EP2105887A1 - Method for diagnosing a gas turbine - Google Patents

Abstract

The method involves using an intake mass flow of a gas turbine (1) as a parameter during the prediction of the additional power. A combustion chamber pressure loss or the pressure loss between an environment and a compressor inlet is determined as the input parameters for determining the intake mass flow of a turbine inlet pressure. Independent claims are included for the following: (1) a gas turbine system with multiple components having a gas turbine and a control system; and (2) a prognosis module for use in a gas turbine system.

Description

The invention relates to a method for the diagnosis of a gas turbine comprising several components, in which the additional power by which the operating power of the gas turbine would increase in the event of a cleaning of one of the components is automatically predicted.

In the case of a gas turbine, its performance and efficiency are reduced during operation by soiling, deposition, erosion and corrosion, which negatively impacts the entire power plant process. Especially the aerodynamic parts of the compressor at the inlet of the gas turbine are affected.

The pollution of the gas turbine is caused by the adhesion of particles to the surfaces. Oil and water mist help dust and aerosols settle on the blades. The most common contaminants and deposits are mixtures of water wetting, water-soluble and water-insoluble materials. In the gas turbine, soiling can occur due to ash deposits and unburned, solid cleaning preparations. Such air pollutants adhere to the components of the flow path of the gas turbine and react with them as scales. It also causes erosion by particle impact and abrasion, commonly referred to as erosion.

Furthermore, pieces of ice that form at the inlet of the gas turbine can dissolve and strike the components in the flow path of the gas turbine. To prevent this, a so-called anti-icing system is used. Here is prevented by air preheating, the temperature of the Air entering the gas turbine does not sink below freezing and thus does not freeze the water.

The described aging processes cause an increased surface roughness of the blades. This leads to comparatively large friction losses in the gas turbine, since laminar boundary layer flows can turn into a turbulent flow and so there is a growing flow resistance. Furthermore, gaps in the gas turbine increase due to abrasion and corrosion. The losses due to the increased gap flow increase and the performance of the system is reduced.

The influence of aging phenomena is at the entrance of the gas turbine, d. H. especially large on the compressor. Geometric changes of the blades due to erosion, deposits and damage cause a reduced efficiency of the gas turbine. Occurring deposits, erosion and corrosion at the inlet lead to changed entrance angles, which have a strong influence on the thermodynamic performance. An aged compressor may cause stalling.

Aging of the compressor negatively affects gas turbine efficiency, gas turbine power and gas turbine exhaust mass flow. To counteract the degradation of the turbine system, regular compressor washes are performed. Compressor blades can be washed in online and offline mode. In the online mode, the turbine system continues to operate during cleaning, the gas turbine load is only slightly lowered. Online washes are mainly used to prevent build up of the dirt layer. Usually online laundry is done once a day with deionized water and every third day with cleansers.

For offline laundry, however, the system is shut down. To avoid thermal stresses, it becomes six Hours cooled with the help of a shaft rotating device. An offline wash is usually done about once a month. If the turbine system has not been cleaned for a comparatively long period of time, it is usually necessary to carry out offline scrubbing for typical systems since the online cleaning method can no longer remove the contaminants.

An offline wash causes a greater performance recovery than an online wash. With the help of an offline laundry, performance gains of several percent can be achieved. Online laundry results in lower performance recovery. The most effective bucket cleaning can be achieved with a combination of online and offline washes. A regular online laundry extends the time intervals between the required offline washes.

The optimal time for off-line laundry is often determined by the operator for purely economic operational considerations, e.g. B. in low load periods. This means that the decision on the timing of elimination of contamination of one of the components of the turbine system, for. B. by a wash of the compressor based only on empirical values under economic aspects or under preliminary studies with fixed boundary conditions.

Alternatively, the determination of the timing of the offline wash may be made based on an up-to-date prediction of the gas turbine power gain expected by the off-line wash. In this case, such a prognosis is usually created based on the development of the compressor efficiency of the gas turbine, which serves as a parameter for the strength of the contamination of the compressor.

However, the measurement data used to determine the compressor efficiency may be provided with relatively high data uncertainty, allowing accurate prediction of the expected performance gain from off-line laundry and thus makes it difficult to determine the cost-optimal time for the operation of the gas turbine for such offline washing.

The invention is therefore based on the object to provide a method of the type mentioned above, which allows a particularly reliable prognosis of the expected performance gain in a cleaning.

This object is achieved according to the invention by using the intake mass flow of the gas turbine as a parameter when forecasting the additional power.

The invention is based on the consideration that the selection of the time of a required to obtain a high operating performance of the gas turbine off-line laundry can be achieved at very low cost by the most accurate forecast of the performance gain by such offline washing the gas turbine. In other words, to determine whether off-line laundry is economically viable at the present time in terms of production downtime due to the gas turbine shutdown, it should be known as accurately as possible at all times how high the expected performance recovery from off-line laundering is.

To increase the accuracy of such a prognosis, the statistical uncertainties should be minimized. This can be done, for example, by improving the measuring equipment or by increasing the number of measurements. However, such an increase only leads to a reduction in the statistical error, but systematic errors in the forecast of the additional service should also be minimized to a large extent. This can be achieved by additionally using additional parameters for forecasting the additional service. One such quantity, which is characteristic of the performance of the gas turbine, is the intake mass flow of the gas turbine.

In a gas turbine, the compressor is on the flow medium side all other components such. B. upstream of the combustion chamber. Accordingly, the compressor is the environmental influences such as incoming dust and dirt particles most exposed component. Advantageously, a cleaning of the compressor is therefore carried out in particular, since this has the highest degree of contamination and thus a corresponding cleaning has a particularly positive influence on the recovery of operating performance of the gas turbine.

The Ansaugmassenstrom as a parameter for the operating performance of the gas turbine is usually not measured directly because of the high cost, the large uncertainty and the risk of damage, but indirectly determined by balancing. Very expensive instruments would have to be used for a direct measurement, because first of all there are very high temperatures, and secondly it is absolutely necessary to avoid the sensors breaking off because of the probably high consequential damage to the turbine blading.

For an energy balance of the entire gas turbine on the one hand and the combustion chamber on the other hand are u as input variables. a. the operating power, the fuel mass flow and the Brennstoffheizwert needed. However, these values are comparatively difficult to determine and involve a very high error. In a combined cycle power plant, in which a gas turbine is operated together with a steam turbine on a shaft, the performance of the gas turbine as a single value only comparatively complex and inaccurate determine, as always only the total power of the entire combined cycle power plant is available. Therefore, to determine the Ansaugmassenstroms advantageously the turbine inlet pressure, the combustion chamber pressure loss and / or the pressure loss between the environment and the compressor inlet are determined as entry characteristics.

In this case, the turbine inlet pressure with the help of the mass pressure equation according to Stodola in a value for the Transfer suction mass flow, while in each case resistance coefficients can be determined from the combustion chamber pressure loss or the pressure loss between environment and compressor inlet, which can be used to determine the Ansaugmassenstroms. Such determination of the Ansaugmassenstroms without resolution of an energy balance is associated with much lower statistical errors and therefore allows an even more accurate forecast of the additional power by which the operating performance of the gas turbine would increase in the case of cleaning one of the components.

In order to further minimize the statistical errors in the determination of the Ansaugmassenstroms, is advantageously determined for determining the Ansaugmassenstroms from a number of input characteristics a preliminary value for the Ansaugmassenstrom, wherein for each provisional value by cross-comparison with the respective other provisional values in each case a validated value is determined. Such a cross-match can be done, for example, based on the VDI2048. This is essentially based on the balancing principle according to Gauss, whose basic idea is not only to use the minimum quantity of measured quantities required for a solution, but also to record all achievable measured variables together with the associated variances and covariances. For the present method, this means that all achievable input parameters are used to each determine a preliminary value for the Ansaugmassenstrom.

Since it is always the same physical intake mass flow, the true values of the input characteristics should be such that all resulting provisional values are the same. On the basis of this assumption, the Gauss method gives consistent values for the actual values of the measured quantities and validated values for the intake mass flow. The validated values for the intake mass flow generated in this way are then averaged and thus form one with a particularly low statistical Error provided characteristic for determining the operating performance of the gas turbine.

To further reduce the statistical and systematic error of the gas turbine, the intake mass flow should not be provided as the sole parameter for determining the operating performance of the gas turbine. In an advantageous embodiment, therefore, the compressor efficiency of the gas turbine is additionally used as a parameter.

When measuring the input parameters, it should be taken into account that in particular thermodynamic parameters of the gas turbine are dependent on the respective ambient conditions such as air pressure and outside temperature. Nevertheless, in order to be able to compare measured values at different times, the respective parameters should be normalized to reference conditions. The norm is the ISO conditions (temperature 15 ° C, pressure 1.013 bar, humidity 60%).

