EVALUATION OF LOW-CYCLE FATIGUE DAMAGE IN STEEL STRUCTURAL COMPONENTS BY A MAGNETIC MEASUREMENT TECHNIQUE

Transcription

1 VALUATION OF LOW-CYCL FATIGU DAMAG IN STL STRUCTURAL COMPONNTS BY A MAGNTIC MASURMNT TCHNIQU Madhav Rao Govindarajul,2, Andrew Stroml,2 David C. Jiles l,2 and S.B. Biner ICenter for ND 2 Ames Laboratory Iowa State University Ames, Iowa 511 INTRODUCTION Fatigue is one of the leading causes of failure in structural components. The development of a viable ND technique which can detect fatigue damage in its early stages of development, monitor its progress and be capable of predicting the onset of catastrophic failure is very essential. If the damage can be detected at an early stage, corrective measures can be taken either in the form of repairs to, or replacement of, the damaged part. Magnetic inspection methods are perhaps the most promising techniques for nondestructive evaluation of the mechanical condition of ferromagnetic materials [1]. These methods depend on the inherent ferromagnetic properties of steels for material evaluation. In general the changes in magnetic properties are very sensitive to mechanical and thermal treatments and are easily measurable. However the magnetic methods have not been fully exploited when compared with other methods due to associated problems, such as the lack of proper instrumentation and difficulty in interpretation of results. In recent years there has been significant progress at the Center for ND at Iowa State University in this direction and the foundations for the basic understanding of the changes in bulk magnetic properties have been laid [2-6]. The use developed by the Center for ND at Iowa State University provides an extremely powerful method for evaluating the condition of steel structural components. In ferromagnetic materials a well established correlation exists between the elastic / plastic deformation and the associated changes in magnetic properties. The magnetic properties are influenced by dislocations and other imperfections such as inclusions and inhomogeneities in the material. It is also well known that during the fatigue process the dislocation density changes very rapidly. These dislocations interact and impede the free movement of magnetic domain walls and lead to changes in Review of Progress in Quantitative Nondestructive valuation, Vol. 12 dited by D.O. Thompson and D.. Chimenti, Plenum Press, New York,

2 magnetic properties such as coercivity. Schumann [7] initially reported on the coercivity changes within a single load cycle in the saturation stage of the fatigue hardening curve. The changes in the magnetic properties during fatigue loading were also studied by Shah and Bose [8] and Bose [9]. Sanford-Francis [1,11] has used the changes in magnetic properties with stress cycling to predict remaining fatigue life. However, in spite of these research efforts made in the past, there is a lack of basic understanding of the variation in microstructure-sensitive magnetic properties with fatigue damage. The present paper reports on recent investigations of a systematic study of microstructural changes due to fatigue damage in structural steels and the corresponding changes in magnetic properties, using the magnetic hysteresis method. The ultimate objective of this investigation is to develop a suitable ND method to monitor remaining fatigue life of structural steel components. XPRIMNTAL PROCDUR In the present study, specimens taken from a 15 ft. truss railroad bridge erected at Barstow, Illinois in 1898 were used. The chemical composition and mechanical properties of the specimens used in the investigation are given in Tables 1 and 2 respectively. Fatigue tests were conducted on a computer controlled, 1 kn, servo hydraulic MTS system. The specimens were subjected to low-cycle fatigue at strain amplitudes of.2 and at a frequency of 2 Hz. The magnetic parameters were measured with the Magnescope, at predetermined intervals in fatigue life under zero load. The magnetic properties recorded were: coercivity, remanence, maximum differential permeability and hysteresis loss. Surface replicas were taken on one specimen to evaluate fatigue damage on the specimen surface leading to nucleation and propagation of fatigue cracks. RSULTS AND DISCUSSION The variations in magnetic properties with percentage of fatigue life are shown in Figure 1. The values of magnetic properties measured, viz., coercivity, remanence, maximum differential permeability and hysteresis loss have shown a significant change with fatigue cycling. A sharp increase in coercivity in the initial stages of fatigue life can be observed. This rapid increase can be attributed to the increase in dislocation interactions with magnetic domain walls which impede the domain wall movement. After its initial rapid increase, coercivity increased very gradually until the crack propagation stage, which corresponded to approximately 85% of fatigue life. The levelling off in the magnetic properties is due to the stabilization of the dislocation structure. Figure 2a is the plot of maximum load necessary to obtain the predetermined strain level vs. % fatigue life. This plot shows a typical fatigue softening occurring in the specimen in the low-cycle fatigue regime. The decrease observed in maximum load recorded at 85 % fatigue life is associated with the stable crack propagation stage. Surface replicas were taken at intervals of approximately 2%, 3%, 58% and 85% of fatigue life, as indicated in Figure 2a. Figure 2b and 2c show the scanning electron 184

