Search Results for neural network model

This study investigates the growth kinetics and spallation behavior of oxide scales formed under steam environments on alloys used in high-temperature plants. The influence of alloy composition is analyzed using two approaches: an empirical model based on the concept of “chromium equivalent” and a neural network model. Both models demonstrate a good correlation with experimental results when sufficient data is available to generate the model parameters. However, there is insufficient data on scale spallation to develop similar models describing the influence of alloy composition on this phenomenon.

This paper outlines a comprehensive UK-based research project (2007-2010) focused on developing fireside corrosion models for heat exchangers in ultra-supercritical plants. The study evaluates both conventional materials like T22 and advanced materials such as Super 304H, examining their behavior under various test environments with metal skin temperatures ranging from 425°C to 680°C. The research aims to generate high-quality data on corrosion behavior for materials used in both furnace and convection sections, ultimately producing reliable corrosion prediction models for boiler tube materials operating under demanding conditions. The project addresses some limitations of existing models for these new service conditions and provides a brief review of the fuels and test environments used in the program. Although modeling is still limited, preliminary results have been presented, focusing on predicting fireside corrosion rates for furnace walls, superheaters, and reheaters under various service environments. These environments include those created by oxyfuel operation, coal-biomass co-firing, and more traditional coal firing.

There is a growing need to automate the gas tungsten arc welding process for fabrication and repair of nuclear components due to an increasing shortage of experienced welders. Therefore, a collaborative effort has been performed in this study to develop a fully autonomous gas tungsten arc welding system with adaptive capabilities. The system employs the application of two neural networks that have been presented in. The first utilizes a vision based convolutional neural network to perform real time control of the filler wire entry position into the weld pool. The second predicts optimal weld parameters and torch positioning for each weld pass deposited within a multi-pass groove. A commercialization path for the technology is in-progress, with the artificial intelligent algorithms currently being incorporated and tested on commercially available equipment.

Laboratory-scale tests are frequently used to generate understanding of high-temperature oxidation phenomena, to characterise and rank the performance of existing, future materials and coatings. Tests within the laboratory have the advantage of being well controlled, monitored and offer the opportunity of simplification which enables the study of individual parameters through isolating them from other factors, such as temperature transients. The influence of pressure on the oxidation of power plant materials has always been considered to be less significant than the effects of temperature and Cr content, but still remains a subject of differing opinions. Experimental efforts, reported in the literature, to measure the influence of steam pressure on the rate of oxidation have not produced very consistent or conclusive results. To examine this further a series of high pressure steam oxidation exposures have been conducted in a high pressure flowing steam loop, exposing a range of materials to flowing steam at 650 and 700 °C and pressure of 25, 50 and 60 bar. Data is presented for ferritic-martensitic alloys showing the effect of increasing pressure on the mass change and oxide thickness of these alloys in the flowing steam loop. In addition the effect observed on the diffusion of aluminium from an aluminised coating in these alloys is also presented and the differences in the extent of diffusion discussed.

Enhanced life assessment methods contribute to the long-term operation of high-temperature components by reducing technical risks and increasing economic benefits. This study investigates creep-fatigue behavior under multi-stage loading, including cold start, warm start, and hot start cycles, as seen in medium-loaded power plants. During hold times, creep and stress relaxation accelerate crack initiation. Creep-fatigue life can be estimated using a modified damage accumulation rule that incorporates the fatigue fraction rule for fatigue damage and the life fraction rule for creep damage while accounting for mean stress effects, internal stress, and creep-fatigue interaction. In addition to generating advanced creep, fatigue, and creep-fatigue data, scatter band analyses are necessary to establish design curves and lower-bound properties. To improve life prediction methods, further advancements in deformation and lifetime modeling are essential. Verification requires complex experiments under variable creep conditions and multi-stage creep-fatigue interactions. A key challenge remains the development of methods to translate uniaxial material properties to multiaxial loading scenarios. Additionally, this study introduces a constitutive material model, implemented as a user subroutine for finite element applications, to simulate start-up and shut-down phases of components. Material parameter identification has been achieved using neural networks.

The highest creep rupture strength of recent 9-12% Cr steels which have seen practical application is about 130 MPa at 600°C and 100,000 h. While the 630°C goal may be realized, much more work is needed to achieve steam temperatures up to 650°C. Conventional alloy development techniques can be slow and it is possible that mathematical models can define the most economical path forward, perhaps leading to novel ideas. A combination of mechanical property models based on neural networks, and phase stability calculations relying on thermodynamics, has been used to propose new alloys, and the predictions from this work were published some time ago. In the present work we present results showing how the proposed alloys have performed in practice, considering long term creep data and microstructural observations. Comparisons are also made with existing enhanced ferritic steels such as Grade 92 and other advanced 9-12%Cr steels recently reported.

Directional coarsening of the γ' phase (rafting) in Ni-based single crystal superalloys during creep at 1273 K was simulated by the phase-field method. The inelastic strain introduced in the γ phase was assumed to be composed of plastic strain (ε p ) and creep strain (ε c ). The simulations were performed with various sets of values of material parameters and the magnitude of external tensile stress. We let a feed-forward neural network learn the simulation data in order to enable fast and exhaustive prediction of the time to rafting, t raft . From the analysis based on the trained neural network, it has been shown that t raft becomes longer with increasing magnitude of γ/γ' lattice misfit, with decreasing creep coefficient, and with increasing yield stress of the γ phase (σγ ys ). The sensitivity of t raft to σ γ ys is high when the ratio of ε p to the total inelastic strain (ε p + ε c ) is high.

This study investigates the influences of product chemistry and grain size on the high-temperature creep properties of 316 stainless steels by analyzing an extensive range of historical and modern literature data. The investigated 316 stainless steel creep property dataset, including more than 160 heats and 2,400 creep testing data, covers a wide spectrum of elemental compositions and product forms. To perform a prudent analysis of the creep property dataset, a statistical overview was first implemented to understand the data distribution relevant to data sources, chemistries, product forms, testing temperatures, and grain sizes. The creep data of 550°C, 600°C, 650°C, 700°C, and 750°C with ±10°C were grouped together, and the analytical study was performed on each sub dataset to investigate the temperature-specific creep performance. The creep strength was evaluated using the average stress ratio (ASR) between the experimental and predicted creep data of tested 316SS heats. The influence of composition and grain size on the creep strength ratio were evaluated using linear correlation analysis. Effects of specified and non-specified elements including C, N, and B were specifically investigated to understand their impacts on the creep strength with regards to the variation of creep temperature. In addition to the literature data, the most recent EPRI creep data of three commercial heats were used to validate the correlations from the historical creep property dataset.

This work concerns a study into the design of creep resistant precipitation hardened austenitic steels for fossil fuel power plants using an integrated thermodynamics based model in combination with a genetic algorithm optimization routine. The key optimization parameter is the secondary stage creep strain at the intended service temperature and time, taking into account the coarsening rate of MX carbonitrides and its effect on the threshold stress for secondary creep. The creep stress to reach a maximal allowed creep strain (taken as 1%) at a given combination of service temperature and time is formulated and maximized. The model was found to predict the behavior of commercial austenitic creep resistant steels rather accurately. Using the alloy optimization scheme three new steel compositions are presented yielding optimal creep strength for various intended service times up to 105 hours. According to the evaluation parameter employed, the newly defined compositions will outperform existing precipitate strengthened austenitic creep resistant steels.

Ni-base superalloys used for hot section hardware of gas turbine systems experience thermomechanical fatigue (TMF), creep, and environmental degradation. The blades and vanes of industrial gas turbines (IGTs) are made from superalloys that are either directionally-solidified (DS) or cast as single crystals (SX). Consequently, designing and evaluating these alloys is complex since life depends on the crystallographic orientation in addition to the complexities related to the thermomechanical cycling and the extent of hold times at elevated temperature. Comparisons between the more complex TMF tests and simpler isothermal low cycle fatigue (LCF) tests with hold times as cyclic test methods for qualifying alternative repair, rejuvenation, and heat-treatment procedures are discussed. Using the extensive set of DS and SX data gathered from the open literature, a probabilistic physics-guided neural network is developed and trained to estimate life considering the influence of crystallographic orientation, temperature, and several other cycling and loading parameters.

In order to enable safe long-term operation, metallic pipes operated in the creep range at high temperatures require considerable wall thicknesses at significant operating pressures, such as those required in thermal power plants of all kinds or in the chemical industry. This paper presents a concept that makes it possible to design such pipes with thinner wall thicknesses. This is achieved by adding a jacket made of a ceramic matrix composite material to the pipe. The high creep resistance of the jacket makes it possible to considerably extend the service life of thin- walled pipes in the creep range. This is demonstrated in the present paper using hollow cylinder specimens. These specimens are not only investigated experimentally but also numerically and are further analyzed after failure. The investigations of the specimen show that the modeling approaches taken are feasible to describe the long-term behavior of the specimen sufficiently. Furthermore, the paper also demonstrates the possibility of applying the concept to pipeline components of real size in a power plant and shows that the used modeling approaches are also feasible to describe their long-term behavior.

During the last decades, new generations of Ni-based superalloys have emerged with judiciously controlled chemistries. These alloys heavily rely on the addition of refractory elements to enhance their mechanical properties at elevated temperatures; however, a clear interpretation of the influence of these minor-element additions on the alloy's high-temperature oxidation behavior is still not well understood, particularly from the standpoint of predicting the transition from internal to external alumina formation. In this context, the present investigation describes a systematic study that addresses the intrinsic effects that minor element additions of Nb, Ta, and Re have on the oxidation behavior of alumina-scale forming γ-Ni alloys. By combining a novel simulation approach with high-temperature oxidation experiments, the present study evidences the generally positive effect associated with 2 at. % addition of Ta and Re as well as the detrimental consequences of Nb additions on the 1100 °C oxidation of (in at. %) Ni-6Al-(0,4,6,8)Cr alloys.

This study investigates the steam oxidation behavior of Alloy 699 XA, a material containing 30 wt.% chromium and 2 wt.% aluminum that forms protective oxide scales in low-oxygen conditions. The research compares four variants of the alloy: conventional bulk material, a laser powder bed fusion (LPBF) additively manufactured version, and two modified compositions. The modified versions include MAC-UN-699-G, optimized for gamma-prime precipitation, and MAC-ISIN-699, which underwent in-situ internal nitridation during powder atomization. All variants were subjected to steam oxidation testing at 750°C and 950°C for up to 5000 hours, with interim analyses conducted at 2000 hours. The post-exposure analysis employed X-ray diffraction (XRD) to identify phase development and scanning electron microscopy with energy dispersive spectroscopy (SEM/EDS) to examine surface morphology, cross-sectional microstructure, and chemical composition. This study addresses a significant knowledge gap regarding the steam oxidation behavior of 699 XA alloy, particularly in its additively manufactured state.