LABORELEC investigated the influence of intermittent operation on the life of components, within the framework of a multidisciplinary project (the GST - Gas Steam Turbine project) dealing with electric combined-cycle power plants. From this point of view, special attention was devoted to the high-pressure steam duct between the recovery boiler and the steam turbine.
The Drogenbos power plant (Belgium) provided flow rate and pressure data, as well as the opening/closing signals of the by-pass and valves. Laborelec fitted the duct between the boiler and the mixing sphere with thermocouples and strain gauges. Data analysis enabled the stress cycles to which the duct is submitted to be determined and the flow of water from the boiler to be pinpointed that causes critical thermal shocks some hours after stops.
Damage caused by water flowing from the boiler in this way severely limits the life of the high-pressure duct and obliged the power plant to add blow-offs and to adapt the stopping procedure. Laborelec is now investigating the impact of these events on the collectors of the recovery boiler.
GST units play an increasingly significant part in Belgium's electricity production. These units are very complex and possess their own specificity and advantages : high efficiency, can cope with frequent rapid start-up, can cope with changing loads, several shafts or different machines (gas turbines and steam turbines) on the same shafting, the environment is protected and can use different types of fuel.
In the multidisciplinary "Gas steam turbine (GST)" project (see Appendix 1), the role of is to follow and help operators in the fields of materials, diagnostics, optimisation of the lifetime, maintenance, availability and flexibility of operation, as well as with its contacts with the manufacturers.
In the context of this project, the objectives of work package 5 (thermal transients in the steam duct) are to study the mechanical and thermal behaviour of the steam duct between the boiler and the turbine (high pressure) and to advice regarding the operation in order to optimise the lifetime of the unit while preserving operational flexibility and the role of the unit in the network.
To meet these objectives, the study comprised the following tasks :
The study was carried out on the Drogenbos power station (460 MW). Operational since 1995, it has two gas turbines of 150 MW and a steam turbine of 175 MW.
- Measurements of temperatures and stresses on the outsides of different sections of the steam duct from June 1997 to December 1998.
- Analysis of the records of start-up and shutdown. Evaluation of the thermal stresses resulting from the variations in temperature over time ("temperature gradient") and space ("temperature stratification" on the periphery of the sections).
- Make an inventory of the critical cycles (cases of stratification on shutdown) and of the fatigue engendered.
- Investigation, with the operators, of the causes of stratification on shutdown.
2 Preliminary evaluation of the lifetime under Creep
The characteristics of high pressure superheated steam duct at Drogenbos are as follows:
|10 Cr Mo 910
|Table 1: characteristics of the high pressure superheated steam duct
The maximum operational stress is the circumference stress due to the pressure.
The lifetime under operating conditions of the steel 10CrMo910, for s = 37.1 N/mm² and T = 525°C is greater than 106h (2), even when taking account of additional stresses of 20% in the bends. It can be concluded that the lifetime under pressure stress does not pose any problems.
This value is low by comparison with the stresses of the thermal gradient, whose order of magnitude is more than twenty times greater. It has therefore been possible to ignore the stresses due to pressure in the course of the study.
3 Measuring equipment
3.1 Initial installation (May 1997)
Four sections were selected and equipped with sheathed thermocouples. They have been fixed by means of a small plate electrically spot welded to the walls. There were also 5 high strain gages that were also held in place by spot welding and placed axially in the lower area of the steam duct. The position and the direction of the strain gages were chosen in such a way that they could measure both the stresses caused by thermal shocks or thermal gradients and by the bending due to normal expansion or stratification in the lower parts of horizontal sections of the steam duct.
The siting of the sections and the measuring points is illustrated in figures 1 and 2 :
- section S1 : just downstream of the isolation valve of the boiler and its by-pass,
- section S2 and section S3 : horizontal elements in the vicinity of bends. These sections were, a priori, susceptible to thermal stratification.
A blow-off had been situated between these two sections since 1993.
- section S4 : just upstream the mixing sphere before the steam turbine.
Fig 1: Localisation of the measurement sections|
Fig 2: sensors location|
3.2 Additional thermocouples (from the 5/12/97)
Following from observations made of the first times water was found to be flowing, 4 additional thermocouples were fitted in November 97 : 2 on the S-2 section (immediately at the outlet of the collector, see figure 1) and 2 on the S-1 section (between the bleed point and the anti-leak check valve, see figure 1).
3.3 Other measurements and recordings
The following signals, characterising the thermodynamic condition of boiler 2, were also recorded : the steam pressure (downstream of the isolation valve), the rate of steam flow (downstream of the isolation valve), the limit signals, open or closed position, of the valve, the limit signals, open or closed position, of the by-pass of the isolation valve.
All the sensors are connected to a recording device that simultaneously updates all readings taken, every 2 seconds.
4 Analysis of the variations in temperature
4.1.1 Classification of starting sequences
In what follows, the following types of starts can be distinguished:
- Cold start-up (after a stop of more than one week) : the temperatures of all pipes are initially £ 100°C (except in section S4, which remains hot if the other boiler - boiler 1 - is in service).
- Hot start-up (after a stop of several hours) : all the initial temperatures are greater than the maximum saturation temperature in the first phase of start-up (T~ 250°C for p = 40 bar).
- Tepid start-up (after a stop of several days) : the initial temperatures of certain sections lie between 100 and 250°C.
4.1.2 Section before the boiler isolation valve (IV)
Once the boiler starts to get hot, the temperatures of the upstream sections (S-2 and S-1) increase progressively. Maximum gradients are observed :
- from 20 to 40 K/min in cold start-up,
- from 5 to 25 K/min in the other cases.
4.1.3 Duct downstream of the boiler isolation valve (IV)
Once the by-pass of the IV has opened, the pressure becomes established more or less rapidly in the downstream duct. Sections S1, S2, S3 (S4) may heat up in different ways according to the type of start-up :
- In cold start-up. In all sections of the duct, the initial temperature of the metal "To" is lower than the temperature of saturation "Tsat" determined by the pressure: the duct heats up rapidly and in a uniform way by condensation of steam.
- gradients of 20 to 40 K/min are observed.
- In hot start-up. In all sections, To is greater than Tsat: there is no condensation, nor thermal shock. The duct heats up by convection on contact of the steam, with moderate gradients:
- gradients on the downstream duct : 7 K/min,
- gradients on the upstream duct : 25 K/min.
- In tepid start-up. The sections that have cooled down thoroughly (normally or as the result of stratification) for which To < Tsat warm up by condensation, with thermal shock:
- amplitude of the shocks DT = 100 to 200 K.
It can be seen that tepid start-up conditions are often more severe than cold start-up conditions; this can be explained by the residual pressure in the boiler and the almost instantaneous increase in pressure in the duct once the IV by-pass has opened. It would appear desirable, to limit the intensity of the thermal shocks, open this by-pass sufficiently early before the IV itself, in order to slow down the pressurisation process.
In addition it was observed that :
- the sections further from the boiler, for which To>
Tsat, are sometimes subjected, before getting hot, to momentary cooling of the lower part of the pipework to the saturation temperature Tsat : as the formation of water in these sections cannot take place (since To>Tsat), this stratification can only be explained by the flow of water formed in the cold sections before them towards the hot sections after them;
- in the hottest sections (away from the boiler), the water coming from upstream does not arrive in the liquid phase and there is no stratification. This is the case particularly for section S4, near to the mixing sphere before the steam turbine when boiler 1 is in service.
4.2.1 Section upstream of the isolation valve
At the beginning of the stop sequence, this section cools quite rapidly but in a homogenous and limited manner :
It is related to the cooling of the boiler and slows down rapidly when the flow is cut off. Several hours after the closure of the valve, a rapid and large drop in temperature is observed (12 stops out of 31) in the lower part of the instrumented sections, to the current saturation temperature in the boiler. This drop occurs first in S-2 and then, several minutes later, in S-1. The cause of this behaviour can only be ascribed to the passage of cold water coming from the boiler.
- gradient of 15 to 20 K/min.
difference in temperature between the high and low parts of the sections concerned can be as high as 100 K in S-2 and, in some cases, lower in S-1; probably because of the bleeding point situated between sections S-2 and S-1 (see figure 1).
4.2.2 Duct downstream of the isolation valve
At the beginning of the stop sequence, moderate cooling is observed:
The rate of cooling slows very sharply after the flow has been cut off.
- gradient of -10 K/min,
- occasionally a stratification shock in S1 of DTmax = - 40 K.
At the moment when the valve is closed, if the boiler pressure is high (greater than 55b), a rapid, momentary drop in the bottom temperature is observed in section S1, and therefore stratification appears:
As the bottom temperature remains higher than 400°C, this occurrence can only be explained by a quantity of water coming back from the boiler in the course of the valve closing. After this shock has been absorbed, cooling continues slowly.
- the DT (thermal shock, stratification) can reach 80 K.
