Research On Production Capacity Forecast Of Coal-measure Gas Co-production Based On Optimized Wavelet Neural Network

In recent years,with my country's attention to the potential and development value of coal-measure unconventional natural gas resources,it has gradually promoted the rapid development of the coal-measure gas industry.Coal-measure gas co-production productivity is a comprehensive index to measure the potential gas production capacity of coal-measure gas wells.The level of productivity directly affects the economic benefits of coal-measure gas projects.Due to the complex overlapping relationship between coal-measure gas reservoirs,the production process is determined by multiple geological factors and the relationship between the factors is complex,so it is difficult to establish precise mathematical expressions to describe the dynamic production process.Based on the genetic algorithm to optimize the wavelet neural network,the paper establishes a coal-measure gas co-production productivity prediction model,and predicts the productivity of the coal-measure gas wells in the study area,which provides a new idea for further research on the coal-measure gas co-production productivity prediction.First of all,on the basis of the geological characteristics that affect the production capacity of coal-measure gas co-production,combined with previous work results,the influencing factors of coal-measure gas co-production capacity are deeply discussed.The sample data of the study area is improved by GIS spatial interpolation,and the gray correlation method is introduced.The factors that have the greatest impact on the co-production capacity are screened out.Then,the wavelet neural network is used as the capacity prediction model,and the genetic algorithm is used to optimize the initial values of the weights and translation scaling factors of each layer in the network,which improves the generalization and convergence of the prediction model.Aiming at the problem that the sample data is small and the diversity affects the model training,the K-Means method is used to cluster the data into 3 categories to ensure the quality and commonality of each type of sample.A model was constructed for each type of sample after clustering,and the initial parameters of the genetic algorithm in the model and the number of nodes in each layer of the wavelet neural network were discussed,and three types of different structures based on coal-measure gas co-production capacity prediction were determined.The genetic algorithm optimizes the wavelet neural network model.Finally,the KNN method is used to classify the 20 coal-measure gas sample data in the selected study area,and then the corresponding genetic algorithm is used to optimize the wavelet neural network model to predict the production capacity,and to compare the prediction results of the other four models.Compare.The results show that the genetic algorithm optimized wavelet neural network model has the best predictive effect regardless of the prediction accuracy,stability or generalization ability,which verifies the applicability and accuracy of the model in the prediction of coal-measure gas co-production capacity degree.The paper has 40 pictures,13 tables,and 95 references.