CN110119690B - Parallel reactor vibration sensitive area selection method based on CRP and RQA - Google Patents
Parallel reactor vibration sensitive area selection method based on CRP and RQA Download PDFInfo
- Publication number
- CN110119690B CN110119690B CN201910316827.6A CN201910316827A CN110119690B CN 110119690 B CN110119690 B CN 110119690B CN 201910316827 A CN201910316827 A CN 201910316827A CN 110119690 B CN110119690 B CN 110119690B
- Authority
- CN
- China
- Prior art keywords
- point
- measuring point
- reactor
- rqa
- measuring points
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/30—Reactive power compensation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Monitoring And Testing Of Nuclear Reactors (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a method for selecting a vibration sensitive area of a shunt reactor based on CRP and RQA, which comprises the following steps: step 1: measuring points are arranged inside and outside the reactor, vibration signals of the measuring points are obtained, and an internal measuring point D with the largest vibration amplitude is selected as a reactor magnetic field concentrated area; step 2: constructing CRP of vibration signals between the internal measuring points D and the external measuring points; and step 3: calculating RQA parameters of vibration signals of the internal measuring point D and the external measuring points; and 4, step 4: constructing the RQA parameters of the internal measuring points D and the external measuring points in the step 3 into characteristic matrixes by adopting a principal component method; and 5: calculating a correlation coefficient between the internal measuring point D and each external measuring point feature matrix; step 6: and judging the vibration sensitive area of the reactor by combining the CRP and the correlation coefficient value. The invention combines CRP and RQA, visually judges the sensitivity of the measuring point from CRP, and quantitatively compares the sensitivity of different measuring points by the correlation coefficient between characteristic matrixes constructed by RQA parameters.
Description
Technical Field
The invention belongs to the field of power equipment vibration signal analysis, and particularly relates to a parallel reactor vibration sensitive area selection method based on CRP and RQA.
Background
The high-voltage shunt reactor plays an important role in stabilizing voltage and reactive compensation in a power system, and is key equipment for smart grid construction and global energy internet strategy. The dynamic characteristics of the mechanical structure of the reactor body are influenced by the change of the parameters of the mechanical structure of the parallel reactor. Because the high-voltage shunt reactor is large in size and complex in structure, the attenuation degree of the internal vibration signal on the surface of the oil tank is greatly different. In field test, a position which can best reflect the vibration characteristic inside the reactor needs to be determined on the surface of the oil tank for measuring point arrangement. Therefore, the correlation between the vibration signal on the surface of the oil tank of the reactor and the internal vibration signal needs to be researched, so that the vibration sensitive area of the reactor is determined, and a foundation is laid for the subsequent health monitoring and fault diagnosis of the reactor. Because the traditional signal analysis is based on a linear theory and the vibration signal of the reactor is nonlinear, the traditional method cannot obtain a good analysis effect. CRP (cross recursion plot) is an important means for characterizing basic characteristics of dynamic systems, nonlinear systems and chaotic systems, and can be used for measuring the dynamic characteristics of the systems through graphs and reproducing the recursive behaviors among the systems.
Disclosure of Invention
The invention aims to: aiming at the problems, the invention provides a parallel reactor vibration sensitive area selection method based on CRP and RQA, wherein CRP compares the similarity of phase tracks of different power systems in a phase space, is suitable for the comparison of dynamic characteristics among different measuring points, and Recursive Quantitative Analysis (RQA) parameters extracted from CRP can quantitatively describe the dynamic characteristics of the system; aiming at the nonlinear characteristics of the vibration signals of the parallel reactor, CRP and RQA are introduced into the vibration signal analysis of the reactor, the extracted RQA parameter is used for calculating a correlation coefficient by adopting a principal component method, and the correlation among the vibration signals is comprehensively measured.
The technical scheme is as follows: in order to realize the purpose of the invention, the technical scheme adopted by the invention is as follows: a method for selecting a vibration sensitive area of a shunt reactor based on CRP and RQA comprises the following steps:
step 1: respectively arranging P measuring points and K measuring points inside and outside the reactor, acquiring vibration signals of the measuring points inside and outside the reactor, and selecting an inside measuring point D with the largest vibration amplitude as a reactor magnetic field concentration area;
and 2, step: respectively constructing a cross recursion graph CRP of the reactor vibration signals between the internal measuring points D and each external measuring point;
and step 3: calculating RQA characteristic parameters of the vibration signals of the internal measuring points D and each external measuring point reactor;
and 4, step 4: constructing the RQA characteristic parameters of the internal measuring points D and each external measuring point in the step 3 into a characteristic matrix H by adopting a principal component method;
and 5: calculating a correlation coefficient between the internal measuring point D and each external measuring point feature matrix according to the feature matrix constructed in the step 4;
step 6: judging a reactor vibration sensitive area by combining the CRP in the step 2 and the correlation coefficient value in the step 5; if the diagonal line structure formed by recursion points in the CRP of the internal measuring point D and the external measuring point W (W is 1,2, … K) is continuous, the value range of the correlation coefficient of the internal measuring point D and the external measuring point W is between p and 1, and p belongs to [0,1], the external measuring point W is located in the reactor vibration sensitive area, and the area surrounded by W (W is less than K) measuring points meeting the condition is used as the reactor vibration sensitive area.
Further, in step 1, setting internal and external measuring points, acquiring vibration signals of all measuring points, and determining a magnetic field concentrated region in the reactor, the steps are as follows:
step 1.1: k piezoelectric acceleration sensors are respectively fixed at the top end of the reactor oil tank, the middle part and the bottom part of the tank wall through magnetic bases to obtain vibration data of the surface of the oil tank;
step 1.2: p piezoelectric sensors are arranged in the internal position of the reactor through a valve at the top of the oil storage tank and used for measuring vibration signals of the positions of the iron core cake and the iron yoke accessory, and the internal position comprises: the middle part of the upper yoke of the iron core, the outer side of the middle part of the side yoke, the upper part of the magnetic shielding coil, the middle part of the upper surface of the side yoke and the upper surface of the iron core cake; the type of the sensor is AC 102-1A;
step 1.3: and measuring to obtain vibration signals of the internal measuring point and the external measuring point of the reactor, and selecting the internal measuring point D with the largest vibration amplitude as a reactor magnetic field concentrated region according to a time domain signal diagram.
Further, in step 2, constructing a cross recursion graph CRP of the reactor vibration signal between the internal measuring point D and each external measuring point, and the steps are as follows:
step 2.1: according to the Takens theorem, for a given time sequence x (i) of the original vibration signal, i is 1,2, … n, n represents the number of sampling pointsSelecting embedding dimension m and delay time tau to carry out phase space reconstruction, reconstructing time sequence x of signal i Comprises the following steps:
x i ＝{u i ,u i+τ ,…,u i+(m-1)τ } (1)
where i is 1,2, …, N is N- (m-1) τ, N represents the number of points in the phase space, and u represents i Representing constituent elements of a reconstructed signal obtained after phase space reconstruction of an original signal;
step 2.2: respectively obtaining the reconstruction time sequences of the internal measuring point D and the external measuring point W according to the step 2.1, and obtaining the phase track of each measuring point on the phase space;
step 2.3: calculating the jth point x on the W-phase space track of the external measuring point j The ith point x on the space track with the internal measuring point D i A distance S between ij ：
S ij ＝||x i -x j || (2)
Wherein, | | · | | is a Euclidean norm of a time series, i ═ 1,2, …, N, j ═ 1,2, …, N;
step 2.4: according to the distance S obtained in step 2.3 ij Calculating a recursive value R ij The formula is as follows:
R ij ＝θ(r-S ij ) (3)
wherein r is a recursive threshold and θ (—) is a Heaviside function;
step 2.5: drawing a cross recursion graph of the internal measuring point D and the external measuring point W by taking the point i as an abscissa and the point j as an ordinate; the recursion state between the point i and the point j is represented by a black point or a white point in a cross-recursion graph (CRP); if R is ij When the value is 0, marking the position of (i, j) in the CRP as a white point, and indicating that no recursion relation exists between the point i and the point j; when R is ij When the value is 1, the position corresponding to the (i, j) is a black point, which indicates that a recursion relation exists between the point i and the point j; therefore, the dynamic characteristics of the system can be accurately displayed, and as the difference of signals is gradually increased, the phenomena that the density of recursion points is reduced and the diagonal structure is weakened occur in CRP;
step 2.6: and repeating the steps 2.3-2.5 until the cross recursion graph between the inner measuring point D and each outer measuring point is completely drawn.
Furthermore, according to the construction process of the cross recursion graph algorithm, the embedding dimension m and the delay time tau have great influence on the recursion phenomenon in the recursion graph, and the value of the embedding dimension m and the delay time tau determines whether the reconstructed phase space can keep the characteristics of the original system; the invention adopts a differential entropy method to simultaneously optimize the embedding dimension m and the delay time tau, and preferably, the embedding dimension is selected to be 3 and the delay time is selected to be 6.
Further, in step 3, calculating RQA characteristic parameters of the reactor vibration signals of the internal measuring point D and each external measuring point, wherein the method comprises the following steps:
step 3.1: selecting 4 commonly used RQA parameters as characteristic quantity calculation, including: recursion Rate (RR), determination rate (DET), average diagonal length (L), Entropy (ENTR);
the Recursion Rate (RR) is a density of recursion points in the CRP, and is used for reflecting the dispersion degree of the recursion points, and the expression is as follows:
the determination rate (DET) is the proportion of recursive points forming a 45-degree diagonal structure, and the DET can measure the regularity of the system, and the expression is as follows:
where P (l) is the probability density of the diagonal distribution of length l, l min Represents the minimum value of the diagonal length;
the average diagonal length (L) is the average length of the diagonal, and represents the average length of time consumed when the phase tracks are close to each other, and the expression is as follows:
the Entropy (ENTR) refers to information entropy distributed by a 45-degree diagonal structure and represents the complexity of a system, and the expression is as follows:
ENTR＝-∑P(l)lnP(l) (7)
step 3.2: and (4) calculating RQA characteristic parameters of the vibration signals of the electric reactors at the internal measuring points D and the K external measuring points according to the formulas (4) to (7) in the step 3.1, wherein the RQA characteristic parameters are used for analyzing the correlation of the vibration signals at different measuring points of the electric reactors, and the RQA parameters can correspondingly change along with the gradual increase of the difference of the signals.
Further, in step 4, constructing the RQA feature parameters of the internal measurement points D and each external measurement point in step 3 as a feature matrix H by using a principal component method, respectively, and performing the following steps:
step 4.1: calculating n groups of RQA parameters corresponding to the n groups of vibration signals of the measuring points according to the step 3;
step 4.2: constructing a characteristic matrix H of the measuring points through the n groups of RQA parameters obtained in the step 4.1, wherein the expression is as follows:
in the formula, RR n 、DET n 、L n 、ENTR n Respectively representing the recursion rate, the determination rate, the average diagonal length and the entropy in the nth group of RQA parameters;
step 4.3: and repeating the steps 4.1-4.2 until the feature matrixes H of the internal measuring points D and the K external measuring points are completely constructed.
Further, in step 5, a correlation coefficient r between the inside measuring point D and each outside measuring point feature matrix H is calculated ij The method comprises the following steps:
step 5.1: and (3) carrying out singular value decomposition on the covariance matrix of the feature matrix H, wherein the formula is as follows:
HH T ＝U∑ 2 U T (9)
where U is an orthogonal matrix, Sigma is a singular value matrix, Sigma 1 Sum-sigma 2 Is a sub-matrix of sigma, and the k-th column vector in the matrix sigma is defined as the k-th principal component H of the feature matrix H k Main component H k Corresponding to a singular value σ k Selecting a first main component construction standard quantity PC, and expressing as follows:
PC＝U 1 H T (11)
in the formula of U 1 As a singular value sub-matrix sigma 1 A corresponding orthogonal matrix;
step 5.2: respectively calculating the PC value PC of the internal measuring point D according to the step 5.1 D PC value PC from external point W (W ═ 1,2, … K) W ；
Step 5.3: the correlation coefficient r between the feature matrices of the measuring points D and W is obtained according to the formula (12) ij The formula is as follows:
wherein E (PC) represents the expectation of PC, D (PC) represents the variance of PC, and the correlation coefficient r ij The value range is between 0 and 1; the magnitude of the correlation coefficient value reflects the strength of the signal correlation degree between the two measuring points, and the larger the correlation coefficient is, the higher the similarity of the signals of the two measuring points is, the smaller the loss of the signals of the measuring points in the reactor at the position of the outer surface is;
step 5.4: and 5.3, repeating the step until the correlation coefficient calculation between the feature matrixes of all the external measuring points and the internal measuring point D is completed.
Has the advantages that: compared with the prior art, the technical scheme of the invention has the following beneficial technical effects: the invention combines CRP and RQA methods, not only can visually judge the sensitivity of the measuring point from CRP, but also can quantitatively compare the sensitivity of different measuring points by comparing the correlation coefficient between the characteristic matrix H constructed by RQA parameters. The research result provides a basis for determining the vibration sensitive area of the extra-high voltage shunt reactor and selecting a vibration measuring point from the nonlinear dynamics angle.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a position of an external measuring point of a shunt reactor;
FIG. 3 is a shunt reactor tank top valve;
FIG. 4(a) is a time domain signal diagram of measuring points No. 1-5 in the shunt reactor;
FIG. 4(b) is a time domain signal diagram of measuring points No. 1-8 outside the shunt reactor;
FIG. 5-1 is a cross-recursion plot of outer station number 1 and inner station number 5;
FIG. 5-2 is a cross-recursion plot of outer station number 2 and inner station number 5;
5-3 is a cross-recursion plot of outer station number 4 and inner station number 5;
5-4 are cross-recursion graphs consisting of outer station No. 5 and inner station No. 5;
5-5 are cross-recursion graphs consisting of outer station No. 7 and inner station No. 5;
fig. 5-6 are cross-recursion graphs of outer point No. 8 and inner point No. 5.
Detailed Description
The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.
The invention relates to a selection method of a vibration sensitive area of a shunt reactor based on CRP and RQA, and the specific flow is shown in figure 1, and the method comprises the following steps:
step 1: respectively arranging 5 measuring points and 8 measuring points inside and outside the reactor, acquiring vibration signals of the measuring points inside and outside the reactor, and selecting the measuring point inside with the largest vibration amplitude as a reactor magnetic field concentration area; the implementation steps are as follows:
step 1.1: respectively fixing 8 piezoelectric acceleration sensors at the top end of the reactor oil tank, the middle part of the tank wall and the bottom part of the tank wall through magnetic bases to obtain vibration data of the surface of the oil tank, wherein measuring points are arranged as shown in figure 2;
step 1.2: 5 piezoelectric sensors enter the reactor through a valve at the top of an oil storage tank in the figure 3 to measure vibration signals of the positions of iron core cakes and iron yoke accessories, the positions of internal measuring points are shown in the table 1, and the measuring points are arranged as shown in the figure 3; the type of the sensor is AC 102-1A;
TABLE 1
Step 1.3: measuring to obtain vibration signals of an internal measuring point and an external measuring point of the reactor, and selecting the internal measuring point 5 with the largest vibration amplitude as a reactor magnetic field concentrated region according to a time domain signal diagram;
step 2: respectively constructing a cross recursion graph CRP of the reactor vibration signals between the internal measuring points 5 and each external measuring point; the implementation steps are as follows:
step 2.1: according to the Takens theorem, for a time sequence x (i) of a given original vibration signal, i is 1,2, … n, n represents the number of sampling points, an embedding dimension m and a delay time tau are selected for phase space reconstruction, in the embodiment, the embedding dimension is selected to be 3, the delay time is selected to be 6, and the time sequence x of a reconstructed signal is selected i Comprises the following steps:
x i ＝{u i ,u i+τ ,…,u i+(m-1)τ } (1)
where i is 1,2, …, N is N- (m-1) τ, N represents the number of points in the phase space, and u represents i Representing constituent elements of a reconstructed signal obtained after phase space reconstruction of an original signal;
step 2.2: obtaining a reconstruction time sequence of the internal measuring point 5 and the external measuring point W (W is 1,2, … 8) according to the step 2.1, respectively, to obtain a phase trajectory of each measuring point on a phase space;
step 2.3: calculating the j point x on the W-phase space track of the external measuring point j The ith point x on the space track with the internal measuring point 5 i A distance S between ij ：
S ij ＝||x i -x j || (2)
Wherein, | | · | | is a Euclidean norm of a time series, i ═ 1,2, …, N, j ═ 1,2, …, N;
step 2.4: according to the distance S obtained in step 2.3 ij Calculating a recursive value R ij The formula is as follows:
R ij ＝θ(r-S ij ) (3)
wherein r is a recursion threshold, and theta is a Heaviside function;
step 2.5: drawing a cross recursion graph of the internal measuring point 5 and the external measuring point W by taking the point i as an abscissa and the point j as an ordinate; the recursion state between the point i and the point j is represented by a black point or a white point in a cross-recursion graph (CRP); if R is ij When the value is 0, marking the position of (i, j) in the CRP as a white point, and indicating that no recursion relation exists between the point i and the point j; when R is ij When the value is 1, the position corresponding to the (i, j) is a black point, which indicates that a recursion relation exists between the point i and the point j; therefore, the dynamic characteristics of the system can be accurately displayed, and as the difference of signals is gradually increased, the phenomena that the density of recursion points is reduced and the diagonal structure is weakened occur in CRP;
step 2.6: and repeating the steps 2.3-2.5 until the cross recursion graph between the inner measuring point 5 and each outer measuring point is completely drawn.
Fig. 4(a) is a time domain signal diagram of measuring points 1-5 inside the shunt reactor, and fig. 4(b) is a time domain signal diagram of measuring points 1-8 outside the shunt reactor. As can be seen from fig. 4(a), the vibration amplitude of the internal measuring point 5 of the shunt reactor, i.e. the upper surface of the iron core cake, is the maximum, and the amplitude is 5 to 10 times of the other measuring points in the shunt reactor, which indicates that the iron core cake is a magnetic field concentration area of the reactor, and the vibration of the surface of the reactor oil tank is mainly generated by the vibration attenuation at the upper surface of the iron core cake; and determining the vibration sensitive area of the surface of the reactor oil tank by comparing the correlation between the external measuring point signal and the measuring point 5, namely the iron core cake upper surface signal. Because the vibration signals at the symmetrical positions of the shunt reactor are identical in propagation rule and extremely high in signal correlation, corresponding cross recursion graphs are very similar.
FIG. 5 is a cross-recursion diagram of the inner 5 measuring points and a part of the outer measuring points. FIG. 5-1 is a cross-recursion diagram formed by the outer measuring point No. 1 and the inner measuring point No. 5, FIG. 5-2 is a cross-recursion diagram formed by the outer measuring point No. 2 and the inner measuring point No. 5, FIG. 5-3 is a cross-recursion diagram formed by the outer measuring point No. 4 and the inner measuring point No. 5, FIG. 5-4 is a cross-recursion diagram formed by the outer measuring point No. 5 and the inner measuring point No. 5, FIG. 5-5 is a cross-recursion diagram formed by the outer measuring point No. 7 and the inner measuring point No. 5, and FIG. 5-6 is a cross-recursion diagram formed by the outer measuring point No. 8 and the inner measuring point No. 5. The strength of the correlation of different measuring points can be reflected visually by the change of the diagonal line structure in the cross recursion graph, the similarity degree of signals of the external measuring points and the internal measuring points can be judged according to the continuity of the diagonal lines in the recursion graph, the similarity degree of the external measuring points 5 is greater than that of other external measuring points, and the fact that the external measuring points 5 belong to a vibration sensitive area can be judged.
