CN116089882A - Cable fault prediction processing method and device and electronic equipment - Google Patents
Cable fault prediction processing method and device and electronic equipment Download PDFInfo
- Publication number
- CN116089882A CN116089882A CN202211735073.6A CN202211735073A CN116089882A CN 116089882 A CN116089882 A CN 116089882A CN 202211735073 A CN202211735073 A CN 202211735073A CN 116089882 A CN116089882 A CN 116089882A
- Authority
- CN
- China
- Prior art keywords
- network model
- data
- partial discharge
- trained
- short
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
- G01R31/1227—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials
- G01R31/1263—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation
- G01R31/1272—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing of components, parts or materials of solid or fluid materials, e.g. insulation films, bulk material; of semiconductors or LV electronic components or parts; of cable, line or wire insulation of cable, line or wire insulation, e.g. using partial discharge measurements
-
- 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)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a cable fault prediction processing method and device and electronic equipment. Wherein the method comprises the following steps: obtaining partial discharge characteristic data of a high-voltage cable; respectively inputting first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training; respectively inputting first test set data in the partial discharge characteristic data into a trained radial basis function network model and a trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; a cable fault prediction model is determined based on the first output result and the second output result. The method solves the technical problems of low cable fault prediction efficiency and poor prediction accuracy caused by poor cable fault prediction model performance in the related technology.
Description
Technical Field
The invention relates to the field of intelligent power grid safety monitoring, in particular to a cable fault prediction processing method and device and electronic equipment.
Background
The high-voltage cable is an important power device of the power system, and the running state of the high-voltage cable influences the safety and reliability of power supply of a power grid. However, insulation defects are inevitably generated in the cable system due to factors such as design defects, process defects in the installation process, external force damage, water tree invasion and the like. Partial discharge PD (Partial Discharge) is both a major cause of insulation degradation and an important characterization of cable insulation defects and insulation degradation. The possible partial discharge faults in the high-voltage cable can be timely and accurately identified, and the method plays an important role in safe and stable operation of the high-voltage cable. In the related art, the high-voltage cable partial discharge fault detection is mainly performed in a manual mode or a neural network prediction mode, so that the problems of high detection cost, poor model performance and the like exist, and the cable fault prediction and recognition efficiency are easy to be low, and the accuracy is poor.
In view of the above problems, no effective solution has been proposed at present.
Disclosure of Invention
The embodiment of the invention provides a cable fault prediction processing method, a device and electronic equipment, which are used for at least solving the technical problems of low cable fault prediction efficiency and poor prediction accuracy caused by poor performance of a cable fault prediction model in the related technology.
According to an aspect of the embodiment of the present invention, there is provided a cable failure prediction processing method, including: obtaining partial discharge characteristic data of a high-voltage cable; respectively inputting the first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training to obtain a trained radial basis function network model and a trained long-short-term memory network model; respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; and determining a cable fault prediction model based on the first output result and the second output result.
According to another aspect of the embodiment of the present invention, there is also provided a cable fault prediction processing device, including: the first acquisition module is used for acquiring partial discharge characteristic data of the high-voltage cable; the training module is used for respectively inputting the first training set data in the partial discharge characteristic data into the initial radial basis function network model and the initial long-short-time memory network model for training to obtain a trained radial basis function network model and a trained long-short-time memory network model; the second acquisition module is used for respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-time memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-time memory network model; and the determining module is used for determining a cable fault prediction model based on the first output result and the second output result.
According to another aspect of the embodiment of the present invention, there is further provided a nonvolatile storage medium, wherein the nonvolatile storage medium stores a plurality of instructions, and the instructions are adapted to be loaded by a processor and execute any one of the cable fault prediction processing methods.
According to another aspect of the embodiment of the present invention, there is further provided an electronic device, which is characterized by including one or more processors and a memory, where the memory is configured to store one or more programs, where the one or more programs, when executed by the one or more processors, cause the one or more processors to implement any one of the cable fault prediction processing methods described above.
In the embodiment of the invention, partial discharge characteristic data of the high-voltage cable are obtained; respectively inputting the first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training to obtain a trained radial basis function network model and a trained long-short-term memory network model; respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; based on the first output result and the second output result, a cable fault prediction model is determined, the purpose of constructing a more accurate cable fault prediction model by integrating various neural network model characteristics and improving the performance of the model is achieved, and therefore the performance of the cable fault prediction model is improved, the technical effects of cable fault prediction efficiency and prediction accuracy are further improved, and the technical problems of low cable fault prediction efficiency and poor prediction accuracy caused by poor cable fault prediction model performance in the related technology are solved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiments of the invention and together with the description serve to explain the invention and do not constitute a limitation on the invention. In the drawings:
FIG. 1 is a schematic diagram of a cable fault prediction processing method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an alternative cable fault prediction processing method according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a cable fault prediction processing device according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The high-voltage cable is an important power device of the power system, and the running state of the high-voltage cable influences the safety and reliability of power supply of a power grid. However, insulation defects are inevitably generated in the cable system due to factors such as design defects, process defects in the installation process, external force damage, water, tree invasion and the like. Partial discharge PD (Partial Discharge) is both a major cause of insulation degradation and an important characterization of cable insulation defects and insulation degradation.
