Geothermal reservoir models are used to understand the current state of a geothermal resource and to predict its response to future utilization. These models are usually calibrated using records of temperature, enthalpy or pressure obtained from tests, or from operation of production and injection wells. While these data provide a continuous record of the evolving reservoir state at the location of the well, less is known about conditions further from the well. One way we fill in the gaps in our understanding is through the use of geophysical methods, e.g., a magnetotelluric (MT) survey, microearthquake (MEQ) monitoring. These provide a semi-continuous, spatial measure of some physical quantity (e.g., electrical resistivity, density of seismic events) throughout the reservoir volume. Although it will not generally be appropriate to insert this quantity directly into a reservoir simulation, conversion to some another property – say, permeability or temperature – can provide a useful constraint on the reservoir model. In this paper, we describe two such approaches for the incorporation of MT and MEQ data. A magnetotelluric survey uses sensitive recordings of the Earth’s electromagnetic fields above a geothermal field to estimate the subsurface distribution of electrical resistivity. Often, a particular focus is placed on imaging the extent, thickness, and depth of the hydrothermally altered clay cap that can form above the upflow zone, and which seals the reservoir from cool groundwater. Here, we present a stochastic methodology that integrates wellbore temperature profiles with a resistivity distribution derived from an MT survey. Our goal is to extrapolate reservoir temperature profiles away from the wellbore, and to probabilistically contour the top and bottom surfaces of the clay cap. To do this, we estimate 1D resistivity contrasts through the clay cap using Markov Chain Monte Carlo (MCMC) and then fit the most probable temperature distribution consistent with lithology and temperatures observed in wells. Clouds of microearthquakes are sometime induced during well stimulation, or around reinjection wells in geothermal fields. In most cases, the event magnitudes are small enough to not raise a safety concern, in which case they are a useful source of information about pressure and stress conditions away from a wellbore. Assuming MEQ density is reflective of local pore pressure changes, we demonstrate an inversion procedure that illuminates the spatiotemporal evolution of permeability about a well during stimulation. As it can be difficult to determine which of several mechanisms are responsible for permeability enhancement, our inversion is agnostic to the underlying physical processes. Instead, we determine permeability changes consistent with the pressure change around the well, the latter of which we infer from an MEQ catalog. Here, we demonstrate application of the method to a synthetic problem. The goal of both methods is to interpret common geophysical datasets in terms of physical quantities relevant to the simulation of geothermal reservoirs. These interpretations may serve as either a soft constraint – say, post-comparison of MT inferred temperatures with those from a reservoir model – or as a direct input to the reservoir model itself, e.g., the spatial distribution of low permeability in the caprock, spatiotemporal permeability changes around a stimulated well.

Press the Back button in your browser, or search again.

Copyright 2019, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.