3 m 5 Q 4. -.J I u ;i. ro : 3.S 2 3 OJ :::J 3.7 () : I-a-- Lmax - He SO 9 1 3(p 7 (a) " 6 (j) 26 C/) :::J 5 ro r - -.J '- en 4 u ro <Ii C 3 -.J OJ 1S C :::J ro )32 OJ. a:.:2 14 I-a-- Lmax - Br SO 9 1 Figure 1. Variation in magnetic properties with expended fatigue life: (a) change in coercivity with % fatigue life, (b) changes in remanence with % fatigue life. (b) 1841

4 D 8 C\l Q) 5 Qj n. Cii +:: 75 -l 4 " C C\l Q) '- Q) 7 3 (5 :J :J I-s- Lmax - Max. Diff. Perm (c) ' 5 z (f) 45 _ l (f) 4 (f) " C\l 4 -l a (f) w 3 -l 35 :J Q) U >- 3 II I 25 1 I-s- Lmax - Hysteresis Loss (d) Figure 1 (contd.) (c) Variation of maximum differential permeability with % fatigue life and (d) changes in hysteresis loss with % fatigue life. 1842

5 Table 1. Chemical composition (in weight % ) of the steel used in the investigation C Mn P S Si Al N2 Fe Bal. Table 2. Mechanical properties of the material used in the investigation Yield.2% Yield Ultimate Young's Hardness, Strength, Strength, Tensile Modulus, R., MPa MPa Strength, MPa GPa micrographs of the surface replicas taken at 2 % and 3%, which show the formation of slip bands on the surface. Figure 2d and 2e are the surface replicas taken at approximately 58 % fatigue life. Figure 2d shows the extensive surface damage due to the formation of persistent slip bands and associated snear localization which will eventually lead to the formation of a stable crack, whereas Figure 2e shows the nucleation of a microcrack of the order of 6-8 I'm at some other location on the sample in the same condition. Figure 2f is the surface replica taken at approximately at 85 % of fatigue life, which shows the growth of a stable surface crack of the order of 5-1 I'm. Magnetic properties measured at these stages of fatigue life show a significant correlation between the microstructural characteristics and the magnetic properties measured. It can be seen that the generation, accumulation and movement of dislocations which resulted in the formation of slip bands is accompanied by a rapid increase in coercivity. Also, a significant reduction in the magnetic properties is noticed with the propagation of a stable microcrack (Figure 2f), which is also accompanied by a drop in the maximum load value. CONCLUSIONS The results of this investigation demonstrate that magnetic hysteresis method is a viable ND technique for detecting the fatigue damage in ferromagnetic structural materials. The results indicate that the changes in the magnetic properties, especially the variations in coercivity and remanence can be used as a measure of fatigue damage. Impending failure can be predicted from the rapid decrease in the magnetic properties such as coercivity and remanence. However, the suitability of this technique as a viable ND method to monitor the progress of fatigue damage throughout the lifetime depends on improving the sensitivity of the system to detect nucleation and propagation of stable microcracks at an earlier stage. Further research is in progress in this direction. The magnetic properties of railroad bridge steel showed a significant change due to fatigue cycling. Among all the magnetic properties measured, coercivity and 1843

6 7 6 z 6 5 2% " al -I 4 ::J % 58% 85% 2 1- Replica Taken Here f \ (a) (b) (c) Figure 2. Scanning electron micrographs of surface replicas taken at various fatigue life intervals, (a) Plot of Maximum load vs. % Fatigue life, arrows show the locations where replicas were taken, (b) replica at 2% fatigue life, (c) replica at 3% fatigue life. 1844

7 (d) (e) (f) Figure 2 (contd). Scanning electron micrographs of surface replicas taken at various fatigue life intervals, (d) and (e) replicas taken at 58% fatigue life and (f) replica taken at 85 % fatigue life. 1845

8 remanence increased rapidly during the early stages of fatigue life due to dislocation movement and their interactions with magnetic domain walls and then remained fairly constant until last 15 % fatigue life. Thereafter, the magnetic properties changed significantly until fracture. This reduction in the magnetic properties could be related to the formation of stable growing microcracks. ACKNOWLDGMNTS This work is supported by National Science Foundation through the Center for ND at Iowa State University, under grant number MSS We thank Mr. David Utrata, Association of American Railroads for providing the samples, Mr. Zhao Jun Chen for his help in replica work. RFRNCS 1. D.C. Jiles, NDT International, 21, 5, p311 (1988) 2. D.C. Jiles and D.L. Atherton, J.Phys.D.17, p1265, (1984) 3. D.C. Jiles, J.Phys.D.21, p1196, (1988) 4. D.C. Jiles, J.Phys.D.21, p1186, (1988) 5. D.C. Jiles, S. Hariharan and M.K. Devine, I Trans. Mag.26, p2577, (199) 6. H.D. Schumann, phys. stat. sol. (a), K27 (1973) 7. M.B. Shah and M.S.C. Bose, phys. stat. sol (a), 86, p275 (1984) 8. M.S.C. Bose, NDT International, 19, 2, p83 (1986) 9. C.H.A. Sanford-Francis, Brit J NDT, 23, p241 (1981) 1. C.H.A. Sanford-Francis, Brit J NDT, 29, p83 (1987) 1846