CN103995530A - Fault detection method based on intelligent power distribution environment - Google Patents

Abstract

The invention discloses a fault detection method based on an intelligent power distribution environment. The problem of singularity-free cloud terminal network fault detection of a singularity-free cloud terminal network structure time-delay control system is researched. Considering the state of interval variable predication on an intelligent power grid, a cloud model network structure time-delay singularity-free cloud terminal network fault predication filter is designed. A perception recognizing method that a non-deterministic non-linear structure time-delay feedback filter structure comprises a cloud theory model used for predicating the relevant information of the intelligent power grid is used. Based on the singularity-free cloud terminal network structure time-delay control system filter method, dependence sufficient conditions existing in the singularity-free cloud terminal network fault detection filter are executed, and an iterative method for designing the system weight of the filter is provided.

Description

A kind of failure prediction method based under intelligent power distribution environment

Technical field

The invention belongs to nonlinear network structure time lag control technology field, relate to a kind of failure prediction method based under intelligent power distribution environment, specifically, relate to a kind of failure prediction method based in uncertain non-linear nonsingular cloud terminal network Control System with Delay under intelligent power distribution environment.

Background technology

When being applied in novel nonsingular cloud terminal network structure Control System with Delay performance for the develop rapidly of nonlinear network structure time lag control technology, also improve the complexity of nonsingular cloud terminal network structure Control System with Delay.For optimizing reliability and stability, science, utilize nonsingular cloud terminal network failure prediction to become the system of nonsingular cloud terminal network structure Control System with Delay indispensability.In addition, the network structure Control System with Delay that numerous scholars and scientific research personnel pay close attention to, few owing to having line, information resources can be shared, and are easy to the advantages such as maintenance and expansion, have become the development trend of nonsingular cloud terminal network structure Control System with Delay.The neural network forecast intelligent grid problem self being short of for network structure Control System with Delay, must be for the feature of network structure Control System with Delay, design meets the nonsingular cloud terminal network failure prediction system of network structure Control System with Delay demand.In prior art, have for the system with state measurement delay and put forward structure Delay-Dependent adequate condition of separating out of joint.In prior art for the nonsingular cloud terminal network failure prediction problem of cloud model network structure lagging network structure Control System with Delay, designed a kind of random measurement of describing and postponed and the measurement model of packet loss problem, and labor implicit Conservative Property in this model.Someone,, for the nonlinear network structure Control System with Delay that has Random Communication delay and packet loss, has designed the uncertain nonlinear organization time lag of cloud model network structure time lag cloud theoretical model wave filter.

Summary of the invention

In order to overcome the defect existing in prior art, the invention provides a kind of failure prediction method based under intelligent power distribution environment, study the nonsingular cloud terminal network failure prediction problem of nonsingular cloud terminal network structure Control System with Delay.Consider to exist the interval state that becomes prediction intelligent grid, the nonsingular cloud terminal network of design cloud model network structure time lag failure prediction wave filter.The cloud theoretical model apperceive identity method of utilizing uncertain nonlinear organization Systems with Time Delay Feedback filter construction to comprise prediction intelligent grid relevant information.Carried out based on nonsingular cloud terminal network structure Control System with Delay filtered method the prediction intelligent grid that nonsingular cloud terminal network failure prediction wave filter exists and relied on adequate condition, and to have provided design of filter be the alternative manner of flexible strategy.Its technical scheme is as follows:

Based on the failure prediction method under intelligent power distribution environment, comprise the following steps:

Based on nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node comprise n nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node, each nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node has n dimension attribute and property value is the prediction of the nonsingular cloud terminal network of numerical value structure Delay-Fault, is designated as ${\mathrm{\χ}(\mathrm{\μ}+1)}_{\mathrm{\ψ}}^{\mathrm{\μT}}=\{\mathrm{\Σ}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}(i+1){\mathrm{\χ}}_1\left(\mathrm{\μ}\right),\mathrm{\Σ}{\mathrm{\χ}}_2\left(\mathrm{\μ}\right),...,\mathrm{\Σ}{\mathrm{\χ}}_{n-i}\left(\mathrm{\μ}\right)\},$ i＝1,2,...,n。The attribute vector of nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node is μ=1,2 ..., n, cloud theoretical model algorithm is gathered is ${\mathrm{\χ}(\mathrm{\μ}+1)}_{\mathrm{\ψ}}^{\mathrm{\μT}}=\{\mathrm{\Σ}{\mathrm{\χ}}_1\left(\mathrm{\μ}\right),\mathrm{\Σ}{\mathrm{\χ}}_2\left(\mathrm{\μ}\right),...,\mathrm{\Σ}{\mathrm{\χ}}_n\left(\mathrm{\μ}\right)\}$ The process of class is:

Step 1: the nonsingular cloud terminal network structure Control System with Delay networking failure prediction transform method based on cloud theoretical model of describing according to system is to former nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node the networking of structure Control System with Delay, obtains the nonsingular cloud terminal network structural network fault perceptual signal prediction set of node forming based on nonsingular cloud terminal network structure Control System with Delay networking failure prediction wherein i=1,2 ..., n.Cloud theoretical model function Y _fselect traditional Y _qmean algorithm realizes.

Step 2: build the responsive similarity measurement of density based on hamming distance, thereby obtain the similar matrix χ (μ-(μ) of nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node Y ^μ) ∈ R ^m, wherein

$W_{\mathrm{ij}}\left\{\begin{array}{c}{\mathrm{\χ}}_F(\mathrm{\μ}+1)=\underset{\mathrm{\μ}-d\left(\mathrm{\μ}\right)}{\overset{\mathrm{\μ}-1}{\mathrm{\Σ}}}\left\{X_F{\mathrm{\χ}}_F\right[\mathrm{\μ}-d\left(\mathrm{\μ}\right)]+Y_F[\mathrm{\μ}-y\left(\mathrm{\μ}\right)\left]\right\}\\ \mathrm{\γ}[\mathrm{\μ}-\left(\mathrm{\μ}\right)]=\underset{\mathrm{\μ}-d\left(\mathrm{\μ}\right)}{\overset{\mathrm{\μ}-1}{\mathrm{\Σ}}}{Z}_F{\mathrm{\χ}}_F[\mathrm{\μ}-d\left(\mathrm{\μ}\right)]\\ {\mathrm{\χ}}_F\left(0\right)=0\\ Q_F\left(0\right)=0\end{array}\right..$

Step 3: the cloud theoretical model matrix description of constructing nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node Y is γ _e[μ-d (μ)]={ [μ-γ (μ)]-[μ-y (μ)] } ^{μ T}, wherein χ _f(μ-d (μ)) ∈ R ⁿfor vector matrix system is described as

$\left\{\begin{array}{c}z(\mathrm{\μ}+1)=\stackrel{\‾}{X}[\mathrm{\μ}-z\left(\mathrm{\μ}\right)]+{\stackrel{\‾}{X}}_{1}z\left[(\mathrm{\μ}-d\left(\mathrm{\μ}\right))\right]+\stackrel{\‾}{Y}[\mathrm{\μ}-\mathrm{\ω}\left(\mathrm{\μ}\right)]\\ {\mathrm{\γ}}_e\left(\mathrm{\μ}\right)=\stackrel{\‾}{Z}z\left(\mathrm{\μ}\right)+{\stackrel{\‾}{Z}}_{1}z\left[(\mathrm{\μ}-d\left(\mathrm{\μ}\right))\right]+\stackrel{\‾}{Q}\mathrm{\ω}\left(\mathrm{\μ}\right)\end{array}\right..$

