CN111579978B - System and method for realizing relay fault identification based on artificial intelligence technology - Google Patents
System and method for realizing relay fault identification based on artificial intelligence technology Download PDFInfo
- Publication number
- CN111579978B CN111579978B CN202010420948.8A CN202010420948A CN111579978B CN 111579978 B CN111579978 B CN 111579978B CN 202010420948 A CN202010420948 A CN 202010420948A CN 111579978 B CN111579978 B CN 111579978B
- Authority
- CN
- China
- Prior art keywords
- layer
- output
- relay
- sensor
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 43
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000005516 engineering process Methods 0.000 title claims abstract description 22
- 238000012544 monitoring process Methods 0.000 claims abstract description 7
- 238000004891 communication Methods 0.000 claims description 31
- 238000013528 artificial neural network Methods 0.000 claims description 30
- 230000006870 function Effects 0.000 claims description 25
- 230000005540 biological transmission Effects 0.000 claims description 23
- 230000004913 activation Effects 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 7
- 239000003990 capacitor Substances 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 238000003062 neural network model Methods 0.000 claims description 4
- 230000001133 acceleration Effects 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000003066 decision tree Methods 0.000 claims description 3
- 238000005265 energy consumption Methods 0.000 claims description 3
- 230000005855 radiation Effects 0.000 claims description 3
- 238000013461 design Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 4
- 230000005611 electricity Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 6
- 230000007257 malfunction Effects 0.000 description 6
- 230000003750 conditioning effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000001143 conditioned effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
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
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
- G01R31/3277—Testing of circuit interrupters, switches or circuit-breakers of low voltage devices, e.g. domestic or industrial devices, such as motor protections, relays, rotation switches
- G01R31/3278—Testing of circuit interrupters, switches or circuit-breakers of low voltage devices, e.g. domestic or industrial devices, such as motor protections, relays, rotation switches of relays, solenoids or reed switches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Remote Monitoring And Control Of Power-Distribution Networks (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The invention discloses a system and a method for realizing relay fault identification based on an artificial intelligence technology, which relate to the technical field of electricity, and solve the technical problem that the relay can be monitored online, remotely and in real time in the operation process of the relay. The invention can realize the monitoring of the working state of the relay without a user on the working site, has high intelligent degree and high precision, and has better practical value.
Description
Technical Field
The invention relates to the technical field of electricity, in particular to a system and a method for realizing relay fault identification based on an artificial intelligence technology.
Background
Relay protection equipment is vital equipment in a power system, and can cut off faults when the power system breaks down so as to ensure reliable and stable operation of the power system. Therefore, in order to ensure the reliability and stability of the operation of the power system, the relay protection equipment is required to have measures for preventing the malfunction and the malfunction in configuration and function, so that the malfunction and the malfunction of any unit cannot be caused by the malfunction and the malfunction of the relay protection equipment.
Patent application number CN201910068008.4 discloses a relay protection device control method and device, which solve the problem of poor relay protection reliability caused by judging according to a single data source when the relay protection device is controlled to act, but have poor intelligence and low precision.
Patent application number CN201621129925.7 discloses an automatic switching device of relay protection outlet loop, when different relay protection devices need outlet transmission and only one breaker mechanism is actually used, the switching device is utilized to automatically switch operation power supplies, so that the operation power supplies of the different relay protection devices are prevented from being mutually connected, and meanwhile, the information of which relay protection device is switched to is transmitted to a monitoring background, thereby being convenient for training and guiding a teacher to control the whole group of transmission tests of on-site training staff, and still not involving intelligent and automatic processing and being accurate
Along with the development of electronic technology and artificial intelligence technology, the reliable operation of a power system is one of the basis and most important technical indexes for realizing the smooth operation of a power grid system, and in order to avoid the loss caused by personnel, power grid, equipment and the like, an artificial intelligence control method is needed to improve the control precision.
Disclosure of Invention
Aiming at the defects of the prior art, the invention discloses a system and a method for realizing relay fault identification based on an artificial intelligence technology, which can realize relay protection in power equipment, and have good intelligent degree and high precision.
The invention adopts the following technical scheme:
a system for implementing relay fault identification based on artificial intelligence techniques, comprising:
the relay state acquisition layer is internally provided with terminal equipment, the terminal equipment is internally provided with a relay, and the relay is connected with a sensor or an overcurrent detection circuit, wherein the sensor at least comprises a photoelectric sensor, an infrared sensor, a speed sensor, an acceleration sensor, a GIS sensor, a vibration sensor, a ripple sensor, a temperature and humidity sensor, an angle sensor, a magnetic field sensor, a rotating speed sensor, an RFID tag, a GPS device, a ray radiation sensor, a heat-sensitive sensor or an energy consumption sensor; the overcurrent detection circuit at least comprises an I/V converter, an amplifier, an integrator, a comparator, a timer, a converter and an output relay, wherein secondary current output by the converter converts an output signal into a voltage signal through the I/V converter, and then the voltage is subjected to direct current output. Comparing the voltage comparator with a manually-set threshold value, and outputting an identification signal if the output direct-current voltage is greater than the manually-set threshold value;
the data transmission layer is internally provided with a wired communication module or a wireless communication module and is used for receiving and transmitting relay state data information perceived by the relay state acquisition layer; wherein: the wired communication module at least comprises an RS485 communication module or an RS232 communication module, and the wireless communication module at least comprises a TCP/IP network system, a ZigBee wireless network, a GPRS communication module or a CDMA wireless communication or Bluetooth communication module;
the system comprises an artificial intelligent learning layer, a relay state acquisition layer, a control layer and a control layer, wherein a computer processing system is arranged in the artificial intelligent learning layer, the computer processing system is provided with an Apriori association rule algorithm model and an improved BP neural network algorithm unit, the Apriori association rule algorithm model is used for searching computer information perceived by the relay state acquisition layer according to different relay fault attributes so as to search an optimal solution, and the attributes at least comprise fault classification such as current, voltage, harmonic wave, magnetic field, vibration, overcurrent, overvoltage or manual operation error information; the improved BP neural network algorithm unit searches the hypothesis space of possible weight vectors by using gradient descent to find the weight vector of the best fitting sample, and moves towards the negative gradient direction of the loss function each time by using the loss function until the loss function obtains the minimum value, so as to improve the artificial intelligence learning precision and the learning accuracy; the output end of the Apriori association rule algorithm model is connected with the output end of the improved BP neural network algorithm unit, the Apriori association rule algorithm model is a decision tree Apriori association rule algorithm model, the improved BP neural network algorithm unit comprises an input layer, a hidden layer, an output layer and an activation layer, the output end of the input layer is connected with the input end of the hidden layer, the output end of the hidden layer is connected with the input end of the output layer, and the input end of the activation layer is respectively connected with the output ends of the input layer, the hidden layer and the output layer;
the data application layer is internally provided with a computer with a remote data transmission interface, and the data information output by the artificial intelligence learning layer is researched, investigated, controlled or applied by the computer;
the remote scheduling layer is internally provided with a remote management computer, the remote management computer is provided with a remote management data receiving and transmitting port, and receives data information output by the data application layer; wherein:
the state database output interface in the relay state acquisition layer is connected with the input end of the data transmission layer, the output end of the data transmission layer is connected with the input end of the artificial intelligence learning layer, the output end of the artificial intelligence learning layer is connected with the input end of the data application layer, and the output end of the data application layer is connected with the input end of the remote scheduling layer.
Further, the overcurrent detection circuit further comprises a current acquisition circuit.
Further, the current sampling circuit comprises a current transformer, the output end of the current transformer is connected with a three-stage LM224 amplifying circuit, and the output end of the three-stage LM224 amplifying circuit is connected with the A/D conversion unit.
The invention also adopts the following technical scheme:
a method for implementing relay fault identification based on artificial intelligence technology, the method comprising the steps of:
(S1) acquiring relay fault data information through a relay state acquisition layer;
(S2) transmitting data information of the relay state acquisition layer through the data transmission layer;
(S3) the artificial intelligence learning layer receives the data information transmitted by the data transmission layer and learns the received data information;
(S4) the data application layer applies the data information output by the artificial intelligence learning layer;
and S5, the remote scheduling layer and the data application layer carry out data communication, so that remote transmission, monitoring and application of data are realized.
Further, the artificial intelligence learning is a learning mode based on an Apriori association rule algorithm and a BP neural network algorithm.
Further, the method of the Apriori association rule algorithm comprises the following steps:
scanning a relay list in an operation working state, and then obtaining parameters of all relay operation states according to the arrangement mode of the style parameters;
according to each attribute value of the relay running state, the attribute values at least comprise working voltage, working current, ripple wave, magnetic field and manual operation, and the weighted value of each attribute value is calculated by using a hierarchical analysis method;
according to the manually set weighting values, calculating the weighting value of each attribute value warning set t, which can be represented by a formula (1);
(4) Calculating the weighted support degree of each attribute value warning according to the weight of each attribute value set, and expressing the weighted support degree by a formula 2;
generating a weighted warning frequent k item set according to a preset minimum support threshold;
(5) Generating a candidate k+1 item set of the warning item by adopting an optimized splicing and branch reducing method according to the prior property of the network weighted item set, calculating the weighted support degree of the candidate warning k+1 item set, and generating a weighted warning frequent k+1 item set;
(6) And (4) repeating the step until the alarm frequent item set cannot be continuously generated, and finally obtaining the optimal solution.
k is between 50 and 100.
Further, the BP neural network algorithm adjusts a model output result by adjusting a weight or a threshold value in the BP neural network model, and then achieves the aim of gradually approaching the model output result, so that an output error can be minimized, and the BP neural network algorithm comprises a forward propagation algorithm model and a backward propagation algorithm model.
Further, the working method of the forward propagation algorithm model is as follows:
assuming that the data input from the input layer is X, the input layer to hidden layer parameters are ω, b 1 The parameters from the hidden layer to the output layer are v, b 2 The activation function is denoted as g 1 、g 2 The model is set as:
the output formula from the input layer to the hidden layer is:
net 1 ＝ω T x+b 1 ,h＝g 1 (net 1 ) (3)
the output formula from the hidden layer to the output layer is:
then there are:
the loss function is:
where k=2.
Further, in the back propagation algorithm model:
the formula for adjusting the weight coefficient of the output layer is as follows:
the implicit layer weight coefficient is adjusted using the following formula:
in addition, in different relay fault data samples, as the input modes are different [22], the corresponding quadratic accurate function models are also different, and the method is expressed by the following formulas:
for N relay fault information samples, the following formula is used for expressing a total accurate function expression:
with the letter Deltaomega ki Output layer weight coefficient for representing BP neural network algorithm model by letterInformation expected output value representing BP neural network algorithm model, using the letter +.>The calculation output is expressed as BP neural network algorithm model, the calculation output is expressed as a constant by letters eta, and the value of eta is between 0.1 and 2.8.
Has the positive beneficial effects that:
the invention can realize the information acquisition and identification of relay faults, can realize the on-line, remote and real-time monitoring of data in the running process, and improves the working performance of the relay. According to the invention, the artificial intelligent algorithm is integrated with the remote communication technology and the online monitoring technology, so that a user can monitor the working state of the relay without being on a working site, the intelligent degree is high, the precision is high, and the practical value is good.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic diagram of a relay fault recognition system based on artificial intelligence technology
FIG. 2 is a schematic diagram of an overcurrent detection architecture in a relay fault recognition system based on artificial intelligence technology according to the present invention;
FIG. 3 is a schematic diagram of an overcurrent protection circuit in a relay fault recognition system based on artificial intelligence technology according to the present invention;
FIG. 4 is a schematic diagram of an embodiment of a relay fault recognition system implemented based on artificial intelligence technology according to the present invention;
FIG. 5 is a schematic flow chart of a relay fault recognition method based on artificial intelligence technology;
FIG. 6 is a schematic diagram of an artificial intelligence learning algorithm in a relay fault recognition method based on an artificial intelligence technique;
fig. 7 is a schematic diagram of a forward algorithm model in a BP neural network algorithm model in a relay fault recognition method based on an artificial intelligence technology.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1 System
As shown in fig. 1-4, a system for implementing relay fault recognition based on artificial intelligence technology includes:
the relay state acquisition layer is internally provided with terminal equipment, the terminal equipment is internally provided with a relay, and the relay is connected with a sensor or an overcurrent detection circuit, wherein the sensor at least comprises a photoelectric sensor, an infrared sensor, a speed sensor, an acceleration sensor, a GIS sensor, a vibration sensor, a ripple sensor, a temperature and humidity sensor, an angle sensor, a magnetic field sensor, a rotating speed sensor, an RFID tag, a GPS device, a ray radiation sensor, a heat-sensitive sensor or an energy consumption sensor; the overcurrent detection circuit at least comprises an I/V converter, an amplifier, an integrator, a comparator, a timer, a converter and an output relay, wherein secondary current output by the converter converts an output signal into a voltage signal through the I/V converter, and then the voltage is subjected to direct current output. Comparing the voltage comparator with a manually-set threshold value, and outputting an identification signal if the output direct-current voltage is greater than the manually-set threshold value;
the data transmission layer is internally provided with a wired communication module or a wireless communication module and is used for receiving and transmitting relay state data information perceived by the relay state acquisition layer; wherein: the wired communication module at least comprises an RS485 communication module or an RS232 communication module, and the wireless communication module at least comprises a TCP/IP network system, a ZigBee wireless network, a GPRS communication module or a CDMA wireless communication or Bluetooth communication module;
the system comprises an artificial intelligent learning layer, a relay state acquisition layer, a control layer and a control layer, wherein a computer processing system is arranged in the artificial intelligent learning layer, the computer processing system is provided with an Apriori association rule algorithm model and an improved BP neural network algorithm unit, the Apriori association rule algorithm model is used for searching computer information perceived by the relay state acquisition layer according to different relay fault attributes so as to search an optimal solution, and the attributes at least comprise fault classification such as current, voltage, harmonic wave, magnetic field, vibration, overcurrent, overvoltage or manual operation error information; the improved BP neural network algorithm unit searches the hypothesis space of possible weight vectors by using gradient descent to find the weight vector of the best fitting sample, and moves towards the negative gradient direction of the loss function each time by using the loss function until the loss function obtains the minimum value, so as to improve the artificial intelligence learning precision and the learning accuracy; the output end of the Apriori association rule algorithm model is connected with the output end of the improved BP neural network algorithm unit, the Apriori association rule algorithm model is a decision tree Apriori association rule algorithm model, the improved BP neural network algorithm unit comprises an input layer, a hidden layer, an output layer and an activation layer, the output end of the input layer is connected with the input end of the hidden layer, the output end of the hidden layer is connected with the input end of the output layer, and the input end of the activation layer is respectively connected with the output ends of the input layer, the hidden layer and the output layer;
the data application layer is internally provided with a computer with a remote data transmission interface, and the data information output by the artificial intelligence learning layer is researched, investigated, controlled or applied by the computer;
the remote scheduling layer is internally provided with a remote management computer, the remote management computer is provided with a remote management data receiving and transmitting port, and receives data information output by the data application layer; wherein:
the state database output interface in the relay state acquisition layer is connected with the input end of the data transmission layer, the output end of the data transmission layer is connected with the input end of the artificial intelligence learning layer, the output end of the artificial intelligence learning layer is connected with the input end of the data application layer, and the output end of the data application layer is connected with the input end of the remote scheduling layer.
Further, the overcurrent detection circuit further comprises a current acquisition circuit.
Further, the current sampling circuit comprises a current transformer, the output end of the current transformer is connected with a three-stage LM224 amplifying circuit, and the output end of the three-stage LM224 amplifying circuit is connected with the A/D conversion unit. When the current sampling design is carried out, a current sensor is adopted at the input end to carry out current input, the input from large current to small alternating voltage is realized through a current transformer, and the A/D module input into the microcontroller is subjected to filtering, amplifying and outputting through a signal conditioning circuit. As shown in fig. 5, R1 and C1 are resistances and capacitances obtained by integrating current transformers. The output voltage signal of R1 is amplified by an amplifying circuit constituted by the operational amplifier LM 224U 1A, U B, and the amplified signal generates an ac voltage of about 2.5V. In the figure, the capacitor C3, the resistor R7, the capacitor C4, and the resistor R6 realize low-pass filtering and high-pass filtering, and the cutoff frequency thereof is 1/2 pi RC. By mathematical calculation, the frequency is selected in the frequency range between 10.61Hz-123.72 Hz. The LM 224U1C amplifier is mainly used for providing a reference dc voltage to the ac voltage generated in the forward direction, and superposing the ac voltage signal and the dc voltage signal to finally increase the voltage value. The information output by V0 is changed into unipolar voltage after being conditioned by a conditioning circuit, and then the voltage is directly input into an A/D unit (LPC 2294), so that current sampling is realized.
Example 2 method
As shown in fig. 5-7, a method for implementing relay fault recognition based on artificial intelligence technology includes the following steps:
(S1) acquiring relay fault data information through a relay state acquisition layer;
(S2) transmitting data information of the relay state acquisition layer through the data transmission layer;
(S3) the artificial intelligence learning layer receives the data information transmitted by the data transmission layer and learns the received data information;
(S4) the data application layer applies the data information output by the artificial intelligence learning layer;
and S5, the remote scheduling layer and the data application layer carry out data communication, so that remote transmission, monitoring and application of data are realized.
In the above embodiment, the artificial intelligence learning is a learning mode based on an Apriori association rule algorithm and a BP neural network algorithm.
In the above embodiment, the method of the Apriori association rule algorithm is as follows:
(1) Scanning a relay list in an operation working state, and then obtaining parameters of all relay operation states according to the arrangement mode of the style parameters;
(2) According to each attribute value of the relay running state, the attribute values at least comprise working voltage, working current, ripple wave, magnetic field and manual operation, and the weighted value of each attribute value is calculated by using a hierarchical analysis method;
(3) According to the manually set weighting values, calculating the weighting value of each attribute value warning set t, which can be represented by a formula (1);
(4) Calculating the weighted support degree of each attribute value warning according to the weight of each attribute value set, and expressing the weighted support degree by a formula 2;
generating a weighted warning frequent k item set according to a preset minimum support threshold;
(5) Generating a candidate k+1 item set of the warning item by adopting an optimized splicing and branch reducing method according to the prior property of the network weighted item set, calculating the weighted support degree of the candidate warning k+1 item set, and generating a weighted warning frequent k+1 item set;
(6) And step four, repeating until the alarm frequent item set cannot be continuously generated, and finally obtaining the optimal solution.
The method for realizing relay fault recognition based on the artificial intelligence technology according to claim 6, wherein the method comprises the following steps: k is between 50 and 100.
In the above embodiment, the BP neural network algorithm adjusts the model output result by adjusting the weight or the threshold in the BP neural network model, and then achieves the purpose of gradually approximating the model output result, so that the output error can be minimized, and the BP neural network algorithm includes a forward propagation algorithm model and a backward propagation algorithm model.
In the above embodiment, the working method of the forward propagation algorithm model is as follows:
assuming that the data input from the input layer is X, the input layer to hidden layer parameters are ω, b 1 The parameters from the hidden layer to the output layer are v, b 2 The activation function is denoted as g 1 、g 2 The model is set as:
the output formula from the input layer to the hidden layer is:
net 1 ＝ω T x+b 1 ,h＝g 1 (net 1 ) (3)
the output formula from the hidden layer to the output layer is:
then there are:
the loss function is:
where k=2.
The data is subjected to linear transformation processing of weight values and bias items from an input layer, and then the output of a hidden layer, namely the input of the next layer, is obtained through an activation layer; the hidden layer is to the output layer, through the linear transformation of the weight value and the bias item, then through the activation layer, the output layer is obtained.
Further, in the back propagation algorithm model:
the formula for adjusting the weight coefficient of the output layer is as follows:
the implicit layer weight coefficient is adjusted using the following formula:
in addition, in different relay fault data samples, the corresponding quadratic accurate function models are different due to different input modes, and the quadratic accurate function models are expressed by the following formulas:
for N relay fault information samples, the following formula is used for expressing a total accurate function expression:
with the letter Deltaomega ki Output layer weight coefficient for representing BP neural network algorithm model by letterInformation expected output value representing BP neural network algorithm model, using the letter +.>The calculation output is expressed as BP neural network algorithm model, the calculation output is expressed as a constant by letters eta, and the value of eta is between 0.1 and 2.8.
When the above formula is used for evaluation, particularly when the fault type information of the complex relay is extracted, the acquired fault type information is required to be evaluated in terms of further improving learning accuracyIs normalized. In the normalization process, if m kinds of relay fault information are input in the model and N samples are number, the input data x is calculated ij The normalization process is performed by the following formula:
in the above formula, i=1, 2, …, N; j=1, 2 …, m, Z in the above formula ij The normalized data is obtained.
The normalized following formula may be:
wherein y is i Outputting a relay fault data sample;
y′ i the information is standardized relay fault data sample information;
y max to output the maximum value in the relay fault information data sample, y min Outputting a minimum value in a relay fault information data sample;
wherein 0.2< q <1.3;0.1< b <1.4, more preferably 0.25< q <1.1;0.15< b <1.3,
then the number of hidden layer nodes is determined to be between 6 and 10, more preferably between 8 and 9, the numerical value from the input layer to the hidden layer is between 0.2 and 0.6, more preferably between 0.25 and 0.45, and the numerical value from the hidden layer to the output layer can be between 0.08 and 0.32, more preferably between 0.1 and 0.25, and a BP neural network model is built in this way, so that the relay fault data type can be quickly identified.
While specific embodiments of the present invention have been described above, it will be understood by those skilled in the art that these specific embodiments are by way of example only, and that various omissions, substitutions, and changes in the form and details of the methods and systems described above may be made by those skilled in the art without departing from the spirit and scope of the invention. For example, it is within the scope of the present invention to combine the above-described method steps to perform substantially the same function in substantially the same way to achieve substantially the same result. Accordingly, the scope of the invention is limited only by the following claims.
Claims (3)
1. A system for realizing relay fault recognition based on artificial intelligence technology is characterized in that: comprising the following steps:
the relay state acquisition layer is internally provided with terminal equipment, the terminal equipment is internally provided with a relay, and the relay is connected with a sensor or an overcurrent detection circuit, wherein the sensor at least comprises a photoelectric sensor, an infrared sensor, a speed sensor, an acceleration sensor, a GIS sensor, a vibration sensor, a ripple sensor, a temperature and humidity sensor, an angle sensor, a magnetic field sensor, a rotating speed sensor, an RFID tag, a GPS device, a ray radiation sensor, a heat-sensitive sensor or an energy consumption sensor; the overcurrent detection circuit at least comprises an I/V converter, an amplifier, an integrator, a comparator, a timer, a converter and an output relay, wherein secondary current output by the converter converts an output signal into a voltage signal through the I/V converter, and then the voltage is subjected to direct current output; comparing the voltage comparator with a manually-set threshold value, and outputting an identification signal if the output direct-current voltage is greater than the manually-set threshold value;
the data transmission layer is internally provided with a wired communication module or a wireless communication module and is used for receiving and transmitting relay state data information perceived by the relay state acquisition layer; wherein: the wired communication module at least comprises an RS485 communication module or an RS232 communication module, and the wireless communication module at least comprises a TCP/IP network system, a ZigBee wireless network, a GPRS communication module or a CDMA wireless communication or Bluetooth communication module;
the system comprises an artificial intelligent learning layer, a computer processing system and a BP neural network algorithm unit, wherein the computer processing system is provided with an Apriori association rule algorithm model and the BP neural network algorithm unit, the Apriori association rule algorithm model is used for searching computer information perceived by the relay state acquisition layer according to different relay fault attributes so as to search an optimal solution, and the attributes at least comprise fault classification such as current, voltage, harmonic wave, magnetic field, vibration, overcurrent, overvoltage or manual misoperation information; the BP neural network algorithm unit searches a possible weight vector hypothesis space by using gradient descent to find the weight vector of the best fitting sample, and moves towards the negative gradient direction of the loss function each time by using the loss function until the loss function obtains the minimum value, so that the artificial intelligence learning precision is improved, and the learning accuracy is improved; the output end of the Apriori association rule algorithm model is connected with the input end of the BP neural network algorithm unit, the Apriori association rule algorithm model is a decision tree Apriori association rule algorithm model, the BP neural network algorithm unit comprises an input layer, a hidden layer, an output layer and an activation layer, the output end of the input layer is connected with the input end of the hidden layer, the output end of the hidden layer is connected with the input end of the output layer, and the input end of the activation layer is respectively connected with the output ends of the input layer, the hidden layer and the output layer;
the data application layer is internally provided with a computer with a remote data transmission interface, and the data information output by the artificial intelligence learning layer is researched, investigated, controlled or applied by the computer;
the remote scheduling layer is internally provided with a remote management computer, the remote management computer is provided with a remote management data receiving and transmitting port, and receives data information output by the data application layer; wherein:
the output interface of the state database in the relay state acquisition layer is connected with the input end of the data transmission layer, the output end of the data transmission layer is connected with the input end of the artificial intelligence learning layer, the output end of the artificial intelligence learning layer is connected with the input end of the data application layer, and the output end of the data application layer is connected with the input end of the remote scheduling layer; the overcurrent detection circuit further comprises a current sampling circuit; the current sampling circuit comprises a current transformer, the output end of the current transformer is connected with a three-stage LM224 amplifying circuit, and the output end of the three-stage LM224 amplifying circuit is connected with an A/D conversion unit; when the current sampling design is carried out, a current sensor is adopted at the input end to carry out current input, and the input from large current to small alternating voltage is realized through a current transformer; the resistor R1 and the capacitor C1 are resistors and capacitors which are integrated by adopting a current transformer, an output voltage signal of the resistor R1 is amplified by an operational amplifier, an amplifying circuit formed by LM 224U 1A, LM U1B is used for amplifying the amplified signal to generate 2.5V alternating voltage, the capacitor C3, the resistor R7, the capacitor C4 and the resistor R6 realize low-pass filtering and high-pass filtering, and the cutoff frequency is 1/2 pi RC; the LM 224U1C amplifier is used for providing a reference direct-current voltage for the alternating-current voltage generated forward, superposing an alternating-current voltage signal and a direct-current voltage signal and finally improving the value of the voltage;
the method of the Apriori association rule algorithm comprises the following steps:
(1) Scanning a relay list in an operation working state, and then obtaining parameters of all relay operation states according to the arrangement mode of the style parameters;
(2) According to each attribute value of the relay running state, the attribute values at least comprise working voltage, working current, ripple wave, magnetic field and manual operation, and the weighted value of each attribute value is calculated by using a hierarchical analysis method;
(3) According to the manually set weighting values, calculating the weighting value of each attribute value warning set t, which can be represented by a formula (1);
(4) Calculating the weighted support degree of each attribute value warning according to the weight of each attribute value set, and expressing the weighted support degree by a formula 2;
generating a weighted warning frequent k item set according to a preset minimum support threshold;
(5) Generating a candidate k+1 item set of the warning item by adopting an optimized splicing and branch reducing method according to the prior property of the network weighted item set, calculating the weighted support degree of the candidate warning k+1 item set, and generating a weighted warning frequent k+1 item set;
(6) Repeating the fourth step until the warning frequent item set cannot be continuously generated, and finally obtaining an optimal solution;
the BP neural network algorithm adjusts a model output result by adjusting a weight or a threshold value in the BP neural network model, then achieves the aim of gradually approaching the model output result, so that an output error can be minimized, and comprises a forward propagation algorithm model and a backward propagation algorithm model;
the working method of the forward propagation algorithm model comprises the following steps:
assuming that the data input from the input layer is x, the input layer to hidden layer parameters are ω, b 1 The parameters from the hidden layer to the output layer are v, b 2 The activation function is denoted as g 1 、g 2 The model is set as:
the output formula from the input layer to the hidden layer is:
net 1 ＝ω T x+b 1 ,h＝g 1 (net 1 ) (3)
the output formula from the hidden layer to the output layer is:
then there are:
the loss function is:
wherein k=2;
in the back propagation algorithm model:
the formula for adjusting the weight coefficient of the output layer is as follows:
the implicit layer weight coefficient is adjusted using the following formula:
in addition, in different relay fault data samples, the corresponding quadratic accurate function models are different due to different input modes, and the quadratic accurate function models are expressed by the following formulas:
for N relay fault information samples, the following formula is used for expressing a total accurate function expression:
with the letter Deltaomega ki Output layer weight coefficient for representing BP neural network algorithm model by letterInformation expected output value representing BP neural network algorithm model, using the letter +.>The calculation output is expressed as BP neural network algorithm model, the calculation output is expressed as a constant by letters eta, and the value of eta is between 0.1 and 2.8.
2. A method for realizing relay fault identification based on artificial intelligence technology, which is applied to the system of claim 1, and is characterized in that: the method comprises the following steps: (S1) acquiring relay fault data information through a relay state acquisition layer;
(S2) transmitting data information of the relay state acquisition layer through the data transmission layer;
(S3) the artificial intelligence learning layer receives the data information transmitted by the data transmission layer and learns the received data information;
(S4) the data application layer applies the data information output by the artificial intelligence learning layer;
and S5, the remote scheduling layer and the data application layer carry out data communication, so that remote transmission, monitoring and application of data are realized.
3. The method for realizing relay fault identification based on artificial intelligence technology according to claim 2, wherein the method comprises the following steps: the k-term set has a k-value between 50 and 100.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010420948.8A CN111579978B (en) | 2020-05-18 | 2020-05-18 | System and method for realizing relay fault identification based on artificial intelligence technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010420948.8A CN111579978B (en) | 2020-05-18 | 2020-05-18 | System and method for realizing relay fault identification based on artificial intelligence technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111579978A CN111579978A (en) | 2020-08-25 |
| CN111579978B true CN111579978B (en) | 2024-01-02 |
Family
ID=72110962
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010420948.8A Active CN111579978B (en) | 2020-05-18 | 2020-05-18 | System and method for realizing relay fault identification based on artificial intelligence technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111579978B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114295941A (en) * | 2021-11-18 | 2022-04-08 | 国家能源(山东)工程技术有限公司 | High-voltage cable insulation data acquisition and fault early warning device and method |
| CN115097296A (en) * | 2022-06-22 | 2022-09-23 | 云南电网有限责任公司电力科学研究院 | Relay reliability evaluation method and device |
| CN116300673A (en) * | 2023-01-11 | 2023-06-23 | 广东鸿业管桩有限公司 | Device for protecting internal circuit of PLC digital quantity input module based on intermediate relay |
Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1403786A (en) * | 2001-09-10 | 2003-03-19 | 松下电工株式会社 | Object detecting apparatus with thermoelectric sensor |
| CN101950382A (en) * | 2010-09-01 | 2011-01-19 | 燕山大学 | Method for optimal maintenance decision-making of hydraulic equipment with risk control |
| CN103439650A (en) * | 2013-08-07 | 2013-12-11 | 王岩 | Method and device used for state monitoring and fault diagnosis of relay |
| CN103533707A (en) * | 2013-09-30 | 2014-01-22 | 广州东霖电子有限公司 | Voltage-stabilization dimmable LED (light emitting diode) driving power supply circuit |
| CN104463706A (en) * | 2014-12-10 | 2015-03-25 | 深圳供电局有限公司 | Method and system for power grid to detect reasons of voltage sag incident |
| CN105677759A (en) * | 2015-12-30 | 2016-06-15 | 国家电网公司 | Alarm correlation analysis method in communication network |
| CN107016507A (en) * | 2017-04-07 | 2017-08-04 | 国网技术学院 | Electric network fault method for tracing based on data mining technology |
| CN107084853A (en) * | 2017-03-06 | 2017-08-22 | 上海大学 | The lower equipment failure prediction method of cloud manufacture |
| CN107408884A (en) * | 2015-03-31 | 2017-11-28 | 古河电气工业株式会社 | The control method of power inverter and power inverter |
| WO2018061539A1 (en) * | 2016-09-30 | 2018-04-05 | ミツミ電機株式会社 | Optical scanning device and retina-scanning-type head-mounted display |
| CN109270407A (en) * | 2018-11-16 | 2019-01-25 | 国网山东省电力公司电力科学研究院 | Extra high voltage direct current transmission line failure cause discrimination method based on Multi-source Information Fusion |
| CN109459662A (en) * | 2018-11-29 | 2019-03-12 | 广东电网有限责任公司 | High-tension cable defect state evaluation system |
| CN110232240A (en) * | 2019-06-12 | 2019-09-13 | 贵州电网有限责任公司 | A kind of improved transformer top-oil temperature prediction technique |
| CN110619386A (en) * | 2019-09-09 | 2019-12-27 | 国家电网有限公司 | TMR operation monitoring and fault intelligent research and judgment method and system |
| CN110908306A (en) * | 2019-10-25 | 2020-03-24 | 大唐水电科学技术研究院有限公司 | Hydroelectric generating set reliability monitoring system based on Internet of things |
| CN111695288A (en) * | 2020-05-06 | 2020-09-22 | 内蒙古电力(集团)有限责任公司电力调度控制分公司 | Transformer fault diagnosis method based on Apriori-BP algorithm |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015002630A2 (en) * | 2012-07-24 | 2015-01-08 | Deloitte Development Llc | Fraud detection methods and systems |
-
2020
- 2020-05-18 CN CN202010420948.8A patent/CN111579978B/en active Active
Patent Citations (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1403786A (en) * | 2001-09-10 | 2003-03-19 | 松下电工株式会社 | Object detecting apparatus with thermoelectric sensor |
| CN101950382A (en) * | 2010-09-01 | 2011-01-19 | 燕山大学 | Method for optimal maintenance decision-making of hydraulic equipment with risk control |
| CN103439650A (en) * | 2013-08-07 | 2013-12-11 | 王岩 | Method and device used for state monitoring and fault diagnosis of relay |
| CN103533707A (en) * | 2013-09-30 | 2014-01-22 | 广州东霖电子有限公司 | Voltage-stabilization dimmable LED (light emitting diode) driving power supply circuit |
| CN104463706A (en) * | 2014-12-10 | 2015-03-25 | 深圳供电局有限公司 | Method and system for power grid to detect reasons of voltage sag incident |
| CN107408884A (en) * | 2015-03-31 | 2017-11-28 | 古河电气工业株式会社 | The control method of power inverter and power inverter |
| CN105677759A (en) * | 2015-12-30 | 2016-06-15 | 国家电网公司 | Alarm correlation analysis method in communication network |
| WO2018061539A1 (en) * | 2016-09-30 | 2018-04-05 | ミツミ電機株式会社 | Optical scanning device and retina-scanning-type head-mounted display |
| CN107084853A (en) * | 2017-03-06 | 2017-08-22 | 上海大学 | The lower equipment failure prediction method of cloud manufacture |
| CN107016507A (en) * | 2017-04-07 | 2017-08-04 | 国网技术学院 | Electric network fault method for tracing based on data mining technology |
| CN109270407A (en) * | 2018-11-16 | 2019-01-25 | 国网山东省电力公司电力科学研究院 | Extra high voltage direct current transmission line failure cause discrimination method based on Multi-source Information Fusion |
| CN109459662A (en) * | 2018-11-29 | 2019-03-12 | 广东电网有限责任公司 | High-tension cable defect state evaluation system |
| CN110232240A (en) * | 2019-06-12 | 2019-09-13 | 贵州电网有限责任公司 | A kind of improved transformer top-oil temperature prediction technique |
| CN110619386A (en) * | 2019-09-09 | 2019-12-27 | 国家电网有限公司 | TMR operation monitoring and fault intelligent research and judgment method and system |
| CN110908306A (en) * | 2019-10-25 | 2020-03-24 | 大唐水电科学技术研究院有限公司 | Hydroelectric generating set reliability monitoring system based on Internet of things |
| CN111695288A (en) * | 2020-05-06 | 2020-09-22 | 内蒙古电力(集团)有限责任公司电力调度控制分公司 | Transformer fault diagnosis method based on Apriori-BP algorithm |
Non-Patent Citations (8)
| Title |
|---|
| Research on Transformer Fault Diagnosis Based on BP Neural Network Improved by Association Rules;Jiang Long等;《2nd International Conference on Electrical Materials and Power Equipment - Guangzhou - China》;20191231;554-559 * |
| Research on Transformer Fault Diagnosis Based on BP Neural Network Improved by Association Rules;Jiang, L等;《 2019 2ND INTERNATIONAL CONFERENCE ON ELECTRICAL MATERIALS AND POWER EQUIPMENT (ICEMPE 2019)》;554-559 * |
| 一种基于概率的加权关联规则挖掘算法;尹群,王丽珍,田启明;《计算机应用》;20050430;805-807 * |
| 加权关联规则的相关研究;刘洋;《科技创新与应用》;20200320;17-18 * |
| 基于加权关联规则挖掘的相关文献推荐;陈祖琴等;《情报分析与研究》;20071031;57-61 * |
| 基于粗糙集的关联规则挖掘在变电设备故障诊断中的应用;贺林晓;《硕士电子期刊》;全文 * |
| 焦李成，刘若辰，慕彩红.《简明人工智能》.西安电子科技大学出版社,2019,253. * |
| 马慧彬.《基于机器学习的乳腺图像辅助诊断算法研究》.湖南师范大学出版社,2016,46-49. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111579978A (en) | 2020-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111579978B (en) | System and method for realizing relay fault identification based on artificial intelligence technology | |
| CN108832619A (en) | Transient stability evaluation in power system method based on convolutional neural networks | |
| CN103728535A (en) | Extra-high-voltage direct-current transmission line fault location method based on wavelet transformation transient state energy spectrum | |
| CN110133393A (en) | A kind of electricity consumption monitoring system and method based on non-intruding monitor technology | |
| CN105022021A (en) | State discrimination method for gateway electrical energy metering device based on the multiple agents | |
| CN113391239B (en) | Mutual inductor anomaly monitoring method and system based on edge calculation | |
| CN116245033A (en) | Artificial intelligent driven power system analysis method and intelligent software platform | |
| CN111932051A (en) | Malicious behavior detection method based on non-invasive power terminal time sequence monitoring | |
| CN102930408B (en) | A kind of 750kV electric grid secondary equipment state appraisal procedure based on information fusion | |
| CN113947011A (en) | Low-voltage direct-current contactor state evaluation method and device | |
| Li et al. | Abnormal signal recognition method of wearable sensor based on machine learning | |
| CN115169405B (en) | Hotel guest room equipment fault diagnosis method and system based on support vector machine | |
| CN112085043A (en) | Intelligent monitoring method and system for network security of transformer substation | |
| CN116151099A (en) | Low-voltage distribution network load state estimation terminal | |
| CN113916302B (en) | Crown block online detection and diagnosis system and method based on artificial intelligence technology | |
| Yang et al. | Transient fault diagnosis of track circuit based on MFCC-DTW | |
| CN114760194A (en) | Fault studying and judging method for Internet of things equipment | |
| Xing et al. | Intelligent Diagnosis Method of Distribution Network Fault for Construction of Digital Twin Coordination System | |
| Zhang et al. | Research on transformer oil level monitoring system based on inspection robot | |
| Fayyad et al. | Transmission line protection using high-speed decision tree and artificial neural network: A hardware co-simulation approach | |
| Wang et al. | Turn-to-turn short circuit of motor stator fault diagnosis using dropout rate improved deep sparse autoencoder | |
| Gong et al. | State detection method of secondary equipment in smart substation based on deep belief network and trend prediction | |
| Huu et al. | Development of warning and predicting system for quality of air in smart cities using RNN | |
| Li | Energy consumption prediction of public buildings based on PCA-RF-AdaBoost | |
| Zakri et al. | Qualified two-hybrid techniques by DWT output to predict fault location |
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 | ||
| CB02 | Change of applicant information |
Address after: 650200 room 801, building 15, Jinfuyuan, Junfu Huacheng community, Yiliu street, Guandu District, Kunming City, Yunnan Province Applicant after: Liu Zhuhua Address before: 510275 Zhongshan University, 135 West Xingang Road, Guangdong, Guangzhou Applicant before: Liu Zhuhua |
|
| CB02 | Change of applicant information | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20231212 Address after: Room 601, Building 3, No. 101 Daxue Road, High tech Zone, Zhuhai City, Guangdong Province, 519000 Applicant after: ZHUHAI SINO SCIENCE AND TECHNOLOGY OF ELECTRIC POWER CO.,LTD. Address before: 650200 room 801, building 15, Jinfuyuan, Junfu Huacheng community, Yiliu street, Guandu District, Kunming City, Yunnan Province Applicant before: Liu Zhuhua |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |