Approaches for accurate prediction of ductile failure of polyethylene (PE) pipe have been developed. The main objective of this thesis is to characterize deformation and mechanical behaviour of PE material under different conditions which contribute to PE failure. For this purpose, two different loading conditions are considered. One is the mostly used compressive loading in plastic pipe industry, named squeeze-off process. The other is a transverse loading which is a part of a test method to characterize environment stress cracking (ESC) of PE.

In the first part of this thesis, squeeze-off of PE pipe is simulated using finite element modelling. Squeeze-off is a widely used industrial procedure to block or reduce fluid flow in PE pipes. A set of experimental testing data was used to tune and extract elastic-plastic and creep material properties of a finite element (FE) model which consists of a pipe specimen and a squeezing bar. Squeezing speeds of 0.01, 1, and 50mm/min that cover common speeds used in the pipe repair or maintenance were used to model the squeeze-off process. A material sensitivity analysis was performed to identify parameters in the constitutive equations for which change of values yields a sensitive response of the deformation behaviour of PE. This study shows that identifying these parameters improves agreement between experimental data and finite element simulation. The FE model was then used to determine stress and strain distribution in the pipe specimen during the squeezed-off process.

In the second part of this thesis, an indentation loading that is used to generate deep stretch in a PE plate is simulated using FE modelling. Development of the indentation loading is part of a project to design a new test method to shorten the time for crack initiation in PE during the exposure to an aggressive agent. This method is used to accelerate time for characterizing PE’s environmental stress cracking resistance (ESCR). Two cylindrical indenters of 13 and 7mm in diameter were modeled to generate the deep stretch in the central part of a circular area of 15mm in diameter. Through the FE modelling, stress variation and distribution are established during the deep stretch, but without the exposure to the aggressive agent. Data from the experimental testing was used to tune and extract information to quantify material behaviour. Three types of material input data were considered, one purely based on elastic-plastic (EP) deformation, another including damage generation, and the third including creep deformation, all of which were to calibrate input stress-strain curve by regenerating the load-stroke curve obtained from the experimental testing. Comparison of input stress-strain curves for the three types of FE modelling reveals the stress drops due to damage and creep, and their differences caused by the indenter size and loading speed used for the testing. The three types of input stress-strain curves also led to establishment of a stress-strain relationship for PE when no stress drop is caused by damage or creep. The study suggests that the stress-strain curve is more sensitive to the loading speed using the 7mm indenter. The FE modelling also shows more frequent stress relaxation during the deep stretch using the 7mm indenter than the 13mm indenter. Therefore, the study concludes that the 13mm indenter is more effective than the 7mm indenter in transforming the crystalline phase to the amorphous phase through the deep stretch, and thus is expected to generate ESCR with less scattering.