Step 4: ask nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node matrix equation $z\left(\mathrm{\μ}\right)={\left[\begin{array}{cc}{\mathrm{\χ}}^{\mathrm{\μT}}\left(\mathrm{\μ}\right)& {\mathrm{\χ}}_F^{\mathrm{\μT}}\left(\mathrm{\μ}\right)\end{array}\right]}^{\mathrm{\μT}}$ Front ω (μ)=[x ^{μ T}(μ) ζ ^{μ T}(μ) f ^{μ T}(μ)] ^{μ T}individual eigenvalue of maximum characteristic of correspondence vector expression is described as

$\{\stackrel{\‾}{X}={\left[\begin{array}{cc}{X}_{\mathrm{\μ}}& 0\\ Y_{F}Z& X_F\end{array}\right]}^{\mathrm{\μT}};{\stackrel{\‾}{X}}_1={\left[\begin{array}{cc}\mathrm{Y\μ}& 0\\ Y_F\mathrm{Q\μ}& 0\end{array}\right]}^{\mathrm{\μT}};\stackrel{\‾}{Y}={\left\{\begin{array}{ccc}-\mathrm{Y\μ}& Y_d& Y_f\\ -Q{Y}_f\mathrm{\μ}& Y_f{Q}_q& Y_f{Q}_f\end{array}\right\}}^{\mathrm{\μT}}$

$\{\stackrel{\‾}{Z}={\left[\begin{array}{cc}-X& Z_F\end{array}\right]}^{\mathrm{\μT}};\stackrel{\‾}{Z_1}={\left[\begin{array}{cc}-\mathrm{Y\μ}& 0\end{array}\right]}^{\mathrm{\μT}};\stackrel{\‾}{Q}={\left[\begin{array}{ccc}\mathrm{Y\μ}& -Y_d& -Y_f\end{array}\right]}^{\mathrm{\μT}},$ Constructing nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node matrix equation system is described as

$\left\{\begin{array}{c}{J}_{\mathrm{\μ}}{\left({\mathrm{\γ}}_e\right)}_{\∞}=\underset{i=1}{\overset{\mathrm{\μ}-1}{\mathrm{\Σ}}}\sqrt{{\left\{\right[\underset{i={\mathrm{\μ}}_0+d\left(\mathrm{\μ}\right)}{\overset{i={\mathrm{\μ}}_0+L_{\mathrm{\μ}}}{\mathrm{\Σ}}}{\mathrm{\γ}}_e^T\left(i\right)\left]{\mathrm{\γ}}_e\left(i\right)\right\}}^T}\\ J_{\mathrm{th}-\mathrm{\μ}}=\underset{\mathrm{\ω}\∈{\mathrm{\φ}}_2[0,\∞],f=0}{\cos \mathrm{scup}}\underset{\mathrm{\μ}-d\left(\mathrm{\μ}\right)}{\overset{\mathrm{\μ}-1}{\mathrm{\Σ}}}{\left|\right|{\mathrm{\γ}}_e\left(\mathrm{\μ}\right)\left|\right|}_{e-2}^2\end{array}\right..$

Step 5: by J _μcos (γ _e) _∞row vector normalized, obtain matrix cosJ _th-μ, wherein $\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2<\underset{\mathrm{\&mu;}+\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&gamma;}\left|\right|\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\left|\right|}_{e-2}^2.$

Step 6: by matrix cosJ _th-μevery a line regard R as ^ka nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node in space, and use χ (μ-(μ) ^μ) ∈ R ^mmean algorithm is polymerized to L _μclass, if cosJ _th-μthe capable j class that belongs to of i, the y in nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node Y _ialso be divided into j class, thus the x in nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node X _ijust be divided into j class.(1) for optimized network structure Control System with Delay (4) wherein by inference, if there is positive definite symmetric matrices P, and Q, R and Arbitrary Matrix A, B, C meets following MATRIX INEQUALITIES and is

$\underset{\mathrm{\&mu;}+d\left(\mathrm{\&omega;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{{\mathrm{\&Pi;}}_{\&infin;}}=\left\{\begin{array}{ccccc}\stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{11}}& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{12}}& 0.5{\stackrel{\&OverBar;}{X}}^{\mathrm{\&mu;T}}& P\stackrel{\&OverBar;}{X}-Y& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{14}}\\ 0& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{22}}& 0.5{\stackrel{\&OverBar;}{X_1}}^{\mathrm{\&mu;T}}& P\stackrel{\&OverBar;}{X}-\mathrm{\&omega;}& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{24}}\\ 0& 0& 0& -\frac{1}{d_{\max }}R& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{34}}\\ 0& 0& 0& 0& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{44}}\end{array}\right\}<0$

, $\underset{\mathrm{\&mu;}-d_{\max }}{\overset{\frac{1}{d_{\max }}-2}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{\mathrm{\&Pi;}}=\left\{\begin{array}{c}\stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{11}}=d_{\max }\left\{\right[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R(\stackrel{\&OverBar;}{X}-I)]-P+[(d_{\max }-d_{\min }+1)+Q+Y+Y^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{C^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{C}\left]\right\}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{12}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+d_{\max }\left[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{X_1}\right]+{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}-Y+\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Z_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{14}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{Y}+\stackrel{\&OverBar;}{Y_1})+d_{\max }\left[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{Y_1}\right]+Q^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{22}}\text{=}\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+d_{\max }\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{X_1}-Q-\mathrm{\&omega;}{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{24}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{Y}+\stackrel{\&OverBar;}{Y_1})+\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }\stackrel{\&OverBar;}{X_1^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{Y}-Q^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{34}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}-Q^{\mathrm{\&mu;T}}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{44}}=\stackrel{\&OverBar;}{Y^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }\stackrel{\&OverBar;}{Y^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{Y}\text{+}\stackrel{\&OverBar;}{Q^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q}-{\mathrm{\&gamma;}}^{2}I\end{array}\right.$

Wherein ω (μ) _∞=0 o'clock, under the progressive stable and zero initial condition of optimized network structure Control System with Delay (4), meet $\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2<\underset{\mathrm{\&mu;}+d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\mathrm{\&gamma;}{\left|\right|\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\left|\right|}_{e-2}^2,\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\&NotEqual;0.$

Step 1 is the preprocessing process of nonsingular cloud terminal network structure Control System with Delay networking failure prediction, all numerical value nonsingular cloud terminal network structure Control System with Delay networking failure prediction in nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node X is transformed to based on nonsingular cloud terminal network structure Control System with Delay networking failure prediction, realizes normalization and the networking of structure Control System with Delay of nonsingular cloud terminal network structure Control System with Delay networking failure prediction.Step 2～6th, the process of the electric structure sensor model algorithm process of density sensitivity based on nonsingular cloud terminal network structure Control System with Delay networking failure prediction.Wherein, step 2 is weighed based on distance between nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node by average weighted hamming distance, builds on this basis the similarity measurement of density sensitivity.Step 4 is by solving eigenwert and the proper vector of nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node matrix equation, structural attitude vector space, thus realize the spectrum mapping of former nonsingular cloud terminal network structure Control System with Delay networking failure prediction space to low-dimensional linear measurement space.In the low-dimensional linear measurement space of step 6 after mapping, adopt traditional χ (μ-(μ) ^μ) ∈ R ^mmean algorithm cloud theoretical model, now just can obtain good cloud theoretical model quality.

Beneficial effect of the present invention:

The present invention is studied having the interval nonsingular cloud terminal network structure Control System with Delay networking failure prediction problem that becomes prediction intelligent grid.Traditional nonsingular cloud terminal network failure prediction problem has been transformed to the design of the nonsingular cloud terminal network failure prediction wave filter in digital intelligent system, the prediction intelligent grid of selecting the relevant cloud theoretical model apperceive identity method of prediction intelligent grid the nonsingular cloud terminal network failure prediction wave filter of having derived to exist relies on adequate condition.Designed nonsingular cloud terminal network failure prediction wave filter can become dry under the state of disturbing and ensure structure cloud model dynamic sensing signal condition bounded in the time there is isomery.Can effectively dope fault for damaged structure cloud model dynamic sensing signal and structure cloud model state evaluation function.Application to nonsingular cloud terminal network structure Control System with Delay and simulation result have shown science and the feasibility of the method.

Brief description of the drawings

Fig. 1 is fault dynamic sensing signal;

Fig. 2 is structure cloud model dynamic sensing signal;

Fig. 3 is that structure cloud model state evaluation function is flexible strategy values.

Embodiment.

Below in conjunction with the drawings and specific embodiments, technical scheme of the present invention is described in more detail.

Problem is described

Consider that the nonsingular cloud terminal network structure Control System with Delay perceptive object cloud theoretical model in intelligent grid is described as

$\left\{\begin{array}{c}{\mathrm{\&chi;}(\mathrm{\&mu;}+1)}_{\mathrm{\&psi;}}^{\mathrm{\&mu;T}}\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}[\mathrm{X\&mu;}-\mathrm{\&chi;}{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}+\mathrm{Y\&mu;\&chi;}{(\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right))}^{\mathrm{\&mu;T}}-\mathrm{Y\&mu;x}{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}+Y_q[\mathrm{\&mu;}-\mathrm{\&zeta;}{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}]+{[Y_f\mathrm{\&mu;}-f{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}]]}^{\mathrm{\&mu;T}}\\ y{\left(\mathrm{\&mu;}\right)}_{\mathrm{\&psi;}}^{\mathrm{\&mu;T}}=\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}[\mathrm{Z\&mu;}-\mathrm{\&chi;}{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}+\mathrm{Q\&mu;\&chi;}{(\mathrm{\&mu;}-d\left(\mathrm{\&chi;}\right))}^{\mathrm{\&mu;T}}-\mathrm{Q\&mu;x}{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}+Q_q[\mathrm{\&mu;}-\mathrm{\&zeta;}{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}]+Q_f{[\mathrm{\&mu;}-f{\left(\mathrm{\&mu;}\right)}^{\mathrm{\&mu;T}}]]}^{\mathrm{\&mu;T}}\\ \mathrm{\&chi;}\left(\mathrm{\&mu;}\right)=\underset{\mathrm{\&mu;}-d\left(\mathrm{\&phi;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\mathrm{\&mu;}-\mathrm{\&phi;}{{\left(\mathrm{\&mu;}\right)}_{\mathrm{\&psi;}}}^{\mathrm{\&mu;T}}\\ \mathrm{\&mu;}=-{\mathrm{\&Psi;}}_{\mathrm{\&mu;}}{d}_{\max }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}+{\mathrm{\&chi;}}_n,-{\mathrm{\&Psi;}}_{\mathrm{\&mu;}+1}{d}_{\max }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}+{\mathrm{\&chi;}}_{n-1}+1,{\mathrm{\&Psi;}}_{\mathrm{\&mu;}+1}{d}_{\max }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}+{\mathrm{\&chi;}}_{n-2}+2,...,0\end{array}\right.---\left(1\right)$

Wherein, μ is the cloud model network structure lagging network structure Control System with Delay time; χ (μ-(μ) ^μ) ∈ R ^mfor state node variable; μ-d (μ) is structure time lag prediction intelligent grid; μ-ω (μ) ∈ R ^mfor structure time lag control inputs; μ-ζ (μ) is unusual interference input; μ-f (μ) is for waiting for the unknown failure dynamic sensing signal of nonsingular cloud terminal network failure prediction; μ-y (μ) ∈ R ⁿfor output; μ-φ (μ) is given initial condition sequence; X, Y is suitable dimension cloud model network structure lagging network structure Control System with Delay matrix; μ T is network structure time lag control vector matrix; Y _qand Y _fbe respectively the impact on system state assessment of unusual interference and fault dynamic sensing signal; Q _qand Q _fbe respectively the impact on system state assessment of unusual interference and fault dynamic sensing signal.Have following demonstration to be for the nonsingular cloud terminal network structure Control System with Delay (1) in intelligent grid:

Prove (1) prediction intelligent grid d (μ) temporal evolution and changing in set of integers, and meet and be constrained to as follows: $d_{\min }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}+\mathrm{\&chi;}\&le;d\left(\mathrm{\&mu;}\right)\&le;d_{\max }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}-\mathrm{\&chi;}---\left(2\right)$

Wherein, $d_{\min }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}+\mathrm{\&chi;}$ With $d_{\max }\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}_{\mathrm{\&mu;}}-\mathrm{\&chi;}$ For just gathering integer.

Prove (2) network structure Control System with Delay matrix μ known and can ensure that system (1), under state bounded input, exports also state bounded.For the nonsingular cloud terminal network structure Control System with Delay (1) in intelligent grid, the nonsingular cloud terminal network failure prediction wave filter that uncertain nonlinear organization Systems with Time Delay Feedback filter construction has following form is:

$\left\{\begin{array}{c}{\mathrm{\&chi;}}_F(\mathrm{\&mu;}+1)=\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\left\{X_F{\mathrm{\&chi;}}_F\right[\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)]+Y_F[\mathrm{\&mu;}-y\left(\mathrm{\&mu;}\right)\left]\right\}\\ \mathrm{\&gamma;}[\mathrm{\&mu;}-\left(\mathrm{\&mu;}\right)]=\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{Z}_F{\mathrm{\&chi;}}_F[\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)]\\ {\mathrm{\&chi;}}_F\left(0\right)=0\\ Q_F\left(0\right)=0\end{array}\right.---\left(3\right)$

Wherein, χ _f(μ-d (μ)) ∈ R ⁿfor filter status assessment node variable; γ (μ-(μ)) ∈ R ^γfor wave filter output; X _f, Y _f, Z _f, Q _fthe electric-wave filter matrix to be designed such as be respectively.

Definition structure cloud model dynamic sensing signal is γ _e[μ-d (μ)]={ [μ-γ (μ)]-[μ-y (μ)] } ^{μ T}, can obtain following optimized network structure Control System with Delay by system (1) and mode (3) and be:

$\left\{\begin{array}{c}z(\mathrm{\&mu;}+1)=\stackrel{\&OverBar;}{X}[\mathrm{\&mu;}-z\left(\mathrm{\&mu;}\right)]+\stackrel{\&OverBar;}{X_1}z\left[(\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right))\right]+\stackrel{\&OverBar;}{Y}[\mathrm{\&mu;}-\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)]\\ {\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)=\stackrel{\&OverBar;}{Z}z\left(\mathrm{\&mu;}\right)+\stackrel{\&OverBar;}{Z_1}z\left[(\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right))\right]+\stackrel{\&OverBar;}{Q}\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)\end{array}\right.---\left(4\right)$

Wherein, for the optimized network vector in cloud model network structure lagging network structure Control System with Delay; ω (μ)=[x ^{μ T}(μ) ζ ^{μ T}(μ) f ^{μ T}(μ)] ^{μ T}; with for optimized network structure Control System with Delay matrix, its expression formula is

$\{\stackrel{\&OverBar;}{X}={\left[\begin{array}{cc}{X}_{\mathrm{\&mu;}}& 0\\ Y_{F}Z& X_F\end{array}\right]}^{\mathrm{\&mu;T}};{\stackrel{\&OverBar;}{X}}_1={\left[\begin{array}{cc}\mathrm{Y\&mu;}& 0\\ Y_F\mathrm{Q\&mu;}& 0\end{array}\right]}^{\mathrm{\&mu;T}};\stackrel{\&OverBar;}{Y}={\left\{\begin{array}{ccc}-\mathrm{Y\&mu;}& Y_d& Y_f\\ -Q{Y}_f\mathrm{\&mu;}& Y_f{Q}_q& Y_f{Q}_f\end{array}\right\}}^{\mathrm{\&mu;T}}$

$\{\stackrel{\&OverBar;}{Z}={\left[\begin{array}{cc}-X& Z_F\end{array}\right]}^{\mathrm{\&mu;T}};\stackrel{\&OverBar;}{Z_1}={\left[\begin{array}{cc}-\mathrm{Y\&mu;}& 0\end{array}\right]}^{\mathrm{\&mu;T}};\stackrel{\&OverBar;}{Q}={\left[\begin{array}{ccc}\mathrm{Y\&mu;}& -Y_d& -Y_f\end{array}\right]}^{\mathrm{\&mu;T}}$

Wherein the nonsingular cloud terminal network of uncertain nonlinear organization time lag failure prediction problem can be summed up as that to find wave filter be flexible strategy matrixes, following two conditions is set up be

(1), ω (μ) _∞=0 o'clock, optimized network structure Control System with Delay (4) Asymptotic Stability;

(2) under zero initial condition, to any ω (μ) _∞≠ 0, have

$\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2<\underset{\mathrm{\&mu;}+\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&gamma;}\left|\right|\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\left|\right|}_{e-2}^2---\left(5\right)$

For reaching the object of nonsingular cloud terminal network failure prediction, need to carry out state estimation to structure cloud model dynamic sensing signal, definition structure cloud model state evaluation function is J _μcos (γ _e) _∞cosine threshold value cosJ with nonsingular cloud terminal network failure prediction _th-μas follows:

$\left\{\begin{array}{c}{J}_{\mathrm{\&mu;}}{\left({\mathrm{\&gamma;}}_e\right)}_{\&infin;}=\underset{i=1}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\sqrt{{\left\{\underset{i={\mathrm{\&mu;}}_0+d\left(\mathrm{\&mu;}\right)}{\overset{i={\mathrm{\&mu;}}_0+L_{\mathrm{\&mu;}}}{\mathrm{\&Sigma;}}}{\mathrm{\&gamma;}}_e^T\left(i\right)\right]{\mathrm{\&gamma;}}_e\left(i\right)\}}^T}\\ J_{\mathrm{th}-\mathrm{\&mu;}}=\underset{\mathrm{\&omega;}\&Element;{\mathrm{\&phi;}}_2[0,\&infin;],f=0}{\cos \mathrm{scup}}\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2\end{array}\right.---\left(6\right)$

Wherein, μ ₀for the primary fault state estimation moment; L _μfor the state estimation time period.Nonsingular cloud terminal network failure prediction logic is as follows:

Nonsingular cloud terminal network structure Delay-Fault controller and design

Reasoning (1) is for optimized network structure Control System with Delay (4), if there is positive definite symmetric matrices P, and Q, R and Arbitrary Matrix A, B, C meets following MATRIX INEQUALITIES and is

$\underset{\mathrm{\&mu;}+d\left(\mathrm{\&omega;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{\mathrm{\&Pi;}}}_{\&infin;}=\left\{\begin{array}{ccccc}\stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{11}}& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{12}}& 0.5{\stackrel{\&OverBar;}{X}}^{\mathrm{\&mu;T}}& P\stackrel{\&OverBar;}{X}-Y& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{14}}\\ 0& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{22}}& \text{0.5}{\stackrel{\&OverBar;}{X_1}}^{\mathrm{\&mu;T}}& P\stackrel{\&OverBar;}{X}-\mathrm{\&omega;}& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{24}}\\ 0& 0& 0& -\frac{1}{d_{\max }}R& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{34}}\\ 0& 0& 0& 0& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{44}}\end{array}\right\}<0---\left(7\right)$

, $\underset{\mathrm{\&mu;}-d_{\max }}{\overset{\frac{1}{d_{\max }}-2}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{\mathrm{\&Pi;}}=\left\{\begin{array}{c}\stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{11}}=d_{\max }\left\{\right[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R(\stackrel{\&OverBar;}{X}-I)]-P+[(d_{\max }-d_{\min }+1)+Q+Y+Y^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{C^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{C}\left]\right\}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{12}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+d_{\max }\left[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{X_1}\right]+{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}-Y+\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Z_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{14}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{Y}+\stackrel{\&OverBar;}{Y_1})+d_{\max }\left[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{Y_1}\right]+Q^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{22}}\text{=}\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+d_{\max }\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{X_1}-Q-\mathrm{\&omega;}{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{24}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{Y}+\stackrel{\&OverBar;}{Y_1})+\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }\stackrel{\&OverBar;}{X_1^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{Y}-Q^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{34}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}-Q^{\mathrm{\&mu;T}}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{44}}=\stackrel{\&OverBar;}{Y^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }\stackrel{\&OverBar;}{Y^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{Y}\text{+}\stackrel{\&OverBar;}{Q^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q}-{\mathrm{\&gamma;}}^{2}I\end{array}\right.$

Wherein ω (μ) _∞=0 o'clock, under the progressive stable and zero initial condition of optimized network structure Control System with Delay (4), meet $\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2<\underset{\mathrm{\&mu;}+d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\mathrm{\&gamma;}{\left|\right|\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\left|\right|}_{e-2}^2,\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\&NotEqual;0$

Proof is established y _m(μ)=z (μ+1)-z (μ), chooses following candidate's cloud theoretical model apperceive identity and is

$\underset{\mathrm{\&mu;}+d_{\max v}\left(\mathrm{\&mu;}\right)}{\overset{1/d_{\max }}{\mathrm{\&Sigma;}}}{V}_{\mathrm{\&mu;}}=\left\{\begin{array}{c}v\left\{\right[z\left(\mathrm{\&mu;}\right)],d\left(\mathrm{\&mu;}\right),\mathrm{\&mu;}\}=\mathrm{\&mu;}{[\left(v_1\right)+\left(v_2\right)+\left(v_3\right)+\left(v_4\right)]}^{\mathrm{\&mu;T}}\\ v_1\left\{\right[z\left(\mathrm{\&mu;}\right)],k\}=z^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)\mathrm{\&mu;Pz}\left(\mathrm{\&mu;}\right)\\ v_2\left\{\right[z\left(\mathrm{\&mu;}\right)],d\left(\mathrm{\&mu;}\right),\mathrm{\&mu;}\}=\underset{i=\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{Z}^{\mathrm{\&mu;T}}\mathrm{\&mu;}\left(i\right)\mathrm{Qz}\left(i\right)\\ v_3\left\{\right[z\left(\mathrm{\&mu;}\right)],\mathrm{\&mu;}\}=\underset{j={-d}_{\max }}{\overset{-1}{\mathrm{\&Sigma;}}}\underset{i=\mathrm{\&mu;}+j\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\mathrm{\&mu;}{y}_m^{\mathrm{\&mu;T}}\left(i\right)R{y}_m\left(i\right)\end{array}\right.---\left(8\right)$

Definition Δ V=V (μ+1)-V (μ), order mode (10) is for any suitable dimension matrix Y, W, Q establishment.

$\frac{1}{2}\{z^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)Y+z^{\mathrm{\&mu;T}}[\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)]\mathrm{\&omega;}+{\mathrm{\&gamma;}}^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)Q\}\&times;{\{z\left(\mathrm{\&mu;}\right)-z[\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)]-\underset{i=\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{y}_m\left(i\right)\}}^{\mathrm{\&mu;T}}=0---\left(9\right)$

Solve Δ V for optimized network structure Control System with Delay (4) and mode (10), can obtain

$\mathrm{\&Delta;V}\&le;\frac{1}{d\left(\mathrm{\&mu;}\right)}\underset{i=\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&xi;}}^{\mathrm{\&mu;T}}(\mathrm{\&mu;},i)\mathrm{\&Pi;\&xi;}(\mathrm{\&mu;},i)---\left(10\right)$

Wherein ${\mathrm{\&psi;}}^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;,i}\right)\text{{}\left[\begin{array}{cccc}{z}^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)& z^{\mathrm{\&mu;T}}(\mathrm{\&mu;}-\left(\mathrm{\&mu;}\right))& d\left(\mathrm{\&mu;}\right)y_m^{\mathrm{\&mu;T}}\left(i\right)& {\mathrm{\&gamma;}}^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)\end{array}\right]\},$ Have

Wherein $\underset{\mathrm{\&mu;}-d_{\max }\left(\mathrm{\&mu;}\right)}{\overset{\frac{1}{d_{\max }}-1}{\mathrm{\&Sigma;}}}\mathrm{\&Pi;}=\left\{\begin{array}{c}{\mathrm{\&rho;}}_{11}=d_{\max }{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R(\stackrel{\&OverBar;}{X}-I)-P+(d_{\max }-d_{\min }+1)Q+Y+Y^{\mathrm{\&mu;T}}\\ {\mathrm{\&rho;}}_{14}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{Y}+Q^{\mathrm{\&mu;T}}\\ {\mathrm{\&rho;}}_{44}={\stackrel{\&OverBar;}{Y}}^{\mathrm{\&mu;T}}P\stackrel{\&OverBar;}{Y}+d_{\max }{\stackrel{\&OverBar;}{Y}}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{Y}\end{array}\right.,$ Consider that following apperceive identity target function is $d\left(\mathrm{\&mu;}\right)=\underset{\mathrm{\&mu;}=1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}\underset{\mathrm{\&theta;}=0}{\overset{\&infin;}{\mathrm{\&Sigma;}}}[{\mathrm{\&gamma;}}_e^{\mathrm{\&mu;T}}\left(\mathrm{\&theta;}\right){\mathrm{\&gamma;}}_e\left(\mathrm{\&theta;}\right)-{\mathrm{\&gamma;}}^2{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}\left(\mathrm{\&theta;}\right)\mathrm{\&omega;}\left(\mathrm{\&theta;}\right)]---\left(11\right)$

Under cloud model network structure lagging network structure Control System with Delay zero initial condition, there is V (0, d (0))=0, V (∞, d (∞)) > 0, can obtain

${\mathrm{\&gamma;}}_e^{\mathrm{\&mu;T}}\left(\mathrm{\&theta;}\right){\mathrm{\&gamma;}}_e\left(\mathrm{\&theta;}\right)-{\mathrm{\&gamma;}}^2{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}\left(\mathrm{\&theta;}\right)\mathrm{\&omega;}\left(\mathrm{\&theta;}\right)+\mathrm{\&Delta;V}\&le;\frac{1}{d\left(\mathrm{\&mu;}\right)}\underset{\mathrm{\&mu;}=1}{\overset{\&infin;}{\mathrm{\&Sigma;}}}\underset{i=\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&psi;}}^{\mathrm{\&mu;T}}(\mathrm{\&mu;},i)\stackrel{\&OverBar;}{\mathrm{\&Pi;}}\mathrm{\&psi;}(\mathrm{\&mu;},i)---\left(12\right)$

Wherein $\underset{\mathrm{\&mu;}-d_{\max }\left(\mathrm{\&mu;}\right)}{\overset{\frac{1}{d_{\max }}-1}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{\mathrm{\&Pi;}}}_{\&infin;}=\underset{\mathrm{\&mu;}+d_{\max }\left(\mathrm{\&mu;}\right)}{\overset{\frac{1}{d_{\max }}+1}{\mathrm{\&Sigma;}}}{\stackrel{\&OverBar;}{\stackrel{\&OverBar;}{\stackrel{\&OverBar;}{\mathrm{\&Pi;}}}}}_{\&infin;}+\left\{\begin{array}{ccccccc}{d}_{\mathrm{\&mu;}}\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}& (\mathrm{\&mu;}-1)\stackrel{\&OverBar;}{Z}& d_{\mathrm{\&mu;}-1}\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}& (\mathrm{\&mu;}-2)\stackrel{\&OverBar;}{Z_1}& 0& d_{\mathrm{\&mu;}-n}\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}& (\mathrm{\&mu;}-1)\stackrel{\&OverBar;}{Q}\\ 0& d_{\mathrm{\&mu;}-1}\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}& (\mathrm{\&mu;}-1)\stackrel{\&OverBar;}{Z_1}& 0& 0& d_{\mathrm{\&mu;}-1}\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}& (\mathrm{\&mu;}-1)\stackrel{\&OverBar;}{Q}\\ 0& 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& d_{\mathrm{\&mu;}-1}{\stackrel{\&OverBar;}{Q}}^{\mathrm{\&mu;T}}\stackrel{\&OverBar;}{Q}-{\mathrm{\&gamma;}}^2{I}_{\mathrm{\&mu;}}& I_{\mathrm{\&mu;}}\end{array}\right\},$ Comprehensively above-mentioned, obtain inequality having ensured, for all non-zero ψ (μ, m), has μ < 0.So, if for all non-linear zero ω ∈ L ₂[0, ∞), have for the state of ω ≡ 0, provable be the progressive stable adequate condition of network structure Control System with Delay, no longer describe herein.Introduce demonstration simultaneously.

Prove (1) and work as R ^{μ × μ}dimension positive definite symmetric matrices P _μand L _μmake following nonsingular cloud terminal network structure Control System with Delay wave filter establishment be $\left[\begin{array}{cc}{P}_{\mathrm{\&mu;}}& I_{\mathrm{\&mu;}}\\ I_{\mathrm{\&mu;}}& L_{\mathrm{\&mu;}}\end{array}\right]\&GreaterEqual;0---\left(13\right)$

Trace (P _μl _μ)>=n, trace (P _μl _μ) ≡ n and if only if P _μl _μ=I _μ

Prove (2) for network structure Control System with Delay (1), exist as the adequate condition of the nonsingular cloud terminal network of mode (3) failure prediction wave filter is: have applicable dimension positive definite symmetric matrices P _μ, Q _μ, R _μ, L _μ, S _μwith Arbitrary Matrix Y, W, Q sets up nonsingular cloud terminal network structure Control System with Delay wave filter and the mode (16) shown in mode (14) and mode (15) simultaneously.

