CN115754790A - Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system - Google Patents
Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system Download PDFInfo
- Publication number
- CN115754790A CN115754790A CN202211434208.5A CN202211434208A CN115754790A CN 115754790 A CN115754790 A CN 115754790A CN 202211434208 A CN202211434208 A CN 202211434208A CN 115754790 A CN115754790 A CN 115754790A
- Authority
- CN
- China
- Prior art keywords
- power supply
- voltage
- supply side
- fault
- current
- 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.)
- Pending
Links
Images
Landscapes
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a power supply side fault diagnosis and automatic switching method of a low-voltage dual-power supply system, which comprises the following steps of: extracting a current/voltage characteristic vector of a power supply side of the low-voltage dual-power supply system; training a fault state classification model by using the extracted feature vectors; optimizing parameters of the fault state classification model by adopting an ant colony algorithm to obtain a current/voltage threshold under the fault condition; the method comprises the steps of taking current/voltage of a power supply side collected in real time as a main data source, comparing the current/voltage with a current/voltage threshold value obtained under the condition of a fault, and judging whether the power supply side of the low-voltage dual-power-supply system has the fault or not; if the fault occurs, the automatic switch switching function of the low-voltage dual-power system is realized by controlling the circuit breaker through the controller. The invention has reasonable design, can quickly and accurately judge whether the power supply has faults or not according to the voltage value, ensures that the power supply can be quickly switched, and can be widely used for quickly switching low-voltage double power supplies on circuit breakers of power systems and large-scale power distribution networks.
Description
Technical Field
The invention belongs to the technical field of power supplies, relates to an upper power supply system, and particularly relates to a power supply side fault diagnosis and automatic switching method of a low-voltage dual-power supply system.
Background
Along with the continuous improvement of the power level in China, people put forward higher standards for power supply quality. In fact, the continuity of the power supply is one of the key evaluation criteria of the quality of the power. The dual power automatic transfer switch refers to a switch that automatically transfers to another power source after power failure for some reason. The application range of the device is very wide, and hospitals, districts, airports, stations, docks, chemical engineering, prevention and treatment and other places which do not allow power failure. Because the power supply has many kinds of faults, the conversion is necessary under the fault conditions, such as phase loss of any phase, overvoltage, undervoltage, frequency deviation, harmonic wave and the like, wherein the necessary conversion for the phase loss of any phase is the lowest requirement, and a high-end controller can even comprehensively check the quality of two paths of power supplies and automatically access one path of power supply with higher electric energy quality. In addition to preventing power supply switching in a fault-free condition, the controller of a dual power automatic transfer switch must be able to recognize momentary fluctuations in various voltages, including brief voltage drops without power failure. For example, the switching of the low-voltage distribution buscouple switch in the transformer room is a normal power interruption, and it is necessary to determine such a "normal" power interruption without determining the power interruption at the time of switching of the buscouple switch as a power failure.
In actual industrial production, most of the existing dual power supply systems are manual dual power supply systems, and the manual dual power supply systems have the disadvantage of long switching time, because the manual dual power supply systems are in a mechanical switching mode, the switching time is usually more than one second, and the long switching time easily causes all powered equipment systems to be interrupted in power supply, so that the equipment systems stop working. Due to the fact that the automatic switching function is not available, the power supply source cannot be switched in time in many times, power loss of the power supply circuit is caused, and power supply is interrupted. In addition, the switching mechanism of the general manual dual-power switching is complex, so that the reliability of the manual dual-power switching system is low, and the defects of high use cost, large installation occupation space and high later maintenance cost exist.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a power supply side fault diagnosis and automatic switching method of a low-voltage dual-power-supply system, which is reasonable in design, and can quickly and accurately realize the automatic switching function of a main power supply and a standby power supply.
The technical problem to be solved by the invention is realized by adopting the following technical scheme:
a power supply side fault diagnosis and automatic switching method of a low-voltage dual-power supply system comprises the following steps:
step 1, extracting a current/voltage characteristic vector of a power supply side of a low-voltage dual-power system by adopting a wavelet packet decomposition method;
step 2, taking the extracted feature vectors as training samples and testing samples, and training an ITSVM fault state classification model;
step 3, optimizing parameters of an ITSVM fault state classification model by adopting an ant colony algorithm to obtain a current/voltage threshold under the fault condition;
step 4, current/voltage of the power supply side collected in real time is used as a main data source, and is compared with a current/voltage threshold value obtained under the condition of a fault, and whether the power supply side of the low-voltage dual-power-supply system has the fault or not is judged;
and step 5, if the power supply side of the low-voltage dual-power system breaks down, controlling the circuit breaker through the controller to realize the automatic switching function of the 0.4kV dual-power system.
Further, the implementation method of step 1 includes the following steps:
performing wavelet packet analysis on a power supply side current/voltage signal to obtain each sub-frequency band;
secondly, performing single reconstruction on the discrete signals of each sub-frequency band to obtain energy characteristics of each reconstructed signal;
thirdly, normalizing the energy characteristics of each reconstruction signal to form an energy characteristic vector t;
and fourthly, selecting the most reflected data characteristic component from T to form a power supply side voltage/current characteristic vector T.
Further, the method for calculating the energy characteristics comprises the following steps: setting the energy characteristic of a power supply side voltage/current signal corresponding to the kth frequency band of the jth layer after wavelet packet decomposition as E jk Then, there is the following formula:
wherein, j wavelet packet decomposition levels; k =0,1,2, · 2 j-1 Is a number of a decomposition band, L isA kth band data length of a jth layer; m is a discrete point of the reconstructed signal of the kth frequency band of the jth layer; x is the number of km To reconstruct the amplitude of the discrete points;
in the step three, the energy characteristic vector t is:
step four, a power supply side voltage/current characteristic vector T is as follows:
further, in the step 2, a radial basis kernel function is selected in the improved dual-support vector machine by using a low-voltage dual-power-supply-system power-supply-side current/voltage signal feature vector sample set, and the sample set is trained by using whether the power supply needs to be converted as a class label, so that an ITSVM fault state classification model is obtained.
Further, the parameters of the ant colony algorithm for optimizing the ITSVM fault state classification model in the step 3 include: penalty parameter c 1 、 c 2 And a Gaussian kernel parameter sigma, the optimization method comprises the following steps:
initializing: size M of Ant colony, maximum number of cycles T max The pheromone volatilization coefficient p belongs to [0,1 ]]Randomly determining the initial position of the ant colony according to the sigma and the c 1 、c 2 The initial pheromone concentration of the ith ant is calculated according to the following formula:
τ(i)＝e -f(x) ；
the path is transferred: randomly selecting p ants from the ant colony, and selecting the ant with the largest pheromone from the p ants as the head ant X obj At a position X best Other ants gather to the head ant position according to pheromone attraction as follows:
Xi＝(1-λ)X i +λX obj
after the search is finished, head ant X generated in the last iteration is obtained best In the vicinity thereofLocal search in the domain:
updating pheromone: after each search, the pheromone concentration tau (i) of the ith ant is updated according to the following formula:
τ(i)＝(1-ρ)τ(i)+Δτ(i)
checking iteration termination conditions, and if the iteration termination conditions are not met, returning to the steps of. Otherwise, the iteration is finished to obtain the optimal sigma and c 1 、c 2 。
Further, the specific implementation method of step 4 is as follows: the method comprises the steps of collecting current/voltage of a power supply side in real time as a main data source, comparing a voltage/current signal of the collected voltage/current of the power supply side at a certain moment with a voltage/current signal of the same moment and the same position in the last cycle, extracting a characteristic vector by adopting a wavelet packet decomposition method after the difference is found, comparing the characteristic vector with a current/voltage threshold value obtained under the fault condition, and judging whether the power supply side of the low-voltage dual-power-supply system is in fault or not.
Further, the specific implementation method of 5 is as follows: after the power supply side fault is judged, the controller sends out an instruction to drive the operating mechanism to control the circuit breaker to be closed or opened, so that the motor is controlled to rotate forwards and backwards.
