 So, without further ado, we're going to continue with the session on sensors. Again, please keep adding your questions to the Slido. So, happy to welcome Ali Massoudi to talk about how we're measuring vibration. Great. Thanks for the introduction. My name is Ali Massoudi. I'm from the University of Southampton Optoelectronic Research Centre. So, in the theme of sensing systems that can be used to monitor environments. So, this is a new kind of technology which not many people are familiar with. So, my aim is to first introduce and apply, first introduce what distributed optical fiber sensor is, what the technology is, how it works, classification, and then I'm going to talk about the conventional distributed acoustic sensor. The naming is important, as I've explained, and then the application of them, and then the new generation, the invitation of the conventional band, then the new generation with a high resolution distributed acoustic sensor, some experimental results, and then results. So, most of the people in this here are familiar with the point sensor, so when you have a point, temperature sensor point, the shrinkage or point like microphone to measure one, the data from single point. So, if I explain that in a conventional system, so we have an interrogator, we have an interrogator. So, let's say, I usually give an example of a pipeline, the fiber is used to take the lights from the interrogator to a point that you want to do the measurements, and then you do the analysis, the data goes back and it gets analyzed, and then you do the measurement. So, if the pipeline has a temperature distribution of this dashed line, you get a single point. Perfect if you know where you want to do the measurements, but if you want to map the whole pipeline, that's the idea. The second concept is was distributed sensor. So, again, some people might be familiar with the concept of fiber bracketing sensors. So, basically, you have you modify the properties of the time at the few locations, and then you can do the measurements at those locations temperature gain vibration, magnetic field, so on so forth. So, similarly, again, you need to know there's a, you can do the measurement at three decimal points, because you need to know where you're putting your sensors so again, if you have this temperature distribution, you put it at few positions and then you can measure those points. But if, in this case, if you have a whole spot between the two spots on your pipeline, then you miss that because you haven't put it at one point. Then we have the concept of a fully distributed optical fiber sensor, which basically use optical fiber to map your measurements all along the fiber with a set spatial resolution. So if you have a one meter spatial resolution, you have 10 kilometer of sensor optical fiber, it means that you can measure 10,000 points with a single fiber from a single point by accessing only a single end of the fiber. And then you can map the whole distribution of your measurements, it can be vibration, temperature, strain, so on so forth. So, my expertise in distributed acoustic sensor, so what you expect at the output of a distributed acoustic sensor is what we call the Waterloo plot or a spectrogram, which shows the vibration. And the main acoustic has been established for unfortunate reasons, it is a vibration distributed vibration sensor what we will start calling it distributed as acoustic sensor that which is easier to pronounce. So basically at the output, you expect to see the location of the vibration, the length of it, and the frequency and the phase of those vibration, and you should be able to, especially resolve those vibrations all around the fiber. So again, going back to the pipeline. So if you have a leakage on the pipeline, if there's someone walking on the pipeline, someone's digging, you should be able to use that as a, like as a strange as a microphone. At those points, measuring the strain that vibration imposes on the fiber, the frequency, the amplitude, all the features. And if you do the Fourier transform, you can do the spectrogram so you can look at the frequency components of those vibrations as we need to distance. And of course, because it's a real touch system, it's a dynamic system, if there's a moving objects in a pipeline, it's the peak going down the fiber, you should be able to have this diagonal line which shows the location of that, the speed of it. And it generates so on and so forth. The principle of cooperation of this system is quite simple. So you have your pulse of light, if you send out the fiber, let's say you have a cross section of the optical fiber. It's much smaller. So the fiber is like 250 marker in diameter. So I just zoomed in. And then we have in homogenesis in the fiber. And when the pulse goes in the fiber, it interacts with these in the genities. And then it, like, gets reflected. And what we do is we look at the means of the lights, scattering from two section of the fiber separated by distance L, which can be a meter, two meters, 10 meters, depending on your spatial resolution. And you can see that if the lights coming back from like the blue and the orange section are in phase, you get a constructive interference, and then you get a high intensity light. If the fiber is stretched, then you send the next pulse and then you compare the face coming from the blue and the orange section, you get a kind of destructive interference. And then you get no light and anything in between. So if you have a vibration and you keep sending pulse and measuring the state of these two sections, you can measure the vibration at this point of the fiber. The same principle you can repeat that over and over. So it's from this point and then you get to the next point next point and then that way you can map the vibration all along the time. And the equation is quite simple. So, as you can see, it's going there. So the face difference between the blue and orange section is given by the fight, which is at this city senior, which is a function of L. And then you send the next boss at different time, and then again you measure the face difference between the blue and red, and then if it's eroded by death L. So if you do this equation, and if you subtract the face difference from time to time, you get a change in the elongation of the fiber as a function of a city. I come to that error later on. In addition to this, we've done a few tests using the system, say one is the obvious one. In this one, we've done it with the fiber down to the geode steps to that. And then we've done the surface waves so we're excited to see how the surface waves, very, very long ways, how they travel on the surface. We've done this in Japan with the subsequent measuring tsunamis measuring earthquakes measuring geophysical safe molecules activities. We put the fiber like telecom fiber next to the road to measure traffic. We put the fiber on the rail to measure the deflection of the beam the condition of the track the condition of the balance subsurface. This is one of these lines is one wheel of the axle of the train. Either last one of this we can put it in a subsea high voltage cable to measure the deflection if there is anchor drop if there is any activity is going in the cable monitoring the health of that. And then in the other discussion last one for the wave monitoring that's similar things that can be done so with the test that we did with input to Japan, when the wave comes in, the fiber was 10 meters subsea was buried and then we can see the waves coming and everything monitor the speed the location, everything about the base in this format so in this class you can see again we have distance time, you can do a frequency get all the information at those points. The issue with the conventional DC with a sense that acoustic sensor is that when you write on this in homogenities. There are two issues one is the signal is extremely weak because this in homogenities are very small. So you are limited the special result of plus one meter so we can resolve a stream every meter, not the final resolution. The second issue is that because these in homogenities are randomly distributed when you stretch the fiber, you discard the location of them and then you get this phase error in your measurements again. And that causes, you cannot have a very fine strain measurement. To address these two issues, we start to work on a new type of fiber which we use a femtosecond laser, they focus in the core and we change the properties of the fiber at the fix location, which again those can be fixed, we keep putting this femtosecond laser and we have this kind of point reflectors at the fixed distance along the fiber. And then if you look at the principle is the same you send the possible like we get a reflection from these two reflectors, you measure the phase based on that we can measure what is going on the fiber how much the five is elongated in this case we can see the phase error function is a linear function of the L, there is no phase error factor, so it is a lot more precise in terms of the measurement and also because we rely on the detector, we can have them space much closer with the simple pulse peak power. So basically this is a setup that we use to put reflectors in the core, I don't go into details, I think it's kind of boring for you, and you can see the reflectors every 10 centimeter in this case along the fiber. And then this is the experimental setup. This part is basically responsible to generate the pulse, so we have a laser we modulated, we have SOA to have a higher extension ratio. So we have a pulse, it goes on the fiber, these are reflectors in this case we put the piezoelectric actuators which we can stretch and compress the fiber. And then we have this detector part, which is basically the reflector right come back in after the pulse gets into half and then we mix the light again to measure the phase between any adjacent detector and based on that we do that analysis. So this there is also when we do the sinusoidal vibration and PZT you can see again the location of that actuations, the amplitude of it, the frequency of it, if you do the 2D cross section of this, you can see the sinusoidal wave, and then the special resolution and this is the FFT, sorry, this is the FFT of that. And then you can see the special resolution how well we can resolve those vibrations. So if you have very high resolution, I think the number that we are achieving is trying to push that down to five centimeter resolution and strain wise we can do soft nano strain. And probably to reach the hundred people strain or load so it is extremely sensitive for measurements both in a normal and high resolution that's. And so in conclusion, I explained the classification of the distributed different kind of optical hypersensors show the principle of the back system. The applications, the limitation of conventional that system and then how we can have an higher resolution that system by inscribing reflectors in the core of the fiber and having a high resolution distributed acoustic sense. Just fine as a plug, we have this system we develop it as a like a semi commercial system so if anyone wants to acquire wants to do the collaboration test the system see how it works. I mean, it should be the tiring stuff. Thanks for your attention. Thank you so much. It's really fascinating stuff I only learned about last night that even existed. I think, could you perhaps the sort of like really obvious industrial applications and infrastructure for some from an ecological point of view and environmental health point of view. Is it possible to use something like that. And this question came from Tom or just look for me. Is it really possible to listen to sounds like movement of animals or how sense to the sound is it is talking about spatial resolution that must be related to the size of that signal as well. Yeah, so at the moment you're working on. So that depends on the density of the meeting. So we try to use the system and we have used the system to to listen to mammals like Subsea mammal so we can listen to like whales with Dr. If it is the environment so it's a space like this in principle we are working on a way to wrap the fiber and make it as a microphone. And we can have many of them because you have this one you wrap and then you go next one next one, and then we can many of them and should be able to have like a hundred ones integrated signal from a single point. Very sensitive as it is doable, but it won't be distributed anymore it's going to be quasi distributed because you have to wrap the fiber and make it point to kind of measurements. But it's possible. I don't think I might think there was any question on the slide. But does anyone else in the room want to raise a question before I thank you. So the question is that can we make it agree? I'm just asking questions for the audience so the question is I can't be put the fiber in different direction. Yes, you can take it in direction to read because this is a 1D environment measurement. So if we do 2D or 3D, you can put it in a like in a in a grid pattern and we can have it 3D to have them the whole map distribution of this. That's possible. The fiber is like 250 mark so it's pretty flexible. I just wondered, do you keep the raw data or do you only keep the process data, the team signal that's been in terms of and how do you code that producing five volumes. Yeah, we try to try to keep the raw data so that anyone can come and later do their own analysis later on. But as you mentioned, it generates huge amount of data like a terabyte per day. So, yeah, sometimes you have to get rid of data. Thank you very much.