CN101458158A - Steam turbine plain bearing failure diagnosis method based on acoustic emission detection and device thereof - Google Patents

Abstract

The invention relates to a method for diagnosing the faults of the sliding bearings of a steam turbine based on acoustic emission detection. The steps are as follows: (1) the arrangement of acoustic emission sensors and the acquisition of emission signals: the acoustic emission sensors are arranged on the sliding bearings of a steam turbine generator unit for receiving acoustic emission signals of the sliding bearings, and the received acoustic emission signals are transmitted to an acoustic emission detection system; (2) the acoustic signals are processed and analyzed, and the characteristics are extracted; (3) fault diagnosis: according to the acquired various characteristics, depending on empirical values or an expert system, the various characteristics are compared respectively, thus the diagnosis of the faults of the sliding bearings of the steam turbine generator unit is finished. The device of the invention comprises the acoustic emission detection system and the acoustic emission sensors, wherein, the acoustic emission sensors are arranged on the sliding bearings of the steam turbine generator unit. The invention provides the method for diagnosing the faults of the sliding bearings of the steam turbine based on the acoustic emission detection and a device thereof which have the advantages that the operation is simple and rapid, the diagnosing precision is high, and higher automation degree can be met.

Description

Steam turbine plain bearing failure diagnosis method and device thereof based on acoustic emission detection

Technical field

The present invention is mainly concerned with the steam turbine field, refers in particular to a kind of steam turbine plain bearing failure diagnosis method and device thereof based on acoustic emission detection.

Background technology

Sliding bearing is the vitals of Turbo-generator Set, and its poor working environment, load weigh, and various faults take place during operation easily, even causes device damage and security incident takes place, and influences safety and economic operation.Therefore, the duty of sliding bearing is carried out on-line monitoring and fault diagnosis is very important.At present, monitoring for sliding bearing in the Turbo-generator Set generally is to be undertaken by artificial mode, the testing staff is by manual detection or parameter or curve by showing on some supplementary instruments, judge the state of sliding bearing by rule of thumb, like this to testing staff's competency profiling height, testing process is intelligent inadequately, inefficiency.

Summary of the invention

The problem to be solved in the present invention just is: at the technical matters that prior art exists, the invention provides a kind of simple and compact for structure, with low cost, easy and simple to handle, quick, diagnostic accuracy is high, can satisfy the steam turbine plain bearing failure diagnosis method and the device thereof based on acoustic emission detection of high automation degree.

For solving the problems of the technologies described above, the solution that the present invention proposes is: a kind of steam turbine plain bearing failure diagnosis method based on acoustic emission detection is characterized in that step is:

(1), calibrate AE sensor is installed, obtained acoustic emission signal: calibrate AE sensor is installed on the sliding bearing of Turbo-generator Set, is used for receiving the acoustic emission signal of sliding bearing, and send the acoustic emission signal that receives to acoustic emission detection system;

(2), acoustic emission signal is handled, analyzed, extract feature: acoustic emission signal is carried out event count calculating, Ring-down count calculating, energy calculating, signal amplitude calculating, centre frequency calculating, the calculating of spectrum energy instability, spectrum analysis and power spectrumanalysis, extract following characteristics simultaneously in the signal after handling, analyzing: event count feature, Ring-down count feature, energy feature, signal amplitude feature, centre frequency feature, spectrum energy instability feature, time-frequency characteristics and power spectrum characteristic;

(3) fault diagnosis: according to each category feature that obtains in the step (2),, each category feature is compared respectively, thereby finished fault diagnosis to the Turbo-generator Set sliding bearing according to empirical value or expert system.

A kind of steam turbine plain bearing failure diagnosis device based on acoustic emission detection, it is characterized in that: it comprises acoustic emission detection system and is installed on calibrate AE sensor on the sliding bearing of Turbo-generator Set, described calibrate AE sensor links to each other with prime amplifier by cable, the output terminal of prime amplifier links to each other with acoustic emission detection system, acoustic emission detection system is used for receiving the acoustic emission signal of sliding bearing, feature is handled, analyzes, extracted to described acoustic emission detection system to acoustic emission signal, and carry out fault diagnosis.

