·Home ·Table of Contents ·Industrial Plants and Structures

Evaluating the condition & remaining life of older power plants Eyckmans Marc - Product Manager Laire Charles- Product Manager D'ambros Laurent - Engineer LABORELEC - BELGIUM - Failure analysis & Material assessment in plants Contact

SUMMARY

Run/repair/replace or refurbishment decision making is the day of today in an open market a more strategical approach than the last decade. The best technical solution for higher assets is a must which can only be achieved by applying a mixed methodology of Remaining Life Time (RLT), Risk Based Inspection (RBI) & Reliability Engineering (RE). This paper gives an overview of the RLT & RBI methodology applied by Laborelec for the Belgian Power generator Electrabel.

As further explained in the paper the Laborelec Remaining Life Time (RLT) methodology; mainly based on corrosion, creep and thermal fatigue principals, is structured in a multi-level approach where the final scientific approach is a mix of design, operation & maintenance history in combination with quantified material characteristics.

The scientific approach results in a theoretical creep & thermal fatigue remaining life estimation extended with recommendations of future predictive inspection intervals or run/repair/replace decisions. These recommendations are determinated by the complementary Risk Based Inspection (RBI) & Reliability Engineering (RE) home developed methodologies. Some practical examples will show the benefit of this "three dimensional" approach.

1. BACKGROUND

The current situation regarding the assessment, testing and inspection of power plants frequently results in the formulation of the following question : How long can power plants be operated safely and cost-effectively while satisfying increased requirements pertaining to operational availability and reduced pollutant emissions after having served their originally intended service life ?
This question is of growing importance when the fact is considered that the percentage of plants currently in operation that are more than 30 years old is rising. In order to answer this question the operational capability of the existing plant must first be investigated.

The availability of a power plant depends on the availability of its non-redundant component. Assurance of a proper operation of these components, so called Key Components, should be, therefore, the main task of a plant Remaining Life Time (RLT) program. The condition of these components can be assessed only by way of a RLT methodology. Based on the RLT results a proper decision can be made as to plant safety and availability :

• maintain in operation as it is
• repair
• replace

Taking into account the economic implications of these three alternatives as well as the economic, social and environmental implications of unscheduled outages due to sudden failures the selection of a proper RLT methodology gain in importance. This methodology should provide the best technical solution to ensure a safe operation of the plant.

2. REMAINING LIFE METHODOLOGY

The art in determining a plant's actual status and how much of its entire service life has expired lies in combining analytic considerations, calculation methods, the relevant non-destructive tests, including strain measurements and the correct selection of those components, in which the respective damage is to be reckoned with.

Unexpected damage may be caused by exceeding thermal, static, and/or dynamic parameters which are used as the basis for calculations.

An important basis for the results is the analysis of the actual operational data (e.g. temperature, pressure, throughput), operational experience and operational tenders.
A systematic integral approach which allows assessment of the plants current operational capability and safety is only possible by correctly drawing a correlation between the operational load and the actual status of the plant or its components, which has been obtained from tests and inspections.

Based on these results, the right measures for future procedures can then be initiated in a reasonable manner.

3. THE LABORELEC RLT METHODOLOGY

The Laborelec RLT methodology is structured in a "Three Level Approach" with following sections :

• Level 1 : Design Data Collection & operation history
• Level 2 : Operating / Maintenance / Inpsection history
• Level 3 : Scientific Approach based on level 1 & 2 data in combination with quantified material properties

Fig :

In accordance with the "Three Level Approach" concept, each section contains a specific program for the condition assessment of the equipment.
The Laborelec 3 level RLT methodology, schematicly presented in underlaying figure, can be applied on the following power plant components :

• pipe lines
• boiler tubes
• boiler large components

4. LEVEL 1 : DESIGN DATA COLLECTION

The question posed during an initial approach is what kind of damage? malfunctions? material failures ? can occur.
A first quick scan will separate the possible & non-critical items in the proces and this based on temperature & wall thickness design data. The establishment of the list of possible critical parts is based on our 30 years field experience for similar equipment. For the final list we can go on with level 2 & 3.
Following table gives a quick overview of possible failure mechanisms for boiler components.

	C	R	TF	F	E	Er	Co
- evaporator			X		H²	X	X
- drum			X				X
- superheater / reheater tubes	X						X
- superheater headers	X		X
- reheater headers,	X		X
- desuperheater nozzles	X		X
- steam lines	X		X
- feedwater lines						X	X
Table 1:

C = creep     R = relaxation     TF = thermal fatigue     F = fatigue
E = embrittlement     Er = Erosion     Co = Corrosion

5. LEVEL 2 : OPERATING / MAINTENANCE & INSPECTION HISTORY

Level 2 gives a specific plant history input so we complete design data with operating & maintenance events which are :
• For the operating part :
• The operating parameters like pressure & temperature
• Incidents, events of failure & repair statistics
• Condition of the plant facility by n° of starts/stops & service hours
• For the maintenance part :
• Review of component replacement & repair
• Review of component geometry
• For the inspection part :
• Non destructive testing results where the most important are :
• Wall & internal oxide thickness measurement by ultrasonics
• Metallographic examionation by replication
• Stress measurements by strain gages
• Destructive material testing like failure analysis, isostress creep testing

The results of the NDE & DE provide one essential input for the component integrity evaluation and life assessment.

5.1. NON DESTRUCTIVE TESTING
Non destructive examinations (NDE) are essential constituents of any residual life assessment programme. The objective of such assessment is to compare the current condition of the material of a given component with its original condition to define the amount of deterioration of the component.

Three major questions have to be answered before starting NDE :

• When ?
• Where ?
• How ?

The used inspection techniques depend on the given component, the location on the component, the damage modes to be looked at and the material used. Some of the regularly used methods to establish the material condition provide data which can be quantified in analyses whereas others can only indicate whether a defect is present or not. Following table correlates the examination methods with various components of fossil fired power stations.
In underlaying short overview we see the NDE-techniques used on the different boiler components.

Components	DC	TM	ME	HT	PT/MT	ET	UT	RT
- economiser headers	X				X		X
- waterwalls	X	X					X
- boilers drums	X			X	X
- lower waterwalls and headers							X
- junction headers	X				X		X
- waterwall risers		X			X		X
- waterwall headers					X
- superheater headers (welds)	X	X	X		X		X
- reheater headers (welds)	X	X	X		X		X
- desuperheaters :
liners							X
nozzles	X				X
- HT superheater tubing	X	X	X	X	X		X
- steam piping	X	X	X	X	X		X	X
- feedwater piping		X						X
Table 2:

DC Dimension checks      TM Thickness measurement      ME Microscopic examination       
HT Hardness testing       PT Liquid penetrant examination      MT Magnetic particle examination
ET Eddy current examination       UT Ultrasonic examination      RT Radiographic examination

5.2. Destructive testing
The scatter band of material properties (creep strength) is an important source of uncertainty for the calculation of the life expenditure. It may therefore be necessary to determine mechanical property values from specimens of material taken from the actual components. Sampling may however not in any way degrade the integrity of the component. Various sampling methods are used : boot samples, trepanned, core samples, through wall trepanning, ....

For some of them subsequent weld repair may be required. Destructive testing can have following objectives

• verification of non-destructive examinations results (if no defects are found)
• direct (quantitative) assessment of the degree of damage (structural & mechanical)
• determination of component material mechanical properties (to reduce the scatter band) : creep testing, fatigue testing, fracture toughness, crack growth rate evaluation, ...
• post failure search for the damage mechanisms and propagation depth.

5.3. Isostress creep testing
Quantitative residual life assessments are performed using the isostress testing technique. Acceleration is obtained by testing at enhanced temperatures, applying the representative service stress.

The decision to choose for increased temperature testing instead of increased stress, is based on the fact that creep is a thermally activated process. For ferritic materials however the maximum temperature is limited to about 720°C (by approaching the AC1 temperature). Test specimens are loaded under tension to a stress equal to the service stress of the component under investigation. Usually 5 to 6 specimens are exposed to mutually different temperatures, chosen such that creep rupture times are invoked in the range of 100 to 1000 hours. The secondary creeprate is deduced from the recorded strain-time evolution. Time to failure and total elongation after failures are determined also. Finally the test results are extrapolated in the co-ordinate system ln (rupture time) versus temperature (°C).

6. Level 3 : Scientific Approach

Based on the information of level 1 & 2, Laborelec finally calculates the theoretical remaining life of the component. This can be done by 2 different scientific approaches :
• theoretical life time consumption calculations mainly based on the TRD 508 recommendations
• qualitative life time assessment based on metallographic investigations

For each of these apporaches we use several home made soft tools like

• recalculation of stresses from internal pressure under static loading & ultrasonic measured wall thickness by our soft programme called "LILCA"
• calculation of metaltemperature by oxide thickness measurement also done by the soft "LILCA"
• recalculation of the stress range under cyclic loads by a home made soft programme called "THERMSTRESS"
• calculate or estimate creep damage level based on metallographic examination results of critical locations by our soft programme "CREEPMAP"
• recommendation of predictive inspection intervals by the Laborelec soft "INTERVAL"

7. Theoretical life assessment based on TRD 508

The aging of the boiler & piping components is mainly manifested by 2 mechanisms :
• creep
• thermal fatigue

Degradation mechanisms other than creep and fatigue are less accessible to useful life prediction by calculation

7.1. Creep
Static load creep degradation can only appear under certain conditions of stress & temperature which has to be superior of 450°C. For each period of time "ti", under a certain stress & temperature, the relative creep life consumption "e" is calculated as follows :

with t "Ri" = time to rupture for that stress & temperature during period "ti"

The time to rupture is calculated based on the creep rupture data curves for the different boiler materials. These curves (lower, mean & maximum material resistance) are specified are specified in following standards :

• NBN 837
• Din 17735

Static load creep life consumptions are calculated by the home made soft programme called "LILCA"; which also allows you eventually to estimate the mean metaltemperature in service.

7.2. High temperature tube life prediction with "LILCA"
To perform assessments and determine the remaining life of high temperature boiler tubing, it is necessary to accurately measure critical tube dimensions and predict the time dependent response of the tube material.

Laborelec measures wall and internal oxide scale thickness to predict remaining tube life with "LILCA" software. Thickness measurements are made using focused UT transducers. Steamside oxide layer thickness measurements allow the evaluation of the average metal temperature of components by using the material oxidation kinetics.
"LILCA", with the possibility for input of external wall loss due to erosion or high temperature corrosion, contains algorithms that allows the user to make deterministic estimates of tube remaining life.

7.3. Fatigue calculations with "Thermstress"
Fatigue only take place under alternating load conditions. During this load alternation 3 types of alternating stresses can be registrated :

• mechanical stresses due to change in pressure
• thermal stresses due to temperature differences through the component wall thickness, these stresses become only significant for wall thiknesses lager than 45 mm
• stratification due to temperature differences on 2 opposite spots on the component

There exist "Wöhler"-fatigue curves for these 3 types of stress and the different materials. These curves predict the total acceptable number of stress cycles. The relative life consumption "fi" for one cycle with a constant stress amplitude is calculated as follows : "Ni" (with "Ni" = maximum allowed number of that typical cycli)

7.1.E. Linear Damage Summation
Creep-fatigue analysis places an important role in the life assessment activity of power plant equipment. In the U.S.A. an approach to the creep-fatigue design and remaining life calculation was develop in the ASME Boiler and Pressure Vessel Code, Section III, Code Case N-47 which is essentially based on the linear damage summation method given by underlaying equation. This method combines the damage summation for creep and for fatigue as follows :


where
• N/N_f - cycling load portion of the life fraction;
• N - number of cycles at a given strain range;
• N_f - fatigue life (number of cycles to failure without creep damage) at that strain range
• t/t_f - time-dependent creep life fraction
• t - time at a given stress;
• t_r - time to rupture at that stress;
• D^' - the cumulative damage index

The damage summation method is very popular because it is easy to use and requires only standard S-N (Wöhler) and creep stress rupture curves.

8. Remaining Life Assessment and Inspection Interval Recommendations Based on Material Microstructure Investigations

8.1. Background
One of the main damage mechanisms affecting power plant components is the creep damage. such damage may occur in different forms : localized or bulk damage.

Localized damage creep may become manifest in the form of cracks. The cracking process is characterized by a time-dependent growth under an approximately constant load. As in the case of fatigue cracking, creep cracking may be characterized by some fracture mechanics parameters.

Bulk damage it can manifest itself in two forms (fig1):

• intergranular creep cavitation and
• microstructural degradation

Fig 1: (a) Ferrite - Pearlitic structure; (b) Carbides Precipitation at the grain boundaries (c) Spheroidization of carbides from pearlite has begun (d) Spheroidization of carbides from pearlite is finished. (e) dispersed carbides (no ferritic - pearlite structure); (f) carbide coalesence;

8.2. Scope of Application


Fig 2: Description of the replication principle

The assessment of microstructural damage under creep conditions is of special interest for the component life assessment. The levels of microstructural degradation or creep cavitation may be evaluated in terms of remaining life or required inspection intervals.

Life assessment techniques based on metallographic methods may be performed destructively by means of sample extraction, preparation and microscopy investigation. Alternatively, metallographic investigations may be done non-destructively by way of replication.

There are two major applications of the replication technique :

• Study of the microstructure (creep cavitation, precipitate spacing, grain size, etc.) using an optical or electron microscope;
• Examination and identification of second-phase particles by extraction technique.
At present, the plastic replication technique is used principally for reproducing surface features such as creep cavities, cracks and gross microstructural features. Field application of carbide extraction replica method will require further development work. Fig. 2 illustrates replication principle. In order to obtain accurate results, a very high surface quality is mandatory.

