Mark Houston and Philip Nash were interviewed for OEReports by Frederick Su.
Why are you looking at multiplexed fiber arrays for geophysical applications?
Houston: Litton has developed a borehole imaging system (Figure 1) that uses fiber optic hydrophones to listen to sound waves underground to locate and image oil and gas reserves. One advantage is that you don't have to use any underground electronics, which rarely survive the corrosive fluids and the high subsurface temperatures, which can exceed 150 deg. C, of the deep earth.
Figure 1. Litton fiber optic borehole system consists of (a) a topside cabinet containing lasers and demodulation electronics and (b) a downhole hydrophone array with no electronics (courtesy Litton Guidance and Control Systems).
A second advantage is weight. Fiber optic sensors don't require copper for electrical power wires, so the cables can be made much smaller and lighter.
Fiber optic sensors will give us a better image of rock formations. This will be done by using large numbers of underground sensors spread over a volume within the oil field. To construct a subterranean sensor net, we need to multiplex or connect a large number of these sensors on a limited number of cables.
Nash: Fiber optic arrays have a number of applications -- on land, in the ocean, and down oil wells. For ocean-based systems, the two main applications are seismic streamers and seabed arrays (Figure 2), which are known as ocean bottom cables (OBCs). Streamers are very long (up to 8 km) line arrays of hydrophones that are towed behind surface vessels during seismic surveys (part of the oil exploration process). They are used as acoustic receivers in conjunction with impulsive sources (usually airguns) to build up a picture of the subsurface geology. Streamers can contain up to 10,000 individual hydrophone channels. OBCs can also be laid onto the seabed for the same purpose. They are normally laid in a grid that is continuously deployed and recovered. OBC surveys are slower to carry out, but usually give higher quality data.
Figure 2. The Defence Evaluation and Research Agency (UK) seabed array (a) laid out and (b) being deployed at sea. (© British Crown Copyright/DERA 2000).
How do you detect these sound waves underground?
Houston: We use hydrophones and geophones to detect seismic waves in the earth. A hydrophone senses P-waves or pressure waves (Figure 3), which are the same as sound waves in the ocean or in air, i.e., variations in pressure. A P-wave has particle motion that is parallel to the direction of wave travel. A standard geophone detects vertical ground motions, which may be induced by either P-waves or S-waves. S-waves or shear waves have particle motion that is perpendicular to their direction of travel and, unlike P-waves, cannot travel through liquid.
Figure 3. Compressional or P-waves transmit through both fluids and solids. Particle motion is parallel to the direction of travel. Shear or S-waves transmit only through solids. Particle motion is perpendicular to the direction of travel (Courtesy Litton Guidance and Control Systems).
Nash: A typical geophone measures the displacement of an inertial mass relative to a rigid case, which is excited by the seismic signal. They are basically a mass-spring system, in which the fiber measures the strain in the spring. The main difference between the hydrophone and the geophone is that the hydrophone measures changes in a surrounding pressure field and, at low frequencies, is omnidirectional, while a geophone is measuring a component of acceleration and is inherently directional at all frequencies. Geophones are normally used in a three-axis configuration, with three orthogonally mounted devices. Typically, geophones will be used at lower frequencies than hydrophones, normally below 100 Hz.
And how does the hydrophone work?
Houston: The Litton hydrophone sensor is configured as a Mach-Zehnder interferometer (Figure 4) that is sensitive to pressure. Laser light enters the sensor and is split into two arms, one sensing and one reference. The sensing arm consists of fiber wound on a hollow air-backed cylinder. The fiber on the reference arm is wound around a solid mandrel that is insensitive to pressure variations. Any variation in the relative path lengths between the sensor and the reference legs will show up as an interference pattern or phase change in the recombined light signal emerging from the sensor. As sound waves or pressure changes occur around the hydrophone, the hollow cylinder changes its diameter, just slightly.
Figure 4. Pressure changes are measured by the change in diameter of a hollow mandrel on the sensing leg of a Mach-Zehnder interferometer. This hydrophone configuration achieves 120- to 140-dB dynamic range (Courtesy Litton Guidance and Control Systems).
Nash: The dimensional change in the mandrel can be very small -- about the diameter of an atom.
Houston: This diameter change stretches the fiber that's wrapped around the cylinder and increases the sensor path length. The reference arm is unaffected, and the path length difference of the recombined light signal results in a phase difference that is proportional to the pressure change. The light signal returns to the surface along the return fiber, and the topside electronics demodulate the phase changes to output 24-bit digital words of sound-wave pressure levels.
Phil, can you talk about military applications?
Nash: Military applications are mainly for antisubmarine warfare and consist of three main areas: (1) seabed arrays (similar to OBCs, but mainly using hydrophones) for area surveillance, either rapidly deployable or fixed systems; (2) towed arrays behind surface ships or submarines (similar to seismic streamers, but usually rather smaller in scale); and (3) submarine hull-mounted arrays. In all cases, the main emphasis is on hydrophones rather than geophones, although geophones are being considered for some applications.
Mark, what's the diameter of the borehole and how deep does it go?
Houston: The borehole often starts out fairly large at the surface, sometimes in excess of 20 inches in diameter. The diameter quickly narrows to as small as 5 inches and is further reduced by the insertion of pipe or casing that prevents unconsolidated rock layers fr om collapsing and blocking the hole. The casing also helps control well fluids. The engineers may need to inject water or gas to extract gas and oil. To accomplish this, they use production tubing and other hydraulic or electrical devices that control the flow of water, gas, and oil. It is helpful to fit a sensor inside these restricted spaces in the production plumbing.
The size of the Litton sensor is about 0.5 in. in diameter. With the armor cabling and packaging to protect it from abrasion and corrosive fluids, the diameter is around 1.5 in. This downhole sensor can slip into some tight places underground.
The current system has a 10,000-ft. lead-in. In the future, we'll build longer cables that can go down 16 to 20,000 ft.
These sensors are hardened against the environment?
Houston: Yes, the sensors are protected by their packaging. The optical fiber can withstand temperatures up to 500 to 600 deg. C. The fiber, couplers, and mandrels need to be protected against CO2, hydrogen sulfide, and oil and gas fluids. We protect the sensors so that they are sensitive to very small variations in pressure (0.05 microbars). The fiber optic hydrophones resist the static pressure at depth and they sense the transient pressure of a passing seismic wave. Our current hydrophone can survive a maximum pressure of 8,500 psi. (about 20,000-ft. depth in water), but we also have designs that will survive to 20,000 psi.
Let's talk more about the multiplexing.
Houston: We have a borehole system that has 96 channels with a single hydrophone for each channel. This configuration is a survey tool for cross-well imaging and tomography.
The 96 hydrophones, spaced at 5-ft. intervals, are packaged in a single, constant-diameter armored hose with an active array of 475 ft.
We use Frequency Division Multiplexing (FDM), where tunable, high-performance, narrowband lasers are adjusted to radiate at slightly different frequencies. We have six lasers for these 96 channels or sensors. Laser number one illuminates 16 sensors by transmitting light of a particular central wavelength and modulation frequency, and the array "wiring" architecture is such that the laser illuminates a unique set of sensors. Similarly, laser number two illuminates a different set of sensors. Note that while one laser illuminates 16 sensors, the number of illumination and gather fibers are not the same, so the average is 8 sensors per fiber pair.
As the laser light traverses the sensors, the sensors impose a pressure-proportionate phase change on the emerging light. The return fibers gather light from different sets of sensors and route them to optical detectors. The combined sensor signals are separated by electronic signal demodulation.
All six Nd:YAG lasers are run continuously. About 20 mW of optical power is transmitted down the hole by each laser.
Nash: A typical architecture for us would use a pair of fibers to carry time/wavelength multiplexed signals to and from the arrays, and would then branch off individual wavelengths into sub arrays of up to 64 channels (Figure 5). By channels, I essentially mean one fiber coil together with a reflective element, which comprises a directional fiber x-coupler with a mirror on one port.
Figure 5. A DERA hydrophone/geophone array using TDM/WDM architecture. Eight pulsed lasers (therefore eight wavelengths) are used, and at each add/drop multiplexer in (a), one optical wavelength is routed to a TDM or array module (b) where up to 64 hydrophones (or geophones) are time multiplexed -- i.e., time of flight is used to discriminate return signals. This gives 8 X 64 or 512 channels. The hydrophones are the discs marked 1,2, , N. The 64 time-multiplexed optical signals coming out of the TDM module (these are the shaded rectangular pulses shown in "b") now carry the acoustic/seismic signal and are recombined by the add/drop multiplexer back onto the telemetry fiber. This routes the returning signals to the receiver, where the different wavelengths are separated using a wavelength demultiplexer, and then the TDM signals are electronically demultiplexed. (© British Crown Copyright/DERA 2000).
