Surface Movement Angle Predicition In The Metal Mines Based On BP Neural Network

The large-scale exploitation of mineral resources brings mankind great economic and social benefits, but it produces a series of negative impact on the survival environment of mankind too. The impact induced by mining subsidence on the environment is an important aspect, so, it's imperative to study the scope of the strata and surface movement. An accurate forecast to the subsidence situation during the mining process is of great significance to the safety of the mines.Aimed at the research actuality of surface movement feature in the metal mines, the process and damage types and forms of strata and surface movement caused by mining are analysed. Based on the contrast of the main differences in underground metal mine and coal mine, geological and mining factors which influence the surface movement angle are stated. The result shows that the influencing factors are various, interactive, certain or uncertain due to mining in difference mines are all different and complicated. Artificial neural network technology has highly nonlinear mapping capability, self-organizing, self-adaptive and self-learning ability, particularly suited to address the complex causal relationship between non-deterministic reasoning, judgment prediction and classification problems. So, using BP neural network to predicate the collapse scope of surface is proposed.It's easy and effective to make the BP neural network model by using the Neural Network Toolbox (NNT) based on MATLAB to evaluate the predictive model. We adopt a three-layer BP neural network. Input-layer contains seven nerve cells (properties of hanging wall and footwall surrounding rock, mining depth, ore body inclination, etc.)and output-layer contains two nerve cells (hanging wall and footwall displacement angles).The training results show that the selecting factors are reasonable, having deep internal relations between the input and output factors. It's proved that the scope of strata and surface movement due to mining could be predicted in theories. Through the test, the result shows the feasibility of the predictive model. So it's valuable and has a bright future.On the basis of prediction, sensitivity analyses are put up to seven quantitative as similar to the above. The results show that, each factor's change will cause haling wall and footwall displacement angles' changes. In this thesis, the similarities and differences between each factor' sensitivity in Metal mines where caving or resuming-fill mining methods are used are aggregated first time.The visualization of forecast to mine subsidence can be realized by GUI interface in MATLAB software, in this way, we can further enhance efficiency, reduce workload and improve the usability and scalability of the network.