Research On Multi-source Heterogeneous Large Data Fusion Algorithms And Visual Analysis Method

Energy is the main material basis for human survival and development.The current energy problem has risen to the main factor affecting the national economic development.In view of China’s important primary energy industry,the coal industry,the state will increase its research efforts to avoid the shortage of coal in short supply.In the entire life cycle of coal mining,the main difficulty lies in how to suppress the occurrence of rock bursts.However,data analysis and research of rockburst experiments are faced with the problems of difficult data collection and insufficient analysis,so that researchers often only analyze the mechanism of rockburst from a single aspect.Rockburst is a complex phenomenon caused by many factors.Therefore,it is of great theoretical and practical significance to introduce the multidisciplinary fusion of multi-source heterogeneous big data fusion technology and visual analysis technology into rockburst experimental analysis and research.The rock burst experiment of the State Key Laboratory of Deep Geotechnical Mechanics and Underground Engineering is taken as the main research goal of this article,aiming at the difficulty of data collection,massive data storage and single data analysis in the rock burst experiment data This paper firstly simulates the complex rockburst geological environment through the deep national heavy laboratory,and designs a rockburst multi-source heterogeneous big data acquisition system,and selects related multi-source heterogeneous The sensor collected the information of rock burst experiment.After analyzing the related influencing factors of the rockburst experiment,a new multi-source heterogeneous big data fusion hybrid framework suitable for rockburst experiment analysis is proposed.For this framework,the rockburst data is pre-processed and proposed and implemented Three algorithms for rockburst data,one is to combine SAE sparse autoencoder algorithm and K-Medians clustering algorithm into a data-level fusion algorithm suitable for rockburst one-dimensional books-SAEM algorithm;second,use CNN The neural network processes the rockburst image data to obtain a CNN rockburst image processing algorithm;third,the above two algorithms are distributed and fused through a fuzzy set algorithm to obtain a new rockburst multi-source heterogeneous big data fusion combination algorithm ——Distributed Rockburst Decision combination algorithm.According to the experimental verification of the above three algorithms,by comparing the effects of each algorithm,it is concluded that the DRD combined algorithm not only has more diversified data sources,but also has a faster loss value reduction and better prediction effect.It solves the single problem of rock burst experiment data analysis source.Finally,a rock burst multi-source heterogeneous big data fusion algorithm and visual analysis platform based on rock burst experiment and safety and energy saving are designed and realized.It not only shows the visualization of rock burst experiment data,but also can be The linkage analysis and query research on the inflection point data.In summary,the research in this paper provides a strong basis for improving the safety of coal mining,and lays a strong foundation for the study of rock burst mechanism.