In order to predict the additional power in the case of a cleaning of one of the components of the gas turbine from the calculated instantaneous operating performance of the gas turbine, a reference value for the operating performance of a just cleaned gas turbine is required. In this case, the operating performance of the gas turbine, apart from its contamination state, is also dependent on the pollution-independent erosion, and therefore essentially on the operating age of the gas turbine. In order to obtain such a reference value, parameters of identical and / or construction-type gas turbines are advantageously used as comparison variables in the prognosis of the additional power. As a result, in particular the operating performance after the cleaning of the gas turbine can be forecast particularly well, and it is altogether achievable to have a more accurate forecast of the additional power by cleaning the gas turbine.

The additional performance of a cleaning of one of the components of the gas turbine is often intended not only for an immediate Cleaning, but also for future periods, to allow long-term planning of cleaning. For this purpose, in an advantageous embodiment, a prognosis of the temporal development of the respective parameter is created. Such a prognosis is possible by several evaluations of the input parameters or measured values at different times.

A particularly cost-optimal operation of the gas turbine is possible if a determination of the timing of an offline gas turbine laundry is taken not only under purely economic aspects, such as in low load periods, but based on an accurate forecast of the operating performance of the gas turbine in the future. For this purpose, it is advantageously determined depending on the value of the determined additional power in balance with the overall economic effort, whether the gas turbine is temporarily stopped to eliminate the pollution and optionally determines an optimal time for the temporary shutdown. By accurately predicting the additional performance achieved by offline washing, a determination of a time for such offline washing can be made on the basis of a much more precise analysis in which the costs and benefits of off-line laundry can be accurately weighed against each other.

Advantageously, the method finds application in a gas turbine plant with a gas turbine comprising several components and with a control system which is connected on the data input side to a number of sensors arranged to determine input parameters in the gas turbine, the control system comprising a prognosis module.

In an advantageous embodiment, data of a database with comparison variables of structurally identical and / or building-like gas turbines can be read into the prognosis module. For this purpose, the forecasting module should have a correspondingly open architecture, which makes such reading possible. This can be, for example by means of a mobile data carrier or via a permanent data connection to the database, ie the database can be stored on a writable memory within the control system or stored on an external server which is connected via a data transmission line to the control system of the gas turbine.

This allows a comparison between the data of identical and / or construction-like gas turbines, whereby a recourse to a particularly large base of experience can be made and thus a lesser statistical error is achieved. Conversely, the data obtained in the gas turbine can also be used to expand the database by making it available to the database and stored there.

Advantageously, a prognosis module for use in a gas turbine plant is suitable for carrying out the method.

The advantages achieved by the invention are in particular that a comparatively precise analysis of the degree of contamination of the gas turbine, in particular its compressor is possible by the additional consideration of Ansaugmassenstroms the gas turbine. As a result, a forward-looking, adapted to the operational and economic circumstances planning the offline washing of the gas turbine is possible, whereby a particularly high efficiency of the gas turbine can be achieved during their term. The method described here also allows the determination of the Ansaugmassenstromes without any knowledge of the fuel data and without solving an associated with high uncertainties energy balance. Incidentally, so that the consideration of the Ansaugmassenstroms in relation to the operating performance of the gas turbine for single-shaft systems, in which a gas turbine and a steam turbine are arranged on a common shaft, in the first place possible.

FIG 1

einen Längsschnitt durch eine Gasturbine,

FIG 2

ein Diagramm des zeitlichen Verlaufs der Betriebsleistung einer Gasturbine, und

FIG 3

eine schematische Darstellung eines Verfahrens zur Prognose der erzielten Zusatzleistung im Falle einer Verdichterreinigung.

An embodiment of the invention will be explained in more detail with reference to a drawing. Show:

FIG. 1

a longitudinal section through a gas turbine,

FIG. 2

a diagram of the time course of the operating performance of a gas turbine, and

FIG. 3

a schematic representation of a method for forecasting the achieved additional performance in the case of a compressor cleaning.

Identical parts are provided with the same reference numerals in all figures.

The gas turbine 1 according to FIG. 1 has a compressor 2 for combustion air, a combustion chamber 4 and a turbine 6 for driving the compressor 2 and a generator not shown in detail or a working machine. For this purpose, the turbine 6 and the compressor 2 are arranged on a common, also referred to as a turbine rotor turbine shaft 8, with which the generator or the working machine is connected and which is rotatably mounted about its central axis 9.

The combustor assembly 4 includes a number of individual burners 10 around the turbine shaft 8 for combustion of a liquid or gaseous fuel.

The turbine 6 has a number of rotatable blades 12 connected to the turbine shaft 8. The blades 12 are arranged in a ring on the turbine shaft 8 and thus form a number of blade rows. Furthermore, the turbine 6 comprises a number of fixed vanes 14, which are also fixed in a ring shape with the formation of rows of vanes inside the turbine 6. The blades 12 serve to drive the turbine shaft 8 by momentum transfer from the turbine 6 flowing through Working medium M. The guide vanes 14, however, serve to guide the flow of the working medium M between two each seen in the flow direction of the working medium M consecutive blade rows or blade rings.

The compressor 2 is the component of the gas turbine 1 which is closest to the air inlet 16. Accordingly, it is most exposed to dirt deposits and the resulting contamination of the gas turbine 1. To prevent a reduction in the operating performance of the gas turbine 1, therefore, the compressor 2 must be cleaned regularly. In this case, so-called online washes can be carried out relatively frequently, for example once a day, for which no standstill of the gas turbine 1 is required. At longer intervals, the turbine should be shut down to remove stubborn dirt for offline washing.

The gas turbine 1 comprises a control system 18 which is connected via a data line 20 to various sensors 22 arranged inside the gas turbine 1. In order to determine the optimum offline washing time, the control system 18 comprises a prognosis module 24, which processes the input parameters detected by the sensors 22 and determines the degree of contamination of the gas turbine and the expected gain in operating performance for an offline washing performed on the basis of this data. In order to improve the quality of forecasting, comparative data of structurally identical or construction-type gas turbines can be read into the forecasting module. For this purpose, the control system is connected via a further data line 20 to a database 26 containing such comparison data. The database 26 may be located on an external database server, not shown in greater detail. Alternatively, the comparison data can also be read without permanent data connection to the database 26 via a mobile data carrier.

FIG. 2 1 shows a graph of the operating power of a typical gas turbine 1. The line L1 shows the operating power of the gas turbine 1 at the time of commissioning 30. The line L2 shows the theoretical maximum power of the gas turbine over its running time, its drop only by aging and irreversible Pollution is generated.

The line L3 shows the additional influence of the reversible pollution on the operating performance of the gas turbine. Section I shows the influence of regular online laundry on the operating performance of the gas turbine. This is carried out at regular intervals 32, for example once a day. This results in a comparatively low increase in performance, but cumulatively contributes to the maintenance of the gas turbine 1 in a not inconsiderable manner via the frequent online washes.

At longer intervals, 34 offline washes are performed at timings to be determined. These offline washes result in a much greater return on performance, but require a much greater effort, since the gas turbine 1 must be shut down, with a not inconsiderable cost arises. Therefore, the time points 34 should be chosen in a forward-looking manner, which can be done on the one hand on the basis of economic criteria such as electricity price or fuel price, on the other hand based on the operational variables of the gas turbine. In particular, the anticipated performance gain from off-line laundry should be known for optimal determination of the time 34 of offline laundry.

FIG. 3 schematically shows the flow of the method for determining the additional power by which the operating performance of the gas turbine 1 would increase in the case of cleaning the compressor. For this purpose, the turbine inlet pressure 40a, the combustion-chamber pressure loss 40b, are initially used as input parameters and pressure loss between environment and compressor inlet 40c measured. From the turbine inlet pressure 40a, a preliminary value for the intake mass flow 42a is determined based on the mass pressure equation according to Stodola. Further, the pressure loss in the combustion chamber 40b and the pressure loss between the ambient and the compressor inlet 40c are transferred via a constant drag coefficient approach to provisional values for the intake mass flow 42b and 42c, respectively.

The different approaches initially provide different preliminary values for the intake mass flow 42a, 42b and 42c. With the secondary condition that all intake mass flows should be the same, data validation is then carried out based on VDI2048. This corrects the readings based on the specified uncertainties so that the preliminary values for the suction mass flow are practically the same. On the one hand, validated values for the intake mass flow 44 are produced from the input characteristics corrected in this way; on the other hand, the validated input parameters can be used as a basis for calculating the compressor efficiency 46.

Averaging results then comparatively accurate values for the Ansaugmassenstrom 48 and the compressor efficiency 50 at a certain time 52. These measurements are taken at several times 52 and stored. In each case, the recorded measured values are converted to ISO reference conditions (temperature 15 ° C., pressure 1.013 bar, atmospheric humidity 60%) using a mathematical function, for example a polynomial, in order to be able to correlate the values recorded under different environmental conditions. From the thus obtained normalized values for intake mass flow 54 and compressor efficiency 56, a time profile of intake mass flow 58 and compressor efficiency 60 can now be extrapolated by means of regression analysis. In order to ensure a sufficient quality of the regression, there should be no less than ten times 52 of the measurement.

For both values, mass flow of suction and compressor efficiencies, the difference 62 between the values after the last offline wash and the current time is formed in each case. Then each of the two results is multiplied by a factor. These factors are the result of a fleet analysis, i. H. a comparison with identical and / or construction-like gas turbines 1. The corresponding data can be supplied from an external database 26. Based on the respective statistical uncertainties, probability values are assigned to the result values.

The two results 62 are then converted to a gas turbine power with the aid of gas turbine type-specific indices 64. The thus obtained forecast of the additional power in the case of a cleaning of the compressor is finally supplied to the output 68.

For a more accurate forecast of the additional power in the event of cleaning of the compressor thus the Ansaugmassenstrom the gas turbine is taken into account, to determine the Ansaugmassenstroms no energy balance is resolved and no information about the gas turbine power and the fuel is required, in particular no indication of its calorific value and its mass flow , By thus having a relatively low uncertainty prognosis, the turbine operator can determine the time 34 for an offline laundry based on company-specific data. Thus, overall a cheaper operation of the gas turbine is possible.

Claims (13)

A method for diagnosing a multiple-component gas turbine (1), wherein the additional power by which the operating power of the gas turbine (1) would increase in the case of cleaning one of the components is automatically predicted, wherein in the forecast of the additional power of the intake mass flow ( 48) of the gas turbine (1) is used as a parameter. The method of claim 1, wherein the additional power in the case of a cleaning of the compressor (2) is predicted. The method of claim 1 or 2, wherein for determining the Ansaugmassenstroms (48) of the turbine inlet pressure (40a), the combustion chamber pressure loss (40b) and / or the pressure loss between the environment and compressor inlet (40c) as input characteristics (40a, 40b, 40c) are determined , Method according to one of claims 1 to 3, wherein for determining the Ansaugmassenstroms (48) from a number of input characteristics (40a, 40b, 40c) respectively a preliminary value for the Ansaugmassenstrom (42a, 42b, 42c) is determined, wherein for each provisional value (42a, 42b, 42c) is determined by cross-balancing with the respective other provisional values in each case a validated value (44), and wherein the characteristic mass for the intake mass flow (48) of the gas turbine (1) characteristic as an average of the validated values (44) 44) is generated. Method according to one of claims 1 to 4, wherein in the forecast of the additional power of the compressor efficiency (50) of the gas turbine (1) is used as a parameter. Method according to one of claims 1 to 5, wherein the respective characteristic is normalized to reference conditions. Method according to one of claims 1 to 6, wherein in the forecast of the additional performance characteristics of identical and / or construction-like gas turbines (64) are used as comparative variables. Method according to one of claims 1 to 7, wherein a prediction of the temporal evolution (58,60) of the respective characteristic is created. Method according to one of claims 1 to 8, in which, depending on the value of the determined additional power in balance with the total economic effort is determined whether the gas turbine (1) is temporarily stopped to eliminate the pollution and optionally an optimal time (34) for the temporary shutdown is determined. Method according to one of claims 1 to 9, wherein the determination of the Ansaugmassenstroms (48) takes place without information about the Brennstoffheizwert and / or fuel mass flow. Gas turbine plant comprising a gas turbine comprising a plurality of components (1) and having a control system (18) which is connected on the data input side to a number of sensors (22) arranged to determine input characteristics (40a, 40b, 40c) in the gas turbine (1) the control system (18) comprises a prognosis module (24) designed for carrying out the method according to one of claims 1 to 9. Gas turbine plant according to claim 11, in which data of a database (26) with comparison variables of structurally identical and / or building-like gas turbines (64) can be read into the prognosis module. A forecasting module (24) for use in a gas turbine plant, wherein the forecasting module (24) is adapted to carry out the method of any one of claims 1 to 9.

Images