Quantitative estimation of permeability enhancement, production and depletion from microseismic data

It is a known fact that cost efficiency is the number one priority across the industry, perhaps even more so in today’s price environment. Operators are seeking technologies to improve profitability for unconventional field development. Getting a reliable estimate of production within days of completing an unconventional well is every operator’s dream. Armed with this knowledge operators would be able to evaluate the effectiveness of treatment and well spacing strategies without having to wait for months of actual production, allowing for them to arrive at an optimum development plan in a more timely fashion. In this article we demonstrate that microseismic monitoring, taken as an insitu, real time field observation of the completion of a well that integrates all the effects of rock properties, tectonic stresses and treatment parameters enables an estimate of well productivity within reasonable limits. In this method, microseismic data is used to describe a deterministic discrete fracture network model, differentiating between propped and un-propped fractures and further defining a fracture intensity in a given rock volume. This fracture intensity is translated into a measure of permeability enhancement, both magnitude and distribution, achieved through the process of hydraulic fracturing. Through reservoir simulation using this more complete model it is possible to estimate each well’s current and future productivity and the reservoir pressure depletion as a function of time. Current practices rely on unrealistic planar fracture models leading to to inaccurate predictions of unconventional wells’ production and drainage patterns. The approach presented here explicitly honours the microseismic measurements and provides a realistic representation of the induced fracture geometry and thus improved production and drainage estimates. These early production predictions are available within a few days after a well’s treatment, eliminating the typical 6-to-12-month waiting time for production results. The ability to create a model that is consistent with actual production, as shown in the case histories presented here, suggests that uncertainty is sufficiently constrained through the workflow outlined in this article.