTs Based on Dynamic Quantile SPC and Prior Knowledge
This paper addresses the challenges of fault diagnosis in gas turbines (GTs) utilized in oil and gas pipeline systems by proposing a novel multiparameter analysis framework that integrates dynamic, quantile-based Statistical Process Control (SPC) with prior domain knowledge. The proposed approach initially employs a dynamic quantile SPC model to establish adaptive control limits, effectively handling the non-stationarity and non-normality of gas turbine operational data. By analyzing parameter variations under typical operating conditions and incorporating expert insights, a multiparameter fault analysis matrix and corresponding weighting factors are constructed to facilitate fault diagnosis with prior knowledge. Furthermore, a fault probability model based on parameter change rates and weighting factors is developed to quantify the likelihood of different fault modes. An operating condition clustering and correction mechanism enables the dynamic adjustment of control limits, thereby preventing misdiagnoses caused by varying operational states. The validity of the proposed method is demonstrated using real data from a domestic pipeline gas turbine, validated by real domestic pipeline GT data, outperforming existing models, with a fault accuracy up to 10%. The approach efficiently estimates fault probabilities and accurately detects both sudden and gradual faults, significantly enhancing intelligent fault diagnosis capabilities for gas turbines.
- Open Access
Drilling and Completion Condition Recognition Algorithm Based on CNN-GNN-LSTM Neural Networks and Applications
Drilling and completion condition identification is of great significance in improving operational efficiency, reducing safety risks and optimizing resource utilization. However, traditional methods rely on experts’ experience and rules and have low recognition accuracy and poor robustness when facing dynamic working condition changes. In recent years, deep learning technology has shown great potential in the field of time series data analysis and multimodal data fusion. In this paper, we propose a hybrid deep learning model (CNN-GNN-LSTM) based on a convolutional neural network (CNN), graph neural network (GNN) and long short-term memory network (LSTM). The model extracts the local spatial features of multi-sensors via a CNN module to reduce the noise interference; models the nonlinear dependency between sensors via a GNN module to capture the complex interaction relationship; and mines the long- and short-term time dependencies via an LSTM module to accurately identify the dynamic change and transition process of the working conditions. This significantly improves the classification accuracy under dynamic changes in multi-conditions. This study compares the performance of four models: a CNN, LSTM, CNN-LSTM, and CNN-GNN-LSTM. The results show that the CNN-GNN-LSTM outperforms the other models in key metrics such as the classification accuracy, recall, F1 score, etc., and is more robust to noise interference and changes in complex working conditions. This study verifies the advantages of the hybrid model in multi-sensor complex scenarios and provides technical support for the intelligent development of drilling and completion condition recognition.
Open AccessDrilling and Completion Condition Recognition Algorithm Based on CNN-GNN-LSTM Neural Networks and Applications
Drilling and completion condition identification is of great significance in improving operational efficiency, reducing safety risks and optimizing resource utilization. However, traditional methods rely on experts’ experience and rules and have low recognition accuracy and poor robustness when facing dynamic working condition changes. In recent years, deep learning technology has shown great potential in the field of time series data analysis and multimodal data fusion. In this paper, we propose a hybrid deep learning model (CNN-GNN-LSTM) based on a convolutional neural network (CNN), graph neural network (GNN) and long short-term memory network (LSTM). The model extracts the local spatial features of multi-sensors via a CNN module to reduce the noise interference; models the nonlinear dependency between sensors via a GNN module to capture the complex interaction relationship; and mines the long- and short-term time dependencies via an LSTM module to accurately identify the dynamic change and transition process of the working conditions. This significantly improves the classification accuracy under dynamic changes in multi-conditions. This study compares the performance of four models: a CNN, LSTM, CNN-LSTM, and CNN-GNN-LSTM. The results show that the CNN-GNN-LSTM outperforms the other models in key metrics such as the classification accuracy, recall, F1 score, etc., and is more robust to noise interference and changes in complex working conditions. This study verifies the advantages of the hybrid model in multi-sensor complex scenarios and provides technical support for the intelligent development of drilling and completion condition recognition.