Dual active bridge converter open-circuit fault prediction method and device based on deep learning

Abstract

The invention discloses a deep learning-based open-circuit fault prediction method and device for a dual-active bridge converter, and the method comprises the steps: building an equivalent circuit simulation model of the dual-active bridge converter, and building an open-circuit fault prediction training data set based on an open-circuit detection signal of the equivalent circuit simulation model of the dual-active bridge converter and a corresponding open-circuit fault result; constructing a dual-active bridge converter open-circuit fault prediction model, and training the dual-active bridge converter open-circuit fault prediction model through the open-circuit fault prediction training data set; and acquiring a real-time open-circuit detection signal of the equivalent circuit simulation model of the dual-active bridge converter, inputting the trained open-circuit fault prediction model of the dual-active bridge converter, and outputting a result which is a corresponding open-circuit fault prediction result of the dual-active bridge converter. According to the invention, based on the deep learning algorithm, the open circuit abnormity model is established for the dual-active bridge converter for analysis, the reliability of the mining electrical system power supply conversion equipment can be improved, and the frequency of abnormity in the use process is reduced.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a method and device for predicting open-circuit faults of a dual active bridge converter based on deep learning. Background Art

[0002] Dual active bridge (DAB) converters feature electrical isolation and bidirectional energy flow, making them widely used in electric vehicles, energy storage systems, distributed power generation, and DC transmission. They perform power conversion and energy transfer, and are often used in electrical equipment as power conversion devices for electrical systems. Reliability is a critical issue for power conversion devices used in mining electrical systems, as failures can cause significant economic losses and personal injury. IGBTs are among the most vulnerable components in mining electrical system power conversion equipment, and IGBT short-circuit failures can cause devastating damage. Summary of the Invention

[0003] The present invention aims to improve the reliability of power conversion equipment in mining electrical systems and save the time required for abnormality analysis by using a deep learning algorithm to quickly analyze open circuits in power device modules in circuits.

[0004] To this end, the first object of the present invention is to propose a method for predicting open-circuit faults of a dual active bridge converter based on deep learning, comprising:

[0005] Build a dual active bridge converter equivalent circuit simulation model, and create an open circuit fault prediction training dataset based on the open circuit detection signal and corresponding open circuit fault results of the dual active bridge converter equivalent circuit simulation model;

[0006] Based on the deep trust network model, a dual active bridge converter open circuit fault prediction model is constructed, and the dual active bridge converter open circuit fault prediction model is trained using the open circuit fault prediction training dataset.

[0007] The real-time open-circuit detection signal of the dual-active-bridge converter equivalent circuit simulation model is obtained, and the trained dual-active-bridge converter open-circuit fault prediction model is input. The output result is the corresponding dual-active-bridge converter open-circuit fault prediction result.

[0008] The equivalent circuit simulation model of the dual-active bridge converter includes two symmetrical full-bridge circuits on the primary and secondary sides, which are controlled by a preset phase-shift strategy. The preset phase-shift control strategies include single phase shift (SPS), extended phase shift (EPS), and triple phase shift (TPS).

[0009] Among them, when the preset phase shift control strategy is single phase shift control, when the dual active bridge converter works normally, the conduction angle of the insulated gate bipolar transistor IGBT on each bridge arm of the primary and secondary full bridge circuit is 180°, complementary conduction, and the IGBTs in the diagonal position are turned on or off at the same time, and the primary and secondary full bridge circuits respectively output a square wave voltage U with a duty cycle of 50%. AB and U CD By adjusting U AB with U CD The size of the phase shift duty cycle between the two can change the size and direction of the transmission power; when the phase shift duty cycle is greater than 0, that is, U AB The phase is ahead of U CD When the phase shift duty cycle increases, the power will be transmitted in the forward direction; otherwise, the power will be transmitted in the reverse direction. The transmission power of the dual active bridge converter increases with the increase of the phase shift duty cycle.

