Research On Economic Evaluation Of Residual Coal Remining Based On D-S Evidence Theory

Coal is an important non-renewable resource.In the long-term mining process,a lot of residual coal resources are left.The remining of residual coal is of great significance to the protection of resources and the sustainable development of society and economy.However,the mining of residual coal is difficult and the economic risk is high,so it is very important to carry out economic evaluation on it.At present,there is a lack of systematic evaluation and summary of economic evaluation algorithms for coal remining in domestic studies,and various related evaluation algorithms also have many problems.Therefore,this paper studies the economic evaluation method of remining of residual coal,and carries out application verification combining with the actual data of the study area.This paper established the economic evaluation system of residual coal remining based on D-S theory,which was composed of three algorithms: fuzzy comprehensive evaluation method,convolutional neural network and input-output method.Firstly,GIS software was used to digitize and perform numerical operations on the original data of the mining area to obtain the data of 24 key indicators related to economy,and a raster data set with a spatial resolution of 5 meters was established.Then,Multi-level fuzzy comprehensive evaluation method,convolutional neural network and input-output methods were used to evaluate the obtainable of residual coal in the mining area.Finally,according to the spatial distribution characteristics of the evaluation results,we set probability transformation function and mapped the evaluation result into the D-S recognition framework.In order to strengthen the advantages of each algorithm above,we modified the evidence fusion rules of D-S evidence theory and established the D-S information fusion system based on conditional judgment.Using this system,the economic evaluation results of each coal seam in the mining area are calculated.The results show that Multi-level fuzzy comprehensive evaluation method can evaluate a coal seam well,but there is no comparison between each coal seam,so this method can only provide the priority information within the same coal seam.Convolutional neural network can effectively extract spatial structure information and provide the most continuous,discriminative and reliable evaluation results,which can be used as the main evidence source in data fusion.The input-output method provides the most clear and intuitive boundary,but the parameter setting is too subjective and sensitive to the accuracy of data.Based on the above analysis,the fusion system established in this paper takes the boundary information of the input-output method as the classification condition of the fusion system,takes the evaluation information of the convolutional neural network algorithm as the core evidence,and introduces the evaluation results of the AHP and the input-output method to modify it.Through the fusion practice of four coal seams in the mining area and the differential analysis before and after,it is proved that this fusion method can fuse the three evidence sources well,fully strengthen the advantages of each algorithm,and can get stable and reliable evaluation results.