Compared with prior art, advantage of the present invention just is: the steam turbine plain bearing failure diagnosis method and the device thereof that the present invention is based on acoustic emission detection, proposition is based on the steam turbine plain bearing failure diagnosis method of acoustic emission detection, utilize the steam turbine plain bearing failure diagnosis system of this method exploitation, can be by acoustic emission signal, the extraction signal characteristic that detects sliding bearing, finish automatic judgement to the steam turbine plain bearing fault, its have easy and simple to handle, quick, diagnostic accuracy is high, can satisfy advantages such as high automation degree.Because acoustic emission testing technology is a kind of detection of dynamic, the generation of detection failure and evolution in real time compared with existing method for diagnosing faults to steam turbine plain bearing, has the out of order advantage of early diagnosis; The present invention proposes the steam turbine plain bearing failure diagnosis method based on acoustic emission detection, acoustic emission signal to steam turbine plain bearing is obtained in real time, signal analysis, feature extraction and utilize certain rule to carry out fault judgement, for trouble-saving generation, safeguard that unit safety has important value.

Description of drawings

Fig. 1 is a workflow synoptic diagram of the present invention;

Arrangement synoptic diagram when Fig. 2 is practical application of the present invention;

Fig. 3 is the synoptic diagram of acoustic emission signal analysis process of the present invention;

Fig. 4 is the synoptic diagram that sound emission signal characteristic of the present invention extracts.

Embodiment

Below with reference to the drawings and specific embodiments the present invention is described in further details.

As shown in Figure 1, Figure 2, Figure 3 and Figure 4, a kind of steam turbine plain bearing failure diagnosis method based on acoustic emission detection the steps include:

(1), install calibrate AE sensor, obtain acoustic emission signal: calibrate AE sensor 1 is installed on the sliding bearing 2 of Turbo-generator Set 3, be used for receiving the acoustic emission signal of sliding bearing 2, and send the acoustic emission signal that receives to acoustic emission detection system 5;

(2), acoustic emission signal is handled, analyzed, and carry out feature extraction: acoustic emission signal is carried out event count calculating, Ring-down count calculating, energy calculating, signal amplitude calculating, centre frequency calculating, the calculating of spectrum energy instability, spectrum analysis and power spectrumanalysis, extract following characteristics simultaneously in the signal after handling, analyzing: event count feature, Ring-down count feature, energy feature, signal amplitude feature, centre frequency feature, spectrum energy instability feature, time-frequency characteristics and power spectrum characteristic;

(3) fault diagnosis: according to each category feature that obtains in the step (2),, each category feature is compared respectively, thereby finished fault diagnosis to sliding bearing 2 on the Turbo-generator Set 3 according to empirical value or expert system.Following table 1 is sliding bearing 2 failure collection and characteristics of Acoustic Emission set on the Turbo-generator Set 3, and each category feature and the failure collection of collecting contrasted one by one, can judge which kind of fault sliding bearing 2 exists on the Turbo-generator Set 3.

Table 1 Turbo-generator Set sliding bearing failure collection and characteristics of Acoustic Emission set

If sliding bearing 2 has n kind status flag S ₁, S ₂..., S _n, and establish every kind of status flag have " having " (representing) with numerical value 1 and " nothing " (representing) with numerical value 0 two states, promptly

So the fault domain is U={ (S ₁, S ₂..., S _n)/S _j=0,1, j≤n}.Can determine relation between Turbo-generator Set sliding bearing fault and the characteristics of Acoustic Emission, to see Table 2.2 according to experimental study and theoretical analysis so obtain the characteristics of Acoustic Emission domain of sliding bearing 2 faults.

The characteristics of Acoustic Emission domain of table 2.2 sliding bearing fault

If sliding bearing 2 has the physical fault feature $u_0=(S_1^{\left(0\right)},S_2^{\left(0\right)},...,S_{20}^{\left(0\right)})$ , here, $0\≤S_i^{\left(0\right)}\≤1.$ With u ₀The substitution formula ${\mathrm{\μ}}_{A_i}\left(u\right)=\cos \sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(S_j-{S^i}_j)}^2/n}$ After, obtain ${\mathrm{\μ}}_{A_1}\left(u_0\right),{\mathrm{\μ}}_{A_2}\left(u_0\right),...,{\mathrm{\μ}}_{A_{21}}\left(u_0\right).$ So by formula ${\mathrm{\μ}}_{A_i}\left(u_0\right)=\max \{{\mathrm{\μ}}_{A_1}\left(u_0\right),{\mathrm{\μ}}_{A_2}\left(u_0\right),\·\·\·,{\mathrm{\μ}}_{A_m}\left(u_0\right)\}$ Choose a degree of membership maximal value

According to maximum membership grade principle, the diagnosable state that goes out sliding bearing 2 is A _i

Bump to rub and be example so that the part to take place in the unit starting, idiographic flow of the present invention is: calibrate AE sensor 1 is installed on the sliding bearing 2 of Turbo-generator Set 3, receives the acoustic emission signal of sliding bearing 2.This signal is sent to after prime amplifier 4 amplifies through cable, enters acoustic emission detection system 5.In acoustic emission detection system 5, finish the processing and the analysis of acoustic emission signal of the sliding bearing 2 of Turbo-generator Set 3, comprise: event count (rate) calculates, Ring-down count (rate) calculates, energy (rate) calculates, and signal amplitude is calculated, and centre frequency is calculated, the spectrum energy instability is calculated, spectrum analysis, power spectrumanalysis, correlation analysis.The feature of the acoustic emission signal of the sliding bearing 2 of extraction Turbo-generator Set 3, comprise: event count (rate) feature, Ring-down count (rate) feature, energy (rate) feature, signal amplitude feature, centre frequency feature, spectrum energy instability feature, time-frequency characteristics, power spectrum characteristic, correlation analysis feature.The part takes place in the unit starting to be bumped when rubbing, characteristics of Acoustic Emission is: event count rate has peak value under the slow-speed of revolution, Ring-down count rate has peak value under the slow-speed of revolution, the spectrum energy instability is increased to high 1 value, maximum coefficient of autocorrelation is increased to high 1 value, specific energy increases to high 1 value, and amplitude increases to high 1 value.At last with the fault signature that extracts $u_0=(S_1^{\left(0\right)},S_2^{\left(0\right)},...,S_{20}^{\left(0\right)}),$ Here, $0\≤S_i^{\left(0\right)}\≤1.$ With u ₀The substitution formula ${\mathrm{\μ}}_{A_i}\left(u\right)=\cos \sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(S_j-{S^i}_j)}^2/n}$ After, obtain ${\mathrm{\μ}}_{A_1}\left(u_0\right),{\mathrm{\μ}}_{A_2}\left(u_0\right),...,{\mathrm{\μ}}_{A_{21}}\left(u_0\right).$ So by formula ${\mathrm{\μ}}_{A_i}\left(u_0\right)=\max \{{\mathrm{\μ}}_{A_1}\left(u_0\right),{\mathrm{\μ}}_{A_2}\left(u_0\right),\·\·\·,{\mathrm{\μ}}_{A_m}\left(u_0\right)\}$ Choose a degree of membership maximal value

According to maximum membership grade principle, diagnosable to go out the sliding bearing state be A ₁, be in the unit starting part and bump and rub.

As shown in Figure 2, steam turbine plain bearing failure diagnosis device based on acoustic emission detection of the present invention, it comprises acoustic emission detection system 5 and is installed on calibrate AE sensor 1 on the sliding bearing 2 of Turbo-generator Set 3, calibrate AE sensor 1 links to each other with prime amplifier 4 by cable, the output terminal of prime amplifier 4 links to each other with acoustic emission detection system 5, acoustic emission detection system 5 is used for receiving the acoustic emission signal of sliding bearing 2, feature is handled, analyzes, extracted to 5 pairs of acoustic emission signals of described acoustic emission detection system, and carry out fault diagnosis.

With the oil film whirl fault is example, and idiographic flow of the present invention is: calibrate AE sensor 1 is installed on the sliding bearing 2 of Turbo-generator Set 3, receives the acoustic emission signal of sliding bearing 2.This signal is sent to after prime amplifier 4 amplifies through cable, enters acoustic emission detection system 5.In acoustic emission detection system 1, finish the processing and the analysis of acoustic emission signal of the sliding bearing 2 of Turbo-generator Set 3, comprise: event count (rate) calculates, Ring-down count (rate) calculates, energy (rate) calculates, and signal amplitude is calculated, and centre frequency is calculated, the spectrum energy instability is calculated, spectrum analysis, power spectrumanalysis, correlation analysis.The feature of the acoustic emission signal of the sliding bearing 2 of extraction Turbo-generator Set 3, comprise: event count (rate) feature, Ring-down count (rate) feature, energy (rate) feature, signal amplitude feature, centre frequency feature, spectrum energy instability feature, time-frequency characteristics, power spectrum characteristic, correlation analysis feature.When oil whirl took place, characteristics of Acoustic Emission was: the spectrum energy instability increases to high 2 values and fluctuation, and maximum coefficient of autocorrelation increases to high 2 values and fluctuation, and specific energy increases to high 2 values and fluctuation, and amplitude increases to high 2 values and fluctuation.At last with the fault signature that extracts $u_0=(S_1^{\left(0\right)},S_2^{\left(0\right)},...,S_{20}^{\left(0\right)}),$ Here, $0\≤S_i^{\left(0\right)}\≤1.$ With u ₀The substitution formula ${\mathrm{\μ}}_{A_i}\left(u\right)=\cos \sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(S_j-{S^i}_j)}^2/n}$ After, obtain

So by formula ${\mathrm{\μ}}_{A_i}\left(u_0\right)=\max \{{\mathrm{\μ}}_{A_1}\left(u_0\right),{\mathrm{\μ}}_{A_2}\left(u_0\right),\·\·\·,{\mathrm{\μ}}_{A_m}\left(u_0\right)\}$ Choose a degree of membership maximal value

According to maximum membership grade principle, diagnosable to go out the sliding bearing state be A ₈, be oil whirl.

