Thursday 22 September 2016

Conference Experience 2: CERI 2016 Galway

Galway conference was the first time in which I had to present my work. The conference was carried out in the School of Civil Engineering of the University of Galway on 29th and 30th August, covering several areas of study: Bridges, Environment, Materials, Concrete, Technology, Teaching and Learning, Geotechnics, Timber, and Structures. Depending on the session there were around 8 presentations. However, some of the topics were extended in many sessions at different times. The paper I presented was titled “Drive-by Bridge Damage Detection Using Curvatures in Uncertain Environments”. In this paper we studied the influence of two main factors that disturb the collection of data when using drive-by monitoring: road profile and noise. The conclusion was that the road profile does not have much influence in the final results. In contrast, the noise can spoil the shape of the curvature, making impossible either quantify or localise the damage. Further work needs to be done for improving noise removal.

Fig 1. Somebody that I don’t know at all answering questions. However, he seems really handsome!

On the one hand, the road profile is not a problem given that it is the same for all the sensors on the vehicle when passing through the bridge. The latter explains that deflections obtained by some sensors match deflections obtained from other sensors when using the Instantaneous Curvature (IC) defined in the Belfast Conference blog. On the other hand, we have noise measurements that, as they are random for each sensor, lead to curvatures that have nothing to do with the results that should be obtained. When using the second derivative of a function, the disturbances are larger than in the original added noise.


Eight TRUSS early stage researchers were there enjoying the sun (irony mode on) and the conference presentations (irony mode off): Sofia (ESR1), Shah Nur (ESR2), Alberto (ESR3), Rui (ESR4), Farhad (ESR7), Barbara (ESR8), JJ (ESR10) and me (ESR12). We will gather together in Dublin in early December again for the Mid-Term Review Project Meeting!


Fig 2. Part of TRUSS ESR group with Dr. Arturo González in the middle

We all enjoyed the conference, that counted with the participation of  many UCD researchers, like my workmates Paul, Alexandra and Paraic. Also the question man, Abdollah was there, with Longwei and many other UCD people. That made the conference to be a bit like home.

Fig 3. Main entrance of the building almost collapsed by conference people and students.

Thursday 15 September 2016

Conference Experience 1: ISHMII Belfast

Belfast conference in Queen’s University was my first experience in this kind of meeting. The conference was held on the 26th and 27th of May, covering several areas of the Structural Engineering:
·         ‘Sensors and Structural Health Monitoring Systems’,
·         ‘Structural Health Monitoring Strategies for Bridge Structures’,
·         ‘Geotechnical Monitoring for Civil Engineering infrastructure’ and
·         ‘Challenges in Practical Application of SHM Systems’.

In each of these sessions, eight presentations were included in addition to the keynote one. Our TRUSS paper titled “Instantaneous Curvature in Bridge Damage Detection” presented by my supervisor, Prof. Eugene O’Brien, was the keynote lecture for the ‘Challenges in Practical Application of SHM Systems’ session.























Figure 1. Prof. Eugene O'Brien presenting keynote lecture.


The paper consisted of using bridge curvatures calculated from the deflection and based on these curvatures, analysing several damage situations  with and without considering noise or road profile. Three points are needed to calculate curvature and a number of definitions of curvature is possible. Instantaneous Curvature (IC) takes into account several positions of the bridge for a fixed location of the force. Moving Reference Curvature (MRC) combines different positions of the force and locations in the bridge. IC doesn’t hinder the damage effect in curvature and allows its localisation. Using IC, we are able to obtain good results, when we are not considering noise, but when we do, even applying a simple denoise method based in repetition of the measurements, does not work properly.


Other TRUSS mates also attended the event: Farhad (ESR7), JJ (ESR10) and Antonio (ESR11). Farhad contributed with another TRUSS paper to the workshop. My project mates and also PhD students in UCD, Paul Fitzgerald and Alexandra Micu, together with the postdoc Enrique Sevillano also enjoyed the conference. We all felt that we had great fun and gathered a very valuable networking experience.

Fig 2. TRUSS people trying to return to life the conference sign after a wind attack.

Fig 3. Conference photo with all participants.

Tuesday 13 September 2016

1 year of PhD Student! Work Activities

On the 21st of August 2015, a Spanish guy first stepped into Ireland with no idea about how the work routine was going to be. After a few days settling in and sorting accommodation issues, the 1st of September that guy started to work. This means that it is the anniversary, and this post is going to give an overall view of what I have done until now. Other posts will be dedicated to cover individual topics in detail.

Four conference papers have been done so far: Tirana (Albania), Belfast (UK), Galway (Ireland) and Cape Town (South Africa). The topics show a development of the research plan:

·         SEMC2016 Conference in Cape Town introduced the ranges of data analysed when using different measurement methods,
·         BCCCE2016 Conference in Tirana showed the curvature method as a bridge damage detector,
·         CERI2016 Conference in Galway studied the influence of the road profile and the noise in the results and
·        CSHM6 Belfast Conference introduced case studies in order to verify the method.


Fig 1. Conference logos.

Learning activities that I have completed during this period include Endnote training, Academic Written English course, seminars and demonstrating to students. I have also helped in a community activity in Willow School with the marshmallow challenge in which we had lots of fun.

Fig 2. In the middle of an Endnote lesson for referencing with Shah Nur.

Finally, I made two visits to our Industrial partner in the project, i.e., Greenwood Engineering. There, I was able to see how this Danish company works and also collect some data information on how the Traffic Speed Deflectometer (TSD) operates. The TSD is intended to measure the road stiffness in order to detect damage. However, if we manage to make the TSD able to measure bridge damage, it will be a significant contribution for monitoring bridges in a cost-efficient manner.

Fig 3. Photo with the Italian TSD.


More posts about these topics will come soon!

Friday 8 April 2016

Contour plot applications: Why should you use them?

Contour plots are a very useful tool that can be used in several fields for showing information. As one image is worth one thousand words, the best way to define a contour plot is the following example:

Figure 1. Isobars map

Everyone has seen these lines when watching the weather forecast. Some lines appear on the map, and they have a certain value assigned to them. In this case the value gives us the pressure that can be found all over the line, which are called isobars. The closer the isobars, the stronger the wind in that region will be. For example a strong wind is expected in Lisbon. The distribution of the isobars helps to forecast some weather behaviours such as anticyclones and cyclones. The first ones are associated with stable weather and the second ones to unstable weather.
Coming back to the original question: what are the contour plots? The contour plots are graphs in which each line shows one value with relevant information for the problem in hand. In the weather forecast, the lines are telling us all the places that have a certain pressure. It can be easily seen that the pressure in Brussels is very similar to the one in Vienna, whose value is around 1012 millibars. Topographical maps also work with the same idea.
Now I am going to expose an example related to bridges which is the structure that I am investigating in my research project. We are going to create a contour plot in which a bending moment will be shown. The bending moment is defined as the distance between one support and the force multiplied by the force, as shown in Figure 2.



Figure 2. Bending Moment (M) definition

Bending moment is going to depend on the location of the bridge section being investigated and on the position of a force. For example, we are going to analyse the behaviour of a 20 m length (L) simply supported bridge for sections every meter, when having a 20 kN force (P) also at every meter. We start putting the force on the left support, obtaining a row matrix of zeros, as x=0, at every point of the bridge. Then we put the force 1 m from the left support (x=1). The bending moment equation for this situation is the following:


where a is the distance from the left support to the position of the bridge we want to analyse. Do not confuse x with a, x is related to the position of the force in the bridge and a is related to the position of the bridge. Therefore, a row is obtained for the force situated at x=1 m, and we place it below the results obtained when the force was on the support (i.e., row with zeros below). We repeat the procedure until x=20 m and we obtain a 21 x 21 matrix like this (in kNm):



The information of this matrix that combines the value corresponding to the position of the force, in every row of the matrix, and the information regarding the position analysed of the bridge, in every column, can be plotted in a contour plot like the following:

Figure 3. Bending moment contour plot of a 20 m bridge under 20 kN point load.


The advantage of this plot is that we can predict the value of points that we have not calculated before and visualise a picture of the bridge response anywhere. For example, if we want to know the value in the bridge position 4.5 m when the force is at 9 m from the left support, we can estimate that the bending moment would be a bit less than 50 kNm. When using the formula, a bending moment of 49.5 kNm is obtained. It means that with a glimpse we can access to all the data. Therefore, if you are frustrated in your work, remember to use this tool. It can make you discover unexpected relationships between parameters and give you hints to move forward. Try it!