Introduction

A long-term period of operation of a pressure vessel – reactor (over 40 yr) causes certain damage to the reactor mantle. The occurrence of this damage demands a thorough inspection of the reactor structure itself, along with repairing of the damaged parts. Reactor repairs included the replacement of a part of the reactor mantle with a new material. The pressure vessel considered in the present work was made of low-alloy Cr–Mo A-387Gr.B steel in accordance with the ASTM standard with 0.8–1.15% Cr and 0.45–0.6% Mo. For the designed operating parameters (p = 35 bar and t = 537°C), the material is in the area where it is prone to the decarbonization of the surface in contact with hydrogen. As a consequence of surface decarbonization, the material strength may be reduced. The reactor represents a vertical pressure vessel with cylindrical mantle. Deep lids are welded on the top and bottom sides of the mantle of the same quality as the mantle itself. Inside the reactor, the most important process takes place in the stage of motor gasoline production and involves platforming in order to change the structure of hydrocarbon compounds and, thus, achieve a higher-octane rating. We tested new and exploited parent metal (PM), the components of welded joints [weld metal (WM) and heat affected zones (HAZ)], as well as the low-alloyed steel from which the reactor is made.

The exploited PM was made of A-387Gr.B steel with a thickness of 102 mm. The chemical compositions and mechanical properties (Table 1) of the exploited and new PM (according to the test documentation) are presented below. The chemical compositions of the exploited and new PM specimens (wt.%) were as follows: specimen E (C – 0.15, Si – 0.31, Mn – 0.56, P – 0.007, S – 0.006, Cr – 0.89, Mo – 0.47, and Cu – 0.027) and specimen N (C – 0.13, Si – 0.23, Mn – 0.46, P – 0.009, S – 0.006, Cr – 0.85, Mo – 0.51, and Cu – 0.035).

Table 1 Mechanical Properties of the Exploited and New PM Specimens
Full size table

Welding of steel sheets made of the exploited and new PM was performed in two stages according to the requirements given in the welding procedure provided by a welding expert. These stages included: root weld of specimen E by using a coated LINCOLN S1 19G electrode (AWS: E8018-B2) and filling by arc welding under the powder protection; here, the wire (denoted by LINCOLN LNS-150) and the powder (denoted by LINCOLN P230) were used as additional materials.

The chemical compositions of the coated electrode LINCOLN S1-19G and the wire LINCOLN LNS-150 (according to the test documentation) are given below. The chemical compositions of the additional welding materials are as follows: LINCOLN S1-19G: C – 0.07, Si – 0.31, Mn – 0.62, P – 0.009, S – 0.010, Cr – 1.17, and Mo – 0.54; LINCOLN LNS-150: C – 0.10, Si – 0.14, Mn – 0.71, P – 0.010, S – 0.010, Cr – 1.12, and Mo – 0.48. Their mechanical properties, also according to the test documentation, are presented in Table 2.

Table 2 Mechanical Properties of Additional Materials
Full size table

A butt welded joint was made with a U-weld. The shape of the groove for welding preparation was chosen based on the sheet thickness, in accordance with the appropriate SRPS EN ISO 9692-1:2012 [2] and SRPS EN ISO 9692-2:2008 [3] standards.

Evaluation of the Fatigue Strength

Metal fatigue is defined as the process of cumulative damage under the action of variable loads, which manifests itself in the occurrence of cracks and fracture. Fatigue strength of welded joints is determined by testing the specimens under variable loads that lead to the initiation of cracks or fracture. In the case of reactors, i.e., pressure vessels operating under increased pressure and temperature conditions, high-cycle fatigue tests are of particular importance. The strength of welded joints under variable loads observed in nonstationary operating modes of the reactors in the course of startups and shutdowns serves as an important characteristic for the evaluation of the integrity and remaining life. At the same time, it should be taken into account that a crack-like damage occurs after a large number of changes in the load under stresses much lower than the yield stress (high-cycle fatigue). Under loads whose levels are lower than yield stress, typical of high-cycle fatigue, it is a common practice to perform tests in the rigid mode.

High-cycle fatigue depends on the properties of constituents of the welded joint. The characteristics of high-cycle fatigue start significant changes only at temperatures above 400°C for steels used for pressure vessels and their welded joints.

Testing of the effects of temperature and exploitation duration on the behavior of the new PM, as well as on the behavior of butt welded joints subjected to variable load was performed to determine the points in the S–N-diagram (drawing of Wohler’s curve) and fatigue strength (Sf) of the welding procedure. Moreover, the specimens were tested according to the ASTM E466, ASTM E467 and ASTM E468 standards [4,5,6]. The shape of the specimen tested under variable load is shown in Fig. 1.

Fig. 1.
figure 1

Specimen for dynamic testing according to ASTM E466.

Full size image

Testing was carried out in a high-frequency pulsator. The values of the achieved frequency ranged between 115 and 165 Hz depending on the load and testing temperature. In order to fully evaluate the behavior of materials subjected to variable loads, while taking into account the dimension of the specimen, the most critical case of variable load was considered with the use of a variable load alternating between tension and pressure (R = – 1).

During this testing, it is a general rule to determine only the number of load changes until fracture under the load with a constant range, while the standard requires only the data about the magnitude of stresses for which crack initiation and fracture do not occur after a specific number of cycles (typically between 106 and 108 cycles). For steel materials, the ASTM E466 standard defines Sstrress, Sf after 107 cycles.

