CN113640596A - Converter transformer abnormity detection method and device, computer equipment and storage medium - Google Patents
Converter transformer abnormity detection method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN113640596A CN113640596A CN202110790810.1A CN202110790810A CN113640596A CN 113640596 A CN113640596 A CN 113640596A CN 202110790810 A CN202110790810 A CN 202110790810A CN 113640596 A CN113640596 A CN 113640596A
- Authority
- CN
- China
- Prior art keywords
- converter transformer
- dimension
- network
- component
- primary 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.)
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Abstract
The application relates to a converter transformer abnormity detection method, a converter transformer abnormity detection device, computer equipment and a storage medium. The method comprises the following steps: respectively acquiring detection data of each dimensionality of the converter transformer from each information source; respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network; performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake. The method improves the reliability of the frequent detection of the commutation variation.
Description
Technical Field
The present application relates to the field of power technologies, and in particular, to a converter transformer abnormality detection method, apparatus, computer device, and storage medium.
Background
How to ensure the stable operation of the power grid has been one of the most important research topics in the power industry. The power converter transformer is used as a hub device for voltage grade conversion, electric energy distribution and transfer in a power grid, and safe and stable operation of the power converter transformer is important for stable operation of the whole power grid. The power converter transformer (complex high-voltage equipment consisting of a plurality of subsystems) faces various problems such as insulation aging and the like under long-term operation, the fault probability of the power converter transformer is gradually increased, and once the converter transformer has an operation accident, major production accidents such as equipment damage, premium and even large-area power failure can be caused. Only by finding out various potential faults of the converter transformer as early as possible and making corresponding maintenance schemes according to different fault characteristics can major accidents be avoided. Therefore, the development of the on-line state evaluation method and the fault diagnosis technical research of the converter transformer has important significance for improving the operation and maintenance level of the converter transformer.
The tradition is to changeing the flow and is used the periodic inspection as the main, and the periodic inspection is patrolled and examined including preventive test, periodic overhaul minor repair and regularly, though this kind of fortune dimension mode can avoid the large tracts of land to have a power failure, the condition of examining and repairing inefficiency afterwards, nevertheless because the periodic inspection needs the off-line to go on, has following not enough:
most preventive tests need to be carried out under the condition of shutdown, and the converter transformer cannot be shut down randomly;
the converter state (such as voltage, temperature and the like) after power failure is not consistent with the actual operation condition, so that the judgment accuracy is influenced;
regular maintenance is not continuously monitored in time, and the converter transformer still has the possibility of accidents in the maintenance interval, so that 'under maintenance' occurs;
regular maintenance is carried out according to a specified maintenance cycle strictly, even if the converter is in a good running state, preventive tests and maintenance are carried out according to a plan, so that huge waste in manpower and material resources is caused, and meanwhile, unnecessary damage to the converter body due to too many disassembly and assembly is caused, namely, so-called 'over-maintenance' occurs, so that the power supply reliability is reduced, and the probability of hidden danger is increased.
Disclosure of Invention
In view of the above, it is necessary to provide a converter transformer abnormality detection method, apparatus, computer device and storage medium that can reliably detect the abnormality of the converter transformer.
A converter transformer abnormality detection method comprises the following steps:
respectively acquiring detection data of each dimensionality of the converter transformer from each information source;
respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network;
performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
In one embodiment, the deep confidence network consists of a plurality of restricted Boltzmann machines at a lower layer and a multi-layer feedforward network at an upper layer; a method of training a deep belief network, comprising:
carrying out forward training on each limited Boltzmann machine successively in an unsupervised mode by inputting samples, and taking the output of the limited Boltzmann machine of the current layer as the input of the network of the previous layer to obtain the initial parameters of each layer;
and inputting the output of the limited Boltzmann machine at the uppermost layer into the multilayer feedforward network, performing supervised learning on the multilayer feedforward network by using the labeling result of the input sample, performing back propagation, and adjusting network parameters of each layer.
