0ne third of the UK’s electricity is generated by coal-fired power stations. These plants currently work with an efficiency of around 40% and an operating temperature of 565°C. Increasing this efficiency by just 10% would pay dividends, both in environmental and economic terms. But the barrier to achieving this improvement is wholly materials-related.
With today’s advanced steels having only been in service for up to 20 years, extrapolating data for a 50-year service life and designing better steels is problematic. Now a nanoscale steel modelling project headed by Professor Roy Faulkner at Loughborough University could provide a great leap forward for research in this area.
Modelling Power Station Steels
The recently validated work on modelling power station steels involves determining small changes at the nanoparticle level using mathematical models to establish the strength of the alloy. The models are then extended to predict the long-term behaviour of steel, as well as to verify other factors such as the effects of heat treatment, or the result of increasing the amount of an element such as chromium in an alloy.
Coal-Fired Power Station Operating Conditions
The main component of a coal-fired power station is miles of steel alloy piping. To increase the efficiency, the steam temperature in the turbines needs to be raised as much as possible - typically up to 650°C for an ultra-supercritical power plant that achieves an efficiency of 50%. Unfortunately, the knock-on effect of this raise in operating temperature is increased pressure and stress placed on the power station structure and currently there are no steels that can run at 650°C for 30-50 years at such high pressures. The Faulkner group model could change this by helping design better steels with improved performance over time.
What are Better Steels?
Better steels have greater creep and yield strengths and consequently a longer life. High corrosion resistance is also a priority. The steels of choice are ferritic ones with low thermal expansion coefficients, often known as ferritic martensitic steels. From the point of view of performance at high temperatures and pressures, austenitic steels (used in chunky castings in power station headers) are superior, but unfortunately their expansion coefficients are twice that of ferritic steels. All boilers work under cyclic situations so large that austenitic steels would ratchet the system to failure with thermal fatigue.
Developing the Model
To model the characteristics of the ferritic martensitic steels used in currently operating power station piping, a random number Monte‑Carlo‑type model was developed by the researchers. A block was taken,, to simulate a complete grain in which to follow the transformation process. The heat treatment and service (ageing) time was then divided up into small time intervals (Δt), and the material was allowed to nucleate particles of the second phase on a random basis, as in real life, allowing a random array of particles. Precipitates form through quenching and tempering, and the growth of each individual particle in its own localised chemical field over Δt was monitored.
Predicting Precipitate Formation
In an idealised microstructure of ferritic martensitic steel, there are different driving forces for nucleation at different locations, which is taken into account by the model. Quenching forms lath packets - martensite plates with different orientations. Tampering at 750°C causes precipitate phases that strengthen the material. Precipitates form within grains as well as on grain, lath and lath packet boundaries, but the majority of the carbide precipitate forms on a grain boundary of some sort. ‘We’re not trying to predict just one set of precipitate phases, we have to predict the behaviour of a number, which is typical of modern alloys,’ says Faulkner. Other precipitates include needles of NX nitride phase and an iron/molybdenum rich laves phase.
Reasons for the Success of the Model
The success of the Faulkner group model over models developed by other groups in predicting microstructural behaviour is mostly due to the detailed inclusion of the important effects of grain boundaries on precipitation, as Faulkner explains. ‘Boundary-based precipitation is so prevalent in this sort of structure. At the boundaries you get completely different fluxes of solute atoms growing the precipitates compared to growth in the centre of the grain where diffusion coefficients are different and segregation occurs. Impurity elements segregate to the boundaries during heat treatment and can affect precipitation rates there. We have built all these effects into the model, giving much more realistic data. The approach seems to be working very well and we have a good correlation between what the model predicts and experimental data for long-term exposure studies. The model was validated against experimental data using heat treated alloys in studies using the transmission electron microscopes.
Modelling, A Time Saving process
‘Modelling offers the solution to a problem that can’t be solved by other means in such a short space of time,’ says Faulkner. This means that designers of alloys, such as Corus, (who has collaborated on the project, along with Powergen, Innogy and Mitsui Babcock) could take a short cut in forecasting how a material will develop, without having to do the lengthy and costly experiments to measure the microstructural evolution.
Developing Better Steels
‘We have developed the models to such a point that we can feed in different compositions to improve strength, and also make suggestions about how heat treatment might be altered. All of this will contribute to making a better steel,’ says Faulkner. So how will the models be used?
The Effects of Different Preciptates
Damage parameters can be fed from the precipitation models continuous damage mechanics (CDM) models to forecast creep rate strain as a function of time to show the effects of different precipitates. For example, it has been shown that using vanadium nitride for strengthening is more effective than M23C6 carbide. The curves can be used to separate out the effects of different precipitates. Boron is one element that has been tested. It affects the creep strength of steel by altering the rate at which precipitates coarsen and grow, and it restricts diffusion to some extent.
Corrosion is another important issue. ‘For 30 years service at 650°C, any ferritic 9% chromium steels will oxidise away completely, so alloy designers have now developed parallel ferritic steel alloys with approximately 12% chromium, to give extra corrosion resistance,’ says Faulkner.
The Prediction of Precipitate Behaviour Over Time
Another use is in the prediction of the volume fraction (which gives the volume of the material occupied by precipitate), particle size, size distributions and interparticle spacing. The latter is probably the most important parameter as it relates to the Orowan criterion - the strengthening effect of the precipitate. Closer spacing of particles gives an increase in yield strength of the steel the desired characteristic. As particles become more widely spaced over time, the steel loses strength, which is what materials engineers are constantly working against. ‘We could also feed in data for the precipitate distributions, sizes and spacings into the CDM model and work out the stress rupture life, the time taken for the material to fail at a given stress due to creep at any given temperature. This is the next area we will be looking at,’ says Faulkner.
Japanese Transmission Electron Microscopy Studies
Several of the Loughborough group are involved in ongoing studies into the effects of grain boundaries and the validity of the model using the JEM-ARM1300 multi beam million volt transmission electron microscope (TEM) at Hokkaido University, Japan. The focus of the studies has been on large atoms such as hafnium and zirconium. Impurities were ion implanted into a thin film specimen of E911 steel using the 400keV ion beam line. The effects were observed in real-time and the steel was then tempered for one hour at 650°C.
The Effect of Large Atom Incorporation
‘The density of precipitates increased compared with hafnium-free steel in a way we never expected, so we got a finer distribution of particles leading to greater creep strength,’ says Faulkner. ‘Creep experiments have been done showing the beneficial effects of these elements but 1 think we now have the answer to why this is happening and we can now use this information with the models to suggest methods for increasing the precipitation density and reducing inter-particle spacing.’
The length of time before steels developed using the model come into use depends on the steel manufacturers. ‘Corus has taken the data and is interested in developing a new composition,’ says Faulkner. ‘It could possibly be two to three years.’ Other groups have achieved good results in this area (for example the Cambridge University group), but Faulkner believes that their research is the ‘most commercially viable’.
Future Uses of the Model
Future uses of the model include research into the characteristics of weld zones and in superalloys. The group is already using a variation of the model in work on aluminium alloys with Qinetiq for fighter and civil aircraft applications.
How Will the Research Affect Future Power Generation?
As for how this research will affect the future of power generation - greater efficiency could lead to a reduction in consumer's electricity bills. Despite the negative environmental impact of burning coal, the fact remains that it is relatively cheap and plentiful at present. So models such as those developed at Loughborough will be helpful in shaping the face of the UK’s power production for the foreseeable future.