Answers are grounded in this paper and cite their sources.
Underground gas storage wells have higher sealing integrity requirements for the casing-cement sheath-formation system due to the characteristics of alternate strong injection-production mode and high gas leakage risk. The cement sheath is crucial to the zonal isolation and production safety as a core component of the wellbore. However, the conventional monitoring methods only assess cement conditions at a specific moment and lack the capability for full-cycle dynamic monitoring. Therefore, the article proposes a new method using Distributed Acoustic Sensors (DAS) to monitor the Top of Cement (TOC) and fluid leakage location. A series of DAS monitoring experiments are conducted to evaluate wellbore integrity, including a 24-h cement slurry hydration process and fluid leakage process within cement. Then the Continuous Wavelet Transform-Convolutional Neural Network-Bidirectional Long Short-Term Memory (CWT-CNN-BiLSTM) network is established to capture the sealing failure characteristics of the cement sheath. The results indicate that the hydration strain change of cement slurry and the position of TOC are accurately determined by the DAS system. The location and flow rate of fluid leakage in cement sheath cracks are determined by analyzing the strain rate profile and acoustic energy. The CWT-CNN-BiLSTM network adaptively extracts the time-frequency information of non-stationary signals and reveals the mapping relationship between signal features and cement sheath failure conditions. Compared with the conventional model, the model demonstrates greater robustness and reliability, achieving a classification accuracy of 99.31 %. In conclusion, the novel monitoring method can determine whether remedial measures are needed by providing information on the cement sheath integrity. • The experimental platform of DAS system monitoring the cement sheath integrity is built, and the feasibility of DAS monitoring the cement sheath integrity is verified. • The Continuous Wavelet Transform-Convolutional Neural Network-Bidirectional Long Short-Term Memory (CWT-CNN-BiLSTM) network is established to capture the sealing failure characteristics of the cement sheath. • Two methods for monitoring the top of cement by DAS are developed. TOC monitoring is realized in the two working conditions of cement hydration process and cement solidification. • The location of fluid leakage inside the cement sheath is accurately obtained, and the relationship between fluid leakage characteristics and optical fiber signals is determined. • Using the experimental data to drive the deep learning model, the monitoring effect of DAS on the integrity of cement sheath is effectively improved.
No takes yet. Share an insight, caveat, or question.
Su et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: