 I would like to welcome you on the presentation for a test case of offshore wind park operation. We have been starting to talk about test cases very early in this action and it has been some time until we really developed a test case and this is what I would like to present now. We are working with structural health monitoring and we are basing our action, our research here on what has been done in the last decades and that's why we can say traditionally, traditionally SHM is focused on loadings or structural reactions to monitor this in structural systems but also internal structural properties. However, in the perspective of that we should address the value of structural health monitoring or the utility in conjunction with the utility of infrastructures. Then we also must say that the value for society is highly related to the functionality and the benefits earned with infrastructures so we need also to include this utility in our modeling and at the same time the value or the utility for society can only be realized when costs and risks are sufficiently low. So that's why our value of information analysis should build upon a benefit cost and risk-informed assessment of the infrastructure performance and this is the perspective I'm going to take here in this presentation and where we have been seeing this in the last presentation I think it was on the first slide. An overview about the framework and related to different system states. We have the intact systems here, we have a system under exposure events under constituent damage states and system damage states accounting for the scenarios and then for each of the system states we can quantify consequences, benefits in terms of economy, health and environment and with a consistent formulation we may be able to formulate such a utility function and to optimize our actions. When we have a closer perspective in terms of the value of information and the structural health monitoring we see that we have various strategies here for monitoring the intact system, hazard threats, constituent damage states and system damage states. So this is symbolized with a mini Bayesian network here with the observation node. This would be a decision node and this is a cost node. So this is symbolizing the SHM decision information precision and cost and what we also have developed is a decision tree for value of information analysis that goes to the fact sheet by Dimitri Waal and Robi Kaspehler I think also and Cheng Singh. So the very basic representation of a value of information analysis and composing experiments or SHM strategies, the outcomes of the SHM then actions and system states. This is the very basic formulation of a value of information analysis, a very basic decision tree. This is what we have to account for when we are doing a value of information analysis and we then calculate the value of information as the optimal utility for doing structural health monitoring or minus the optimal utility when not doing structural health monitoring or we can also have the related value of information which is then just divided by the utility optimal utility without doing structural health monitoring. So maybe we have both. We have a visualization of the framework. First my first slide and now we have this decision tree. So where is the connection? The connection is the system states. So it's this branch here and it's this system stage which we find in the framework here also. So this was basically the introduction, shortly outlining the framework how we can use it and now we take the perspective of offshore wind park operation. That's a very actual problem as there have been substantial investments in Europe in wind parks in the last decade and now the focus of the industry is on operation because that's crucial for energy efficiency. This is very important for society but it's also about the economy of operating the wind parks and for earning money, for earning benefits. And the last point is there's a large number of identical structures so that's unusual to what we normally have. We have rather unique structures. There's not a high number of identical structures in civil engineering and exactly this was something I came across when I started working in the field of offshore wind energy 10 years ago. So there's a large number of identical structures and there is dependencies. To define a decision situation or decision scenario we can think of that an operator wants to assess in the commissioning phase of an offshore wind park what amount of data needs to be collected to facilitate a service life extension. We will see in a few slides why a service life extension is something which the industry is after but that's the usual situation. We have been in contact with a few wind park operators which are also included here in our in our action. So it's the commissioning phase the investment has been done and of course the operator aims at the highest utility throughout the service life that's the decision objective and where the solution is SHM systems but which one and then there are all kinds of questions which is which SHM system at how many turbines at which locations and at which locations in one wind turbine so this is all open questions. So let's take basis in this decision scenario and let's have a look at the intact system state. The intact system state is characterized by power production and operation costs. So there's an income due to power production so here we have the parameters like the capacity of the wind turbine the nominal capacity then availability factor of this wind turbine and then there is a factor of the nominal capacity when it is available because that depends on the wind speed on the mean wind speed and then there is there's numbers yeah