The term "partial discharge" refers to a discharge in which only a partial region of the insulation system is discharged without forming a penetrating discharge path under the action of an electric field. The main reason for partial discharge is that when the electrolyte is not uniform, the electric field intensity born by each region of the insulator is not uniform, and the electric field intensity reaches the breakdown field intensity in some regions to generate discharge, while other regions still maintain the insulating property. The insulation structure of the large-scale electrical equipment is complex, the used materials are various, and the electric field distribution of the whole insulation system is uneven. Because of imperfect design or manufacturing process, the insulation system contains air gaps, or insulation is damped in long-term operation, and water is decomposed under the action of an electric field to generate gas so as to form bubbles. Because the dielectric constant of air is smaller than that of insulating material, even if the insulating material is under the action of a not too high electric field, the field intensity of air gap bubbles can be very high, and partial discharge can occur after the field intensity reaches a certain value. In addition, defects or various impurities are mixed in the insulation, or some electrical connection defects exist in the insulation structure, so that local electric fields are concentrated, and solid insulation surface discharge and floating potential discharge can possibly occur at the places where the electric fields are concentrated. The types of partial discharge can be broadly classified into air gap discharge, surface (surface) discharge, corona discharge, and levitation discharge. In the related art, the high-voltage cable partial discharge fault detection is mainly performed in a manual mode or a neural network prediction mode, so that the problems of high detection cost, poor model performance and the like exist, and the cable fault prediction and recognition efficiency are easy to be low, and the accuracy is poor.
In view of the foregoing, embodiments of the present invention provide a method embodiment for cable fault prediction processing, it should be noted that the steps illustrated in the flowchart of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flowchart, in some cases, the steps illustrated or described may be performed in an order other than that illustrated herein.
Fig. 1 is a flowchart of a cable fault prediction processing method according to an embodiment of the present invention, as shown in fig. 1, the method includes the steps of:
step S102, partial discharge characteristic data of the high-voltage cable are obtained.
In an alternative embodiment, the acquiring partial discharge characteristic data of the high voltage cable includes: acquiring partial discharge fault history data of a high-voltage cable; carrying out noise reduction treatment on the partial discharge fault historical data to obtain noise-reduced partial discharge fault historical data; and carrying out feature extraction processing on the noise-reduced partial discharge fault historical data to obtain the partial discharge feature data.
Alternatively, but not limited to, during actual operation of the high voltage cable, partial discharge fault history data of the high voltage cable may be obtained, i.e. the partial discharge fault history data includes operation data before the current operation time of the high voltage cable.
Optionally, the above-mentioned noise-reduced officeThe partial discharge fault history data is subjected to feature extraction processing to obtain the partial discharge feature data, and the method comprises the following steps: performing feature extraction processing on the noise-reduced partial discharge fault historical data to obtain feature sample data; and normalizing the characteristic sample data to obtain the partial discharge characteristic data. The cable Partial Discharge (PD) features included in the partial discharge feature data may include, but are not limited to: maximum discharge distribution positive half cycle skew
Negative half-cycle skew of maximum discharge distribution +.>
Maximum discharge distribution positive half cycle prominence +.>
Maximum discharge distribution negative half cycle prominence +.>
Maximum discharge amount distribution asymmetry degree Q m Correlation degree CC of maximum discharge distribution m Positive half-cycle skew of average discharge capacity distribution>
Negative half-cycle skewness of average discharge capacity distribution +.>
Average discharge amount distribution positive half cycle prominence +.>
Negative half cycle prominence of average discharge capacity distribution +.>
Average discharge amount distribution asymmetry degree Q a Average discharge amount distribution correlation degree CC a Positive half-cycle skew of discharge frequency distribution>
Negative half cycle skewness of discharge frequency distribution
Positive half cycle prominence of discharge frequency distribution >
Negative half cycle prominence of discharge frequency distribution>
Degree of asymmetry Q of discharge frequency distribution n Degree of correlation CC of discharge frequency distribution n 18 kinds.
Through the mode, the partial discharge fault historical data of the high-voltage cable is subjected to noise reduction processing, so that the signal to noise ratio of the data is improved, and the partial discharge characteristic data obtained by characteristic extraction on the basis has higher data quality.
In an optional embodiment, the noise reduction processing is performed on the partial discharge fault history data to obtain noise-reduced partial discharge fault history data, and the noise-reduced partial discharge fault history data includes: judging whether missing data exists in the partial discharge fault historical data; when the missing data exists in the partial discharge fault historical data, filling the missing data by adopting an interpolation method to obtain filled partial discharge fault historical data; and carrying out noise reduction treatment on the filled partial discharge fault historical data to obtain the noise-reduced partial discharge fault historical data.
According to the method, under the condition that missing data exists in the partial discharge fault historical data, filling data are obtained by adopting an interpolation method based on other historical data with the same missing data characteristics in the partial discharge fault historical data, the missing data are filled with the filling data, and the filled partial discharge fault historical data are obtained, so that the integrity of training data is ensured.
