Time domain induced polarization method: Time constant mapping with full-waveform inversion.

Time domain induced polarization method: Time constant mapping with full-waveform inversion.

Pole-dipole forward and reverse IP survey was conducted along line of interest. Main purpose of detailed survey to understand the nature of the two IP anomalies found during previous profiling. Time constant mapping may be useful, as example, for sulphide and graphite objects separation.

Survey description: forward and reverse pole-dipole. 5 meters between electrodes. Min spacing 7.5 meters, max spacing 250 meters. Length pulse/pause 1 sec., 20 windows. 3000 records for resistivity, 60000  total for IP.

Time constant is least sensitive from cole-cole parameters, so tau-mapping looks as complicated problem. Generally our approach can be described as step by step divide and conquer strategy.

At the beginning resistivity data was inspecting and IP decays denoising and smoothing for every record. Data quality is satisfied enough for full-waveform inversion. About 5% of records were removed.

1.    Resistivity inversion. Smoothness constraint algorithm.

No alt text provided for this image


2.    Chargeability (Static IP) inversion. Minimum structure algorithm was used to get compact objects only. 

No alt text provided for this image

3.     Apparent-tau pseudosection shows larger values for anomaly at the left part of line. We can already assume that this has  higher tau.

No alt text provided for this image

4.     To reduce number of inverted parameters and equivalence in full-waveform inversion, static IP model multiplied by some factor (inverted) was used instead chargeability every cell. Thus, the total number of parameters involved in the full-wave inversion was ncells*2+1, 4501. Additionally, to keep resulting models close to static IP  structure, cross-gradient operator was applied.

No alt text provided for this image

Full-waveform inversion took about 40 min for 10 iterations (5% final RMS). Results confirms higher tau value for anomaly in the left part of line.



Jesus Avila García

Senior Geophysicist | Applied Seismics and Geophysics | Structural Health Monitoring | Site effect Assessment | Data Science & Data Analysis

5 年
Eliana Yanez

Ingeniero Geologo, Msc. Recursos Naturales

5 年

Estoy interesada, deseo conocer mas sobre el proyecto

回复
Diego Badilla Elizondo

Geophysicist, PMP, Electronics Engineer

5 年

Very interesting results Alex, we did some dipole-dipole IP profiles and could nice to get some feedback from you! Cheers!

回复

Webinar from teoric items to software use??

回复

要查看或添加评论,请登录

Alex Kaminsky的更多文章

社区洞察

其他会员也浏览了