 So this case is about bridge again, steel bridge and it's about the condition assessment problem let's say and I have co-authors who helped a lot and did a lot of work on this, so I really like to acknowledge their help here. Yes, so just very briefly because this has been presented already talking about bridge in Stockholm in a very central location. So it's a huge traffic demand on this bridge more than 500 trains per day and if you have to close down this bridge that will have quite significant consequences for the entire city and of course costs a lot of money. It's a steel bridge and it was built in 50s, so fatigue is a real problem and there are numerous fatigue critical details and there has been fatigue cracks and there is in this case we're focusing on a more interesting detail where one would expect fatigue cracks according to some theoretical considerations but they couldn't find any signs of fatigue so then what what happens then that's the question should we do something or is it still safe or do we need to do some more advanced assessment do more inspection with some devices or monitoring system simply yeah simply just try to have a better more sophisticated model that could tell us that it's still okay but that is a typical problem with the condition assessment of any structures and rather this well-known flowchairs how to do this how to do initial assessment and then have to decide if you you need to do some further inspection and then go in the stook that was also shown in a previous presentation and then show this in Dublin we have this without this brilliant idea that it's not that try to structure these possibilities like basically these three main possibilities that focusing on having a better model or include more information by collecting information or somehow improve the the consideration of uncertainties and here we had some examples this time for this case like why are there the options for this bridge how to to look at the fatigue problem or and moving from data mystic analysis to really be based assessment on direct consideration of consequences or just using data that is also already available from the design or or actually do something on the bridge and then even showed some results from previous assessment and that the idea is really in all these cases that if you can somehow remove the consumptiveness of your estimation then you expect to save money because then you don't have to do this costly intervention so you don't have to repair or close down the bridge but but you don't know it certainly there is a chance that you will improve your assessment or the prediction but but it might not worse do it there's just some some numerical examples of a going and then usually you combine these movement along the axis and that that makes sense so try to find the best ways to to remove these consumptiveness how this can be used here and how you can use the video analysis so last time we did this chart it's very easy to see but we we identified what are the typical actions that one could do and also indicators in this case it was quite clear maybe i shouldn't show this again i would just do show you what we have been doing since the last meeting and the main difficulty is actually someone who is not familiar with these methods is you look at this that has been discussed in the planning school as well and yeah how i would really use this for a real case so i'm trying to show this to you and yesterday we heard about different software tools there's a very very efficient software tool called excel and i think you can understand a lot so we tried to simplify the problem and then the state of nature here in this case it was there is damage or no damage in this detail and we had some prior probabilities based on the previous assessment let's say from visual inspection then we consider two type of actions interventions here repair or no repair we don't include all the possible consequences just to to make it simple and the total costs maybe you shouldn't call it total costs but that's the cost of the intervention and the cost of a possible failure and then there is a combination of those if you have no damage and you don't repair then you don't pay anything right if you don't repair and it fails then you have to pay more then if you repair then then that has a certain cost and then what we are after now is to decide on what kind of assessment we should do so we just assumed it's a very generic formulation if you have three options do no further assessment that has no cost and we consider two different assessment options one is cheaper but let's say less reliable you can sit here in this conditional probabilities and the other one is more expensive but gives you better indication of of damage so we call this outcomes improvement or no improvement it's not so clear that you have a doubt and after the assessment you still have a doubt or you improve your your assessment about the the fatigue life for example I hope it's clear if it's not then you can take a closer look after the meeting so then we just did this similar examples that are presented in the training school for those who were here and then you can you can calculate the so you can do this without assessment you could calculate the expected benefit in this case just costs and then you can do this with this less good or cheaper I would say cheaper assessment and with a more expensive assessment and then it's very simply you can select which one to choose if you have one shot so it doesn't really include this successive steps in this case but I will conclude later and then we thought that probably we shouldn't do this by hand all the time so then we looked into this influence we created an influence diagram of this simple problem and then did this with genie and you see the numbers are numbers are the same so we hope that everything was correct here and then then it's just much easier and you could then calculate if you don't assign costs to these to these possibilities here then you could just calculate the the value of information how much you you should pay for for this method or for this method all right yes and then what we are working on right now is to to include these successive steps so it's very difficult to see but so so like first you have to you only consider two options either you go with no assessment or you go to level one and then if you are if you do this level one then you have another decision problem you either satisfy with your results and make a decision based on that so these are the intervention notes here and these are decisions on further assessment or you can go move to the next level but this just gets more complicated they try to put this in genie but i don't have didn't show you any numbers yet because i'm a little uncertain if it's correct maybe this is something we could get some help from someone who is more familiar with this but this was our first idea how to do this in genie and then you could you could decide on a second or you could compare different strategies or how to how to move to the next and that was it and we have we have some publications already published something about this okay have a good rest