CN111060815A - GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method - Google Patents
GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method Download PDFInfo
- Publication number
- CN111060815A CN111060815A CN201911303417.4A CN201911303417A CN111060815A CN 111060815 A CN111060815 A CN 111060815A CN 201911303417 A CN201911303417 A CN 201911303417A CN 111060815 A CN111060815 A CN 111060815A
- Authority
- CN
- China
- Prior art keywords
- layer
- hidden layer
- input
- output
- fault diagnosis
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000003745 diagnosis Methods 0.000 title claims abstract description 37
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000012544 monitoring process Methods 0.000 claims abstract description 20
- 238000012360 testing method Methods 0.000 claims abstract description 8
- 238000012545 processing Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 27
- 238000013528 artificial neural network Methods 0.000 claims description 23
- 230000008569 process Effects 0.000 claims description 19
- 230000004913 activation Effects 0.000 claims description 16
- 230000002457 bidirectional effect Effects 0.000 claims description 13
- 230000002068 genetic effect Effects 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 6
- 125000004122 cyclic group Chemical group 0.000 claims description 6
- 230000035772 mutation Effects 0.000 claims description 6
- 230000009466 transformation Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000013473 artificial intelligence Methods 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000010353 genetic engineering Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims 1
- 125000006850 spacer group Chemical group 0.000 claims 1
- 238000012423 maintenance Methods 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 23
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical group [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 8
- 230000000306 recurrent effect Effects 0.000 description 5
- 230000007246 mechanism Effects 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 230000004907 flux Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
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
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
- G01R31/3271—Testing of circuit interrupters, switches or circuit-breakers of high voltage or medium voltage devices
- G01R31/3275—Fault detection or status indication
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Image Analysis (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention discloses a GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method, which specifically comprises the following steps: s1, real-time monitoring is carried out by using a switching-on and switching-off coil current on-line monitoring system to obtain switching-on and switching-off coil current data, and the data are divided into a training set and a testing set which are used as input variables together; s2, initializing weights, inputting the sample data of the training set into the Bi-RNN, optimizing and updating the characteristic information parameters of each generation by adopting GA as error back propagation, taking GA as input, taking mean square error as fitness, taking a certain number of iterations as a model termination condition, selecting the optimal combination of predicted characteristic quantities, and completing model training; and S3, inputting the obtained test set sample data into a trained fault diagnosis model, and processing the input opening and closing coil current data by the fault diagnosis model to finish fault diagnosis and classification of the high-voltage circuit breaker. The method can more accurately and effectively judge the fault type of the circuit breaker, and further efficiently finish the maintenance.
Description
Technical Field
The invention belongs to the technical field of online monitoring of faults of high-voltage circuit breakers, and particularly relates to a GA-Bi-RNN-based fault diagnosis method for a high-voltage circuit breaker.
Background
The high-voltage circuit breaker is the most main control and protection device of a power system and is related to the reliability and safety of power transmission, power distribution and power utilization. High voltage circuit breakers can achieve a variety of operations in the event of system faults and non-fault conditions. The breaker can close, bear and break the normal current of the operation loop, and can also close, bear and break the specified overload current within the specified time. High-voltage circuit breakers generally use an electromagnet as a first control element, and most of the operating mechanisms are direct-current electromagnets. When current passes through the coil, magnetic flux is generated in the magnet, and the movable iron core is influenced by the magnetic force, so that the breaker is opened or closed. The opening and closing coil current can be used as rich information for diagnosing the mechanical fault of the high-voltage circuit breaker.
There are many existing methods for diagnosing faults of a high-voltage circuit breaker, wherein various artificial intelligence algorithms are involved, such as: fuzzy control can clarify fuzzy concepts or natural languages by using an accurate mathematical tool, but certain human factors exist in the determination process of membership functions and fuzzy rules of the fuzzy control; the radial basis function neural network provides a better structural system for the fault diagnosis problem of the circuit breaker, but has the defects that the self reasoning process and reasoning basis cannot be explained, and the neural network cannot work normally when the data is insufficient.
The Bi-directional recurrent neural network (Bi-RNN) is a neural network modeling data sequences, and the processing mode is essentially different from that of a feedforward neural network, and the Bi-directional recurrent neural network only processes a single input unit and hidden layer information at the previous time point. The bidirectional cyclic neural network can freely and dynamically acquire input information without being limited by a fixed-length input space, and has good fault-tolerant capability, parallel processing capability and self-learning capability. But the learning process is slightly too single, and the defect of incomplete training may exist in the training process; therefore, the problem can be solved by optimizing the bidirectional recurrent neural network by using a Genetic Algorithm (GA), and the weight of the bidirectional recurrent neural network is updated until the weight is within a set error range, so that the problem can be effectively solved, and meanwhile, the fault can be classified more accurately and rapidly.
