CN117607756A - Fuse performance test platform based on antagonistic neural network - Google Patents
Fuse performance test platform based on antagonistic neural network Download PDFInfo
- Publication number
- CN117607756A CN117607756A CN202410076051.6A CN202410076051A CN117607756A CN 117607756 A CN117607756 A CN 117607756A CN 202410076051 A CN202410076051 A CN 202410076051A CN 117607756 A CN117607756 A CN 117607756A
- Authority
- CN
- China
- Prior art keywords
- data
- module
- neural network
- network
- fuse
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 60
- 230000003042 antagnostic effect Effects 0.000 title claims abstract description 49
- 238000011056 performance test Methods 0.000 title claims abstract description 35
- 238000012360 testing method Methods 0.000 claims abstract description 73
- 238000012549 training Methods 0.000 claims description 16
- 238000013500 data storage Methods 0.000 claims description 11
- 238000000034 method Methods 0.000 claims description 11
- 238000005070 sampling Methods 0.000 claims description 9
- 230000007613 environmental effect Effects 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 239000007787 solid Substances 0.000 claims 2
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 230000002349 favourable effect Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 7
- 239000002131 composite material Substances 0.000 description 5
- 238000004088 simulation Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000010998 test method Methods 0.000 description 2
- QYAPHLRPFNSDNH-MRFRVZCGSA-N (4s,4as,5as,6s,12ar)-7-chloro-4-(dimethylamino)-1,6,10,11,12a-pentahydroxy-6-methyl-3,12-dioxo-4,4a,5,5a-tetrahydrotetracene-2-carboxamide;hydrochloride Chemical compound Cl.C1=CC(Cl)=C2[C@](O)(C)[C@H]3C[C@H]4[C@H](N(C)C)C(=O)C(C(N)=O)=C(O)[C@@]4(O)C(=O)C3=C(O)C2=C1O QYAPHLRPFNSDNH-MRFRVZCGSA-N 0.000 description 1
- 102100026127 Clathrin heavy chain 1 Human genes 0.000 description 1
- 101000912851 Homo sapiens Clathrin heavy chain 1 Proteins 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M17/00—Testing of vehicles
- G01M17/007—Wheeled or endless-tracked vehicles
-
- 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
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/74—Testing of fuses
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The invention discloses a fuse performance test platform based on an antagonistic neural network, which comprises an information acquisition module, a scene reconstruction module, a circuit model and a test module, wherein the scene reconstruction module is respectively connected with the information acquisition module, the circuit model and the test module, and the information acquisition module is used for acquiring real circuit operation parameters; the scene reconstruction module comprises an antagonistic neural network and generates synthetic data of the real-like data by optimizing the antagonistic neural network; the circuit model is used for providing input data for the scene reconstruction module; the test module is used for executing the performance test of the fuse. Compared with the prior art, the invention has the beneficial effects that: the test platform realizes reconstruction of complex working scenes of the electric automobile through the antagonistic neural network, and can be applied to performance test of the electric automobile fuse, and the test platform can meet test requirements of special high-performance fuses. The test platform is favorable for developing high-performance electric automobile fuses.
Description
Technical Field
The invention relates to the technical field of power electronics, in particular to a fuse performance test platform based on an antagonistic neural network.
Background
The development of electric automobile products is focused mainly on improving the journey and power parameters, so that various manufacturers are studying how to provide larger capacity batteries and how to provide higher power. However, while a larger capacity battery and higher power are being pursued, the load of the automobile circuit is inevitably increased, and the requirements for the fuse of the electric automobile are increasing.
The electric automobile fuse is a necessary device for protecting the electric automobile and components thereof, and is also an important guarantee for improving the electric automobile technology. In order to protect important parts such as a motor and a battery pack of an electric vehicle from serious damage due to surge and malfunction, it is necessary to select an appropriate fuse. In other words, the fuse is an important part of the power train of the electric vehicle.
