Rockburst Prediction Based On Neural Network And Microseismic Information

At present, many underground project or space have been near or above1,000meters in domestic, and have entered the stage of deep operations. The large depth will cause adverse effect for the project to implement successfully and the safety of personnel and equipment. Especially, the rockburst caused by high ground stress is undoubted the top priority. This paper is depended on previous studies, combined with the rockburst problems in the excavation process of the diversion and drainage tunnel in Jinping â…¡, and applied the method of microseismic monitoring combined with Neural Network to predict the rockburst. The main work and research include:(1) Rockburst is a highly complex dynamic phenomenon, which affected by many factors, such as high stress, structural plane, excavation and so on. But, in this paper, the microseismic information parameters of rockburst is major found from the aspect of the microseismic monitoring, and the encountered instances of rockburst and the related information is collected in the project, which,is the foundation of the rockburst prediction.(2) The microseismic monitoring and rockburst prediction in the excavation process of the great depth tunnel. According to the conditions of project site of the Jinping â…¡, microseismic monitoring system has been established. Through the real-time transmission of the micro-seismic signals, the microseismic data is analyzed timely, and the potential field of rockburst will be preliminary predicted. By analyzing and processing each parameter in microseismic information compared with the actual rockburst, the evolution of microseismic information before rockburst is explored, and the basic norm of rockburst prediction is also determined. With the specific example, the lag rockburst in deep underground project is analyzed, and the characteristics and internal rules of it are also determined and identified. Studies above have shown that, microseismic monitoring can be used as a major technique of rockburst prediction, especially in engineering, necessary measures can be taken in a timely manner in the potential area of rockburst, in order to achieve the purpose of preventing or reducing the hazard of rockburst. But, the prediction of the specific time of rockburst, especially the time of lag rockburst, is needed to be researched still.(3) Using the Neural Network model to predict rockburst. By analyzing the impact of parameters of microseismic information on rockburst, we can determine the parameters which are independent relatively, and establish the sample database of rockburst prediction based on microseismic information and Neural Networks. By using the different periods of the samples in microseismic information, we can create two different Neural Network models. Then compaired with their learning results and predictive power, we can make error analysis. Finally we can choose the model of rockburst prediction which is fit for the large depth tunnel of Jinping â…¡.These results can be learned to prevent and control the rockburst of Jinping â…¡ hydropower and the similar projects.