Design and Analysis of Fatigue Resistant Welded Structures

Dieter Radaj

• Publisher: Woodhead Publishing
• Year: 1990
• Print ISBN:
  9781855730045
• eBook ISBN:
  9781845698751
• Topic: Technology & Engineering
• Subtopic: Civil Engineering
• Index: Technology & Engineering

An English version of a sucessful German book. Both traditional and modern concepts are described.

1

Introduction

1.1 The phenomenon of fatigue

Fatigue strength* is a particularly important application characteristic of welded structures, which can only be maintained at a safe level by broad knowledge and experience as well as by calculation and testing. In this book the phenomenon of fatigue is used as a starting point. The basis for the formal assessment of fatigue strength in the second half of the book (from Chapter 7 onwards) is produced via the large amount of technical and procedural data presented in the first half. This assessment is carried out in several different ways (on the basis of nominal stress, structural stress, notch stress, safety as determined by fracture mechanics, and limit load determined by testing) and for each of these ways only some of the necessary data are certain, otherwise they are estimated – material parameters, loading conditions, weld joint geometries and defect characteristics.

In the presence of time variable, frequently repeated (cyclic) loading, plastic deformation occurs at the microscopic and macroscopic level, which decreases future ability to withstand stress and initiates cracks. The cracks, initially microscopic, propagate, finally leading to a (monotonic) final fracture. Therefore, fatigue is the initiation and propagation of cracking, whereby, counted in cycles, the phase involving stable propagation of cracking can include a major part of the total life. Microscopic and macroscopic notches with their localised notch stress concentrations, and changes of section with their more extensive increases in structural stress are decisive here. Surface and environmental influences (surface roughness, corrosion, temperature) also have a particularly marked effect. A large number of further parameters such as design, material and manufacture also influences the fatigue process in a variety of combinations.

Fatigue viewed as a multiple parameter problem which cannot be seen in isolation not only results in the well known scatter of strength values within a series of tests and between various laboratories, but above all it prevents the quantitative forecast of the phenomenon which is the objective. Service fatigue strength predictions carried out in engineering on the basis of general published data and theory, without direct tests of fatigue strength under service conditions, are little more reliable than the medium term weather or economic forecasts, which are generally known to be problematical and error prone. However, unlike weather and economic forecasts, predictions of service performance can be improved in particular cases by detailed, well planned test methods, applied to the actual component with the precision usual in the mechanical sciences. Consequently, the generalised forecast with its quantitative uncertainty as mentioned above can only be used as an indicator for appropriate technical action as regards dimensioning, design, manufacture and operation. The interesting theoretical and experimental testing procedures involved in specific problems of service fatigue strength cannot be investigated in detail here.

From the point of view of the physics of metals, too, the phenomenon of fatigue appears to be extremely complicated. The effect of piled up groups of dislocations, which defines the static strength characteristics, does not explain fatigue fracture, which is known to occur well below the tensile strength. However, the formation and movement of dislocations are essential in fatigue fracture too as the first incipient microcracks are observed on the surface of slipbands. As a consequence, repeated etching to remove the slipbands increases the fatigue strength considerably. On the other hand, the microcracks combine to form macrocracks when subjected to further loading. Subsequent crack growth, which is stable at first, is described by slip processes at the tip of the crack, which occur in differing slip planes during loading and unloading.

1.2 Basic tests

The basic engineering fatigue strength test is the cyclic load test introduced by A Wöhler (1819-1914), in which a smooth or notched test specimen or the component itself is subjected to a periodically repeated (usually sinusoidal) load of constant amplitude (tension, pressure, bending or torsion). The number of cycles, N, which is endured before crack initiation or complete fracture occurs, is plotted against the load or stress amplitude (S-N curve), Fig.1. The stress amplitude which can be endured for any desired length of time without a fracture (the endurance limit) is called the fatigue strength for infinite life (for more than 2 × 10⁶ to 10⁹ cycles). Anything which fractures sooner, when subjected to a higher stress amplitude, exhibits the high or low cycle fatigue strength for finite life (the latter for more than 10³ cycles) which in the limit tends to the ‘static’ strength at 0.5 cycles. Depending on the magnitude of the static mean load, differing S-N curves are produced. Most important are the curves for zero mean load (fatigue strength under alternating load) and for zero minimum load (fatigue strength under pulsating load). In the low cycle fatigue strength range, it is more appropriate to carry out cyclic tests within limits of displacement or strain than of load or stress. Instead of plotting the load or stress endured, the displacement or strain endured is plotted over the number of cycles (up to crack initiation).

1 S-N curve for structural steel, St 37, number of cycles in linear (a) and logarithmic (b) scales, after Stüssi.15.

The Wöhler test with a constant amplitude is not adequate for real conditions in which load sequences with variable amplitude are predominant and frequently there are also aperiodic processes. The random amplitude test has been introduced as an alternative to the Wöhler test under constant amplitude loading; in the former a strictly random load sequence is applied – with a fixed amplitude range and frequency content – until crack initiation or fracture occurs. However, the results of such tests are less suitable for generalisation, as can be seen from the aforementioned specification of amplitude range and frequency content and from the fact that the loads are indeed frequently aperiodic in practice, but are not strictly random. Fatigue tests with blocks of stepped amplitudes are of further assistance here (multilevel test, block program test), as are load history tests (follower tests).

The cyclic stress-strain curve of a smooth test specimen has proved to be an important basis for the macroscopic description of fatigue processes. The extreme points of the stationary hysteresis loops (Fig.2) which develop in a partly softening, partly hardening manner after a fairly large number of cycles (about 10% of the number of cycles to crack initiation) in a fatigue test between load or displacement limits delineate the cyclic stress-strain curve which may be used as the basis of a stress, strain and damage analysis at the root of notches subjected to cyclic loading.

2 Hysteresis loop of the cyclic stress-strain curve.

1.3 Terms and definitions

The most important terms and definitions used in describing fatigue are explained below with reference to the German standard DIN 50100.¹

Fatigue includes damage to materials, crack initiation and crack propagation under frequently repeated loading of variable amplitude. The loading sequence can be both determinate, periodic or aperiodic, and also stochastic (random). It can show a static mean stress σ_m. Fatigue strength is the ability to withstand load without crack initiation or propagation, expressed by the level and number of cycles of the loading endured. Terms and definitions, illustrated in Fig.3 and 4, are drawn from the fatigue test which is carried out using a periodic stress sequence with a constant amplitude. Applied stress values are indicated by lower case letter indices, strength values by upper case letter indices. The fatigue strength is the stress amplitude*, σ_A, endured without limitation on cycles at a preset mean stress or the corresponding upper stress^†, σ_Up. Fatigue strength for finite life (high or low cycle fatigue strength) is the (higher) stress amplitude endured with a limited number of cycles at a preset mean stress or the corresponding upper stress. The fatigue strength is dependent on the mean stress, the finite life fatigue strength is additionally dependent on the number of cycles, N, (see Fig. 1 and 5). The fatigue strength**, σ_A0, under alternating load is characterised by zero mean stress, the fatigue strength**, σ_p, under pulsating load is characterised by the lower stress equal to zero. The stress amplitude is constant in a one level fatigue test; in a multilevel test it changes stepwise in a predetermined manner. In a simulated service fatigue test it follows a sequence which is similar to an operational one, i·e. completely or partially random. The service fatigue s...

Table of contents

• Cover image
• Title page
• Table of Contents
• Copyright page
• Preface
• List of symbols
• 1: Introduction
• 2: Fatigue strength for infinite life of welded joints in structural steel
• 3: Fatigue strength for finite life and service fatigue strength of welded joints
• 4: Fatigue strength of welded joints in high tensile steels and aluminium alloys
• 5: Fatigue strength of welded components, design improvements
• 6: Fatigue strength of spot, friction, flash butt and stud welded joints
• 7: Design codes, assessment of nominal and structural stress
• 8: Notch stress approach for assessment of fatigue strength of seam welded joints
• 9: Fracture mechanics approach for assessment of fatigue strength of seam welded joints
• 10: Structural stress, notch stress and stress intensity approach for assessment of fatigue strength of spot welded joints
• 11: Corrosion and wear resistance of welded joints
• 12: Example for the development of a fatigue resistant welded structure
• Bibliography
• Index