Step S104, the first training set data in the partial discharge characteristic data are respectively input into an initial radial basis function network model and an initial long-short-time memory network model for training, and a trained radial basis function network model and a trained long-short-time memory network model are obtained.
It should be noted that, the radial basis function (Radial Basis Function, RBF) neural network is a forward network with good performance, has the performance of optimal approximation, simple training, fast learning convergence speed and overcoming the problem of local minimum, and has been proved to be capable of approximating any continuous function with any accuracy. It has been widely used in the fields of pattern recognition, nonlinear control, image processing, and the like. The long-short-term memory network (Long Short Term Memory Network, LSTM) is an improved cyclic neural network, which can solve the problem that RNNs cannot cope with long-distance dependence, namely, the gradient disappearance problem of RNNs can be avoided. The embodiment of the invention comprehensively utilizes the two models when constructing the cable fault prediction model, so that the cable fault prediction model obtained by subsequent construction has better performance.
Optionally, the partial discharge characteristic data is first subjected to data division to obtain first training set data and first test set data. The partial discharge characteristic data comprises 6000 groups of sample data, the data are randomly disordered, the data are split according to 8:2, the first 4800 groups are first training set data, and the later 1200 groups are first test set data. It should be noted that, the number of the groups of the sample data used in the embodiment of the invention is not fixed, in the subsequent operation process, the cable partial discharge fault data is remotely obtained by the server to predict, and the collected data is continuously written into the original data, so that the self-updating of the sample data is realized, the richness of the sample data is ensured, and meanwhile, the prediction accuracy is gradually improved.
In an optional embodiment, before the first training set data in the partial discharge feature data is input to the initial radial basis function network model and the initial long-short-term memory network model to perform training, to obtain a trained radial basis function network model, and before the trained long-short-term memory network model is obtained, the method further includes: determining the characteristic type corresponding to the partial discharge characteristic data of the high-voltage cable; based on the feature types, determining initialization super parameters corresponding to the initial radial basis function network model and the initial long-short time memory network model respectively, wherein the initialization super parameters at least comprise: the initial radial basis function network model and the initial long-short time memory network model respectively correspond to the initial node number and the initial learning rate.
Before model training, features such as the number of neurons, the number of initial nodes, the initial learning rate and the like, which are respectively corresponding to the initial radial basis function network model and the initial long-short-term memory network model, are determined according to the data features of the partial discharge feature data. For example, since the cable PD characteristic type is used as an input node of the input layer, the node number is 18, the neuron number of the output layer is 4, which is the type of cable partial discharge failure (corona discharge, floating discharge, air gap discharge, creeping discharge), and the learning rate is set to 0.002.
Alternatively, the following formula may be used as an activation function for the initial radial basis function network model (i.e., RBF model), but is not limited to:
wherein mu is t As the center point, sigma t The radial basis width determines how fast the radial basis function decreases. The activation function of gate network gate in the initial long-short time memory network model (i.e. LSTM model) is defined as sigmoid function, and the activation function corresponding to the output layer is tanh function. The sigmoid function is a mathematical function with an S-shaped curve, which can be understood as a squeezing function, the output of which is limited between 0 and 1, thus making the function very useful in probability prediction. the tanh function is one of hyperbolic functions, and tanh () is a hyperbolic tangent. Mathematically, the hyperbolic tangent "tanh" is derived from the basic hyperbolic function hyperbolic sine and hyperbolic cosine.
Step S106, the first test set data in the partial discharge characteristic data are respectively input into the trained radial basis function network model and the trained long-short-time memory network model, and a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-time memory network model are obtained.
By the method, the trained radial basis function network model and the trained long-short-term memory network model are respectively tested based on the first test set data in the partial discharge characteristic data, so that output results respectively output by the two models are obtained and used for evaluating the performance of the models and constructing a cable fault prediction model according to comparison of the output results of the two models.
And step S108, determining a cable fault prediction model based on the first output result and the second output result.
In an optional embodiment, when the first test set data includes a plurality of sets of test data, and the first output result and the second output result are both plural, the determining a cable fault prediction model based on the first output result and the second output result includes: determining comparison results between the plurality of first output results and the corresponding second output results, wherein the comparison results are as follows: a first comparison result in which the first output result is consistent with the corresponding second output result, or a second comparison result in which the first output result is inconsistent with the corresponding second output result; determining a total number of first results in comparison results between the plurality of first output results and corresponding second output results respectively, and the number of the first comparison results; determining a first proportion of the number of the first comparison results to the total number of the first results; and under the condition that the first proportion is larger than a preset proportion threshold value, constructing the cable fault prediction model according to the trained radial basis function network model and the trained long-and-short-term memory network model.
By the above mode, whether the output results (namely the first output result and the second output result) are consistent or not is judged, if the output results are consistent, the prediction is correct, and if the output results are inconsistent, the prediction is incorrect. And then calculating the accuracy of the model according to a first proportion of the number of the predicted correct results to the total number of the results. And further, whether the accuracy meets the requirement or not, namely whether the accuracy is larger than a preset accuracy threshold (or whether the first proportion is larger than a preset proportion threshold) is judged, and if so, a cable fault prediction model is built according to the combination of the trained RBF network model and the trained LSTM network model.