The effect of exploitation duration and temperature on Sstrress, Sf, i.e., the maximum dynamic stress under which there is no initiation of crack-like defects in shapes with smooth structures, is graphically shown in the form of Wohler’s curves (S–N-diagrams) in Fig. 2a for specimens with butt welded joints and in Fig. 2b for specimens cut out from the new PM [1].

Fig. 2.
figure 2

S–N-diagram for the specimens cut out from the butt welded joint (a) and from the new PM (b) and tested at room and working temperatures: (□) 20°C; (○) 540°C.

Full size image

Testing of specimens of the PM was not performed because all specimens of welded joint cracked in the zone of the exploited PM and, hence, these tests provided the characteristics of the welded joint and PM.

In order to draw Wohler’s curve and determine the fatigue strength, it is necessary to test the specimens for 6–7 different levels of loading.

According to the ASTM E 466 standard, three specimens were tested for each load level, which made a total of 21 specimens. For this reason, the proposed tests are very cost-intensive and, hence, are justified only when the data are necessary for designing, first of all from the aspect of fatigue and fracture mechanics; in other words, when the parts subjected to long-term variable loads are designed as a part of the total designed structure life.

The effect of testing temperature on Sstrress, Sf obtained by testing the specimens cut out from a butt welded joint and the new PM is shown in Fig. 3 [1].

Fig. 3.
figure 3

Behavior of the Sf values for specimens cut out from the new PM and a butt welded joint depending on temperature: (□) new PM; (○) welded joint.

Full size image

Macro- and Microstructural Testing

For the successful application of A-387Gr.B steel and in order to obtain its maximal creep resistance, guaranteed mechanical properties are requested at higher temperatures, as well as the creep resistance at exploitation temperatures within a period that can be longer than 150,000 h. These properties are obtained with the proper thermal treatment providing the structure consisting of ferrite and bainite. Very fine carbides that start to sediment during this thermal treatment segregate on the grain boundary, as well as within the grain, which can be seen at high magnifications [7, 8].

Carbide precipitation, which begins during thermal treatment for the relaxation of residual stresses, continues in the course of operation at the exploitation temperatures and pressures [9, 14]. The appearance of these brittle phases can be ascertained by the data of metallographic analysis obtained under high magnification. This testing was conducted in order to evaluate the period of operation the parent metal and the components of welded joints depending on the variations of the microstructural properties. The macrorecording of a butt welded joint of the new PM and exploited PM is presented in Fig. 4 [1].

Fig. 4.
figure 4

Macrorecording of welded joint.

Full size image

After etching of the butt welded joint, we can clearly distinguish the new and exploited parent metal; the heat affected zones on both sides, and the weld metal with well-marked groove filling zone.

Both parent metals show even structures that consist of bright polygonal ferrite crystals and transformed areas that can be analyzed under high magnifications. These transformed areas represent dark surfaces of perlite that looks like a compact dark microconstituent. The microstructure of the PM after operation for more than 40 yr is shown in Fig. 4 and the PM microstructure is shown in Fig. 5 [1]. The difference is in the grain size. A newly installed parent metal has a structure with grain size 5 according to the ASTM scale, while the exploited material has a structure with grain size 3 according to the ASTM scale.

Fig. 5.
figure 5

Microstructures of the exploited PM (a) and new PM (b); ferrite-perlite structure.

Full size image

It is clear that, at a × 100 magnification, it is impossible to detect a noticeable difference between the degraded and new material, except in the grain size.

The microstructures of the HAZ on the degraded and new PM sides are shown in Fig. 6 [1]. They consist of ferrite, bainite and perlite. A bainite in the HAZ is formed as a consequence of a higher cooling rate of a part of the parent metal heated to the austenitizing temperature in the course of welding. A bainite level reduces the increase in the distance from the joint line.

Fig. 6.
figure 6

HAZ microstructure for the used (a) and new PM (b).

Full size image

A weld-metal structure with large dendrite created as a consequence of the foundry bath size and dimensions of the welded plates is shown in Fig. 7 [1].

Fig. 7.
figure 7

WM microstructure; dendrite structure of the weld metal.

Full size image

Higher magnifications (× 500 and more) enabled us to reveal the differences in the structural properties of the exploited and new PM. The period of operation of more than 40 yr affected the presence of carbides on the grain boundaries and within the grains. The amount of carbides in the new PM is much lower and the carbides are smaller (Fig. 8) [1].

Fig. 8.
figure 8

Microstructures of the degraded PM (a) and new PM (b).

Full size image

CONCLUSIONS

The resistance of the material to crack initiation is determined as a result of fatigue-strength testing of materials. This is the maximum value of stress for which crack initiation does not occur for smooth specimens. The higher the ratio of fatigue strength to yield stress, the better its resistance to crack initiation. By analyzing the results obtained as a result of high-cycle fatigue testing of smooth specimens for the purposes of plotting Wohler’s curves and evaluation of the fatigue strength, it can be shown that the duration of exploitation has a predominant effect on the obtained values of fatigue strength. In case of testing of the specimens of welded joints made of the exploited PM at room temperature, the ratio of the fatigue strength to yield stress is 0.68, i.e. the obtained level of fatigue strength represents 68% of the yield stress. All specimens of welded joints broken in the course of the tests carried out under loads higher than the fatigue load suffered cracking either in the exploited PM or in the HAZ on the side of the exploited PM. The effect of testing temperature is such that an increase in temperature reduces the fatigue strength. In this case, the specimens of welded joint failed either in the exploited PM or in its HAZ. The crack initiation resistance for testing at 540°C decreases, i.e., the trend toward brittle fracture becomes more pronounced. The test results and their analysis justify the choice of welding technology for the purpose of replacing of a part of the reactor mantle.