TABLE OF CONTENTS Regardless of how it is made, a weld constitutes a discontinuity in a structure and as such may diminish its integrity. The weld itself may well be, and often is, stronger than the parts that it joins, but there are many secondary effects such as reduced resistance to fatigue or to corrosion which must be countered in one way or another. Much of this book has been concerned with the nature of these effects and the means of overcoming them. It is appropriate at this stage to take a look at the general problem of ensuring the reliable behaviour of welded equipment. Four main aspects will be covered; firstly, the means of making a quantitative measurement of reliability; secondly, the character and incidence of failures; thirdly, the available methods of assessing the reliability of flawed equipment, and finally the means of assuring reliability during construction and in service.
According to classical reliability theory the behaviour of a large number of identical working parts in service can be described by the bathtub curve illustrated in Fig. 6.1. When first put into operation this group suffers a relatively high rate of startup failures. Such failures are due to defective material or a lack of quality control, and the components in question are eliminated during a relatively short period of operation. At the end of the operating period there is again a relatively high rate of failure due to wear because the parts have exceeded their normal operating...