[0010] Among them, the open circuit detection signal of the dual active bridge converter equivalent circuit simulation model is the transformer leakage inductance current. Due to the symmetry of the dual active bridge converter, when the power switching devices in the diagonal direction of the dual active bridge converter equivalent circuit simulation model are abnormal, the transformer leakage inductance current i L The waveforms show the same behavior. If the switching device anomalies in the diagonal direction are classified into the same open-circuit fault result category, then the dual active bridge converter equivalent circuit simulation model has a total of five open-circuit fault result categories.

[0011] Among them, in the dual active bridge converter equivalent circuit simulation model, when the dual active bridge converter is not open-circuited, according to Kirchhoff's voltage law, formula (1) is obtained:

[0012] U AB =kU CD +u L +ri L (1)

[0013] in: k is the transformer ratio;

[0014] Define the ideal switching functions QA, QB, QC, and QD as:

[0015]

[0016] Among them, S1~S8 are IGBTs; D1~D8 are diodes;

[0017] The output voltage of the primary and secondary full-bridge circuit is represented by the converter's input and output voltages U1 and U2:

[0018]

[0019] The mathematical model formula of the equivalent circuit of the dual active bridge converter is expressed as:

[0020]

[0021] Transformer leakage inductance current i L The formula is:

[0022]

[0023] Among them, the open circuit fault prediction model of the dual active bridge converter adopts the deep belief network DBN, which includes:

[0024] Multi-layer restricted Boltzmann machine (RBM), used for unsupervised layer-by-layer pre-training and feature extraction;

[0025] The feedforward back propagation network is used to add a classifier on the top layer of the deep belief network DBN and fine-tune the parameters of the multi-layer restricted Boltzmann machine RBM;

[0026] The open circuit fault prediction model of the dual active bridge converter is a deep belief network DBN with a two-layer RBM structure.

[0027] Among them, the step of constructing the dual active bridge converter open circuit fault prediction model also includes the step of optimizing the deep learning network model DBN in the dual active bridge converter open circuit fault prediction model using a particle swarm algorithm, an ant colony algorithm or a genetic algorithm.

[0028] A second object of the present invention is to provide a device for predicting open-circuit faults of a dual active bridge converter based on deep learning, comprising:

[0029] A circuit simulation model construction module is used to construct an equivalent circuit simulation model of a dual active bridge converter and create an open circuit fault prediction training data set based on the open circuit detection signal and corresponding open circuit fault results of the dual active bridge converter equivalent circuit simulation model;

[0030] A prediction model building module is used to build a dual active bridge converter open circuit fault prediction model based on a deep trust network model, and train the dual active bridge converter open circuit fault prediction model using an open circuit fault prediction training dataset;

[0031] The prediction module is used to obtain the real-time open circuit detection signal of the dual active bridge converter equivalent circuit simulation model, input the trained dual active bridge converter open circuit fault prediction model, and output the corresponding dual active bridge converter open circuit fault prediction result.

[0032] The third object of the present invention is to provide an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute each step in the method of the aforementioned technical solution.

[0033] A fourth object of the present invention is to provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute each step in the method according to the aforementioned technical solution.

[0034] Different from the prior art, the present invention's deep learning-based dual-active bridge converter open-circuit fault prediction method, by building a dual-active bridge converter equivalent circuit simulation model, based on the open-circuit detection signal of the dual-active bridge converter equivalent circuit simulation model and the corresponding open-circuit fault result, creates an open-circuit fault prediction training data set; based on a deep trust network model, builds a dual-active bridge converter open-circuit fault prediction model, and trains the dual-active bridge converter open-circuit fault prediction model using the open-circuit fault prediction training data set; obtains the real-time open-circuit detection signal of the dual-active bridge converter equivalent circuit simulation model, inputs the trained dual-active bridge converter open-circuit fault prediction model, and outputs the corresponding dual-active bridge converter open-circuit fault prediction result. Based on a deep learning algorithm, the present invention conducts open-circuit anomaly modeling and analysis on the dual-active bridge converter, which can improve the reliability of power conversion equipment in mine electrical systems and reduce the frequency of anomalies during use. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The present invention and / or additional aspects and advantages will become apparent and readily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0036] Figure 1 This is a flow chart of a deep learning-based open-circuit fault prediction method for a dual active bridge converter provided by the present invention.

[0037] Figure 2 This is a schematic diagram of the topological structure of the equivalent circuit simulation model of the dual active bridge converter in the deep learning-based open circuit fault prediction method of the dual active bridge converter provided by the present invention.

[0038] Figure 3 This is a schematic diagram of the equivalent circuit structure of the dual active bridge converter in the dual active bridge converter open circuit fault prediction method based on deep learning provided by the present invention when no abnormality occurs in the equivalent circuit simulation model.

[0039] Figure 4 This is a structural diagram of a deep belief network in a deep learning-based open-circuit fault prediction method for a dual active bridge converter provided by the present invention.

[0040] Figure 5 This is a structural schematic diagram of a deep learning-based dual active bridge converter open circuit fault prediction device provided by the present invention.

[0041] Figure 6 It is a structural schematic diagram of a non-transitory computer-readable storage medium storing computer instructions provided by the present invention. DETAILED DESCRIPTION

[0042] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention, but are not to be construed as limiting the present invention.

[0043] like Figure 1 As shown in FIG, a method for predicting an open-circuit fault of a dual active bridge converter based on deep learning is provided in an embodiment of the present invention, comprising:

[0044] S110: Constructing a dual active bridge converter equivalent circuit simulation model, and creating an open circuit fault prediction training data set based on an open circuit detection signal and a corresponding open circuit fault result of the dual active bridge converter equivalent circuit simulation model.

[0045] The circuit topology of the dual active bridge converter equivalent circuit simulation model constructed by the present invention is as follows: Figure 2 shown.

[0046] exist Figure 2 In the figure, U1 is the input voltage of the converter; C1 is the filter capacitor on the input side of the converter; L is the sum of the transformer leakage inductance and the external series inductance; i L is the inductor current; k is the transformer ratio; C2 is the filter capacitor on the output side of the converter; R L is the equivalent resistance of the converter output side; U2 is the output voltage of the converter; U AB and U CD is the output voltage of the original secondary side H-bridge; S1~S8 are IGBTs; D1~D8 are diodes.

[0047] pass Figure 2 As can be seen, the dual active bridge converter is structurally symmetrical, thus enabling bidirectional power flow. The dual active bridge converter consists of two symmetrical full-bridge circuits on the primary and secondary sides, controlled by a preset phase-shift control strategy.

[0048] Specifically, the preset phase shift control strategies include single phase shift (SPS), extended phase shift (EPS), and triple phase shift (TPS) control.

[0049] Phase-shift control is simpler and easier to implement than pulse-width modulation control. Therefore, most dual-active bridge converters operate under phase-shift control. Taking single-phase-shift control as an example, when the dual-active bridge converter is operating normally, the insulated gate bipolar transistors (IGBTs) on each arm of the primary and secondary full-bridge circuits are 180° apart and complementary, and the IGBTs in diagonal positions are turned on or off at the same time. The primary and secondary full-bridge circuits each output a square wave voltage U with a duty cycle of 50%. AB and U CD By adjusting U AB with U CD The size of the phase shift duty cycle between the two can change the size and direction of the transmission power; when the phase shift duty cycle is greater than 0, that is, U AB The phase is ahead of U CD When the phase shift duty cycle increases, the power will be transmitted in the forward direction; otherwise, the power will be transmitted in the reverse direction. The transmission power of the dual active bridge converter increases with the increase of the phase shift duty cycle.

[0050] The open circuit detection signal of the dual active bridge converter equivalent circuit simulation model is the transformer leakage inductance current. Due to the symmetry of the dual active bridge converter, when the power switching devices in the diagonal direction of the dual active bridge converter equivalent circuit simulation model are abnormal, the transformer leakage inductance current i L The waveforms appear identical, and if diagonally opposite switching device anomalies are categorized as the same open-circuit fault result category, the dual-active bridge converter equivalent circuit simulation model has a total of five open-circuit fault result categories. Specifically, S1 and S4, S2 and S3, D1 and D4, and D2 and D3 all fall into the same anomaly category, resulting in a total of five open-circuit categories, including the normal case.

