Future for fossils

Materials World magazine
,
1 Aug 2009

Professor Rachel Thomson of the Department of Materials at Loughborough University, UK, describes methodologies for the prediction of power plant lifetimes.

There is growing recognition of the need for conventional fossil fuel power generation systems to be part of the technologies used during the transition to a lower carbon economy. The energy White paper issued in May 2007 stated, ‘We will continue to need fossil fuels as part of a diverse energy mix for some time to come. But in order to meet our carbon reduction goals, sources such as coal and gas must become cleaner’. To guarantee security of supply during the introduction of new technologies, it is essential to maintain existing fossil-fired plant and, where possible, extend their life, while minimising carbon dioxide emissions.

Many of the conventional fossil-fired power plant design codes were conservative, providing opportunities to develop R&D tools and condition monitoring methods to extend life and reduce maintenance costs. These tools can also developlifetime prediction methods to increase plant flexibility, such as by enabling co-firing with waste and biomass. In the future, more emphasis will be placed on new-build, high efficiency power plant, including CO2 capture and storage technologies, possibly via retrofit. These plants will operate under more aggressive environments and conditions than conventional plant and require new life assessment methodologies for critical components.

A UK consortium comprising the Universities of Bristol, Cranfield, Loughborough and Nottingham, along with 10 industrial partners from the power generation supply chain, is working to deliver and validate an integrated suite of lifetime prediction tools and methodologies that will allow the extension in lifetime of current fossil fuel-fired power generation plant. A number of underpinning technologies are required, such as an understanding of microstructural degradation in critical components, condition monitoring tools, modelling of mechanical behaviour and surface engineering to understand and mitigate environmental degradation.

Achievements include the development of
• a holistic structural integrity assessment procedure to predict weld failures in power plants
• models for complex coatings systems which unify previously separate strands of chemical, metallurgical, microstructural and mechanical predictions
• unique quantitative data and models for fireside corrosion of heat exchanger materials for coal/biomass fuels and operating conditions
• an online monitoring method for hot corrosion
• a miniaturised sample extraction technique for small-scale sampling of power plant components. These tools can be used in steam and gas systems.

Cracked header – case study

A header is a major component in a coal-fired power station, combining the flow of steam from the elements in the boiler into one flow driving the turbines of the power station. It is a complex construction comprising a number of cylindrical steel barrels butt-welded together, each with an outside diameter of 450mm and a wall thickness of 50mm. Several hundred connections are welded onto the header, which itself is over 25m long and weighs 18t.

This case study concerns a header designed to BS1113:89 for a steam pressure of 17.5MPa, an operating temperature of 580ºC and a service life of 150,000hrs. The header was manufactured from Grade 91 steel, which contains a high concentration of alloying elements, particularly chromium, and was introduced into the UK power industry in the late 1980s. It replaced components made from lower alloy steels, which were experiencing problems with cracking.

This particular header was a retrofit component installed in a power station in 1992. Routine maintenance in 2004 found cracks after only 58,000hrs of service. The cracks were small so it was possible to repair them safely. However, they had reappeared by 2006, leading to the complete replacement of the header assembly in 2008. An inspection showed that Type IV cracking had occurred. This is where creep cracks initiate and grow within the heat-affected zone adjacent to the welds.

Premature cracking of this type is caused by service stresses and temperatures, and the properties of the steel itself. Detailed investigation showed that the temperatures or stresses were unlikely to be the sole contributing factors. Work carried out at Loughborough University has established that the concentration of minor elements, such as nitrogen and aluminium, in Grade 91 steel is important in determining its mechanical properties. The test certificates show the cracks occurred in parts of the component with the lowest nitrogen to aluminium ratio. This provides a simple tool to identify components that might be susceptible to premature cracking and need more detailed inspections.

