The objective of the work presented is to compare results from numerical optimization with experimental data and to highlight and discuss the differences between two fundamentally different optimization methods. The problem domain is minimization of acoustic emission and the structure used in the work is a closed cylinder with forced vibration of one end. The optimization method used in this paper is simulated annealing (SA), a stochastic method. The results are compared with those from a gradient-based method used on the same structure in an earlier paper (see Tinnsten 2000).

Residual deformations of pipe-flange joints due to thermal strains caused by the multi-pass welding procedure in a manufacturing process are investigated both experimentally and numerically (FEM). To avoid any risk of leakage between two mounted pipes using flange joints, geometrical distortions (twisting) of the flange must be prevented. How the flange is distorted depends both on the groove shape and the weld sequence used However, to be spared from machine work after the welding process, the distortions have to be minimized. A casting-technique has been developed in order to measure the twisting of the flange after each weld pass. Also, measurements of the residual radial deflection of the pipe have been performed.

Four examples are used to illustrate the current ability to calculate residual stress and distortion in situations of varying complexity. The methodology used is capable of allowing for interactions between thermal, mechanical, and transformation processes. Experimental values for a 2.25Cr-1Mo ferritic stainless steel were used in the simulation. It seems possible to simulate most welding situations given sufficient computer resources

The quenching process of thin plates is studied. In many industrial applications of quenching, there is a need for simple tests for evaluation of analysis methods and material models for simulations. The presented study is part of a program for the simulation of the combined pressing and quenching process of thin-walled structures. This type of manufacturing process is used in, e.g., the production of side-impact protection beams for cars. By the use of an experiment with one-sided spray-cooling of a thin strip, large deformations can be studied and followed in time. The prediction of the deformation of the strip can be regarded as a one-dimensional problem where the curvature of the strip is sought. The thermo-mechanical behavior of the material can be evaluated throughout the time-history of the tests. In order to match the deformation of the plate, the material model must predict the temperature distribution through the thickness of the plate in each point in time, the phase transformations taking place in each layer of the thin plate and the stress and strain components in each layer. Experiments have been performed with a material (Inconel 600) without phase transformations in the actual temperature and time scale and with a material with only martensitic transformations. Analyses with a thermo-elastic-plastic material model with transformation plasticity included have been performed

A computational model for simulation of flame-cutting has been developed. Two different cutting conditions are compared with respect to the residual stresses. The first case is a plate cut at room temperature and the other is the cutting of a preheated plate. The model depends on measured micro-structure to estimate the required heat input. The model is verified by comparing the calculations for the preheated plate with experimentally obtained residual stresses. Lower residual stresses in the preheated plate prevents subsequent cracking

This paper describes a study of the multi-pass butt-welding of thick-walled pipes. Two different groove shapes were used: single-U and narrow-gap grooves. Strains were measured on both the pipe outer and inner surfaces during and after welding. Residual stresses and radial deflections were measured on the outer surface after welding. Measured parameters of the two different groove shapes are compared and discussed. Residual stresses at the weld center and radial deflections in the narrow-gap pipe were somewhat larger than those in the single-U pipe. The residual stress fields of both types of groove were found to be rotationally nonsymmetrical. The experimentally obtained residual stresses and radial deflections are compared with corresponding analytical and empirical results based on methods described in the literature and with numerical results for the same pipe from a separate study. Deviations between the measurements presented in this study and published results are discussed.