Research And Application Of Automatic Recognition And Location Technology Of Optical Fiber Microseismic Signals In Mines

Microseismic monitoring technology is one of the effective technical means in the courses of monitoring mining safety.Source location is the core of microseismic monitoring technology,and the automatic pick-up of arrival time of first arrival is an important prerequisite for source location.Therefore,the accurate and rapid pick-up of arrival time of first arrival is the key to achieve source location in a low SNR environment.Among them,the accuracy of source location directly determines the effect of microseismic monitoring system,but in actual mining engineering,a series of problems such as inaccurate source location of microseismic monitoring system,low efficiency have emerged due to adverse environment.Consequently,the source of high precision positioning plays a crucial part in the mine safety monitoring.At the same time,after the seismic event information quantity become large,need to be unified information management,but the efficiency of traditional seismic event information management is not high and a lot of work need to be done,and timeliness of information and sharing cannot meet the actual demand,so it is imperative to develop a set of seismic event information management system based on Web.In this paper,the optical fiber microseismic monitoring system is taken as the research object,aiming at three aspects of optical fiber microseismic signal automatic first arrival wave picking,highly precise source location and microseismic event information management system based on Web.The following work need to be completed:(1)Aiming at the low SNR,the traditional STA/LTA method has low accuracy of picking up first arrival waves,on the contrary,an automatic picking up first arrival waves based on random forest was proposed.Firstly,the characteristics of microseismic data were extracted,and each feature sample was labeled by feature category.Then,the microseismic waves were distinguished by the random forest classifier.Finally,the arrival time of the first arrival wave was accurately picked by combining the corresponding output probability value of each feature sample.The experimental results show that the random forest method has higher accuracy for first arrival picking,and the classification accuracy of microseismic feature samples is up to 98.5%,and the minimum picking error is 3.2 milliseconds.(2)In order to address the issue of instability and large error in the process of source location,the particle swarm and Newton iteration location methods are determined.Firstly,the initial solution of source location is obtained by using particle swarm optimization,and then the solution is used as the initial value of Newton iterative algorithm to carry out accurate iterative operation.Finally,the final location result is realized until the convergence of the results.The experimental results show that the method is hard to diverge and stable in the later solution process,and the seismic source location obtained is close to the real blasting location,with an error of 12.8 meters.(3)To tackle the problems of low efficiency of traditional microseismic event information management,a Web-based microseismic event information management system is designed.The system is based on B/S architecture,j Query+ Ajax framework is employed by the front end to achieve microseismic information display,Spring Boot+My Batis Plus+SQL Serveris is applied by the back end to accomplish microseismic data management,the data suggests that the system's availability and stability meet the operating requirements through the experimental test.(4)Taking the actual test data of Ningtiaota Coal Mine as an example,the proposed methods of first arrival picking and source location are applied and verified,and the microseismic data information management is realized by using the microseismic event information management system based on Web.The feasibility and reliability of the research contents is verified.