RU2297613C2 - Method of diagnosing gas-turbine engine - Google Patents

Abstract

FIELD: testing of turbomachines.

SUBSTANCE: method can be used for inspection of technical condition of turbomachines due to comparison of spectral analysis data of vibration of working engine with initial data. Before testing spectral analysis in wide range, the spectral analysis of envelope of vibration signal is made; the signal is selected at characteristic frequencies. Main source resulting to changes in vibration are found from maximal deviations due to comparison amplitudes of components of received spectrum with the same values at initial state of engine. After that the defect is localized due to measurement and comparison with initial modulation components of vibration inside ranges of only those carrier frequencies, which frequencies are multiple to frequencies with highest deviation from initial values.

EFFECT: improved truth of results of diagnostics.

5 dwg

Description

The invention relates to the field of testing turbomachines, in particular to methods for monitoring their technical condition using vibration methods, and can be applied when testing the engine on a bench or in operating conditions (by plane or helicopter) after one of the forms of maintenance, as well as diagnosing the regulator and individual units, for example high-pressure plunger pumps.

As the initial signals capable of characterizing the technical condition of the elements of a gas turbine engine, first of all, the elements of the gas-air duct (GVT), high-frequency vibration of the engine housing generated by the aerodynamic interaction of rotating and fixed blade grids in a running engine can serve. To describe and analyze this interaction, models are usually used that take into account the pulsed nature of the impact of edge tracks on the blades following the flow.

Under the influence of various factors that arise during operation, such as, for example, damage to the blades, coking of the fuel nozzles, the precession movement of the rotor, local changes in the velocity vector of the air flow coming out of the blade grill take place. In this case, a distortion of the edge tracks presented in the form of a pulse sequence occurs, which can be represented as modulation of the initial sequence of pulses in amplitude or frequency. The main carrier frequency of such a pulse sequence, for example, from the impeller blades, is the repetition frequency of its blades relative to a fixed point (the product of the rotor speed and the number of blades). In the spectrum of vibration created by the impeller, not one but several carrier frequencies may be present, including multiples of the main carrier, the presence of which is due to the shape of real pulses.

The main diagnostic information appears during the interaction of the air flow with the elements of the engine’s GW and is contained in the parameters characterizing the modulation of the edge tracks. The characteristics of the defect (s) determine the parameters of the modulated signal. By analyzing the change in the vibration signal, defects can be diagnosed.

In previous solutions, to monitor the technical condition of the turbomachine, various options were used to measure the vibration components at the blade speed and rotor frequencies and multiples of them, followed by the calculation of combined parameters. When a large-scale defect occurs in the engine’s GW, leading to modulation of the edge traces of the working blades, the vibration spectrum around each carrier frequency changes. The components of the spectrum at frequencies that are multiples of the rotor speed increase or decrease due to a change in the unevenness of the impellers. In addition, components appear that are offset in frequency from the carrier due to phase distortion in the pulse train.

Adopted as a prototype, a method for diagnosing a gas turbine engine according to the application No. 13275 of the Republic of Latvia provides both the determination of the state of the GWT of the engine as a whole and the localization of the source of deterioration of its performance. This is achieved by the fact that after measuring and spectral analysis of the vibration of the working engine, using the vibration transducer installed on its body, the signal spectrum energy is measured above and below each carrier frequency of all engine stages in side bands with a width less than the rotor speed, and the obtained values are compared with the same values measured in the initial state of the engine. In this case, the state of the gas-air path of the engine as a whole is judged by the deviation of the energy values of the spectra measured in the side frequency bands at the carrier frequencies of all engine stages from the original ones. To determine the source of engine performance deterioration, the energy values measured in the side frequency bands at the carrier frequencies of each engine stage are compared with the same values in its initial state.

Modern gas turbine engines are made according to a 2- or even 3-shaft scheme and each rotor has from 1-2 to 10-12 stages. In each stage there is an impeller and a guide or straightening apparatus with an inherent number of blades. In the vibration spectrum of a gas turbine engine, in this case, there can be several tens or even hundreds of types of carrier frequencies. The amplitude-frequency characteristics of the vibrational path of the engine are very complex, so the nature of the change in the modulation components in different ranges will be different, which will affect the reliability of the diagnosis.