In one embodiment, the step of inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain the primary diagnosis result of each dimension includes:
respectively inputting the detection data of each dimension into a corresponding trained deep confidence neural network, and extracting the probability distribution characteristics of the detection data through a limited Bolmann machine of the deep neural network;
and the multilayer feedforward network of the deep confidence neural network obtains a primary diagnosis result of the dimension according to the probability distribution characteristics.
In one embodiment, performing decision fusion according to the component dimensions of the converter transformer according to the multidimensional primary diagnostic result to obtain the abnormal detection result of the converter transformer, includes:
respectively obtaining primary diagnosis results related to all components of the converter transformer from the primary diagnosis results of all dimensions; the converter transformer comprises: the device comprises a body, a sleeve, a tap switch, a cooling system and a non-electric quantity protection pile;
fusing the related primary diagnosis result of each component to obtain the state of each component;
and when all the components are in a normal state, determining that the overall state of the converter transformer is in a normal state.
In one embodiment, fusing the primary diagnostic results associated with each component to obtain the status of each component includes:
and acquiring the weight of each dimension in the primary diagnostic result related to each component, and fusing the primary diagnostic results related to each component to obtain the state of each component.
In one embodiment, the method further comprises: and setting different weights for the data dimensionality of the sensing collection related to each part by using the distance of the sensor.
In one embodiment, the method further comprises: and setting different weights for the data dimensionality acquired by the sensors related to each part by using the importance and contribution degree of the data dimensionality acquired by the sensors to the fusion result.
An abnormality detection device for a converter transformer, comprising:
the data acquisition module is used for respectively acquiring detection data of each dimensionality of the converter transformer from each information source;
the primary diagnosis module is used for respectively inputting the detection data of each dimension into the corresponding trained depth confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network;
the fusion diagnosis module is used for performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric protection stake
A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:
respectively acquiring detection data of each dimensionality of the converter transformer from each information source;
respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network;
performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:
respectively acquiring detection data of each dimensionality of the converter transformer from each information source;
respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network;
performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
The converter transformer abnormity detection method, the device, the computer equipment and the storage medium are introduced into the converter transformer state evaluation and fault diagnosis system by utilizing the powerful data search, processing and decision-making capabilities of the artificial intelligence technology, so that the defect that a data source is too extensive in the traditional evaluation and fault diagnosis method can be effectively overcome, and the real state and the fault type of the converter transformer can be accurately and effectively analyzed from massive and complex converter transformer characteristic data. The method has important theoretical significance and engineering value for improving the reliability of the converter transformer and improving the state evaluation and maintenance strategy of the converter transformer.
Drawings
FIG. 1 is a diagram illustrating an exemplary embodiment of an application of a converter transformer anomaly detection method;
FIG. 2 is a schematic flow chart of a converter transformer anomaly detection method according to an embodiment;
FIG. 3 is a schematic diagram of a fusion process in one embodiment;
FIG. 4 is a schematic diagram of a deep belief network in another embodiment;
FIG. 5 is a diagram illustrating a deep belief network training process in one embodiment;
FIG. 6 is a graph of a hierarchy of transformer evaluations in one embodiment;
fig. 7 is a block diagram showing a structure of an abnormality detection apparatus for a converter transformer in one embodiment;
FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The converter transformer abnormity detection method provided by the application can be applied to the application environment shown in fig. 1. The terminal 104 is connected to the plurality of sensors 102 through a network, acquires monitoring data of each sensor, and stores rheological test data. The terminal respectively obtains detection data of each dimensionality of the converter transformer from each information source; respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network; performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake. The terminal 102 may be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices.
In an embodiment, as shown in fig. 2, a converter transformer abnormality detection method is provided, which is described by taking the example that the method is applied to the terminal in fig. 1, and includes the following steps:
step 202, respectively obtaining detection data of each dimension of the converter transformer from each information source.
