CN105373701A - Electromechanical equipment association degree determination method - Google Patents

Abstract

The invention relates to an electromechanical equipment association degree determination method. The determination method comprises the following steps: step 1, determining a reference sequence and a comparison sequence, wherein the reference sequence is statistical data or an optimal single target optimized value in different moments, and the comparison sequence is a sequence compared with the reference sequence; step 2, performing dimension removal on the reference sequence and the comparison sequence; step 3, determining correlation coefficients between the dimensionless reference sequence and the dimensionless comparison sequence; and step 4, calculating and obtaining the association degree according to the at least one correlation coefficient obtained in the step 3. According to the electromechanical equipment association degree determination method, the problem of low electromechanical equipment fault diagnosis accuracy is solved.

Description

A kind of electromechanical equipment degree of association defining method

Technical field

The present invention relates to field of electromechanical technology, particularly relate to a kind of field of diagnosis about equipment fault.

Background technology

Fault diagnosis is exactly the Given information determination unknown message utilizing electromechanical equipment, carries out comprehensive evaluation and analyze judging to the state of electromechanical equipment.Existing comprehensive evaluation analysis pattern is more, but reliability and accuracy are all not good enough.

Method provided by the invention, appropriate characteristic parameter is extracted from status signal, formation can reflect the reference mode of the normal and various malfunction feature of electromechanical equipment, then the degree of association between pattern to be checked (treating diagnostic state) and each reference mode is calculated, according to most relevance degree principle, judge that state to be checked is maximum with associating of which kind of normal condition, carry out state recognition, first guided maintenance personnel check the fault that most probable occurs.Solve in prior art, the problem of fault diagnosis reliability and poor accuracy.

Summary of the invention

In view of above-mentioned analysis, the present invention aims to provide a kind of electromechanical equipment degree of association defining method, in order to solve the not high problem of existing electromechanical equipment fault diagnosis accuracy.

Object of the present invention is mainly achieved through the following technical solutions:

A kind of electromechanical equipment degree of association defining method, it is characterized in that, said method comprising the steps of: the first step, determine reference sequences and comparative sequences, wherein reference sequences is not statistics in the same time or optimum single object optimization value, and comparative sequences is the sequence compared with reference sequences; Second step, goes dimension process to reference sequences and comparative sequences; 3rd step, determine nondimensional reference sequences and nondimensional comparative sequences correlation coefficient:

${\mathrm{\ξ}}_{ij}\left(k\right)=\frac{\underset{i}{\min }\underset{k}{\min }|Y_j\left(k\right)-X_i\left(k\right)|+a*\underset{i}{\max }\underset{k}{\max }|Y_j\left(k\right)-X_i\left(k\right)|}{|Y_j\left(k\right)-X_i\left(k\right)|+a*\underset{i}{\max }\underset{k}{\max }|Y_j\left(k\right)-X_i\left(k\right)|}$

Wherein, i=0,1,2 ..., m; J=1,2,3 ..., n; K=1,2,3 ..., p, a jth dimensionless comparative sequences Y _j(k) and i-th dimensionless reference sequences X _i(k) in a kth characteristic parameter place minimum value, a jth dimensionless comparative sequences Y _j(k) and i-th dimensionless reference sequences X _ik (), in a kth characteristic parameter place maximal value, a is resolution ratio, 1>=a>=0; 4th step, calculates the degree of association according at least one correlation coefficient that the 3rd step obtains.

Optionally, the degree of association of described 4th step is the mean value of multiple correlation coefficient.

Optionally, the degree of association of described 4th step is the weighted mean value of multiple correlation coefficient.

Optionally, the degree of association of described 4th step is:

$r_{Bi}=\frac{1}{1+\frac{1}{n}{d}_{oi}^0+\frac{1}{n-1}{d}_{oi}^1+\frac{1}{n-2}{d}_{oi}^2}$

In formula,

$d_{oi}^0=\underset{k=1}{\overset{n}{\Σ}}|X_i\left(k\right)-Y_j\left(k\right)|;d_{oi}^1=\underset{k=1}{\overset{n}{\Σ}}\left|\right(X_i(k+1)-Y_j(k+1))-(X_i\left(k\right)-Y_j\left(k\right)\left)\right|;$

n is n group dimensionless reference sequences and dimensionless comparative sequences, and X is dimensionless reference sequences, and Y is dimensionless comparative sequences.

Optionally, reference sequences is multiple, and comparative sequences is multiple, then the degree of association of the 4th step is a degree of association matrix.

Optionally, when the degree of association is greater than preset value, the fault type of electromechanical equipment can be determined.

Beneficial effect of the present invention is as follows: the reliability and the accuracy that improve electromechanical equipment fault diagnosis.

Other features and advantages of the present invention will be set forth in the following description, and, becoming apparent from instructions of part, or understand by implementing the present invention.Object of the present invention and other advantages realize by structure specifically noted in write instructions, claims and accompanying drawing and obtain.

Accompanying drawing explanation

Accompanying drawing only for illustrating the object of specific embodiment, and does not think limitation of the present invention, and in whole accompanying drawing, identical reference symbol represents identical parts.

The electromechanical equipment degree of association defining method schematic diagram that Fig. 1 provides for the specific embodiment of the invention.

Embodiment

Specifically describe the preferred embodiments of the present invention below in conjunction with accompanying drawing, wherein, accompanying drawing forms the application's part, and together with embodiments of the present invention for explaining principle of the present invention.

The object of grey correlation analysis is to find a kind of quantization method can weighing degree of association size between each factor, to find out the key factor of influential system developing state, thus grasps the principal character of things.The quantitative description of System Development changing trend and comparative approach are the Fundamentals of Mathematics according to Space Theory, determine reference sequence (female ordered series of numbers) and somely compare correlation coefficient between ordered series of numbers (subnumber row) and the degree of association.

The step of grey correlation analysis: the first step, determines reference sequences and comparative sequences; Second step, nondimensionalization process; 3rd step, determines correlation coefficient; 4th step, determines the degree of association.

One, reference sequences and comparative sequences is determined

First association analysis will determine reference sequences and comparative sequences, but before determining reference sequences and comparative sequences, first provides the concept of grey correlation factor set.

X is made to be sequence sets,

If for x _i(k) and x _j(k), the same order of magnitude, and dimensionless, then claim X to be the comparable sequence sets of quantity; M is the maximal value of m as shown in Equation 1, and K is the maximal value of n, n be more than or equal to arbitrarily 2 integer, ijk represents i-th, jth, a kth x, is positive integer.

If there is not parallel sequence in X, then X is claimed to be that numerical value can approximating sequence collection, if X has following character: numerical value accessibility; Quantity comparability; The non-negative factor; Then X is claimed to be grey correlation factor set, or grey correlation sequence collection, claim the sequence in X to be the factor.

Provide the definition of reference sequences and comparative sequences below: so-called " reference sequences ", is often designated as X _0j, it is made up of not statistics in the same time, or optimum single object optimization value.Reference sequences X _0jcan be expressed as:

$X_{0j}=\left\{\begin{array}{c}{x}_{0j}|x_{0j}=(x_{0j}\left(1\right),x_{0j}\left(2\right),...,x_{0j}\left(n\right)),j=1,2,...J\\ x_{0j}\left(k\right)\∈x_{0j},k\∈K,K=\{1,2,...,n\},n\≥2\end{array}\right\}---\left(2\right)$

Wherein, k represents not in the same time.Concerning multiple-objection optimization, k is target indicator.

Do with reference sequences in association analysis " subnumber arranges " that correlation degree compares, be referred to as " comparative sequences ", be also called " system correlative factor behavior sequence ", be designated as X.Comparative sequences can be expressed as:

$X=\left\{\begin{array}{c}{x}_i|i\∈M,M=\{1,2,...,m\},m\≥1,\\ x_i=(x_i\left(1\right),x_i\left(2\right),...,x_i\left(n\right)),\\ x_i\left(k\right)\∈x_i,k\∈K,K=\{1,2,...,n\},n\≥2\end{array}\right\}---\left(3\right)$

Two, nondimensionalization process

Because the physical significance of factor each in system is different, or measurement unit is different, thus causes the dimension of data different, and the order of magnitude of numerical value sometimes differs greatly.Between so different dimension, varying number level, inconvenience is compared, or is difficult to obtain correct result relatively time.For the ease of analyzing, simultaneously for ensureing that data have equivalence and same sequence, just need the data processing of raw data being carried out to nondimensionalization before each factor compares.Make it dimension data processing and usually have following several mode:

Accumulating generation, is designated as AGO;

Inverse accumulated generating, is designated as IAGO;

First value, is designated as INGO;

Equalization, is designated as MGO;

Interval value, is designated as QGO;

Measurementization;

Modelling.

From functional connotation, these methods can be divided three classes: i.e. hierarchical transformation type, numerical transformation type, reversal type.The method belonging to hierarchical transformation type has Accumulating generation and inverse accumulated generating.Regressive is cumulative inverse generation.Ordered series of numbers one by one adds up by so-called Accumulating generation AGO exactly, and this kind of conversion is hierarchical, and the object changing level is to find rule.Accumulating generation has the effect disclosing potential rule.

Scale data processing is referred to as in equalization, first value, interval value etc., belongs to numerical transformation.Its function be by those because dimension, the order of magnitude are different the object without comparability, after quantitative transformations, make it become dimensionless and there is same order, thus " not comparable " is converted into " comparable ".Quantitative transformations is usually used in grey correlation analysis and GM (1, N) modeling.

Estimating, refers to the measure of merit of the upper limit, lower limit, intermediate value, and its function is the polarity changing data, and the sample making target inconsistent is unified in polarity, is convenient to compare and computing.Estimating data variation is mainly used in grey situation.Modelling, refers to the data variation obtained by certain model.

In grey correlation analysis, conventional ash generation method mainly scale data processing method, especially equalization, first value data processing method.

First value process

All data of an ordered series of numbers are all removed by its first number, thus the method obtaining a new ordered series of numbers is called just value process.This new ordered series of numbers to show in original data series that value is not in the same time relative to the multiple of first moment value.This ordered series of numbers has common starting point, dimensionless.

Make the formation sequence that x ' is x,

x＝(x(1),x(2),…,x(n))

x′＝(x′(1),x′(2),…,x′(n))

If meet

x′(k)＝x(k)/x′(1)

x(k)∈x,x(1)∈x

The initial value process sequence then claiming x ' to be x.

Note just value is treated to INGO, then

INGO:x→x′

INGO:x(k)→x′(k)

x′(k)＝x(k)/x(1)。

Equalization process

All data of an ordered series of numbers are all removed with its mean value, thus the method obtaining a new ordered series of numbers is called equalization process.This new ordered series of numbers to show in original data series that value is not in the same time relative to the multiple of mean value.

Make the formation sequence that x ' is x,

x＝(x(1),x(2),…,x(n))

x ^m＝(x ^m(1),x ^m(2),…,x ^m(n))

If meet

$x^m=x\left(k\right)/\stackrel{\‾}{x}$

$\stackrel{\‾}{x}=\frac{1}{n}\underset{k=1}{\overset{n}{\Σ}}x\left(k\right)$

Then claim x ^mfor the average value processing sequence of x.

Note just value is treated to MGO, then

MGO:x→x ^m

MGO:x(k)→x ^m(k)

$x^m=x\left(k\right)/\stackrel{\‾}{x}$

$\stackrel{\‾}{x}=\frac{1}{n}\underset{k=1}{\overset{n}{\Σ}}x\left(k\right)$

Three, correlation coefficient is determined:

Correlation degree between system or between factor judges that whether its contact is tight according to the similarity degree of geometric configuration between curve, and therefore, the size of difference between curve, can as the yardstick of correlation degree.

X is made to be association factor collection

$X=\left\{\begin{array}{c}{x}_i|i\∈M,M=\left\{1,2,\mathrm{...},m\right\},m\≥2,\\ x_i=\left(x_i\left(1\right),x_i\left(2\right),\mathrm{...},x_i\left(n\right)\right),\\ x_i\left(k\right)\∈x_i,k\∈K,K=\left\{1,2,\mathrm{...},n\right\},n\≥2\end{array}\right\}---\left(4\right)$

If reference vector integrates as X ₁(p), X ₂(p) ..., X _mp (), vector set to be checked is Y ₁(p), Y ₂(p) ..., Y _n(p). wherein benchmark mould adds up to m, and mould to be checked adds up to n, and each benchmark mould comprises p characteristic parameter.Then | Y _j(k)-X _i(k) | for jth mould to be checked and i-th benchmark mould are in the relative quantity of a kth characteristic parameter place difference, Y _j(k) and X _ik () is respectively Y _jwith X _iin the data of kth point.If there is nonnegative real number ξ _ij(k) on X under certain environment Y _j(k) and X _ithe comparison measure of (k), | Y _j(k)-X _i(k) | less, ξ _ijwhen () is larger k, claim ξ _ijk () is X _ik () is to Y _jk () is at the correlation coefficient of k point.Correlation coefficient is:

