AMPP Store - Products tagged with 'machine learning'

Available for download

Estimating Corrosion Rate Risk Distributions using Machine Learning and Geospatial Analytics

Product Number: 51320-14640-SG

Author: Joseph Mazzella, Thomas Hayden, Haralampos Tsaprailis, Len Krissa

Publication Date: 2020

$20.00

Estimating corrosion growth rates for underground pipelines is a challenging problem. There are confounding variables with complex interaction effects that may result in unexpected outcomes. For instance, the relationship between soil conditions and AC interference is highly non-linear and challenging to model. This work expands upon prior work using a suite of machine learning tools to estimate corrosion rates. However, instead of estimating a single corrosion growth rate for a single girth weld address (GWA), this work estimates a distribution of potential corrosion growth rates. Modeling distributions provide a more effective risk-measurement framework, especially concerning high volatility or areas of severe tail risk.

This work relies heavily on machine learning and geospatial tools - particularly artificial neural networks and gradient boosted trees to estimate the corrosion rates and non-linear processes. Building upon prior work using data from a North American Operator, the models in this paper use additional variables from recent research in AC interference and microbiologically influenced corrosion to construct a higher accuracy and distribution-based model of pipeline corrosion risk.

Available for download

How Not to Do Machine Learning for Corrosion Prediction

Product Number: 51323-19310-SG

Author: Thomas Hayden, Joseph Mazzella, Len Krissa, Paul Murray

Publication Date: 2023

$20.00

The intention of this work is to pose epistemic questions about corrosion measurement, statistical inference, and the role of machine learning in predicting corrosion growth. The audience of this work is practitioners implementing inferential algorithms or tools for corrosion prediction. In this work, an algorithm consists of a process for estimating the presence and severity of corrosion.

Available for download

Improving Cathodic Protection Pipeline Integrity Monitoring Data in the Time of IIoT and Big Data

Product Number: 51321-16259-SG

$20.00

Available for download

IR 4.0 Integrity Management Using Data Analytics

Product Number: MPWT19-15487

Author: Dr. Haaken Ahnfelt, Dr. Luis Caetano, Dr. Hilde Aas Nøst, Dr. Knut Nordanger, Reidar Kind, Dr. Zeeshan Lodhi, Dr. Lay Seong Teh

Publication Date: 2019

$0.00

One of the pillars of the fourth industrial revolution (4IR) is to let machines make decisions on behalf of humans; this paper describes new technology that allows machines to decide inspection programs and field validation and testing of results. The technology described is a part of integrity management, and uses data, statistics and expert decisions to design inspection programs. These inspection programs are an important part of the safeguarding of equipment to maintain production and safety.
This technology is a data-driven predictive model of material loss from corrosion, based on domain expert input and historical data in the form of non-destructive testing (NDT) tests. The technology trends is based on historical data and SME input, while accounting for uncertainties in NDT measurements, with uncertainties in historical trends and uncertainties in future trends. This produces a more realistic failure prediction to enhance existing RBIs and adds safety by improving on early detection of trends in data. In total, this enables the machine to update inspection plans autonomously, reducing the number of inspections significantly.
The paper also describes how the technology can be developed further to use production data and integrity operating windows to improve predictions, deal with localised corrosion and assess if the test points on a corrosion circuit are sufficient, can be reduced in number or should be manually evaluated by adding more test points.

Available for download

Leveraging Physical and Virtual On-Aircraft Sensors to Inform Maintenance Practices

Product Number: 51324-20921-SG

Author: Rebecca Skelton Marshall; Dan Christy; Floyd Steele; Fritz Friedersdorf; Sam Kunselman

Publication Date: 2024

$40.00

Time-based inspection and maintenance intervals are a conventional method of corrosion monitoring on aircraft. However, as corrosion processes do not necessarily occur in scheduled events, these conventional maintenance practices can lead to over- or under-estimation of costly inspections. A shift toward evidence-based and data driven predictive maintenance using a combination of on-asset monitoring devices and component-level models could improve efficiency and reduce total ownership costs. In particular, a “virtual sensor”, i.e., a trained model to predict the corrosion at a given location on the aircraft, can be leveraged to optimize the placement of physical real-time monitoring devices. This digital twin process can be applied to determine the corrosion susceptibility of a single aircraft, or to conduct a fleet-wide analysis. In this work, sensing device measurements deployed at varying locations will be used to demonstrate the applicability of severity tracking, through data-driven machine learning models. In particular, models will be trained on environmental parameters and leveraged to predict current (corrosion rate).

Available for download

Machine learning based NDT data fusion to detect corrosion in reinforced concrete structures by robot-assisted inspection

Product Number: 51321-16392-SG

Author: Patrick Pfändler/Ueli Angst

Publication Date: 2021

$20.00

Available for download

Multitask Learning Framework For Screening Of Corrosion Inhibitors For Mild Steel

Product Number: 51322-17757-SG

Author: Abhishek Agarwal, Pradeep Rathore, Dharmendr Kumar, Sri Harsha Nistala, S Sanjana, Vinay Jain, Beena Rai

Publication Date: 2022

$20.00

Corrosion of metallic structures is a ubiquitous problem in industries such as power generation, oil and gas, pulp and paper, metals processing etc. which also results in significant financial losses. According to the National Association of Corrosion Engineers (NACE) International report, the global cost of corrosion was ~ 2.5 trillion USD in 2013 - close to 3.4 percent GDP of the entire world. The use of corrosion inhibitors is one of the most effective and economical ways to mitigate corrosion of metal and alloy components. Corrosion inhibitors are substances that are added in small quantities in corrosive media to protect metal and alloy components from corrosion.

Available for download

Pitting Corrosion Detection by Ultrasound Monitoring

Product Number: 51324-20810-SG

Author: Magnus Wangensteen; Ali Fatemi; Tonni Franke Johansen; Erlend Magnus Viggen

Publication Date: 2024

$40.00

Early diagnosis is essential for successful mitigation of pitting corrosion, a localized form of corrosion that causes cavities and structural failure in metallic materials. Although ultrasonic inspection techniques are effective in detecting uniform wall thinning, they have challenges in accurately identifying pitting corrosion. The present work proposes a technique for early-stage pitting detection utilizing time-lapse pulse-echo signals. The generation of two-dimensional time-lapse images of ultrasonic reflectivity may be achieved by capturing several ultrasonic traces over a period of time. These images can then serve as input for a neural network trained specifically for the purpose of pitting diagnostics. To obtain a substantial training dataset required for training the machine learning model, a drilling experiment was conducted, and random time-ordered combinations of the pulse-echo measurements produced the time-lapse images. A classification neural network was trained to detect the presence of pits, while a separate regression network was trained to estimate the pit depth. Test data from a previously unseen transducer indicates that the pit depth estimations exhibit a mean absolute error of less than 0.2 mm. All pits are reliably identified when they exceed the defined pitting threshold of 0.5 mm by a depth of 0.1 mm.

Available for download

Predicting Corrosion Severity of Pipeline Steels in Supercritical CO2 Environments Using Supervised Machine Learning

Product Number: 51324-20803-SG

Author: Emily Seto; Meifeng Li; Jing Liu

Publication Date: 2024

$40.00

The importance of effective corrosion management in carbon capture, utilization, and storage (CCUS) networks has significantly increased. Captured CO2 is often transported in the supercritical state (s-CO2) and can contain impurities like H2O, O2, SOx, or NOx. While repurposing existing oil and gas pipelines for s-CO2 transport has been suggested, further testing and risk assessment is required to validate this strategy and its associated risks. Given the substantial amount of corrosion data available from recent corrosion studies, machine learning (ML) has emerged as a promising tool for corrosion prediction and management. This study aims to utilize supervised ML techniques to predict the corrosion severity of pipeline steels operating in s-CO2 systems. The selected algorithms, random forest (RF), K-nearest neighbor (KNN), and support vector machine (SVM) were trained on a comprehensive data set of X-series pipeline steels which includes corrosion rates, impurity levels, temperatures, pressures, and exposure times. Additional testing data set and error and accuracy scores were used to determine the most accurate algorithm. An additional experimental testing was performed to verify the predictions of the model. It was found that the RF model had the best accuracy of 65.3% out of the three tested models and KNN had the worst accuracy of 59.2%. In multiple impurity environments the RF model was able to accurately predict corrosion severity but overestimated corrosion severity in environments with short exposure times.

Available for download

Predicting Long-Term Exposure Performance of Galvanized Rebar Based on Artificial Intelligence and Electrochemical Methods

Product Number: 51324-21166-SG

Author: Deeparekha Narayanan; Yi Lu; Victor Ponce; Homero Castaneda

Publication Date: 2024

$40.00

In this work, we carried out electrochemical studies on ASTM A615 (bare steel rebar), ASTM A767 (steel rebar with hot dip galvanized zinc coating), and ASTM A1094 (steel rebar with continuously galvanized zinc coating) rebars exposed to two different environments. In one condition, the samples were exposed to a simulated concrete pore solution (SCPS) containing 3.5 wt.% NaCl. Over a period of 12 months, the electrochemical properties of the samples were regularly assessed through open circuit potential (OCP), linear polarization resistance (LPR) and electrochemical impedance spectroscopy (EIS) on a weekly basis. In the other condition, steel rebars were embedded in concrete with water-to-cement ratio of 0.53. A controlled surface area of the cast concrete block was exposed to a 3.5 wt% NaCl solution using a dam mounted on it. This method allowed for the introduction of chlorides into the reinforced concrete while maintaining control over the exposure process. Under these conditions, the rebars were continuously monitored by carrying out OCP and EIS tests for a period of up to three years since curing. Based on the experimental results obtained, we developed a mathematical framework that combines mechanistic and machine-learning concepts for analyzing the behavior of the rebars in both conditions. EIS analysis was utilized to quantify the transports processes, activation, and interface interaction of the rebars with the corrosive environments in each condition. EIS served as tool to quantify the transports processes, activation mechanisms, and interface interaction of the rebars within corrosive environments across diverse conditions. We conducted this analysis using a Time Series Prediction (TPS) approach of several phase angle plots along 300 days of rebars in pore solution and 900 days of rebars in reinforced concrete, which leveraged recurrent neural networks techniques to predict corrosion mechanisms. This approach allowed us to learn dynamically from real-time measurements, eliminating the sole reliance on domain expertise for parameter optimization. Finally, we utilized our comprehensive experimental-theoretical framework, which integrated Electrochemical Impedance Spectroscopy testing, to make long-term predictions for the performance of the rebars using neural networks techniques. These predictions spanned several years and were based on rigorous analysis. To validate the accuracy and reliability of our framework, we compared the predictions with the experimental results, thereby confirming the accuracy and reliability of our predictions.

Available for download

Predictive Analytics and Machine Learning in Integrated External Corrosion Management

Product Number: 51323-19393-SG

Author: Thomas Hayden, Joseph Mazzella, Keith Parker, Christophe Baete

Publication Date: 2023

$20.00

Managing external corrosion, especially for underground assets, is a significant challenge dating back to the first underground pipeline in 1865. The very first issue of the journal, CORROSION, featured a headline story on this subject. This subject is fundamental for corrosion engineers and pipeline operators.

Available for download

Towards Automated Development and Evaluation of Optimal Corrosion Inhibitor Products

Product Number: 51323-18783-SG

Author: Hector Klie, Arturo Klie, Ana Ferrer, Yolanda De-Abreu

Publication Date: 2023

$20.00

In most engineering and scientific applications, machine learning (ML) or artificial intelligence (AI) methods in general, are primarily oriented to design a statistical/heuristic procedure to predict the outcome of a system under new conditions. This mechanism aims at exploring non-evident correlations between inputs and outputs that are embedded in the data. However, a large body of this effort relies on black-box function approximations (e.g., neural networks) that have shown limitations to elucidate additional insights from the underlying physical process that generated the data. Thus, this type of knowledge is generated in a data-driven manner without fully explaining the physics governing the problem.

Products tagged with 'machine learning'

Association for Materials Protection and Performance