Disclosure of Invention
The invention aims to provide a GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method, which adopts a bidirectional cyclic neural network to analyze fault characteristic signals and combines a genetic algorithm to carry out parameter optimization, so that the fault type of a circuit breaker can be judged more accurately and effectively while the defect of artificial neural network diagnosis is overcome, and further, the maintenance is completed efficiently.
The technical scheme adopted by the invention is that the GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method is implemented according to the following steps:
step 1, a switching-on and switching-off coil current online monitoring system is used for monitoring in real time to obtain switching-on and switching-off coil current data, and the data are divided into a training set and a testing set which are used as input variables together;
step 2, initializing weights, inputting the sample data of the training set into the Bi-RNN, optimizing and updating the characteristic information parameters of each generation by adopting GA as error back propagation, taking GA as input, taking mean square error as fitness, taking a certain number of iterations as a model termination condition, selecting an optimal combination of predicted characteristic quantities, and completing model training;
and 3, inputting the test set sample data obtained in the step 1 into the fault diagnosis model trained in the step 2, and processing the input opening and closing coil current data by the fault diagnosis model to finish fault diagnosis and classification of the high-voltage circuit breaker.
The present invention is also characterized in that,
in the step 1, the on-line current monitoring system for the opening and closing coil comprises a process layer, a spacing layer and a station control layer;
the method comprises the steps that a process layer collects and extracts characteristic information parameters of the opening and closing coil current of the high-voltage circuit breaker, an online monitoring system is utilized to preprocess collected data, and analysis and calculation of the characteristic information parameters are completed; the bay level consists of a substation breaker IED and an Ethernet, the bay level transmits characteristic information parameters monitored and processed by the process level to the substation breaker IED by utilizing communication between a CAN bus and the substation breaker IED, and data are uploaded to a station level monitoring center by the Ethernet by adopting IEC 61850 series standard protocols; the station control layer is used for remotely monitoring equipment in the station, receiving characteristic information parameters transmitted by the spacing layer and carrying out real-time fault diagnosis on the circuit breaker by combining an artificial intelligence neural network.
The step 2 is specifically implemented according to the following method:
step 2.1, initializing weights, and initializing all weights into a random number [0,1 ];
step 2.2, after the step 2.1, extracting a sample X from the training set, inputting the sample X into the bidirectional circulation neural network, giving out a target output vector of the sample X, and recording the target output vector as O;
the following functional relationship exists between the input of the input layer and the output of the hidden layer:
wherein,
the input value of the hidden layer is input in the forward direction for time t,
i (t) is an input value reversely input into the hidden layer at the time t, I (t) is a time node U of the opening and closing coil current and the opening and closing coil current changing along with the time t, S (t) is a vector of h multiplied by 1,
representing the output of the forward hidden layer at time t,
for the output of the reverse hidden layer at time t,
is an input vector with h elements for representing the output of the forward input hidden layer at the time t-1, h is the dimension of the hidden layer,
reversely inputting the output of the hidden layer at the time of t-1;
respectively represent input layers I (t),
U is connected to the weight matrix of the forward input hidden layer,
respectively represent input layers I (t),
U is connected to the weight matrix of the reverse input hidden layer; wforwardFor the forward input of a transformation weight matrix, W, of hidden layer statesbackwardInputting a transformation weight matrix of hidden layer states for the reverse direction;
wherein f () is sigmoid function:
the following functional relationship exists between the output S (t) of the hidden layer and the output O (t) of the output layer:
O(t)＝g(YS(t)) (7)
where Y is the weight matrix for the connection of the hidden layer to the output layer, g () is the softmax function:
wherein x is the input value of the hidden layer, i is the number of nodes of the hidden layer, and a weight matrix is randomly generated
WforwardAnd Wbackward；
Step 2.3, after step 2.2, sequentially calculating from the front layer to the back layer to obtain an output value o (t) of the bidirectional cyclic neural network, wherein an activation function netj (t) of a certain node at a certain moment of the hidden layer is expressed by a formula:
wherein n represents the number of input layer nodes, i (t) represents the number of hidden layer nodes at time t, VjiA weight matrix, theta, representing the level of the layer to which the node is connected at that momentjRepresenting a bias parameter, and updating the calculation mode of the hidden layer node activation function:
hj(t)＝f(netj(t)) (11)
where m denotes the total number of hidden layer nodes, l(t-1)Denotes the hidden layer node, V, at time t-1jlA weight matrix representing the layer connected with the node at the moment; hj (t) represents the last time the activation function of the hidden layer node is updated;
activation function netk (t) of output layer:
yk(t)＝g(netk(t)) (13)
wherein j (t) represents hidden layer node at time t, thetakRepresents aAn offset parameter, WkjA weight matrix representing a layer connected with the node at the moment is shown, and yk (t) represents an activation function of the node of the output layer;
and 2.4, after the step 2.3, adopting a genetic algorithm as error back propagation optimization, taking the updated characteristic information parameters of each generation after optimization as input, taking the mean square error as fitness, taking certain iteration times as model termination conditions, and selecting the optimal combination of the predicted characteristic quantities.
In step 2.4, the specific process of genetic error back propagation is as follows:
one standard genetic algorithm is SCA ═ C, E, P0M, phi, delta, psi, T), where C is the GA encoding method, E is the fitness function of GA, P0Is an initial population, M is the size of the population, phi is the selection operation, delta is the crossover operation of GA, psi is the mutation operation of GA, and T is the termination operation condition of GA; to prevent entry into local optimality;
(a) and encoding:
according to the required precision, 11-bit binary numbers are adopted to encode the connection weights and the threshold, wherein the 1 st bit is a sign bit, and the corresponding relation of the rest 10-bit encoding is as follows:
wherein δ (1.0-0.0)/(210-1) is 0.00098;
(b) and genetic manipulation:
in order to improve the running speed and the convergence capability of the model, the cross rate P is calculatedcAnd the rate of variation PmThe method comprises the following steps:
in the formula (f)maxTo the maximum fitness of the individual, favgTo average individual fitness, f' to perform cross-trainingMaking the maximum fitness of the individual, wherein f is the maximum fitness of the individual performing the mutation operation;
(c) objective function
Using the minimum of the sum of the differences between the output of the model and the expected output of the training samples as an objective function, i.e.
In the formula, YBi-RNN-GAIs the output value of the Bi-RNN model, YdataN is the expected output of the training samples and is the number of the samples;
(d) individual fitness
In the formula, CmaxThe individual fitness with the maximum population is selected.
The invention has the beneficial effects that:
(1) the method selects two angles from an optimized neural network model and optimized characteristic parameters to improve the fault diagnosis capability of the model, and the bidirectional cyclic neural network is used as a deep neural network model, so that more abstract and representative characteristics can be extracted from original data, input information can be more freely and dynamically acquired without being limited by a fixed-length input space, and the method has good fault tolerance capability, parallel processing capability and self-learning capability;
(2) the method optimizes the connection weight by using the genetic algorithm, updates the weight until the weight is within a set error range, has better local and global search capability, and can effectively improve the speed and the accuracy of fault diagnosis of the high-voltage circuit breaker;
(3) according to the method, a Dropout technology is introduced in the bidirectional loop network training process to prevent overfitting, and the generalization capability of the model is enhanced.
Drawings
FIG. 1 is a structural diagram of an on-line current monitoring system of a switching-on/switching-off coil adopted in the GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method of the invention;
FIG. 2 is a flow chart of the GA-Bi-RNN based high voltage circuit breaker fault diagnosis method of the present invention;
FIG. 3 is a fault diagnosis model of the GA-Bi-RNN based high-voltage circuit breaker fault diagnosis method of the invention;
fig. 4 is a characteristic curve of the opening/closing coil current according to example 1 of the method of the present invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention provides a GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method, which is implemented according to the following steps as shown in figures 1-3:
step 1, a switching-on and switching-off coil current online monitoring system is used for monitoring in real time to obtain switching-on and switching-off coil current data, and the data are divided into a training set and a testing set which are used as input variables together;
in the step 1, the on-line current monitoring system for the opening and closing coils comprises a process layer, a spacing layer and a station control layer as shown in figure 1; the method comprises the steps that a process layer collects and extracts characteristic information parameters of the opening and closing coil current of the high-voltage circuit breaker, an online monitoring system is utilized to preprocess collected data, and analysis and calculation of the characteristic information parameters are completed; the bay level consists of a substation breaker IED and an Ethernet, the bay level transmits characteristic information parameters monitored and processed by the process level to the substation breaker IED by utilizing communication between a CAN bus and the substation breaker IED, and data is uploaded to a station control level monitoring center by the Ethernet by adopting an IEC 61850 series standard protocol; the station control layer is used for remotely monitoring equipment in the station, receiving characteristic information parameters transmitted by the spacing layer and carrying out real-time fault diagnosis on the circuit breaker by combining an artificial intelligence neural network.
Step 2, initializing weights, inputting the sample data of the training set into the Bi-RNN, optimizing and updating the characteristic information parameters of each generation by adopting GA as error back propagation, taking GA as input, taking mean square error as fitness, taking a certain number of iterations as a model termination condition, selecting an optimal combination of predicted characteristic quantities, and completing model training;
the step 2 is specifically implemented according to the following method:
step 2.1, initializing weights, and initializing all weights into a random number [0,1 ];
step 2.2, after the step 2.1, extracting a sample X from the training set (taking the example 1 as an example, five groups of data therein are taken as the training set), inputting the sample X into the bidirectional recurrent neural network, giving out a target output vector of the sample X, and recording the target output vector as O;
the following functional relationship exists between the input of the input layer and the output of the hidden layer:
wherein,
the input value of the hidden layer is input in the forward direction for time t,
i (t) is an input value reversely input into the hidden layer at the time t, I (t) is a time node U of the opening and closing coil current and the opening and closing coil current changing along with the time t, S (t) is a vector of h multiplied by 1,
representing the output of the forward hidden layer at time t,
for the output of the reverse hidden layer at time t,
is an input vector with h elements for representing the output of the forward input hidden layer at the time t-1, h is the dimension of the hidden layer,
reversely inputting the output of the hidden layer at the time of t-1;
respectively represent input layers I (t),
U is connected to the weight matrix of the forward input hidden layer,
respectively represent input layers I (t),
U is connected to the weight matrix of the reverse input hidden layer; wforwardFor the forward input of a transformation weight matrix, W, of hidden layer statesbackwardInputting a transformation weight matrix of hidden layer states for the reverse direction;
wherein f () is sigmoid function:
the following functional relationship exists between the output S (t) of the hidden layer and the output O (t) of the output layer:
O(t)＝g(YS(t)) (7)
where Y is the weight matrix for the connection of the hidden layer to the output layer, g () is the softmax function:
wherein x is the input value of the hidden layer, i is the number of nodes of the hidden layer, and a weight matrix is randomly generated
WforwardAnd Wbackward；
Step 2.3, after step 2.2, sequentially calculating from the front layer to the back layer to obtain an output value o (t) of the bidirectional cyclic neural network, wherein an activation function netj (t) of a certain node at a certain moment of the hidden layer is expressed by a formula:
wherein n represents the number of input layer nodes, i (t) represents the number of hidden layer nodes at time t, VjiA weight matrix, theta, representing the level of the layer to which the node is connected at that momentjRepresenting a bias parameter, and updating the calculation mode of the hidden layer node activation function:
hj(t)＝f(netj(t)) (11)
where m denotes the total number of hidden layer nodes, l(t-1)Denotes the hidden layer node, V, at time t-1jlA weight matrix representing the layer connected with the node at the moment; hj (t) represents the last time the activation function of the hidden layer node was updated.
Activation function netk (t) of output layer:
yk(t)＝g(netk(t)) (13)
wherein j (t) represents tInscribing hidden layer nodes, thetakDenotes a bias parameter, WkjThe weight matrix of the layer connected with the node at the moment is represented, and yk (t) represents the activation function of the node of the output layer (which may be the same activation function as h of the node of the hidden layer).
And 2.4, after the step 2.3, adopting a genetic algorithm as error back propagation optimization, taking the updated characteristic information parameters of each generation after optimization as input, taking the mean square error as fitness, taking certain iteration times as model termination conditions, and selecting the optimal combination of the predicted characteristic quantities.
In step 2.4, the specific process of genetic error back propagation is as follows:
one standard genetic algorithm is SCA ═ C, E, P0M, phi, delta, psi, T), where C is the GA encoding method, E is the fitness function of GA, P0Is an initial population, M is the size of the population, phi is the selection operation, delta is the crossover operation of GA, psi is the mutation operation of GA, and T is the termination operation condition of GA; to prevent entry into local optimality;
(a) and encoding:
according to the required precision, 11-bit binary numbers are adopted to encode the connection weights and the threshold, wherein the 1 st bit is a sign bit, and the corresponding relation of the rest 10-bit encoding is as follows:
wherein δ (1.0-0.0)/(210-1) is 0.00098;
(b) and genetic manipulation:
in order to improve the running speed and the convergence capability of the model, the cross rate P is calculatedcAnd the rate of variation PmThe method comprises the following steps:
in the formula (f)maxTo the maximum fitness of the individual, favgF' is the maximum fitness of the individuals performing the cross operation, and f is the maximum fitness of the individuals performing the mutation operation;
(c) objective function
Using the minimum of the sum of the differences between the output of the model and the expected output of the training samples as an objective function, i.e.
In the formula, YBi-RNN-GAIs the output value of the Bi-RNN model, YdataN is the expected output of the training samples and is the number of the samples;
(d) individual fitness
In the formula, CmaxThe individual fitness with the maximum population is selected.
And 3, inputting the test set sample data obtained in the step 1 into the fault diagnosis model trained in the step 2, and processing the input opening and closing coil current data by the fault diagnosis model to finish fault diagnosis and classification of the high-voltage circuit breaker.
Examples
With t0Extraction of a fault characteristic parameter I for the zero point of a command time1，I2，I3，t1，t2，t3，t4，t5Monitoring the state of the circuit breaker to obtain ten groups of fault sample data, wherein the ten groups of fault sample data comprise a normal mechanism (A), an overlow operating voltage (B), a jamming (C) at the initial stage of a closing iron core, a jamming (D) of the operating mechanism and an overlarge idle stroke (E) of the closing iron core, and the data acquisition condition is specifically shown in table 1;
TABLE 1 Fault sample data
As shown in fig. 4, the characteristic curve of the opening/closing coil current indicates:
(1) stage I, t ═ t0～t1(ii) a Coil at t0Starting to supply power at time t1The iron core starts to move at the moment; t is t0The time for issuing the opening and closing command of the circuit breaker is the starting point of the opening and closing timing of the circuit breaker; t is t1The current and the magnetic flux in the coil rise to be enough to drive the iron core to move, namely the iron core starts to move; the characteristic of this stage is that the current rises exponentially, the iron core is static; the time of this phase is related to the control supply voltage and the coil resistance.
(2) Stage II, t ═ t1～t2(ii) a At this stage, the core starts to move and the current drops; t is t2To control the valley point of the current, it represents that the core has significantly slowed or stopped moving by the load of the operating machine.
(3) Stage III, t ═ t2～t3(ii) a At this stage the core stops moving and the current rises exponentially.
(4) Stage IV, t ═ t3～t4(ii) a This phase is a continuation of phase iii, the current reaching an approximately steady state.
(5) Stage V, t ═ t4～t5(ii) a And in the current breaking stage, the auxiliary switch is broken, an arc is generated between the contacts of the auxiliary switch and is elongated, the voltage of the arc is rapidly increased, and the current is rapidly reduced until the arc is extinguished.
From the analysis of the current waveform in FIG. 4, t0～t1The time current may reflect the state of the coil (e.g., whether the resistance is normal). t is t1～t2The change of time current represents the condition of the change of mechanical load of the iron core movement structure, such as jamming, tripping and energy release; t is t2Generally, the moment when the moving contact starts to move, from t2Then, the mechanism drives the moving contact to switch on and off through a transmission system, namely the moving process of the moving contact; t is t4The moment when the auxiliary contact of the circuit breaker is cut off; t is t0～t4The change of time and current can reflect the mechanical operating machineAnd constructing the working condition of the transmission system.
The output of the fault type is represented by a binary number, which is specifically shown in table 2:
TABLE 2 Fault type output settings
According to the GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method provided by the invention, the bidirectional circulating neural network is adopted to analyze the fault characteristic signals, the parameter optimization is carried out by combining the genetic algorithm, the deficiency of artificial neural network diagnosis is made up, and the fault type of the circuit breaker can be judged more accurately and rapidly.
Claims (5)
1. The GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method is characterized by comprising the following steps of:
step 1, a switching-on and switching-off coil current online monitoring system is used for monitoring in real time to obtain switching-on and switching-off coil current data, and the data are divided into a training set and a testing set which are used as input variables together;
step 2, initializing weights, inputting the sample data of the training set into the Bi-RNN, optimizing and updating the characteristic information parameters of each generation by adopting GA as error back propagation, taking GA as input, taking mean square error as fitness, taking a certain number of iterations as a model termination condition, selecting an optimal combination of predicted characteristic quantities, and completing model training;
and 3, inputting the test set sample data obtained in the step 1 into the fault diagnosis model trained in the step 2, and processing the input opening and closing coil current data by the fault diagnosis model to finish fault diagnosis and classification of the high-voltage circuit breaker.
2. A GA-Bi-RNN based high voltage circuit breaker fault diagnosis method according to claim 1, wherein in step 1, the on-line open/close coil current monitoring system comprises a process layer, a spacer layer and a station control layer.
3. The GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method of claim 2, wherein the process layer collects and extracts characteristic information parameters of the opening and closing coil current of the high-voltage circuit breaker, and an online monitoring system is used for preprocessing the collected data and completing analysis and calculation of the characteristic information parameters; the bay level consists of a substation breaker IED and an Ethernet, the bay level transmits characteristic information parameters monitored and processed by the process level to the substation breaker IED by utilizing communication between a CAN bus and the substation breaker IED, and data are uploaded to a station level monitoring center by the Ethernet by adopting IEC 61850 series standard protocols; the station control layer is used for remotely monitoring equipment in the station, receiving characteristic information parameters transmitted by the spacing layer and carrying out real-time fault diagnosis on the circuit breaker by combining an artificial intelligence neural network.
4. A GA-Bi-RNN based high voltage circuit breaker fault diagnosis method according to claim 1, wherein step 2 is specifically performed according to the following method:
step 2.1, initializing weights, and initializing all weights into a random number [0,1 ];
step 2.2, after the step 2.1, extracting a sample X from the training set, inputting the sample X into the bidirectional circulation neural network, giving out a target output vector of the sample X, and recording the target output vector as O;
the following functional relationship exists between the input of the input layer and the output of the hidden layer:
wherein,
the input value of the hidden layer is input in the forward direction for time t,
i (t) is an input value reversely input into the hidden layer at the time t, I (t) is a time node U of the opening and closing coil current and the opening and closing coil current changing along with the time t, S (t) is a vector of h multiplied by 1,
representing the output of the forward hidden layer at time t,
for the output of the reverse hidden layer at time t,
is an input vector with h elements for representing the output of the forward input hidden layer at the time t-1, h is the dimension of the hidden layer,
reversely inputting the output of the hidden layer at the time of t-1;
respectively represent input layers I (t),
U is connected to the weight matrix of the forward input hidden layer,
respectively represent input layers I (t),
U is connected to the weight matrix of the reverse input hidden layer; wforwardFor the forward input of a transformation weight matrix, W, of hidden layer statesbackwardInputting a transformation weight matrix of hidden layer states for the reverse direction;
wherein f () is sigmoid function:
the following functional relationship exists between the output S (t) of the hidden layer and the output O (t) of the output layer:
O(t)＝g(YS(t)) (7)
where Y is the weight matrix for the connection of the hidden layer to the output layer, g () is the softmax function:
wherein x is the input value of the hidden layer, i is the number of nodes of the hidden layer, and a weight matrix is randomly generated
WforwardAnd Wbackward；
Step 2.3, after step 2.2, sequentially calculating from the front layer to the back layer to obtain an output value o (t) of the bidirectional cyclic neural network, wherein an activation function netj (t) of a certain node at a certain moment of the hidden layer is expressed by a formula:
wherein n represents the number of input layer nodes, i (t) represents the number of hidden layer nodes at the time of tNumber, VjiA weight matrix, theta, representing the level of the layer to which the node is connected at that momentjRepresenting a bias parameter, and updating the calculation mode of the hidden layer node activation function:
hj(t)＝f(netj(t)) (11)
where m denotes the total number of hidden layer nodes, l(t-1)Denotes the hidden layer node, V, at time t-1jlA weight matrix representing the layer connected with the node at the moment; hj (t) represents the last time the activation function of the hidden layer node is updated;
activation function netk (t) of output layer:
yk(t)＝g(netk(t)) (13)
wherein j (t) represents hidden layer node at time t, thetakDenotes a bias parameter, WkjA weight matrix representing a layer connected with the node at the moment is shown, and yk (t) represents an activation function of the node of the output layer;
and 2.4, after the step 2.3, adopting a genetic algorithm as error back propagation optimization, taking the updated characteristic information parameters of each generation after optimization as input, taking the mean square error as fitness, taking certain iteration times as model termination conditions, and selecting the optimal combination of the predicted characteristic quantities.
5. A GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method according to claim 4, wherein in step 2.4, the specific process of genetic error back propagation is as follows:
one standard genetic algorithm is SCA ═ C, E, P0M, phi, delta, psi, T), where C is the GA encoding method, E is the fitness function of GA, P0Is the initial population, M is the population size, phi is the selection operation, and delta is the crossover operation of GAMaking psi the mutation operation of GA, and T the termination operation condition of GA; to prevent entry into local optimality;
(a) and encoding:
according to the required precision, 11-bit binary numbers are adopted to encode the connection weights and the threshold, wherein the 1 st bit is a sign bit, and the corresponding relation of the rest 10-bit encoding is as follows:
wherein δ is (1.0-0.0)/(2)10-1)＝0.00098；
(b) And genetic manipulation:
in order to improve the running speed and the convergence capability of the model, the cross rate P is calculatedcAnd the rate of variation PmThe method comprises the following steps:
in the formula (f)maxTo the maximum fitness of the individual, favgF' is the maximum fitness of the individuals performing the cross operation, and f is the maximum fitness of the individuals performing the mutation operation;
(c) objective function
Using the minimum of the sum of the differences between the output of the model and the expected output of the training samples as an objective function, i.e.
In the formula, YBi-RNN-GAIs the output value of the Bi-RNN model, YdataN is the expected output of the training samples and is the number of the samples;
(d) individual fitness
In the formula, CmaxThe individual fitness with the maximum population is selected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911303417.4A CN111060815B (en) | 2019-12-17 | 2019-12-17 | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911303417.4A CN111060815B (en) | 2019-12-17 | 2019-12-17 | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111060815A true CN111060815A (en) | 2020-04-24 |
| CN111060815B CN111060815B (en) | 2021-09-14 |
Family
ID=70302107
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911303417.4A Active CN111060815B (en) | 2019-12-17 | 2019-12-17 | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111060815B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111913103A (en) * | 2020-08-06 | 2020-11-10 | 国网福建省电力有限公司 | Fault detection method for spring energy storage operating structure circuit breaker |
| CN112729411A (en) * | 2021-01-14 | 2021-04-30 | 金陵科技学院 | Distributed drug warehouse environment monitoring method based on GA-RNN |
| CN113358157A (en) * | 2021-06-10 | 2021-09-07 | 国网甘肃省电力公司兰州供电公司 | RST-PNN-GA-based power equipment temperature rise detection and early warning method |
| CN113469222A (en) * | 2021-06-10 | 2021-10-01 | 国网甘肃省电力公司兰州供电公司 | RST-PNN-GA-based high-voltage circuit breaker fault detection method |
| CN115389812A (en) * | 2022-10-28 | 2022-11-25 | 国网信息通信产业集团有限公司 | Artificial neural network short-circuit current zero prediction method and prediction terminal |
| CN116087692A (en) * | 2023-04-12 | 2023-05-09 | 国网四川省电力公司电力科学研究院 | Distribution network tree line discharge fault identification method, system, terminal and medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1153488A1 (en) * | 1999-02-12 | 2001-11-14 | Deutsche Telekom AG | Method for monitoring the transmission quality of an optical transmission system, notably an optical wavelength-division multiplex network |
| CN106291351A (en) * | 2016-09-20 | 2017-01-04 | 西安工程大学 | Primary cut-out fault detection method based on convolutional neural networks algorithm |
| CN107490760A (en) * | 2017-08-22 | 2017-12-19 | 西安工程大学 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
| CN108569607A (en) * | 2018-06-22 | 2018-09-25 | 西安理工大学 | Elevator faults method for early warning based on bidirectional valve controlled Recognition with Recurrent Neural Network |
| CN109270442A (en) * | 2018-08-21 | 2019-01-25 | 西安工程大学 | High-voltage circuitbreaker fault detection method based on DBN-GA neural network |
| CN109726200A (en) * | 2018-12-06 | 2019-05-07 | 国网甘肃省电力公司信息通信公司 | Grid information system fault location system and method based on two-way deep neural network |
| US20190146474A1 (en) * | 2016-05-09 | 2019-05-16 | Strong Force Iot Portfolio 2016, Llc | Methods and systems of industrial production line with self organizing data collectors and neural networks |
| CN110118928A (en) * | 2018-02-05 | 2019-08-13 | 西安交通大学 | A kind of circuit breaker failure diagnostic method based on Back Propagation Algorithm |
| CN110261109A (en) * | 2019-04-28 | 2019-09-20 | 洛阳中科晶上智能装备科技有限公司 | A kind of Fault Diagnosis of Roller Bearings based on bidirectional memory Recognition with Recurrent Neural Network |
-
2019
- 2019-12-17 CN CN201911303417.4A patent/CN111060815B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1153488A1 (en) * | 1999-02-12 | 2001-11-14 | Deutsche Telekom AG | Method for monitoring the transmission quality of an optical transmission system, notably an optical wavelength-division multiplex network |
| US20190146474A1 (en) * | 2016-05-09 | 2019-05-16 | Strong Force Iot Portfolio 2016, Llc | Methods and systems of industrial production line with self organizing data collectors and neural networks |
| CN106291351A (en) * | 2016-09-20 | 2017-01-04 | 西安工程大学 | Primary cut-out fault detection method based on convolutional neural networks algorithm |
| CN107490760A (en) * | 2017-08-22 | 2017-12-19 | 西安工程大学 | The circuit breaker failure diagnostic method of fuzzy neural network is improved based on genetic algorithm |
| CN110118928A (en) * | 2018-02-05 | 2019-08-13 | 西安交通大学 | A kind of circuit breaker failure diagnostic method based on Back Propagation Algorithm |
| CN108569607A (en) * | 2018-06-22 | 2018-09-25 | 西安理工大学 | Elevator faults method for early warning based on bidirectional valve controlled Recognition with Recurrent Neural Network |
| CN109270442A (en) * | 2018-08-21 | 2019-01-25 | 西安工程大学 | High-voltage circuitbreaker fault detection method based on DBN-GA neural network |
| CN109726200A (en) * | 2018-12-06 | 2019-05-07 | 国网甘肃省电力公司信息通信公司 | Grid information system fault location system and method based on two-way deep neural network |
| CN110261109A (en) * | 2019-04-28 | 2019-09-20 | 洛阳中科晶上智能装备科技有限公司 | A kind of Fault Diagnosis of Roller Bearings based on bidirectional memory Recognition with Recurrent Neural Network |
Non-Patent Citations (1)
| Title |
|---|
| 黄新波 等: "智能断路器机械特性在线监测技术和状态评估", 《高压电器》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111913103A (en) * | 2020-08-06 | 2020-11-10 | 国网福建省电力有限公司 | Fault detection method for spring energy storage operating structure circuit breaker |
| CN111913103B (en) * | 2020-08-06 | 2022-11-08 | 国网福建省电力有限公司 | Fault detection method for spring energy storage operating structure circuit breaker |
| CN112729411A (en) * | 2021-01-14 | 2021-04-30 | 金陵科技学院 | Distributed drug warehouse environment monitoring method based on GA-RNN |
| CN112729411B (en) * | 2021-01-14 | 2022-09-13 | 金陵科技学院 | Distributed drug warehouse environment monitoring method based on GA-RNN |
| CN113358157A (en) * | 2021-06-10 | 2021-09-07 | 国网甘肃省电力公司兰州供电公司 | RST-PNN-GA-based power equipment temperature rise detection and early warning method |
| CN113469222A (en) * | 2021-06-10 | 2021-10-01 | 国网甘肃省电力公司兰州供电公司 | RST-PNN-GA-based high-voltage circuit breaker fault detection method |
| CN115389812A (en) * | 2022-10-28 | 2022-11-25 | 国网信息通信产业集团有限公司 | Artificial neural network short-circuit current zero prediction method and prediction terminal |
| CN116087692A (en) * | 2023-04-12 | 2023-05-09 | 国网四川省电力公司电力科学研究院 | Distribution network tree line discharge fault identification method, system, terminal and medium |
| CN116087692B (en) * | 2023-04-12 | 2023-06-23 | 国网四川省电力公司电力科学研究院 | Distribution network tree line discharge fault identification method, system, terminal and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111060815B (en) | 2021-09-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111060815B (en) | GA-Bi-RNN-based high-voltage circuit breaker fault diagnosis method | |
| CN109270442B (en) | DBN-GA neural network-based high-voltage circuit breaker fault detection method | |
| CN106291351B (en) | High-voltage circuitbreaker fault detection method based on convolutional neural networks algorithm | |
| CN113505458B (en) | Method, system, equipment and storage medium for predicting critical trigger branch of cascading failure | |
| CN112051481A (en) | Alternating current-direct current hybrid power grid fault area diagnosis method and system based on LSTM | |
| CN114661905B (en) | BERT-based power grid fault diagnosis method | |
| CN110646708B (en) | 10kV single-core cable early state identification method based on double-layer long-and-short-term memory network | |
| CN113536509B (en) | Micro-grid topology identification method based on graph convolution network | |
| CN113219328A (en) | Intelligent fault diagnosis method for circuit breaker operating mechanism based on information fusion | |
| CN116205265A (en) | Power grid fault diagnosis method and device based on deep neural network | |
| CN112651519A (en) | Secondary equipment fault positioning method and system based on deep learning theory | |
| CN114091549A (en) | Equipment fault diagnosis method based on deep residual error network | |
| CN106526465A (en) | High-voltage circuit breaker fault intelligent diagnosis method based on improved fuzzy Petri network | |
| CN115688574A (en) | Transient stability online evaluation method and system for power system | |
| CN113326663A (en) | Railway relay running state evaluation method based on extreme learning machine | |
| CN109901064A (en) | Fault Diagnosis for HV Circuit Breakers method based on ICA-LVQ | |
| Teng et al. | On-line fault diagnosis technology and application based on deep learning of fault characteristic of power grid | |
| CN112085043B (en) | Intelligent monitoring method and system for network security of transformer substation | |
| CN114839523A (en) | Circuit breaker fault detection system and method based on CNN combined LSTM algorithm | |
| CN117057623A (en) | Comprehensive power grid safety optimization scheduling method, device and storage medium | |
| CN115456106A (en) | High-voltage circuit breaker fault diagnosis model optimization method | |
| CN115130767A (en) | Alternating current contactor electric service life prediction method of characteristic selection improved gating circulation unit | |
| Chen et al. | CNN-Informer-Based Remaining Useful Life Prediction for Electrical Devices | |
| CN115684912A (en) | Energy storage system breaker fault diagnosis method and system based on ANFIS | |
| CN114578195A (en) | IEMD and GA-BP based online intelligent diagnosis method for insulation fault of switching-on and switching-off coils of ring main unit switch |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |