Technology for a longer life

There are several case histories on high temperature components such as boiler headers and turbines, which are subject to creep – due to on-load temperatures and pressures, and to fatigue – due to start up and shut down thermal transient operation.

Powergen has developed several technologies for the life extension of power plants. These include:

• LEO (Life extension and optimisation) software. The key purpose of this is to calculate daily damage rates in boiler, turbine and feedwater components by the analysis of plant operations (temperature and pressure information). It uses advanced algorithms and finite element models to assess both transient and steady-state operations. It allows variations in real operations to be quantified, and optimised operations to be attained for safe life extension.

• ARCMAC plant creep gauges and software. These ascertain the remaining life in high temperature components, such as main steam pipework, more accurately.

Transformer condition monitoring. Ageing transformers are a high risk area on power plants. The system monitors the key areas of tap changer condition, using advanced algorithms, winding temperatures, partial discharge and dissolved oil-in-gas.

Creep and fatigue damage

If a component is operated at 14°C above its intended temperature, its life will be halved, so it is important that operators are aware of the impact that temperature can have on creep life consumption rates.

Practical power plant operations result in variations in transients and damage due to a variety of factors including:

• Different periods off-load.

• Operational differences between different operators.

• Different shutdown procedures.

• Different commercial pressures.

• Lack of feedback on best practice from a plant damage perspective.

Life monitoring and simulator analysis allows best practice to be identified and consistently attained.

The requirement to design against creep places components at a significantly increased risk of failure from thermal fatigue and creep fatigue where creep and fatigue mechanisms interact. This increasing emphasis on flexible rather than base-load operation means that asset managers need to consider the number of cycles and load changes.

Thermal fatigue is associated with temperature transients that occur during start-up, load changes and shutdown of plant.

How can we approach the problem of creep-fatigue in plant and judge the risks to components? The answer requires some idea of the factors that influence the driving forces and resistance to creep-fatigue damage, which inevitably means: a knowledge of geometry and stress-raising features; an understanding of thermal transients acting on the component of interest, and an understanding of material properties.

Damage monitoring system

Many strategic plant components may require replacement, possibly more than once, throughout the life of the plant.

Extensive studies of flexible operation have shown ways of identifying costs and developing improved procedures to optimise operation of the generating assets. Following such studies, effort has been spent developing appropriate systems for quantifying creep and creep-fatigue damage to plant.

Power Technology developed a suite of computer software applications, one of which, LEO, performs this role.

The LEO system quantifies creep and creep-fatigue damage and provides a facility for plant operators to optimise their operating procedures, thereby obtaining the best balance between taking advantage of commercial opportunities to increase MWe income and minimisation of plant damage.

The system calculates both creep-fatigue and creep damage separately. Thermal stress histories are evaluated using a Greens function approach, which calculates thermal stress histories using steam or pseudo- steam temperatures, either from thermocouples mounted to thin-walled sections such as stub tubes or near-surface deep-drilled positions.

Creep damage is calculated using standard polynomial rupture laws with specific reference stresses that are calculated for the component, preferably from limit load analyses or otherwise from inverse code methods, where applicable. This places operators in a more informed position if they encounter pressure from OEMs to replace plant, which may have significant residual life. Other useful features of LEO include the ability to analyse user-defined cycles. This provides a useful and systematic means of iterating towards optimal start-up procedures and gauging plant damage associated with different types of cycle. LEO is also modular, enabling components to be added if and when necessary. The three figures below show typical outputs from LEO that can be produced to help operators understand damage.

The elastic stress history for a component gives a daily snapshot of stress peaks and troughs, and a sensible guideline could be to try to limit the stress range to within plus/minus the yield strength. The damage accrued on a daily basis would then be represented on a fatigue damage trend plot shown in Figure 1. Here, damage is displayed against a normalised against fatigue index, which can be set by the user. In the figure, damage is normalised against 1000 cycles, which means that a cycle that is calculated to give 1000 cycles to crack initiation would correspond to an index of 1.

Figure 2 illustrates the possible trending that can be used to determine whether daily operating temperatures are within an acceptable limits and identifies temperatures that are required to meet user-defined criteria in terms of operating hours.

Studies of cyclic damage rate variations have demonstrated significant ranges in the levels of damage from cycle to cycle. Figures 3 and 4 illustrate a bad cycle and a good cycle taken from a superheater outlet header. The bad cycle was calculated to be 30 times worse than the good cycle. This factor has been calculated by examining extremes of operation, but an order of magnitude improvement in damage rates can be realistically attained for some components, particularly those whose lives tend to be dominated by downshock cycles. High and low temperature headers tend to fall into this category.

The employment of damage

monitoring systems to investigate damage need not only be restricted to fatigue. Often there are commercial gains possible where enhanced boiler firing may be necessary to deliver increased MW output or conservative constraints may be in place that can be re-assessed using life monitoring. Here, a system like LEO can quantify the threat to plant in terms of accelerated creep life consumption and enable operators to be provided with feedback of modified operations. In many cases, it can be shown that faster start-up rates can be implemented which cause less damage than previous un-monitored operations. Figure 2 provides an illustration of creep life trending where the effects of enhanced boiler firing can be assessed.

The use of damage monitoring has been outlined and demonstrated to give substantial benefits to operators in terms of damage reduction and qualitative life assessment. This enables asset managers to investigate the effects of changing plant operations on associated damage and optimise use of plant.