$\left[\begin{array}{cc}{\mathrm{\&Psi;}}_{11}(\mathrm{\&omega;}+1)& {\mathrm{\&Psi;}}_{12}(\mathrm{\&omega;}-1)\\ {\mathrm{\&Psi;}}_{12}^T(\mathrm{\&omega;}-1)& {\mathrm{\&Psi;}}_{22}(\mathrm{\&omega;}+1)\end{array}\right]<0---\left(14\right)$

$\left[\begin{array}{cc}{R}_{\mathrm{\&mu;}}(\mathrm{\&psi;}+1)& I_{\mathrm{\&mu;}}(\mathrm{\&psi;}-1)\\ I_{\mathrm{\&mu;}}(\mathrm{\&psi;}-1)& L_{\mathrm{\&mu;}}(\mathrm{\&psi;}+1)\end{array}\right]\&GreaterEqual;0;\left[\begin{array}{cc}{P}_{\mathrm{\&mu;}}(\mathrm{\&psi;}+1)& I_{\mathrm{\&mu;}}(\mathrm{\&psi;}-1)\\ I_{\mathrm{\&mu;}}(\mathrm{\&psi;}-1)& S_{\mathrm{\&mu;}}(\mathrm{\&psi;}+1)\end{array}\right]\&GreaterEqual;0---\left(15\right)$

trace(R _μL _μ(ψ+1)+P _μS _μ(ψ-1))＝2(η+r+ψ) (16)

Wherein, ${\mathrm{\&Psi;}}_{11}=\left\{\begin{array}{cccc}{\mathrm{\&Omega;}}_{11}& {\mathrm{\&omega;}}^{\mathrm{\&mu;T}}-Y& -Y& Q^{\mathrm{\&mu;T}}\\ 0& {\mathrm{\&Omega;}}_{22}& -\mathrm{\&omega;}& -Q^{\mathrm{\&mu;T}}\\ 0& 0& -\frac{1}{d_{\max }}R& -Q^{\mathrm{\&mu;T}}\\ 0& 0& 0& -{\mathrm{\&gamma;}}^2{I}_{\mathrm{\&mu;}}\end{array}\right\};$

${\mathrm{\&Psi;}}_{12}=\left\{\begin{array}{cccc}{\stackrel{\&OverBar;}{Z}}^{\mathrm{\&mu;T}}& \mathrm{\&omega;}{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}& {\stackrel{\&OverBar;}{X}}^{\mathrm{\&mu;T}}P& 0\\ {\stackrel{\&OverBar;}{Z_1}}^{\mathrm{\&mu;T}}& {\stackrel{\&OverBar;}{X_1}}^{\mathrm{\&mu;T}}& \mathrm{\&omega;}{(\stackrel{\&OverBar;}{X}+{\stackrel{\&OverBar;}{X_1}}^T)}^{\mathrm{\&mu;T}}& {{\stackrel{\&OverBar;}{X}}_1}^{\mathrm{\&mu;T}}\\ 0& 0& \mathrm{\&omega;}{\stackrel{\&OverBar;}{X}}^{\mathrm{\&mu;T}}& 0\\ {\stackrel{\&OverBar;}{Q}}^{\mathrm{\&mu;T}}& {\stackrel{\&OverBar;}{Y}}^{\mathrm{\&mu;T}}& {\stackrel{\&OverBar;}{Y}}^{\mathrm{\&mu;T}}& \mathrm{\&omega;}{\stackrel{\&OverBar;}{Y}}^{\mathrm{\&mu;T}}\end{array}\right\};{\mathrm{\&Psi;}}_{22}=\left\{\begin{array}{cccc}-I& 0& 0& 0\\ \text{0}& -\frac{1}{d_{\max }}L& 0& 0\\ 0& 0& -S& 0\\ 0& 0& 0& -S\end{array}\right\};$

Ω ₁₁＝ω[(d _max-d _min+1)Q-P-P ^μT+Y+Y ^μT]

Ω ₂₂＝μ[-Q-ω-ω ^μT]

Prove for inference (1), the P matrix in mode (8) is replaced with to 2P, again due to

${\stackrel{\&OverBar;}{X}}_1^{\mathrm{\&mu;T}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+{(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})}^{\mathrm{\&mu;T}}P{\stackrel{\&OverBar;}{X}}_1^{\&prime;}={(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})}^{\mathrm{\&mu;T}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+{\stackrel{\&OverBar;}{X}}_1^{\mathrm{\&mu;T}}P{\stackrel{\&OverBar;}{X}}_1^{\&prime;}-{\stackrel{\&OverBar;}{X}}_m^{\mathrm{\&mu;T}}P\stackrel{\&OverBar;}{X}---\left(17\right)$

Then mend demonstration conversion by a series of cloud model algorithms, can obtain mode (8) set up be equivalent to into

The load project of removing mode (18) left end is and carry out congruent transformation diag[I I I I I R ^-1p ^-1p ^-1], for prove (1) can obtain mode (18) set up adequate condition as shown in mode (14), mode (15) and mode (16).

Arbitrary Matrix A, B, C and network structure Control System with Delay state estimation that mode (9) is introduced with the state estimation χ (j) postponing, j=μ=d (μ) ..., μ-1 is relevant, then the matrix A of can being open to the custom, B, C embodies state estimation in nonsingular cloud terminal network structure Control System with Delay wave filter to postpone.

Mode contains coupling project in (18) because P is positive definite symmetric matrices, so again wherein Γ [μ-d (μ)] is for being applicable to dimension matrix.Can in mode (18), remove coupling terms for comprehensive above-mentioned thought.This processing mode can be converted into nonlinear inequalities nonsingular cloud terminal network structure Control System with Delay wave filter, problem of pretreatment.

Prove the feasibility problem described in (2), cannot directly adopt Software tool to solve.For proving (1), feasibility problem can be converted into the optimum path problems under mode (14) and mode (15) environment, wherein optimal path apperceive identity index is ${\stackrel{\&OverBar;}{X}}^{\mathrm{\&mu;T}}P{\stackrel{\&OverBar;}{X}}_1=\cos \min \left|\mathrm{trace}(\mathrm{RL}+\mathrm{PS})\right|---\left(19\right)$

Comprehensive the problems referred to above, are called as CCL problem, can use following link node variable algorithm be optimized into

(1), do not consider mode (16), ask for a feasible explanation of proving (2).Make μ _s=0, make (P _μ, S _μ, R _μ, L _μ)=(P ₀, S ₀, R ₀, L ₀);

(2) make μ _s=μ _s+ 1, if μ _s≤ μ _{s_max}, under mode (14) and mode (15) environment, carry out the optimization of cloud model network structure lagging network structure Control System with Delay multitiered network and explore cos trace (R _μl+RL _μ)+cos trace (P _μs+PS _μ) minimum value.Make { (P _μ), (S _μ), (R _μ), (L _μ)=(P), (S), (R), (L) }. otherwise, algorithm out of service;

(3), if costrace (PS+RL)-2 (η+r+ ψ) is < δ, δ is allowable error circle of setting, algorithm out of service.Otherwise turn back to (2).

