Maximal information coefficient and geodetector coupled quantification model: a new data-driven approach to coalbed methane reservoir potential evaluation

In coalbed methane (CBM) development, it is imperative to evaluate the reservoir’s potential and identify the favorable areas to achieve higher production. This paper presents a novel quantified model inspired by the reservoir hydrodynamic heterogeneity. Treating the groundwater hydrodynamic zones as the computation units, the maximal information coefficient (MIC) and the geodetector model (GDM) were applied to measure the correlation between reservoir parameter and average daily gas production. An evaluation index system was then established, and its weight vector was obtained from a game theory model which couples the MIC scores with the GDM q-values. Taking the Gujiao mining area in the northern Qinshui Basin as a case study, the proposed model demonstrates a reliable classification of reservoir potential in alignment with the prior studies. The result indicates that the proposed model provides an insight into the relationship between CBM production and the reservoir heterogeneity, e.g., the distinction of hydrodynamics leads to the importance variation of the primary reservoir parameters and results in the uneven CBM production. Moreover, the model offers a quantified measurement of the correlation between reservoir heterogeneity and CBM production, and eliminates the subjective influence from the conventional methods.