Of the three categories for geothermal energy development, electrical power generation is somewhat more restrictive within the state. While Texas does not have active volcanoes or geysers, the state does have a very large number of oil and gas wells (~600,000) that were drilled, some of which have gone sufficiently deep to encounter high temperatures for the use in a binary-cycle or hybrid-cycle power plant system. This discussion will focus on the potential for West Texas developing geothermal power plants from these deep wells. The button labeled 'Geothermal Investigations - Delaware & Val Verde Basins' will be used to discuss data and ideas developed during a recently started Federal (DOE) and State of Texas (SECO) sponsored investigation.
In Texas, five regions display potential for electrical energy production due to access to high temperatures. Sedimentary basins, such as the Delware-Val Verde, Anadarko-Ardmore, East Texas, and the Gulf Coast, have wells drilled sufficiently deep to encounter high bottom hole temperatures (BHTs) that can be used in geothermal power plants. Additionally, the Trans Pecos of the far West Texas region displays higher temperatures closer to the surface of the Earth. A brief discussion of the Delaware-Val Verde Basins and the Trans Pecos is given here to demonstrate the geothermal well potential in these regions.
DELAWARE & VAL VERDE BASINS
Past investigations (Swift and Erdlac, 1999; Erdlac and Swift, 2004) have documented that the deepest part of the oil and gas rich Permian Basin may be adaptable for future geothermal energy production, especially involving electrical power production.
The Delaware-Val Verde Basins are highly complex in nature, with numerous folds, stratigraphic pinchouts and eroded highs (push-up structures), local basins (pull-apart), and thrust, strike-slip, and local normal fault orientations. A number of wells have reached depths of over 8 km (>26,000'). In previous studies (Swift and Erdlac, 1999; Erdlac and Swift, 2004) 3,625 BHT readings were plotted as depth in meters versus temperature in degrees C, representing over 2,000 wells incorporated in these studies. A large number of the BHTs are over 100oC, with several readings possibly reaching as high as 210oC in the deepest of wells. Thus temperatures exist within the range for use by conventional binary plant technology.
Among the interesting discoveries using this initial database was the observation that the subsurface distribution of temperature is a bit more complex than expected. In fact the temperature gradient appeared to be nonlinear in form. This was discovered empirically by plotting the data in both normal-normal and log-normal manner and then defining the statistically best-fit lines (linear and nonlinear) that described the data. The data was analyzed using three approaches: 1) testing the data as defined by a logarithmic function; 2) testing the data as representative of a shallow linear and deep linear best fit straight line; and 3) testing the data as defined by a shallow logarithmic function and a deep linear equation. When comparing the correlation coefficients (R) and the coefficients of determination (R2) for all approaches, the best fit was for the logarithmic function, giving R and R2 values of 95.6% and 91.39% respectively. Coefficient values for the other two analyses gave values within the 70-80% range, not bad statistically, but not as good as the logarithmic approach.
Why is this important? Fourier's law of heat conduction states that dT/dz = -q/k, where dT/dz is the change of temperature with distance (depth in this case), q is the heat flux (flow of heat / unit area / unit time), k is the coefficient of thermal conductivity, z is the coordinate in the direction of temperature variation, and the - sign indicates flow in the direction of decreasing T (Turcotte and Schubert, 1982). This equation is linear in its format. For a known heat flux and conductivity, -q/k is a constant and dT/dz defines a straight line. Similarly, if a surface temperature and a shallow temperature (100-300 m depth) measurement is taken, then a subsurface projection can be done to determine the dT/dz line. And if an average thermal conductivity can be guessed at, then the heat flux can be calculated for some particular depth.
This equation works fine in the laboratory where blocks of rock or metal or other materials are being investigated for their thermal conductivity. However, the subsurface BHT data suggests that nature makes the process a bit more complex when dealing with thousands of meters of rock thickness when compared to laboratory materials that measure only in cm thickness. In particular, the empirical data suggests a logarithmic equation of the form z = 3372.5 [ln (T)] - 11,083 as the best fit for the data. This equation can be algebraically manipulated to derive T = e 0.00029652 (z + 11,083). By taking the first derivative of this equation, two equations can be derived for the temperature gradient. The first equation, written as a function of depth, is given as dT/dz = (2.9652 x 10-4) e 0.00029652 (z + 11083). And by substituting for T, the equation can be written as a function of T, given by dT/dz = (2.9652 x 10-4) T. Thus for a given depth or subsurface temperature within the range of the data, a temperature gradient can be calculated directly from either of these values. And if the thermal conductivity of a particular rock type is known at a specific depth within the region, the heat flux can then be calculated within the formation of interest.
In the most simple of these analyses, assuming shallow and deep linear regression functions for the BHT data, the data demonstrated that the "deep" data had temperature gradients that were on the order of twice the shallow temperature gradient. Thus for a 'quick' estimate of how temperature varies in the subsurface, the data supports the contention that higher temperature gradients are present within the deeper parts of the basin when compared to shallower depths. However, for pinpointing the temperature gradient at a specific depth, the logarithmic function is more accurate, and will be of greater importance to industry for future planning and development of geothermal power plants in the region.
TRANS PECOS REGION
Recently, some preliminary investigations (Trentham and Erdlac, unpublished data) have used oil and gas wells drilled in a part of the Trans Pecos region called the "Rim Rock" country as a comparison with the Delaware-Val Verde Basin data. This "Rim Rock" data covers oil and gas wells drilled in primarily in Presidio County, but does include some wells from Jeff Davis and Hudspeth Counties as well as scattered oil and gas wells south of Alpine in Brewster County. A subset of the shallow wells provided by the SMU Geothermal Laboratory was also incorporated with the oil and gas well data.
While this data could be plotted as a normal-normal distribution, the data was also plotted in a log-normal manner. In the normal-normal plot, shallow and deep best fit linear fits gave thermal gradients of 21.5oC/km and 44.5oC/km respectively. These data were similar to the Delaware-Val Verde Basin data in that the deep gradient was at least twice that of the shallow data. However, the log-normal plot of the data gave a slightly better overall statistical fit. Two log-normal equations were again derived, the first as a function of depth giving dT/dz = (5.1334 x 10-4) e 0.0051334 (z + 6197.8), and the second as a function of temperature given by dT/dz = (5.1334 x 10-4) T. Thus both data sets appear to best be described as log-normal distributions. The reasons for this type of distribution is presently unknown.
References
Erdlac, Jr., R.J., and D.B. Swift, 2004, Deep permeable strata geothermal energy (DPSGE): Tapping giant heat reservoirs within deep sedimentary basins - an example from Permian Basin carbonate strata, in Shook, G.M., Technical Program Chairman, Geothermal Energy - The Reliable Renewable: Geothermal Resources Council Transactions, v. 28, p. 327-331.
Swift, D.B, and R.J. Erdlac, Jr., 1999, Geothermal energy overview and deep permeable strata geothermal energy (DPSGE) resources in the Permian Basin, in Grace, D.T. and Hinterlong, G.D., ed., The Permian Basin: Providing Energy For America: West Texas Geological Society Fall Symposium, Publication 99-106, p. 113-118.
Turcotte, D.L., and G. Schubert, 1982, Geodynamics: Applications of continuum physics to geological problems: John Wiley & Sons, New York, 450 p.
4901 E. University . Odessa, Texas 79762 . (432) 552-2020