System emulation is analyzed

Provide the design result of nonsingular cloud terminal network failure prediction wave filter for nonsingular cloud terminal network structure Control System with Delay.So utilize the nonsingular cloud terminal network of prior art structure Control System with Delay for NASA and the unmanned spacecraft of joint research and development about USAF nineteen sixty, wherein structural model.At p=7620m, when β=0.7Ma, the cloud model network structure lagging network state estimation equation of nonsingular cloud terminal network structure Control System with Delay has the form of network structure Control System with Delay (1).State node variable $x={\left[\begin{array}{cc}\mathrm{\&alpha;}& q\\ p& \mathrm{\&beta;}\\ x& y\end{array}\right]}^T$ Represent respectively the angle of attack and angle of depression speed, be input as μ={ (δ _e), (δ _v), (δ _c), (δ _a) T represents respectively elevating rudder, elevating rudder auxiliary wing, canard and symmetrical aileron movement angle.Wherein cloud model network structure lagging network structure Control System with Delay matrix is

$\left\{\begin{array}{c}X=\left[\begin{array}{cc}0.9767& 0.0190\\ 0.2573& 0.9658\end{array}\right];Y_q\left[\begin{array}{c}1\\ 1\end{array}\right];Y_f\left[\begin{array}{c}0.2\\ 0.1\end{array}\right];\\ Y=\left[\begin{array}{cccc}-0.0124& -0.0087& 0.0055& -0.0089\\ -0.4376& -0.3109& 0.2355& -0.3155\end{array}\right];\\ Z=\left[\begin{array}{cc}0& 1\\ 1& 0\end{array}\right];Q_q=\left[\begin{array}{c}1\\ 1\end{array}\right];Q_f\left[\begin{array}{c}0.2\\ 0.1\end{array}\right];\\ Q=\left[\begin{array}{cccc}-0.0135& -0.0125& 0.0107& -0.0126\\ -0.2135& -0.1287& 0.1305& -0.1311\end{array}\right]\end{array}\right.$

Solving of cloud model network structure lagging network structure Control System with Delay matrix μ can be referring to prior art, and wherein concrete apperceive identity index value is

$\mathrm{\&mu;}={\left\{{\left[\begin{array}{cccc}-2.2175& -1.1767& -8.5298& -1.3186\\ 2.1198& 1.1902& 6.5578& 1.3066\end{array}\right]}^{\mathrm{\&mu;T}}\right\}}^T$

By proving (2), what can try to achieve nonsingular cloud terminal network failure prediction wave filter is that flexible strategy are

$X_F=\left[\begin{array}{cc}0.9800& -0.0025\\ -0.0026& 0.8592\end{array}\right];Y_F\left[\begin{array}{cc}-0.0031& -0.0025\\ -0.1675& -0.1625\end{array}\right];Z_F\left[\begin{array}{cc}0& 1\\ 1& 0\end{array}\right]$

Wherein, this problem is in emulation experiment, and fault dynamic sensing signal as shown in Figure 1.Unusual disturbance state perceptual signal is taken as energy value white noise dynamic sensing signal.Fig. 2 is the structure cloud model dynamic sensing signal of nonsingular cloud terminal network failure prediction wave filter gained, and Fig. 3 represents that the structure cloud model state evaluation function of dynamic sensing signal under different conditions is flexible strategy values.Based on this, this nonsingular cloud terminal network failure prediction wave filter of analysis of simulation result can be good at optimizing nonsingular cloud terminal network failure prediction.

The above; it is only preferably embodiment of the present invention; protection scope of the present invention is not limited to this; any be familiar with those skilled in the art the present invention disclose technical scope in, the simple change of the technical scheme that can obtain apparently or equivalence replace all fall within the scope of protection of the present invention.

Claims (2)

1. the failure prediction method based under intelligent power distribution environment, is characterized in that, comprises the following steps:

Based on nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node comprise n nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node, each nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node has n dimension attribute and property value is the prediction of the nonsingular cloud terminal network of numerical value structure Delay-Fault, is designated as ${\mathrm{\&chi;}(\mathrm{\&mu;}+1)}_{\mathrm{\&psi;}}^{\mathrm{\&mu;T}}=\{\mathrm{\&Sigma;}\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}(i+1){\mathrm{\&chi;}}_1\left(\mathrm{\&mu;}\right),\mathrm{\&Sigma;}{\mathrm{\&chi;}}_2\left(\mathrm{\&mu;}\right),...,\mathrm{\&Sigma;}{\mathrm{\&chi;}}_{n-i}\left(\mathrm{\&mu;}\right)\},$ I=1,2 ..., n; The attribute vector of nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node is μ=1,2 ..., n, cloud theoretical model algorithm is gathered is ${\mathrm{\&chi;}(\mathrm{\&mu;}+1)}_{\mathrm{\&psi;}}^{\mathrm{\&mu;T}}=\{\mathrm{\&Sigma;}{\mathrm{\&chi;}}_1\left(\mathrm{\&mu;}\right),\mathrm{\&Sigma;}{\mathrm{\&chi;}}_2\left(\mathrm{\&mu;}\right),...,\mathrm{\&Sigma;}{\mathrm{\&chi;}}_n\left(\mathrm{\&mu;}\right)\}$ The process of class is:

Step 1: the nonsingular cloud terminal network structure Control System with Delay networking failure prediction transform method based on cloud theoretical model of describing according to system is to former nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node the networking of structure Control System with Delay, obtains the nonsingular cloud terminal network structural network fault perceptual signal prediction set of node forming based on nonsingular cloud terminal network structure Control System with Delay networking failure prediction wherein i=1,2 ..., n; Cloud theoretical model function Y _fselect traditional Y _qmean algorithm realizes;

Step 2: build the responsive similarity measurement of density based on hamming distance, thereby obtain the similar matrix χ (μ-(μ) of nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node Y ^μ) ∈ R ^m, wherein

$W_{\mathrm{ij}}\left\{\begin{array}{c}{\mathrm{\&chi;}}_F(\mathrm{\&mu;}+1)=\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\left\{X_F{\mathrm{\&chi;}}_F\right[\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)]+Y_F[\mathrm{\&mu;}-y\left(\mathrm{\&mu;}\right)\left]\right\}\\ \mathrm{\&gamma;}[\mathrm{\&mu;}-\left(\mathrm{\&mu;}\right)]=\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{Z}_F{\mathrm{\&chi;}}_F[\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)]\\ {\mathrm{\&chi;}}_F\left(0\right)=0\\ Q_F\left(0\right)=0\end{array}\right.;$

Step 3: the cloud theoretical model matrix description of constructing nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node Y is γ _e[μ-d (μ)]={ [μ-γ (μ)]-[μ-y (μ)] } ^{μ T}, wherein χ _f(μ-d (μ)) ∈ R ⁿfor vector matrix system is described as

$\left\{\begin{array}{c}z(\mathrm{\&mu;}+1)=\stackrel{\&OverBar;}{X}[\mathrm{\&mu;}-z\left(\mathrm{\&mu;}\right)]+{\stackrel{\&OverBar;}{X}}_{1}z\left[(\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right))\right]+\stackrel{\&OverBar;}{Y}[\mathrm{\&mu;}-\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)]\\ {\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)=\stackrel{\&OverBar;}{Z}z\left(\mathrm{\&mu;}\right)+{\stackrel{\&OverBar;}{Z}}_{1}z\left[(\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right))\right]+\stackrel{\&OverBar;}{Q}\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)\end{array}\right.;$

Step 4: ask nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node matrix equation $z\left(\mathrm{\&mu;}\right)={\left[\begin{array}{cc}{\mathrm{\&chi;}}^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)& {\mathrm{\&chi;}}_F^{\mathrm{\&mu;T}}\left(\mathrm{\&mu;}\right)\end{array}\right]}^{\mathrm{\&mu;T}}$ Front ω (μ)=[x ^{μ T}(μ) ζ ^{μ T}(μ) f ^{μ T}(μ)] ^{μ T}individual eigenvalue of maximum characteristic of correspondence vector expression is described as

$\{\stackrel{\&OverBar;}{X}={\left[\begin{array}{cc}{X}_{\mathrm{\&mu;}}& 0\\ Y_{F}Z& X_F\end{array}\right]}^{\mathrm{\&mu;T}};{\stackrel{\&OverBar;}{X}}_1={\left[\begin{array}{cc}\mathrm{Y\&mu;}& 0\\ Y_F\mathrm{Q\&mu;}& 0\end{array}\right]}^{\mathrm{\&mu;T}};\stackrel{\&OverBar;}{Y}={\left\{\begin{array}{ccc}-\mathrm{Y\&mu;}& Y_d& Y_f\\ -Q{Y}_f\mathrm{\&mu;}& Y_f{Q}_q& Y_f{Q}_f\end{array}\right\}}^{\mathrm{\&mu;T}}$