percentages of what needs to be invested in the wind park operation yearly and then there is a discount rate so with these numbers this is very on a very generic level but still we can do here a cost benefit analysis and we see that there is a return over investment over 20 years operation of 15 percent and 25 years it's 30 percent so but this is this is exactly why operators are requesting service life extension also already in the commissioning state and this goes exactly to the challenge I've outlined so in the last years these systems can be very efficient and can help to to earn money so when we just take this cost benefit analysis then structural health monitoring is seen as something which goes against the benefits so why should we do this and this this is the situation I've been experienced over years that the discussions are always going to the direction yeah why should I use SHM then there's a lot of people which can do SHM which can implement very well systems which can analyze data this is very very developed but the benefits the utility of SHM is unclear so but we we also need to look at the other system states like the exposures so I'm considering here extreme exposure events and fatigue so I narrow down the context for this test case there's of course more yeah structural performance models there's also corrosion and so on but let's look to fatigue and extreme exposure events where fatigue is caused by the environmental loadings winds and waves and in conjunction with the operation and extreme events may be caused also by extreme winds in combination with extreme waves normal and abnormal operation so this is some of the basic assumptions the wind turbine fatigue damage is characterized with the fatigue design factor of 2.5 so that should be that should characterize a wind turbine fatigue damage we we have many components there and so the usual fatigue design factor is between two and three and we also characterized the wind turbine failure with the failure probability of five times 10 to the power of minus four this is in line with the JCS basis of design for minor consequences because wind park has only economic consequences and we we can on this level neglect any consequences in regarding to human safety and also the environment so well in wind park any of the wind turbine can be damaged or can fail so we are considering a wind park of 50 wind turbines so that's why we have to take here a serious system model to model the fact that there can be a fatigue damage or a failure at any of the system constituents and we account also for a correlation of 0.8 so when one turbine is damaged or has failed then there can be a failure of the wind park or damage or a failure of a wind park and we are also accounting for this cascading event with the brittle daniel system also utilizing a correlation of 0.8 and it's brittle because there is production capacity loss after failure the consequences in case that the wind park fails is that when it is damaged it's part of the wind park investment that's one percent if it was if it fails then the consequences is the wind park investment and there's also the production loss as a consequence so this is the damage or failure probability here on this axis and this is service life in years so we have it until 25 years here and where the probabilities of complete system damage or system failure are significantly lower than the failures and damages of one of these wind turbines now coming to the SHM strategies so we can think of an SHM strategy also highly simplified here we are monitoring each of the of the wind turbine we can we may be able to yeah monitor the fatigue loading so that goes to the exposure at each wind turbine location and we can utilize here a model to for the fatigue loading in the context of the pre-posterior decision analysis meaning that we are assuming that the model uncertainties have realized here for the fatigue loading in the fatigue damage model the second SHM strategy can be hot spot hot spot monitoring so here we would also measure very locally at each of the wind turbine structure and here we assume that more model uncertainties have been realized or realizing no have been realized because the structure is there so then we say that the loading model uncertainty the far field stress uncertainty and the hot spot stress uncertainty are realized so this is the way of modeling the SHM data in conjunction with the design models which we are already having and we can think of a further third SHM strategy where we model the extreme system loading the SHM precision is accounted for so we have yeah limited precision I think it's a standard deviation of 10% normally distributed and we also account for the statistical uncertainties so again a simplified model so we are assuming to start monitoring in the first year of the service life but if you just have one year of measurement and we do a statistical analysis we have very high statistical uncertainties because we just have one observation and then the statistical uncertainty decreases as I'm collecting in the consecutive service life more observations so this is accounted here in this model before and then structural monitoring is about costs so we have been assuming an investment of 500 thousand euros installation also 500 000 euros operation 20 000 euros per annum and we also assumed here a replacement interval of 10 years I've been involved in in tenders for SHM systems where this is was not accounted for but it's rather obvious that 10 years is rather high lifetime for SHM system okay SHM strategies so then we come to the decision tree so now we are back into our value of information analysis and the generic decision tree we have here now SHM that's easier that's our prior decision