Failure Pressure Prediction of Defective Pipeline Using Finite Element Method and Machine Learning Models
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Oil and gas pipeline failure and leakage can seriously damage people's lives and the ecosystem. The prediction of failure pressure for pipelines with damage is one of the most important and challenging tasks faced by industry, which affects the assessment of pipeline safety. Previous studies widely used industrial models or the finite element (FE) method to predict the failure pressure. However, the industrial models may give limited information, and the FE method has much heavy computation burden. In this work, three machine learning models - artificial neural network (ANN), XGBoost (XGB) and CatBoost (CAT) are developed for forecasting the failure pressure of pipelines with defects. Firstly, the simulation results of the FE method are validated by real failure pressure and compared with the calculation results of industrial models (ASME-B31G and DNV). Then 180 pipeline samples including pipeline attributes and defect sizes collected from real in-line inspection data in a pipeline company and the corresponding FE simulation results of failure pressure of these 180 defective pipelines are used for the training and testing of the machine learning models. The results show that the simulation accuracy of the FE method is higher than the calculation accuracy of the industrial models, and the FE simulation results are suitable to be the outputs of machine learning models. The three machine learning methods all provide satisfactory prediction accuracy in failure pressure. Specifically, CAT is the best machine learning method in this study for its lowest relative error (3.11% on average), mean absolute error (0.53), root mean square error (0.78) and highest coefficient of determination (R2) up to 98% in testing. Moreover, the machine learning models present better performance on average relative errors compared to the industrial models. CAT shows higher accuracy than the industrial models and FE simulation on minimum and average relative errors. Finally, the prediction result of CAT is used to discuss the effect of input features on failure pressure of pipelines, which demonstrates that the importance of features follows the order of pipeline thickness > pipeline outside diameter > defect depth > defect length > defect width. Once the above machine learning methods are used in industry, more and more real data will be collected to train a model and make it more accurate. In this way, these methods will provide an efficient way to evaluate the safety of defective pipelines. In addition, the failure pressure of pipeline could be estimated to help operators figure out a pipeline condition and further prioritize the pipelines for maintenance.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it