When major stratification arose during some stopping sequences, the following conditions obtained:
||From 6 to 19 h after the stop
||from 1/2 to 9 h after S1
||once: 2h30 after S2
||85 to 250 K
||90 to 250 K
|Table 2: conditions associated with major stratification
These events, and particularly the thermal shocks related to them, contribute greatly to fatigue.
5 Evaluation of the lifetime under fatigue
The abrupt variations of temperature observed can usefully be compared to perfect thermal shocks (variations of internal temperature DT in an infinitely small period of time). The calculation of internal thermal stresses is in that case determined simply by the amplitude measured DT and by the characteristics of the material (modulus of elasticity E, coefficient of thermal expansion a, coefficient of lateral contraction m).
Measurements of external stresses remain useful for the direct evaluation of stresses of a static nature or for the calculation of internal thermal stresses due to moderate gradients (for example £
30 K/min) that last a long time.
The internal stress diagram for the most critical shutdown-start-up cycle is calculated as follows :
- diagram of the relative deformations e
measured and corrected,
- the stresses that are static in nature due to the expansion and the stratified state, are obtained by the simple formula :
- Superimposing the virtual gradient stresses corresponding to thermal shocks of amplitude DT onto the static stresses :
- thermal shock of heating:DT = + 160K
- thermal shock of cooling:DT = - 240K
- the formula is the following:
By introducing the coefficient of concentration 2 required for the thermal gradient stresses (that constitute the essentials of the cycle) in the discontinuous areas (orifices, etc.), one obtains:
Ds = 2×(-605)×957 N/mm2 = 3124 N/mm2
According to the fatigue curves (1), for amplitude of 3124 N/mm², at the average temperature T = 200°C, the number of cycles causing the initiation of fissures is equal to 210.
Although starting points are inevitable, it is essential to avoid as much as possible the points observed on stopping.
- concerning stresses of a static nature, strain gages readings show that:
- the stress during flow is low (approx. 9 N/mm²)
- the bending stress due to the stratified condition on stopping is moderate (122 N/mm²); it does not pose any creep problems, since the temperature is low.
6 Investigation of the causes of major thermal shocks on stopping
Figure 3 below gives the diagram needed for understanding the observations. The sections where measurements were made upstream (S-2 and S-1) and downstream (S1) of the isolation valve (IV) are shown on it.
Fig 3: round Isolation Valve|
Of the 31 shutdowns analysed, 12 have given rise to major thermal shocks by stratification downstream of the IV.
6.3 Analysis of a typical situation
The shutdown of the 07/02/98 with thermal shocks upstream and downstream of the IV, was composed of the following phases:
Interpretation. Some water that has condensed in the boiler, at the saturation temperature, flows into the bottom section of the duct. At first, the water is vaporised on contact with the hot piping and cools it until it reaches the saturation temperature. The stream of water then reaches the following section and, section by section, the whole of the bottom part of the duct upstream of the IV cools to the level of the saturation temperature in the boiler. It takes 11 minutes for the sections S-2 and S-1 about 1m away to cool.
- shutdown according to the normal procedure :
- 3h15 : started reducing the pressure, the rate of flow, and the temperature.
- 5h00 : closed the IV (p boiler: 38 bar; all the T° pipework at approximately 350°C).
- thermal shocks upstream of the IV :
- 6h04 : thermal cooling shock in S 2 (DT=-60K).
- 6h15 : thermal cooling shock in S 1 (DT=-60K).
- after these shocks, the bottom temperature of sections S-2 and S-1 stabilised at T=240°C, which corresponded to the saturation temperature at the pressure in the boiler (» 33 bar).
Interpretation. The propagation of the water coming from the boiler spreads across the IV, which is not watertight, and reaches the bottom of the duct downstream, which is under vacuum. The water starts by vaporising and cools the duct until it gets to the saturation temperature prevailing downstream of the IV (T£ 100°C). The stream then reaches the following section, and section by section, the whole of the bottom part of the duct cools to »
100°C. In this case there is an interval of 45 minutes until the parts of sections S1 and S2 at a distance of »
30 m cool. The thermal shocks downstream of the IV are greater than the upstream shocks mainly because of the lower saturation temperature downstream (under vacuum).
- thermal shocks downstream of the IV :
- until 9h00 : progressive cooling in the downstream sections takes place as the boiler pressure and saturation temperature fall.
- 9h05 : thermal cooling shock along the bottom of section S1 (DT=-200K).
- 9h15 : the bottom temperature of section S1 stabilises at 100°C.
- 10h00 : thermal shock along the bottom of section S2 (DT=-150K).
6.4 Effect of the blow-off upstream isolation valve
In all the incidents of major thermal shocks downstream of the IV, we observed that the preceding thermal shocks before the IV are identical in sections S-2 and S-1, that is to say before and after the blow-off. Conversely, in the majority of cases without major thermal shock downstream of the IV, the preceding thermal shock in S-1 was markedly lower than that in S-2.
We can therefore assume that when the bleeding point is open and can remove the greater part of the water coming from the boiler, the risk of thermal shock throughout the duct downstream of this bleeding point is low.
The preceding analysis has revealed two important facts :
- the water causing the thermal shocks comes from the boiler and penetrates downstream of the Isolation valve (IV) because it is not watertight;
- the thermal cooling shock in the bottom part of the duct downstream of the IV is more serious than that upstream, because the saturation temperature is lower downstream (pressure of the condenser). Therefore keeping it at boiler pressure could reduce the thermal shock in the duct downstream of the IV.
The operators are examining the possibilities of eliminating the thermal shocks; two methods can be envisaged :
- progressive controlled cooling of the boiler and the steam duct;
- the use of the blow-off to eliminate the propagation of water beyond that point.
- TRD - "Technische Regelm für Dampfkessel" [Technical regulations for steam boilers], 1981 edition, Carl Heymans Verlag - Beuth (Berlin)
- "Courbes de fluage isothermes (DIN 17175)" [Isothermal flow curves (DIN 17175)] (In French)
- "Meting thermische stratificatie in de stoomleidingen - TGV Drogenbos - Beschrijving van de meetketens" [Measurement of thermal stratification in pipes carrying steam - GST Drogenbos - Description of the measuring sequences], TRACTEBEL (M. De Smet), no. 4408 (23/09/93). (In Dutch)
- Marc WOLFS "Centrale TGV de Drogenbos - Mesures des transitoires de température sur la tuyauterie de vapeur surchauffée haute pression" [Drogenbos GST power station - Measurements of temperature transients on the pipework of high pressure superheated steam], LABORELEC-MATER-STR-97-0012/R-MW-cg; 24/09/1997. (In French)
- LABORELEC - "Production Classique - Séminaire-Exposé de quelques faits marquants en projets et service de l'année 1997" 10/02/1998 ("Contraintes thermiques transitoires dans une tuyauterie vapeur de centrale TGV" - Marc WOLFS)" [Traditional Production - Seminar-Exposé of some salient facts concerning projects and service in the year 1997" 10/02/1998 ("Transient thermal stresses in the steam pipework of a GST power station" - Marc WOLFS)] (In French)
- Olivier VAN ROYE "Etude de la fatigue thermique des composants à paroi épaisse dans l'industrie : élaboration d'un logiciel convivial et application en centrale thermique" [Study of the thermal fatigue of thick walled components in industry: development of a user-friendly software package and application in a thermal power station]. Paper at the end of a study, I.S.I.B. (June 1998). (In French)
- PICH R.
"Die Berechnung der elastischen, instationären Wärmespannungen in Platten, Hohlzylindern und Hohlkugeln mit quasistationären Temperaturfeldern" [Calculation of the elastic, non-stationary temperature stresses in sheets, hollow cylinders and hollow spheres with quasi-stationary temperature zones], Mitteilungen VGB, Hefte 87-88, Dez. 1963, Febr. 1964. (In German)
- H. DIETMANN, St. ISSLER
"Zur Bestimmung von Dehnungserhöhungsfaktoren" [Determination of factors increasing expansion], VGB Kraftwerkstechink 9/98 (see p.122) (In German)
Appendix 1: The Work Package of the GST multidisciplinary project
WP1: Aspiration of air and dirt from the compressor.
WP2: The combustion system.
WP3: Materials for high temperature applications.
WP4: The mechanical behaviour of combustion turbines.
WP5: Thermal transients in the steam circuit.
WP6: The gas circuit of the recovery boiler and polluting emissions.
WP7: The air-cooled condenser.
WP8: Anticipation of the problems posed by intermittent operation.
WP9: Superficial stress of the water in vessels of GST units.
WP10: Monitoring the operating conditions of the power stations.
WP11 : Study of lubricants for gas turbines and gas-fuelled engines.
WP12: Regulation frequency - power of a GST.
WP13: Criteria for de-icing the air at the intake of a gas turbine.
WP14: Measurement of the temperature and the speed of gas leaving a gas turbine.