And step 3: calculating RQA characteristic parameters of the vibration signals of the internal measuring points 5 and each external measuring point reactor; the method comprises the following specific steps:
step 3.1: selecting 4 commonly used RQA parameters as characteristic quantity calculation, including: recursion Rate (RR), determination rate (DET), average diagonal length (L), Entropy (ENTR);
the Recursion Rate (RR) is a density of recursion points in the CRP, and is used for reflecting the dispersion degree of the recursion points, and the expression is as follows:
the determination rate (DET) is the proportion of recursive points forming a 45-degree diagonal structure, and the DET can measure the regularity of the system, and the expression is as follows:
where P (l) is the probability density of the diagonal distribution of length l, l min Represents the minimum value of the diagonal length;
the average diagonal length (L) is the average length of the diagonal, and represents the average length of time consumed when the phase tracks are close to each other, and the expression is as follows:
the Entropy (ENTR) refers to information entropy distributed by a 45-degree diagonal structure and represents the complexity of a system, and the expression is as follows:
ENTR＝-∑P(l)lnP(l) (7)
step 3.2: and (4) calculating RQA characteristic parameters of the reactor vibration signals of the internal measuring points 5 and the 8 external measuring points according to the formulas (4) to (7) in the step 3.1, wherein the RQA characteristic parameters are used for analyzing the correlation of the vibration signals of different measuring points of the reactor, and the RQA parameters can correspondingly change along with the gradual increase of the signal difference.
In order to quantitatively describe the recursion phenomenon in fig. 5, RQA parameters of the vibration signals at the external points 1 to 8 of the high-voltage shunt reactor were calculated, and the results are shown in table 2.
TABLE 2
And 4, step 4: respectively constructing RQA characteristic parameters of the internal measuring points 5 and each external measuring point in the step 3 into a characteristic matrix H by adopting a principal component method; the method comprises the following specific steps:
step 4.1: calculating n groups of RQA parameters corresponding to the n groups of vibration signals of the measuring point according to the step 3;
step 4.2: constructing a characteristic matrix H of the measuring points through the n groups of RQA parameters obtained in the step 4.1, wherein the expression is as follows:
in the formula, RR n 、DET n 、L n 、ENTR n Respectively representing the recursion rate, the determination rate, the average diagonal length and the entropy in the nth group of RQA parameters;
step 4.3: and repeating the steps 4.1-4.2 until the feature matrix H of the internal measuring point 5 and the 8 external measuring points is completely constructed.
And 5: calculating a correlation coefficient between the internal measuring point 5 and each external measuring point feature matrix according to the feature matrix constructed in the step 4; the method comprises the following specific steps:
step 5.1: and (3) carrying out singular value decomposition on the covariance matrix of the feature matrix H, wherein the formula is as follows:
HH T ＝U∑ 2 U T (9)
where U is an orthogonal matrix, Sigma is a singular value matrix, Sigma 1 Sum sigma 2 Is a sub-matrix of sigma, and the k-th column vector in the matrix sigma is defined as the k-th principal component H of the feature matrix H k Main component H k Corresponding to a singular value σ k Selecting a first main component construction standard quantity PC, and expressing as follows:
PC＝U 1 H T (11)
in the formula of U 1 As a singular value sub-matrix sigma 1 A corresponding orthogonal matrix;
step 5.2: respectively calculating the PC value PC of the internal measuring point 5 according to the step 5.1 D PC value PC from external point W (W is 1,2, … 8) W ；
Step 5.3: the correlation coefficient r between the feature matrices of the inner measurement point 5 and the outer measurement point W is obtained from equation (12) ij The formula is as follows:
wherein E (PC) represents the expectation of PC, D (PC) represents the variance of PC, and the correlation coefficient r ij The value range is between 0 and 1; the magnitude of the correlation coefficient value reflects the strength of the signal correlation degree between the two measuring points, and the larger the correlation coefficient is, the higher the similarity of the signals of the two measuring points is, the smaller the loss of the signals of the measuring points in the reactor at the position of the outer surface is;
step 5.4: and (5.3) repeating the step until the calculation of the correlation coefficients between the feature matrixes of all the external measuring points and the internal measuring points 5 is completed.
Step 6: judging a reactor vibration sensitive area by combining the CRP in the step 2 and the correlation coefficient value in the step 5; if the diagonal structure formed by recursion points in CRP of the internal measuring point 5 and the external measuring point W (W is 1,2, … 8) is continuous, the correlation coefficient range of the internal measuring point 5 and the external measuring point W is between p and 1, and p belongs to [0,1], the external measuring point W is positioned in a reactor vibration sensitive area, and the area surrounded by W (W is less than 8) measuring points meeting the condition is used as the reactor vibration sensitive area.
Table 3 shows the correlation coefficients of the outer points 1 to 8 and the inner points 5. As can be seen from the calculation results in Table 3, the correlation coefficient values corresponding to the vibration signals at different measuring points also show distinct results. As the similarity of the two sets of signals decreases, the correlation coefficient also decreases. r is ij Can be used as a comprehensive evaluation index for measuring the similarity of two groups of signals.