In order to make hybrid vehicles and electric vehicles compete with internal combustion engine vehicles, a major technical problem to be overcome is to develop fuses with excellent performance. With the popularization of electric vehicles and hybrid electric vehicles, the requirements on the electric vehicles will be larger and larger, and the power, speed and endurance mileage of the electric vehicles will be larger and larger. These requirements present many challenges to the electric vehicle industry and device manufacturers must address the performance issues of fuses.
Fuses fall into two categories, one being a relatively inexpensive universal fuse, however such fuses do not provide critical protection during their application; another type is a specially designed, dedicated fuse, which is typically implemented into the system early in the design cycle to best protect all devices around the fuse. In the long term, dedicated fuses would be more suitable for use in electric vehicles than general purpose fuses, and while dedicated designs would be more expensive, these costs are worth considering the inexpensive battery pack of electric vehicles. The general fuse can meet general safety requirements of the electric automobile, but gradually damages expensive components in the electric automobile, and increases maintenance costs.
However, there is no performance testing platform for testing dedicated fuses in the prior art. The test requirements of dedicated fuses have not been met using test methods and equipment for universal fuses.
Therefore, there is a need to develop a more advanced dedicated fuse test platform for performing the performance test of the electric automobile fuse under the complex working conditions.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a fuse performance test platform based on an antagonistic neural network, which comprises an information acquisition module, a scene reconstruction module, a circuit model and a test module, wherein the scene reconstruction module is respectively connected with the information acquisition module, the circuit model and the test module, and the information acquisition module is used for acquiring real circuit operation parameters; the scene reconstruction module comprises an antagonistic neural network, and generates synthesized data of the real-like data by optimizing the antagonistic neural network; the circuit model is used for providing input data to the scene reconstruction module; the test module is used for executing performance test of the fuse.
Preferably, the countermeasure neural network includes a generator network for generating synthetic data that resembles real data, and a discriminator network for discriminating whether the data is real data or synthetic data, and the generator network and the discriminator network are capable of iterative training.
Preferably, the cross entropy loss function of the criteria used to train the arbiter network is:
wherein,,/>is the identifier network +.>Output data of->Generator network parameters, < >>Is a discriminator network parameter.
Preferably, the error function of the antagonistic neural network is:
in the method, in the process of the invention,representing the number of real data points +.>Representing the number of synthetic data points, +.>Representing the arbiter network for synthetic data input from the generator network>Output of->Representing the arbiter network for real data +.>Is provided.
Preferably, the training mode of the generator network and the discriminator network is a random gradient descent end-to-end mode, and the gradients are evaluated by back propagation.
Preferably, the system further comprises a data storage module, wherein the data storage module is respectively connected with the information acquisition module and the scene reconstruction module, and is used for storing data obtained from the information acquisition module.
Preferably, the device further comprises a data output module, wherein the data output module is respectively connected with the scene reconstruction module and the test module, the data output module is used for outputting data generated by the scene reconstruction module to the test module, and the output data of the data output module are voltage data and current data.
Preferably, the real circuit operating parameters include current, voltage and ambient temperature data at the sampling points. The sampling points are nodes in the electric automobile circuit, the information acquisition module comprises a current collector and a voltage collector, and the information acquisition module is used for acquiring data of the sampling points when the electric automobile operates under different working conditions.
Preferably, the working conditions comprise European endurance test working conditions, global unified light vehicle test circulation working conditions, chinese light vehicle driving working conditions, heavy commercial vehicle driving working conditions, U.S. environmental protection agency working conditions, power system short circuit working conditions, collision working conditions, vehicle turning working conditions, wading and heavy rain working conditions and battery overcharge and discharge working conditions.
Preferably, the test module is used for testing the working temperature, rated voltage, rated current, short-circuit interception capability, fusing characteristic and joule integral of the fuse, and comprises a solid-state switch and a test circuit, wherein the solid-state switch can control a power supply, a resistor and an inductor to be connected into the test circuit.