[0051] In the equivalent circuit simulation model of the dual active bridge converter, when the dual active bridge converter is not open-circuited, according to Kirchhoff's voltage law, formula (1) is obtained:

[0052] U AB =kU CD +u L +ri L (1)

[0053] in: k is the transformer ratio;

[0054] Define the ideal switching functions QA, QB, QC, and QD as:

[0055]

[0056] Among them, S1~S8 are IGBTs; D1~D8 are diodes;

[0057] The output voltage of the primary and secondary full-bridge circuit is represented by the converter's input and output voltages U1 and U2:

[0058]

[0059] The mathematical model formula of the equivalent circuit of the dual active bridge converter is expressed as:

[0060]

[0061] During the actual operation of the dual active bridge converter, both the power module and the diode may open circuit, and different open circuit anomalies may have different characteristics and lead to different consequences. Figure 3 It can be seen that the circuit topology of the dual active bridge converter is a structure in which the primary and secondary full-bridge circuits are symmetrical to each other. Therefore, the IGBTs and diodes of the primary and secondary full-bridge circuits have similar situations and characteristics when an open circuit abnormality occurs.

[0062] The study focuses on the situation when the IGBT and diode of the primary full-bridge circuit of the dual active bridge converter are open-circuited. When the circuit is open, S1 is in a non-conducting state. In this case, the time when S1 is open is between t0 and t1 or t3 and t6. The open circuit phenomenon does not begin until S1 is about to be turned on. At this time, i L The positive direction increases, and since S1 is open, the circuit topology changes, i L Will flow through S4, D 2 .

[0063] When an IGBT in a dual-active-bridge converter opens, some or even all variables in the converter will behave differently than when no abnormality occurs. The basic concept behind the dual-active-bridge converter open-circuit diagnosis method based on inductor current is to analyze the system inductor current state residual, generated by comparing the actual system with the established dual-active-bridge converter mathematical model. When the dual-active-bridge converter is operating normally, the system inductor current state residual is approximately zero. However, when the dual-active-bridge converter opens, the magnitude of the system inductor current state residual increases significantly, deviating from zero. Therefore, detecting the system inductor current state residual can be used to analyze open-circuit conditions.

[0064] The principle of the open circuit judgment method of the dual active bridge converter based on inductor current is known from the mathematical model of formula (4). The state of the converter is determined by the ideal switching function, input and output voltages and inductor current. Rewriting formula (4) can obtain the transformer leakage inductance current iL The formula is:

[0065]

[0066] The transformer leakage inductance current iL is selected as the open circuit detection signal. Due to the symmetry of the converter, when an abnormality occurs in the diagonally opposite power switching device, the waveform of the converter's transformer leakage inductance current iL is the same. Therefore, the abnormalities of the diagonally opposite switching devices are classified into the same abnormality category, that is, S1 and S4, S2 and S3, D1 and D4, and D2 and D3 are respectively classified into the same abnormality category. The circuit has a total of five abnormality categories, including the normal situation.

[0067] Based on the open circuit detection signal and the corresponding open circuit fault results of the dual active bridge converter equivalent circuit simulation model, an open circuit fault prediction training data set is created. Through the dual active bridge converter equivalent circuit simulation model, different open circuit detection signals are input into the simulation model to obtain the corresponding types of abnormal categories. The open circuit detection signal and the corresponding types of abnormal categories are used as the training set of the model to proceed to the next step.

[0068] S120: Based on the deep trust network model, construct a dual active bridge converter open circuit fault prediction model, and train the dual active bridge converter open circuit fault prediction model using an open circuit fault prediction training data set.

[0069] like Figure 4 As shown in Figure 1, the open circuit fault prediction model of the dual active bridge converter adopts the deep belief network DBN, which includes:

[0070] Multi-layer restricted Boltzmann machine (RBM), used for unsupervised layer-by-layer pre-training and feature extraction;

[0071] The feedforward back propagation network is used to add a classifier on the top layer of the deep belief network DBN and fine-tune the parameters of the multi-layer restricted Boltzmann machine RBM;

[0072] The open circuit fault prediction model of the dual active bridge converter is a deep belief network DBN with a two-layer RBM structure.