The invention has the advantages and positive effects that:
1. according to the invention, voltage/current information of the power supply side of the low-voltage dual-power supply is acquired in real time, WPA operation is carried out to obtain a voltage/current characteristic vector, then the voltage/current characteristic vector is transmitted into an ITSVM algorithm optimized by ACO to train to obtain a current/voltage threshold value of a power supply side fault, and the power supply side fault is judged and judged through an ITSVM fault classification model, so that the 10ms automatic switching function of the low-voltage main power supply and the low-voltage standby power supply is realized, and the method can be widely used for quickly switching the low-voltage dual-power supply on a breaker of a power system and a large-scale power distribution network.
2. According to the invention, the WPA algorithm is adopted to obtain the detailed information of the distribution of the original signal on different frequency bands and position the time point of the signal mutation, so that the characteristic vector is effectively extracted; by adopting the ITSVM algorithm, the learning and training speed can be improved, and the classification performance is further improved; the ACO algorithm is used for optimizing ITSVM parameters, optimizing the classification precision and generalization capability of the fault classification algorithm, and quickly and accurately judging whether the power supply has faults or not according to the voltage value, so that the power supply can be quickly switched.
Drawings
Figure 1 is a schematic diagram of a two-layer WPA;
FIG. 2 is a flow chart of extracting feature vectors by using a WPA method according to the present invention;
FIG. 3 is a flow chart of the algorithm for optimizing ITSVM using ACO according to the present invention;
FIG. 4 is a fault diagnosis flow diagram of the present invention;
FIG. 5 is a voltage and current waveform diagram of the voltage-loss switching of the resistive-inductive load;
fig. 6 is a flow chart of the automatic dual power switching of the present invention.
Detailed Description
The present invention will be described in detail with reference to the drawings.
The time design idea of the invention is as follows: analyzing the acquired voltage/current signals by adopting a WPA algorithm to obtain a characteristic vector as a training sample, and judging whether a power supply needs to be converted or not as a class label; establishing a low-voltage (0.4 kV) dual-power-supply-system power-supply-side fault diagnosis model (in the embodiment, a 0.4kV dual-power-supply system is taken as an example for explanation), training an ITSVM algorithm to obtain a training sample, optimizing the ITSVM algorithm by using an ACO algorithm to obtain a current/voltage threshold value under a fault condition, optimizing model parameters by using the ACO algorithm, and outputting optimal parameters sigma and c 1 、c 2 And ensuring that an ITSVM algorithm establishes an SVM classification model according to the optimal parameters, wherein an ACO algorithm optimizes the ITSVM as shown in figure 3. And finally, performing WPA operation on the voltage/current signals acquired in real time to obtain a characteristic vector, transmitting the characteristic vector into a trained fault state classification model, comparing the characteristic vector with a current/voltage threshold value obtained under the fault condition, and judging whether the power side of the 0.4kV dual-power system fails or not, so that the power supply is switched within 10 ms. The model canThe functions of health state monitoring and fault judgment are realized, and the stable operation time and the operation reliability of the system are improved.
Based on the design idea, the invention provides a power supply side fault diagnosis and automatic switching method of a low-voltage dual-power supply system, which comprises the following steps of:
step 1, extracting a current/voltage characteristic vector of a power supply side of a low-voltage dual-power supply system by adopting a wavelet packet decomposition (WPA) method.
The Wavelet Packet decomposition (WPA) algorithm can decompose not only low-frequency signals of a 0.4kV dual-power-supply mutual-throw switch, but also high-frequency signals at the dual power supplies, and compared with Wavelet decomposition, WPA is a fine decomposition technology, can decompose frequency bands in a multi-level mode, improves time-frequency resolution, and can extract fault characteristic fault characteristics of the power supply side of a 0.4kV dual-power-supply system through refining the frequency bands. In WPA, i in node (i, j) represents the number of layers of wavelet packet decomposition; j represents a sequence of wavelet packet decomposition of the same layer, and each node corresponds to the characteristics of the fault signal on the power supply side. The two-layer wavelet packet structure is shown in fig. 1.
The power supply side fault signal of the 0.4kV dual-power system belongs to a non-stable signal, and the WPA algorithm is as the following formula (1):
in the formula (d) i,j,m Representing the voltage/current wavelet packet decomposition of the top layer power supply side; d is a radical of i,j,2m And d i,j,2m+1 Represents further power supply side voltage/current wavelet packet decomposition; i represents the size of the wavelet packet; j represents the decomposition position of the wavelet packet; m represents the power supply side voltage/current wavelet packet frequency; l represents a variable, h 0 And h 1 Respectively representing the coefficients of the multi-resolution filter.
The wavelet packet reconstruction algorithm is as the formula (2):
the wavelet packet energy spectrum refers to the result of wavelet packet decomposition expressed in terms of energy. The signal x (t) may be subjected to the following WPA operation, as in equation (3):
in the formula, f ij (t j ) A wavelet packet representing the power supply side current/voltage signal decomposes the reconstructed signal at the i-th layer node (i, j). In case of a fault on the power supply side, the energy of the signals in the different frequency bands will change significantly. Therefore, the fault diagnosis can be performed on the power supply side according to the energy spectrum in different frequency bands.
The energy spectrum of a WPA can be expressed as formula (4):
in the formula, E ij (t j ) Representing the frequency band energy value corresponding to the jth node of the signal on the ith layer; x is the number of ij Representing the reconstructed signal f ij (t j ) The amplitude corresponding to the discrete points of (a).
Therefore, the total energy of the power supply side signal can be expressed as formula (5):
when the power supply side voltage/current signal is decomposed on the ith layer in a wavelet mode, the percentage of energy corresponding to different frequency bands in the total energy is as shown in a formula (6):
based on the WPA principle, the specific method for extracting the current/voltage characteristic vector of the power supply side of the low-voltage dual-power system is shown in figure 2 by taking a 0.4kV dual-power system as a research object, and comprises the following steps:
(1) And carrying out wavelet packet analysis on the current/voltage signal at the power supply side to obtain each sub-frequency band.
(2) And performing single-branch reconstruction on each sub-band discrete signal to obtain the energy of each reconstructed signal. Setting the power supply side voltage/current signal energy characteristic corresponding to the kth frequency band of the jth layer after wavelet packet decomposition as E jk Then, there is the following formula:
wherein, j wavelet packet decomposition levels; k =0,1,2, · 2 j-1 Is the serial number of the decomposition frequency band, L is the data length of the kth frequency band of the jth layer; m is a discrete point of the reconstructed signal of the kth frequency band of the jth layer; x is a radical of a fluorine atom km To reconstruct the amplitude of the discrete points.
(3) Normalizing the energy characteristics of the power supply side voltage/current signals in each frequency band of the jth layer to form the following energy characteristic vector t:
(4) And (3) selecting the components which can reflect the data characteristic most from T, and forming a power supply side voltage/current characteristic vector T as follows:
and 2, taking the extracted feature vectors as training samples and test samples, taking whether the power supply needs to be converted or not as class labels for training, selecting a radial basis kernel function in the ITSVM, and training the ITSVM fault state classification model.
An Improved double Support Vector Machine (ITSVM) algorithm is an Improved TSVM algorithm based on a least square function and an equality constraint. A double Support Vector Machine (TSVM) is used as a two-classification algorithm, and is different from a traditional two-classification algorithm SVM (Support Vector Machine), when the two-classification problem is solved (whether a power supply side of a 0.4kV dual-power system fails or not), the TSVM generates two non-parallel hyperplanes by solving two quadratic programming problems with smaller scales, and each hyperplane is close to one type of sample as much as possible and is far away from the other type of sample at the same time; the standard SVM obtains 1 classification hyperplane by solving 1 quadratic programming problem, and the category of the data sample on which side of the hyperplane is the category corresponding to the side; because the larger-scale quadratic programming problem in the traditional SVM is converted into two smaller-scale quadratic programming problems, the training time of the TSVM is greatly shortened to be about 1/4 of that of the traditional SVM, and the influence of unbalanced samples on the traditional SVM is overcome. As an extension of TSVM, the ITSVM uses equality constraint to replace inequality constraint in the original quadratic programming problem, and converts the quadratic programming problem into a least square form, thereby further reducing the computational complexity and shortening the training time.
In the binary problem, a given m multiplied by n dimension 0.4kV dual-power system power supply side current/voltage signal feature vector sample set X belongs to R m×n Wherein, positive class (the characteristic vector of the power supply side non-fault voltage/current) and negative class (the characteristic vector of the power supply side fault voltage/current) sample sets are respectively recorded as:
the non-linear ITSVM can be obtained by solving the following optimization problem:
ITSVM1：
s.t.-(Bω 1 +e 2 b 1 )+q＝e 2
ITSVM2：
s.t.-(Aω 2 +e 1 b 2 )+q＝e 1
wherein: omega 1 ，ω 2 ∈R n Is a hyperplane normal vector; b 1 ,b 2 Belongs to R, and is a bias value, and q is a relaxation variable; e.g. of the type 1 ,e 2 Is a full 1 vector, wherein
Penalty parameter c 1 ,c 2 >0。
Respectively solving ITSVM1 and ITSVM2 to obtain [ omega ] 1 b 1 ]And [ omega ] 2 b 2 ]And further 2 non-parallel hyperplanes are constructed:
x T ω 1 +b 1 ＝0 (9)
x T ω 2 +b 2 ＝0 (10)
the category of the voltage/current characteristic vector x can be judged according to the distance between the voltage/current characteristic vector x and the two hyperplanes.
For non-linear ITSVM, kernel function is introduced
And mapping the data samples to a high-dimensional feature space which can be subjected to linear classification, thereby solving a classification hyperplane. At this point, the problem of demand solutions translates into
s.t.-(K(B,C) T )u 1 +e 2 γ 1 )+q＝e 2 (11)
s.t.-(K(A,C) T )u 2 +e 1 γ 2 )+q＝e 1 (12)
In the formula: u. of 1 ,u 2 ∈R n Is a hyperplane normal vector; gamma ray 1 ,γ 2 E R is a biasA value; c T ＝[A B] T 。
The constraint conditions are substituted into the objective functions (11), (12) to obtain [ u 1 γ 1 ]And [ u ] 2 γ 2 ]And then 2 non-parallel hyperplanes are obtained:
K(x T ,C T )u 1 +γ 1 ＝0 (13)
K(x T ,C T )u 2 +γ 2 ＝0 (14)
the class to which the voltage/current eigenvector x belongs can be determined by:
ITSVM is an improvement on SVM algorithm, when a kernel function is introduced, RBF has the same advantage, so the radial basis kernel function is selected as follows:
step 3, adopting Ant Colony Optimization (ACO) to classify the parameters (sigma, c) of the ITSVM fault state classification model 1 ，c 2 ) And optimizing to obtain the current/voltage threshold under the fault condition.
The Ant Colony Optimization (ACO) is an Optimization parameter algorithm. The classification performance of the ITSVM classifier directly influences the overall classification accuracy, and the classification performance of each ITSVM classifier is in turn associated with a penalty parameter c 1 、c 2 There is a large relationship with whether the gaussian kernel parameter σ is properly selected. Therefore, in the training process of the classification model, the optimization of the parameters of the ITSVM is very important. The ACO algorithm has obvious advantages in solving the problem of target optimization and is applied to ITSVM parameter optimization to obtain the best classification effect. Therefore, the invention adopts the ACO algorithm to carry out parameter (sigma, c) on the ITSVM 1 ，c 2 ) Optimizing is carried out to improve the accuracy of the fault analysis of the power supply side of the 0.4kV dual-power system.
The ACO optimization ITSVM algorithm is shown in fig. 3 and comprises the following steps:
(1) Parameter initialization
Scale M including ant colony, maximum number of cycles T max The pheromone volatilization coefficient p belongs to [0,1 ]]Isoparametric, randomly determining the initial position of the ant colony, according to (sigma, c) 1 ，c 2 ) The initial pheromone concentration of the ith ant is calculated according to the parameter range:
τ(i)＝e -f(x) (16)
(2) Path transfer
Randomly selecting p ants from the ant colony, and selecting the maximum pheromone (the minimum objective function) from the p ants as the head ant X obj At a position X best And the other ants gather to the head ant position according to the pheromone attraction degree according to the formula (17):
Xi＝(1-λ)X i +λX obj (17)
after the search is completed, the head ant X generated in the last iteration can be obtained best Local search is performed in its vicinity:
(3) Pheromone update
After each search, the pheromone concentration tau (i) of the ith ant needs to be updated:
τ(i)＝(1-ρ)τ(i)+Δτ(i) (19)
checking iteration termination conditions, and if the iteration termination conditions are not met, returning to the steps (1) and (2); otherwise, the iteration ends and the optimal (sigma, c) is output 1 ，c 2 )。
And 4, collecting current/voltage of the power supply side in real time as a main data source, comparing the collected voltage/current signal of the power supply side at a certain moment with the voltage/current signal of the same moment and the same position in the last cycle, extracting a characteristic vector by WPA (wavelet packet access) after finding that the voltage/current signal and the voltage/current signal are different, comparing the characteristic vector with a current/voltage threshold value obtained under the condition of a fault, and judging whether the power supply side of the 0.4kV dual-power-supply system is in fault or not.
To ensure reliable operation of the 0.4kV dual-power-supply fast mutual-throw switch, the rated voltage of the main circuit should be ensured to operate within a normal operating range of 85-110% ue.
Dividing a cycle signal of the acquired power supply side voltage into 96 points, comparing a voltage signal at a certain moment with a voltage signal at the same moment and the same position in the last cycle, and judging the difference between the two signals and a current/voltage threshold domain obtained by a fault model. When the model threshold is exceeded, the subsequent ten or more points of the current cycle are continuously tracked while the current signal is detected. When the current signal is detected, similar to the voltage, the next more than ten points in the current cycle are also detected and compared with the current signal at the same time and the same position of the previous cycle. When the voltage is reduced, if the detected current signal rises, the internal fault (load side short circuit) is judged to be possible, and in this case, the upper-port low-voltage circuit breaker is disconnected. When the voltage is reduced and the current is kept unchanged or reduced, the circuit has voltage loss or voltage reduction faults, the fault model judges that the main power supply has faults, the dual-power switch is switched to the standby power supply, and when the voltage of the common power supply is recovered to be normal, the common power supply is switched back to the common power supply. When the voltage rises, the circuit has an overvoltage fault, the fault model judges that the main power supply has a fault, the dual-power switch is switched to the standby power supply, and when the voltage of the common power supply is recovered to be normal, the common power supply is switched back to the common power supply. Fig. 5 is a voltage-current waveform diagram of three-phase voltage falling from 220V to 170V simultaneously under the resistive load, wherein a channel 1 is a voltage wave and a channel 2 is a current wave.
And step 5, if the power supply side of the low-voltage dual-power system breaks down, controlling the circuit breaker through the controller to realize the automatic switching function of the 0.4kV dual-power system.
As shown in fig. 6, after the power supply side fault is determined, the controller sends a command to drive the operating mechanism to control the circuit breaker to close or open, so as to control the motor to rotate forward and backward. Normally, a common power supply is used for supplying power to a load, and when a fault of the common power supply is detected, whether switching processing is performed or not is judged. Power supply side failures are generally divided into three cases:
in the first case, the utility power fails and the backup power is normal. When the controller detects the fault of the common power supply, the corresponding indicator lamp is lightened and gives an alarm, and the main power supply is switched back after the power supply is detected to be recovered to be normal.
In the second case, the normal power supply is normal and the backup power supply fails. When the controller detects that the common power supply is normal. When the standby power supply fails, the controller can send a control command, light the corresponding indicator lamp and give an alarm, and when the standby power supply is detected to be recovered to be normal, the alarm is turned off, and the corresponding indicator lamp is not switched.
In the third situation, if the fault of the common power supply or the standby power supply is detected, a corresponding instruction is sent, a corresponding indicator lamp is lightened, the change-over switch is switched to a zero position, and the common power supply and the standby power supply are cut off and give an alarm.
It should be emphasized that the embodiments described herein are illustrative and not restrictive, and thus the present invention includes, but is not limited to, the embodiments described in the detailed description, as well as other embodiments that can be derived by one skilled in the art from the teachings herein.
Claims (7)
1. A low-voltage dual-power-supply system power-supply-side fault diagnosis and automatic switching method is characterized in that: the method comprises the following steps:
step 1, extracting a current/voltage characteristic vector of a power supply side of a low-voltage dual-power system by adopting a wavelet packet decomposition method;
step 2, taking the extracted feature vectors as training samples and testing samples, and training an ITSVM fault state classification model;
step 3, optimizing parameters of the ITSVM fault state classification model by adopting an ant colony algorithm to obtain a current/voltage threshold under the fault condition;
step 4, current/voltage of the power supply side collected in real time is used as a main data source, and is compared with a current/voltage threshold value obtained under the condition of a fault, and whether the power supply side of the low-voltage dual-power-supply system has the fault or not is judged;
and step 5, if the power supply side of the low-voltage dual-power-supply system fails, controlling the circuit breaker through the controller to realize the automatic switching function of the 0.4kV dual-power-supply system.
2. The method for power supply side fault diagnosis and automatic switching of the low-voltage dual-power supply system according to claim 1, characterized in that: the implementation method of the step 1 comprises the following steps:
performing wavelet packet analysis on a power supply side current/voltage signal to obtain each sub-frequency band;
secondly, performing single reconstruction on each sub-band discrete signal to obtain the energy characteristics of each reconstructed signal;
thirdly, normalizing the energy characteristics of the reconstruction signals to form an energy characteristic vector t;
and fourthly, selecting the most reflected data characteristic component from T to form a power supply side voltage/current characteristic vector T.
3. The method for diagnosing and automatically switching the power supply side fault of the low-voltage dual-power supply system according to claim 2, wherein the method comprises the following steps: the method for calculating the energy characteristics comprises the following steps: setting the power supply side voltage/current signal energy characteristic corresponding to the kth frequency band of the jth layer after wavelet packet decomposition as E jk Then, there is the following formula:
wherein, j wavelet packet decomposition levels; k =0,1,2, · 2 j-1 The sequence number of the decomposed frequency band is L, and the data length of the kth frequency band of the jth layer is L; m is a discrete point of the reconstructed signal of the kth frequency band of the jth layer; x is the number of km To reconstruct the amplitude of the discrete points;
the step three is that the energy characteristic vector t is as follows:
step four, a power supply side voltage/current characteristic vector T is as follows:
4. the method for diagnosing and automatically switching the power supply side fault of the low-voltage dual-power supply system according to claim 1, wherein the method comprises the following steps: and 2, selecting a radial basis kernel function in the improved double-support vector machine by using a low-voltage double-power-supply system power supply side current/voltage signal characteristic vector sample set, and training the sample set by using whether the power supply needs to be converted as a class label to obtain an ITSVM fault state classification model.
5. The method for diagnosing and automatically switching the power supply side fault of the low-voltage dual-power supply system according to claim 1, wherein the method comprises the following steps: the parameters of the ant colony algorithm for optimizing the ITSVM fault state classification model in the step 3 comprise: penalty parameter c 1 、c 2 And a Gaussian kernel parameter sigma, the optimization method comprises the following steps:
initializing: ant colony size M, maximum number of cycles T max The pheromone volatilization coefficient p belongs to [0,1 ]]Randomly determining the initial position of the ant colony according to sigma and c 1 、c 2 The initial pheromone concentration of the ith ant is calculated according to the following formula:
τ(i)＝e -f(x) ；
path transfer: randomly selecting p ants from the ant colony, and selecting the ant with the largest pheromone from the p ants as the head ant X obj At a position X best Other ants gather to the head ant position according to pheromone attraction as follows:
Xi＝(1-λ)X i +λX obj
after the search is finished, the head ant X generated in the last iteration is obtained best Local search is performed in its vicinity:
updating pheromone: after each search, the pheromone concentration tau (i) of the ith ant is updated according to the following formula:
τ(i)＝(1-ρ)τ(i)+Δτ(i)
checking an iteration termination condition, and if the iteration termination condition is not met, returning to the steps of. Otherwise, the iteration is finished to obtain the optimal sigma and c 1 、c 2 。
6. The method for diagnosing and automatically switching the power supply side fault of the low-voltage dual-power supply system according to claim 1, wherein the method comprises the following steps: the specific implementation method of the step 4 comprises the following steps: the method comprises the steps of collecting current/voltage of a power supply side in real time as a main data source, comparing a voltage/current signal of the collected voltage/current of the power supply side at a certain moment with a voltage/current signal of the same moment and the same position in the last cycle, extracting a characteristic vector by adopting a wavelet packet decomposition method after the difference is found, comparing the characteristic vector with a current/voltage threshold value obtained under the fault condition, and judging whether the power supply side of the low-voltage dual-power-supply system is in fault or not.
7. The method for diagnosing and automatically switching the power supply side fault of the low-voltage dual-power supply system according to claim 1, wherein the method comprises the following steps: the concrete implementation method of the step 5 is as follows: after the power supply side fault is judged, the controller sends an instruction to drive the operating mechanism to control the circuit breaker to be closed or opened, so that the motor is controlled to rotate forwards and backwards.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211434208.5A CN115754790A (en) | 2022-11-16 | 2022-11-16 | Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211434208.5A CN115754790A (en) | 2022-11-16 | 2022-11-16 | Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115754790A true CN115754790A (en) | 2023-03-07 |
Family
ID=85371904
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211434208.5A Pending CN115754790A (en) | 2022-11-16 | 2022-11-16 | Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115754790A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117129904A (en) * | 2023-10-27 | 2023-11-28 | 深圳市大易电气实业有限公司 | Industrial power supply rapid switching monitoring method based on data analysis |
-
2022
- 2022-11-16 CN CN202211434208.5A patent/CN115754790A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117129904A (en) * | 2023-10-27 | 2023-11-28 | 深圳市大易电气实业有限公司 | Industrial power supply rapid switching monitoring method based on data analysis |
| CN117129904B (en) * | 2023-10-27 | 2023-12-22 | 深圳市大易电气实业有限公司 | Industrial power supply rapid switching monitoring method based on data analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109270442B (en) | DBN-GA neural network-based high-voltage circuit breaker fault detection method | |
| CN112051481B (en) | Alternating current-direct current hybrid power grid fault area diagnosis method and system based on LSTM | |
| Sun et al. | Fault diagnosis of power transformers using computational intelligence: A review | |
| WO2015158198A1 (en) | Fault recognition method and system based on neural network self-learning | |
| CN106856322B (en) | A kind of flexible direct current power distribution network intelligent protection system based on neural network | |
| CN104535865A (en) | Comprehensive diagnosing method for operation troubles of power transformer based on multiple parameters | |
| CN112348071B (en) | Open-circuit fault diagnosis method for switching tube of submodule of modular multilevel converter | |
| CN113111591B (en) | Automatic diagnosis method, device and equipment based on internal fault of modular power distribution terminal | |
| CN112883634B (en) | DC measurement system state prediction method and system based on multi-dimensional analysis | |
| CN112949874B (en) | Power distribution terminal defect characteristic self-diagnosis method and system | |
| CN115754790A (en) | Power supply side fault diagnosis and automatic switching method for low-voltage dual-power system | |
| CN113219328B (en) | Intelligent fault diagnosis method for circuit breaker operating mechanism based on information fusion | |
| CN116244617A (en) | Intelligent high-voltage switch cabinet fault diagnosis method and system based on heterogeneous graph structure learning | |
| CN108304931A (en) | A kind of Condition-based Maintenance of Substation Equipment method for diagnosing faults | |
| Li et al. | Physics-informed learning for high impedance faults detection | |
| Bishal et al. | ANN based fault detection & classification in power system transmission line | |
| Singh et al. | Power system fault diagnosis using fuzzy decision tree | |
| CN112926620B (en) | Defect type identification method and system for partial discharge of transformer bushing | |
| CN113702767A (en) | Island direct-current microgrid fault diagnosis method based on wavelet sliding window energy | |
| CN109635430A (en) | Grid power transmission route transient signal monitoring method and system | |
| Shihabudheen et al. | Detection of high impedance faults in power lines using empirical mode decomposition with intelligent classification techniques | |
| CN115146739A (en) | Power transformer fault diagnosis method based on stacked time series network | |
| Wang et al. | Synthetic data in DC microgrids: Label creation for ensemble learning for fault isolation | |
| CN115201632A (en) | Multi-terminal direct-current transmission line fault identification method | |
| CN110866652B (en) | Online PMU data error correction method and system based on LSTM model |
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 |