在对立井爆破过程中井壁振动信号进行准确采集的基础上,采用匹配追踪(matching pursuit,MP)算法建立Gabor原子库,实现了爆破信号重构分量和残差分量的有效分离。通过信号短时傅里叶变换(short time Fourier transform,STFT)和魏格纳分布...在对立井爆破过程中井壁振动信号进行准确采集的基础上,采用匹配追踪(matching pursuit,MP)算法建立Gabor原子库,实现了爆破信号重构分量和残差分量的有效分离。通过信号短时傅里叶变换(short time Fourier transform,STFT)和魏格纳分布(Wigner-Ville distribution,WVD)对重构信号的时频分析及霍夫(Hough)变换时频聚集性对比,验证了WVD时频分析的优势。结果表明:MP-WVD组合信号分析方法对信号变化的适应性强,能够精确捕捉信号的局部细节,适合用于爆破信号非线性时频特征的提取过程。可为爆破方案调整和参数优化提供参考。展开更多
In mine ventilation networks, the reasonable airflow distribution is very important for the production safety and economy. Three basic problems of the natural, full-controlled and semi-controlled splitting were review...In mine ventilation networks, the reasonable airflow distribution is very important for the production safety and economy. Three basic problems of the natural, full-controlled and semi-controlled splitting were reviewed in the paper. Aiming at the high difficulty semi-controlled splitting problem, the general nonlinear multi-objectives optimization mathematical model with constraints was established based on the theory of mine ventilation networks. A new algorithm, which combined the improved differential evaluation and the critical path method(CPM) based on the multivariable separate solution strategy,was put forward to search for the global optimal solution more efficiently. In each step of evolution,the feasible solutions of air quantity distribution are firstly produced by the improved differential evolution algorithm, and then the optimal solutions of regulator pressure drop are obtained by the CPM.Through finite steps iterations, the optimal solution can be given. In this new algorithm, the population of feasible solutions were sorted and grouped for enhancing the global search ability and the individuals in general group were randomly initialized for keeping diversity. Meanwhile, the individual neighborhood in the fine group which may be closely to the optimal solutions were searched locally and slightly for achieving a balance between global searching and local searching, thus improving the convergence rate. The computer program was developed based on this method. Finally, the two ventilation networks with single-fan and multi-fans were solved. The results show that this algorithm has advantages of high effectiveness, fast convergence, good robustness and flexibility. This computer program could be used to solve large-scale generalized ventilation networks optimization problem in the future.展开更多
基金financially supported by the National Natural Science Foundation of China (No. 51134023)
文摘In mine ventilation networks, the reasonable airflow distribution is very important for the production safety and economy. Three basic problems of the natural, full-controlled and semi-controlled splitting were reviewed in the paper. Aiming at the high difficulty semi-controlled splitting problem, the general nonlinear multi-objectives optimization mathematical model with constraints was established based on the theory of mine ventilation networks. A new algorithm, which combined the improved differential evaluation and the critical path method(CPM) based on the multivariable separate solution strategy,was put forward to search for the global optimal solution more efficiently. In each step of evolution,the feasible solutions of air quantity distribution are firstly produced by the improved differential evolution algorithm, and then the optimal solutions of regulator pressure drop are obtained by the CPM.Through finite steps iterations, the optimal solution can be given. In this new algorithm, the population of feasible solutions were sorted and grouped for enhancing the global search ability and the individuals in general group were randomly initialized for keeping diversity. Meanwhile, the individual neighborhood in the fine group which may be closely to the optimal solutions were searched locally and slightly for achieving a balance between global searching and local searching, thus improving the convergence rate. The computer program was developed based on this method. Finally, the two ventilation networks with single-fan and multi-fans were solved. The results show that this algorithm has advantages of high effectiveness, fast convergence, good robustness and flexibility. This computer program could be used to solve large-scale generalized ventilation networks optimization problem in the future.
