遗传算法反演龙门山断裂带断层三维滑动参数研究

利用位错理论模型将遗传算法运用于断层三维滑动参数反演问题的求解,采用川西地区2004—2007年GPS观测数据对龙门山断裂带主要断层的三维滑动速率进行计算分析。结果表明：反演结果在量值上整体较小,与地质结果具有较好的一致性,走滑分量︱U1︱〈3.2 mm/a,倾滑分量︱U2︱〈1.54 mm/a,张开分量︱U3︱〈2.5mm/a,低滑动速率反映断层的闭锁及应力的积累及大地震发生的危险性;局部分量较地质结果偏大,反映实测GPS数据反演结果体现的是断层实时活动状态;遗传算法的全局收敛、不依赖初值等优点使结果更加稳定,而张开分量显示局部不规律性,表现出各子断层空间活动不均匀性。

Inversion of Three-dimensional Slip Parameters of Fault in Longmenshan Fault Zone Using Genetic Algorithm

Based on the dislocation theory model,genetic algorithm was used to inverse the three-dimensional slip parameter of fault.The GPS data from 2004 to 2007 in the western of Sichuan were used to calculate the three-dimensional slip rates of the main faults in Longmenshan fault zone.The results showed that the inversion,which accorded with the geological result,was low on the whole,strike-slip component ︱U1︱3.2 mm/a,dip-slip component ︱U2︱1.54 mm/a,stretch component ︱U3︱2.5 mm/a;low slip rate indicated the closure of the fault,the accumulation of stress and the probability of megaseism;part component,which was lager than the geological result,indicated that the actual GPS data inversion displayed the real-time active state of the fault;the results were stable because of the global convergence and independent initial value of genetic algorithm;stretch component indicated that the part was irregular and the space activities of different sub-faults were inhomogeneous.