$\{\stackrel{\&OverBar;}{Z}={\left[\begin{array}{cc}-X& Z_F\end{array}\right]}^{\mathrm{\&mu;T}};\stackrel{\&OverBar;}{Z_1}={\left[\begin{array}{cc}-\mathrm{Y\&mu;}& 0\end{array}\right]}^{\mathrm{\&mu;T}};\stackrel{\&OverBar;}{Q}={\left[\begin{array}{ccc}\mathrm{Y\&mu;}& -Y_d& -Y_f\end{array}\right]}^{\mathrm{\&mu;T}},$ Constructing nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node matrix equation system is described as

$\left\{\begin{array}{c}{J}_{\mathrm{\&mu;}}{\left({\mathrm{\&gamma;}}_e\right)}_{\&infin;}=\underset{i=1}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\sqrt{{\left\{\right[\underset{i={\mathrm{\&mu;}}_0+d\left(\mathrm{\&mu;}\right)}{\overset{i={\mathrm{\&mu;}}_0+L_{\mathrm{\&mu;}}}{\mathrm{\&Sigma;}}}{\mathrm{\&gamma;}}_e^T\left(i\right)\left]{\mathrm{\&gamma;}}_e\left(i\right)\right\}}^T}\\ J_{\mathrm{th}-\mathrm{\&mu;}}=\underset{\mathrm{\&omega;}\&Element;{\mathrm{\&phi;}}_2[0,\&infin;],f=0}{\cos \mathrm{scup}}\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2\end{array}\right.;$

Step 5: by J _μcos (γ _e) _∞row vector normalized, obtain matrix cosJ _th-μ, wherein $\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2<\underset{\mathrm{\&mu;}+\mathrm{\&omega;}\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\mathrm{\&gamma;}\left|\right|\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\left|\right|}_{e-2}^2;$

Step 6: by matrix cosJ _th-μevery a line regard R as ^ka nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node in space, and use χ (μ-(μ) ^μ) ∈ R ^mmean algorithm is polymerized to L _μclass, if cosJ _th-μthe capable j class that belongs to of i, the y in nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node Y _ialso be divided into j class, thus the x in nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node X _ijust be divided into j class; (1) for optimized network structure Control System with Delay (4) wherein by inference, if there is positive definite symmetric matrices P, and Q, R and Arbitrary Matrix A, B, C meets following MATRIX INEQUALITIES and is

$\underset{\mathrm{\&mu;}+d\left(\mathrm{\&omega;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{{\mathrm{\&Pi;}}_{\&infin;}}=\left\{\begin{array}{ccccc}\stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{11}}& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{12}}& 0.5{\stackrel{\&OverBar;}{X}}^{\mathrm{\&mu;T}}& P\stackrel{\&OverBar;}{X}-Y& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{14}}\\ 0& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{22}}& 0.5{\stackrel{\&OverBar;}{X_1}}^{\mathrm{\&mu;T}}& P\stackrel{\&OverBar;}{X}-\mathrm{\&omega;}& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{24}}\\ 0& 0& 0& -\frac{1}{d_{\max }}R& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{34}}\\ 0& 0& 0& 0& \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{44}}\end{array}\right\}<0$

, $\underset{\mathrm{\&mu;}-d_{\max }}{\overset{\frac{1}{d_{\max }}-2}{\mathrm{\&Sigma;}}}\stackrel{\&OverBar;}{\mathrm{\&Pi;}}=\left\{\begin{array}{c}\stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{11}}=d_{\max }\left\{\right[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R(\stackrel{\&OverBar;}{X}-I)]-P+[(d_{\max }-d_{\min }+1)+Q+Y+Y^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{C^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{C}\left]\right\}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{12}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+d_{\max }\left[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{X_1}\right]+{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}-Y+\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Z_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{14}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{Y}+\stackrel{\&OverBar;}{Y_1})+d_{\max }\left[{(\stackrel{\&OverBar;}{X}-I)}^{\mathrm{\&mu;T}}R\stackrel{\&OverBar;}{Y_1}\right]+Q^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{Z^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{22}}\text{=}\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{X}+\stackrel{\&OverBar;}{X_1})+d_{\max }\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{X_1}-Q-\mathrm{\&omega;}{\mathrm{\&omega;}}^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{24}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P(\stackrel{\&OverBar;}{Y}+\stackrel{\&OverBar;}{Y_1})+\stackrel{\&OverBar;}{{X_1}^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }\stackrel{\&OverBar;}{X_1^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{Y}-Q^{\mathrm{\&mu;T}}+\stackrel{\&OverBar;}{{Z_1}^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q_1}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{34}}=0.5\stackrel{\&OverBar;}{X^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}-Q^{\mathrm{\&mu;T}}\\ \stackrel{\&OverBar;}{{\mathrm{\&rho;}}_{44}}=\stackrel{\&OverBar;}{Y^{\mathrm{\&mu;T}}}P\stackrel{\&OverBar;}{Y}+d_{\max }\stackrel{\&OverBar;}{Y^{\mathrm{\&mu;T}}}R\stackrel{\&OverBar;}{Y}\text{+}\stackrel{\&OverBar;}{Q^{\mathrm{\&mu;T}}}\stackrel{\&OverBar;}{Q}-{\mathrm{\&gamma;}}^{2}I\end{array}\right.$

Wherein ω (μ) _∞=0 o'clock, under the progressive stable and zero initial condition of optimized network structure Control System with Delay (4), meet $\underset{\mathrm{\&mu;}-d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2{\left|\right|{\mathrm{\&gamma;}}_e\left(\mathrm{\&mu;}\right)\left|\right|}_{e-2}^2<\underset{\mathrm{\&mu;}+d\left(\mathrm{\&mu;}\right)}{\overset{\mathrm{\&mu;}-1}{\mathrm{\&Sigma;}}}\mathrm{\&gamma;}{\left|\right|\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\left|\right|}_{e-2}^2,\mathrm{\&omega;}{\left(\mathrm{\&mu;}\right)}_{\&infin;}\&NotEqual;0.$

2. the failure prediction method based under intelligent power distribution environment according to claim 1, it is characterized in that, step 1 is the preprocessing process of nonsingular cloud terminal network structure Control System with Delay networking failure prediction, all numerical value nonsingular cloud terminal network structure Control System with Delay networking failure prediction in nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction set of node X is transformed to based on nonsingular cloud terminal network structure Control System with Delay networking failure prediction, realize normalization and the networking of structure Control System with Delay of nonsingular cloud terminal network structure Control System with Delay networking failure prediction, step 2～6th, the process of the electric structure sensor model algorithm process of density sensitivity based on nonsingular cloud terminal network structure Control System with Delay networking failure prediction, wherein, step 2 is weighed based on distance between nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node by average weighted hamming distance, builds on this basis the similarity measurement of density sensitivity, step 4 is by solving eigenwert and the proper vector of nonsingular cloud terminal network structure Control System with Delay networking fault perceptual signal prediction node matrix equation, structural attitude vector space, thus realize the spectrum mapping of former nonsingular cloud terminal network structure Control System with Delay networking failure prediction space to low-dimensional linear measurement space, in the low-dimensional linear measurement space of step 6 after mapping, adopt traditional χ (μ-(μ) ^μ) ∈ R ^mmean algorithm cloud theoretical model.