Spatio-Temporal Characteristics Of Coalbed Methane Production Based On DTW Clustering

Coalbed methane(CBM)is a high-quality clean energy.Mining CBM before coal mining can reduce the occurrence of accidents such as gas explosions during the mining process,and enable the rational use of CBM resources.With the development of China's CBM industry,a large amount of data has been accumulated in the process of CBM mining,which contains a wealth of information.Therefore,data-driven knowledge mining and analysis is of practical significance for exploring the genesis of high production of CBM wells.At present,the research on CBM is mainly carried out on some static attribute data.CBM drainage is a dynamic process in a continuous period of time,during which changes in gas output,geological factors,and drainage control parameters may conceal important drainage modes and productivity influence rules.Furthermore,during the mining process,CBM wells located close to each other will have different levels of linkage effects,and their characteristic information in space cannot be ignored.It is of great significance to study the temporal and spatial characteristics of CBM production to explore the comprehensive effects of CBM production capacity.The main research content and results of this paper are summarized as follows.1)The improved DTW algorithm is used to measure the similarity of each parameter curve among CBM wells,and then the characteristics of each parameter curve are extracted by spectral clustering.In this paper,the CBM vertical wells in Fanzhuang block of Qinshui Basin,whose actual gas production time is more than two years,are taken as the research object.Through the extraction of the morphological characteristics of CBM production curves,water production curves and casing pressure curves,the process characteristics of CBM production,water production and casing pressure control in Fanzhuang block are studied,and the causes of productivity differences are analyzed from the perspective of drainage process.In the study,the gas production curves are classified into three types,and the water production and casing pressure curves are classified into two types.The results show that the improved DTW algorithm can correctly identify the fluctuation characteristics of parameter series and effectively measure the similarity between two curves with length difference and position offset.CBM wells with gentle change of daily gas production and long-term high-yield have better productivity benefits.In the process of production control,the use of stable mode is more conducive to produce ideal productivity benefits.2)From the perspectives of geological conditions and drainage control,the main influencing factors of the CBM production characteristics in the study area are extracted.Drainage parameters are selected as daily water production and casing pressure,which are extracted into two curve characteristic modes through DTW and spectral clustering.There are many geological factors that affect productivity.Through the calculation of the maximum information coefficient,the correlation degree of the various relationships between the qualitative factors and the productivity benefits of each region is fairly measured and calculated,and the 6 main geological influencing factors in the study area are extracted.In addition,inter-well interference is a comprehensive manifestation of various geological factors and the results of changes in the drainage control process during the drainage and production process.In this paper,based on the spatial and temporal information of CBM well drainage,the degree of inter well interference is calculated by using transfer entropy.The results show that the calculated inter well interference roughly covers the influence of some fault structures on productivity.3)On the basis of nine influencing factors,C4.5 decision tree algorithm is used to construct CBM productivity benefit decision tree and CBM production mode decision tree,from which 10 strong rules of productivity benefit and 9 strong rules of production mode are extracted.Integrating all the rules of the two decision tree models,applying appropriate casing pressure control mode and drainage mode to different geological conditions is more helpful to achieve the ideal productivity benefit.The positive effects of the two casing pressure and water production modes in the process of drainage and production are as follows: the casing pressure mode with smooth speed drop and zero trend and the bimodal stable water production mode are mainly to control the steady progress of the drainage and production process.Multi peak transformation casing pressure mode can timely release the output hysteresis state in some time periods in the process of drainage.Oscillation jump water production model can accelerate the migration and accumulation of gas in coal seam,and make the drainage process smoothly.