Surface oxides as well as decarburized zones must be removed prior to replication of the component surface.

8.3. Remaining Life Assessment and Recommended Inspection


Fig 3:

Neubauer and Webel have published the first attempt to correlate the creep life consumption of plant components to cavitation. They collected data on steam pipes from numerous German power plants. According to their theory, the creep damage level can be classified in accordance with the number of cavities and their orientation.
Thus, they separated four classes of degradation (see fig. 3) :
Class A - Isolated cavities;
Class B - Oriented cavities;
Class C - microcracks;
Class D - macrocracks;

They suggested also corresponding actions for each damage stage :

• Class A - No remedial actions required;
• Class B - Replica tests at specified intervals;
• Class C - Limited service until repair;
• Class D - Immediate repair.

Because of the high conservatism included into this theory, it is actually used as a monitoring technique, rather than a life prediciton method. Nevertheless, worldwide acceptance with power plant operators because of its simplicity.

Laborelec also suggested a mixted method of assessment where creep degradation and aging proces are taken into account in combination with there repercussions on the component remaining life. Creep damage process is not the only one affecting power plant components operated at elevated temperatures and high pressures. There are many other damage mechanisms that result in an increase of the damage rate of these components and consequently in a decrease remaining life.

The author proposes a parameter to define a correlation between the following variables :

• Creep damage state evaluated by means of microstructural investigations
• Component expended life;
• Recurrent inspection intervals

It is accomplished based on a logarithmic correlation between the recurrent inspection intervals and component expended life, for each damage level.

9. Risk assessment & Trending analysis

Several methodologies in the last 2 decades have promised to allow obtaining "risk knowledge from data". Virtually all of them were able to prove that they can do it "in principle", but many failed to assure their place in daily practice. The reasons for this are not the methodologies themselves but the high investment price, the very consuming implementation time, the poor measurable benefit and the late break even point of the investment into the "risk knowledge management" (fig 4). Laborelec developed a risk based inspection concept avoiding the above pitfalls and resulting in measurable benefits (fig 5).

Fig 4:

Fig 5:

9.1. The Laborelec RBI methdology


Fig 6:

As for each existing RBI-methodology, Laborelec also started from the basic definition of risk as a produkt of probability & consequence of failure. This probability and consequence for each proces-item are schematicly presented in a 5x5 risk matrix allowing us to identify on a clear manner the high, medium & low risk items. The aim of this Laborelec Risk matrix is a reducing the non-availability of the high risk items and enlargement of the inspection intervals of the low risk items (fig 6).

9.2. Identification of probability and consequence
Laborelec uses a quick scan tool in first instance for the item probability ranking. This quick scan is based on a mix of design & proces parameters and operation & maintenance history. Fine tuning is afterwards possible by the implementation of a integrity factor.
The consequence factor is build of 4 subdiaries. These are the costs of security, environment, maintenance and productionloss due to stop time. Only the highest value is taken into account for the consequence ranking

9.3. Trend analysing & inspection intervall modelling
Trend analysing & inspection intervall modelling is only possible fro trending degradation mechanism like uniform corrosion, erosion, creep & fatigue. For each of these degradation mechanisms, the inspection intervall is defined as the remaining lifetime devide by a security factor. At this time Laborelec proceeds for a statistical approach for the erosion & corrosion degradation mechanism. For creep evaluation we prefer a deterministic approach such as a sudden death risk analysis method.

Concluding comments

None of the worldwide Remaining lifetime methodologies takes into account all uncertainties like scatter in material properties and additional system stresses. This paper demonstrates that the Laborelec 3 dimensional remaining life time methodology; based on creep, thermal fatigue & erosion or corrosion; allows to take into account a maximum of these uncertainties. Some practical examples for use of our home made RLT soft tools are shown.

This RLT-methodology in combination with our RBI-methodology, a risk ranking approach also discussed in this paper, allows the client to identify the optimum inspection interval for all proces items that are sensible for trending degradation mechanisms. The proces equipment affected by non trending degradation mechanisms can only be optimised by changes in design, maintenance or operation conditions.

As a mix of trending & non trending degradation mechanisms will always be present in the entire proces, the author is of the opinion that the most effective manner to manage "trending & non-trending" is to implement a progressive but flexible inspection plan worked out by a team build of material- & NDT-specialists + maintenance & operation people of the plant. This plan must be continuously adjusted according to new information and evolving technological + economical factors.

© AIPnD , created by NDT.net

|Home|    |Top|