In an alternative embodiment, the method further comprises: under the condition that the first proportion is not greater than a preset proportion threshold value, respectively updating the trained radial basis function network model and the trained long-short-time memory network model by adopting a gradient descent algorithm to obtain a new radial basis function network model and a new long-short-time memory network model; respectively inputting second training set data in the partial discharge characteristic data into the new radial basis function network model and the new long-short-time memory network model for training to obtain a new trained radial basis function network model and a new trained long-short-time memory network model; respectively inputting second test set data in the partial discharge characteristic data into the new trained radial basis function network model and the new trained long-short-term memory network model to obtain a third output result output by the new trained radial basis function network model and a fourth output result output by the new trained long-short-term memory network model; and determining comparison results between the plurality of third output results and the corresponding fourth output results respectively under the condition that the second test set data comprises a plurality of groups of test data, wherein the third output results and the fourth output results are all a plurality of, and the comparison results are: a third comparison result, in which the third output result is consistent with the corresponding fourth output result, or a fourth comparison result, in which the third output result is inconsistent with the corresponding fourth output result, respectively; determining the total number of second results in comparison results between the plurality of third output results and the corresponding fourth output results respectively, and the number of the third comparison results; determining a second proportion of the number of the third comparison results to the total number of the second results; and under the condition that the second proportion is larger than the preset proportion threshold value, constructing the cable fault prediction model according to the new trained radial basis function network model and the new trained long-and-short-term memory network model.
By the method, under the condition that the first proportion is not larger than a preset proportion threshold value, the trained radial basis function network model and the trained long-short-term memory network model do not meet accuracy requirements, a gradient descent algorithm is introduced to update initial weights and initial neuron biases in the initial LSTM network model and the initial RBF network model at the moment, a new radial basis function network model is obtained, model training and model testing operation are conducted again on the new long-short-term memory network model, under the condition that output results corresponding to the new trained radial basis function network model and the new trained long-short-term memory network model respectively meet accuracy requirements, a cable fault prediction model is built based on the new trained radial basis function network model and the new trained long-short-term memory network model, and if the accuracy requirements are still not met, the gradient descent algorithm is continuously adopted to update weights and the neuron biases in the two models respectively. Therefore, the model precision is effectively improved, and the accuracy of cable fault prediction is further improved.
In an alternative embodiment, after determining the cable fault prediction model based on the first output result and the second output result, the method further includes: acquiring to-be-measured characteristic data of the to-be-measured cable; and inputting the characteristic data to be tested into the cable fault prediction model for testing to obtain the discharge type corresponding to the cable to be tested.
By the method, the cable fault prediction model is deployed to the field operation environment to perform field monitoring on the cable to be tested, after the characteristic data to be tested of the cable to be tested are obtained, the characteristic data to be tested are input to the cable fault prediction model to be tested, and then the discharge type corresponding to the cable to be tested can be obtained, so that the real-time monitoring of the partial discharge of the high-voltage cable is realized.
Through the steps S102 to S108, training set data in the cable partial discharge characteristic data are transmitted into a prediction model formed by combining the RBF neural network and the LSTM neural network for training, the trained model is further tested by adopting test set data in the cable partial discharge characteristic data, and the cable fault prediction model meeting the accuracy is obtained after the test. The method achieves the aims of constructing a more accurate cable fault prediction model by integrating various neural network model characteristics and improving the model performance, thereby realizing the technical effects of improving the cable fault prediction model performance, improving the cable fault prediction efficiency and the prediction accuracy, and further solving the technical problems of low cable fault prediction efficiency and poor prediction accuracy caused by poor cable fault prediction model performance in the related technology.
It should be noted that, in the embodiment of the invention, the RBF and LSTM combined neural network is utilized to process and train the cable partial discharge fault data, so that the problems of unstable prediction, low accuracy, large workload, complex flow and the like in the prior art are solved, the cable partial discharge fault data is automatically updated and the prediction result is automatically output, and meanwhile, a gradient descent algorithm is introduced to learn and update the cable partial discharge fault data when the network prediction model updates the weight. The accuracy of prediction is effectively improved, the ability of the workers to prejudge the faults of the transformer is improved, and the intelligent level of the power transformation equipment is also improved. The risk and the artificial misjudgment rate caused by field operation are avoided, a large amount of manpower is saved, the recognition speed is accelerated, and the recognition accuracy is improved. The safe operation of the high-voltage cable and related equipment is effectively ensured.
Based on the foregoing embodiment and the optional embodiments, the present invention proposes an optional implementation, and fig. 2 is a flowchart of an optional cable fault prediction processing method according to an embodiment of the present invention, as shown in fig. 2, where the method includes:
step S1, cable partial discharge fault historical data are obtained to serve as original data, noise reduction processing is conducted, partial discharge characteristic data are extracted, and normalization preprocessing is conducted. The method specifically comprises the following substeps:
Step (1.1): the historical data (namely the partial discharge fault historical data) of the high-voltage cable partial discharge defect site is obtained through past historical monitoring and is used as the original data. With the continuous operation of the cable, the original data can be written in from the update and gradually increased, and the continuously enriched partial discharge fault historical data can gradually improve the prediction accuracy.
Step (1.2): the integrity of the partial discharge failure history data is checked, if there is a missing data, and if there is a missing data, interpolation is used to fill. For example, the missing data is filled with the median of all the values of the characteristic gas which is the same as the missing data in the partial discharge fault history data, and the filled partial discharge fault history data is obtained.
Step (1.3): and carrying out wavelet noise reduction processing on the filled partial discharge fault historical data to obtain noise-reduced partial discharge fault historical data, thereby improving the data signal-to-noise ratio and further obtaining model training input data with higher quality. For example, fourier noise reduction processing is performed on the partial discharge failure history data, so that the signal-to-noise ratio of the signal is improved. Fourier transform formula:
wherein ω represents frequency, t represents time, e -iωt As a complex function.
Step (1.4): and carrying out feature extraction processing on the noise-reduced partial discharge fault historical data, and extracting feature data in the partial discharge fault historical data to obtain feature sample data of the high-voltage cable. Cable Partial Discharge (PD) features in the feature sample data may include, but are not limited to: maximum discharge distribution positive half cycle skew
Negative half-cycle skew of maximum discharge distribution +.>
Maximum discharge distribution positive half cycle prominence +.>
Maximum discharge distribution negative half cycle prominence +.>
Maximum discharge amount distribution asymmetry degree Q m Correlation degree CC of maximum discharge distribution m Positive half cycle skew of average discharge capacity distribution
Negative half-cycle skewness of average discharge capacity distribution +.>
Average discharge amount distribution positive half cycle prominence +.>
Negative half cycle prominence of average discharge capacity distribution +.>
Average discharge amount distribution asymmetry degree Q a Average discharge amount distribution correlation degree CC a Positive half-cycle skew of discharge frequency distribution>
Negative half-cycle skewness of discharge frequency distribution>
Positive half cycle prominence of discharge frequency distribution
Negative half cycle prominence of discharge frequency distribution>
The number of discharge times is divided intoDegree of cloth asymmetry Q n Degree of correlation CC of discharge frequency distribution n The 18 kinds of extracted PD sample data will be sample feature data.
Step (1.5): and carrying out normalization processing on the sample characteristic data to obtain partial discharge characteristic data of the high-voltage cable, so that the training speed can be increased. The normalization formula is:
Wherein x is i For sample characteristic data, y i And (3) normalizing the sample characteristic data (namely, partial discharge characteristic data).
Step S2, respectively determining initialization super parameters (hidden layer, node number and learning rate) corresponding to the initial radial basis function network model and the initial long-short-term memory network model according to the partial discharge characteristic data; the weights and neuron bias are initialized.
Step S3, transmitting one path of first training set data in the partial discharge characteristic data together with the set super-parameters, the initial weight and the initial neuron bias into the RBF network model, transmitting the other path of the first training set data into the LSTM network model, and performing model training to obtain a trained RBF network model and a trained LSTM network model; and respectively inputting the first test set data in the partial discharge characteristic data into the trained RBF network model and the trained LSTM network model to obtain a first output result output by the trained RBF network model and a second output result output by the trained LSTM network model.
And S4, judging whether the output results (namely the first output result and the second output result) are consistent, if so, predicting correctly, and if not, predicting incorrectly. And then calculating the accuracy of the model according to the proportion of the number of the predicted correct results to the total number of the results.
S5, judging whether the accuracy meets the requirement, namely whether the accuracy is larger than a preset accuracy threshold, if so, constructing a cable fault prediction model according to the trained RBF network model and the trained LSTM network model; if not, the initial weights and initial neuron biases in the initial LSTM network model and the initial RBF network model are respectively updated by using a gradient descent algorithm, and then the initial weights and initial neuron biases are respectively transmitted into the RBF network model and the LSTM network model along paths 1 and 2. It should be noted that paths 1 and 2 are performed simultaneously.
And S6, circularly executing the steps S3 to S5 until the prediction result meets the accuracy requirement, and constructing a cable fault prediction model based on the finally obtained new trained radial basis function network model and the new trained long-short-term memory network model.
Step S7, deploying a cable fault prediction model to a field operation environment to perform field monitoring on the cable to be tested, and acquiring the characteristic data to be tested of the cable to be tested; and inputting the characteristic data to be tested into a cable fault prediction model for testing, so as to obtain the discharge type corresponding to the cable to be tested, and realizing the real-time monitoring of the partial discharge of the high-voltage cable. It should be noted that, before outputting the predicted result, the predicted result needs to be subjected to inverse normalization processing, so as to obtain the discharge type corresponding to the cable to be tested.
It should be noted that, the embodiment of the invention is applied to the field of prediction diagnosis of large power grids, a new prediction model is obtained by combining an RBF network model and an LSTM model, and the type of the partial discharge fault is predicted by collecting and self-updating cable partial discharge fault history data and inputting the data as a model. Compared with the traditional neural network model, the model can increase and decrease the PD feature type number according to actual conditions, so that a large amount of complicated workload during data processing is reduced, the problems of low accuracy and low stability of a prediction result of the traditional neural network are avoided, a stable and reliable basis is provided for power station staff to the running state of the transformer, and the safety and stability of the whole power grid are ensured.
The embodiment also provides a cable fault prediction processing device, which is used for implementing the foregoing embodiments and preferred embodiments, and is not described in detail. As used below, the terms "module," "apparatus" may be a combination of software and/or hardware that implements a predetermined function. While the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.
According to an embodiment of the present invention, there is further provided an apparatus embodiment for implementing the above-mentioned cable fault prediction processing method, and fig. 3 is a schematic structural diagram of a cable fault prediction processing apparatus according to an embodiment of the present invention, as shown in fig. 3, where the above-mentioned cable fault prediction processing apparatus includes: a first acquisition module 300, a training module 302, a second acquisition module 304, a determination module 306, wherein:
the first obtaining module 300 is configured to obtain partial discharge characteristic data of the high voltage cable;
the training module 302, coupled to the first obtaining module 300, is configured to input first training set data in the partial discharge feature data to an initial radial basis function network model and an initial long-short-term memory network model, respectively, to perform training, so as to obtain a trained radial basis function network model and a trained long-short-term memory network model;
the second obtaining module 304 is connected to the training module 302, and is configured to input the first test set data in the partial discharge feature data to the trained radial basis function network model and the trained long-short-term memory network model, respectively, to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model;
The determining module 306 is connected to the second obtaining module 304, and is configured to determine a cable fault prediction model based on the first output result and the second output result.
In the embodiment of the present invention, the first obtaining module 300 is configured to obtain partial discharge characteristic data of the high voltage cable; the training module 302, coupled to the first obtaining module 300, is configured to input first training set data in the partial discharge feature data to an initial radial basis function network model and an initial long-short-term memory network model, respectively, to perform training, so as to obtain a trained radial basis function network model and a trained long-short-term memory network model; the second obtaining module 304 is connected to the training module 302, and is configured to input the first test set data in the partial discharge feature data to the trained radial basis function network model and the trained long-short-term memory network model, respectively, to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; the determining module 306 is connected to the second obtaining module 304, and is configured to determine a cable fault prediction model based on the first output result and the second output result, thereby achieving the purpose of building a more accurate cable fault prediction model by integrating various neural network model features and improving the performance of the model, and further achieving the technical effects of improving the performance of the cable fault prediction model, improving the cable fault prediction efficiency and the prediction accuracy, and further solving the technical problems of low cable fault prediction efficiency and poor prediction accuracy caused by the poor performance of the cable fault prediction model in the related art.
It should be noted that each of the above modules may be implemented by software or hardware, for example, in the latter case, it may be implemented by: the above modules may be located in the same processor; alternatively, the various modules described above may be located in different processors in any combination.
Here, the above is to be noted: the first acquisition module 300, the training module 302, the second acquisition module 304, and the determination module 306 correspond to steps S102 to S108 in the embodiment, and the above modules are the same as the examples and application scenarios implemented by the corresponding steps, but are not limited to those disclosed in the above embodiments. It should be noted that the above modules may be run in a computer terminal as part of the apparatus.
It should be noted that, the optional or preferred implementation manner of this embodiment may be referred to the related description in the embodiment, and will not be repeated herein.
The cable fault prediction processing device may further include a processor and a memory, where: the first acquisition module 300, the training module 302, the second acquisition module 304, the determination module 306, etc. are stored in the memory as program modules, which are executed by the processor to implement the corresponding functions.
The processor comprises a kernel, the kernel accesses the memory to call the corresponding program module, and the kernel can be provided with one or more than one. The memory may include volatile memory, random Access Memory (RAM), and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM), among other forms in computer readable media, the memory including at least one memory chip.
According to an embodiment of the present application, there is also provided an embodiment of a nonvolatile storage medium. Optionally, in this embodiment, the nonvolatile storage medium includes a stored program, where the device in which the nonvolatile storage medium is located is controlled to execute any one of the cable failure prediction processing methods when the program runs.
Alternatively, in this embodiment, the above-mentioned nonvolatile storage medium may be located in any one of the computer terminals in the computer terminal group in the computer network or in any one of the mobile terminals in the mobile terminal group, and the above-mentioned nonvolatile storage medium includes a stored program.
Optionally, the program controls the device in which the nonvolatile storage medium is located to perform the following functions when running: obtaining partial discharge characteristic data of a high-voltage cable; respectively inputting the first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training to obtain a trained radial basis function network model and a trained long-short-term memory network model; respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; and determining a cable fault prediction model based on the first output result and the second output result.
According to an embodiment of the present application, there is also provided an embodiment of a processor. Optionally, in this embodiment, the processor is configured to run a program, where any one of the cable fault prediction processing methods is executed when the program runs.
According to an embodiment of the present application, there is also provided an embodiment of a computer program product adapted to perform a program initialized with the steps of any one of the cable fault prediction processing methods described above when executed on a data processing device.
Optionally, the computer program product mentioned above, when executed on a data processing device, is adapted to perform a program initialized with the method steps of: obtaining partial discharge characteristic data of a high-voltage cable; respectively inputting the first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training to obtain a trained radial basis function network model and a trained long-short-term memory network model; respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; and determining a cable fault prediction model based on the first output result and the second output result.
The embodiment of the invention provides an electronic device, which comprises a processor, a memory and a program stored on the memory and capable of running on the processor, wherein the following steps are realized when the processor executes the program: obtaining partial discharge characteristic data of a high-voltage cable; respectively inputting the first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training to obtain a trained radial basis function network model and a trained long-short-term memory network model; respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model; and determining a cable fault prediction model based on the first output result and the second output result.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
In the foregoing embodiments of the present invention, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.
In the several embodiments provided in the present application, it should be understood that the disclosed technology content may be implemented in other manners. The above-described embodiments of the apparatus are merely exemplary, and the division of the modules may be a logic function division, and there may be another division manner when actually implemented, for example, a plurality of modules or components may be combined or may be integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with respect to each other may be through some interface, module or indirect coupling or communication connection of modules, electrical or otherwise.
The modules described above as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, i.e., may be located in one place, or may be distributed over a plurality of modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist alone physically, or two or more modules may be integrated into one module. The integrated modules may be implemented in hardware or in software functional modules.
The integrated modules described above, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable non-volatile storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a non-volatile storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of the embodiments of the present invention. And the aforementioned nonvolatile storage medium includes: a U-disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a removable hard disk, a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.
Claims (10)
1. A cable fault prediction processing method, characterized by comprising:
obtaining partial discharge characteristic data of a high-voltage cable;
respectively inputting first training set data in the partial discharge characteristic data into an initial radial basis function network model and an initial long-short-term memory network model for training to obtain a trained radial basis function network model and a trained long-short-term memory network model;
respectively inputting first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-term memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-term memory network model;
and determining a cable fault prediction model based on the first output result and the second output result.
2. The method of claim 1, wherein in the case where the first test set data includes a plurality of sets of test data, the first output result and the second output result are each a plurality, the determining a cable fault prediction model based on the first output result and the second output result includes:
Determining comparison results between the plurality of first output results and the corresponding second output results, wherein the comparison results are as follows: a first comparison result in which the first output result is consistent with the corresponding second output result, or a second comparison result in which the first output result is inconsistent with the corresponding second output result;
determining a first result total number in comparison results between the plurality of first output results and corresponding second output results respectively, and the number of the first comparison results;
determining a first proportion of the number of the first comparison results to the total number of the first results;
and under the condition that the first proportion is larger than a preset proportion threshold value, constructing the cable fault prediction model according to the trained radial basis function network model and the trained long-and-short-term memory network model.
3. The method according to claim 2, wherein the method further comprises:
under the condition that the first proportion is not greater than a preset proportion threshold value, respectively updating the trained radial basis function network model and the trained long-short-time memory network model by adopting a gradient descent algorithm to obtain a new radial basis function network model and a new long-short-time memory network model;
Respectively inputting second training set data in the partial discharge characteristic data into the new radial basis function network model and the new long-short-time memory network model for training to obtain a new trained radial basis function network model and a new trained long-short-time memory network model;
respectively inputting second test set data in the partial discharge characteristic data into the new trained radial basis function network model and the new trained long-short-term memory network model to obtain a third output result output by the new trained radial basis function network model and a fourth output result output by the new trained long-short-term memory network model;
and under the condition that the second test set data comprises a plurality of groups of test data, and the third output results and the fourth output results are all a plurality of, determining comparison results between the plurality of third output results and the corresponding fourth output results, wherein the comparison results are as follows: a third comparison result, in which the third output result is consistent with the corresponding fourth output result, or a fourth comparison result, in which the third output result is inconsistent with the corresponding fourth output result, respectively;
Determining a total number of second results in comparison results between the plurality of third output results and corresponding fourth output results, respectively, and the number of third comparison results;
determining a second proportion of the number of third comparison results to the total number of second results;
and under the condition that the second proportion is larger than the preset proportion threshold value, constructing the cable fault prediction model according to the new trained radial basis function network model and the new trained long-and-short-term memory network model.
4. The method of claim 1, wherein before the first training set of data in the partial discharge feature data is input to an initial radial basis function network model and an initial long-short-term memory network model for training, respectively, to obtain a trained radial basis function network model and a trained long-short-term memory network model, the method further comprises:
determining the characteristic type corresponding to the partial discharge characteristic data of the high-voltage cable;
based on the feature types, determining initialization hyper-parameters corresponding to the initial radial basis function network model and the initial long-short time memory network model respectively, wherein the initialization hyper-parameters at least comprise: the initial radial basis function network model and the initial long-short time memory network model respectively correspond to the initial node number and the initial learning rate.
5. The method of claim 1, wherein after the determining a cable fault prediction model based on the first output result and the second output result, the method further comprises:
acquiring to-be-measured characteristic data of the to-be-measured cable;
and inputting the characteristic data to be tested into the cable fault prediction model for testing to obtain the discharge type corresponding to the cable to be tested.