${\mathrm{\ξ}}_{ij}\left(k\right)=\frac{\underset{i}{\min }\underset{k}{\min }|Y_j\left(k\right)-X_i\left(k\right)|+a*\underset{i}{\max }\underset{k}{\max }|Y_j\left(k\right)-X_i\left(k\right)|}{|Y_j\left(k\right)-X_i\left(k\right)|+a*\underset{i}{\max }\underset{k}{\max }|Y_j\left(k\right)-X_i\left(k\right)|}---\left(5\right)$

Wherein i=0,1,2 ..., m; J=1,2,3 ..., n; K=1,2,3 ..., p, be jth mould to be checked and m benchmark mould at p characteristic parameter place, in like manner it is maximal value.

Definition: Δ ij=︱ Y _j(k)-X _i(k) ︱

As Δ ij=Δ min, the previous value of correlation coefficient is ξ _ij(k)=1

As Δ ij=Δ max, the next time value of correlation coefficient is:

${\mathrm{\ξ}}_{ij}\left(k\right)=\frac{{\mathrm{\Δ}}_{\min }+{\mathrm{a\Δ}}_{\max }}{{\mathrm{\Δ}}_{max}+{\mathrm{a\Δ}}_{\max }}=\frac{1}{1+a}\·\left(a+\frac{{\mathrm{\Δ}}_{\min }}{{\mathrm{\Δ}}_{max}}\right)---\left(6\right)$

In A type degree of association formula, α is resolution ratio, and the resolution of the degree of association is relevant with the value size of resolution ratio, and being multiplied by α is to reduce the impact of extreme value on result of calculation, thus improves resolution.

If α=1, then the span of correlation coefficient is: 0.5<=ξ _ijk () <=1, span is less, and resolution is lower.

If α=0.1, then the span of correlation coefficient is: 0.09<=ξ _ijk () <=1, span is comparatively large, and resolution is higher.

It can thus be appreciated that:

Resolution ratio α can regulate size and the constant interval of correlation coefficient, ξ _ijk the next time value of () increases with the increase of α, but be less than 1 all the time.

By minimum information principle, when α=0.5436 than being easier to the change observing degree of association resolution, therefore, generally get α=0.5.

Four, the degree of association is determined:

(1) the A type degree of association

Between two systems or two factors, the tolerance of relevance size, is called the degree of association.The degree of association describes in systems development process, and the situation of change relatively between factor, namely changes the relativity of size, direction and speed etc.If both are in evolution, change is basically identical relatively, then think that both degrees of association are large, otherwise both degrees of association are little.The essence of association analysis, carries out the comparison of geometric relationship exactly to ordered series of numbers curve.If two ordered series of numbers curve co-insides, then relevance is good, and namely correlation coefficient is 1, and so the two ordered series of numbers degrees of association also equal 1.Meanwhile, two ordered series of numbers curves can not be vertical, i.e. onrelevant, so correlation coefficient is greater than 0, therefore the degree of association is also greater than 0.Because correlation coefficient is a tolerance of curve geometry correlation degree, in relatively overall process, more than one of correlation coefficient.Therefore the mean value getting correlation coefficient is as the tolerance of correlation degree comparing overall process.

$r_{Aij}=\frac{1}{p}\underset{k=1}{\overset{p}{\&Sigma;}}{\mathrm{\&xi;}}_{ij}\left(k\right)---\left(7\right)$

Claim r _aijfor Y _jk () is to X _ithe grey relational grade of (k).

