Research shows that polarimetric optical-fiber sensors embedded in composite laminates can monitor structural integrity and detect damage while the structure remains in service. An experimental arrangement can be created to monitor strain in composite specimens. Light from a linearly polarized helium-neon laser is converted into circularly polarized light using a quarter-wave plate, and is coupled into a polarization-maintaining optical fiber with a microscope objective lens and three-axis positioner. In order to ensure that the intensity modulation was a maximum, the input beam was polarized at 45 degrees to the axes of the fiber. Although this study used polarization-maintaining fibers, similar results also have been seen for standard single-mode fibers.
As fiber-reinforced polymer composites become increasingly popular, damage detection in these materials has become an important issue. Most of the conventional damage-assessment and nondestructive- inspection methods require structures to be taken out of service. These methods are time-consuming and need skilled technicians, and are often difficult to implement on hard-to-reach parts of the structure.
For these reasons, a built-in assessment system must be developed to monitor constantly the structural integrity of critical components. Optical-fiber sensors have shown a potential to serve this purpose. Dimensionally similar to reinforcing fibers in composites and immune to electromagnetic interference, these small, lightweight, corrosion-free fibers can also be used at high temperatures.
Several schemes of damage detection using optical-fiber sensors are being investigated. They range from simple threshold load-detection sensors requiring very little data processing to sophisticated strain-monitoring sensors that detect damage from abnormal strain response.
Depending on which scheme is used, the sensors can be categorized as intensity, polarimetric, and interferometric. Intensity sensors are the most rugged but the least sensitive; interferometric sensors are very sensitive but the least rugged. Polarimetric sensors rely on interference principles, but since the two interfering bean1s pass through the same path, external influences affect both beams equally. Thus they appear to be the best choice for applications in which global response is sought. They have better sensitivity than intensity sensors, yet are more rugged than interferometric sensors.
An experimental arrangement can be created to monitor strain in composite specimens. Light from a linearly polarized helium-neon laser is converted into circularly polarized light using a quarter-wave plate, and is coupled into a polarization-maintaining optical fiber with a microscope objective lens and three-axis positioner.
Light output from the fiber is collimated, then passed through a polarization beam splitter. This device splits the input beam into two beams with orthogonal polarization. Each beam is detected with a photodiode. Another half-wave plate is used at the output to align the polarization so both detectors are illuminated equally if the light output from the fiber is circularly polarized, and only one is illuminated if the light output is linearly polarized.
Voltage signals D1 and D2 given by the two detectors, proportional to the light illuminating them, are represented by
D1 (εz) = |Ex(O)|2 sin2 Φ; D2(εz) = |Ex(O)|2 cos2 ε
where ε = πL/Lb(ε); and Lb(εz) is the strain-induced beat length. The state of polarization (SOP) of the light output from the fiber is represented by a quantity 5, defined as
S = ± 1 corresponds to linear polarization; 5 = 0 corresponds to circular polarization. Voltage signals D1 and D2 are recorded with a personal computer using analog-to-digital data-acquisition hardware, and the SOP is calculated. Tills method of obtaining the SOP is immune to changes in the power coupled into the fiber.
Specimens are composite laminates with eight layers of 0.254-rnillimeter-thick woven glass fabric in epoxy matrix- Epocast 50A resin and 946 cold-curing hardener from Ciba-Geigy (now Novartis) in Basel, Switzerland. The specimens measure 120 by 20 by 2.1 millimeters. High-birefringence optical fiber (Fibercore HB-600) is embedded at the center of each specimen but at different distances from the beam's neutral axis. Defects are introduced in some specimens at the time of fabrication. Delamination is simulated by introducing a piece of release film. Cutting the glass fabric simulates the fiber break.
Specimens are progressively loaded in a three-point bending arrangement, and the SOP is recorded against both applied load and central displacement 8. The relationship between applied load and displacement 8 has al- ways been linear. As expected, 5 varies sinusoidally with respect to 8. When the phase Φ is unwrapped from these data and plotted with respect to 8, the relationship is again. found to be linear. The accompanying tables give the load values required to produce a phase change of 1t in the sensor output for specimens with different defects. For each type of defect, measurements were repeated on at Ieast three specimens.
The load required to produce a given phase change in the output of fiber-optic polarmetric sensors is reduced in the presence of defects. Three samples were used for each test shown in the tables, and the data represent an average of these samples. The scatter of these data is reasonable. To ensure that the intensity modulation was a maximum, the input beam was polarized at 45 degrees to the axes of the fiber. Although this study used polarization- maintaining fibers, similar results also have been seen for standard single-mode fibers.
This article is adapted from a paper presellted at the ASME Asia '97 Congress and Exhibition, held in Singapore Sept. 3-Oct. 2, 1997.