6. The method according to any one of claims 1 to 5, wherein the acquiring partial discharge characteristic data of the high voltage cable comprises:
acquiring partial discharge fault history data of a high-voltage cable;
carrying out noise reduction treatment on the partial discharge fault historical data to obtain noise-reduced partial discharge fault historical data;
and carrying out feature extraction processing on the noise-reduced partial discharge fault historical data to obtain the partial discharge feature data.
7. The method of claim 6, wherein the denoising the partial discharge fault history data to obtain denoised partial discharge fault history data comprises:
judging whether missing data exists in the partial discharge fault historical data;
Under the condition that the missing data exists in the partial discharge fault historical data, filling the missing data by adopting an interpolation method to obtain filled partial discharge fault historical data;
and carrying out noise reduction treatment on the filled partial discharge fault historical data to obtain the noise-reduced partial discharge fault historical data.
8. A cable fault prediction processing device, comprising:
the first acquisition module is used for acquiring partial discharge characteristic data of the high-voltage cable;
the training module is used for respectively inputting the first training set data in the partial discharge characteristic data into the initial radial basis function network model and the initial long-short-time memory network model for training to obtain a trained radial basis function network model and a trained long-short-time memory network model;
the second acquisition module is used for respectively inputting the first test set data in the partial discharge characteristic data into the trained radial basis function network model and the trained long-short-time memory network model to obtain a first output result output by the trained radial basis function network model and a second output result output by the trained long-short-time memory network model;
And the determining module is used for determining a cable fault prediction model based on the first output result and the second output result.
9. A non-volatile storage medium storing a plurality of instructions adapted to be loaded by a processor and to perform the cable fault prediction processing method of any one of claims 1 to 7.
10. An electronic device comprising one or more processors and a memory for storing one or more programs, wherein the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the cable fault prediction processing method of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211735073.6A CN116089882A (en) | 2022-12-30 | 2022-12-30 | Cable fault prediction processing method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211735073.6A CN116089882A (en) | 2022-12-30 | 2022-12-30 | Cable fault prediction processing method and device and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116089882A true CN116089882A (en) | 2023-05-09 |
Family
ID=86200408
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211735073.6A Pending CN116089882A (en) | 2022-12-30 | 2022-12-30 | Cable fault prediction processing method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116089882A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116303094A (en) * | 2023-05-10 | 2023-06-23 | 江西财经大学 | Multipath coverage test method based on RBF neural network and individual migration |
| CN117791856A (en) * | 2023-12-20 | 2024-03-29 | 武汉人云智物科技有限公司 | Power grid fault early warning method and device based on inspection robot |
-
2022
- 2022-12-30 CN CN202211735073.6A patent/CN116089882A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116303094A (en) * | 2023-05-10 | 2023-06-23 | 江西财经大学 | Multipath coverage test method based on RBF neural network and individual migration |
| CN117791856A (en) * | 2023-12-20 | 2024-03-29 | 武汉人云智物科技有限公司 | Power grid fault early warning method and device based on inspection robot |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116089882A (en) | Cable fault prediction processing method and device and electronic equipment | |
| CN111210024A (en) | Model training method and device, computer equipment and storage medium | |
| CN111756034B (en) | Transient voltage stability evaluation method for power system based on graph space-time network | |
| CN109507535B (en) | Method and device for predicting operation stage and operation life of transformer substation grounding grid | |
| CN111628501A (en) | AC/DC large power grid transient voltage stability assessment method and system | |
| CN112990330A (en) | User energy abnormal data detection method and device | |
| CN116401532B (en) | Method and system for recognizing frequency instability of power system after disturbance | |
| CN111414703A (en) | Method and device for predicting residual life of rolling bearing | |
| CN112565187A (en) | Power grid attack detection method, system, equipment and medium based on logistic regression | |
| CN112288147A (en) | Method for predicting insulation state of generator stator by BP-Adaboost strong predictor | |
| CN114839492A (en) | Method and device for identifying GIS partial discharge type based on MOBILE NETV3 | |
| CN112949201B (en) | Wind speed prediction method and device, electronic equipment and storage medium | |
| CN112613191A (en) | Cable health state evaluation method and device, computer equipment and storage medium | |
| CN115902557A (en) | Switch cabinet fault diagnosis processing method and device and nonvolatile storage medium | |
| CN113052060B (en) | Bearing residual life prediction method and device based on data enhancement and electronic equipment | |
| CN113469257B (en) | Distribution transformer fault detection method and system | |
| CN112287751B (en) | Excitation surge current identification method, device, computer equipment and storage medium | |
| CN115795360A (en) | Cable fault detection method based on artificial neural network | |
| CN114019296A (en) | Distribution line ground fault identification method based on BP neural network | |
| CN110797906B (en) | Method and device for determining maximum installed capacity of wind power base and storage medium | |
| CN113507389A (en) | Power grid key node identification method and system | |
| CN111612233A (en) | Method and device for obtaining importance scores of power distribution network line re-jump influence factors | |
| CN115034380B (en) | Overhead line power failure prediction neural network compression method and system | |
| CN116304846B (en) | CVT internal insulation abnormality online assessment method based on self-supervision learning | |
| CN117454820A (en) | Effectiveness verification method of transformer winding simulation model based on frequency division frequency response |
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 |