PSI - Issue 48

7

Štěpán Major / Procedia Structural Integrity 48 (2023) 230 – 237 Major / Structural Integrity Procedia 00 (2019) 000 – 000

236

When using this approach, it is first necessary to introduce a dimensionless internal variable φ ≥ 0 and propose an evolution equation for φ in the form:   , , 0 P A x t      (13) where is an equivalent plastic shear strain rate. The letter A is accumulation modulus during plastic flow. Accumulation modulus depend on the current values of lattice hydrogen concentration C L and the local stress state. Then accumulation modulus during plastic flow can be expressed as: (14) where σ M is mean normal stress, and ‡“ is an equivalent shear stress. Overall, this modeling method is considerably more complex than the previous two, but due to the fact that we do not have enough space in this article to fully describe it, and also because this method is described in detail in the article by its authors, and in our cause, it was only applied to chosen example, the reader can be referred to the original post. 5. Experimental study Load experiments and fractographic studies were carried out on welded samples with dimensions of 4 20 90   mm, which were always created by welding two half-length parts. The geometry of the samples is evident from Fig. 1 and Fig. 2a. The material used was 16Mo3 alloy steel suitable for pressure vessels (Minimal yield strength σ y,min = 275 MPa, tensile strength σ U = 590 MPa). During load tests, the samples were subjected to bending stress as shown in Fig. 1. Optical measurements designed to determine COD were carried out on a Microprof -100 optical measuring device. The effect of hydrogen on the welded part was simulated by immersing it in an acid solution. 6. Results and conclusion The relative decrease in weld life due to hydrogen embrittlement is as follows: a) we consider the life defined as the number of load cycles N at the load amplitude σ a and the amplitude of the same life defined by the number of load cycles N ; b) now, from the experimental curve for the sample that was exposed to hydrogen, we determine the σ aH amplitude at which the same lifetime value N was measured as in the previous case. c) The proportion of these two amplitudes for different values of lifetime N will form a curve that will characterize the relative decrease in lifetime due to the action of hydrogen. We will then be able to assess the effectiveness of the individual models using this experimental curve of the relative decrease in load amplitude and an identical curve in which the value of σ aH is replaced by the predicted load amplitude of σ aH (predicted) for the given lifetime under the action of hydrogen N and on the weld. The size of the deviation of both curves then indicates the quality of the model. These ratios of amplitudes therefore correspond to the values on the vertical axis of the Wohler curve, see Fig.4. If we look at the curves in Fig. 4, we see that model I and model III are approaching and deviating from the experimental curve, both roughly the same, but each to a different side. The first model would generally predict a higher allowable load for a given lifetime and therefore the values determined by it would be in the dangerous area. From this point of view, the advantage of model III is obvious, by which the calculated values would be rather in a safe area, i.e. below the yellow curve determined experimentally. At the same time, this model is more accurate than model II. It should also be noted that models I and II work with data obtained by fractographic analysis of the fracture, i.e. they strive, especially in the case of Model I, to reconstruct the fracture formation process and at the same time have limited use for predicting the fracture process affected by hydrogen embrittlement. From this point of view, the approach proposed by the three authors Ananand et. al. appears to be the most accurate and effective. , 0 M L eq A A C            