CN106596110A - Online data-based hydroelectric generating set hydraulic unbalance fault automatic analysis and diagnosis method - Google Patents

Abstract

The invention provides an online data-based hydroelectric generating set hydraulic unbalance fault automatic analysis and diagnosis method. The method includes the following steps that: 1) the 1X component and blade frequency component of a swing and vibration signal under a no-load condition are obtained through calculation; 2) the 1X components and blade frequency components of a plurality of swing and vibration signals under different load conditions are obtained through calculation; and 3) the vector difference amplitude of the 1X components of the plurality of swing and vibration signals and the 1X component of the swing and vibration signal under the no-load condition and the vibration vector difference amplitude of the blade frequency components and the blade frequency component under the no-load condition are obtained through calculation; 4) the linear correlation coefficient of the vibration vector difference amplitude of the 1X components of the swing and vibration signals and loads and the linear correlation coefficient of the vibration vector difference amplitude of the blade frequency components and the loads are calculated; and 5) the correlation coefficients are compared with preset values, and factors resulting in a hydraulic unbalance fault are judged according to comparison results. The online data-based hydroelectric generating set hydraulic unbalance fault automatic analysis and diagnosis method of the present invention is easy to operate. According to the method, users do not need to perform complicated operation such as setting and data selection, and one-key completion type software operation is adopted.

Description

Examined based on the automatically analyzing for turbine-generator units waterpower imbalance fault of online data Disconnected method

Technical field

The present invention relates to a kind of turbine-generator units waterpower imbalance fault based on online data automatically analyzes diagnosis Method.

Background technology

The real-time diagnosis of Hydropower Unit running status are directly connected to safe and stable operation, power quality and the electricity in power station The important economic benefits indicator such as power production cost, its social benefit is huge.With power plant scale and monitoring aid system not Break and expand, the control of unit and Monitoring Data quantity of information are increasing, run operator's having in real time to operating states of the units Effect monitoring, equipment fault to be made quickly and accurately judge become more and more difficult, therefore, the intelligent Hydropower Unit of research and development Fault diagnosis system is very important.There are unavoidably various abnormal conditions in Hydropower Unit, same in running Abnormal phenomena may have different producing causes, and the failure for occurring has randomness, many of which thing power station staff Cannot in advance, directly detect, typically to search according to the personal experience of staff and to the analysis of Monitoring Data therefore The reason for barrier occurs and position, thus have certain subjectivity and limitation.Therefore, it is that the normal safe for ensureing Hydropower Unit is transported OK, its running status is detected, failure symptom is found in time, it is engineering circles dream to accomplish that " maintenance forecasting " prevents trouble before it happens Sleep in the hope of ideal, be also the developing direction of large-scale power station unit maintenance.Fault Diagnosis of Hydroelectric Generating Set is to rise in recent years One identification operating states of the units science, be the integrated system of a complexity, the professional range being related to is wide.Timely and accurately The state and failure of Hydropower Unit are diagnosed to be, are significant to improving Hydropower Unit job stability, security reliability.Pass The fault diagnosis of system turbine-generator units mainly has following 4 kinds of methods

1st, the diagnostic method based on signal processing

It is basis and the premise of various diagnostic methods based on the diagnostic method of signal processing, examines as a kind of traditional failure Disconnected method, according to system the eigenvalue of signal, such as amplitude, the dependency between phase place, frequency, variance, and signal can be surveyed, Certain relation existed between the source of trouble, by signal processing and feature extraction, finds out its mapping between the source of trouble and closes It is carrying out the fault diagnosis of operation equipment.At present conventional signal processing method include temporal analysiss, frequency domain analysises, when Frequency domain analysises etc..

2nd, the diagnostic method based on analytical model

It is the diagnostic method based on control theory based on the diagnostic method of analytical model.Equipment is regarded as one to have Certain input, the dynamical system of output relation, according to the input of system, output relation sets up mathematical expression or Analytical Expression mould Type, produces residual error, according to model using methods such as observer, wave filter, equivalent space equation, parameter model estimation and identifications Residual error is judging the probability for breaking down.Hydropower Unit is the nonlinear and time-varying system of a complexity, and preparation is compared in foundation Mathematical model is often hardly resulted in, and which also limits diagnostic method the answering in Approach for Hydroelectric Generating Unit Fault Diagnosis based on analytical model With.At present conventional analytical model method includes method for parameter estimation, method for estimating state, Parity space approach etc..

3rd, the diagnostic method based on Heuristicses

The knowledge representation for being related to research field, the method for processing and applying, are referred to as Knowledge based engineering diagnostic method, as before Barrier sign tree method, expert system method etc., they the characteristics of be that Professional knowledge is carried out into table by way of semantic and framework Reach, the diagnosis of failure is then carried out using reasoning and reasoning process, such as production rule reasoning, logical reasoning, Illegible knowledge illation Deng.At present the diagnostic method of conventional Heuristicses includes signed digraph analysis method, Fault Tree Analysis, specialist system Method of diagnosis etc..

4th, the diagnostic method based on data-driven

It is exactly the online or offline status data for utilizing equipment long term accumulation based on the method for diagnosing faults of data-driven, Without accurate analytical mathematical models, applied statistics analytical model is identified, nerual network technique or support vector machine The methods such as category of model, are learnt by data and are modeled, and equipment state is identified and is classified, and the system of finally giving may be deposited Failure.At present conventional data-driven method includes Neural Network Diagnosis Method, support vector machine diagnostic method, statistical Analysis diagnostic method, fuzzy diagnosis method etc..

The defect of traditional fault diagnosis technology：

The domestic research in terms of fault diagnosis technology is started late, and twentieth century begins one's study and attempts answering the end of the seventies With diagnostic techniquess, twentieth century proceeds by the research work of Intelligent fault diagnosis the nineties, and research method concentrates on fuzzy Logical approach, FTA, expert system technology, artificial neural network technology etc., wherein expert system technology and nerve net Network technology is the focus of application.Many monitoring diagnosis systems also begin to come into operation, but mostly concentrate on steam turbine and other The monitoring and fault diagnosis of rotating machinery, the application for Hydropower Unit is little.This is low mainly due to Hydropower Unit rotating speed, Safe operation to unit is not given to enough attention so that the research of Hydropower Unit on-line monitoring and fault diagonosing technology falls After other (large-scale) rotating machineries.

In recent years, the domestic research in Approach for Hydroelectric Generating Unit Fault Diagnosis field has made some progress, but main also in reason The effect of the intelligent diagnosis system application being equipped with by conceptual phase, the also engineering system without successful Application, the country is not Ideal, wherein preferably having reliable data storage management and analytic function, does not reach the function mesh of on-line automatic diagnosis Mark, it is impossible to provide effective diagnostic result to operations staff and production management personnel.Far from meeting Condition Maintenance for Hydraulic Power Plant It is actually needed.

The content of the invention

For above-mentioned technical problem, the invention provides a kind of turbine-generator units waterpower based on online data is uneven The automatic analyzing and diagnosing method of failure, step includes：

1) by monitoring record data on-line, calculate and obtain idle condition lower swing, the 1X components of vibration signal and blade Frequency component；

2) different loads are set, by monitoring record data on-line, is calculated and is obtained under different loads operating mode, multiple pendulum Dynamic, the 1X components of vibration signal and blade frequencies component；

3) swing, the 1X components of vibration signal under the zero load and different loads operating mode that are obtained according to step 1 and step 2 and Blade frequencies component, the arrow under calculating different loads operating mode, under multiple swings, the 1X components and idle condition of vibration signal Amount spread value and multiple blade frequencies components and the phasor difference amplitude under idle condition；

4) the phasor difference amplitude drawn by step 3 calculate swing, the 1X component vector spread values of vibration signal with The linearly dependent coefficient and blade frequencies component vector spread value of load and the linearly dependent coefficient of load；

5) correlation coefficient that the phasor difference amplitude and step 4 for being drawn by step 3 is drawn is carried out respectively with preset value Contrast, the factor for causing water conservancy imbalance fault is judged according to comparable situation.

Preferred version is：Select the data under close or identical head.

Preferred version is：Select the last idle condition lower swing, the 1X components of vibration signal and blade frequencies component.

Preferred version is：Calculated by FFT and obtain idle condition lower swing, the 1X components of vibration signal and blade frequencies point Amount.

Preferred version is：When the load for setting is as nominal load, under swing, the 1X components and idle condition of vibration signal Phasor difference amplitude change amplitude patient more than or equal to minimum and swing, the 1X component vector spread values of vibration signal and machine The linearly dependent coefficient of group load is more than or equal to preset value, then judge because runner body blade profile is inconsistent, leaf road is inconsistent, The waterpower imbalance fault of the inconsistent formation of each length of blade of runner.

Preferred version is：Vector spread when the load for setting is as nominal load, under blade frequencies component and idle condition The linear correlation of value change amplitude patient more than or equal to minimum and blade frequencies component vector spread value and unit load Coefficient be more than or equal to preset value, then judge due to runner envelope circularity irregular shape into waterpower imbalance fault.

Preferred version is：It is P that load is retrieved from the data record of on-line monitoring storage, and head is H⁰Data record, By FFT, calculate and obtain unloaded lower bearing bracket vibration, the 1X components of throwWith blade frequencies component(n is runner blade Piece number), i=i+1 is set, loading condiction is P=i*0.1Pr, and Pr is rated load：I=1,2 ... ... 10, thenForForP be 0.1Pr, 0.2Pr0.3Pr ... Pr；Meter Calculation drawn under different loads operating mode, phasor difference amplitude under multiple swings, the 1X components and idle condition of vibration signal and multiple Phasor difference amplitude under blade frequencies component and idle condition：

I=1, then 2 ... ... 10, Aⁱ _1XForAⁱ _nXFor

When the load of setting is more than or equal to nominal load, swing, the 1X component vector spread values of vibration signal are calculated With the linearly dependent coefficient and blade frequencies component vector spread value and the linearly dependent coefficient of load of load：

I=1,2 ... ... 10, if

Y_1X={ 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 }

Y_nX={ 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 }；

Wherein,Change lower limit for minimum patient throw, the 1X amplitudes under full-load conditions of frame vibration, Generally select 0.4 times to 0.7 times related GB operational shock amplitude；

C_{l_1X}For the correlation coefficient of 1X components, more than 0.7 value is generally selected；

For minimum patient throw, the nX amplitudes under full-load conditions of frame vibration change lower limit, the value Need to be determined according to the test case of actual set；

C_{l_nX}For the correlation coefficient of nX components, more than 0.7 value is generally selected.

Preferred version is：Step 5) include：

IfAnd C_1X≥C_{l_1X}, then exist because runner body blade profile is inconsistent, leaf road is inconsistent, The waterpower imbalance fault of the inconsistent formation of each length of blade of runner；

IfAnd C_nX≥C_{l_nX}, then exist due to runner envelope circularity irregular shape into waterpower it is uneven Weighing apparatus failure.

Beneficial effects of the present invention are as follows：

(1) it is easily operated.User of service need not be configured, select the complex operations such as data, adopt " one-key operation " formula Software operation.

(2) data selection, calculating, decision process automatization.All garbled datas and calculated according to failure or defect model Process, analysis ratiocination, the process that judges completed by computer, without the need for interactive operation in the middle of operator.

(3) clear and definite analyzing and diagnosing conclusion and possible maintenance suggestion are provided in reporting.

(4) user interface is arrived in output in the form of reporting, and report can be automatically converted to the forms such as WORD.

Description of the drawings

Fig. 1 leads throw 1X with active change curve for water；

Fig. 2 top covers vibrate 1X with active change curve；

Fig. 3 is the flow chart of the embodiment of the present invention；

Fig. 4 is the basic procedure of embodiment of the present invention report generation flow process.

Specific embodiment

Below the present invention is described in further detail, with make those skilled in the art with reference to specification word being capable of evidence To implement.

It should be appreciated that it is used herein such as " have ", "comprising" and " including " term do not allot one or many The presence or addition of individual other elements or its combination.

1st, the unbalanced failure mechanism of waterpower

Waterpower imbalance is a kind of most complicated out-of-balance force.

Waterpower out-of-balance force produce mechanism be：Flow skewness circumferentially.

The unbalanced reason of waterpower can be produced may following several：

(1) blade profile is inconsistent, if blade is than larger with the probability of the appearance problems of traditional template method processing；

(2) leaf road is inconsistent, if blade angle is inconsistent or blade servomotor stroke is inconsistent can produce this kind of problem；

(3) turning wheel interval is inconsistent, if each length of blade of runner is inconsistent, inflow between unit each stator can be caused to differ Cause, it is uneven so as to produce waterpower；

(4) guide vane servomotor stroke is inconsistent, if the connecting lever length of operation blade is inconsistent, it may appear that waterpower is uneven Weighing apparatus.

(5) runner envelope circularity is irregular, causes flow along the circumferential direction uneven, it may appear that waterpower is uneven.

2nd, the unbalanced fault signature of waterpower and Identification of parameter

Because the irregular and caused waterpower imbalance of runner envelope circularity can mainly cause the leaf of throw, top cover radial vibration Piece frequency and its multiple frequency change, and linearly increase with the increase of unit flow.

Due to water caused by the failure such as runner body blade profile is inconsistent, leaf road is inconsistent, each length of blade of runner is inconsistent Force unbalance power then causes unit throw, the 1X of top cover radial vibration to change, and is in the increase of unit flow It is linearly increasing.

For turbine-generator units, waterpower imbalance fault mainly by near runner to machine Group throw measuring point and frame in radial vibration all have an impact, and throw and top cover vibration (or water pilot bearing vibration) are especially led on water to be affected Substantially.Therefore identification waterpower imbalance fault mainly passes through throw and the 1X components of frame vibration and blade frequencies component Carry out feature identification.

As shown in Figure 1 and Figure 2, one typically due to waterpower imbalance fault caused by leaf road is inconsistent：Water lead throw and The horizontal 1X components of top cover increase and existing increase with unit load.

In actual on-line monitoring system, system can automatically select throw under idle condition, 1X point of radial vibration Amount vector is selected in identical head, throw, the 1X component vectors of radial vibration and benchmark 1X components under different load as benchmark The change of vector come recognize unit whether there is because runner body blade profile is inconsistent, leaf road is inconsistent, each length of blade of runner The waterpower imbalance fault of inconsistent formation；And by comparing different load under throw, radial vibration blade frequencies component arrow Amount with idle condition lower blade frequency component change come recognize unit with the presence or absence of due to runner envelope circularity irregular shape into Waterpower imbalance fault.

The data selected under close or identical head are limited, is because under same head, unit load and unit The close linear relationship of flow, easily carries out feature identification.

The invention provides a kind of automatically analyzing for turbine-generator units waterpower imbalance fault based on online data is examined Disconnected method, step includes：

1) by monitoring record data on-line, calculate and obtain idle condition lower swing, the 1X components of vibration signal and blade Frequency component；

2) different loads are set, by monitoring record data on-line, is calculated and is obtained under different loads operating mode, multiple pendulum Dynamic, the 1X components of vibration signal and blade frequencies component；

3) swing, the 1X components of vibration signal under the zero load and different loads operating mode that are obtained according to step 1 and step 2 and Blade frequencies component, the arrow under calculating different loads operating mode, under multiple swings, the 1X components and idle condition of vibration signal Amount spread value and multiple blade frequencies components and the phasor difference amplitude under idle condition；

4) the phasor difference amplitude drawn by step 3 calculate swing, the 1X component vector spread values of vibration signal with The linearly dependent coefficient and blade frequencies component vector spread value of load and the linearly dependent coefficient of load；

5) correlation coefficient that the phasor difference amplitude and step 4 for being drawn by step 3 is drawn is carried out respectively with preset value Contrast, the factor for causing water conservancy imbalance fault is judged according to comparable situation.

Preferred version is：Select the data under close or identical head.

Preferred version is：Select the last idle condition lower swing, the 1X components of vibration signal and blade frequencies component.

Preferred version is：Calculated by FFT and obtain idle condition lower swing, the 1X components of vibration signal and blade frequencies point Amount.

Preferred version is：When the load for setting is as nominal load, under swing, the 1X components and idle condition of vibration signal Phasor difference amplitude change amplitude patient more than or equal to minimum and swing, the 1X component vector spread values of vibration signal and machine The linearly dependent coefficient of group load is more than or equal to preset value, then judge because runner body blade profile is inconsistent, leaf road is inconsistent, The waterpower imbalance fault of the inconsistent formation of each length of blade of runner.

Preferred version is：Vector spread when the load for setting is as nominal load, under blade frequencies component and idle condition The linear correlation of value change amplitude patient more than or equal to minimum and blade frequencies component vector spread value and unit load Coefficient be more than or equal to preset value, then judge due to runner envelope circularity irregular shape into waterpower imbalance fault.

Embodiment

As shown in figure 3, a kind of the invention provides turbine-generator units waterpower imbalance fault based on online data Automatic analyzing and diagnosing method, step includes：

1) select to monitor record data on-line under the last idle condition；

2) head data H are recorded⁰, by FFT, calculate and obtain unloaded lower bearing bracket vibration, the 1X components of throwWith Blade frequencies component(n is runner bucket number)；

3) it is P load to be retrieved from the data record of on-line monitoring storage, and head is H⁰Data record, by FFT become Change, calculate and obtain unloaded lower bearing bracket vibration, the 1X components of throwWith blade frequencies component(n is runner bucket number), if Determine i=i+1, loading condiction is P=i*0.1P_r, P_rFor rated load.For example：I=1,2 ... ... 10, thenForForP is 0.1P_r、0.2P_r0.3P_r,,……P_r；

4) vector under calculating different loads operating mode, under multiple swings, the 1X components and idle condition of vibration signal Spread value and multiple blade frequencies components and the phasor difference amplitude under idle condition

Such as i=1, then 2 ... ... 10, Aⁱ _1XForAⁱ _nXFor

5) when the load of setting is more than or equal to nominal load, swing, the 1X component vector spreads of vibration signal are calculated Value and the linearly dependent coefficient and blade frequencies component vector spread value of load and the linearly dependent coefficient of load

Such as i=1,2 ... 10

If

Y_1X={ 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 }

Y_nX={ 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 }；

6) after Characteristic parameter identification, the decision condition of waterpower imbalance fault becomes simple, and actual conditions is such as Under：

IfAndSo exist because runner body blade profile is inconsistent, leaf road differs Cause, the waterpower imbalance fault of the inconsistent formation of each length of blade of runner；

IfAnd C_nX≥C_{l_nX}, then exist due to runner envelope circularity irregular shape into waterpower it is uneven Weighing apparatus failure；

In above formula:

1)For minimum patient throw, the 1X amplitudes under full-load conditions of frame vibration change lower limit, lead to Often select 0.4 times to 0.7 times related GB operational shock amplitude；

2)C_{l_1X}For the correlation coefficient of 1X components, more than 0.7 value is generally selected；

3)For minimum patient throw, the nX amplitudes under full-load conditions of frame vibration change lower limit, it is somebody's turn to do Value needs to be determined according to the test case of actual set；

4)C_{l_nX}For the correlation coefficient of nX components, more than 0.7 value is generally selected；

7) diagnosis report is automatically analyzed

Automatically analyze diagnostic function be system according to failure mechanism, suitable data are chosen automatically, carry out automatically analyzing and Statistic analysis, and provide assay diagnostic result.Whole analysis procedures system can be automatically performed and need not manually operate.Report The product process basic procedure of announcement is as shown in Figure 4.

Although embodiment of the present invention is disclosed as above, it is not restricted to listed in description and embodiment With, it can be applied to completely various suitable the field of the invention, for those skilled in the art, can be easily Other modification is realized, therefore under the general concept limited without departing substantially from claim and equivalency range, the present invention is not limited In specific details and shown here as.

Claims (8)

1. a kind of automatic analyzing and diagnosing method of the turbine-generator units waterpower imbalance fault based on online data, its feature exists In step includes：

1) by monitoring record data on-line, calculate and obtain idle condition lower swing, the 1X components of vibration signal and blade frequencies Component；

2) different loads are set, by monitoring record data on-line, is calculated and is obtained under different loads operating mode, multiple swings, shaken The 1X components and blade frequencies component of dynamic signal；

3) according to the swing under the zero load and different loads operating mode of step 1 and step 2 acquisition, the 1X components of vibration signal and blade Frequency component, the phasor difference under calculating different loads operating mode, under multiple swings, the 1X components and idle condition of vibration signal Amplitude and multiple blade frequencies components and the vibration vector spread value under idle condition；

4) the phasor difference amplitude drawn by step 3 calculate swing, the 1X components vibration vector spread value of vibration signal with The linearly dependent coefficient and blade frequencies component vibration vector spread value of load and the linearly dependent coefficient of load；

5) correlation coefficient that the phasor difference amplitude and step 4 for being drawn by step 3 is drawn is contrasted respectively with preset value, The factor for causing water conservancy imbalance fault is judged according to comparable situation.

2. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 1 is examined Disconnected method, it is characterised in that select the data under close or identical head.

3. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 1 is examined Disconnected method, it is characterised in that select the last idle condition lower swing, the 1X components of vibration signal and blade frequencies component.

4. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 1 is examined Disconnected method, it is characterised in that calculated by FFT and obtain idle condition lower swing, the 1X components of vibration signal and blade frequencies point Amount.

5. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 1 is examined Disconnected method, it is characterised in that when the load for setting is as nominal load, under swing, the 1X components and idle condition of vibration signal Phasor difference amplitude change amplitude patient more than or equal to minimum and swing, the 1X component vector spread values of vibration signal and machine The linearly dependent coefficient of group load is more than or equal to preset value, then judge because runner body blade profile is inconsistent, leaf road is inconsistent, The waterpower imbalance fault of the inconsistent formation of each length of blade of runner.

6. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 1 is examined Phasor difference amplitude under disconnected method, it is characterised in that when the load for setting is as nominal load, blade frequencies component and idle condition The linear correlation system of change amplitude patient more than or equal to minimum and blade frequencies component vector spread value and unit load Number be more than or equal to preset value, then judge due to runner envelope circularity irregular shape into waterpower imbalance fault.

7. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 1 is examined Disconnected method, it is characterised in that it is P to retrieve load from the data record of on-line monitoring storage, and head is H⁰Data record, lead to FFT is crossed, is calculated and is obtained unloaded lower bearing bracket vibration, the 1X components of throwWith blade frequencies component(n is runner bucket Number), i=i+1 is set, loading condiction is P=i*0.1Pr, and Pr is rated load：I=1,2 ... ... 10, thenForForP be 0.1Pr, 0.2Pr0.3Pr, ... Pr；Meter Calculation drawn under different loads operating mode, phasor difference amplitude under multiple swings, the 1X components and idle condition of vibration signal and multiple Phasor difference amplitude under blade frequencies component and idle condition：

I=1, then 2 ... ... 10, Aⁱ _1XForAⁱ _nXFor

When the load of setting is more than or equal to nominal load, calculates swing, the 1X component vector spread values of vibration signal and bear The linearly dependent coefficient and blade frequencies component vector spread value of lotus and the linearly dependent coefficient of load：

$C_{1X}=\frac{\underset{i=1}{\overset{10}{\Σ}}(X_{1X}^i-\stackrel{\‾}{X})(Y_{1X}^i-\stackrel{\‾}{Y})}{\sqrt{\underset{i=1}{\overset{10}{\Σ}}{(X_{1X}^i-\stackrel{\‾}{X})}^2}\sqrt{\underset{i=1}{\overset{10}{\Σ}}{(Y_{1X}^i-\stackrel{\‾}{Y})}^2}}$

$C_{1X}=\frac{\underset{i=1}{\overset{10}{\Σ}}(X_{1X}^i-\stackrel{\‾}{X})(Y_{1X}^i-\stackrel{\‾}{Y})}{\sqrt{\underset{i=1}{\overset{10}{\Σ}}{(X_{1X}^i-\stackrel{\‾}{X})}^2}\sqrt{\underset{i=1}{\overset{10}{\Σ}}{(Y_{1X}^i-\stackrel{\‾}{Y})}^2}}$

I=1,2 ... ... 10, if

$X_{nX}=\{A_{nX}^1,A_{nX}^2,A_{nX}^3,A_{nX}^4,A_{nX}^5,A_{nX}^6,A_{nX}^7,A_{nX}^8,A_{nX}^9,A_{nX}^{10}\}$

Y_1X={ 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 }

$X_{nX}=\{A_{nX}^1,A_{nX}^2,A_{nX}^3,A_{nX}^4,A_{nX}^5,A_{nX}^6,A_{nX}^7,A_{nX}^8,A_{nX}^9,A_{nX}^{10}\}$

Y_nX={ 0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0 }；

Wherein,Change lower limit for minimum patient throw, the 1X amplitudes under full-load conditions of frame vibration, generally Select 0.4 times to 0.7 times related GB operational shock amplitude；

C_{l_1X}For the correlation coefficient of 1X components, more than 0.7 value is generally selected；

For minimum patient throw, the nX amplitudes under full-load conditions of frame vibration change lower limit, value needs Determined according to the test case of actual set；

C_{l_nX}For the correlation coefficient of nX components, more than 0.7 value is generally selected.

8. the automatically analyzing for turbine-generator units waterpower imbalance fault based on online data according to claim 7 is examined Disconnected method, it is characterised in that step 5) include：

IfAnd C_1X≥C_{l_1X}, then exist because runner body blade profile is inconsistent, leaf road is inconsistent, runner The waterpower imbalance fault of the inconsistent formation of each length of blade；

IfAnd C_nX≥C_{l_nX}, then exist due to runner envelope circularity irregular shape into waterpower it is uneven therefore Barrier.