TABLE 3
The RQA measurement corresponding to the external No. 5 and No. 6 measuring points is obviously superior to other measuring points, and the RQA measurement shows that the vibration signal attenuation at the position is minimum, the similarity degree with the signal on the surface of the iron core cake inside the reactor is highest, and the RQA measurement can be used as a vibration sensitive area of the high-voltage parallel reactor. The reason for this is that the measuring points No. 5 and No. 6 are located in the middle of the front surface of the oil tank, and at this position, only the vibration generated by the radial electromagnetic force of the winding perpendicular to the wall of the oil tank is transmitted to the surface of the oil tank through the insulating oil, so the measuring points should be considered preferentially when being arranged. The measuring points No. 1 and No. 3 are positioned at the top of the oil tank, and the signal attenuation degree is more serious than that of the measuring points No. 5 and No. 6, but is better than that of other measuring points. The measuring point No. 2 is positioned on the casing, the signal attenuation degree is more serious than that of the measuring point No. 1 and the measuring point No. 3, and the measuring points are not recommended to be arranged at the position. 7. No. 8 measurement points are arranged on a reinforcing rib structure on the surface of the oil tank, and compared with a normal flat plate structure, a vibration signal is seriously lost at the reinforcing rib, so that the monitoring of a reactor signal is not facilitated, and the reinforcing rib structure is avoided when the measurement points are arranged. No. 4 measuring point is located the reactor base position, and the signal attenuation degree is great, and the vibration characteristic of signal is difficult to be analyzed here, so the bottom position of reactor also should not be considered when setting up the measuring point.
Claims (6)
1. A parallel reactor vibration sensitive area selection method based on CRP and RQA is characterized in that: the method comprises the following steps:
step 1: respectively arranging P measuring points and K measuring points inside and outside the reactor, acquiring vibration signals of the measuring points inside and outside the reactor, and selecting an inside measuring point D with the largest vibration amplitude as a reactor magnetic field concentration area;
step 2: respectively constructing a cross recursion graph CRP of the reactor vibration signals between the internal measuring points D and each external measuring point;
and step 3: calculating RQA characteristic parameters of the vibration signals of the internal measuring points D and each external measuring point reactor;
and 4, step 4: constructing the RQA characteristic parameters of the internal measuring points D and each external measuring point in the step 3 into a characteristic matrix H by adopting a principal component method;
and 5: calculating a correlation coefficient between the internal measuring point D and each external measuring point feature matrix according to the feature matrix constructed in the step 4;
step 6: judging a reactor vibration sensitive area by combining the CRP in the step 2 and the correlation coefficient value in the step 5; and if the diagonal structure formed by recursion points in CRP of the internal measuring point D and the external measuring point W is continuous, W is 1,2, … K, the value range of the correlation coefficient of the internal measuring point D and the external measuring point W is between p and 1, and p belongs to [0,1], the external measuring point W is positioned in a reactor vibration sensitive area, the area surrounded by W measuring points meeting the condition is used as the reactor vibration sensitive area, and W is less than K.
2. The method for selecting the vibration sensitive area of the shunt reactor based on the CRP and the RQA as claimed in claim 1, wherein the method comprises the following steps: in the step 1, setting internal and external measuring points, acquiring vibration signals of all measuring points, and determining a magnetic field concentrated region in the reactor, wherein the steps are as follows:
step 1.1: k piezoelectric acceleration sensors are respectively fixed at the top end of the reactor oil tank, the middle part and the bottom part of the tank wall through magnetic bases to obtain vibration data of the surface of the oil tank;
step 1.2: p piezoelectric sensors are arranged in the internal position of the reactor through a valve at the top of the oil storage tank and used for measuring vibration signals of the positions of the iron core cake and the iron yoke accessory, and the internal position comprises: the middle part of the upper yoke of the iron core, the outer side of the middle part of the side yoke, the upper part of the magnetic shielding coil, the middle part of the upper surface of the side yoke and the upper surface of the iron core cake;
step 1.3: and measuring to obtain vibration signals of the internal measuring points and the external measuring points of the reactor, and selecting the internal measuring points D with the largest vibration amplitude as the concentrated area of the magnetic field of the reactor according to the time domain signal diagram.