Thus, to determine the state of the engine and, moreover, to localize a defect, it is necessary to examine dozens or even hundreds (if several sensors or measurement directions are used) of frequency ranges in order to identify and measure in them vibration components that are multiples of the rotor speed or band components vibrations near carrier frequencies. For example, in the case of applying the above method for diagnosing the TVZ-117 engine, it is necessary to take into account 11 stages differing in the number of working blades, the number of sensors is 2 (one each on the front support housing and the turbine housing) and the number of measurement directions is 3 (3-axis vibration transducers) . Moreover, if we restrict ourselves to only two frequency ranges for each first carrier frequency (lower and higher than the carrier), the number of study ranges will be 120. Table 1 illustrates the required number of measurements (marked with an x) without taking into account the paired measurement ranges near each carrier in the ranges frequencies corresponding to the first carrier frequencies (in relation to the TVZ-117 engine).

Table 1 Sensor Installation Location Front support Turbine Direction of vibration overload Step number Number of blades frequency Hz axial radial tangential axial radial tangential Compressor one 37 11396 x x x x x x 2 43 13244 x x x x x x 3 59 18172 x x x x x x four 67 20636 x x x x x x 5 73 22484 x x x x x x 6 81 24948 x x x x x x 7 89 27412 x x x x x x 8 89 27412 9 89 27412 10 89 27412 eleven 89 27412 12 89 27412 compressor turbine one 133 40964 x x x x x x 2 101 31108 x x x x x x free turbine 3 64 16000 x x x four 51 12750 x x x

Given the need to diagnose not only GW, but also other parts of the engine, it is necessary to further expand the list of carrier frequencies, including, in addition to the repetition frequencies (and multiples of them) of the impeller blades, also the frequencies generated by rotating mechanisms with a pulsed nature of the interaction of elements including: frequencies of rolling of rolling elements in bearings, re-engagement of teeth in gears, operation of plungers in plunger pumps.

Conducting such a study in a wide range of frequencies under operating conditions is difficult, firstly, due to the limited frequency range of the vibration measuring and analysis instruments, and secondly, due to the need for an operative decision on the use of the engine. In view of the foregoing, in practice, it is necessary to limit the number of frequency ranges under study, which significantly narrows the reliability of engine monitoring in real-life operating conditions.

The aim of the present invention is to increase the reliability of engine diagnostics in operating conditions.

This goal is achieved by the fact that in a diagnostic method for engines based on measuring vibration of a running engine using a vibration transducer installed on its body and comparing data obtained as a result of spectral analysis of vibration with the same values measured in the initial state of the engine, spectral analysis of the envelope is first performed vibration signal emitted at characteristic frequencies, measure the amplitudes of the components of the obtained spectrum in the range from zero to rotor speed, having the highest rotation speed, the obtained values are compared with the values in the initial state, and the place of the dominant sources of vibration changes is judged by the frequencies of the components having the largest deviations of the measured values from the original ones. After determining the engine aggregates, which are the sources of modulation of the carrier frequencies, the defect is localized to a separate unit of this unit by vibration spectra in a wide frequency range by measuring and comparing with the initial values of the modulation components of vibration in the ranges of only those carriers whose frequencies are multiples of the frequencies of the dominant sources.