The information source refers to the source of the detection data of each dimension. Specifically, the information source of the monitoring data may be the monitoring data of the sensor, or may be experimental data obtained by performing an experiment on the converter transformer equipment. Wherein, the monitoring dimension and the experiment dimension of sensor have the multiple, therefore the detection data of converter flow has a plurality of dimensions, include: oil chromatogram, oil temperature, winding temperature, other temperatures (smart robot), voltage (grid side/valve side), current (grid side/valve side), neutral point current, core ground current, SER data, noise, vibration, and oil chromatogram electrification monitoring data, no-load pressurization impedance data, handover test and preventive test data, etc.
Step 204, respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network.
As shown in fig. 3, each dimension corresponds to a trained deep confidence network, and the detection information of each information source is input into the corresponding deep confidence network, so as to obtain the primary diagnosis result of the corresponding dimension.
Specifically, the primary diagnostic result is whether the detection data of one dimension is normal or not. The deep confidence network is obtained by training according to historical data of one dimension. The deep confidence network is used for extracting the characteristics of the detection information of each information source to obtain a primary diagnosis result of the dimension. For example, for each dimension: the method comprises the steps of obtaining a primary diagnosis result of each dimension by respectively adopting a corresponding depth confidence network through oil chromatography, oil temperature, winding temperature, other temperatures (intelligent robot), voltage (network side/valve side), current (network side/valve side), neutral point current, iron core grounding current, SER data, noise, vibration, oil chromatography electrification monitoring data, no-load pressurization impedance data, handover test and preventive test data and the like. E.g., primary detection of oil chromatogram dimension, primary detection of oil temperature dimension, etc.
Step 206, performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
That is to say, the corresponding dimensions of each information source are used for determining the dimension of the component to which the information source belongs, such as the primary diagnosis results of each information source corresponding to the body, the sleeve, the tap switch, the cooling system and the non-electric quantity protection pile respectively, and the component dimensions are taken as the unit for fusion to obtain the diagnosis results of each component, so that the abnormal detection result of the converter transformer is obtained.
The converter transformer abnormity detection method is introduced into the converter transformer state evaluation and fault diagnosis system by utilizing the powerful data searching, processing and decision-making capabilities of the artificial intelligence technology, can effectively overcome the defect that the data source of the traditional evaluation and fault diagnosis method is too one-sided, and can accurately and effectively analyze the real state and fault type of the converter transformer from massive and complex converter transformer characteristic data. The method has important theoretical significance and engineering value for improving the reliability of the converter transformer and improving the state evaluation and maintenance strategy of the converter transformer.
In another embodiment, as shown in FIG. 4, the deep belief network consists of a lower level of multiple Restricted Boltzmann Machines (RBMs) and an upper level of a multi-layer feed-forward network (BPNN).
A method of training a deep belief network, comprising: carrying out forward training on each limited Boltzmann machine successively in an unsupervised mode by inputting samples, and taking the output of the limited Boltzmann machine of the current layer as the input of the network of the previous layer to obtain the initial parameters of each layer; and inputting the output of the limited Boltzmann machine at the uppermost layer into the multilayer feedforward network, performing supervised learning on the multilayer feedforward network by using the labeling result of the input sample, performing back propagation, and adjusting network parameters of each layer.
In particular, a deep belief neural network is one of deep neural networks, which is structurally a neural network with more than one hidden layer. Compared with the traditional shallow neural network, the deep neural network can also provide a modeling basis for a complex nonlinear system, but in order to achieve the purpose of improving the training capability of the model as much as possible, a plurality of layers are added on the basis of the shallow neural network, so that the requirement of the system on higher layers is met.
And (3) constructing a deep confidence network, and firstly training a limited Boltzmann machine. A Restricted Boltzmann Machine (RBM) is a component of a Deep Belief neural network (DBN), and the RBM is mainly used for training a model by an unsupervised learning method and is an energy-based model. The method is composed of a hidden layer H and a visual layer V, a random neural network model which is symmetrically connected and has no feedback is adopted, the input of a data source is input in the visual layer V, and the hidden layer H is used for extracting the characteristics of data. Each layer is represented by a vector, and the number of neurons in each layer is represented by the dimension of the vector. The RBN visual is in the form of a bipartite graph, and the visual layer V and the hidden layer H are binary variables taking 0 or 1. Only the visual layer neurons and the hidden layer neurons are in a fully connected relationship, and no connection relationship exists between the hidden layer neurons or between the visual layer neurons.
The maximum probability distribution of a training sample can be obtained in the RBM training process, the decisive factor of the sample is the weight W, and the scheme adopts a learning algorithm of contrast divergence to train the RMB network so as to improve the calculation speed and precision.
As shown in fig. 5, the training of the deep confidence network is mainly divided into two steps:
the first step is forward training of a Deep Belief Network (DBN), each limited Boltzmann machine (RBM) is trained in an unsupervised mode successively by inputting training samples, each layer performs feature recognition and extraction on data of the previous layer, and then the output data is used as an input part of the network of the previous layer.
The second step is the phase of fine tuning the whole structure, the fine tuning phase of the DBN is complementary to the first step. The data of the entire network training may be over-fitted only if the training is not fine-tuned. It can be seen from fig. 4-5 that the BP network is located at the upper layer of the whole DBN, the input of the BP network is the final output data from the lower layer RBM, and the BP network performs supervised learning on the model, performs back propagation and adjusts and optimizes the parameters of each layer, so that the structure of the network reaches the overall optimal structure, and the phenomenon that the training time of a large data source is too long is avoided.
Correspondingly, the detection data of each dimension are respectively input into the corresponding trained deep confidence network to obtain the primary diagnosis result of each dimension, and the method comprises the following steps: respectively inputting the detection data of each dimension into a corresponding trained deep confidence neural network, and extracting the probability distribution characteristics of the detection data through a limited Bolmann machine of the deep neural network; and the multilayer feedforward network of the deep confidence neural network obtains a primary diagnosis result of the dimension according to the probability distribution characteristics.
The RBM training process is to calculate a maximum probability distribution which can generate a training sample, respectively input the detection data of each dimension into a corresponding trained deep confidence neural network, and extract the probability distribution characteristics of the detection data through a limited Bolmann machine of the deep neural network. After training, the model can obtain corresponding diagnosis results according to the predicted probability distribution characteristics of the detection data aiming at the distribution characteristics of the data in four states of 'good', 'common', 'attention' and 'severe' of each detection data. The diagnosis results include "good", "normal", "attention" and "severe".
In another embodiment, performing decision fusion according to component dimensions of the converter transformer according to the multidimensional primary diagnostic result to obtain an abnormal detection result of the converter transformer, includes: respectively obtaining primary diagnosis results related to all components of the converter transformer from the primary diagnosis results of all dimensions; the converter transformer comprises: the device comprises a body, a sleeve, a tap switch, a cooling system and a non-electric quantity protection pile; fusing the related primary diagnosis result of each component to obtain the state of each component; and when all the components are in a normal state, determining that the overall state of the converter transformer is in a normal state.
Specifically, the converter transformer evaluation is performed as a three-stage evaluation process, as shown in fig. 6. The first stage is a component layer which comprises a body, a sleeve, a tap switch, a cooling system and a non-electric quantity protection pile, and when the states of all the components are normal, the whole state of the converter transformer is a normal state;
the second level is a project level, namely parameter types required by state evaluation of each component are mainly divided into three categories, namely state performance parameter data, operation inspection information data and technical performance parameter data;
the third level is an index layer, namely specific index parameters included by different types of parameter systems in the second level, such as qualitative or quantitative indexes of electrical performance, test data, routing inspection information, parameter information and the like.
And determining a diagnosis result of one level by using the result of the previous level respectively from the third level to the first level based on the division of the three-level evaluation.
Specifically, primary diagnosis results related to all parts of the converter transformer, primary diagnosis results related to a body, primary diagnosis results related to a casing, primary diagnosis results related to a tap changer, primary diagnosis results related to a cooling system and primary diagnosis results related to a non-electric-quantity protection pile are obtained from the primary diagnosis results of all dimensions.
And further, by taking the components as units, fusing the primary diagnosis results of the relevant dimensions to obtain the states of the components, determining the overall state of the converter transformer according to the states of the functional components of the converter transformer, and determining the overall state of the converter transformer to be the normal state when the states of the components are normal.
By adopting the method, more accurate diagnosis results can be provided compared with a single information source.
Wherein, fusing the primary diagnosis result related to each component to obtain the state of each component, comprising: and acquiring the weight of each dimension in the primary diagnostic result related to each component, and fusing the primary diagnostic results related to each component to obtain the state of each component.
The structure of the converter transformer is complex, data of the converter transformer are inconsistent, some state changes are not caused by one reason, and another factor can promote the state changes, so that the factors for comprehensively evaluating the state of the converter transformer system have great ambiguity, how to comprehensively analyze inconsistent information of the converter transformer by using related knowledge, and a clear explanation is given by performing related fusion.
In this embodiment, fusion is achieved by setting different weights for the primary diagnostic result associated with each component during fusion.
In one embodiment, the distance of the sensors is used to set different weights for the data dimension of the sensing collection associated with each component.
In particular, the data weight is corrected by using the distance function, and the data collected by the sensor is not necessarily completely reliable in the operation process of the equipment. In order to obtain a better fusion result, the sensor data needs to be further checked for consistency, and the item adopts a distance function to check: if the distance between the two sensors is close, the support degree of each other is high, and conversely, the support degree is low.
In one embodiment, different weights are set for the data dimensions of the sensing collection related to each component by using the importance and contribution degree of the data dimensions collected by the sensors to the fusion result.
Specifically, the weight of the sensor is corrected by using a Delphi method, a certain weight w (si) is set for the sensor according to the experience of an expert system and field workers and the prior knowledge of an actual application scene, and different weights represent the importance and contribution degree of the sensor measurement value to a fusion result in the current state.
According to the method, different weights are set for the data dimensionality acquired by the sensor based on actual conditions, fusion is carried out based on the weights, and a fused abnormal detection result can be obtained.
It should be understood that, although the steps in the flowchart of fig. 2 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a portion of the steps in fig. 2 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.
In one embodiment, as shown in fig. 7, there is provided a converter transformer abnormality detection apparatus including:
a data obtaining module 702, configured to obtain detection data of each dimension of the converter transformer from each information source respectively.
The primary diagnosis module 704 is used for respectively inputting the detection data of each dimension into the corresponding trained depth confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network.
The fusion diagnosis module 706 is configured to perform decision fusion according to the component dimensions of the converter transformer according to the multidimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
The converter transformer abnormity detection device is introduced into the converter transformer state evaluation and fault diagnosis system by utilizing the powerful data searching, processing and decision-making capabilities of the artificial intelligence technology, can effectively overcome the defect that the data source of the traditional evaluation and fault diagnosis method is too extensive, and can accurately and effectively analyze the real state and fault type of the converter transformer from massive and complex converter transformer characteristic data. The method has important theoretical significance and engineering value for improving the reliability of the converter transformer and improving the state evaluation and maintenance strategy of the converter transformer.
In another embodiment, the deep belief network consists of a lower level plurality of restricted boltzmann machines and an upper level multi-layer feed-forward network; the flow transformer abnormality detection device further includes:
the training module is used for carrying out forward training on each limited Boltzmann machine successively in an unsupervised mode through input samples, and taking the output of the limited Boltzmann machine of the current layer as the input of the network of the previous layer to obtain the initial parameters of each layer; and inputting the output of the limited Boltzmann machine at the uppermost layer into the multilayer feedforward network, performing supervised learning on the multilayer feedforward network by using the labeling result of the input sample, performing back propagation, and adjusting network parameters of each layer.
In another embodiment, the primary diagnosis module is used for respectively inputting the detection data of each dimension into a corresponding trained deep confidence neural network, and extracting probability distribution characteristics of the detection data through a restricted Boltmann machine of the deep neural network; and the multilayer feedforward network of the deep confidence neural network obtains a primary diagnosis result of the dimension according to the probability distribution characteristics.
In another embodiment, the fusion diagnosis module is used for respectively obtaining primary diagnosis results related to all components of the converter transformer from the primary diagnosis results of all dimensions; the converter transformer comprises: the device comprises a body, a sleeve, a tap switch, a cooling system and a non-electric quantity protection pile; fusing the related primary diagnosis result of each component to obtain the state of each component; and when all the components are in a normal state, determining that the overall state of the converter transformer is in a normal state.
In another embodiment, the fusion diagnosis module is further configured to obtain weights of dimensions in the primary diagnosis result related to each component, and fuse the primary diagnosis result related to each component to obtain the state of each component.
Wherein, the distance of the sensor is utilized to set different weights for the data dimension of the sensing collection related to each part.
The importance and contribution degree of the data dimension acquired by the sensor to the fusion result are utilized to set different weights for the data dimension acquired by the sensor related to each component.
For specific limitations of the converter transformer abnormality detection device, reference may be made to the above limitations on the converter transformer abnormality detection method, and details are not described herein again. All or part of the modules in the converter transformer abnormality detection device can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as shown in fig. 8. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to realize a converter transformer abnormality detection method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the converter transformer abnormality detection method of the above embodiments when executing the computer program.
In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the steps of the converter transformer abnormality detection method of the above embodiments.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A converter transformer abnormality detection method comprises the following steps:
respectively acquiring detection data of each dimensionality of the converter transformer from each information source;
respectively inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network;
performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
2. The method of claim 1, wherein the deep belief network is comprised of a lower level plurality of restricted boltzmann machines and an upper level multi-layer feed forward network; a method of training a deep belief network, comprising:
carrying out forward training on each limited Boltzmann machine successively in an unsupervised mode by inputting samples, and taking the output of the limited Boltzmann machine of the current layer as the input of the network of the previous layer to obtain the initial parameters of each layer;
and inputting the output of the limited Boltzmann machine at the uppermost layer into the multilayer feedforward network, performing supervised learning on the multilayer feedforward network by using the labeling result of the input sample, performing back propagation, and adjusting network parameters of each layer.
3. The method of claim 2, wherein inputting the detection data of each dimension into the corresponding trained deep confidence network to obtain the primary diagnosis result of each dimension comprises:
respectively inputting the detection data of each dimension into a corresponding trained deep confidence neural network, and extracting the probability distribution characteristics of the detection data through a limited Bolmann machine of the deep neural network;
and the multilayer feedforward network of the deep confidence neural network obtains a primary diagnosis result of the dimension according to the probability distribution characteristics.
4. The method according to claim 1, wherein performing decision fusion according to component dimensions of the converter transformer according to the multidimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer, comprises:
respectively obtaining primary diagnosis results related to all components of the converter transformer from the primary diagnosis results of all dimensions; the converter transformer comprises: the device comprises a body, a sleeve, a tap switch, a cooling system and a non-electric quantity protection pile;
fusing the related primary diagnosis result of each component to obtain the state of each component;
and when all the components are in a normal state, determining that the overall state of the converter transformer is in a normal state.
5. The method of claim 4, wherein fusing the primary diagnostic results associated with each component to obtain the status of each component comprises:
and acquiring the weight of each dimension in the primary diagnostic result related to each component, and fusing the primary diagnostic results related to each component to obtain the state of each component.
6. The method of claim 5, further comprising: and setting different weights for the data dimensionality of the sensing collection related to each part by using the distance of the sensor.
7. The method of claim 5, further comprising: and setting different weights for the data dimensionality acquired by the sensors related to each part by using the importance and contribution degree of the data dimensionality acquired by the sensors to the fusion result.
8. An abnormality detection device for a converter transformer, comprising:
the data acquisition module is used for respectively acquiring detection data of each dimensionality of the converter transformer from each information source;
the primary diagnosis module is used for respectively inputting the detection data of each dimension into the corresponding trained depth confidence network to obtain a primary diagnosis result of each dimension; each dimension corresponds to a trained deep confidence network;
the fusion diagnosis module is used for performing decision fusion according to the component dimensions of the converter transformer according to the multi-dimensional primary diagnosis result to obtain an abnormal detection result of the converter transformer; wherein the component dimensions include a body, a sleeve, a tap changer, a cooling system, and a non-electric quantity protection stake.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110790810.1A CN113640596A (en) | 2021-07-13 | 2021-07-13 | Converter transformer abnormity detection method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110790810.1A CN113640596A (en) | 2021-07-13 | 2021-07-13 | Converter transformer abnormity detection method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113640596A true CN113640596A (en) | 2021-11-12 |
Family
ID=78417260
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110790810.1A Pending CN113640596A (en) | 2021-07-13 | 2021-07-13 | Converter transformer abnormity detection method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113640596A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113888061A (en) * | 2021-12-07 | 2022-01-04 | 中国南方电网有限责任公司超高压输电公司广州局 | Method and device for evaluating running state of converter transformer |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106841846A (en) * | 2016-12-19 | 2017-06-13 | 广东电网有限责任公司电力调度控制中心 | A kind of transformer state analysis and fault diagnosis method and system |
| CN108196143A (en) * | 2017-12-11 | 2018-06-22 | 囯网河北省电力有限公司电力科学研究院 | Power Transformer Faults depth diagnostic method and terminal device |
| CN109669087A (en) * | 2019-01-31 | 2019-04-23 | 国网河南省电力公司 | A kind of method for diagnosing fault of power transformer based on Multi-source Information Fusion |
| CN109885948A (en) * | 2019-02-28 | 2019-06-14 | 沈阳天眼智云信息科技有限公司 | Method for diagnosing fault of power transformer |
| CN111259834A (en) * | 2020-01-20 | 2020-06-09 | 武汉大学 | Transformer fault diagnosis method and system based on integrated deep belief network |
| CN111325233A (en) * | 2019-10-29 | 2020-06-23 | 国网辽宁省电力有限公司电力科学研究院 | Transformer fault detection method and device |
| WO2020140022A1 (en) * | 2018-12-27 | 2020-07-02 | Guruprasad Srinivasan | System and method for fault detection of components using information fusion technique |
| CN111650451A (en) * | 2020-04-28 | 2020-09-11 | 中国南方电网有限责任公司超高压输电公司广州局 | Converter transformer fault reason identification method and system |
| CN112580254A (en) * | 2020-11-30 | 2021-03-30 | 中国南方电网有限责任公司超高压输电公司广州局 | Multi-dimensional analysis and early warning method for converter transformer based on multi-level decision fusion |
| CN112633493A (en) * | 2020-12-01 | 2021-04-09 | 北京理工大学 | Fault diagnosis method and system for industrial equipment data |
-
2021
- 2021-07-13 CN CN202110790810.1A patent/CN113640596A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106841846A (en) * | 2016-12-19 | 2017-06-13 | 广东电网有限责任公司电力调度控制中心 | A kind of transformer state analysis and fault diagnosis method and system |
| CN108196143A (en) * | 2017-12-11 | 2018-06-22 | 囯网河北省电力有限公司电力科学研究院 | Power Transformer Faults depth diagnostic method and terminal device |
| WO2020140022A1 (en) * | 2018-12-27 | 2020-07-02 | Guruprasad Srinivasan | System and method for fault detection of components using information fusion technique |
