Abstract
This paper presents results of tests involving four different strategies used to automatically select the optimal set of array configurations that generate the maximum amount of information regarding a subsurface with 2D electrical imaging surveys invoking the same number of measurements. The first strategy (Compare R) calculates the improvement in the model resolution for each array configuration added to the set of measurements. This strategy generates the best results but requires the greatest amount of time to complete. The second (BGS) and third (ETH) strategies use linear approximations to calculate a goodness function. Both strategies are less time consuming than the Compare R strategy; however, the BGS strategy produces array configurations with model resolution values that are slightly less accurate than the Compare R strategy, while the ETH strategy produces results of the poorest quality. The fourth strategy (Combined BGS-CR) uses a combination of the BGS and Compare R strategies and produces results that are almost as accurate as the Compare R strategy but requires significantly less computing time. The four optimization strategies were tested using model resolution calculations, synthetic models and field surveys. The results obtained using the different optimization strategies were also compared with conventional arrays. The various tests show that arrays selected via the optimization strategies result in subsurface inversion models that have better resolution and identify more accurate depths and shapes compared to conventional arrays.