Skip to main content

Modelling of Interface Evolution in Advanced Welding

Final Report Summary - MINTWELD (Modelling of Interface Evolution in Advanced Welding)

Executive Summary:
Welding is the most economical and effective way to join metals permanently, and it is a vital component of our manufacturing economy. In welding, work-pieces are mixed with filler materials and molten, to form a pool of metal that upon solidification becomes a strong, permanent joint. Our ability to weld a metal to itself and to other materials is determined by the chemistry at the interface and by the complex morphology of the individual crystals at the weld centre. These boundaries are the critical regions where most catastrophic failures occur.

Mintweld consortium made a number of technological advances and step changes in modelling the principal phenomena at different length scales, and integrated our modelling efforts across the full range of welding. We established the capability to design and engineer welding processes with a multi-scale, multi-physics computational modelling approach. An integrated suite of modelling software has been developed and validated, able to describe the key phenomena of the welding process at all relevant length scales, with a special emphasis on the solid-liquid interface evolution, including the description of macro-scale mass flow and thermal profile in weld pool, meso-scale solid/liquid interface movements during melting and solidification of weld joints, micro/nano-scale grain boundary and morphology evolution in the solidified joints, mechanical integrity, and service life of the welded products.

A unique feature of the Mintweld project is the prediction of interface evolution in industrially relevant systems, such as steel/steel and steel/Ni-based alloys. Validation has been ensured by state-of-the-art experimental techniques, including electron microscopy and atom probe measurements to characterise chemistry and morphology in grain boundaries, and real-time synchrotron X-ray imaging, to observe morphological evolution and internal weld flow during welding.

The modelling strategy, software programme and data obtained in the Mintweld project can find widespread application in the relevant European industry for penetrating novel markets of high economic and strategic importance. It is hoped that the developed capability for intelligent design of high performance welded systems and interfaces can help to ensure that Europe maintains its competitiveness in welding and manufacturing industry.
Project Context and Objectives:
Welding is the most economical and effective way to join metals permanently, and it is a vital component of our manufacturing economy. It is estimated that over 50% of global domestic and engineering products contain welded joints and in Europe the welding industry has traditionally supported a diverse set of companies across the shipbuilding, pipeline, automotive, aerospace, defence and construction sectors. Although no official data exists on the revenue of EU welding industry due to the diverse applications of the welds, revenue from welding equipment and consumable markets reached €3.5 billion in Europe in 2007 [1]. However, advancing market maturity and mounting competition from countries in East Asia pose an immediate challenge to the European welding industry. There is no alternative but to compete by generating new knowledge to ensure transformation from a resource-intensive to a knowledge-intensive industry to maintain our world-leading position.

In welding, work-pieces are mixed with filler materials and melted, to form a pool of metal that upon solidification becomes a strong, permanent joint. During solidification of the weld pool, crystals grow from the partially melted work pieces towards the weld centre. The ability to weld a metal to itself (similar welding) and to other materials (dissimilar welding) strongly depends on the chemical segregation and morphology of the advancing solidification interfaces, in particular at grain boundaries where individual crystals meet near the weld centre. The advancing solidification interfaces and the resulting grain boundaries have critical nano and sub-micron/scale features. The boundaries have different alloy concentrations due to chemical segregation during solidification, and are the most dangerous regions where failure occurs by (1) cracking during solidification of the weld pool (hot cracking) or (2) cracking in service (cold cracking), e.g. hydrogen embrittlement. Cracking is the most common failure mode in welds, and many expensive failures have occurred in welded components. For example, during the period 1996 to 2002, the cost due to the hydrogen induced cracking failure in subsea pipelines on the Norwegian shelf is estimated over €400 million Euros [2]. To meet the increasingly challenging end-user demands, a better understanding and control of the interface evolution and the formation of resulting grain boundaries via alloy design and process optimisation are urgently needed.

Welding interfaces and grain boundaries play a crucial role, but many process variables affect the properties of welds. Therefore, a comprehensive model of processing and in-service behaviour of materials needs to pass information across various spatial and temporal scales. To correctly model the interface evolution during welding and the in-service failure of weld components, multi-scale approaches which link atomistic scale ab-initio and Molecular Dynamic models (interface properties and issues of chemistry in the crack) with micro/nanoscale Phase-Field Models (grain boundary and interface chemistry and structure), meso-scale Front-Tracking models (crystal growth, grain size distribution) and macro-scale Computational Fluid Dynamics models (heat and mass flow) are required.