3. A method for selecting a vibration sensitive area of a shunt reactor based on CRP and RQA as claimed in claim 1 or 2, characterized by: in step 2, constructing a cross recursion graph CRP of the reactor vibration signal between the internal measuring point D and each external measuring point, and comprising the following steps:
step 2.1: according to Takens' theorem, given the time sequence x (i) of the original vibration signal, i is 1,2, … n, n represents the number of sampling points, the embedding dimension m and the delay time tau are selected for phase space reconstruction, and the time sequence x of the signal is reconstructed i Comprises the following steps:
x i ＝{u i ,u i+τ ,…,u i+(m-1)τ } (1)
where i is 1,2, …, N is N- (m-1) τ, N represents the number of points in the phase space, and u represents i Representing constituent elements of a reconstructed signal obtained after phase space reconstruction of an original signal;
step 2.2: respectively obtaining the reconstruction time sequences of the internal measuring point D and the external measuring point W according to the step 2.1, and obtaining the phase track of each measuring point on the phase space;
step 2.3: calculating the jth point x on the W-phase space track of the external measuring point j The ith point x on the space track with the internal measuring point D i A distance S between ij ：
S ij ＝||x i -x j || (2)
Wherein, | | · | | is a Euclidean norm of a time series, i ═ 1,2, …, N, j ═ 1,2, …, N;
step 2.4: according to the distance S obtained in step 2.3 ij Calculating a recursive value R ij The formula is as follows:
R ij ＝θ(r-S ij ) (3)
wherein r is a recursion threshold, and theta is a Heaviside function;
step 2.5: drawing a cross recursion graph of the internal measuring point D and the external measuring point W by taking the point i in the step 2.3 as an abscissa and the point j as an ordinate; the recursion state between the point i and the point j is represented by a black point or a white point in the cross recursion graph CRP; if R is ij The value is 0, and the position of (i, j) in the CRP is marked as a white point, which indicates that no recursion relation exists between the point i and the point j; when R is ij When the value is 1, the position corresponding to the (i, j) is a black point, which indicates that a recursion relation exists between the point i and the point j;
step 2.6: and repeating the steps 2.3-2.5 until the cross recursion graph between the inner measuring point D and each outer measuring point is completely drawn.
4. The method for selecting the vibration sensitive area of the shunt reactor based on the CRP and the RQA as claimed in claim 1, wherein the method comprises the following steps: in step 3, calculating RQA characteristic parameters of the reactor vibration signals of the internal measuring points D and each external measuring point, wherein the method comprises the following steps:
step 3.1: selecting 4 commonly used RQA parameters as characteristic quantity calculation, including: a recursion rate RR, a determination rate DET, an average diagonal length L and an entropy ENTR;
the recurrence rate RR is the density of recurrence points in CRP, and is used for reflecting the dispersion degree of the recurrence points, and the expression is as follows:
the determination rate DET refers to the proportion of recursive points forming a 45-degree diagonal structure, and the DET can measure the regularity of a system, and the expression is as follows:
where P (l) is the probability density of the diagonal distribution of length l, l min Represents the minimum value of the diagonal length;
the average diagonal length L is the average length of the diagonal, and represents the average length of time consumed when the phase tracks are close to each other, and the expression is as follows:
the entropy ENTR is information entropy distributed by a 45-degree diagonal structure and represents the complexity of a system, and the expression is as follows:
ENTR＝-∑P(l)lnP(l) (7)
step 3.2: and (4) calculating RQA characteristic parameters of the vibration signals of the reactors at the internal measuring points D and the K external measuring points according to the formulas (4) to (7) in the step 3.1.
5. The method for selecting the vibration sensitive area of the shunt reactor based on the CRP and the RQA as claimed in claim 1,2 or 4, wherein the method comprises the following steps: in step 4, respectively constructing the RQA characteristic parameters of the internal measuring points D and each external measuring point in step 3 into a characteristic matrix H by using a principal component method, which comprises the following steps:
step 4.1: calculating n groups of RQA parameters corresponding to the n groups of vibration signals of the measuring points according to the step 3;
step 4.2: constructing a characteristic matrix H of the measuring points through the n groups of RQA parameters obtained in the step 4.1, wherein the expression is as follows:
in the formula, RR n 、DET n 、L n 、ENTR n Respectively representing the recursion rate, the determination rate and the average diagonal angle in the Nth group of RQA parametersLine length, entropy;
step 4.3: and repeating the steps 4.1-4.2 until the feature matrixes H of the internal measuring points D and the K external measuring points are completely constructed.
6. The method for selecting the vibration sensitive area of the shunt reactor based on CRP and RQA according to claim 1,2 or 4, characterized by comprising the following steps: in step 5, calculating a correlation coefficient r between the internal measuring point D and each external measuring point characteristic matrix H ij The method comprises the following steps:
step 5.1: and (3) carrying out singular value decomposition on the covariance matrix of the feature matrix H, wherein the formula is as follows:
HH T ＝U∑ 2 U T (9)
where U is an orthogonal matrix, Sigma is a singular value matrix, Sigma 1 Sum-sigma 2 Is a sub-matrix of sigma, and the k-th column vector in the matrix sigma is defined as the k-th principal component H of the feature matrix H k Main component H k Corresponding to a singular value σ k Selecting a first main component construction standard quantity PC, and expressing as follows:
PC＝U 1 H T (11)
in the formula of U 1 As a singular value sub-matrix sigma 1 A corresponding orthogonal matrix;
step 5.2: respectively calculating the PC value PC of the internal measuring point D according to the step 5.1 D PC value PC of external measuring point W W ，W＝1,2,…K；
Step 5.3: the correlation coefficient r between the feature matrices of the measuring points D and W is obtained according to the formula (12) ij The formula is as follows:
wherein E (PC) represents the expectation of PC, D (PC) represents the variance and phase of PCCoefficient of correlation r ij The value range is between 0 and 1;
step 5.4: and (5.3) repeating the step until the calculation of the correlation coefficients between the feature matrixes of all the external measuring points and the internal measuring point D is completed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910316827.6A CN110119690B (en) | 2019-04-19 | 2019-04-19 | Parallel reactor vibration sensitive area selection method based on CRP and RQA |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910316827.6A CN110119690B (en) | 2019-04-19 | 2019-04-19 | Parallel reactor vibration sensitive area selection method based on CRP and RQA |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110119690A CN110119690A (en) | 2019-08-13 |
| CN110119690B true CN110119690B (en) | 2022-08-16 |
Family
ID=67521237
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910316827.6A Active CN110119690B (en) | 2019-04-19 | 2019-04-19 | Parallel reactor vibration sensitive area selection method based on CRP and RQA |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110119690B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113566947A (en) * | 2021-06-24 | 2021-10-29 | 国网电力科学研究院武汉南瑞有限责任公司 | High-voltage reactor vibration signal analysis method, non-transient readable recording medium and system |
| CN113759188A (en) * | 2021-07-02 | 2021-12-07 | 国网河北省电力有限公司电力科学研究院 | Parallel reactor fault detection method and device and terminal |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102778633A (en) * | 2012-08-07 | 2012-11-14 | 绍兴电力局 | Method for selecting vibration sensitive area of power transformer tank surface winding |
| CN107328467A (en) * | 2017-07-26 | 2017-11-07 | 河海大学 | A kind of Transformer Winding thrust change detecting method based on recurrence quantification analysis |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5648157B2 (en) * | 2011-12-28 | 2015-01-07 | 株式会社日立ハイテクノロジーズ | Semiconductor manufacturing equipment |
-
2019
- 2019-04-19 CN CN201910316827.6A patent/CN110119690B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102778633A (en) * | 2012-08-07 | 2012-11-14 | 绍兴电力局 | Method for selecting vibration sensitive area of power transformer tank surface winding |
| CN107328467A (en) * | 2017-07-26 | 2017-11-07 | 河海大学 | A kind of Transformer Winding thrust change detecting method based on recurrence quantification analysis |
Non-Patent Citations (2)
| Title |
|---|
| Dynamical assessment of physiological systems and states using recurrence plot strategies;WEBBER C L;《Journal of Applied Physiology》;19941230;第965-973页 * |
| 交叉递归图在变压器铁芯压紧力变化检测中的应用;郭洁;《高电压技术》;20101130;第127-134页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110119690A (en) | 2019-08-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106443316B (en) | Multi-information detection method and device for deformation state of power transformer winding | |
| CN106845010B (en) | Low-frequency oscillation dominant mode identification method based on improved SVD noise reduction and Prony | |
| CN103558529B (en) | A kind of mode identification method of three-phase cartridge type supertension GIS partial discharge altogether | |
| CN111487046A (en) | Fault diagnosis method for circuit breaker voiceprint and vibration entropy feature fusion | |
| CN110119690B (en) | Parallel reactor vibration sensitive area selection method based on CRP and RQA | |
| CN110991481B (en) | High-voltage shunt reactor internal loosening fault diagnosis method based on cross wavelet transformation | |
| CN110174255B (en) | Transformer vibration signal separation method based on radial basis function neural network | |
| CN110032827B (en) | Current element three-dimensional inversion method based on algebraic elastic network regularization method | |
| CN110703151A (en) | Transformer fault diagnosis method based on vibration blind source separation and Bayesian model | |
| CN117748507B (en) | Distribution network harmonic access uncertainty assessment method based on Gaussian regression model | |
| CN110276094B (en) | Current element three-dimensional inversion method based on Bayesian elastic network regularization method | |
| CN111666912B (en) | Partial discharge fusion feature extraction method considering electrical feature quantity and graphic feature | |
| CN101614799B (en) | Signal separation and selection method applied in current sensor angular difference on-line monitoring system | |
| CN117289097A (en) | Power equipment partial discharge detection method, model training method, device and equipment | |
| CN112881839B (en) | Transformer diagnosis method based on mutual information of frequency concentration and vibration stability | |
| CN114924157B (en) | Parallel reactor state monitoring method and system based on 5G transmission | |
| CN113655534B (en) | Nuclear magnetic resonance FID signal noise suppression method based on multi-linear singular value tensor decomposition | |
| CN108334822B (en) | Kalman and modified wavelet transform filtering method based on electric vehicle charging nonlinear load characteristics | |
| CN112989966B (en) | Improved analog circuit signal noise reduction method | |
| CN111239557B (en) | Partial discharge positioning method based on online calibration | |
| CN111985408A (en) | Porcelain post insulator fault identification method and device | |
| CN111983398A (en) | Power transformer partial discharge TDOA (time difference of arrival) positioning method and system under condition of considering NLOS (non line of sight) | |
| CN116626760B (en) | Stratum discontinuity detection method and device based on self-adaptive high-order maximum entropy WVD | |
| CN117872039B (en) | Line fault location method and system based on improved RBF network | |
| CN115407128B (en) | Electric power system harmonic wave and inter-harmonic wave frequency estimation method based on inter-mass sampling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |