Subsurface imaging by electrical resistivity tomography (ERT) is increasingly used in geotechnical, geo-environmental, and hydrogeological investigation and monitoring. Cross-hole ERT (CHERT) is often used to avoid loss of resolution with increasing depth. However, symmetric artifacts may be induced because the measurement sensitivity around boreholes for some electrode configurations is fairly symmetric. This study aimed to investigate the best practice of ERT for monitoring underground construction or process through an actual scenario of ground improvement with jet grouting columns. The limitation of conventional surface ERT method was first vividly illustrated in an easy-to-understand context. Characteristics and performances of distinct types of CHERT configurations were then studied considering different borehole spacings. Further improvements were explored by evaluating various mixed arrays, model resolution-optimized array, and the comprehensive array. The results show the necessity of mixing two distinct types of electrode configuration. Among various combinations, a combination referred to as the nominal optimal (NOPT) is recommended when imaging resolution, the symmetric effect, and the measurement efficiency are all factored in. Unexpectedly, the resolution-based optimization does not lead to better results and could even induce negative effects. The results also show that the minimum aspect ratio of 2 may not be large enough for imaging complex underground conditions due to reduced sensitivity and increased 3D effect. The above findings are justified by the eigenvalue spectrum of the Hessian matrix, which is considered a better appraisal index than the model resolution for CHERT.
- Cross-hole ERT
- Electrode configuration
- Optimized array
- Underground construction monitoring