When spending by above formula compute associations, average power process being done to each index or space, is considered as of equal importance by each index or space.But in practice, but there is the situation of much inequality power, namely think that some index is even more important.Therefore, according to actual conditions, grey relational grade can be asked for grey incidence coefficient as weighted mean.If the importance in each index or space is differentiated, corresponding weights λ (k) should be given by importance size, k={1,2 ..., p} and

According to Shannon (shannon) information theory, if variable x value x immediately _i(i=1,2 ..., n, n are finite value) probability be z _i(z _i> 0), when condition under, the entropy of stochastic variable x is defined as:

$H\left(x\right)=E\&lsqb;log\frac{1}{z_i}\&rsqb;=-C\underset{i=1}{\overset{p}{\&Sigma;}}{z}_i\log z_i---\left(8\right)$

Wherein specify 0log (0)=0

Utilize above-mentioned theory, grey incidence coefficient weight can be expressed as:

$\mathrm{\&lambda;}\left(k\right)=\frac{1-s\left(k\right)}{\underset{k=1}{\overset{p}{\&Sigma;}}(1-s\left(k\right))}---\left(9\right)$

In formula $s\left(k\right)=-C\underset{i=1}{\overset{m}{\&Sigma;}}{r}_{Ai}\left(k\right)log\left(r_{Ai}\left(k\right)\right)$

$r_{Ai}\left(k\right)=\frac{x_i\left(k\right)}{\underset{i=1}{\overset{m}{\&Sigma;}}{x}_i\left(k\right)}$

$C=\frac{1}{\log m}$

Last weighted association degree is:

$r_{Aij}=\underset{k=1}{\overset{p}{\&Sigma;}}\mathrm{\&lambda;}\left(k\right){\mathrm{\&xi;}}_{ij}\left(k\right)---\left(10\right)$

(2) the Type B degree of association

Except the Deng Shi degree of association, more typical and conventional degree of association quantitative model also has Type B grey relational grade at present, this degree of association, although go back defectiveness (as produced ordinal number effect) in theory, but do not affect them in social economy and the application in producing, the solution particularly for some particular problems receives good effect.

$r_{Bi}=\frac{1}{1+\frac{1}{n}{d}_{oi}^0+\frac{1}{n-1}{d}_{oi}^1+\frac{1}{n-2}{d}_{oi}^2}---\left(11\right)$

In formula,

$d_{oi}^0=\underset{k=1}{\overset{n}{\&Sigma;}}|X_i\left(k\right)-Y_j\left(k\right)|$

$d_{oi}^1=\underset{k=1}{\overset{n-1}{\&Sigma;}}\left|\right(X_i(k+1)-Y_j(k+1))-(X_i\left(k\right)-Y_j\left(k\right)\left)\right|$

Wherein represent that overall displacements is poor, description be " proximity " between curve; for single order difference coefficient, represent the single order slope differences (velocity contrast) of curve; for second order difference coefficient, represent the second order slope differences (acceleration is poor) of curve; what characterize is " similarity " between curve.

So the Type B degree of association can reflect things development comprehensively " proximity " with " similarity ".So keep putting in order of each characteristic parameter of benchmark mould and mould to be checked consistent in the fault diagnosis of internal combustion engine, make diagnostic model sequential, just can well the state of combustion motor make disconnected.

(3) incidence matrix

If more than one of reference sequences, and comparative sequences is more than one, then each comparative sequences forms grey correlation matrix to the grey relational grade of each reference sequences.

If: n reference sequences: x ₁, x ₂, x ₃x _nn ≠ 1

M sequence to be checked: x ₁, x ₂, x ₃x _nn ≠ 1

Then the degree of association of each sequence pair reference sequences to be checked is respectively:

(r ₁₁,r ₁₂,……r _1m)

(r ₂₁,r ₂₂,……r _2m)

.

.

.

(r _n1,r _n2,……r _n1m)

If by r _ij(i=1,2 ..., n, j=1,2 ..., m) suitably arrange, just obtain grey relational grade matrix:

or

One of embodiment: diesel engine failure diagnosis grey correlation diagnostic techniques research.

First, the choosing of diesel engine Diagnostic parameters:

Diesel engine supervision can monitoring parameter many, the thermal parameter that this research is mainly chosen in table 1 is diagnosed for grey correlation.

Grey correlation diagnostic characteristic parameter

These parameters can be divided into two large classes:

Performance parameter comprises rotating speed supercharger speed, each cylinder delivery temperature, rate of fuel consumption, power, pressure parameter etc.Also have lubricating oil to enter out-of-machine temperature in addition, fresh water enters out-of-machine temperature, charge air cooler inflow temperature, temperature after charge air cooler, temperature before and after pneumatic plant, turbine front exhaust temperature (delivery temperature of each group can be measured for exhaust pulse pressure-charging), turbine rear exhaust temperature, air pressure before and after pneumatic plant, air pressure (the exhaust air pressure of each group can be measured for exhaust pulse pressure-charging) before turbine, atmospheric pressure, crankcase pressure, working connection oil pressure, fuel oil and lubricating oil pressure etc.

