• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Fiber Optic Colorimetry Technique For In-Situ Measurement Of Corrosion In Civil Structures

Posted on: Thursday, 11 March 2004, 06:00 CST

KEY WORDS: Fiber Optic Sensors, Corrosion, Civil Structures, Highways, Colorimetry

Corrosion of civil structures (bridges, buildings) & roadways is a persistent problem. Both steel and reinforced concrete structures are subject to corrosive attack. However, due to the difficulty in repairing reinforced concrete, its corrosion problems are perhaps more critical. Engineers have been concerned with the manner in which the pre-stressed concrete behaves during corrosion. The reinforcement steel on corrosion swells to a volume that is about five times larger than that of the original steel. This swelling causes spalling and cracking of the concrete, which further exposes the reinforcement to corrosion and can promote serious structural damage. In the USA alone, annual losses in time, labor, materials and systems due to corrosion problems have been estimated in the billions of dollars. Besides the economics of corrosion damage to buildings and highway infrastructures, there is a safety concern regarding structures with reinforced concrete ex- posed to harsh corrosive environments. Therefore periodical inspections are often carried out to monitor their safety.

Over the years, several methods of monitoring corrosion have been devised with some success. At present, there are only a few methods of determining, non-destructively, whether a corrosion process is underway inside a given concrete structure; until the corrosion has progressed to where visible damage such as spalling has occurred1,2. Therefore, a persistent need exists for developing non-destructive, in-situ and cost-effective techniques for this purpose. Optical fiber based techniques are capable of non-destructive testing (NDT) and in-situ measurements and offer several other at- tractive features such as: small size, flexibility & geometric versatility, light weight, inherent immunity to electromagnetic interference/ radio frequency interference (EMI/RFI), low-loss, large bandwidth, increased sensitivity, resistance to corrosive & hazardous environments, non-intrusive, distrib- uted sensing and compatibility with optical fiber data telemetry systems. They are currently being investigated to address the problem of high service costs involved in maintenance of structures. Corrosion of critical steel mem- bers that are not easily subjected to visual inspection are excellent candidates for embedded or surface attached fiber optic corrosion sensors. A distributed and on-line optical fiber sensing system for corrosion detection will help reduce the costs associated with corrosion damage and extend the lifetime of existing assets.

Fiber optics based direct absorption spectroscopic techniques investigated by some research groups, for the estimation of corrosion, have relied on using single fiber elements for recording the light signal reflected off the specimen at different wavelengths3. The present paper reports a simple and alternate technique based on the color matching principle of colorimetry for detection of corrosion induced color changes. In contrast to single fiber elements, the present technique employs a thin Y-shaped optical fiber lightguide/bundle, which increases the quantity of light energy coupled from a white light source resulting in ease of optical signal processing. The technique actually involves detection of the brown color of rust formed both on steel members and reinforced concrete in structures. For applications where accuracy requirement is not very critical, fiber optic sensing techniques based on multimode optical fibers/bundles and components offer the advantages of simple construction, enhanced coupled optical power, ease of alignment and implementation with cost-effectiveness.

PRINCIPLE

The principle of colorimetry is well-established but by owing to the use of fiber optics, this technique has now become quite versatile and useful. In fact, the utilization of optical fiber technology for color sensing represents a major advance in visible/ near infrared (IR) photometry. It is now possible to perform in- situ/on-line non-contact color measurements of samples4. Through one arm of the optical fiber Y-guide, white light coupled from the source is made incident on the sample surface to be monitored for color changes and the reflected signal is carried by another arm of the Y-guide to a photo-detector. A complementary color filter is used before the photo-detector because color of the light absorbed is complimentary to the color of the corroded specimen and it is the intensity of the complimentary color that varies with color concentration of the sample. The accuracy of color measurement using this technique depends on the proper selection of the complimentary filter. The reflected light signal consequently varies in proportion to the corrosion-induced color of the sample and, thus it is possible to assess the extent of corrosion damage of the surfaces using this technique.

The technique is based on the principle of light being refleeted off the target surface and then received back through a receive branch of a bifurcated fiber optic bundle (Y-guide) onto a photodiode. Keeping the displacement between the common end of the bundle and the target specimen surface as fixed and equal to that for the peak sensitivity value, the quantity of light reflected by the surface gives a direct measure of the state of corrosion of the surface and it is displayed in the form of an electrical signal.

In the bifurcated fiber bundle developed and employed, one group of fibers transmits light from a stable source to a target reflector surface. The other group receives the light reflected from the target and transports it back to a photodetector which converts it to an electrical signal proportional to the intensity of the reflected light as depicted in Fig. l(a), while Fig. l(b) depicts the action of an adjacent pair of fiber optic light transmitters and receivers as used to detect the position and or quality/condition of the surface relative to the ends of the optical fibers.

As the reflecting surface moves away from the set of optical fibers, the area (A) illuminated by the transmitting fiber optic element becomes larger and larger. The illuminated area of interface (Bl), which is providing light to the surface of the receiving element (B2), also grows increasingly larger. There is rapid and linear growth in the signal output as more of surface (C) is illuminated (e.g. as the probe to target displacement increases, increasing amounts of light are captured by the receiving fibers.) This portion of the response curve is referred to as the 'front slope' region. The point at which the entire surface (C) becomes covered with light is referred to as the "optical peak" point of the response curve as shown in fig. l(c). This is a well-known experimental response curve of Y-guide type reflective sensors, which are quite popular amongst the intensity modulated fiber optic sensors.

As the surface moves further away, the size of area (B2) becomes larger than area (C) reducing the detected light intensity. The light intensity-monitoring photosensitive detector output then begins to decrease, resulting in the back slope characteristic where the signal decreases proportional to the square of the displacement between the probe and the target surface. A continued increase in displacement will cause the diverging field of reflected light to exceed the field of view of the receiving/detector fibers, which produces a reversal in the output vs. displacement signal relationship. This also becomes clear from the geometry of the arrangement used. As the target is moved away from the probe, area of the illuminated spot on the target increases in direct proportion to the square of displacement while interception of the reflected light by the fiber probe correspondingly decreases in proportion to the square of displacement. Actually, these displacement dependent measurements of light coupling are based on the interaction between the field of illumination of the transmitting/source fibers and the field of view of the receiving fibers5,6.

Fig. I (a): Arrangement of an optical fiber Y-guide (bifurcated bundle) for reflection modulated fiber optic sensors

Fig. l(b): Reflected light monitoring mechanism of adjacent optical fibers

Fig. l(c): Optical fiber Y-guide transducer response curve

The extremely rapid signal rise in the front slope region of the response curve permits high sensitivity measurements while the back slope region is used for measurements at greater displacements, where sensitivity, linearity, and accuracy requirements are less demanding. The "optical peak" region is used for optical inspection and comparison of surface conditions since at this position the output signal is more sensitive to light intensity variations than changes in displacement. The distribution of fibers at the common (distal) end of the bundle is a major factor determining displacement range and slope sensitivity of the probe. For example, greater displacement sensitivity is obtained with a staggered/ random fiber array7-9.

EXPERIMENTAL DETAILS

The experimental arrangement realized in the laboratory for detection of corrosion in structures is schematically depicted in Fig. 2. The set up employs an indigenously built thin Y-sha\ped fiber optic bundle with half portion (a semi-circle) of its common end made from the source fibers and the other half from the detector fibers. The length of the fiber lightguide used is about 0.5 m, with both the source and detector arms being 15cm long. The diameter of the common end of the lightguide is 5mm while the source arm and detector arm end faces have a diameter of about 1.8 mm. This light- guide was developed indigenously using optical glass fibers of 60/ 120 microns in size. Light from a 50-watt quartz halogen lamp driven by a 12V (DC) stabilized power supply, was coupled to the source arm of the bundle using suitable optics.

Fig. 2: Experimental schematic of in-situ fiber optic colorimetry technique for detection of corrosion

The corroded specimens were prepared in the laboratory by dipping 25mm 25mm size MS sheets in water solution containing varying concentrations of sodium chloride (NaCl). NaCl was added to the water to speed up the process of corrosion and a 0.5% NaCl solution means 0.5gm of NaCl was dissolved in 100 cc of water. The specimen is located very close to the common end of the Y-guide at a displacement of 5mm which corresponds to the peak sensitivity of the response of the lightguide and the reflected light picked up by the other arm of the bundle is made to pass through a complementary filter (Greenish-Yellow) before it is incident on a photo-detector. The complimentary filter used in the investigation is a commercially available interference filter with a typical bandwidth of 15nm. The detector signal is processed and displayed by a high sensitivity optical power meter. Corrosion measurement on the MS specimens was thus carried out both for ambient and concrete . embedded conditions10.

RESULTS

Data was recorded for a fixed displacement between the fiber end and sample surface both for the case of corrosion under ambient and embedded conditions and the results were found to be reproducible. A typical variation between concentrations of NaCl in the water solution and the reflected optical power is shown in Fig. 3. It clearly indicates that at low concentrations, there is a linear increase in absorption (e.g. a decrease in reflected optical power) of the complemen- tary color with an increase of concentration of NaCl, which tries to saturate at relatively higher concentration. An increase in NaCl, concentration means more corrosion in a given time period. Also, on highways and bridges, it is the penetration of chloride used for deicing in the structures that causes corrosion and its state can be assessed from mea- surement of chloride ion concentration present in the structure11 and or by employing the present colorimetry based technique.

Mild steel sheets were embedded in a concrete slab prepared by using a mixture of cement, concrete and a small amount of NaCl solution with varied concentrations. The state of corrosion of the embedded specimens was monitored at regular intervals and the typical trend between reflected optical power and the state of corrosion of the specimen, as depicted in Fig. 4., was observed to be identical as in the ambient case. For the case of embedding, the common end of the fiber light guide located close to the target surface was covered with a teflon sleeve with a step in diameter so that at one end it fit tightly on the fiber bundle's common endface. After the endface, the diameter is wider so it is important not to restrict the cone of light emerging from the bundle, which becomes incident on the target surface.

Fig. 3: Typical variation of reflected optical power with concentration of NaCI solution for corroded MS specimens

DISCUSSION

This simple technique, based on the principle of colorimetry employing fiber optic bundles, enables in-situ monitoring of the onset of corrosion and its further growth in structures. Such colorimetric fiber optic probes can be embedded into surfaces or attached to the structures at pre-determined sites for conveniently monitoring the state of corrosion from a remote control station. The technique can be easily calibrated using standard corroded specimens. However, due to non-availability of standard corroded specimens, such a calibration could not be undertaken. This technique is very useful to study the kinetics of corrosion and further investigations are planned for the future. The technique can also be used for remote/in-situ monitoring of the state of corrosion of interiors of boilers and other critical structures located in harsh environments such as: the chemical industry, nuclear reactor buildings, mines etc.

Fig. 4: Typical variation of reflected optical power with embedment period for a concrete embedded MS specimen

ACKNOWLEDGMENT

The authors are thankful to Mr. Sushil Kumar, Mr. D.P. Chhachhia, Mrs. Asha Kumar, Mr. G.G. Poddar & Rajneesh Talwar for their cooperation in this work. Thanks are also due to Mrs. Madhu Mehta for typing the paper manuscript. This work has been carried out under the Indo-Italian Programme of Co-operation sponsored by DST, New Delhi. The authors gratefully acknowledge the financial assistance provided by DST for this programme.

References

1. Huston, D.R. and Fuhr, P.L., Fiber Optic Smart Structures (ed. Eric Udd), John Wiley & Sons, Inc., New York, pp 647-665, 1995.

2. Miller, M. "Optical Fiber Based Corrosion Sensor System," Doctor of Philosophy Dissertation, The Bradley Deptt. of Electrical Engineer, Virginia Tech, 1995.

3. Fuhr, P.L. et. al., "Fiber Optic Corrosion Sensing for Bridges and Roadway Surfaces," SPIE(USA), vol. 2446, pp 2-8, March 1995.

4. Singh, N. et.al., "Development and Experimental Investigations of a Fiber Optic Color Sensing Probe," J SCI IND RES, vol. 58, pp 359-363, May, 1999.

5. Intensity Modulated Fibre Optic Sensors (Chapter 23), "Fundamcntals of Fibre Optics in Telecommunication and Sensor Systerns," Ed. Bishnu, P. Pal, Wiley Eastern Limited, New Delhi, 1992.

6. Hoogenboom, L., et.al., "Theoretical and Experimental Analysis of a Fibre Optic Proximity Probe," Proc. SPIE Fiber Optic and Laser Sensors II, vol. 478, pp 46-57. 1984.

7. Davis, C.M., "Fiber Optic Sensor: An Overview," Optical Engg. 24(2), pp 347-351. 1985.

8. Kissinger, C., Measurement and Control (M&C), April 1988.

9. Lawrence, L.J., et. al., "Non-Contact Displacement and Vibration Measurement Systems Employing Fiber Optic and Capacitance Transducers", ISA Reprint, Presented at the 23rd Int. Instrumentation Symposium, May 1-5(1977), Las Vegas, Nevada, USA.

10. Singh, N. et. al,, "A Simple Fiber Optic Technique for In- Situ Corrosion Sensing in Structures," SPIE (USA), Non-destructive Evaluation of Aging Materials and Composites, Vol. 3993, pp 201- 205, 2000.

11. Fuhr, PL., et. al., "Embedded Fiber Optic Sensors for Bridge Deck Chloride Penetration Measurement," Optical Engineering, vol. 37 (4), pp 1221-1228, April 1996.

N. Singh, S.C. Jain, N.V. Aulakh, J.K. Chhabra, M.L. Singla, A.K. Aggarwal and R.P. Bajpai are affiliated with Central Scientific Instruments Organisation, Chandigarh, India.

Copyright Society for Experimental Mechanics, Inc. Jan/Feb 2004

More News in this Category