[0073] The model's input layer takes in the open-circuit detection signal from the training set, and the output layer outputs the five abnormal modes of the converter. In the first-layer RBM, the input layer is the visual layer v, consisting of n visual units, and the output layer is the hidden layer h, consisting of m hidden units, with v∈{0,1}n and h∈{0,1}, where m represents the inactive state and 1 represents the active state. The input layer of the second-layer RBM is the output layer of the first-layer RBM. RBMs are energy-based models, and the energy function E(v,h|θ) is defined as follows:

[0074]

[0075] Where, θ=[a i ,β j ,w ij ] is the parameter of each RBM, a i and β j It is v i and h j Bias, w ij Is the connection v i and h j The weight of .

[0076] In the process of constructing the open circuit fault prediction model of the dual active bridge converter, the particle swarm algorithm, ant colony algorithm or genetic algorithm is used to optimize the deep learning network model DBN in the open circuit fault prediction model of the dual active bridge converter.

[0077] S130: obtaining a real-time open circuit detection signal of the dual active bridge converter equivalent circuit simulation model, inputting the trained dual active bridge converter open circuit fault prediction model, and outputting a corresponding dual active bridge converter open circuit fault prediction result.

[0078] like Figure 5 As shown, the present invention provides a dual active bridge converter open circuit fault prediction device 500 based on deep learning, comprising:

[0079] A circuit simulation model construction module 510 is used to construct an equivalent circuit simulation model of a dual active bridge converter and create an open circuit fault prediction training data set based on the open circuit detection signal and the corresponding open circuit fault result of the dual active bridge converter equivalent circuit simulation model;

[0080] A prediction model building module 520 is used to build a dual active bridge converter open circuit fault prediction model based on a deep belief network model, and train the dual active bridge converter open circuit fault prediction model using an open circuit fault prediction training data set;

[0081] The prediction module 530 is used to obtain the real-time open circuit detection signal of the dual active bridge converter equivalent circuit simulation model, input the trained dual active bridge converter open circuit fault prediction model, and output the corresponding dual active bridge converter open circuit fault prediction result.

[0082] In order to implement the embodiment, the present invention also proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute each step in the method of the aforementioned technical solution.

[0083] like Figure 6As shown, the non-transitory computer-readable storage medium 900 includes a memory 910 of instructions and an interface 930, and the instructions can be executed by a processor 920 to complete the method. Alternatively, the storage medium can be a non-transitory computer-readable storage medium, for example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0084] In order to implement the embodiments, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method according to the embodiments of the present invention is implemented.

[0085] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0086] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0087] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0088] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0089] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the embodiments described, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0090] Those skilled in the art will understand that all or part of the steps of the method for implementing the embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0091] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0092] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it is understood that the embodiments are exemplary and are not to be construed as limiting the present invention. Those skilled in the art may make changes, modifications, substitutions, and variations to the embodiments within the scope of the present invention.

Claims

1. A method for predicting open-circuit faults in dual active bridge converters based on deep learning, characterized in that: include: Constructing a dual active bridge converter equivalent circuit simulation model, and creating an open circuit fault prediction training data set based on an open circuit detection signal and a corresponding open circuit fault result of the dual active bridge converter equivalent circuit simulation model; Based on the deep belief network model, a dual active bridge converter open circuit fault prediction model is constructed, and the dual active bridge converter open circuit fault prediction model is trained using the open circuit fault prediction training data set; A real-time open circuit detection signal of the dual active bridge converter equivalent circuit simulation model is obtained, and the trained dual active bridge converter open circuit fault prediction model is input, and the output result is the corresponding dual active bridge converter open circuit fault prediction result.

2. The method for predicting open-circuit faults of dual active bridge converters based on deep learning according to claim 1, characterized in that: The dual active bridge converter equivalent circuit simulation model includes two full-bridge circuits with symmetrical primary and secondary sides, and the full-bridge circuits are controlled by a preset phase shift strategy; the preset phase shift control strategy includes single phase shift (SPS), extended phase shift (EPS), and triple phase shift (TPS) control.

3. The method for predicting open circuit faults of dual active bridge converters based on deep learning according to claim 2, characterized in that: When the preset phase-shift control strategy is single phase-shift control, when the dual active bridge converter works normally, the conduction angle of the insulated gate bipolar transistors (IGBTs) on each bridge arm of the primary and secondary full-bridge circuits is 180°, and they are complementary, and the IGBTs in the diagonal positions are turned on or off at the same time, and the primary and secondary full-bridge circuits respectively output a square wave voltage U with a duty cycle of 50%. AB and U CD By adjusting U AB with U CD The size of the phase shift duty cycle between the two can change the size and direction of the transmission power; when the phase shift duty cycle is greater than 0, that is, U AB The phase is ahead of U CD When the phase shift duty cycle increases, the power will be transmitted in the forward direction; otherwise, the power will be transmitted in the reverse direction. The transmission power of the dual active bridge converter increases with the increase of the phase shift duty cycle.

4. The method for predicting open-circuit faults of dual active bridge converters based on deep learning according to claim 1, characterized in that: The open circuit detection signal of the dual active bridge converter equivalent circuit simulation model is the transformer leakage inductance current. Due to the symmetry of the dual active bridge converter, when the power switch device in the diagonal direction in the dual active bridge converter equivalent circuit simulation model is abnormal, the transformer leakage inductance current i L The waveforms show the same behavior, and the abnormalities of the switching devices in the diagonal direction are classified into the same open-circuit fault result category. Therefore, the dual active bridge converter equivalent circuit simulation model has a total of five open-circuit fault result categories.

5. The method for predicting open circuit faults of dual active bridge converters based on deep learning according to claim 3, characterized in that: In the dual active bridge converter equivalent circuit simulation model, when the dual active bridge converter is not open-circuited, according to Kirchhoff's voltage law, formula (1) is obtained: U AB =kU CD +u L +ri L (1) in: k is the transformer ratio; Define the ideal switching functions QA, QB, QC, and QD as: Among them, S1~S8 are IGBTs; D1~D8 are diodes; The output voltage of the primary and secondary full-bridge circuit is represented by the converter's input and output voltages U1 and U2: The mathematical model formula of the equivalent circuit of the dual active bridge converter is expressed as: Transformer leakage inductance current i L The formula is: Where: r is the winding resistance of the equivalent transformer, and L is the inductance of the equivalent transformer.

6. The method for predicting open-circuit faults of dual active bridge converters based on deep learning according to claim 1, characterized in that: The dual active bridge converter open circuit fault prediction model adopts a deep belief network (DBN), including: Multi-layer restricted Boltzmann machine (RBM), used for unsupervised layer-by-layer pre-training and feature extraction; A feedforward back propagation network is used to add a classifier on the top layer of the deep belief network DBN to supervise and fine-tune the parameters of the multi-layer restricted Boltzmann machine RBM; wherein, The dual active bridge converter open circuit fault prediction model is a deep belief network (DBN) with a two-layer RBM structure.

7. The method for predicting open-circuit faults of dual active bridge converters based on deep learning according to claim 1, characterized in that: The step of constructing the dual active bridge converter open circuit fault prediction model also includes the step of optimizing the deep learning network model DBN in the dual active bridge converter open circuit fault prediction model using a particle swarm algorithm, an ant colony algorithm or a genetic algorithm.

8. A deep learning-based open circuit fault prediction device for dual active bridge converters, characterized in that: include: A circuit simulation model construction module is used to construct an equivalent circuit simulation model of a dual active bridge converter, and create an open circuit fault prediction training data set based on the open circuit detection signal and the corresponding open circuit fault result of the dual active bridge converter equivalent circuit simulation model; A prediction model building module is used to build a dual active bridge converter open circuit fault prediction model based on a deep belief network model, and train the dual active bridge converter open circuit fault prediction model using the open circuit fault prediction training data set; The prediction module is used to obtain the real-time open circuit detection signal of the dual active bridge converter equivalent circuit simulation model, input the trained dual active bridge converter open circuit fault prediction model, and output the corresponding dual active bridge converter open circuit fault prediction result.

9. An electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform each step in the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to enable the computer to execute each step of the method according to any one of claims 1 to 7.

Images