Pre?posterior Bayesian analyse toegepast op inspecties van waterbouwkundige constructies

Pre?posterior Bayesian analyse toegepast op inspecties van waterbouwkundige constructies

Al-Baz, S.

Vrijling, J.K. (mentor)
Vrouwenvelder, A.C.W.M. (mentor)
Van Gelder, P.H.A.J.M. (mentor)
Willems, A. (mentor)
Van der Wiel, W.D. (mentor)

Civil Engineering and Geosciences

Hydraulic Engineering

2011-12-16

Existing structures are deteriorating. The deterioration depends on many factors, including workmanship, and therefore it is difficult to predict the exact deteriorating in time. By inspections it is possible to measure the deterioration and create a model with statistics and curve-fitting. Because of the high consequences C of failure of hydraulic structures, it is important to keep the probability of failure PF low. This follows from the definition of risk: Risk = PF?C Here is PF not the true probability of failure, because this is unknown, but it represents our perception about the probability of failure. The probability of failure consists of aleatoric and epistemic uncertainties about the resistance of and the load on the construction. The epistemic uncertainties can be reduced by inspections, so we will be more certain about the reliability, without changing the structure physically. A low probability of failure for existing structures can be realized by inspections and maintenance. With the inspection results one can determine the condition of the structure and the necessity for maintenance. But what determines the necessity for inspection? This is an interesting question when the inspections are destructive or very costly. When it comes to relatively cheap inspections, this will also play an important role, if they have to be executed a great number of times. Especially if a lot of unnecessary inspections are executed. With the pre-posterior Bayesian analysis one can determine value of certain types and numbers of inspections beforehand. By weighing the value of inspection against the inspection costs, it is possible to develop an optimal inspection strategy. The method above is applied to the reinforcement bars in a concrete plate of a bridge. The effective diameter of the reinforcement bar is uncertain due to the presence of corrosion. Therefore the strength of the reinforcement bar is uncertain. With the present conditions the risk can be to such a degree, that it is cost efficient to inspect and/or maintain the bridge. The effective diameter can be determined by a destructive and expensive method, of which a number of n can be executed. The total costs exist of the inspections costs and the expected maintenance costs belonging to a number of n inspections. The optimal number of inspections n follows from the minimization of the total costs. Because the inspection results are unknown beforehand, these results are treated as random variables, according to prior knowledge (the deterioration model). Then the cost optimal decision is chosen belonging to a number of inspections n . This is possible because we know that the diameter of the reinforcement bar is uncertain, and that this uncertainty can be reduced by inspections.

http://resolver.tudelft.nl/uuid:04f89dec-42ba-4ea0-ac98-16931d4cf930

Student theses

master thesis

(c) 2011 Al-Baz, S.