The sensing properties of the humble optical fibre lends its use to measuring strain and temperature variations along structures like oil pipelines stretching hundreds of kilometres, as Greg Blackman finds out
The optical fibre might have transformed the telecoms industry, being faster than copper cabling, but its capacity as an optical sensor is proving highly useful in detecting strain and temperature variations in buildings, oil pipelines, dams, and bridges, and even aircraft are incorporating fibres to monitor the integrity of the body, engine and wings.
The sensing capabilities of the humble optical fibre were originally discovered because of non-linear effects detrimental to the signal. While troublesome for transmitting data, the effects, essentially stimulated Brillouin and Raman effects, turned out to be valuable for sensing because of their dependence on temperature and strain. Brillouin was found to be sensitive to both temperature and strain, whereas Raman is sensitive to only temperature. This means you can lay an optical fibre, the same optical fibre now ubiquitous in the telecoms industry, over the length of an oil pipeline, say, stretching 150km and use it to monitor temperature or strain along its length.
This is distributed sensing, whereby every section of the fibre can be interrogated and changes in temperature and strain detected according to the light scattering within the fibre. This property of optical fibres makes them good long range sensors, where writing fibre Bragg gratings (FBGs), the alternative method of probing an optical fibre for its response to certain physical effects, would be impractical over these sorts of distances.
The Optoelectronics Research Centre (ORC) at the University of Southampton carries out work on distributed sensing, one of its projects being R&D work for the Sensor division of Schlumberger to produce sensors for BP and National Grid. ‘We recorded sensing capability over 150km, as BP wanted to monitor the pipeline between two oil stations stretching 100-150km,’ explains Dr Mohammad Belal, a research fellow at the ORC involved in the project. ‘You can’t write gratings over these distances and, even if you could, it’s a quasi-distributed sensor with dead space [between the gratings] where you can’t interrogate.’
The distributed optical fibre sensors developed at ORC can ascertain temperature differences at an accuracy of 2°C over a sensing length of 100km and with a spatial resolution of less than 10m. This, however, is over long range, making static temperature and strain measurements. The researchers at ORC are also using optical fibres to measure dynamic effects, such as mapping a moving train, for instance, over short range (less than 1km). In this instance, Rayleigh scattering within the fibre is used to detect dynamic disturbances. The sensors could be installed in structures such as dams, bridges, and railway tracks providing real-time monitoring of the structures’ integrity.
Returning to the BP project, ‘great care has to be taken when transporting oil from a reservoir under the sea to the surface or delivery point’, explains Dr Belal. ‘You want to resolve where focus stresses and strains are taking place, which could result in rupture.’ The pipelines have multiple joints, at which more information is required at a greater resolution. Therefore, the distributed sensor passing along the length of the pipe is couple with a fully discrete fibre Bragg grating (FBG) sensor at the joints.
An FBG sitting on the joint of the pipe will be calibrated according to pre-induced stress or strain, which acts as a background measurement. ‘The FBG is so sensitive that the smallest swings, due to ocean swell or pressure differences from oil moving along the pipe, would influence the Bragg wavelength associated with the grating,’ states Dr Belal.
Individual fibre Bragg gratings are discrete sensors. Quasi-distributed sensing involves writing FBGs at set intervals over a short distance of fibre, providing excellent strain and temperature resolution.
Gratings are essentially a refractive index contrast in the fibre, analogous to opening blinds on a window to let sunlight in, (the blinds are high contrast, while the gaps are low contrast). The grating is written by essentially washing off layers of glass to achieve the refractive index contrast. The grating therefore weakens the fibre in a way by forming corrugation, which makes it sensitive to the tiniest physical deformations, be that thermal or strain related. They are written for a specific wavelength of light, with any disturbances in the fibre changing the wavelength passing through the grating, which can be measured.
‘FBGs give unrivalled resolution capabilities,’ says Dr Belal, ‘in the order of micron or nanometre levels of disturbance.’ ORC has written gratings up to approximately 1m in length, but anything beyond that has to be concatenated, which will always result in a dead space, hence the term quasi-distributed sensing.
FBGs are also used for medical purposes, with the devices placed inside the human body to ascertain damage to tissue. In addition, a radiation-hard fibre can be used to wrap around a nuclear reactor to monitor its structural integrity over a period of time.
‘Fibre is not only being used to send information from one point to another, but at the same time to probe what is happening along the fibre’s length,’ comments Dr Belal. ‘The advantage is that while the fibre is being used to carry information, it is also used as a sensor.
‘This is in some ways an unrivalled technology,’ he continues, ‘because its contemporaries are electronic sensors, which are affected by electromagnetic radiation. Fibre optics is a passive technology with regards to electromagnetic radiation, which is why the sensors are very attractive for the nuclear industry, because of the electromagnetic fields present there.’
To acquire very accurate or high-resolution measurements for small changes in temperature requires a detector that can monitor very small alterations in the light signal – the detector has to have a high signal-to-noise ratio. Hamamatsu provides detectors for optical fibre sensing, including near-infrared enhanced avalanche photodiodes (APDs), which, according to Jack Bennett, technical sales engineer at Hamamatsu, are excellent detectors for optical fibres.