Further research has used advanced characterisation techniques to examine why the concentration of these minor elements has such an influence on creep life. This has led to an in-depth understanding of the evolution of the fine scale structure of the steel. Furthermore, the structures observed in laboratory experiments under accelerated conditions have been benchmarked against those obtained under real service conditions and the results from miniaturised mechanical testing, giving confidence in ranking the behaviour of different components.

Technologies to improve understanding of the Type IV cracking have been developed in four areas –
• A comprehensive understanding of the factors controlling the mechanical properties of different Grade 91 steel casts by Loughborough University.
• Finite element simulation of the temperature, stress and microstructures developed during welding.
• The development of models of creep damage and cracking in service at Nottingham University.
• Direct experimental techniques to measure the stress state in-situ and remove small samples from deep within a welded component to assess mechanical properties at Bristol University.
These methodologies are being applied in the development and lifing of new steels, which will be used in future plant.

Coatings for gas turbine blades – case study
Gas turbine engines are widely used in industrial power generation and will continue to be a part of UK energy. The demands placed on the engines are changing, such as the need to cycle more frequently, to use ‘dirty’ fuels and, most importantly, to operate with increased efficiencies to reduce emissions. These engines operate at gas temperatures of 1,150-1,450ºC, necessitating materials that can withstand high temperatures and aggressive operating environments.

Nickel-based superalloys are usually employed for turbine blades because they have excellent temperature resistance and retain high strength at operating temperatures. However, to provide additional protection against oxidation and corrosion, and minimise the risk of service failures, these alloys are often coated.

The coating systems comprise a metallic coating rich in elements such as chromium and aluminium, which is used on its own for corrosion resistance or as a bond between the substrate and a ceramic thermal barrier coating. During service, in an engine, many changes can occur in the coated blade system due to the mixing of elements between the coating and the superalloy. This can affect the mechanical properties and useable service lifetime of the component. Therefore, it is necessary to optimise the coating and blade material to suit the operating conditions.

The aim has been to model and characterise experimentally the environmental degradation of selected coated turbine blade systems to improve life prediction and failure assessment. Service conditions were simulated at Cranfield University by thermally treating samples in combustion gas environments. The resulting oxide at the outer surface of the metallic coating has been quantified at Bristol University.

A model has been developed at Loughborough that is capable of predicting the microstructure of a coated blade system as a function of time. The key components of this model are a one-dimensional finite-difference based diffusion solver to calculate the distribution of alloying elements, a power-law based growth rate model for predicting surface oxidation and a thermodynamic equilibrium calculation routine to determine the phase evolution with time and space within the substrate-coating system. In addition to forecasting the evolution of concentration and phase profiles in the system after a given thermal history, the model can estimate the losses caused by oxidation and the remaining life of the coating based on a concentration and/or phase fraction dependent failure criteria.

The model has been successfully applied to a number of practical systems to simulate their behaviour under service conditions, and validated by experimental quantification using advanced analytical techniques for laboratory aged and ex-service materials. The microstructural model has been used as an input into a model for mechanical behaviour at Nottingham University, implemented using a finite element scheme that takes into account the properties of the various phases, complex blade geometries and applied external loads. This allows, through the definition of a failure criterion, prediction of the likely remaining life of the coated system.

An advanced and holistic approach has been developed to model complex coatings systems, unifying previously separate strands of chemical, metallurgical, microstructural and mechanical predictions. It is possible to forecast, using parameters such as the chemical compositions of the coating and superalloy system, coating thickness and the exposure history as inputs, the evolution of the microstructure across the coating/substrate system, to estimate the losses due to oxidation and most importantly the likely remaining life of the component based on a specified failure criterion.

The models developed are also being used in reverse to develop new, improved coatings and to better tailor the coating chosen to the superalloy substrate in order to optimise the useful service life in the increasingly harsh environments where they need to operate. This could reduce development costs as only the most promising solutions require experimental testing, and reduce the time to market of new materials systems for industrial power generation.

Further information: Professor Rachel Thomson