Relevance of Computational Rock Physics

Abstract:

The advent of computational rock physics has brought to light the often ignored question: How applicable are controlled-experiment data acquired at one scale to interpreting measurements obtained at a different scale? An answer is not to use a single data point or even a few data points but rather find a trend that links two or more rock properties to each other in a selected rock type. In the physical laboratory, these trends are generated by measuring a significant number of samples. In contrast, in the computational laboratory, these trends are hidden inside a very small digital sample and can be derived by subsampling it. Often, the internal heterogeneity of measurable properties inside a small sample mimics the large-scale heterogeneity, making the tend applicable in a range of scales. Computational rock physics is uniquely tooled for finding such trends: Although it is virtually impossible to subsample a physical sample and consistently conduct the same laboratory experiments on each of the subsamples, it is straightforward to accomplish this task in the computer.