ekenje, z njasih izgreda, ima do prezentacija, da je predpe, skupne pravne veselje nesec in da se občasno trenete o nekomotivama progrzje se tukaj da ima neoma vse neskega nekaj za j годini očilujem na našaj from the concept of structural health monitoring. It is essential, that in our analysis, the inspection of the tank technical condition with the toilet measurement of the thickness loss of corraded sheet must be performed according to the polished low at least every six years if the tank service life exceeds 30 years. To brez s nzem iznočenja, in s drugih način, in v complexesu, in tudi način, in vse način, začne se stavila, tačne čitvarje, in vse najske delavnih pradov, evoza, da je pasa vsak takt je vodilena, taj saj je od nekaj, obstavila, začno tega, način je, in v 5 karakteristikih pojnja, ki je zelo vzvečen v nekaj geometričnih sej, in v 4 kornečne vse še. Zvom, da je, kako je, repetitivno, še inspekčno, stativno, skupno, prijevno, ko je, kot objevno, zelo vzvečen, na vsej, ko je pristil, na vsej, kako je, obzerv, karakteristički pojnj, v kaj je korozija proces vse rekoči v tegničnih inspekčnih. Kako se analizuje in preformačuje, te dve basici, random variabili, je spetifizovati, in nekaj nekaj nekaj nekaj. The first one is the variable action effect, which is interpreted here as a tensile force in a steel shell, which is the function of the hydrostatic pressure and the random over pressure, and in this function the radius of the shell is treated as a fully deterministic parameter. Please note that in such approach both the loads seems to be the quantities capable to be monitored, even continuously monitored, not only during the technical inspection. The second random variable is the random resistance of a shell, which is the function of the random steel limit point and the random thickness of the shell, both the steel limit point and the thickness of the shell are described by the application of the log normal probability distribution, then also the random resistance can be characterized by log normal probability distribution. The median value of the steel point is taken directly from the appropriate standard because the characteristic value of such limit point is known in advance and because of the lack of the empirical experiment, parameter is taken from the experiment. Moreover, the normal coefficient of variance is adopted here to be equal to 0.1, which is in accordance with many statistical estimation and also is Poland. Please note that the random parameters of the thickness are here, the empirical parameters, because these are obtained according to directly from the measurements taken during the technical inspection. The simplified approach was tested by the authors at the first stage of our analysis and according to this analysis, the durability prediction of a single tank shell plate was performed and ideal geometry of the shell was assumed and the bending moment free stress state in the shell was analyzed. Let us see the figure shown in the left side in this slide, in which we can observe the random process of the resistance, which is decreasing in time because of the influence of the corrosion progress and also we have the random process of the variable action effect, which is assumed here to be stationary according to this assumption. The mean value of this process is assumed to be constant during the whole time of the tank service. In each time of the tank service, the failure probability can be calculated and the failure is identified at the point in time when such failure probability reaches the level described by the ultimate value, omega ultimate, which is maximum possible to accept for the tank user. In the first stage of our algorithm, the time-dependent safety margin is specified, then this random variable is standardized and finally the failure probability is calculated through the Laplace function phi. So the safety condition can be written in a way shown in the bottom of this slide, which can be rearranged to the other one. This is a comparison of the reliability index beta, which is the time-dependent reliability index and such reliability index should be greater than the minimum required value of the beta index. Next, to improve our analysis, the advanced approach was performed, according to which the durability analysis for the whole tank sheeting was performed. According to this analysis, the first step was the replication of a real tank shell geometry and the geometry non-linear stress analysis was performed. Please observe in this slide the real geometry, which was measured during the technical inspection on the full field tank with geodetic methods. So as a result of such measurement, we can start with the structural performance analysis and this analysis was performed for comparative purposes in three ways. The first, most simple, was the approach according to the perfect geometry was analyzed in the geometrical linear analysis of the shell and the second approach was the real geometry and the geometrical linear analysis of the shell and the third, which seems to be the most interesting, according to which the real geometry was analyzed and non-linear analysis of the shell was performed. On the right bottom, we can observe the comparison between the von Mises equivalent stresses obtained according to the each of these methods. This is the last slide and the next point of our analysis was the interpolation of the measured values of the thickness of corroded steel sheet, which allows to adjust the experimental results to the numerical models prepared by the authors. At the current stage of analysis, the authors are striving to extend the numerical analysis model of the tank, already including the real geometry and non-linear reaction of the tank material to applied loads to cover the sheet in corrosion state, changing during service life, measured during the technical inspections and projected to the future based on identified statistical adherence and to sum up, I would like to tell that these results obtained in this way seems to be helpful in the reliable evaluation of the durability. However, it should be validated and quite different concept was chosen by the authors for the validation. This concept is based on the Bayesian approach combining the uproaried information with the uproaried measurement data and according to this approach the state of knowledge available to the investigator in the beginning of the consecutive technical state inspection as treated as the uproaried structural state and the new empirical data which are gathered as the results of the technical inspection will be treated as the information gathered a priori and finally the posterior distribution of the parameters which are of interest of the investigator can be evaluated and validated the results obtained by the methods mentioned previously. Thank you.