时频电磁法多分量联合反演及在川西火成岩勘探中的应用

Multi-component joint inversion for time-frequency electromagnetic(TFEM)method:A case study of igneous rock prospecting in western Sichuan

近年来,非地震勘探在川西地区火成岩、深层结构和断裂研究方面发挥了一定的指导作用,特别是时频电磁法通过弱异常信息提取、精细反演,在寻找火山通道、探索火山岩形成机制方面效果良好。由于人工场源的存在,其产生的非平面波效应、阴影效应以及场源附加效应会造成资料反演的多解性,使数据反演成像的质量偏低,甚至造成假异常。多分量电磁场反演是克服上述困难的一种有效手段。从时频电磁勘探的原理出发,推导了二维模型三维源九点差分结构的有限差分正演公式。根据正则化反演原理,构造了多分量联合反演目标函数,推导了适合时频电磁法的多分量联合反演算法。为了平衡不同分量之间数量级的差异,在反演过程中增加了动态平衡系数。理论模型测试结果表明,相对于单分量数据反演,多分量联合反演对模型恢复的效果更好。加噪模型数据测试结果表明,该算法正确、可靠,同时具有较好的抗噪性。利用该方法对川西地区实测时频电磁数据进行处理,反演结果与钻井信息和其他已知信息吻合,证明该算法实用且有效。

In recent years,non-seismic exploration has played a guiding role in the study of igneous rocks,deep-seated results,and fractures in western Sichuan,especially in the extraction and fine inversion of weak anomaly information by the time-frequency electromagnetic(TFEM)method,which has a good effect in searching for volcanic conduits and exploring the forming mechanism of volcanic rocks.However,the non-plane wave effect,shadow effect,and the additional effect of the artificial field source result in the multi-solutions to the data inversion,which leads to the low quality of inversion data imaging or even false anomalies.Multicomponent electromagnetic data inversion is an effective method to overcome the difficulties mentioned above.On the basis of the principle of timefrequency electromagnetic exploration,the finitedifference forward modeling formula for the ninepoint finite-difference structure of the two-dimensional model and three-dimensional source was derived.According to the principle of regularization inversion,a multi-component joint inversion objective function was constructed,and a multi-component joint inversion algorithm suitable for the timefrequency electromagnetic method was derived.To balance the difference in the order of magnitude between different components,we added the dynamic equilibrium coefficient to the inversion process.The theoretical model test results show that the multi-component joint inversion is more effective than single-component data inversion,and the noisy model test results indicate that the algorithm is correct and reliable and has great anti-noise performance.This method was applied to the observed time-frequency data from western Sichuan,and the inversion results were in good agreement with drilling information and other known information,which shows that the algorithm is practical and effective.