多通道瞬变电磁m序列全时正演模拟与反演

传统瞬变电磁方法主要用于金属矿勘查,无法满足油气资源高阻目标体的勘探需要.多通道瞬变电磁(MTEM)系统的出现解决了这一问题.该方法采用伪随机序列发射波形和拟地震观测方式,测量同线电场分量,记录全时发射电流及多道观测数据,实现对高阻薄层的高精度探测.鉴于国内对此方法的研究还处于理论探索阶段,尚未进行相应的仿真模拟和数据处理工作,本文针对m序列发射波形多通道瞬变电磁法的全时正演模拟和反演解释进行研究,为国内正在进行的MTEM仪器系统研发及数据解释提供理论指导.我们利用方波响应移位叠加和电流导数与阶跃响应褶积两种方法实现理论m序列和实际发射波形的全时正演模拟;再通过相关辨识技术,削弱噪声影响,计算脉冲响应;最后对积分得到的阶跃响应进行共中心点道集数据联合反演,获取地下电性分布信息.

Multi-transient EM full-time forward modeling and inversion of m-sequences

The traditional transient electromagnetic method is mainly used in mineral exploration. It is difficult to meet the demand for oil and gas exploration. The multi-transient electromagnetic (MTEM) system can solve this problem due to its grounded wire as transmitter and in-line electric field component, resulting in the underground resistive targets can be more easily detected. However, the domestic research for this method is still at the stage of theoretical derivation, lacking forward modeling and data processing. We work on full-time forward modelling and inversion for m-sequence MTEM method to provide theoretical basis for the on-going development of MTEM system as well as its data processing and interpretation.The MTEM system adopts quasi-seismic observation, measuring in-line electric field component of pseudo-random binary sequence (PRBS) waveforms. Full-wave transmitting current and multi-channel observation data are recorded, so that the resistive thin layers can be detected. We perform full-wave forward modeling for a theoretical m-sequence and actual transmitter waveforms based on superposition of square wave responses and convolution of step responses with derivative of transmitting current, respectively. Taking advantages of the correlation identification technique, the impulse response is calculated while the noise is suppressed. Finally, we invert the CMP data by the Occam's algorithm to obtain underground electrical information.Firstly, Fourier transform is used to analyze the spectral components of square waveforms, 5-order 2ⁿ-sequence and 5-order m-sequence. It shows that the m-sequence has a larger frequency bandwidth with an equal interval between spectral lines than another two waveforms. Secondly, in order to verify the full-wave forward modelling results, a homogenous half-space of 30 Ωm is set up with a suit of 4-order m-sequence transmitting current. The results obtained by superposition of square wave responses, convolution of step responses with derivate of transmitting current and convolution of impulse response with transmitting current are compared. It shows that the first two results are consistent. We further add 30 dB Gaussian noise to the responses. The relative error after correlation between modeling response and transmitting waveforms is reduced to 1/4 of the original one. The step response obtained by integrating the impulse response is very close to the theoretical value with relative error of just 0.08%. Finally, we design two three-layer models to demonstrate the capability of the MTEM method to resolve shallow and deep hydrocarbon reservoirs. In the first model, a reservoir of 500 Ωm and 50 m thickness is set at the depth of 300 m in a homogenous half-space of 50 Ωm. The survey is done respectively at offsets of 900, 1000, 1100, 1200 and 1300 m. For the second model, we assume the hydrocarbon reservoir is moved to 2000 m depth, the survey is done at 5000, 5200, 5500, 5700 and 6000 m offsets, respectively. The inversion results for the two models by the Occam's method show that the MTEM method can well resolve both shallow and deep hydrocarbon reservoirs. The CMP data have better resolution than one-offset data. The MTEM method using m-sequence as transmitting waveforms, measuring multi-channel in-line electric fields, turns out to have a good resolution to hydrocarbon reservoirs. Spectral analysis shows that the m-sequence has a wide frequency band with an equal interval between spectrum lines. The techniques based on superposition of square-wave responses and convolution of step responses with derivate of transmitting current are proved to be very effective for modeling theoretical m-sequence and actual full-wave responses. Using correlation identification technique, the influence from transmitting waveforms is removed in the calculation of impulse response while the noise is suppressed. The Occam's inversion of CMP data obtained by integration of impulse response can well resolve underground resistive targets like reservoirs.