CN111679166A - Switch cabinet partial discharge fault multi-source information fusion detection early warning system and method based on wireless transmission technology - Google Patents
Switch cabinet partial discharge fault multi-source information fusion detection early warning system and method based on wireless transmission technology Download PDFInfo
- Publication number
- CN111679166A CN111679166A CN202010719186.1A CN202010719186A CN111679166A CN 111679166 A CN111679166 A CN 111679166A CN 202010719186 A CN202010719186 A CN 202010719186A CN 111679166 A CN111679166 A CN 111679166A
- Authority
- CN
- China
- Prior art keywords
- switch cabinet
- module
- signals
- partial discharge
- detection
- 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
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1209—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing using acoustic measurements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/18—Status alarms
- G08B21/185—Electrical failure alarms
-
- G—PHYSICS
- G08—SIGNALLING
- G08C—TRANSMISSION SYSTEMS FOR MEASURED VALUES, CONTROL OR SIMILAR SIGNALS
- G08C17/00—Arrangements for transmitting signals characterised by the use of a wireless electrical link
- G08C17/02—Arrangements for transmitting signals characterised by the use of a wireless electrical link using a radio link
Abstract
The invention relates to a multi-source information fusion detection early warning system for local discharge faults of a switch cabinet based on a wireless transmission technology, which belongs to the technical field of electrical safety and comprises a detection terminal, a power supply module, a wireless communication module and a system background module; the detection terminal is connected with the system background module through the wireless communication module; the detection terminal comprises a transient ground voltage detection sensor, an ultrahigh frequency sensor and an ultrasonic sensor; the system background module comprises a data processing module, a central processing unit, a storage module and a display module. The invention also provides a multi-source information fusion detection early warning method for the local discharge fault of the switch cabinet based on the wireless transmission technology. The invention improves the reliability of the detection result, meets the requirement of rapid communication between the detection terminal and the system background, and obtains a more accurate fault detection result by adopting a BP neural algorithm and a multi-Bayesian estimation method.
Description
Technical Field
The invention belongs to the technical field of electrical safety, and relates to a multi-source information fusion detection early warning system and method for a local discharge fault of a switch cabinet based on a wireless transmission technology.
Background
In recent years, the number of power accidents caused by the deterioration of the insulation characteristic of a switch cabinet in an urban distribution network rises year by year, so that the problems of poor reliability, instability and the like of power supply of a distribution network system are caused. With the increasing capacity of power equipment and the increasing number of switch cabinets, the advanced regular inspection and maintenance cannot meet the current rapidly-increasing power demand, and the maintenance mode is developing into state maintenance. Compared with the traditional preventive power failure test, the electrified detection technology is used for acquiring information such as insulation degradation, partial discharge and the like of equipment in operation by adopting a sensor with high sensitivity, and the operation state of the equipment is truly reflected by adopting a computer network based on software support for processing and identifying information quantity, so that the comprehensive diagnosis of equipment faults or defects is realized. Among the most used non-invasive methods, the very high frequency method, the transient voltage earth voltage detection method, and the ultrasonic detection method are currently used. When ultrahigh frequency partial discharge detection is carried out, the ultrahigh frequency partial discharge detection is easily influenced by ultrahigh frequency electromagnetic interference in the environment, and the external sensor cannot detect all-metal closed power equipment. Although the ultrahigh frequency technology is widely and mature applied to the GIS equipment, the ultrahigh frequency technology has certain limitation in switch cabinet equipment mainly based on air insulation, and is only sensitive to air and suspension discharge (plate-to-plate) by a tip, and other types of defects can not detect abnormal signals or detected abnormal signals have no typical discharge characteristics and are not beneficial to defect judgment. The ultrasonic technology has obvious advantages on defect identification, the amplitude is increased along with the increase of the severity of the defect, but under the condition of field detection, the vibration and the sound of some equipment are serious, and an ultrasonic signal is easily submerged into background noise; some equipment have good leakproofness, and ultrasonic signal comes out through the cubical switchboard gap from the source of discharging, has great decay, and this has led to the reduction of defect detectable rate to a certain extent. Compared with a single partial discharge source detection method, the combined detection method has the advantages of high detection sensitivity, high precision and the like.
The related technical scheme comprises the following steps: CN 201910703631.2-switch cabinet partial discharge detection system and switch cabinet partial discharge detection method; cn201911364295. x-a method for detecting partial discharge of a switchgear by using a robot; CN201910907648. X-a high tension switchgear partial discharge detection circuit and detection method thereof.
The existing single detection technology has certain defects: the transient earth voltage method detection method has lower detection sensitivity to most defects; when ultrahigh frequency partial discharge detection is carried out, the ultrahigh frequency partial discharge detection is easily influenced by ultrahigh frequency electromagnetic interference in the environment, and an external sensor cannot detect all-metal closed power equipment; ultrasonic wave technique has obvious advantage to defect identification to amplitude increases along with defect severity's increase, but under the on-the-spot detection condition, some equipment vibration sound is comparatively serious, and ultrasonic signal easily submerges in the background noise, and some equipment have good leakproofness, and ultrasonic signal passes out through the cubical switchboard gap from the source of discharging, has great decay, has led to the reduction of defect detectable rate to this degree.
Disclosure of Invention
In view of the above, the present invention provides a switch cabinet local discharge fault multi-source information fusion detection early warning system and method based on a wireless transmission technology, which solve the problems of poor diagnosis stability and high misjudgment rate of a single sensing technology, and avoid the disadvantage of low safety caused by too close distance between devices of an operator in the previous detection process.
In order to achieve the purpose, the invention provides the following technical scheme:
on one hand, the invention provides a multi-source information fusion detection early warning system for local discharge faults of a switch cabinet based on a wireless transmission technology, which comprises a detection terminal, a power supply module, a wireless communication module and a system background module; the detection terminal is connected with the system background module through the wireless communication module;
the detection terminal comprises a transient ground voltage detection sensor, an ultrahigh frequency sensor and an ultrasonic sensor;
the system background module comprises a data processing module, a central processing unit, a storage module and a display module.
Further, the transient ground voltage detection sensor is used for detecting signals including a ground voltage amplitude signal and a ground voltage pulse signal, the measured amplitude is in the range of 0 to 80dBmV, and the frequency is in the range of 75MHz to 500 MHz.
Furthermore, an Archimedes spiral antenna is arranged in the ultrahigh frequency sensor, the working frequency band is 400 MHz-800 MHz, the voltage standing wave ratio in a pass band is less than 2, the total return loss in the pass band is less than-10 dB, and the receiving gain in the pass band is greater than-60 dB.
Further, the ultrasonic sensor is used for acquiring real-time pulses of ultrasonic signals and detecting the correlation of 50Hz and 100Hz pulse signals.
Further, the wireless communication module adopts NB-IoT technology for communication, QPSK modulation and demodulation is adopted in uplink, and BPSK or QPSK modulation and demodulation is adopted in downlink.
Further, the data processing module is used for performing wavelet denoising and filtering on the acquired signals, selecting a Daubechies wavelet basis as a basis function of wavelet packet decomposition, and filtering the data signals with the frequency meeting 400 MHz-800 MHz.
Further, the central processing unit is used for realizing processing and judgment of fault states and information, extracting and classifying frequency band energy and time domain characteristics of the partial discharge signals through a BP neural network, fusing multi-detection information by adopting a multi-Bayesian estimation method, establishing a partial discharge multi-signal fusion identification model, judging whether each signal acquired at the detection position of the switch cabinet exceeds a safety threshold value, and making response judgment and prediction.
On the other hand, the invention provides a multi-source information fusion detection early warning method for local discharge faults of a switch cabinet based on a wireless transmission technology, which comprises the following steps:
s1: the detection terminal judges whether an instruction for using a certain characteristic sensor is received; if receiving the instruction, executing the acquisition work of the corresponding signal according to the instruction; if no special instruction exists, the acquisition modules of the ground voltage, the ultrahigh frequency and the ultrasonic partial discharge signal work simultaneously, detect the signals at different positions of the switch cabinet respectively, and transmit the acquired signals to a data processing module of a system background in real time through the wireless transmission module;
s2: the data processing module carries out wavelet denoising and filtering processing on the discharge signal;
s3: the central processor performs characteristic acquisition and modeling analysis on the processed signals through a BP neural network and a multi-Bayesian estimation method, and judges the operation condition, fault sites and whether a partial discharge fault trend exists or not through comparison with an internal partial discharge characteristic signal information base and historical operation data of the switch cabinet;
s4: the signals and results processed by the central processor are stored in a memory connected with the central processor;
s5: the decision structure of the central processor is sent to the display module through the wireless transmission module, and if a partial discharge alarm signal is received, a partial discharge position is displayed; if the partial discharge alarm signal is not received, the prediction result of the future operation state of the switch cabinet is displayed.
Further, the BP neural network is respectively used as an identification classifier of ultrasonic wave, ultrahigh frequency and earth electric wave detection signals, the topological structure comprises three layers which are respectively an input layer, a hidden layer and an output layer, information is transmitted between the layers through the connection right of the neurons, and the neurons in the same layer are mutually independent;
for the identification classification of the ultrasonic signal, 4 neurons of the input layer are provided, and correspond to the feature quantities of the ultrasonic signal: periodic maximum, effective, 50Hz correlation and 100Hz correlation; for the identification and classification of the ultrahigh frequency signals, the number of neurons of the input layer is 3, and the neurons correspond to the feature quantities of the ultrasonic signals: a periodic maximum, a valid, and a frequency; for the identification and classification of the earth electric wave signals, the number of neurons of an input layer is 8, the neurons respectively correspond to transient voltage signals at different positions of a switch cabinet, the number of neurons of an output layer corresponds to defect type codes and is set to be 2, and the neurons respectively correspond to a normal state and an abnormal state;
the operation process is as follows: the neuron of the input layer transmits the input characteristic quantity to the hidden layer through the connecting ring, the input information is continuously amplified, attenuated or inhibited, and the optimal connection weight is obtained in training, so that an optimal classifier model is established.