Compared with the prior art, the invention has the following beneficial effects:
the test platform based on the antagonistic neural network provided by the invention realizes reconstruction of complex working scenes of the electric automobile by adopting the antagonistic neural network, can be applied to performance test of the electric automobile fuse, and can meet the test requirement of a special high-performance fuse. The fuse performance test platform based on the countermeasure to neural network can be applied to fuse performance tests of various scenes, meets the diversified demands of products, realizes a test means close to reality, has practicability, and can be popularized and applied on a large scale. The novel test platform is favorable for developing the high-performance electric automobile fuse.
Drawings
The invention is further described below with reference to the accompanying drawings:
FIG. 1 is a schematic diagram of a test platform for testing performance of a fuse based on an antagonistic neural network according to a first embodiment of the present invention;
FIG. 2 is a schematic diagram of an antagonistic neural network according to an embodiment of the present invention;
FIG. 3 is a graph showing the running speed of an electric vehicle under different conditions according to the first embodiment of the present invention;
fig. 4 is an ampere-second characteristic diagram of an electric automobile fuse in the first embodiment of the invention;
fig. 5 is a graph of power loss of an electric automobile fuse according to an embodiment of the present invention;
FIG. 6 is a graph showing the temperature derating of an electric vehicle according to an embodiment of the invention;
fig. 7 is a schematic structural diagram of a fuse performance test platform based on an antagonistic neural network in a second embodiment of the present invention;
fig. 8 is a schematic structural diagram of a fuse performance test platform based on an antagonistic neural network in a third embodiment of the present invention.
Reference numerals:
the system comprises an information acquisition module 1, a scene reconstruction module 2, a circuit model 3, a test module 4, a data storage module 5 and a data output module 6.
Detailed Description
The above and further technical features and advantages of the present invention are described in more detail below with reference to the accompanying drawings.
In the description of the present invention, it should be understood that the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In the description of the present invention, the meaning of "plurality" is at least two unless explicitly defined otherwise.
In the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," "secured," and the like are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally formed; may be mechanically connected, may be electrically connected or may be in communication with each other; either directly or indirectly, through intermediaries, or both, may be in communication with each other or in interaction with each other, unless expressly defined otherwise. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention.
In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.
In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.
In order to illustrate the technical scheme of the invention, the following description is made by specific examples.
Example 1
As shown in fig. 1, the invention provides a fuse performance test platform based on an antagonistic neural network, which comprises an information acquisition module 1, a scene reconstruction module 2 and a circuit model 3.
The information acquisition module 1 is used for acquiring real circuit operation parameters. Preferably, the actual circuit operating parameters include current, voltage and ambient temperature data at the sampling points. In this embodiment, it is further preferable that the sampling point is a key node in the electric automobile circuit. The information acquisition module 1 comprises a current collector and a voltage collector, the current collector and the voltage collector are connected into a plurality of key nodes in an electric automobile circuit, and the electric automobile is enabled to run under different working conditions, so that corresponding circuit current and voltage data are obtained.
Different working conditions include: european endurance test conditions (NEDC conditions), global unified light vehicle test cycle conditions (WLTC conditions), chinese light vehicle driving conditions (CLTC conditions) and heavy commercial vehicle driving conditions (CHTC conditions), us environmental protection agency conditions (EPA conditions), and endurance conditions for each subdivision class. It also includes: typical dangerous working conditions of the electric automobile such as a power system short-circuit working condition, a collision working condition, a rollover working condition, a wading and heavy rain working condition, a battery overcharge and overdischarge working condition and the like. As shown in fig. 3, the running speed curves of the electric automobile under different working conditions are different, and correspondingly, different current changes exist in the electric automobile circuit. The ordinate of fig. 3 is the running speed, and the abscissa is the time.
The scene reconstruction module 2 comprises an antagonistic neural network. The scene reconstruction module 2 generates synthetic data that resemble real data by optimizing the antagonistic neural network. In other words, the antagonistic neural network is used to generate synthetic data that is identical to the real data. The information acquisition module 1 is connected with the scene reconstruction module 2, and the real data acquired by the information acquisition module 1 are used for training against the neural network. As shown in fig. 2, the antagonistic neural network includes a generator network and a discriminator network. The generator network is used to generate the composite data. The discriminator network is used to discriminate between real data and synthetic data.
The circuit model 3 is connected to the scene reconstruction module 2. The circuit model 3 is used to provide input data to the scene reconstruction module 2. Preferably, the circuit model 3 is connected to the generator network. The input data calculated by the circuit model 3 will be passed to the generator network, which generates the composite data. The discriminator network is connected with the generator network and the information acquisition module 1 respectively. The arbiter network receives both the real data from the information acquisition module 1 and the composite data from the generator network.
In this embodiment, the circuit model 3 is preferably an electric automobile circuit model, and the circuit model 3 is based on a circuit model established by an actual electric automobile circuit, and each parameter is measured on a specific electric automobile or obtained from a factory according to a product manual. The circuit model is obtained according to actual parameters of the components such as the battery, the motor, the electric control system and the like.
The scene reconstruction module 2 is provided with an antagonistic neural network, and the data calculated by the circuit model 3 and the actual circuit current and voltage data acquired from the information acquisition module 1 are used for training the antagonistic neural network. The goal of the antagonistic generator network in the neural network is to produce as realistic as possible dummy data. The objective of the arbiter network in the antagonistic neural network is to distinguish as much as possible between spurious data. The generator and the discriminator are trained iteratively, the more the training times are, the more the data generated by the generator is close to the real data, and the more the discriminator has strong false judging capability.
The generator network and the arbiter network are trained using standard cross entropy loss functions. First, binary target variables are defined as follows, whenWhen the corresponding data is real data; when->And when the corresponding data is synthesized data. The arbiter network has a single-value output unit and a logical sigmoid activation function, the output of the arbiter network representing the input data vector +.>Is the real data (+)>) Is a probability of (2). The output of the arbiter network is:
wherein the data vectorRepresenting the input to the arbiter network. />Is a discriminator network parameter. />Then represent data vector +.>Probability of being real data.
The cross entropy loss function for training the criterion of the arbiter network is:
wherein,,/>is the identifier network +.>Is used for the output data of the (a),is a function of the normalized processing according to the total number N of data points. />Generator network parameters. />Is a discriminator network parameter.
The training data set for training against the neural network comprises real data from the information acquisition module 1And by the generator network->Output data given, wherein ∈>Is a potential spatial distribution->Is a random data of the same. Due to->Corresponding to real data->Corresponding to the composite data, the error function against the neural network may finally be written in the form,
number of real data points inEqual to the number of synthetic data points +.>。/>Synthetic data representing a discriminator network for input from a generator network>An output of (2) representing the composite dataProbability of being judged as real data. />Representing the discriminator network for real data +.>An output of (2) representing real data +.>Probability of being judged as real data.
The training mode of the generator network and the discriminator network is an end-to-end mode of random gradient descent, and the gradient is evaluated by adopting counter propagation. In training against neural networks, the following is relevantThe error of (2) is to be minimal and about +.>The error of (2) is maximized. About->The maximization of the error of (c) may be achieved by standard gradient-based methods, but the sign of the gradient is reversed, so that the update of the parameters is as follows:
in the method, in the process of the invention,generator network parameters. />Is the error of the network parameters of the discriminator. />Representing errors at the nth data point or small batches of data points.
Note that since the arbiter uses a decreasing error rate, and the generator uses an increasing error rate for training,andwith different signs. In actual operation, the training process continuously updates the parameters of the generator network and the parameters of the arbiter network, the two are performed in turn, each time updating uses a small batch of data points, a gradient descent step is executed, and then a series of new synthesized data is generated. If the generator successfully finds a perfect method, the arbiter network will not be able to distinguish between real data and synthetic data, so the arbiter network will always output a value of 0.5. When the antagonistic neural network training is completed, the arbiter network is removed, and the generator network can then synthesize new data in the data space by sampling the potential space and passing the sample through the already trained network. It can be seen that the generator network and the arbiter network have infinite freedom, so that a fully optimized antagonistic neural network will generate synthetic data that is identical to real data.
The test platform for testing the performance of the fuse based on the antagonistic neural network in the embodiment further comprises a test module 4, wherein the test module 4 is connected with the scene reconstruction module 2. The scene reconstruction module 2 outputs the synthesized data identical to the real data to the test module 4. The test module 4 performs a performance test of the fuse based on the input data from the scene reconstruction module 2. After the test module 4 obtains the voltage and current data after scene reconstruction, the test module 4 controls the voltage and current of the test circuit by the solid-state switch according to the data, and tests various performances of the electric automobile fuse. The solid-state switch controls the voltage and the current of the test circuit, and the solid-state switch is used for setting the test circuit by connecting the power supply, the resistor and the inductor. Testing various performance of an electric automobile fuse includes testing various performance parameters such as operating temperature, rated voltage, rated current, short circuit interception capability, fusing characteristics, joule integral and the like of the fuse in different simulation scenes.
The invention provides a fuse performance test platform based on an antagonistic neural network, which is suitable for testing an electric automobile circuit fuse, and is used for realizing scene reconstruction based on the antagonistic neural network, so as to realize performance test of the electric automobile fuse, and the basic principle of the test platform technical scheme is as follows: training the countermeasure neural network by adopting real data and circuit model data, and generating data by using the trained neural network, so that electric vehicle current and voltage data close to various real scenes can be obtained for testing the electric vehicle fuse. The test method is very close to physical reality, can provide more accurate and reliable results than the prior art, and has more reference significance. As shown in fig. 4, 5 and 6, the technical scheme in the prior art can only measure the fusing time of the fuse at specific current, power loss and fusing time correction of different environment temperatures, and the technical scheme of the invention can obtain the accurate fusing time and other fuse performance parameters of the fuse under specific scenes such as collision working conditions.
The invention provides an application example of a fuse performance test platform based on an antagonistic neural network, which comprises the following steps: the fuse performance test platform is used for testing the power loss of the fuse of the electric automobile under the highway cruising working condition, wherein the fuse is positioned at the battery pack interface. And establishing a circuit model according to an actual circuit of the electric automobile, obtaining current and voltage data at a battery pack interface in the circuit model, setting an output scene of the scene reconstruction module 2 as expressway cruising, and generating simulation data by the scene reconstruction module 2. And finally, the test module tests the performance of the electric automobile fuse according to the generated simulation data.
The test platform for the fuse performance based on the antagonistic neural network provided by the invention realizes reconstruction of the complex working scene of the electric automobile by adopting the antagonistic neural network, can be applied to the test of the fuse performance of the electric automobile, and can meet the test requirement of a special high-performance fuse. The test platform is not limited to the performance test of the electric automobile fuse, and can also be used for testing fuses of other types of circuits. The fuse performance test platform based on the countermeasure to neural network can be applied to fuse performance tests of various scenes, meets the diversified demands of products, realizes a test means close to reality, has practicability compared with the prior art, and can be popularized and applied on a large scale.
Example two
Fig. 7 is a schematic structural diagram of a fuse performance test platform based on an antagonistic neural network in the present embodiment. As shown in fig. 7, the present embodiment is different from the embodiment in that the present embodiment provides a fuse performance test platform based on an antagonistic neural network, and further includes a data storage module 5 for storing data obtained from the information acquisition module 1. The actual circuit operating parameters can be stored by means of the data storage module 5. The real circuit operating parameters are used to train the anti-neural network. Current, voltage and ambient temperature data in real electric vehicle lines may be stored in the data storage module 5. The data storage module 5 is also connected to the scene reconstruction module 2. The scene reconstruction module 2 is able to invoke the stored data in the data storage module 5. The scene reconstruction module 2 is indirectly connected with the information acquisition module 1 through the data storage module 5.
Example III
Fig. 8 is a schematic structural diagram of a fuse performance test platform based on an antagonistic neural network in the present embodiment. As shown in fig. 8, the difference between the present embodiment and the first and second embodiments is that the present embodiment provides a fuse performance test platform based on an antagonistic neural network, which further includes a data output module 6, where the data output module 6 is connected to the scene reconstruction module 2 and the test module 4 respectively. The data output module 6 is configured to output the generated data of the scene reconstruction module according to the requirement, that is, the simulation data of the driving, fault, and other scenes of the electric vehicle. The data output module directly uses the trained generation network in the antagonistic neural network to generate data.
The above is only a specific embodiment of the present invention, but the technical features of the present invention are not limited thereto. Any simple changes, equivalent substitutions or modifications made on the basis of the present invention to solve the substantially same technical problems and achieve the substantially same technical effects are encompassed within the scope of the present invention.
Claims (10)
1. The fuse performance test platform based on the antagonistic neural network is characterized by comprising an information acquisition module, a scene reconstruction module, a circuit model and a test module, wherein the scene reconstruction module is respectively connected with the information acquisition module, the circuit model and the test module, and the information acquisition module is used for acquiring real circuit operation parameters; the scene reconstruction module comprises an antagonistic neural network, and generates synthesized data of the real-like data by optimizing the antagonistic neural network; the circuit model is used for providing input data to the scene reconstruction module; the test module is used for executing performance test of the fuse.
2. The fuse performance test platform of claim 1, wherein the countermeasure neural network comprises a generator network for generating synthetic data that resembles real data and a discriminator network for discriminating whether the data is real data or synthetic data, the generator network and the discriminator network being capable of iterative training.
3. The antagonistic neural network based fuse performance test platform of claim 2, wherein the cross entropy loss function of the criteria used to train the arbiter network is:
wherein,,/>is the identifier network +.>Output data of->Generator network parameters, < >>Is a discriminator network parameter.
4. A fuse performance test platform based on an antagonistic neural network according to claim 3, characterized in that the error function of the antagonistic neural network is:
in the method, in the process of the invention,representing the number of real data points +.>Representing the number of synthetic data points, +.>Representing the arbiter network for synthetic data input from the generator network>Output of->Representing the arbiter network for real data +.>Is provided.
5. A fuse performance test platform based on an antagonistic neural network according to any of claims 2-4, characterized in that the training pattern of the generator network and the arbiter network is a random gradient descent end-to-end pattern, the gradients being evaluated using counter-propagation.
6. The antagonistic neural network based fuse performance test platform of claim 1, further comprising a data storage module coupled to the information acquisition module and the scene reconstruction module, respectively, the data storage module configured to store data obtained from the information acquisition module.
7. The test platform of claim 1, further comprising a data output module, wherein the data output module is respectively connected with the scene reconstruction module and the test module, the data output module is used for outputting data generated by the scene reconstruction module to the test module, and output data of the data output module are voltage data and current data.
8. The test platform of claim 1, wherein the real circuit operation parameters comprise current, voltage and environmental temperature data of sampling points, the sampling points are nodes in an electric automobile circuit, the information acquisition module comprises a current collector and a voltage collector, and the information acquisition module is used for acquiring the data of the sampling points when the electric automobile operates under different working conditions.
9. The antagonistic neural network based fuse performance test platform of claim 8, wherein the operating conditions include a european endurance test operating condition, a global unified light vehicle test cycle operating condition, a chinese light vehicle driving operating condition and a heavy commercial vehicle driving operating condition, a us environmental agency operating condition, a power system short circuit operating condition, a collision operating condition, a rollover operating condition, a wading and heavy rain operating condition, and a battery overcharge/discharge operating condition.
10. The antagonistic neural network based testing platform of claim 1, wherein the testing module is configured to test the operating temperature, voltage rating, current rating, short circuit shut-off capability, fusing characteristics, and joule integral of the fuse, the testing module comprising a solid state switch and a testing circuit, the solid state switch being configured to control the access of power, resistance, and inductance to the testing circuit.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410076051.6A CN117607756B (en) | 2024-01-18 | 2024-01-18 | Fuse performance test platform based on antagonistic neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410076051.6A CN117607756B (en) | 2024-01-18 | 2024-01-18 | Fuse performance test platform based on antagonistic neural network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117607756A true CN117607756A (en) | 2024-02-27 |
CN117607756B CN117607756B (en) | 2024-04-05 |
Family
ID=89950215
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410076051.6A Active CN117607756B (en) | 2024-01-18 | 2024-01-18 | Fuse performance test platform based on antagonistic neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117607756B (en) |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101776720A (en) * | 2010-01-04 | 2010-07-14 | 苏州热工研究院有限公司 | Multiple-factor detection system for aging condition of fuse and assessment method for aging condition |
CN101915887A (en) * | 2010-07-05 | 2010-12-15 | 苏州热工研究院有限公司 | Fuse life characteristic data test and evaluation system and evaluation method thereof |
CN205301492U (en) * | 2016-01-21 | 2016-06-08 | 山东理工大学 | Fuse test system for car |
CN108696331A (en) * | 2018-03-19 | 2018-10-23 | 西安电子科技大学 | A kind of signal reconfiguring method based on generation confrontation network |
CN108845252A (en) * | 2018-08-23 | 2018-11-20 | 浙江方圆电气设备检测有限公司 | A kind of automobile fuse performance detection device |
CN214703888U (en) * | 2021-03-15 | 2021-11-12 | 银隆新能源股份有限公司 | Electric automobile high voltage direct current fuse life-span automatic checkout device |
WO2022160900A1 (en) * | 2021-01-29 | 2022-08-04 | 华为技术有限公司 | Test environment construction method and device |
CN115113086A (en) * | 2022-05-18 | 2022-09-27 | 中国第一汽车股份有限公司 | Vehicle power distribution test system and method based on vaHil environment simulation |
CN115236272A (en) * | 2022-06-07 | 2022-10-25 | 辽宁警察学院 | Gas sensor fault diagnosis method and device under multi-working condition and storage medium |
CN115902449A (en) * | 2022-09-19 | 2023-04-04 | 威凯检测技术有限公司 | Intelligent integrated testing equipment for thermal fuse link based on machine vision and neural network |
CN116106773A (en) * | 2022-12-15 | 2023-05-12 | 重庆邮电大学 | Mixed model lithium battery health state monitoring method based on data enhancement |
CN116454823A (en) * | 2022-01-06 | 2023-07-18 | 宇通客车股份有限公司 | Overload fuse matching protection method for vehicle battery system |
CN116819204A (en) * | 2023-06-27 | 2023-09-29 | 福建宁德核电有限公司 | Nuclear power plant high-voltage current-limiting fuse tolerance test method |
CN116930769A (en) * | 2023-07-24 | 2023-10-24 | 国网山东省电力公司营销服务中心(计量中心) | Lithium battery modeling method based on bidirectional generation type antagonistic neural network |
-
2024
- 2024-01-18 CN CN202410076051.6A patent/CN117607756B/en active Active
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101776720A (en) * | 2010-01-04 | 2010-07-14 | 苏州热工研究院有限公司 | Multiple-factor detection system for aging condition of fuse and assessment method for aging condition |
CN101915887A (en) * | 2010-07-05 | 2010-12-15 | 苏州热工研究院有限公司 | Fuse life characteristic data test and evaluation system and evaluation method thereof |
CN205301492U (en) * | 2016-01-21 | 2016-06-08 | 山东理工大学 | Fuse test system for car |
CN108696331A (en) * | 2018-03-19 | 2018-10-23 | 西安电子科技大学 | A kind of signal reconfiguring method based on generation confrontation network |
CN108845252A (en) * | 2018-08-23 | 2018-11-20 | 浙江方圆电气设备检测有限公司 | A kind of automobile fuse performance detection device |
WO2022160900A1 (en) * | 2021-01-29 | 2022-08-04 | 华为技术有限公司 | Test environment construction method and device |
CN214703888U (en) * | 2021-03-15 | 2021-11-12 | 银隆新能源股份有限公司 | Electric automobile high voltage direct current fuse life-span automatic checkout device |
CN116454823A (en) * | 2022-01-06 | 2023-07-18 | 宇通客车股份有限公司 | Overload fuse matching protection method for vehicle battery system |
CN115113086A (en) * | 2022-05-18 | 2022-09-27 | 中国第一汽车股份有限公司 | Vehicle power distribution test system and method based on vaHil environment simulation |
CN115236272A (en) * | 2022-06-07 | 2022-10-25 | 辽宁警察学院 | Gas sensor fault diagnosis method and device under multi-working condition and storage medium |
CN115902449A (en) * | 2022-09-19 | 2023-04-04 | 威凯检测技术有限公司 | Intelligent integrated testing equipment for thermal fuse link based on machine vision and neural network |
CN116106773A (en) * | 2022-12-15 | 2023-05-12 | 重庆邮电大学 | Mixed model lithium battery health state monitoring method based on data enhancement |
CN116819204A (en) * | 2023-06-27 | 2023-09-29 | 福建宁德核电有限公司 | Nuclear power plant high-voltage current-limiting fuse tolerance test method |
CN116930769A (en) * | 2023-07-24 | 2023-10-24 | 国网山东省电力公司营销服务中心(计量中心) | Lithium battery modeling method based on bidirectional generation type antagonistic neural network |
Also Published As
Publication number | Publication date |
---|---|
CN117607756B (en) | 2024-04-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Mesbahi et al. | Dynamical modeling of Li-ion batteries for electric vehicle applications based on hybrid Particle Swarm–Nelder–Mead (PSO–NM) optimization algorithm | |
Vaideeswaran et al. | Battery management systems for electric vehicles using lithium ion batteries | |
Morello et al. | Hardware-in-the-loop platform for assessing battery state estimators in electric vehicles | |
Rizoug et al. | Development of new improved energy management strategies for electric vehicle battery/supercapacitor hybrid energy storage system | |
CN109507981B (en) | Vehicle testing method and device and machine-readable storage medium | |
Yao et al. | Sliding mode-based H-infinity filter for SOC estimation of lithium-ion batteries | |
Huo et al. | Research on parameter identification and state of charge estimation of improved equivalent circuit model of Li‐ion battery based on temperature effects for battery thermal management | |
Masih-Tehrani et al. | A novel power distribution system employing state of available power estimation for a hybrid energy storage system | |
Barcellona et al. | A simple battery model for EV range prediction: Theory and experimental validation | |
Bui et al. | An advanced hardware-in-the-loop battery simulation platform for the experimental testing of battery management system | |
Khalfi et al. | An electric circuit model for a lithium-ion battery cell based on automotive drive cycles measurements | |
Rizoug et al. | Aging of high power Li‐ion cells during real use of electric vehicles | |
CN113608140A (en) | Battery fault diagnosis method and system | |
Song et al. | Verification of battery system model for environmentally friendly vehicles using a battery hardware‐in‐the‐loop simulation | |
Kai et al. | Development of Accurate Lithium‐Ion Battery Model Based on Adaptive Random Disturbance PSO Algorithm | |
CN117607756B (en) | Fuse performance test platform based on antagonistic neural network | |
Barreras et al. | Functional analysis of Battery Management Systems using multi-cell HIL simulator | |
Shet et al. | Design of a thermal management system for a battery pack in an electric vehicle using Dymola | |
Yang et al. | Advanced Battery Management System for Electric Vehicles | |
Lehner et al. | Disparity in initial and lifetime parameters of lithium‐ion cells | |
Jia et al. | Temperature enhanced early detection of internal short circuits in lithium-ion batteries using an extended Kalman filter | |
Chen et al. | Charging strategy design of lithium-ion batteries for energy loss minimization based on minimum principle | |
JP2002142370A (en) | Battery pack | |
Bian et al. | An improved particle filter method to estimate state of health of lithium-ion battery | |
Hametner et al. | Local model network based dynamic battery cell model identification |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |