Oil drilling is a messy business. Mud, oil, and grime are a routine and unavoidable part of life on an oil-drilling platform. The well borehole is a 3- to 8-in. diameter column of opaque drilling mud and ground earth. The completed well produces a turbulent, multiphase brew of dark crude oil, briny water, and methane gas. At the bottom of the hole, pressures may range up to 20,000 psi, and the temperatures may be 175°C.
Compared to the air-conditioned rooms where polished optics are meticulously cleaned with isopropyl alcohol and the labs in which optical analytical instruments are more typically used, these conditions are outrageously hostile. On first exposure to an oil-drilling platform, an optical engineer might conclude that optics cannot encroach upon this market, at any price.
This impression, however, isn't entirely true. In the late 1980s, optical instruments began to find niches within the oil and gas industry. Today, optical sensors are used commercially for spectroscopy, gas-flow imaging, discrete measurements of pressure and temperature, distributed Raman temperature measurements, fiber-optic telemetry from platforms to sub-sea production centers or between individual marine seismic sensors, and video pipeline inspections.1,2 The applications that became commercially successful leveraged optical technology to provide novel physical measurements rather than simply replacing existing electrical sensors.
what's at the bottom?
Figure 1. At the bottom of an oil well, a fluid-absorption spectrometer obtains spectra to determine the mixture of oil and water in the well output. The spectrometer must work at high temperatures and pressures. The instrument also contains an optical refractometer to detect the presence of gas in the mixture.
The most successful of these applications has been downhole fluid interrogation via time-resolved optical spectroscopy (see figure 1). The instrument uses a fluid absorption spectrometer, designed for operation at 175°C and 20,000 psi, to determine the water-and-oil fraction at the bottom of the borehole. Because the results from the spectrometer presume a fluid, the instrument must also includes a gas-phase detection system based on optical refractometry. The spectrometer/gas detection system operates in the borehole unattended, except for a 5000- to 20,000-ft power and telemetry cable that allows real-time analysis of the spectrometer's output at the surface.
The optical density, D, of a fluid is expressed as
where the transmittance Tλ of the fluid at wavelength λ is given by Tλ = 10-aλL, where aλ is the absorption coefficient of the fluid and L is the optical path length of the sample. The spectrometer can measure optical densities over a range of 3.5 with a resolution to within 0.1.
The spectrometer's primary function is to make a real-time measurement of the fraction of oil and water in the downhole and determine the colorand therefore, the compositionof the oil. It also quantitatively informs the oil-field operator when to collect a sample in a high-pressure bottle for later retrieval to the surface. Because the properties of fluids change when they are brought to the surface, a realistic measurement of these fluids at borehole pressures and temperatures is required.
Figure 2. Plot of optical density versus wavelength shows the optical absorption curves for water and several crude oils over a 2-mm path length. The path length is short because the mixture can be highly absorbing.
The absorption features in the near-IR spectral region are determined by molecular vibrational absorption, and the electronic absorption in the visible spectral band determines the "color" or visible opacity of the fluid (see figure 2). The spectrometers used in the oilfield cover a range from approximately 400 nm through 2 µm.
The measured sample is a mixture of oil and water, so the absorption curves in figure 2 are only ideal. Because oil and water do not mix, the sample may be segregated or stratified across the optical path length L. However, the oil and water phases may mix to a certain degree, depending on the fluid flow rate across the optical path length. The absorption spectrum measured during any time interval will therefore be a unique mathematical combination of the transmittance Tλ of oil and water:
where aOL and aWL are calibrated terms to account for the absorption due to molecular vibrations, and c OLeκ/λ and c WLeκ/λ are terms that account for the fluid "color" from electronic absorption. A study of a large sample of crude oils determined that the optical density from electronic absorption could be modeled by wavelength-dependent exponential term ceκ/λ, where κ is empirically determined and c is the coloration coefficient.3The spectrometer
The spectrometer in figure 1 includes a tungsten halogen lamp as the white-light source for the measurement. The light is split into a reference path and a measurement path via a fiber-optic cable. The measurement path transmits light through the absorption cell. The reference path allows the users to monitor lamp intensity and spectral variation during the measurement period, which can last as long as several hours. Because the absorption cell cannot be filled with a known fluid for calibration during use, the reference path is critical for robust and accurate measurements under these hostile conditions.
The measurement cell is constructed from thick sapphire windows to withstand pressures up to 20,000 psi. The sapphire windows are part of a tubular metal flow line in the tool. The formation output typically flows through the absorption cell at rates on the order of 700 cc/min. The sapphire windows are separated by 2 mm (the path length L) to keep the optical density within the dynamic range of the spectrometer. The visible and near-IR spectra of crude oils and water can be extremely opaque, and the coloration (opacity in the visible spectral band) can affect the optical density of the absorption in the near-IR region.
After the white light transits the measurement cell, a fiber-optic cable collects it and relays it to a spectral distribution system. A remotely controlled shutter permits the selection of either the reference or sample path. Instead of measuring the entire spectral band from 400 to 2000 nm, the system contains 10 narrow-band optical filters that allow it to sample only useful portions of the transmitted light. The design engineer can change these wavelength channels to reflect the spectroscopic interests of the oil specialist. During each acquisition cycle, the instrument scans the silicon and near-IR photodetectors and processes the data. The 10-channel spectral output and the calculated oil/water fraction are displayed graphically and numerically as a time log to the operator at the surface.
The instrument normalizes the optical spectrum with the reference leg, which compensates for all high-temperature variations in the lamp intensity, photodetector response, and filter characteristics. The software analysis accounts for pressure and temperature variations in the absorption spectra.
In order to conserve power and increase the reliability of the instrument, none of the spectrometer's components are cooled. It was a significant challenge to realize a laboratory-quality spectrometer that can operate for hundreds of hours unattended in such a hostile environment and can measure samples that humans don't in any way control or prepare for analysis. The gas-phase sensor
A great deal of empirical and analytical research was performed to realize a simple and robust model that covers a range of flow rates and oil/water fractions. A further complication, however, arises because the sample is not entirely liquid: The contents of the flow line may include particulates such as debris or sand, which introduce optical scattering, and gas, which may evolve from the solution if the pressure in the line drops. The model can handle scattering, but a three-phase flow that includes gas makes determining the oil/water fraction very difficult. With the application of a suitable algorithm and gas suppression by the surface operator, the oil/water fraction can be determined with this time-resolved spectroscopic technique to better than 10% at a 1 kHz acquisition rate. In other words, if the algorithm suggests a 60/40 ratio of oil to water, the accuracy of this ratio is 60 ±5% oil and 40 ±5% water.
Our gas-phase detector is based on the principle that the intensity of reflection near the Brewster angle and the critical angle at an interface differ significantly between oil, water, and gas. Methane gas at 20 kpsi, however, has an index of refraction nearer to that of water than of air. Discerning the difference between a gas and liquid thus requires a refractometer that can measure slight differences between critical angles and reflection intensities.
Light from a near-IR LED is p-polarized and transmitted through a sapphire prism onto a sapphire/flow-line interface. The reflected light travels to a linear detector array that determines the intensity and angle of the reflected light. The surface operator can detect gas and keep it dissolved in the pressurized fluid via a downhole pump, which greatly improves the robustness of the two-phase oil/water algorithm. Pipeline inspection
Sometimes solved problems yield insights to solutions in another application. Video logging or pipeline inspection is constricted by the visible opacity of crude oil when it sticks to optical surfaces. A novel pipeline inspection tool can record simultaneously a 360° side view of the pipe and also down the pipe interior. It utilizes the relative transparency of a thin film of crude oil at near-IR wavelengths that can be inferred from the absorption spectra in figure 2. At these wavelengths, Rayleigh scattering is also reduced, increasing visual clarity. The 880-nm sensitivity of certain CCD imagers, as well as imagers that go deeper into the near-IR spectral region (to beyond 1000 nm), can be used.
The video imager is designed into a 1-11/16-in.-diameter package, which can be thought of as an accessory lens to a standard video camera. The solid fused-silica cylindrical optical housing includes an internal reflective spherical mirror with a small hole in the center of the mirror coating. The combination of the common optical elements and the spherical mirror yield a ±45° side view in oil and water. This provides a side field-of-view of 6 in. on the wall of a 6-in.-diameter pipe casing. The hole in the spherical mirror coating allows light to travel to a pair of sapphire lenses that enlarges the forward field-of-view. This restores the forward image to its original angular view of ±35° in oil or water.
The solid cylindrical optical housing and the thick sapphire lens allow the optical imager to withstand pressures up to 10,000 psi at temperatures of 150°C. Although optimized for use with near-IR wavelengths in a fluid, the system performs very well with color imagers. The short effective focal length provides focused images at distances from lens contact to near infinity, even in air. A software algorithm corrects the side view for magnification distortion and converts the image to rectilinear coordinates. Because moving mechanical parts tend to cause problems in oil-field applications, this fixed-focus dual-view system is ideal.
These are only three of the optical sensors that have been designed specifically for the oil and gas market. By supplying new physical measurements while remaining competitive, robust, and reliable, optical tools will continue to capture market share from electrical sensors. oe
1. R. Schroeder, et al., "Fiber Optic Sensors for the Oilfield," Ch. 31, Handbook of Fibre Optic Sensing Technology, ed. J. M. López-Higuera, Wiley (2002).
2. P. Christie, et al., "Raising the Standards of Seismic Data Quality," Oilfield Review Magazine, Summer 2001.
3. A. Smits, et al., paper #SPE 26496, presented at the 68th Annual Tech. Conf. and Exhibition, Houston, TX (1993).
Robert Schroeder is a senior research scientist with Schlumberger-Doll Research, Ridgefield, CT.