Difference Inversion Of Ert Data: A Fast Inversion Method For 3-D In-Situ Monitoring

A new three-dimensional (3-D) inversion algorithm has been developed for electrical<br>resistivity tomography (ERT). The new algorithm is optimized for in-situ monitoring<br>applications. Instead of direct inversion of electric potential data, our new inversion<br>algorithm inverts the difference between the background data and the subsequent data sets.<br>The resistivity obtained by the inversion of background data serves as the apriori model in<br>the difference inversion. There are several advantages to this method. First, convergence is<br>fast since the inverse routine needs only to find small perturbations about a good initial<br>guess. Second, systematic errors such as those due to errors in field configuration and<br>discretization errors in the forward modeling algorithm tend to cancel. The result is that<br>we can fit the difference data far more closely than the individual potentials. Better data<br>fits often equate to better resolution with fewer inversion artifacts.<br>The newly-developed difference inversion technique was applied to monitoring in-situ<br>steam remediation in Portsmouth, Ohio and monitoring of flow in fluid fractures at the Box<br>Canyon site near the Idaho National Engineering Laboratory. Small changes of<br>conductivity were better resolved using the difference inversion method. Difference<br>inversion produced high-quality images with fewer artifacts, and only took 25% to 50%<br>run time of standard Occam’s inversion in the Box Canyon case.