In order to define a favorable oil and gas accumulation area, this study focused on reservoir recognition which is based on logging data of old wells. The Gaofengchang structure in eastern Sichuan is used as a test area to discuss the necessity and feasibility of curve construction by combining new and old wells. Analysis of the reasons for the inaccuracy of the traditional curve reconstruction method is also provided. Given the interdependence of the well log in the depth domain sample sequence, a new intelligent construction method (BI-LSTM) based on the cyclic neural network is proposed. A discussion on the effect of data increments on prediction accuracy is also provided. The following four conclusions were achieved through curve reconstruction experiments: a high-precision CNL pseudo-curve was obtained; an underdetermined equation in optimization logging interpretation method needed to be extended to a positive definite equation; the quantitative processing of the complex lithologic reservoir parameters for the old wells was realized; and the processing result of the lithology physical property were basically consistent with the core data. Therefore, the BI-LSTM proposed in this paper could improve the accuracy of logging curve construction and has a good promotion significance for the old well review.

Currently, high-yield gas reservoirs have been discovered by exploring Lower Paleozoic tight carbonate oil and gas reservoirs in the Changqing area. These reservoirs, however, are impacted by subsequent karstification and diagenesis processes, leading to the formation of complex pore spaces with varying pore diameters ranging from microns to millimeters. Conventional well curves, with a resolution at the decimeter level, can only provide an overview of the reservoir porosity, making it challenging to evaluate the pore structure. On the other hand, electrical imaging logging data has a sampling interval of 0.1 inches and a resolution as fine as 5 mm, providing valuable information for pore structure evaluation in tight reservoirs. Since formation resistivity is influenced by factors like porosity, mud filtrate resistivity, and cementation, the cementation index serves as a reflection of the pore structure. Therefore, by calibrating the cementation index using data obtained from experiments at different scales, the pore structure parameters of tight carbonate reservoirs can be evaluated. This study conducted a comparison between numerous core experiments at various scales and both conventional and imaging log data. A method was proposed for extracting pore size distribution characteristics and calculating the pore size spectrum based on electrical imaging logging, ultimately leading to the establishment of a quantitative evaluation model for the pore structure of tight carbonate reservoirs. The results demonstrated improved accuracy and consistency in interpretation, as confirmed by well test data.

Accurate inversion of the T₂ spectrum of shale oil reservoir fluids is crucial for reservoir evaluation. However, traditional nuclear magnetic resonance inversion methods face challenges in extracting features from multi-exponential decay signals. This study proposed an inversion method that combines autoencoder (AE) and Fourier transform, aiming to enhance the accuracy and stability of T₂ spectrum estimation for shale oil reservoirs. The autoencoder is employed to automatically extract deep features from the echo train, while the Fourier transform is used to enhance frequency domain features of multi-exponential decay information. Furthermore, this paper designs a customized weighted loss function based on a self-attention mechanism to focus the model’s learning capability on peak regions, thereby mitigating the negative impact of zero-value regions on model training. Experimental results demonstrate significant improvements in inversion accuracy, noise resistance, and computational efficiency compared to traditional inversion methods. This research provides an efficient and reliable new approach for precise evaluation of the T₂ spectrum in shale oil reservoirs.

Water injection allocation is critical for maintaining pressure support in mature offshore waterflooding reservoirs, but its optimization is complicated by delayed injection–production responses, interwell interference, limited intervention windows, and incomplete field labels for injector–producer connectivity. This study proposes a connectivity-aware optimization framework that couples an attention-based connectivity identification network, a group-level long short-term memory (LSTM) production surrogate, and particle swarm optimization (PSO). The methodological novelty lies in using prescribed connectivity labels in a field-informed semi-synthetic benchmark to quantitatively test whether dynamic injection–production sequences and static well-pair attributes can be transformed into interpretable connectivity estimates for injection allocation decision support. The benchmark contains five injectors, ten producers, daily injection and production histories, static well-pair attributes, response lags, and normalized connectivity coefficients generated under practical injection rate, lag, water cut, and adjustment constraints. The attention model recovered the dominant injector–producer relationships with MAE = 0.0146, RMSE = 0.0240, R² = 0.9835, cosine similarity = 0.9962, and top-three overlap = 100%. The group-level LSTM achieved MAE = 4.524 m³/d, RMSE = 5.963 m³/d, MAPE = 1.255%, and R² = 0.964 on the chronological test set. Across 15 optimization cases, the PSO module generated feasible injection reallocations under single-well rate, total-injection balance, and +/−15% adjustment constraints. The results should be interpreted as controlled methodological validation rather than direct field deployment; further testing with anonymized field data is required.