analysis and then we have the three SHM strategies one two and three would also follow here we by how we modeled the SHM results or outcomes we have no indications so if indications are related to probability of detection so we don't have this branch here but we have the action branch so this is a zero is service life 20 years as it was designed and a one is a service life extension to 25 years and then we have the desystem states so that's the overlap with the framework we have a safe state we have wind turbine damage or failure or wind park damage or failure so this is the decision tree we are going to analyze and when we have done that we can quantify the relative value of information so for SHM strategies 18 percent two is 31 percent and three would be 0.7 percent how does it look like in more detail so that's the decision tree only considering the experiments so SHM no SHM here SHM strategy one two and three here and we see the utilities for each of the of the actions and it's interesting to observe here the utility for our prior decision analyzes here it is not optimal to extend the service life because we'll see this in the slide later the risks are too high here here in the prior decision analyzes taking now SHM information the a zero is optimal here and the same number like here should also be here sorry that's a mistake and well when we look at SHM strategy one so that was fatigue load monitoring then we see it has been it was shifted so it's now the highest utility is is here in this branch related to action a1 to service life extension so looking further at SHM strategies two and three we have the same picture here for also for SHM strategy optimal action is a1 service life extension and for SHM strategy a three so that was the extreme system load monitoring we see that the optimal action is not to extend the service life so that's that's further breakdown of the results we have the expected benefits over 20 years that's the utility for the optimal prior decision so we have the expected benefits here we have indirect risks here going to system damage and failure and the direct risks here going to wind turbine damage and failure so when we look at the utility for the optimal preposterior decision so that's related to SHM strategy two then we observe first now that we have a service life of 25 years and that the indirect risks and the direct risks have been substantially reduced especially the indirect risks here in comparison to the previous slide so well this was rather complete for SHM strategies one to three but well looking at one of the first slides here for the intact system we see that we have factors like the wind park operation here it's also the availability factor so I thought it would be good to explore what if we could also reduce the wind park operation costs or we could provide a higher availability of the turbine so if our SHM systems can also influence actions related to the availability and to the wind park operation so I I calculated here for more wind more efficient wind park operation there would be a related benefit of 2.6 percent and if we can achieve lower down times or a higher availability by 10 percent then we can achieve a value of information of 15 percent and the interesting thing to note here is also that as we are in the in the parameters of the cost benefit analysis then we see also in the cost benefit analysis for the intact system the effect of SHM strategies of course when we think more about that more efficient wind park operation so lower number of inspections for instance we also have an interrelation by the observations we are accounting for and the risks meaning the direct and indirect risks so well this is a modeling issue which we have not been accounting for here but in other publications so to conclude this is a test case to be further developed by adding complexity and and to calibrate to assess and to illustrate issues of interests in the future to calibrate also to case studies from the test case here from the basic modeling I've been presenting we can conclude that the wind park service life extension may only be optimal with relevant information with a relative high precision and I think this is also a very important aspect there's a wide range of SHM strategies not not just about the loading and internal structural properties but taking a wider perspective including different system states the intact system how can structural monitoring be beneficial for the operation I think there's a large potential I would also like to explicitly address the working group interaction so this is what I was referring to in the in the opening presentation so we should also work together across working groups so we hope for working group two this test case may be an orientation and inspiration of how we can build SHM strategies and structural performance models together for working group three and four it may be starting points for each of their challenges one for working group three is tools and and also adding complexity to the models and working group four is for the case studies so what we will do further is to use our network to get feedback so this is this is one this is why I'm standing here and requesting feedback but we will also address wind park operators here in our network which are not here today where we will further develop and then I think there should be further test cases we have identified at the very beginning reinforced concrete structures so we would need a test case here as a starting point but I've been observing in the last workshops that there is a lot of people working from different sites for concrete dams and also earth dams and levees so this could be another field for developing test cases thank you for your attention