Session: 11-03-02 Data Science Applications in Drilling

Paper Number: 128345

128345 - Early Lost Circulation Monitoring Using a Hybrid Cnn-Lstm Model 

   Lost circulation is one of the most common issues affecting drilling safety, known for its sudden occurrence. Traditional human assessment methods heavily rely on expert experience, exhibiting a high degree of subjectivity and lag. Conventional machine learning approaches struggle to fully capture the spatial and temporal variations within the multidimensional data of lost circulation, leading to insufficient accuracy. To address the challenge, this paper proposes an early intelligent monitoring approach for lost circulation by concatenating a Convolutional Neural Network (CNN) and a Long Short-Term Memory Neural Network (LSTM). Initially, a sliding window method is employed to structure historical data of lost circulation into time series samples. Subsequently, these time series samples are input into a one-dimensional CNN network to extract spatial feature vectors. Following this, the extracted feature vectors are input into an LSTM network to uncover temporal feature information. Finally, the Softmax function is applied at the network's output layer for classification. The proposed model is tested on a real dataset of lost circulation. The results indicate that the model outperforms conventional methods like CNN and LSTM, with accuracy and recall rates both exceeding 90%. Compared to human assessment, the model detects lost circulation risk earlier, significantly improving its timeliness. This research holds substantial significance in safeguarding safety and enhancing drilling efficiency.

Presenting Author: Liwei Wu 1. China University of Petroleum, Beijing; 2. Tianjin Branch of CNOOC(China) Co.,Ltd.

Presenting Author Biography: Ph.D. student at the College of Petroleum Engineering, China University of Petroleum, Beijing, China, with research interests in the intersection of drilling engineering and artificial intelligence

Authors:

Liwei Wu 1. China University of Petroleum, Beijing; 2. Tianjin Branch of CNOOC(China) Co.,Ltd.
Liang Han China University of Petroleum, Beijing
Gensheng Li China University of Petroleum, Beijing
Xianzhi Song China University of Petroleum, Beijing
Qilong Zhang 1. China University of Petroleum, Beijing; 2. Tianjin Branch of CNOOC(China) Co.,Ltd.
Rui Zhang College of Artificial Intelligence, China University of Petroleum, Beijing
Ziyue Zhang China University of Petroleum, Beijing