Analysis And Prediction Of Surface Movement And Deformation Of Mining Subsidence Based On Kalman Filter

In the deformation monitoring of coal mining, surface subsidence forecasting plays an important role. In this article, we will use ZhuJidong ore repeated mining as an example, the 1111 (3) and 1122 (1) working mining face, on the main shaft coal pillar may have an impact. To ensure the safety of coal mining under three and refers to the main railway building guide next, under water, etc., We have to predict the future in a moment of surface movement deformation.This article introduced in detail of the ZhuJidongkuang 1111 (3) and 1122 (1) working face of the natural geographical conditions, geological and mining conditions, the observation station finished the situation, through the establishment of Kalman filter model, prediction function by Kalman the continuous filter Zhu Jidong mine shaft coal pillar monitoring line location and elevation plane dynamic forecast.The prediction accuracy of Kalman filter is dependent on the accuracy of the model and the statistical characteristic of random disturbance signal. The state transition equation and the observation equation are used in the Kalman filtering algorithm, which is estimated by the time. At the beginning stage of mining surface subsidence, the state parameters (velocity, acceleration, etc.) can be modified in real time, which has good adaptability. However, with the development of mining, the ground monitoring point began to enter the active phase, just using the basic Kalman filtering algorithm cannot usually get the desired results. At this time, we need to use the adaptive interactive multi model algorithm (IMM), to adjust the Kalman filtering model. Interacting multiple model algorithm was first used in the field of maneuvering target tracking, and the state of the work process was described by two or more models, and finally the system state estimation was carried out by a weighted fusion. To a large extent, it overcomes the problem that the single model estimation error is large. Data shows that in the period of entry into the active period and the time from the active period to the decline period, using IMM Kalman filter prediction accuracy 3cm within a point of the total points to 98%,far higher than the standard Kalman filter model, which greatly improving the prediction accuracy.In order to improve the operating efficiency and use of the powerful graphics capabilities, this article uses MATLAB r2009a as the development platform, develops a mining subsidence prediction system, forecast for the future of surface movement point plane location and elevation, forecast information automatically.