 My name is Alder Sosa and I came here today mainly to share a perspective of a bridge owner about a long-term monitoring system that is employed in one of the longest bridges in Europe that is the Lazare Bridge and I came here also today to focus mainly in the past experience current states and Twitter challenge related to this monitoring system. This presentation is going to follow. I will try to just present Brizza, who is Brizza then present the long-term monitoring system of the bridge and then present you what is actual performance after nine years of operation and finally the steps towards a more efficient management of the Lazare Bridge based on this monitoring system and I conclude this presentation with the main conclusions. Who is Brizza? Brizza is the major highway agency in Portugal and is one of the largest total motorway operation in the world with concessions in the United States and operation in the Netherlands and India. On the other hand, this is the largest transport infrastructure group in Portugal, mainly focusing on the asset management of roads and railways. They have some relevant investments in ACGM, mainly in two bridges, the Saraje Bridge and the Lazare Bridge which is near the capital Lisbon and the main focus of this presentation is the Lazare Bridge, which is this nice photo. As I said, it is one of the longest bridge in Europe. There are a lot of sensors on this bridge. So briefly the structure is composed by main three sub structures, so the approach viaduct and the main bridge. The longest structure is the sub viaduct with almost more than nine kilometers. The solutions are quite different. The main bridge is the Cantilever Bridge, the south bridge, the south viaduct is a partial pre-crafted solution. The foundations are very similar and as I said, the construction process was quite different for each one. Now regarding the monitoring system, well briefly, there are a lot of parameters that are being measured. They are represented and I will try just to show you how this is implemented from a bridge owner perspective. So if we focus on the sensor system, so the measurements, we have in fact different subsystems in this sensor system. We have a static acquisition system that is devoted to the long-term performance of the bridge and then we have some complementary subsystems, a dynamic one and optical one, to complement information that is collected by the static acquisition system. And finally, all this information is collected, centralized in the acquisition of the bridge, sent to the database, hosted at BRISA. But I will say that from a bridge owner perspective, it is also important to contextualize what is this sensor system. And now if this is integrated, in fact, so what I presented before is the sensor system, all sensors that are installed there. But in fact, we have more subsystems, which is below the power supply, above we have the network communication and finally the database and data processing. This architecture is very important from a bridge owner perspective because if there are problems that are related to one of these subsystems, for them it's easier to identify the expertise that is necessary to solve those problems. Now, regarding the assessment of the bridge, based on these collected measurements, so as we see here, we have all these parameters. I describe here the sensor type, the acquisition systems that are used, the measuring frequency, what is the proposal of each measurement. But mainly these measurements are being compared with the thresholds that were previously defined by the bridge designer. Nevertheless, there are some parameters that were not defined yet, mainly strain gauges, because this was decided that it was necessary to follow the first years of the bridge to be able to define those thresholds. So for this, it was developed under my PhD, a refined fit and element model for the bridge. So I will present mainly for the main bridge, so this I'm not presenting all the analysis for the bridge, and I want to present you with the many results that were collected. So these are measurements during the construction. It's a strain gauge located in the top layer of this section, near of the pyre, and overlap here are the numerical results, which match very well with what was observed. And in addition, after that, for the loading test, it was excellent to calibrate, not calibrate, because the model was not calibrated. What we did was to implement in the model the real materials, properties, the time history, the environment conditions, and also shrinkage and creep properties measured in specific specimens. And what I want to show here is that the curvatures of these sections, what is something that is quite difficult to match, were very well predicted by the numerical model. And finally, if we enlarge even more this window, we see here, so this graph is the next extension of this one, but now for the following years, and we see that the challenge here is to predict these strengths that are mainly due to shrinkage and creep effects. Here the results match very well, but there are cases that this was not so straightforward. So regarding how to show you all this, the condition of the system after nine years, because this kind of information is scarce in the literature. I just want to share with you that after nine years of operation, the system is in very good condition and without signs of analysis. I just to show you pictures that were collected in 2007 and the pictures that were collected last year. And you can see that the condition of the system is very, very good. Now regarding the function of the embedded sensors, this is also something that we don't find too much in the literature, is that after nine years of operation, we have 93% of them operational. But we are aiming to put it in a 97% performance, because there are some sensors that are not working, because mainly due to the fiber optic sensors, due to the scheme that is used as a serial system, if one fails, the remaining ones in the way that you measure cannot be reached. So this will be recovered. And about also the quality of the collective measures, because we cannot do analysis if you don't trust in the measurements. Based on these nine years, we will say that the most reliable measures that I see in this system are the vibrating wise frame gauges, the thermometers and fiber optic sensors. They show you very stable and very reliable measurements along time. Regarding external sensors, I will say that we have some problems in terms of maintenance, because they are exposed to the degradation. And mainly the sonar altimeters that are in the river, due to the hard conditions that they are exposed. Well, once we have this system in good shape, we have a model that predicts nicely the long-term behavior. There are some challenges that we need to face in the near future. From the bridge general perspective, definitely these kind of systems need some maintenance. As I said, the systems that I described before, here are some maintenance actions that are required for each one. And therefore, the bridge owners need to be aware that this, in fact, is not putting a system. Okay, this will be working and I don't need to be bothered with that. No, we need it. And mainly that is something that is necessary, that is updating the data processing routines continuously. For a long-term system, the advances in the science and in the technology, this is fundamental. Now, regarding more in a research perspective, I present here again that result that I present in the previous. So I see here the trends that are predicted very well with the measurements. But nevertheless, there are some challenges here to be faced, because depending on the models that are used for shrinkage and shrinkage, because these trends are mainly due to shrinkage and kip effects, these trends might change significantly. And so, if you are talking about here five years, but you want to predict 100 years, these trends will produce different thresholds. Okay, so this is mainly due to poor materials models for kip and shrinkage. There is no agreement in terms of which one is better and even the results, the values that they give. There are different rates of shrinkage and kip, because depending on the element thickness, this changes a lot. Numerical model simplifications, always. And also absence of monitoring approaches that are able to map shrinkage and kip deformation. Okay, as strain gauge measure a punctual deformation, but for analyzing this problem properly, we will need something that will be able to map instead of a punctual measurement. And finally, and I'm concluding, recently we won a European project, the Brees and the University of Surrey. And so we want to bring together two worlds that are not, don't communicate too much, that is structure modeling and reliability modeling, and benefiting from monitoring data, as this case study is able to give. Finally, as conclusions, I will have three main messages. One for providers of usage systems, that I say that monitoring projects as part of the Brees design tasks, leads to optimized solutions. The system architecture design, in the perspective of the Brees one, is very important, so it leads to research optimization. And monitoring systems are complex like this one, this example. For the Brees owners, I would say that their feedback and their satisfaction is vital to succeed in the implementation of these kind of systems. And also systems like this one that they expected to work for the lifetime of the Brees, the maintenance is also very important. And finally, and a contribution for working group two, I would say that the integration of physical and mechanical with probability models, by benefiting from ACGM to support asset management throughout the pressure lifetime, mainly creep and shrinkage indicators, based on point and section strength profiles. Thank you very much. If you want to find some additional information about this work, there are a lot of publications. Sorry.