Affiliation:
1. Henan Polytechnic University
Abstract
Abstract
The conclusion that the column vector of the phase space after reconstruction of the mine water inflow sequence has a clear geological meaning is proposed and verified by an example, and then a prediction model of the mine water inflow coupled with chaos theory and artificial neural network is established. The delay time τ is determined by the mutual information method, and the embedded dimension m is determined by Cao method, and the phase space reconstruction of the water inflow sequence is carried out; Regression analysis and Pearson correlation test are used to analyze the correlation between the elements in each column of phase space and the main controlling factors of water inflow; The chaos theory is coupled with Elman neural network (ENN), and the Chaos-ENN water inflow prediction model is established; The gradient prediction method is used to verify the model; Four evaluation indexes are used to evaluate the prediction results. The results show that the embedding dimension of phase space is equal to the number of main control factors of mine water inflow; The column vector of the phase space is linearly related to the mined out area, the development length and the water level of the three main aquifers; The prediction accuracy of Chaos-ENN model is 97.91% and the root mean square error is 10.81m3/h; The prediction accuracy of this model is equivalent to that of the single ENN model established after the main control factors and their values are determined. The results show that the column vector of phase space after reconstruction of water inflow sequence has clear geological meaning; For different mines, the geological meaning of column vector in phase space will be different, but they all represent a main controlling factor of water inflow; The number of ENN input layers and their values in the water inflow prediction model can be quantified by using chaos theory, so the Chaos-ENN model for mine water inflow prediction can be established only by the sequence value of water inflow. The establishment of the model avoids the difficulty of determining and quantifying the main controlling factors in the prediction of water inflow, and the model has high prediction accuracy, so the Chaos-ENN model has high promotion value.
Publisher
Research Square Platform LLC
Reference23 articles.
1. A new method for computing Lyapunov exponents;Barana G;Phys Lett A,1993
2. Practical method for determining the minimum embedding dimension of a scalar time series;Cao LY;Physica D,1997
3. Preditction Study on chaotic time series of mine water inflow (in Chinese);Chen GF;Min Res Dev,2015
4. Study on mine water consumption by phase space reconstruction and chaotic genetic neural network fusion (in Chinese);Deng G;World Sci Technol Res Dev,2016
5. Characterization of strange attractors;Grassberger P;Phys Rev Lett,1983