| Popis: |
A technical review is presented of corrosion assessment models for determining the remaining strength of pipelines containing metal loss or corrosion defects. Initially, burst pressure prediction models for defect-free pipes, including strength and flow theory determined burst solutions and their experimental validations were considered. Next more accurate burst pressure predictions that cover a wide range of pipeline grades were discussed by comparing thick wall burst solutions and machine learning models with full-scale burst data that cover a wide range of pipeline grades up to X120. Corrosion assessment models were categorized into three generations with respect to the reference stress. The 1st, 2nd, and 3rd generation models correspond to the flow stress, UTS, and both UTS and strain hardening rate. Seventeen corrosion models were evaluated and compared with full-scale burst data for a wide range of pipeline steels from low to high grades. The focus was on validating two recently developed 3rd-generation models that can determine more accurate burst strength for corroded thin-wall pipes. As the 4th-generation models, three typical thick-wall corrosion assessment models and machine learning models were introduced to predict more accurate results for corroded thick-wall pipes. After that, advances in corrosion model development were discussed, including constraint effect, bulging factor and defect width effect. Finally, major technical challenges in development of corrosion assessment models were discussed with regard to full-scale experimental tests, numerical simulations, material failure criteria and real corrosion defect assessment. |
