Book Description

Acquisition of downhole temperature measurements, in addition to production data, is routine in production systems. The temperature measurements, which are currently being used for pressure data correction, are cheap to acquire, accurate and have good resolutions. The answer to the question of how useful these temperature measurements can be, beyond the current utilization for pressure correction, was the goal of this research work. In the first part of this work, a mechanistic multiphysics and multiscale model for thermal transport process in a porous medium was developed, accounting for compressibility and viscous dissipation effects like Joule-Thomson and adiabatic expansion phenomena. To validate the model, a laboratory experiment was designed to allow for a controlled flow of air through a porous core, while measuring the temperature changes at different locations. The data acquired were used to verify the model and perform sensitivity studies, and the results showed functional dependencies of the model on useful reservoir parameters such as porosity, flow velocities and thermal properties of the rock and fluid; and these functional dependencies revealed the potential of temperature data as an additional source of constraining data in temporal and distributed reservoir parameter estimation. In addition, the temperature model was well suited for the application of a number of analytical tools that lead to the extraction of these useful reservoir characteristic information. In the second part, using multiresolution methods based on the second derivative of the Gaussian kernel, temperature measurements were combined with pressure data to improve the identification of transients in data as well as yield better behavioral filtering. Until now, only pressure measurements are used and this has shown to be unreliable. The approach developed here exploited the independence between the pressure and temperature measurements to constrain the estimation of the location of the breakpoints. The third segment of this research exploited the convective nature of thermal transport during flow to characterize near wellbore properties such as the extent of damage around a well (or extent of stimulation). The model lent itself to the application of the semianalytical Operator Splitting decomposition technique and as a result, the solution of the advection component could be separated and used to estimate near-wellbore structures such as damage or stimulation radius and permeability. As temperature measurements are an independent source of measurements, a joint inversion of production data and temporal temperature measurements, taken from multiwell production systems, showed a significant improvement in the reservoir state estimation problem, using state space estimation techniques like the Ensemble Kalman filter. This marked improvement was over the results from current approaches which match only production data. Results showed that introducing temperature improved the resolution of both permeability and porosity fields significantly. The last part of this research dealt with the estimation of flowrate, using only temperature measurements. The temperature model showed a strong functional dependence of temperature on flowrates at high Peclet number. By deconvolution, the advective flow kernel was separated from the diffusion part, and the complete flowrate history reconstructed from this kernel. Results showed that in synthetic and field cases, this extracted flowrate compared well with the true flowrate measurements. The philosophical significance of this work is that low-cost temperature measurements, which are measured routinely in producing wells, are a promising source of additional data for further constraining of reservoir characterization and optimization problems.