The invention can be illustrated by graphs, which represent:

figure 1 is a spectrum of vibration overload (a sensor on the housing of the front support of the rotor mounted in the radial direction); 1a - in the initial state of the engine, 1b - with the introduced imbalance;

figure 2 - spectrum vibration (sensor on the turbine housing, mounted in the radial direction): 2A - in the initial state of the engine, 2b - with the introduced imbalance;

figure 3 - spectrum envelope of vibration overload at the 1st characteristic frequency - the repetition frequency of the blades of the last stage of the free turbine (sensor on the turbine housing mounted in the radial direction): 3A - in the initial state of the engine, 3b - with the introduced imbalance;

figure 4 - spectrum envelope of vibration overload at the 2nd characteristic frequency - the repetition frequency of the blades of the 3rd stage of the compressor (the sensor on the front support housing mounted in the radial direction): 4a - in the initial state of the engine, 4b - with the imbalance introduced;

figure 5 - spectrum envelope of vibration overload at the 3rd characteristic frequency - the repetition frequency of the blades of the 4th stage of the compressor (sensor on the front support housing mounted in the radial direction): 5a - in the initial state of the engine, 5b - with the introduced imbalance.

Figure 1-5 adopted the following conventions:

TC - component of the vibration spectrum at the rotor speed of the turbocompressor; ST is a component of the vibration spectrum at the rotor speed of a free turbine; cp is a component of the vibration spectrum at the rotational speed of the central drive; RP - the component of the vibration spectrum at the frequency of rotation of the drive springs; 2рп - component of the vibration spectrum at the frequency of the 2nd harmonic of rotation of the drive spring; PNRM. - component of the vibration spectrum at the rotational speed of the pump-regulator drive.

The inventive method is implemented as follows.

On each rotor of the engine and in its units indicated above, one of the main carrier frequencies can be distinguished one characteristic frequency associated with the node, optimal for control. The concept of characteristic frequency includes the value of the central frequency, which is determined by the product of the rotor (node) rotation frequency by the number of elements (for example, blades or plungers) and the measurement frequency bandwidth. The measurement frequency band is set as a percentage of the center frequency value. Before starting the tests, a list of characteristic frequencies for the test engine is determined, then a vibration transducer is installed on the gas turbine engine body, the frequency range of which is sufficient to measure vibration in a wide range: from several Hz to frequencies exceeding the frequencies of the working blades of the controlled engine stages.

The engine is brought to the operating mode, a spectral analysis of the envelope of vibration emitted at the characteristic frequencies is carried out, and the components of this spectrum are measured in the range from zero to the rotational speed of the rotor having the highest rotation speed. The obtained values are compared with the values of the same components measured in the initial state of the engine. The frequencies of the components having the largest deviations of the measured values from the original ones determine the dominant (main) sources of vibration change. The defect is localized according to the vibration spectra measured over a wide range, by measuring and comparing with the initial values of the modulation components of the vibration only those carriers whose frequencies are multiples of the frequencies of the main sources.

An example implementation of the method

The method was implemented to assess changes in the state of the helicopter gas turbine engine TVZ-117.

During engine tests at the plant’s bench, vibration was measured on the front support housing (compressor) —the front sensor and on the turbine housing — the rear sensor, in the frequency range from 0 Hz to 25 kHz. 3-axis vibration transducers measured vibration at each point simultaneously in 3 directions. A comparative change in the vibrational characteristics was observed when comparing the test results of the same engine in good condition and with additional unbalanced masses simulating possible damage introduced into the compressor turbine and into a free turbine. In this case, the imbalance created in the compressor turbine was created obviously greater than in a free turbine.

When tested in good condition, the characteristics of the engine corresponded to the technical conditions, and the values of the vibration parameters of the engine were taken as the initial ones.

At the first stage of processing the results, the method of spectral analysis of the envelope of vibration was used. To analyze the envelope, characteristic frequencies were selected that corresponded to the working frequencies of the blades of the 3rd (18172 Hz) and 4th (20636 Hz) compressor stages and the last stage of a free turbine (12750 Hz). In parentheses are the values of the characteristic frequencies in bold in table 1, corresponding to the nominal operating mode of the engine.

Vibration parameters were measured at the rated engine operating mode. The rotor speed of the turbocompressor in this mode was 308 s ^-1 , while the rotation speed of a free turbine was 250 s ^-1 . According to the test results, a spectral analysis of vibration was initially carried out in the frequency range corresponding to the rotational speed of the turbocompressor rotor. The analysis results are shown in figure 1. On the front support, the vibration at the rotational frequencies of both rotors did not practically change when a defect was introduced (figa - initial state, figb - when making an imbalance). The vibration at the rotor speed of the turbocompressor, measured by the sensor on the rear support, increased by less than 50% (figa - initial state, figb - when making an imbalance). Since the indicated change in vibration is within the range of vibration with the rotor speed from start to start, it is not possible to determine the source of disturbance only by changing the level in the steady state.

At the next stage, the envelope of vibration from the sensor on the turbine was isolated at the first characteristic frequency of 12750 Hz - the repetition frequency of the working blades of the last stage of the free turbine and its spectral analysis was carried out in the band 0-400 Hz. (The vibration envelope at the frequencies of the compressor turbine blades was not investigated, since the indicated frequencies are outside the frequency range of the available equipment.)

Figure 3 shows the obtained spectra of the vibration envelope, from which it can be seen that, with the practical absence of noticeable narrow-band components, in the initial state (Fig. 3a), the appearance of a component with the rotational speed of the turbocompressor rotor in Fig. 3b clearly indicates the dominant source of vibration excitation - mass imbalance turbocharger rotor.

Figures 4 and 5 show the spectra of the envelope of vibration measured by the sensor on the front support of the compressor and allocated at other characteristic frequencies - the repetition frequencies of the blades of the 3rd (19212 Hz) and 4th (21817 Hz) stages of the compressor. Due to the proximity of the vibration sensor to the drives, here, gears driven by the rotor of the turbocompressor also act as modulation sources.

The envelope spectra show an obvious difference between the structure of the envelope spectrum in the initial state and when an imbalance is introduced. For example, in the 3rd stage, the component at the rotational speed of the turbocompressor rotor in its initial state (Fig. 4a) practically does not stand out from “noise” (random broadband vibration), but when an imbalance is introduced (Fig. 4b) it increases by more than two times. In the 4th stage (figa - the initial state and figb - when making an imbalance), the component at the rotor frequency when making the imbalance also changes almost twice, however, in a smaller direction. Therefore, the dominant nature of the change in rotor vibration of the rotor of the turbocharger is obvious. In this case, the remaining components associated with the vibration of the gears driven by the rotor of the turbocompressor rotate to a much lesser extent and their change is within the usual range of the scatter components.

The change in vibration at the turbocharger rotor speed measured on the compressor turned out to be smaller than on the turbine, which allows us to conclude that the turbocompressor rotor is the dominant source of vibration excitation in the engine, but also partially localize the defect, indicating that this is compressor turbine.

Thus, a change in the spectrum of the envelope of vibration (when introducing an imbalance) indicated that:

- the most likely source of vibration is the turbocharger rotor,

- The most likely location for the defect is the compressor turbine.

Based on the conclusions made, the continuation of the vibration study was reduced due to the exclusion from the total processing volume of the ranges measured by the front vibration sensor and the ranges associated with the frequencies of the compressor turbine blades.

Thus, the use of the proposed method allowed us to reduce the volume of vibration spectrum studies necessary for detecting a defect by 20 times (in table 1, 3 paired ranges in which it is necessary to conduct a detailed study are highlighted in bold instead of all 60 indicated in it) and provide increasing the efficiency of diagnostics in operating conditions.

Claims (1)

A method for diagnosing an engine based on measuring vibration of a running engine, spectral analysis of vibration and comparing the received data with the same values measured in the initial state of the engine, characterized in that they conduct spectral analysis of the envelope of the vibration signal emitted at characteristic frequencies, and the amplitudes of the components of the obtained spectrum are measured in the range from zero to the rotor speed having the highest rotation speed, the obtained values are compared with the same values in and a similar state, the place of the main sources of vibration changes is judged by the frequencies of the components having the largest deviations of the measured values from the original ones, and the defect is localized by vibration spectra in a wide frequency range by measuring and comparing with the initial values of the modulation components of vibration in the ranges of only those bearing whose frequencies are multiples of the frequencies of the main sources.

Images