The APD is made of silicon, which typically is sensitive in the visible spectrum and doesn’t extend into the NIR. However, a microelectromechanical system (MEMS) structure is fabricated on the back of the detector, which gives it much higher quantum efficiency or sensitivity in the NIR region (up to 1100nm). ‘Depending on the specific wavelength, the detector is between two and three times more sensitive than a conventional detector,’ states Bennett. ‘In addition, the APD has a much higher gain (a gain of 50 or 100) than a traditional photodiode, as it is operated under high voltage.’ Once the charge is created in the detector it accelerates down the voltage and knocks off electrons resulting in an avalanche effect and a higher gain. The fact that it’s an APD combined with the NIR sensitivity means it provides a much larger signal than usual.
‘When selecting any detector, the most important parameter is the signal-to-noise ratio – you do want low noise, but you also want high signal as well,’ explains Bennett. ‘These APDs are made of silicon as opposed to InGaAs, which is typically used for the 1µm range, but a much more expensive material. Silicon is much cheaper and it also provides high sensitivity and low noise.’
ORC is also developing new fibre designs, such as photonic band-gap fibres, which have 30 per cent reduced latency, making them ideal for transporting large volumes of data in near real time. Particle physics laboratories like CERN, which generate massive amounts of data, or financial institutes, might use these microstructured fibres for high-speed data transport.
With regards to sensing, the holes in the fibre can also be filled with material sensitive to certain properties. ORC has been working on putting magnetostrictive material inside the fibres, a substance that deforms in response to a magnetic field. Any deformation would change the shape of the fibre and therefore alter the light signature travelling through it.
ORC also has the capability to fabricate micro-fibres, optical fibres that are tapered down to widths of 10nm, in some instances (the scientists can achieve a fibre 150nm wide over 5cm). ‘By tapering the fibre, the sensor becomes very compact and also much more sensitive,’ says Dr Belal. Tapering a fibre down to a certain level causes light to propagate outside the fibre, which increases its sensitivity, although also makes it more difficult to handle. ‘Because the light is accessible on the boundary, you can do a lot of interesting sensing with it, including gas sensing where you’ll get an absorption profile from the light travelling on the surface of the fibre,’ he adds.
‘We’ve also made resonators where the light circulates around a fibre loop multiple times. Therefore you can get ridiculously high sensitivities based on micro-ring resonators.’
The tapering technique can also be transferred to specialised fibres to improve their sensitivity, although the economics for these are less viable compared to sensing with conventional optical fibres. Nevertheless, work on photonic band-gap fibres could see future high-speed networks and throw up unique sensing properties.
The range of devices and applications using optical sensors is huge, from ambient light control in homes and offices to gas sensing to the latest car models with enhanced sensing capabilities. Volvo recently demonstrated a convoy of self-driven cars, which journeyed 125 miles on a Spanish motorway. The cars were wirelessly linked and mimicked the movements of a lead vehicle essentially to drive themselves. Audi will be integrating 3D time-of-flight sensors from PMD Technologies in its future models for driver assistance and safety systems.
Stefan Schwope at TriDiCam, a German company producing time-of-flight sensors, explains that the sensors have to be robust to be used in cars: ‘The package and how you implement it and the electronics is more difficult for the automotive sector. The sensor could be a safety system, so you have to rely on it.’
Time-of-flight sensors generate 3D depth information by pulsing light, either from a laser diode or LED illumination, and measuring the time taken for the light scattering off objects to return to the detector. TriDiCam is currently developing sensors for both driver safety and for use within the car for things like gesture recognition to control a navigation system, for instance. Development is ongoing and Schwope believes that commercial vehicles with these sensors won’t be on the roads until around 2015.
Fuel and exhaust monitoring is another area where optical sensors are playing a role, with sensors integrated for real-time emission readings to monitor the performance of the engine as well as what’s coming out of the exhaust. Hamamatsu has developed fuel-quality sensors, linear arrays and infrared LEDs, photodiodes for occupant detection, and photodiodes and avalanche photodiodes for position and distance control through lidar. ‘We have tens of millions of devices operating in the field already built into cars,’ states Jack Bennett, technical sales engineer at Hamamatsu.
‘All devices that go into cars have to meet strict reliability standards and the sensors need to be rugged and reliable, but still inexpensive,’ he adds.
Elsewhere, pyroelectric sensors are used for gas detection as well as to control lighting in buildings, where they are able to detect body temperature and movement of occupants in a room and automatically switch lights on and off. Pacer, a supplier of optoelectronic products, offers a range of optical sensing devices, including pyro detectors, thermopiles and photodiodes. ‘The pyro has to be sensitive enough to detect very small, almost micro movements, because if the person is sitting still you don’t want the lights to switch off,’ explains Paul Knight, divisional manager of the opto and sensors division at Pacer. Sensing applications use a mix of optical technology that includes the pyro as well as some cleverly designed lenses that zone IR sources and provide for both micro movements while at the same time covering the entire scanned area.
Knight points out that modern digital pyro detectors can now be surface mounted within the PCB, which makes the devices much more compact than previous through-hole detectors. They can therefore be integrated into a range of portable equipment.
Knight adds that health monitoring and remote monitoring of the elderly or those at risk is another application for optical sensors. A thermopile sensor can be used to monitor a patient’s body temperature, while an LED and a photodiode can monitor oxygen levels in the blood, called pulse oximetry. Health monitoring, according to Knight, is a relatively new market that has ‘only recently reached a kick-off point’ due to the availability of technology that’s more easily integrated within smaller devices.
Optical detectors made from novel materials are also being developed for extended sensing range, but applications for standard pyro detectors and thermopiles remain wide-ranging.