Deep coalbed methane (CBM) is rich in resources and is an important replacement resource for tight gas in China. Accurate prediction of post-fracture production and dynamic change characteristics of fractured wells of partial CBM is of great significance in predicting the final recovery rate. In terms of predicting time-series production, the problem one encounters is low prediction accuracy and poor generalisation ability under limited sample conditions. In this paper, we propose a hybrid deep neural network (AT-GRU-MTL) production prediction model based on the combination of an attention mechanism gated recurrent neural network (GRU) and multi-task learning (MTL), where the AT-GRU is responsible for capturing the nonlinear pattern of the production change, while introducing an MTL method that includes a cross-stitch network (CSN) and a weighted loss using homoskedasticity uncertainty to automatically determine the degree of sharing between multiple tasks and the weighting ratio of the total loss function. The model is applied to several typical deep CBM fracturing wells in China, and the accuracy of gas production prediction reaches 90%, while the accuracy of water production prediction is 68%. The experimental results show that, for the blocks with a very large difference in the order of magnitude of the gas and water production, it is very easy for a certain small order of magnitude to be suppressed from learning during the two-way multi-task learning process, which leads to deterioration of its prediction effect; at the same time, the adaptability of the model is evaluated, and it is found that the model is more advantageous for the wells that have been produced for approximately one year. Meanwhile, the evaluation of the model adaptability shows that the model is more dominant in the prediction of wells with production of about one and a half years. Based on the two test wells with shorter (380 days) and longer (709 days) spans, the results indicate that the model may have insufficient sensitivity to the sudden change of the ratio of gas to water and the failure of the dynamic generalisation of the matrix shrinkage–desorption coupling, and the introduction of physical constraints (such as bottomhole flow pressure, etc.) or the division of the data into the production stages may be attempted to deal with the case subsequently. The research results in this paper provide a theoretical basis for dynamic production prediction and analysis in oil and gas field sites.

Deep coalbed methane development faces technical challenges, such as high in situ stress and low permeability. The dynamic evolution of fractures after hydraulic fracturing and the flowback mechanism are crucial for optimizing productivity. This paper focuses on the inversion of post-fracturing fracture volume parameters and dynamic analysis of the flowback in deep coalbed methane wells, with 89 vertical wells in the eastern margin of the Ordos Basin as the research objects, conducting systematic studies. Firstly, through the analysis of the double-logarithmic curve of normalized pressure and material balance time, the quantitative inversion of the volume of propped fractures and unpropped secondary fractures was realized. Using Pearson correlation coefficients to screen characteristic parameters, four machine learning models (Ridge Regression, Decision Tree, Random Forest, and AdaBoost) were constructed for fracture volume inversion prediction. The results show that the Random Forest model performed the best, with a test set R² of 0.86 and good generalization performance, so it was selected as the final prediction model. With the help of the SHAP model to analyze the influence of each characteristic parameter, it was found that the total fluid volume into the well, proppant intensity, minimum horizontal in situ stress, and elastic modulus were the main driving factors, all of which had threshold effects and exerted non-linear influences on fracture volume. The interaction of multiple parameters was explored by the Partial Dependence Plot (PDP) method, revealing the synergistic mechanism of geological and engineering parameters. For example, a high elastic modulus can enhance the promoting effect of fluid volume into the well and proppant intensity. There is a critical threshold of 2600 m³ in the interaction between the total fluid volume into the well and the minimum horizontal in situ stress. These findings provide a theoretical basis and technical support for optimizing fracturing operation parameters and efficient development of deep coalbed methane.