This project concerns welded structures with joints of complex geometry, methods for determining the fatigue life of these joints and finally post-weld treatment for improving the fatigue strength. The project is carried out in cooperation with the Danish company HMF A/S, who manufactures truck-mounted loader cranes. HMF A/S is in constant development, trying to improve their existing and future products. With truck-mounted loader cranes, this primarily involves reduction of the self weight, and at the same time, maintaining a high lifting capacity. This requires thin plate structures that are reinforced with additional welded plates where necessary. The reinforcement plates are often of complex geometry and there is a need for an efficient engineering tool to estimate the fatigue of the critical welded joints. The theory of Linearly Elastic Fracture Mechanics LEFM is one way to determine the fatigue life of a structure, but it requires knowledge of several parameters to give a correct result. A small change in initial crack length or crack opening stress has great influence on the fatigue life. Measurements of the initial crack length are hard to obtain without using destructive tests. LEFM is therefore rarely used by engineers in design practise. Other methods, such as the Nominal Stress method or the Hot-Spot Stress method are available, but they are hard to apply on the complex geometries. This report contains examples of applying both of these methods on a simple joint. A new method is emerging, called the Effective Notch Stress (ENS) method. This method is easily applied to complex 3D geometries and the required stress ranges are usually achieved by performing a finite element analysis. The ENS-method is, however, still very new and lacks documentation with basis in practical experiments. This report gives an evaluation of the applicability of the method by use of practical experiments compared to the analyses results. The practical experiments are performed on a test specimen that comprises thin plates and complex geometry. The ENS-method and the experimental work correlate well and it is found that this method is an effective tool for fatigue estimations of welded joints. However, more experimental data is required to give a broader statistical safety of the conclusion. Experiments on different post-weld treated specimens are performed to give the basis of a comparison of the methods and their effect of the lifetime for the weldings. Two methods have been investigated in this report; TIG dressing and grinding of the weld toe. Both methods improve the fatigue for the welded joints compared to the as welded test series. Additionally this report recommends fatigue classes for both methods. Due to a limited number of experiments, the recommendations can only be seen as a line of direction. Finally, the report gives an example of how the ENS-method can be applied to make a design improvement of a real cane component. A FE-analysis of the original crane component gave indications to which areas improvement could be made. The improved crane component is designed to fulfil requirements in Danish standards. The design improvement reduces the weight of the original component with about 18 % and maintains the fatigue strength of the original component. In addition to this report there is an appendix section and an enclosure CD with additional documentation that have been used and developed during the project.