Further, the multi-bayes estimation method comprises the following steps:
(1) and determining output nodes and input nodes of the network according to the feature vector structure and the fault mode classification, wherein the number of the input nodes is 3, and the input nodes are respectively as follows: ultrahigh frequency signal state, ultrasonic signal state and earth electric wave signal state, output node is 5: normal operation, impending partial discharge, tip partial discharge, surface partial discharge and metal particle partial discharge;
(2) initializing parameters in the model, and setting the parameters as random small numerical values according to experience;
(3) selecting a switch cabinet signal sample data set, wherein the switch cabinet signal sample data set comprises a larger part of unmarked data samples and a smaller part of label sample data, standardizing the data, and dividing the label sample data into a tuning set and a test set according to a certain proportion;
(4) selecting unlabeled sample data as input of a training stage, and then carrying out fine tuning processing on the divided tuning set in the tuning stage so as to complete training of the whole multi-Bayesian model and store the trained multi-Bayesian network;
(5) and testing the saved multi-Bayesian network by the divided test set, saving the test result and completing diagnosis.
The invention has the beneficial effects that:
1. the system can synchronously monitor transient ground voltage signals, ultrasonic signals and ultrahigh frequency signals of the switch cabinet to be detected so as to comprehensively analyze the partial discharge state in the switch cabinet and prevent errors caused by only using a certain detection technology. In addition, the system can execute certain specific detection according to the instruction, meet the specific requirements in the actual fault detection of the switch cabinet and improve the reliability of the detection result.
2. The system adopts a wireless communication information transmission mode, avoids complex detection circuits in the detection process, also ensures the safe distance between a worker and high-voltage equipment, and better meets the requirement of rapid communication between a detection terminal and a system background.
3. The system adopts the solar cell and the storage battery to supply power jointly, thereby better ensuring the requirement of the system on working power consumption and meeting the requirement of energy conservation to the maximum extent.
4. The system can realize the positioning of the fault point of the switch cabinet, and particularly, two or more ultrahigh frequency sensors simultaneously detect the fault switch cabinet, and the time difference, the signal amplitude and the phase attenuation degree of the ultrahigh frequency signals detected by different sensors are processed and compared to determine the position of the fault point.
5. The system adopts a wavelet packet method to carry out denoising processing on the acquired signals so as to eliminate environmental interference to the maximum extent and obtain more accurate frequency domain characteristics of the local discharge signals of the switch cabinet.
6. The system adopts a BP neural algorithm to analyze and reprocess the acquired data in the central processor to obtain a characteristic value reflecting the condition of the high-voltage switch cabinet and provide useful data information for early warning diagnosis; and the post-fault early warning and diagnosis part compares and analyzes the data processed by the algorithm with historical data and other information by a multi-Bayesian estimation method, predicts and diagnoses and early warns the fault condition and fault node of the high-voltage switch cabinet, and takes corresponding measures when needed.
7. The detection terminal of the system is provided with a metal shell so as to eliminate interference signals in the environment through metal shielding.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention may be realized and attained by the means of the instrumentalities and combinations particularly pointed out hereinafter.
Drawings
For the purposes of promoting a better understanding of the objects, aspects and advantages of the invention, reference will now be made to the following detailed description taken in conjunction with the accompanying drawings in which:
FIG. 1 is a schematic structural diagram of a multi-source information fusion detection early warning system for partial discharge faults of a switch cabinet;
FIG. 2 is a schematic structural block diagram of signal processing of a multi-source information fusion detection early warning system for partial discharge faults of a switch cabinet;
FIG. 3 is a schematic diagram of a multi-information fusion recognition method based on a multi-Bayesian estimation method;
FIG. 4 is a flow chart of a multi-source information fusion detection early warning method for partial discharge faults of a switch cabinet;
FIG. 5 is a flow chart of signal pre-training of a multi-Bayesian feature set processing model.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention in a schematic way, and the features in the following embodiments and examples may be combined with each other without conflict.
Wherein the showings are for the purpose of illustrating the invention only and not for the purpose of limiting the same, and in which there is shown by way of illustration only and not in the drawings in which there is no intention to limit the invention thereto; to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there is an orientation or positional relationship indicated by terms such as "upper", "lower", "left", "right", "front", "rear", etc., based on the orientation or positional relationship shown in the drawings, it is only for convenience of description and simplification of description, but it is not an indication or suggestion that the referred device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes, and are not to be construed as limiting the present invention, and the specific meaning of the terms may be understood by those skilled in the art according to specific situations.
As shown in fig. 1-3, a multi-source information fusion detection and early warning system for local discharge fault of a switch cabinet based on wireless transmission technology comprises a detection terminal, a power supply module, a wireless communication module and a system background module. The detection terminal comprises a transient ground voltage detection sensor, an ultrahigh frequency sensor and an ultrasonic sensor; the power supply module comprises a solar cell and a storage battery; the system background comprises a data processing module, a central processing unit, a storage module and a display module.
The transient ground voltage detection sensor is a double-layer copper-clad plate with the thickness of 1mm, the detection signal comprises a ground voltage amplitude signal and a ground voltage pulse signal, the amplitude range of the detection signal is 0-80 dBmV, and the frequency range is 75 MHz-500 MHz.
The Archimedes spiral antenna selected in the ultrahigh frequency sensor meets the following requirements: the working frequency band is 400 MHz-800 MHz, the voltage standing wave ratio in the pass frequency band is less than 2, the total return loss in the pass frequency band is less than-10 dB, and the receiving gain in the pass frequency band is more than-60 dB.
The ultrasonic sensor is a piezoelectric waterproof ultrasonic sensor, and collects real-time pulses of ultrasonic signals during detection, mainly based on the correlation of 50Hz and 100Hz pulse signals.
The hardware of the detection terminal is additionally provided with a metal shell, and aims to eliminate interference signals through metal shielding.
The power supply module is composed of a solar cell and a storage battery together so as to achieve the aim of energy conservation and expect to ensure the power utilization requirement of the system to the maximum extent.
The wireless communication module adopts NB-IoT technology, the loss bandwidth is about 180KHz, QPSK modulation and demodulation is adopted in the uplink, BPSK or QPSK modulation and demodulation is adopted in the downlink, and the information transmission requirement between the detection terminal and the system background can be met.
The data processing module realizes the denoising and filtering of the acquired signals. Specifically, a Daubechies wavelet basis is selected as a basis function of wavelet packet decomposition, and the Daubechies wavelet basis has the characteristics of excellent frequency domain characteristics, good smoothness and low signal distortion rate; and an FIR (finite Impulse response) series filter is selected to filter the data signals with the frequency of 400 MHz-800 MHz.
The central processing unit selects an AMD series processor for realizing the processing and judgment of the fault state and the information, extracts and classifies the frequency band energy and the time domain characteristics of the partial discharge signals through a BP neural network, then realizes the fusion of multi-detection information by adopting a multi-Bayesian estimation method, establishes a partial discharge multi-signal fusion recognition model, judges whether each signal acquired at the detection position of the switch cabinet exceeds a safety threshold value, and makes response judgment and prediction.
Bayesian estimation is a common method for fusing high-level information of multiple sensors in an environment, sensor information is combined according to a probability principle, measurement uncertainty is expressed by conditional probability, and when observation coordinates of a sensor group are consistent, data of the sensors can be directly fused, but in most cases, data fusion of the sensor measurement data is carried out by adopting Bayesian estimation in an indirect mode. The multi-Bayesian estimation takes each sensor as a Bayesian estimation, combines the associated probability distribution of each individual object into a combined posterior probability distribution function, provides the final fusion value of the multi-sensor information by using the likelihood function of the combined distribution function as the minimum, and provides a feature description of the whole environment by the fusion information and a prior model of the environment.
The storage module is connected with the central processing unit and used for storing data.
The display module adopts an STN dot matrix liquid crystal display module, is connected with the central processing unit and is used for displaying the decision result of the central processing unit.
An embodiment of the present application provides a method for detecting partial discharge of a switch cabinet, which is implemented by using the multi-source information fusion detection and early warning system for partial discharge faults of the switch cabinet, as shown in fig. 4, specifically as follows:
(1) the detection terminal judges whether an instruction for using a certain characteristic sensor is received;
(2) if receiving the instruction, executing the acquisition work of the corresponding signal according to the instruction;
(3) if no special instruction exists, the acquisition modules of the ground voltage, the ultrahigh frequency and the ultrasonic partial discharge signal work simultaneously, detect the signals at different positions of the switch cabinet respectively, and transmit the acquired signals to a central processor of a system background in real time through the wireless transmission module;
(4) the data processing module carries out wavelet denoising and filtering processing on the discharge signal;
(5) the central processor performs characteristic acquisition and modeling analysis on the processed signals through a BP neural network and a multi-Bayesian estimation method, and judges the operation condition, fault sites and whether a partial discharge fault trend exists or not through comparison with an internal partial discharge characteristic signal information base and historical operation data of the switch cabinet;
(6) the signals and results processed by the central processor are stored in a memory connected with the central processor;
(7) the decision structure of the central processor is sent to the display module through the wireless transmission module, and if a partial discharge alarm signal is received, a partial discharge position is displayed; if the partial discharge alarm signal is not received, the prediction result of the future operation state of the switch cabinet is displayed.
The three sensors of the invention collect different types of partial discharge signals, each having advantages. The ground voltage sensor and the ultrahigh frequency sensor have higher sensitivity, but are easy to be subjected to some electromagnetic interference on site, so that fault information can be confirmed through the ultrasonic sensor. In addition, the ground voltage signal and the ultrahigh frequency signal have different frequencies, and only one detection method is not enough to completely detect the discharge signal, so that the local discharge fault can be more accurately monitored by combining multiple methods.
The central processor of the system background can integrate and analyze the acquired signals, comprehensively judge whether the partial discharge signals sent by the discharge detection module are abnormal or not, and send warning signals to abnormal equipment;
the partial discharge signal acquisition module is used for carrying out partial discharge detection on the switch cabinet to generate a partial discharge signal, and the acquired signal is sent to the system background module through the wireless communication module.
The BP neural network is respectively used as an identification classifier of ultrasonic wave, ultrahigh frequency and earth electric wave detection signals. The classifier is based on the basic composition principle and the information processing mode of the nervous system, and adopts a proper learning rule to solve and correct the model in parameter training, so as to establish a final model for solving the problem. The topological structure comprises three layers, namely an input layer, a hidden layer and an output layer, wherein information is transmitted between the layers through the connection rights of neurons, and the neurons in the same layer are independent. For the identification classification of the ultrasonic signal, 4 neurons of the input layer are provided, and correspond to the feature quantities of the ultrasonic signal: periodic maximum, effective, 50Hz correlation and 100Hz correlation; for the identification and classification of the ultrahigh frequency signals, the number of neurons of the input layer is 3, and the neurons correspond to the feature quantities of the ultrasonic signals: a periodic maximum, a valid, and a frequency; for the identification classification of the earth electric wave signals, the number of neurons of the input layer is 8, and the neurons respectively correspond to the transient voltage signals at different positions of the switch cabinet. The output layer neurons correspond to the defect type codes and are set to be 2, and the output layer neurons correspond to a normal state and an abnormal state respectively. The number of hidden neurons is generally determined empirically or by testing, and too large a number easily results in complex training and slower convergence speed, resulting in over-learning. The hidden layer neurons for all three recognition classifiers were empirically set to 8. The operation process of the whole neural network is as follows: the neuron of the input layer transmits the input characteristic quantity to the hidden layer through the connecting ring, the input information is continuously amplified, attenuated or inhibited, and the optimal connection weight is obtained in training, so that an optimal classifier model is established. And (3) applying MATLAB to design a BP neural network classifier, and designing the structure NN (input layer node number, hidden layer node number and output layer node number) of the network. The transfer function of the output layer is logsig () function, the transfer function of the hidden layer is Sigmoid type, and the learning step length of the neural network is set to be 0.02.
The bottom of the multi-Bayesian model is stacked by a plurality of layers of Bayesian regularized feature sets, the model extracts each layer of feature set from bottom to top, retains important information as much as possible, uses the output of each layer as the input of the previous layer, and finally uses a layer of BP neural network as the output layer. When the Bayesian method is used for fault diagnosis, the fault diagnosis can be divided into two stages: pre-training and tuning. In the pre-training stage, unlabeled samples are used as input of the network, initialization of parameters of a plurality of layers at the bottom is completed through a Bayesian regularization algorithm, and the algorithm flow is shown in FIG. 5. Firstly, inputting training sample data, taking a sample randomly selected from a training data set as an example, after pre-training is completed, the feature set of each layer can obtain parameter values to form a basic framework of the multi-Bayesian model, and in order to better optimize the parameters obtained by each layer, tuning is needed. And in the tuning stage, the label sample is adopted to fine tune the whole network containing the BP layer. The parameter value learned in the unsupervised learning stage (pre-training stage) is used as the initial value of the parameter in the supervised learning stage (tuning stage), and the parameter value is updated through training again, so that the parameter is further optimized, and the network is optimized. The basic flow of the two stages is as follows:
(1) selecting sample data, standardizing the sample data, and inputting the unlabeled sample data in the network.
(2) In unsupervised learning, each layer is subjected to Bayesian regularization training, and each parameter value of each layer is stored.
(3) In supervised learning, label samples are input, adjusted and optimized by using a BP algorithm, and parameter values are updated until the network is converged.
The method for diagnosing the fault of the switch cabinet based on the multi-Bayesian estimation comprises the following specific implementation steps:
(1) and determining output and input nodes of the network according to the feature vector structure and the fault mode classification. The number of input nodes is 3, which are respectively: an uhf signal state, an ultrasonic signal state, and a ground wave signal state. The output nodes are 5: normal operation, impending partial discharge, tip partial discharge, surface partial discharge and metal particle partial discharge;
(2) initializing parameters in the model, and setting the parameters as random small numerical values according to experience;
(3) selecting a switch cabinet signal sample data set, wherein the switch cabinet signal sample data set comprises a larger part of unmarked data samples and a smaller part of label sample data, standardizing the data, and dividing the label sample data into a tuning set and a test set according to a certain proportion;
(4) selecting unlabeled sample data as input of a training stage, and then carrying out fine tuning processing on the divided tuning set in the tuning stage so as to complete training of the whole multi-Bayesian model and store the trained multi-Bayesian network;
(5) and testing the saved multi-Bayesian network by the divided test set, saving the test result and completing diagnosis.
Finally, the above embodiments are only intended to illustrate the technical solutions of the present invention and not to limit the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all of them should be covered by the claims of the present invention.
Claims (10)
1. The utility model provides a cubical switchboard partial discharge trouble multisource information fusion detects early warning system based on wireless transmission technique which characterized in that: the system comprises a detection terminal, a power supply module, a wireless communication module and a system background module; the detection terminal is connected with the system background module through the wireless communication module;
the detection terminal comprises a transient ground voltage detection sensor, an ultrahigh frequency sensor and an ultrasonic sensor;
the system background module comprises a data processing module, a central processing unit, a storage module and a display module.
2. The multi-source information fusion detection early warning system for the local discharge fault of the switch cabinet based on the wireless transmission technology as claimed in claim 1, wherein: the transient ground voltage detection sensor is used for detecting signals including a ground voltage amplitude signal and a ground voltage pulse signal, the measured amplitude ranges from 0dBmV to 80dBmV, and the frequency ranges from 75MHz to 500 MHz.
3. The multi-source information fusion detection early warning system for the local discharge fault of the switch cabinet based on the wireless transmission technology as claimed in claim 1, wherein: the ultra-high frequency sensor is internally provided with an Archimedes spiral antenna, the working frequency band is 400 MHz-800 MHz, the voltage standing wave ratio in a pass band is less than 2, the total return loss in the pass band is less than-10 dB, and the receiving gain in the pass band is more than-60 dB.
4. The multi-source information fusion detection early warning system for the local discharge fault of the switch cabinet based on the wireless transmission technology as claimed in claim 1, wherein: the ultrasonic sensor is used for acquiring real-time pulses of ultrasonic signals and detecting the correlation of 50Hz and 100Hz pulse signals.
5. The multi-source information fusion detection early warning system for the local discharge fault of the switch cabinet based on the wireless transmission technology as claimed in claim 1, wherein: the wireless communication module adopts NB-IoT technology for communication, QPSK modulation and demodulation is adopted in the uplink, and BPSK or QPSK modulation and demodulation is adopted in the downlink.
6. The multi-source information fusion detection early warning system for the local discharge fault of the switch cabinet based on the wireless transmission technology as claimed in claim 1, wherein: the data processing module is used for performing wavelet denoising and filtering on the acquired signals, selecting a Daubechies wavelet basis as a basis function of wavelet packet decomposition, and filtering the data signals with the frequency meeting 400 MHz-800 MHz.
7. The multi-source information fusion detection early warning system for the local discharge fault of the switch cabinet based on the wireless transmission technology as claimed in claim 1, wherein: the central processing unit is used for realizing the processing and judgment of fault states and information, extracting and classifying the frequency band energy and time domain characteristics of the partial discharge signals through a BP neural network, realizing the fusion of multi-detection information by adopting a multi-Bayesian estimation method, establishing a partial discharge multi-signal fusion recognition model, judging whether each signal acquired at the detection position of the switch cabinet exceeds a safety threshold, and making response judgment and prediction.
8. A multi-source information fusion detection early warning method for local discharge faults of a switch cabinet based on a wireless transmission technology is characterized by comprising the following steps: the method comprises the following steps:
s1: the detection terminal judges whether an instruction for using a certain characteristic sensor is received; if receiving the instruction, executing the acquisition work of the corresponding signal according to the instruction; if no special instruction exists, the acquisition modules of the ground voltage, the ultrahigh frequency and the ultrasonic partial discharge signal work simultaneously, detect the signals at different positions of the switch cabinet respectively, and transmit the acquired signals to a data processing module of a system background in real time through the wireless transmission module;
s2: the data processing module carries out wavelet denoising and filtering processing on the discharge signal;
s3: the central processor performs characteristic acquisition and modeling analysis on the processed signals through a BP neural network and a multi-Bayesian estimation method, and judges the operation condition, fault sites and whether a partial discharge fault trend exists or not through comparison with an internal partial discharge characteristic signal information base and historical operation data of the switch cabinet;
s4: the signals and results processed by the central processor are stored in a memory connected with the central processor;
s5: the decision structure of the central processor is sent to the display module through the wireless transmission module, and if a partial discharge alarm signal is received, a partial discharge position is displayed; if the partial discharge alarm signal is not received, the prediction result of the future operation state of the switch cabinet is displayed.
9. The multi-source information fusion detection early warning method for the local discharge fault of the switch cabinet based on the wireless transmission technology, according to claim 8, is characterized in that: the BP neural network is respectively used as an identification classifier of ultrasonic wave, ultrahigh frequency and earth electric wave detection signals, the topological structure comprises three layers, namely an input layer, a hidden layer and an output layer, information is transmitted between the layers through the connection right of neurons, and the neurons in the same layer are mutually independent;
for the identification classification of the ultrasonic signal, 4 neurons of the input layer are provided, and correspond to the feature quantities of the ultrasonic signal: periodic maximum, effective, 50Hz correlation and 100Hz correlation; for the identification and classification of the ultrahigh frequency signals, the number of neurons of the input layer is 3, and the neurons correspond to the feature quantities of the ultrasonic signals: a periodic maximum, a valid, and a frequency; for the identification and classification of the earth electric wave signals, the number of neurons of an input layer is 8, the neurons respectively correspond to transient voltage signals at different positions of a switch cabinet, the number of neurons of an output layer corresponds to defect type codes and is set to be 2, and the neurons respectively correspond to a normal state and an abnormal state;
the operation process is as follows: the neuron of the input layer transmits the input characteristic quantity to the hidden layer through the connecting ring, the input information is continuously amplified, attenuated or inhibited, and the optimal connection weight is obtained in training, so that an optimal classifier model is established.
10. The multi-source information fusion detection early warning method for the local discharge fault of the switch cabinet based on the wireless transmission technology, according to claim 8, is characterized in that: the multi-Bayesian estimation method comprises the following steps:
(1) and determining output nodes and input nodes of the network according to the feature vector structure and the fault mode classification, wherein the number of the input nodes is 3, and the input nodes are respectively as follows: ultrahigh frequency signal state, ultrasonic signal state and earth electric wave signal state, output node is 5: normal operation, impending partial discharge, tip partial discharge, surface partial discharge and metal particle partial discharge;
(2) initializing parameters in the model, and setting the parameters as random small numerical values according to experience;
(3) selecting a switch cabinet signal sample data set, wherein the switch cabinet signal sample data set comprises a larger part of unmarked data samples and a smaller part of label sample data, standardizing the data, and dividing the label sample data into a tuning set and a test set according to a certain proportion;
(4) selecting unlabeled sample data as input of a training stage, and then carrying out fine tuning processing on the divided tuning set in the tuning stage so as to complete training of the whole multi-Bayesian model and store the trained multi-Bayesian network;
(5) and testing the saved multi-Bayesian network by the divided test set, saving the test result and completing diagnosis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010719186.1A CN111679166A (en) | 2020-07-23 | 2020-07-23 | Switch cabinet partial discharge fault multi-source information fusion detection early warning system and method based on wireless transmission technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010719186.1A CN111679166A (en) | 2020-07-23 | 2020-07-23 | Switch cabinet partial discharge fault multi-source information fusion detection early warning system and method based on wireless transmission technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111679166A true CN111679166A (en) | 2020-09-18 |
Family
ID=72457959
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010719186.1A Pending CN111679166A (en) | 2020-07-23 | 2020-07-23 | Switch cabinet partial discharge fault multi-source information fusion detection early warning system and method based on wireless transmission technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111679166A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112198400A (en) * | 2020-09-30 | 2021-01-08 | 国网福建省电力有限公司漳州供电公司 | High-voltage switch cabinet partial discharge online detection method based on spectrum sensing characteristic |
| CN112698160A (en) * | 2020-12-01 | 2021-04-23 | 广东电网有限责任公司广州供电局 | Switch cabinet partial discharge fault identification method and device, computer equipment and storage medium |
| CN112748317A (en) * | 2021-03-23 | 2021-05-04 | 国网河南省电力公司电力科学研究院 | Switch cabinet partial discharge fault detection method and system based on multiple monitoring data |
| CN112924833A (en) * | 2021-03-26 | 2021-06-08 | 国网新疆电力有限公司电力科学研究院 | Switch cabinet partial discharge detection device and method based on ultrasonic sensor array |
| CN113291350A (en) * | 2021-04-12 | 2021-08-24 | 中车唐山机车车辆有限公司 | EMUs high-pressure tank early warning device and EMUs that discharges |
| CN113484698A (en) * | 2021-05-21 | 2021-10-08 | 国网浙江省电力有限公司湖州供电公司 | Partial discharge detection system based on environmental data feedback |
| CN113568353A (en) * | 2021-07-30 | 2021-10-29 | 国网四川省电力公司电力科学研究院 | Switch cabinet partial discharge monitoring and early warning method based on reduced state feature set |
| CN113706840A (en) * | 2021-08-12 | 2021-11-26 | 广东电网有限责任公司广州供电局 | Partial discharge ultrahigh frequency monitoring grading alarm circuit and alarm device and method thereof |
| CN114441914A (en) * | 2022-04-08 | 2022-05-06 | 山东中安电力科技有限公司 | Switch cabinet partial discharge detection system based on data processing |
| CN114878994A (en) * | 2022-07-11 | 2022-08-09 | 杭州世创电子技术股份有限公司 | Partial discharge signal detection method and system based on space ultrahigh frequency sensor |
| CN115865043A (en) * | 2022-11-04 | 2023-03-28 | 浙江黑卡电气有限公司 | Dynamic denoising circuit of multi-dimensional signal and processing method thereof |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105676085A (en) * | 2016-01-31 | 2016-06-15 | 国家电网公司 | Extra-high voltage GIS partial discharge detection method based on multi-sensor information fusion |
| CN109444682A (en) * | 2018-11-02 | 2019-03-08 | 国网四川省电力公司广安供电公司 | The construction method of partial discharge of switchgear diagnostic system based on multi-information fusion |
| CN110687404A (en) * | 2019-08-27 | 2020-01-14 | 西安电子科技大学 | Portable partial discharge detection equipment |
-
2020
- 2020-07-23 CN CN202010719186.1A patent/CN111679166A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105676085A (en) * | 2016-01-31 | 2016-06-15 | 国家电网公司 | Extra-high voltage GIS partial discharge detection method based on multi-sensor information fusion |
| CN109444682A (en) * | 2018-11-02 | 2019-03-08 | 国网四川省电力公司广安供电公司 | The construction method of partial discharge of switchgear diagnostic system based on multi-information fusion |
| CN110687404A (en) * | 2019-08-27 | 2020-01-14 | 西安电子科技大学 | Portable partial discharge detection equipment |
Non-Patent Citations (1)
| Title |
|---|
| 张勇: "顶板动态监测监控及信息融合技术研究", 煤炭工业出版社, pages: 26 - 31 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112198400A (en) * | 2020-09-30 | 2021-01-08 | 国网福建省电力有限公司漳州供电公司 | High-voltage switch cabinet partial discharge online detection method based on spectrum sensing characteristic |
| CN112198400B (en) * | 2020-09-30 | 2024-04-09 | 国网福建省电力有限公司漳州供电公司 | High-voltage switch cabinet partial discharge online detection method based on frequency spectrum sensing characteristics |
| CN112698160A (en) * | 2020-12-01 | 2021-04-23 | 广东电网有限责任公司广州供电局 | Switch cabinet partial discharge fault identification method and device, computer equipment and storage medium |
| CN112748317B (en) * | 2021-03-23 | 2022-03-25 | 国网河南省电力公司电力科学研究院 | Switch cabinet partial discharge fault detection method and system based on multiple monitoring data |
| CN112748317A (en) * | 2021-03-23 | 2021-05-04 | 国网河南省电力公司电力科学研究院 | Switch cabinet partial discharge fault detection method and system based on multiple monitoring data |
| CN112924833A (en) * | 2021-03-26 | 2021-06-08 | 国网新疆电力有限公司电力科学研究院 | Switch cabinet partial discharge detection device and method based on ultrasonic sensor array |
| CN113291350A (en) * | 2021-04-12 | 2021-08-24 | 中车唐山机车车辆有限公司 | EMUs high-pressure tank early warning device and EMUs that discharges |
| CN113291350B (en) * | 2021-04-12 | 2022-10-14 | 中车唐山机车车辆有限公司 | EMUs high-pressure tank early warning device and EMUs that discharges |
| CN113484698A (en) * | 2021-05-21 | 2021-10-08 | 国网浙江省电力有限公司湖州供电公司 | Partial discharge detection system based on environmental data feedback |
| CN113568353A (en) * | 2021-07-30 | 2021-10-29 | 国网四川省电力公司电力科学研究院 | Switch cabinet partial discharge monitoring and early warning method based on reduced state feature set |
| CN113706840A (en) * | 2021-08-12 | 2021-11-26 | 广东电网有限责任公司广州供电局 | Partial discharge ultrahigh frequency monitoring grading alarm circuit and alarm device and method thereof |
| CN113706840B (en) * | 2021-08-12 | 2023-08-08 | 广东电网有限责任公司广州供电局 | Partial discharge ultrahigh frequency monitoring and grading alarm circuit and alarm device and method thereof |
| CN114441914A (en) * | 2022-04-08 | 2022-05-06 | 山东中安电力科技有限公司 | Switch cabinet partial discharge detection system based on data processing |
| CN114441914B (en) * | 2022-04-08 | 2022-06-14 | 山东中安电力科技有限公司 | Switch cabinet partial discharge detection system based on data processing |
| CN114878994A (en) * | 2022-07-11 | 2022-08-09 | 杭州世创电子技术股份有限公司 | Partial discharge signal detection method and system based on space ultrahigh frequency sensor |
| CN114878994B (en) * | 2022-07-11 | 2022-09-27 | 杭州世创电子技术股份有限公司 | Partial discharge signal detection method and system based on space ultrahigh frequency sensor |
| CN115865043A (en) * | 2022-11-04 | 2023-03-28 | 浙江黑卡电气有限公司 | Dynamic denoising circuit of multi-dimensional signal and processing method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111679166A (en) | Switch cabinet partial discharge fault multi-source information fusion detection early warning system and method based on wireless transmission technology | |
| CN106556781A (en) | Shelf depreciation defect image diagnostic method and system based on deep learning | |
| WO2019015311A1 (en) | Vibration signal support vector machine-based gil fault online monitoring system | |
| CN111523660A (en) | Audio-visual-thermal integrated anomaly detection and alarm method based on artificial intelligence | |
| CN107390097A (en) | A kind of acoustoelectric combined shelf depreciation simulation detection system of GIS and its detection method | |
| CN108802525A (en) | Equipment fault intelligent Forecasting based on small sample | |
| CN102841297B (en) | Self-adaptive partial-discharge monitoring method and device for sampling frequency | |
| CN109649432B (en) | System and method for monitoring integrity of steel rail of cloud platform based on guided wave technology | |
| CN103323749A (en) | Multi-classifier information fusion partial discharge diagnostic method | |
| CN110163075A (en) | A kind of multi-information fusion method for diagnosing faults based on Weight Training | |
| CN105203929A (en) | Fixed switch cabinet partial discharge online detection device and method | |
| CN102023278A (en) | Method and device for identifying partial discharge interference of GIS equipment | |
| CN112307950A (en) | Method for extracting and intelligently analyzing detail characteristic quantity of GIS vibration defect identification | |
| Shakiba et al. | Application of machine learning methods in fault detection and classification of power transmission lines: a survey | |
| CN105300528A (en) | Infrared image diagnosis method and infrared image diagnosis system for transformer station equipment | |
| CN110703078A (en) | GIS fault diagnosis method based on spectral energy analysis and self-organizing competition algorithm | |
| CN112986870A (en) | Distributed power transformer winding state monitoring method and system based on vibration method | |
| CN115797772A (en) | Classified early warning method and system for icing thickness of overhead transmission line | |
| CN108508297A (en) | A kind of fault arc detection method based on change coefficient and SVM | |
| Chang et al. | Assessment of the insulation status aging in power cable joints using support vector machine | |
| CN116070140B (en) | Power distribution substation safe operation state monitoring system and method | |
| CN116244600A (en) | Method, system and equipment for constructing GIS intermittent discharge mode identification model | |
| CN112526404A (en) | Power transformer fault on-line positioning system based on ultrahigh frequency external sensor | |
| CN114062839A (en) | Railway power line fault positioning device and method thereof | |
| Balouji et al. | A deep learning approach to earth fault classification and source localization |
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 |