Predicting the slow crack growth resistance of pe resins by means of full notch creep testing in air : confrontation between experiments and modellization

摘要

Nowadays it is well known that Slow Crack Growth (SCG) resistance is of primary importance when selecting a PE for gas or water applications, and there is a lot of laboratory tests which are designed to assess for SCG resistance (e.g. PENT, Cone, Notch,....).rnEach test provides the user with more or less reliable data but the problem remains the complexity when searching for a scientific correlation between all these tests due essentially to the differing conditions from one to another (plane stress, plain strain, surfactant-aided SCG,...) and to the lack of a consistent model which would permit to go from one test to another indifferently.rnIn the early nineties, the Full Notch Creep Test (FNCT) was designed in Japan and since then it has proved to be an excellent candidate to assess for SCG resistance of PE resins. Thus FNCT is now used worldwide, supported by an ISO standard in preparation. In this area Gaz de France R.&D. Division has acquired a significant experience regarding FNCT on both experimental and theoretical levels. On the experimental level, intensive FNC testing have put in light some of the major factors of influence which have to be controlled drastically, among them for example the maximum relevant temperature of the test or the geometry of the specimen or the loading history of the specimen. Under such controlled conditions FNC testing on several representative PE resins currently used onrnthe distribution networks worldwide gives reliablerndata and allows one to obtain data similar to that ofrnhydrostatic pressure tests (HPT), meaning ductilernand brittle failures depending of therntemperature/load conditions, but in significantlyrnshorter test durations compared to HPT, asrnexpected.rnThe experimental approach allows one to rankrnqualitatively the materials in a fairy consistent wayrnconsidering solely the failure time.rnOn the theoretical level, a modelization of the testrnhas been made on the basis of the so-called "localrnapproach" which allows one to describe the SCGrnphenomena very locally around the crack tip.rnAn original 6-parameter elastic-viscous-plasticrnbehavior law - elaborated at Gaz de France in thernmid-nineties and presented elsewhere formerly - isrnconsidered in the modelization.rnPicking up a critical displacement on thernexperimental FNCT curve which corresponds tornthe incubation time, it becomes possible torncalculate a value of the anelastic strain whichrncorresponds to the damage at the crack tip andrnwhose critical value represents the "trigger" forrnSCG.rnTherefore it becomes possible to rank the PErnmaterials more quantitatively considering a hiddenrnrheological parameter which represents a meter ofrnthe internal damage before crack growth.rnThen such quantity can be transported in a plainrnpipe with a preexisting defect to predict its longrnterm resistance under HPT.rnThe overall approach allows one to correlate bothrntests on experimental and theoretical levels.