Parameter can be surveyed and comprise maximum explosive pressure, mean indicated pressure (MIP), compression pressure, detonation pressure correspondence position, the swelling pressure etc.

Generally speaking, these parameters can be divided into fuel oil supply system, air inlet system and exhaust system, lubricating and cooling system parameter etc.This research has taken into full account the working order of intelligent diesel while monitoring conventional thermodynamic parameter, monitors the intelligent control system parameter of diesel engine.Below in conjunction with the characteristic parameter selected by diesel engine failure diagnosis system, the thermal parameter in summary analysis 4-8, because these parameters comprehensively can reflect the various faults of diesel engine.

Useful power: the Effective power done in the diesel engine unit interval, conventional maximum-continuous rating (MCR) evaluates the power performance of boat diesel engine.

Mean effective pressure: mean effective pressure can be regarded as imaginary, an on average constant pressure and acts on piston top, makes piston move a stroke institute work and equals the done Effective power that often circulates.Mean effective pressure weighs the important index of of diesel powered performance.

Effective specific fuel consumption: effective specific fuel consumption is the fuel consumption of unit Effective power, it is the important indicator weighing engine fuel economy energy.

Explosion pressure: explosion pressure is the maximum combustion pressure in cylinder, can be tried to achieve by load-position diagram.Explosion pressure is the important indicator weighing diesel powered performance, and meanwhile, it also reflects the working environment of parts and components of diesel engine, can as the foundation judging the mechanical load that parts and components of diesel engine bears.Explosion pressure mainly by the impact of the factors such as diesel load, ratio of compression and injection timing, can reflect the fault such as diesel fuel injection system and valve timing deterioration.

Detonation pressure position: the value of detonation pressure position and crank angle corresponding to explosion pressure, the position of its general After Top Center 5-15 crank angle degrees.Explosion pressure means that explosion pressure is very large in advance usually, and roughness noise is delayed, then mean that burning is soft, but may with afterburning.

Compression pressure: compression pressure is taken as in-cylinder pressure corresponding when diesel engine pressure stroke piston runs to top dead centre, and it is an important performance characteristic of diesel engine.The difference of compression pressure and actual adiabatic compression pressure is less than 3%, if deviation is excessive, except admission pressure reduces, also may be caused by following fault: vent valve gas leakage, piston-cylinder excessive wear, injection lag etc.

The swelling pressure: the swelling pressure are taken as in-cylinder pressure when diesel engine pressure stroke piston runs to after top dead center 40 crank angle degrees.The size of the swelling pressure is relevant with the combustion process of diesel engine, and after-burning, gas leakage etc. all can cause the change of the swelling pressure.

Maximum pressure raises ratio: maximum pressure raises than trying to achieve by then getting maximal value to load-position diagram differentiate, and it weighs the important index of of the rough degree of diesel combustion.

In cold forward and backward intake air temperature: in cold forward and backward intake air temperature can the work efficiency of concentrated expression charge air cooler, diesel load situation, the boost conditions of supercharger and the cooling power of ICS intercooler system.

The forward and backward delivery temperature of turbine: the forward and backward delivery temperature of turbine can the working condition of concentrated expression supercharger, the work efficiency of supercharger, pneumatic plant, the circulating flushing tolerance of diesel engine, each cylinder delivery temperature etc.They and the thermodynamic performance such as boost pressure, fuel consumption are comprehensively analyzed, and tentatively can determine the performance of diesel engine.

Inlet and outlet manifold pressure: inlet and outlet manifold pressure is by the impact of multiple factors, but they can reflect the operational situation of diesel engine comprehensively, are the key character parameters judging whether diesel engine breaks down.

As can be seen here, the thermal parameter of diesel engine contains the information of gas in-cylinder process quality, in addition, also contains the information such as port timing, injection timing.Therefore can be posed etc. by the work of thermal parameter to diesel engine and make an appraisal, can judge that whether diesel engine is working properly by the difference of characteristic parameter, and under being operated in which type of fault mode.

The present invention, under the prerequisite not destroying 7RTF-60C diesel engine, simulates the 7 class diesel engine most common failures such as single cylinder catches fire, vent valve late release, air cooler efficiency fault, injection advance, efficiency of turbocharger fault, ratio of compression reduces, distributive value increases.When the arranging of fault, in conjunction with the model of fault simulation, for the single cylinder parameter needing to change only with the 6th cylinder for benchmark.Measuring the characteristic parameter under various state when keeping various external condition constant respectively, setting up the vector set to be checked of diagnostic system, and writing down sequence number corresponding to all kinds of fault.Can find in conjunction with actual diesel engine fault simulating test order, diagnostic result and virtual condition fit like a glove.

In sum, embodiments provide a kind of electromechanical equipment degree of association defining method, the method can realize the Accurate Diagnosis of electromechanical equipment fault.

It will be understood by those skilled in the art that all or part of flow process realizing above-described embodiment method, the hardware that can carry out instruction relevant by computer program has come, and described program can be stored in computer-readable recording medium.Wherein, described computer-readable recording medium is disk, CD, read-only store-memory body or random store-memory body etc.

The above; be only the present invention's preferably embodiment, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; the change that can expect easily or replacement, all should be encompassed within protection scope of the present invention.

Claims (6)

1. an electromechanical equipment degree of association defining method, it is characterized in that, said method comprising the steps of: the first step, determine reference sequences and comparative sequences, wherein reference sequences is not statistics in the same time or optimum single object optimization value, and comparative sequences is the sequence compared with reference sequences; Second step, goes dimension process to reference sequences and comparative sequences; 3rd step, determine nondimensional reference sequences and nondimensional comparative sequences correlation coefficient:

${\mathrm{\&xi;}}_{ij}\left(k\right)=\frac{\underset{i}{\min }\underset{k}{\min }|Y_j\left(k\right)-X_i\left(k\right)|+a*\underset{i}{\max }\underset{k}{\max }|Y_j\left(k\right)-X_i\left(k\right)|}{|Y_j\left(k\right)-X_i\left(k\right)|+a*\underset{i}{\max }\underset{k}{\max }|Y_j\left(k\right)-X_i\left(k\right)|}$

Wherein, i=0,1,2 ..., m; J=1,2,3 ..., n; K=1,2,3 ..., p, a jth dimensionless comparative sequences Y _j(k) and i-th dimensionless reference sequences X _i(k) in a kth characteristic parameter place minimum value, a jth dimensionless comparative sequences Y _j(k) and i-th dimensionless reference sequences X _ik (), in a kth characteristic parameter place maximal value, a is resolution ratio, 1>=a>=0; 4th step, calculates the degree of association according at least one correlation coefficient that the 3rd step obtains.

2. electromechanical equipment degree of association defining method according to claim 1, is characterized in that, the degree of association of described 4th step is the mean value of multiple correlation coefficient.

3. electromechanical equipment degree of association defining method according to claim 1, is characterized in that, the degree of association of described 4th step is the weighted mean value of multiple correlation coefficient.

4. electromechanical equipment degree of association defining method according to claim 1, is characterized in that, the degree of association of described 4th step is:

$r_{Bi}=\frac{1}{1+\frac{1}{n}{d}_{oi}^0+\frac{1}{n-1}{d}_{oi}^1+\frac{1}{n-2}{d}_{oi}^2}$

In formula,

$d_{oi}^0=\underset{k=1}{\overset{n}{\&Sigma;}}|X_i\left(k\right)-Y_j\left(k\right)|;d_{oi}^1=\underset{k=1}{\overset{n-1}{\&Sigma;}}|\left(X_i\right(k+1)-Y_j(k+1\left)\right)-\left(X_i\right(k)-Y_i(k\left)\right)|;$

n is n group dimensionless reference sequences and dimensionless comparative sequences, and X is dimensionless reference sequences, and Y is dimensionless comparative sequences.

5. electromechanical equipment degree of association defining method according to claim 1, is characterized in that, reference sequences is multiple, and comparative sequences is multiple, then the degree of association of the 4th step is a degree of association matrix.

6. electromechanical equipment degree of association defining method according to claim 1, is characterized in that, when the degree of association is greater than preset value, can determine the fault type of electromechanical equipment.