A fiber laser is normally run CW at around 1550 nm, but the signal is then pulsed using either acousto-optic or electro-optic modulators to produce the pulses required for time multiplexing. The pulse length is typically in the region of 100 ns-1 µm.
Houston: The next generation Litton system will use Time Division Multiplexing (TDM) as well as Wavelength Division Multiplexing (WDM). Instead of having 8 sensors per fiber pair, as in the borehole systems, we plan to have more than 100 sensors per fiber pair. With this technology, connectivity and the number of lead-in fibers will be less of an issue, and electronics will become more cost efficient on a per channel basis. Practical channel counts will increase from hundreds to thousands, which will enable reservoir managers to do a significantly better job of volumetric sampling and imaging.
For FDM systems, the lasers are on all the time for all sensors; frequency and electronic demodulation separate the channels. For TDM/WDM systems, the lasers send out intermittent bursts of light; signal-packet time-of-flight and optical demodulation are used to separate returns from different sensors. That is, a light packet from the laser travels sequentially to each sensor in the group, and the first signal to return from the group on the gather fiber will be from the "nearest" sensor with the shortest travel path and so on. Each sensor or channel will have a unique time-slot within a sensor group. By combining WDM with TDM we can "double-up" on common TDM time-slots and use wavelength to separate the return signals from different sensor groups. We can get more channels for less cost and fewer fibers.
Why do you need a pair of fibers for each sensor set?
Houston: With separate illumination and gather fibers, you can control the optical power distribution among all sensors in a group so that all sensors have the same apparent brightness. Otherwise, the brightest channel in the set would dominate the noise for all channels and only the brightest channel would have the best signal to noise performance. Secondly, a sensor telemetry that uses a common illumination and gather fiber, on average, reflects half the power back toward the source. High amplitude internal reflections can increase the optical noise floor and degrade the dynamic range. Thirdly, the MZ sensors can be sampled over a very short time window; other sensor configurations may require an average over a larger time window.
How do you determine which rocks contain oil and/or gas?
Houston: We use geophones as well as hydrophones because three-component geophones allow us to record shear waves as well as P-waves. P-waves travel through both fluids and solids and are affected by both the rock matrix and the fluid content of the rocks. The rock matrix is the structural part of the rock exclusive of any fluid inclusions. S-waves, which travel only through solids, are affected only by the properties of the rock matrix. By recording both types of waves, we can begin to separate the effects of different rock types from fluids within the rocks, and the asset manager can better describe the underground reservoir, rock properties, and fluids within the rocks, all of which is what we want to do for active reservoir management.
The wave characteristics help determine the presence or absence of fluids in the rock, whether it is oil, water, or gas (Figure 6). Seismic attributes such as P- and S-wave attenuation with frequency, changes in P- and S-wave velocities, "splitting" of S-waves, and P- and S-wave reflection amplitude and phase variations all reflect subsurface reservoir patterns of rocks and fluids. However, there is an overlap in seismic attributes that are dependent on rock types and fluid content, so prospects require a judgement based on the geologic context to properly assess whether oil and gas are in place or whether it is just plain water.
. P-wave amplitude map draped over the time structure of a single horizon. Darker areas indicate higher amplitudes. An East-West low amplitude skinny valley within the upper half of the map (circled) can be interpreted as a high-porosity channel of sand that could be a likely oil trap. Additional attribute maps derived from S-wave data would reduce the interpretation ambiguities. This example has been provided by Wayne Pennington of the Michigan Technological University from the SPOT (Seismic Petrophysics: Observation and Theory) program. Other images with a more complete explanation can be viewed at www.geo.mtu.edu/spot/
Subsurface seismic sensors -- fiber optic sensors -- will give us the higher frequencies, better resolution, and volumetric sampling that we need to reduce the development risks and to actively manage our oil and gas reservoirs. We'll be more efficient, so, hopefully, it will make the price of oil and gas more stable.
Frederick Su is a freelance writer based in Bellingham, WA. Web: www.bytewrite.com.
Mark H. Houston is the Director for Business Development, Fiber Optic Applied Technology for Litton Guidance & Control Systems. He has a PhD in geophysics/geology from Rice Univ. and has worked within the offshore exploration and production community for more than 25 years. He has held management positions in operations, field support, R&D, and sensor/instrumentation manufacturing. His expertise is the development and application of new technologies for the improvement of data acquisition.
Phil Nash currently leads the Fiber Optic Sensors Group at DERA Winfrith, where his principal research interests are in fiber optic hydrophones. He has been working in the fiber optic sensing field for 14 years, first at Plessey Naval Systems and at DERA since 1994.