Research On Intelligent Analysis Of Microseismic Activity Based On Cloud Computing

Microseismic monitoring technology is necessary for ground pressure disaster monitoring in mining.However,many problems arise in the current application of microseismic monitoring technology.These problems mainly include a large amount of data collected during microseismic monitoring,intense pressure and difficulty in data processing,inaccurate calculation of source parameters,unclear analysis mechanism of microseismic activity,and different standards.Based on cloud computing technology and intelligent method,the above technical difficulties can be effectively overcome at the "cloud-edge-end," The intelligent analysis of microseismic activity can be jointly realized,providing technical support for underground production safety.Therefore,to promote the application of microseismic monitoring technology in the mining area and ensure mine production safety based on cloud computing and intelligent method,the following work is carried out to improve the timeliness and reliability of microseismic activity analysis:(1)Research on intelligent processing technologies of microseismic monitoring data.Aiming at the primary processing of the original data collected by the terminal microseismic monitoring system,a joint algorithm based on STA/LTA and Faster-RCNN is proposed to detect suspected microseismic events from the continuous waveforms.UACNet,a deep network model without data preprocessing,was presented to identify microseismic events,and the recognition accuracy was 95.62%.Caps Net,which can be used in limited data sets,was proposed as a supplement so that the mine can complete the task based on any dataset.The PSNSeg Net model separated the microseismic signals into noise,P wave,and S wave based on mask generation for the identified microseismic events.The arrival errors of the P/S wave were all below 80 ms.Meanwhile,based on Cycle GAN,arbitrary measurements of collected signals,such as acceleration,velocity,and displacement,are generated to realize data fusion of microseismic signals in the data layer.Finally,the collected microseismic signals can complete basic processing and data collation at the terminal level without manual intervention,providing data support for subsequent analysis.(2)Source location intelligent inversion based on the three-dimensional(3D)velocity fields.In order to take the anisotropy of rock mass into full consideration when calculating source parameters,an implicit modeling method based on borehole data and deep learning was proposed to construct a lithologic grid model and map it to the P/S wave propagation velocity field.Based on the velocity fields,a seismic source scanning algorithm considering the P/S wave energy function is proposed.The characteristics of the seismic source spatiotemporal energy field are obtained based on the fast travel method,waveform forward modeling,and octree method.A source search reinforcement learning based on spatiotemporal energy field characteristics is constructed.Finally,accurate source location inversion of microseismic events based on a 3D anisotropic model can be realized at the edge.The error of source location inversion results is controlled within 10 m to realize the preliminary analysis of microseismic activities and deep data processing.(3)Microseismic activity analysis based on spatiotemporal data mining.In order to analyze microseismic events caused by different rupture sources,the focal mechanism of microseismic events is taken as its characteristic critical attribute and combined with the spatial attribute.The DBSCAN clustering algorithm eliminates traditional microseismic clustering analysis’ s single spatial attribute constraint.The microseismic events generated by the same rupture source are integrated to realize the cluster analysis based on the focal mechanism.Aiming at the microseismic event group,Conv LSTM,a spatial-temporal data mining algorithm,was used to excavate the internal law of microseismic activities in mines and predict the events with high intensity,to achieve the purpose of early warning of ground pressure disaster events.(4)Design and development of cloud computing system for intelligent analysis of microseismic activities.Based on the cloud computing architecture idea of "cloud-edge-end," a new microseismic monitoring system platform is designed.A corresponding relational database is constructed to realize the micro-seismic monitoring system’s existing data processing automation and intelligent data analysis.At the same time,the visualization and interaction of microseismic monitoring systems are realized in the form of the Web application.It reduces the difficulty of obtaining the system and is helpful to the popularization and application of microseismic monitoring technology.Finally,based on the above intelligent technology and cloud platform architecture,it has been applied in the Xitieshan lead-zinc mine of Western Mining and achieved good results.