The use of buried pipelines is increasing rapidly in the transportation of oil and gas in both offshore and onshore conditions. In order to increase the performance of the pipeline. the temperature and pressure of the gas or oil have been increased. This produces a compressive axial load in the pipeline and tends to cause the pipeline to buckle in the vertical direction when buried (upheaval buckling). Despite a large body of recently developed theoretical models, no experimental work has been reported (at least in public domain) relating to upheaval buckling of buried pipe in a real soil for verifying such models. This report summaries an experimental and theoretical investigation into the upheaval buckling behaviour of pipelines at the school of Engineering, University of Manchester. 1995-1998, utilising the geotechnical centrifuge in the Peter W Rowe Laboratory. The experimental set-up has been outlined in the report, believed to involve the longest pipe sections yet tested in a centrifuge (1800mm) which utilise the maximum length available in the Manchester machine. Thin-walled stainless steel pipes with 6mm diameter have been heated in a rig with a straight pipe length of 1800mm between fixed end points. The pipes have been heated by the passage of water pumped from a reservoir which can be heated while the centrifuge is running and generating different accelerations. At sustained accelerations during cycles of heating and cooling the axial load, temperature distribution and vertical profile of the pipes have been recorded. The experimental database established from the centrifuge modelling covers unburied models for different initial out-of-line pipe imperfection in various accelerations and buried nlodels for depths of loose and dense cohesion less soil cover up to three pipe diameters at 1 g and multi g conditions. Also the submergence of the model has been examined. Some pull-out tests have been carried out to measure the uplift resistance of soil cover directly. A number of recently developed analytical models have been reviewed and modified to apply to the restrained pipe situation. Some of the models have been validated agamst experimental data. Attempts were however made to qualify and quantify the transverse loading provided by the soil on the pipes. since this is a critical parameter in the selectIOn of depth of soil cover in the field and one which is the subject of much debate III the literature.