With the oil whip fault is example, and idiographic flow of the present invention is: calibrate AE sensor 1 is installed on the sliding bearing 2 of Turbo-generator Set 3, receives the acoustic emission signal of sliding bearing 2.This signal is sent to after prime amplifier 4 amplifies through cable, enters acoustic emission detection system 5.In acoustic emission detection system 5, finish the processing and the analysis of acoustic emission signal of the sliding bearing 2 of Turbo-generator Set 3, comprise: event count (rate) calculates, Ring-down count (rate) calculates, energy (rate) calculates, and signal amplitude is calculated, and centre frequency is calculated, the spectrum energy instability is calculated, spectrum analysis, power spectrumanalysis, correlation analysis.The feature of the acoustic emission signal of the sliding bearing 2 of extraction Turbo-generator Set 3, comprise: event count (rate) feature, Ring-down count (rate) feature, energy (rate) feature, signal amplitude feature, centre frequency feature, spectrum energy instability feature, time-frequency characteristics, power spectrum characteristic, correlation analysis feature.When oil whirl took place, characteristics of Acoustic Emission is: event count rate increased suddenly, and Ring-down count rate increases suddenly, and the spectrum energy instability increases to high 3 values, and maximum coefficient of autocorrelation increases to high 3 values, and specific energy increases to high 3 values, and amplitude increases to high 3 values.At last with the fault signature that extracts $u_0=(S_1^{\left(0\right)},S_2^{\left(0\right)},...,S_{20}^{\left(0\right)}),$ Here, $0\≤S_i^{\left(0\right)}\≤1.$ With u ₀The substitution formula ${\mathrm{\μ}}_{A_i}\left(u\right)=\cos \sqrt{\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(S_j-{S^i}_j)}^2/n}$ After, obtain

So by formula ${\mathrm{\μ}}_{A_i}\left(u_0\right)=\max \{{\mathrm{\μ}}_{A_1}\left(u_0\right),{\mathrm{\μ}}_{A_2}\left(u_0\right),\·\·\·,{\mathrm{\μ}}_{A_m}\left(u_0\right)\}$ Choose a degree of membership maximal value

According to maximum membership grade principle, diagnosable to go out the sliding bearing state be A ₉, be oil whip.

The above only is a preferred implementation of the present invention, and protection scope of the present invention also not only is confined to the foregoing description, and all technical schemes that belongs under the thinking of the present invention all belong to protection scope of the present invention.Should be pointed out that for those skilled in the art in the some improvements and modifications that do not break away under the principle of the invention prerequisite, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (2)

1, a kind of steam turbine plain bearing failure diagnosis method based on acoustic emission detection is characterized in that step is:

(1), calibrate AE sensor is installed, obtained acoustic emission signal: calibrate AE sensor is installed on the sliding bearing of Turbo-generator Set, is used for receiving the acoustic emission signal of sliding bearing, and send the acoustic emission signal that receives to acoustic emission detection system;

(2), acoustic emission signal is handled, analyzed, and carry out feature extraction: acoustic emission signal is carried out event count calculating, Ring-down count calculating, energy calculating, signal amplitude calculating, centre frequency calculating, the calculating of spectrum energy instability, spectrum analysis and power spectrumanalysis, extract following characteristics simultaneously in the signal after handling, analyzing: event count feature, Ring-down count feature, energy feature, signal amplitude feature, centre frequency feature, spectrum energy instability feature, time-frequency characteristics and power spectrum characteristic;

(3) fault diagnosis: according to each category feature that obtains in the step (2),, each category feature is compared respectively, thereby finished fault diagnosis to the Turbo-generator Set sliding bearing according to empirical value or expert system.

2, a kind of steam turbine plain bearing failure diagnosis device based on acoustic emission detection, it is characterized in that: it comprises acoustic emission detection system (5) and is installed on calibrate AE sensor (1) on the sliding bearing (2) of Turbo-generator Set (3), described calibrate AE sensor (1) links to each other with prime amplifier (4) by cable, the output terminal of prime amplifier (4) links to each other with acoustic emission detection system (5), acoustic emission detection system (5) is used for receiving the acoustic emission signal of sliding bearing (2), described acoustic emission detection system (5) is handled acoustic emission signal, analyze, extract feature, and carry out fault diagnosis.

Images