A significant volume of research has been carried out to date on modelling welding phenomena at macro- and micro-scales that incorporate the formation of the weld pool and the advance of the solidification interface, which take into account the effect of various welding process parameters. However, many models do not incorporate the real effects of alloy chemistry and process parameters on the properties of welding interfaces where cracking occurs. The field is fragmented and research activities are largely uncoordinated; thermal models, computational fluid dynamics, solid mechanics of welding have all been developed separately, but are not yet integrated [5] or coupled to microstructural models to provide a more comprehensive model of the welding process.

The objectives of the MINTWELD project are to make a number of technological advances & step changes in the above single-scale models and then to integrate and validate the modelling efforts across the multi-scale range. A major aim of this project is to replace this empiricism by an integrated set of physically based models.

The Mintweld project took advantage of the fluid flow, solidification, phase field and front tracking models to provide thermodynamic, kinetic and element segregation information relevant to the solidification interface evolution and resulting grain boundaries. Ab-initio and molecular dynamics studies provided surface and interface energies. This information was combined within the framework of the Cocks-Suo variational principle [3] to predict component life. Validation of the predictions have been ensured by novel experiments at each scale, including real-time synchrotron imaging to observe morphological evolution of the welding fronts and internal flow in the weld pool, electron probe micro-analysis and atom probe to characterise alloy chemistry near grain boundaries. Strategies for intelligent design of new weld materials with improved properties of the nano-organized interfaces and grain boundaries will be developed and passed on to the welding industry. Controlling the structure and properties in this way offers the opportunity to drastically improve the performance of welds, opening new markets for the EU welding industry. A unique aim of this project will be the prediction for realistic systems, i.e. industrially important steel/steel (structural applications, similar welding) and steel/Ni-based alloys (pipeline applications, dissimilar welding) used in high speed arc and laser welding applications. This will ensure that the outcomes will be of the highest relevance and usefulness to industry.

The output of the research can provide welding industries with an integrated tool to guide the discovery of new materials and/or process routes to manufacture weld components with improved interface properties which drastically improve their performance, through the choice of appropriate components and an improved tuning of the welding process. MINTWELD research outputs is capable of environmentally sustainable development in Europe by enabling intelligent design of new materials and process routes leading to a reduction of cost and energy consumption in the welding industry. The application of the models has been demonstrated in Europe's most advanced welding technological industries and welding institutions. The potential impact of the MINTWELD project would be substantial to the European welding industry.

With the assistance of the new technologies from this project, welds with drastically improved interface/grain-boundary properties, and hence in-service performance, can be produced for demanding applications such as in the oil and gas industry. Due to the diverse application of the welds, it is difficult to estimate the amount of (1) total savings of cost and (2) the increase of revenue or profit margin resulting from this project at this stage. However, in the oil and gas sector, more than 2,400 new sub-sea wells have been installed by 2010 [4], and many of which are deeper or ultra-deep water high pressure facilities. New alloys, such as API X65-X120 steels and 9wt.%Ni-steel, have been developed, but the application of these new alloys in pipelines has been currently limited because of welding cracking. The typical development time for a new welded pipeline product is usually 2 to 3 years in an European leading steel-making industry, during which welding development is a major issue. The overall cost involved in the development is several million Euros. When these pipeline products are in use, the cost for the construction of a pipeline is often up to several billion Euros and the integrity of the pipeline has huge implications for the local energy supply and hence economic prosperity. The integrity of a pipeline is dictated by the welds which are deemed to be the weakest part since welds are prone to interface related defects. Being able to predict and control interface properties during welding, and hence to produce welds with radically improved properties will certainly help improve the productivity of pipeline products and the integrity of the constructed pipeline by using new alloys in conjunction with advanced welding technologies. The economic impact of this project will, therefore, be significant. Another big gain for the European fabrication industry is the possibility to increase the yield rate during fabrication which can have huge cost/energy savings and reduction in CO2 emission, hence giving potential benefits to society and the environment.



Specific objectives:

• A framework for linking and integrating the multiscale modelling efforts which will enable a smooth technology transfer to industry.
• Development of a new computational toolbox of accurate models for modelling interface evolution in welding and for predicting the failure of components under in-service loading conditions.

• Modelling welding processes covering all principal phenomena during melting, mixing and solidification with particular emphasis on the analysis of interfacial phenomena, where most failures occur.

• Modelling of cracking during welding and in-service, especially hydrogen embrittlement of weld component.

• New and unique synchrotron X-ray real-time imaging experiments to provide enhanced model validation at the appropriate length and time scales.

• Evaluating the effect of interface structure and grain-boundary chemistry on the failure (cracking and hydrogen embrittlement) of welded components through detailed Electron Microscopy and Atom Probe Tomography (APT)

• Development of a database of key welding process attributes and thermodynamic properties for selected realistic multi-component alloy systems (steels and Ni-based alloys).

• Strategies for intelligent design of new weld materials and processes under real working conditions.
The project has been managed through co-ordination and integration of modeling efforts within the Mintweld team by distributing the partner tasks within 8 discrete work-packages, with defined deliverables.

The major deliverables include:
• An integrated modelling tool to accelerate the development of new welding fillers and processes at a reduced cost.

• Capability to predict optimal chemistry for welding materials (filler) and to design welding processes

• Improved understanding of microstructure development and interface evolution during welding, and the effect of alloy chemistry and process variables on the structure development and the interface evolution.

• Mechanism of failure in welded components, in particular for hydrogen embrittlement.

• Database of material property data, models and knowledge developed through the project


References

1.Growth Partnership Service: Industrial Automation and Process Control, London Business Wire, June 17, 2008

2.Norwegian Petroleum Directorate: Overview report on - Risk Level Development- North Sea, Norwegian sector. April 2003

3.Cocks ACF, Gill SPA, Advances in Applied Mechanics, (1999), 36, 81-162.
4.Infield Systems of London: Sub-sea Market Update To 2010
Project Results:
Please see the attached pdf file
Potential Impact:
The developed and validated models will provide a comprehensive description of key phenomena in welding, and the influence of them on the performance of weld components. The availability of fully characterised relationships between interface, boundary and macroscopic properties, together with the ability to control interface structure, will allow tailoring of alloys and welding processes to specific engineering applications. Realistic systems (Steel, Ni-base alloys) have been studied by Mintweld industrial partners. When coupled with the data generated from the experiments in the project, the models is able to guide the development of new alloys and result in radically improved properties and processes across a number of industrial sectors that are very important to the European economy.

The principal industrial beneficiaries will be the European welding industry and the manufacturing chains that use welding. As the producers and end-users of advanced fabricated engineering systems, they stand to gain both financially, through reduced part rejection and improved performance of welded components, and technologically, through increased understanding and control of the dominant interface and grain boundary phenomena. Competitive advantage will be gained via resultant increased weld component life and, crucially, via the avoidance of costly (including sub-sea) failures of novel welded systems. The proposed research also contributes to all applications of metal processing by providing a greater fundamental understanding and prediction of interface evolution and grain boundary formation.
In the following sections, we will describe Mintweld’s scientific, industrial, economic, and social impacts.

1. Scientific impact
Our project has established the capability to design and engineer welding interfaces with a multi-scale, multi-physics computational modelling approach. An integrated suite of modelling software has been developed and validated, able to describe the key phenomena of the welding process at all relevant length scales, with a special emphasis on the solid-liquid interface evolution, including the description of macro-scale mass flow and thermal profiles, meso-scale solid/liquid interface movements, micro/nano-scale grain boundary and morphology evolution, mechanical integrity, and service life of the welded product. The MINTWELD project will provide an enhanced understanding of the physical and chemical processes taking place at liquid/solid interfaces during welding and the resulting grain boundaries in welds. In particular, the influence of individual elements and process variables on hot cracking during welding and hydrogen embrittlement in service will be identified. Scientific impacts from modelling further development of individual scale models and the integration of these models are described below:
At macro-scale, Computational Fluid Dynamics (CFD) modelling will take into account a wide range of physical processes to relate the controllable variables to the characteristics of the weld pool. In particular, the influence of hydrodynamics on the formation of fusion weld pools will be modelled, treating the instabilities and turbulence observed in the liquid surface. The dynamics of Solid/Liquid (S/L) interface evolution will be solved to give the local chemistry and the flow of molten liquid in the weld pool, through linking the macro-scale models with the meso-scale solidification model. This development will provide an accurate description of temperature, flow and S/L interface movement in the weld pool.
At meso-scale, grain size and distribution will be predicted using a Front-Tracking (FT) model with the input of thermal and flow data from the CFD modelling. In our study, the prediction of crystal growth rate will be validated by real-time synchrotron X-ray imaging. This will be the first in-situ validation for welding (high-rate solidification). Accurate description of grain morphology and meso-scale chemistry segregation can be obtained for micro-scale solidification modelling.
At micro-nano scale, an adaptive meshing technique will be used in Phase Field (PF) models. The adaptive meshing method is not well developed in Europe. Prof Dantzig, who recently joined EPFL from the USA, will fill this deficit. The PF models will be extended to multi-component real alloy systems to study chemical segregation and morphology evolution during solidification with special emphasis on the late stages of solidification when the hot cracking occurs. Adaptive meshing PF models can simulate solidification microstructures in parameter regimes previously inaccessible through conventional fixed-grid techniques, enabling a complete description of solidification interface evolution. Predicted data will be validated by the measured results of chemistry near/at grain boundaries from (S)TEM and Atom Probe. Accurate chemistry and morphology information at the solidification front can be passed on to atomistic scale Molecular Dynamic models. An adaptive meshing Phase Field Crystal (PFC) model will be used to bridge the length scales between micro-scale continuum models and atomistic-models.
A realistic description of processes which take place during the solidification of the weld pool and formation of grain boundaries, as well as cracking and embrittlement, cannot be achieved without atomistic modelling of solid-liquid and solid-solid interfaces. The developed Molecular Dynamics models can provide detailed information about interface structure and thermodynamic properties of alloy systems. The precision of these models is sufficient to resolve the anisotropy in the computed quantities, which is crucial to the PF prediction of the morphology of the advancing interfaces during solidification. Ab-initio models will be developed through MINTWELD for more complicated multi-component systems, since the use of molecular simulation techniques is limited by the lack of appropriate description of inter-atomic interactions, or force fields. In such cases the ab-initio approach provides a way forward, both in terms of quantum mechanical descriptions of local solidification and cracking events, as well as for the development of more accurate force fields for classical molecular models. Highly efficient computational models will be developed through MINTWELD based on the established framework of density-functional theory (DFT), which will allow performance of fully quantum mechanical simulations on complex metallic systems to be performed, containing hundreds of atoms. The data generated through the project will be valuable for welding industry to design new materials.
The predicted weld pool geometrical and mechanical restraints, grain and phase structure, chemistry and morphology at the interface, inter-atomic interactions will be used to predict hot cracking during welding and hydrogen embrittlement in-service. We will develop (1) the first physically and thermodynamically consistent models of hot cracking during weld pool solidification, and (2) the first models of hydrogen embrittlement which take into account the competition between cleavage and strain localization / micro-voiding which is currently being disputed. The availability of each of these models will provide a step change in our understanding of hot cracking and hydrogen embrittlement and will promote the replacement of current empirical approaches to alloy design and process selection in industry with the more effective knowledge-based approach.
The models and strategy developed in this project will be useful for similar efforts in a wide range of academic research programmes in other FP projects.

2. Industrial and economic impact
Although no official data exists on the revenue of EU welding industry due to the diverse applications of the welds, it is estimated that over 50% of global domestic and engineering products contain welded joints, and welding equipment and consumable markets reached €3.5 billion in Europe in 2007 [1]. In Europe the welding industry has traditionally supported a diverse set of companies across the shipbuilding, pipeline, automotive, aerospace, defence and construction sectors. The demand for improved properties of welds is set to rise substantially. For example, in the oil & gas sector, some 2,400 new sub-sea wells are anticipated to be installed by 2010 [2], many of which will be deeper or ultra-deep water &/or high pressure facilities. Total expenditure is likely to be >€50bn [3] on these latter facilities alone in the period 2007-2011.
The potential impact of MINTWELD project will be substantial on the European welding and manufacturing industry. With the assistance of the integrated model and strategy developed in this project, welds with drastically improved interface (fusion line) properties and hence in-service performance will be produced for demanding applications such as in the automotive and oil & gas industry. For example, the typical development time for a new submerged arc welded (SAW) linepipe product could easily take up to 2 to 3 years in a leading European steel industry. It is estimated that the overall cost involved could run up to several million Euro for each new product development. This is sometimes virtually impossible to confirm due to confidentiality. It should be noted that the welding industry still employs empirical rules and trial-and-error approaches in the design of weld materials and the optimisation of process parameters, even in a leading EU steel industries. A direct up-take of the output will be facilitated by the participation of industry members and welding institutes. If the development time can be shortened by half with the assistance of the developed tool, then a couple of million Euros can be saved for each product development. In Europe, we have about 2000 welding industries, so the total savings will be significant.
Although no data for the in-service failure of welded components in EU are publicly available, various reports quote the cost of environmentally-induced cracking (hydrogen embrittlement for sub-sea pipelines) to major western economies (USA, UK, Japan) as between 3-5% GDP. In the US, an economy of similar size to that of the EU, the annual cost of cracking and corrosion failure is estimated at $276bn [2]. A significant proportion of this cost can be attributed to component failure in extreme environment, resulting from the increasing number of deepwater oil and gas field developments. This decade alone, several major hydrogen embrittlement failures have occurred in major oilfields resulting in lost production valued in several billions Euros, whilst exposing the EU to increased petroleum prices & increasing EU dependency on oil & gas supplies from geo-politically unstable regions.
As an example, the cost for the construction of a pipe line is often enormous to the amount of up to several billion Euros. Currently there is a big challenge for welding industry in meeting low temperature (down to –40 °C) requirements (e.g. minimum mean toughness of 45J for new steel grades) for thick walled welded linepipes. Being able to predict and control the formation of critical interfaces during welding, and hence to produce welds with radically improved properties will certainly help raise the productivity and definitely change the concept and mentality of engineering design with respect to the use of new high strength steels and alloys in conjunction with conventional as well as advanced welding technologies. It was estimated [4] that billions of Euros can be saved through building cost-effectively more efficient & safer sub-sea infrastructure by means of avoiding hydrogen failure for sub-sea pipelines.

3 Impact on Environment and Legislation
In order to respect the Kyoto Protocol on the reduction of greenhouse gases in an effort to prevent anthropogenic climate change [5], the European Commission (EC) has taken many climate-related initiatives to limit carbon dioxide (CO2) emissions and improve energy efficiency [83]. For example, a new legislative proposal recently developed by the Commission defines that the average emissions of CO2 from new passenger cars (which account for about 12% of the European Union's carbon emissions) in the EU will be reduced from currently around 160 grams per kilometre to 130 grams per kilometre by 2012. That will translate into a 19% reduction of CO2 emissions and will place the EU among the world leaders for fuel efficient cars [4]. This has since posed a great challenge to the EU car manufacturers and steel makers in developing and adopting new high strength steels in order to reduce the body weight of new generation cars. One of the main difficulties faced by the industry is the weld-ability of the new materials, which has incurred, and will continue to do so, considerable amount of research and development effort. It is foreseen that the innovative multi-scale modelling tools to be developed under the current project will enable European car manufacturers have to speed up the adaptation procedure of new materials by solving their weld-ability issues, and therefore meet the legislation requirements.
In gas and oil sector, hydrogen embrittlement in pipelines has been responsible for significant delays in hydrocarbon production threatening the energy availability and security for European nations. In 2005, gas supplies to the UK, for example, became worryingly low, prompting heating warnings and limitations placed on industrial users. Energy supply to the EU relies upon the timely exploitation of existing and new oil and gas wells. Delays caused by unexpected failures can result in energy crises such as that noted above, with subsequent impact on basic quality of life. Loss of pipeline containment is believed to be responsible for some 68,624 barrels of oil spilled globally every day [3], with significant environmental impact.
The application of the project will facilitate significant growth in employment, turnover and profitability amongst the partnership and other EU industry. Because of the services they will be able to sell across all the process industries not only throughout the EU, but to the wider global market. Specifically, the participating Welding Institutes and Industries will be able to sell high value added solutions with advanced functionality and performance in a sustainable and growing market. The significant knowledge base assembled by the partnership will maintain a high level of competitiveness, thus maximising job creation within a high technology, high value fabrication industry.
Europe can no longer afford to rely upon an Edisonian style trial-and-error based optimisation of its products; instead, it must rely increasingly upon computer-based simulation methods which, provided they are based upon the correct physics, are carefully validated, and are backed up by databases, will bring knowledge-inspired decision making to production routes. In short, Europe must translate its technological heritage in welding theory and process innovation into knowledge-based computer-aided manufacturing simulation tools if it is to maintain its world-leading position within the welding industry.

References
1. Growth Partnership Service: Industrial Automation and Process Control, London Business Wire, June 17, 2008
2. Infield Systems of London: Sub-sea Market Update To 2010
3. Infield Systems of London: Deep & Ultra-deepwater Oil & Gas Market Update Report 2007/11
4. Norwegian Petroleum Directorate: Overview report on - Risk Level Development- North Sea, Norwegian sector. April 2003
5. http://unfccc.int/kyoto_protocol/items/2830.php
List of Websites:
www.le.ac.uk

Professor Hongbiao Dong
Department of Engineering
University of Leicester
Leicester, LE1 7RH
UK

Tel: +44 116 2522528
Fax: +44 116 2522525
E-mail: h.dong@le.ac.uk
www.le.ac.uk/eg/hd38