| CN109669087A (en) * | 2019-01-31 | 2019-04-23 | 国网河南省电力公司 | A kind of method for diagnosing fault of power transformer based on Multi-source Information Fusion |
| CN109885948A (en) * | 2019-02-28 | 2019-06-14 | 沈阳天眼智云信息科技有限公司 | Method for diagnosing fault of power transformer |
| CN111325233A (en) * | 2019-10-29 | 2020-06-23 | 国网辽宁省电力有限公司电力科学研究院 | Transformer fault detection method and device |
| CN111259834A (en) * | 2020-01-20 | 2020-06-09 | 武汉大学 | Transformer fault diagnosis method and system based on integrated deep belief network |
| CN111650451A (en) * | 2020-04-28 | 2020-09-11 | 中国南方电网有限责任公司超高压输电公司广州局 | Converter transformer fault reason identification method and system |
| CN112580254A (en) * | 2020-11-30 | 2021-03-30 | 中国南方电网有限责任公司超高压输电公司广州局 | Multi-dimensional analysis and early warning method for converter transformer based on multi-level decision fusion |
| CN112633493A (en) * | 2020-12-01 | 2021-04-09 | 北京理工大学 | Fault diagnosis method and system for industrial equipment data |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113888061A (en) * | 2021-12-07 | 2022-01-04 | 中国南方电网有限责任公司超高压输电公司广州局 | Method and device for evaluating running state of converter transformer |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111709447A (en) | Power grid abnormality detection method and device, computer equipment and storage medium | |
| Sharma et al. | Review of artificial intelligence techniques application to dissolved gas analysis on power transformer | |
| Li et al. | A transformer fault diagnosis method based on parameters optimization of hybrid kernel extreme learning machine | |
| CN107656518A (en) | A kind of system failure inference method | |
| Xie et al. | Fault isolating and grading for li-ion battery packs based on pseudo images and convolutional neural network | |
| Chernov et al. | Intelligent decision support for power grids using deep learning on small datasets | |
| CN103675518B (en) | Oil-immersed and high-voltage device maintenance method and device | |
| CN113640596A (en) | Converter transformer abnormity detection method and device, computer equipment and storage medium | |
| Esmaeili Nezhad et al. | A review of the applications of machine learning in the condition monitoring of transformers | |
| Zhang et al. | Condition assessment of stay cables through enhanced time series classification using a deep learning approach | |
| CN105741184A (en) | Transformer state evaluation method and apparatus | |
| Xie et al. | The application of echo state network and recurrent multilayer perceptron in rotating machinery fault prognosis | |
| Huang et al. | Data mining for oil‐insulated power transformers: an advanced literature survey | |
| CN116596301A (en) | Cable multi-state variable evaluation model design method and device based on zero-order learning method | |
| Negri et al. | Artificial Intelligence and Smart Grids: the Cable Joint Test Case | |
| Chokr et al. | Feature extraction-reduction and machine learning for fault diagnosis in PV panels | |
| Firos et al. | Fault Detection in Power Transmission Lines Using AI Model | |
| Jose et al. | Fuzzy logic based fault Diagnosis in Induction motor | |
| Alyunov et al. | Identification of parameters of power transformer models using artificial intelligence methods | |
| Zhang et al. | Neural Network Based on Health Monitoring Electrical Equipment Fault and Biomedical Diagnosis | |
| Dineva et al. | Insulation state assessment of rotating electrical machines by employing generalized additive models | |
| Qu et al. | Remaining Useful Life Prediction of High-Speed Railroad Contact Network Based on Stacking Integrated Attention-LSTM-CNN Deep Learning | |
| Dunwen et al. | A trainsient voltage stability evaluation model based on morphological similarity distance online calculation | |
| Hadiki et al. | Transformers Faults Prediction Using Machine Learning Approach | |
| Bai et al. | Abnormal Detection Scheme of Substation Equipment based on Intelligent Fusion Terminal |
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 |