Method and device for determining working mode of charging station and storage medium

Abstract

The invention discloses a method and a device for determining a working mode of a charging station and a storage medium, and relates to the technical field of power dispatching. The method comprises the following steps: acquiring a harmonic signal of a to-be-detected charging station; the harmonic signals are input into a target detection model to be analyzed, amplitude data and phase angle data of the harmonic signals are obtained, and the target detection model is a tree-shaped two-layer combination model obtained through training by means of an amplitude data sample and a phase angle data sample of historical harmonic signals of the charging station and structural characteristics and operation characteristics of the charging station; and determining the working mode of the charging station based on the amplitude data and the phase angle data. The technical problem that the working mode of the charging station cannot be determined due to low harmonic detection efficiency of the charging station is solved.

Description

Technical Field

[0001] The present invention relates to the field of harmonic detection technology, and in particular to a method, device and storage medium for determining an operating mode of a charging station. Background Art

[0002] At present, the S2G (Station-to-Grid) charging station is an advanced charging infrastructure that allows the charging station to interact with the power grid and optimize the distribution and use of power resources through an intelligent management system. However, the operation of the S2G charging station may also have a harmonic impact on the power grid. As a typical nonlinear load during the charging process, the electric vehicle charging station generates a large amount of harmonics, which flow into the distribution network through the public connection point, reducing the power factor of the charger. When the power department measures the meter, part of the reactive current in the value is calculated as active current, causing a certain error. Therefore, it is of utmost importance to detect harmonics at the charging station to determine the working mode of the charging station to ensure the normal operation of the charging station.

[0003] In the related technology, commonly used harmonic detection methods include instantaneous reactive power theory and Hilbert-Huang transform analysis. The instantaneous reactive power theory initially defined the pq theory based on instantaneous real power p and instantaneous imaginary power q, and then gradually improved this theory, thus generating i p -i q Detection method is used to detect fundamental and harmonic components in three-phase current. This method has good real-time performance and simple principle. It can detect fundamental, harmonic and other frequency components and compensate reactive power, but it is only applicable to harmonic detection and fundamental reactive current detection of three-phase circuits with three-phase sinusoidal symmetry. Fourier transform analysis method can map time domain signals to frequency domain to accurately detect each harmonic component, but there is loss of time domain information, and it is impossible to achieve coexistence of time and frequency domain information. Transient interference, non-stationary signals, and asynchronous sampling will lead to spectrum leakage and fence effect. Therefore, there is a technical problem that the working mode of the charging station cannot be determined due to the low efficiency of harmonic detection of the charging station.

[0004] With respect to the above-mentioned technical problem that the working mode of the charging station cannot be determined due to the low efficiency of harmonic detection of the charging station, no effective solution has been proposed yet. Summary of the invention

[0005] Embodiments of the present invention provide a method, device and storage medium for determining the working mode of a charging station, so as to at least solve the technical problem that the working mode of a charging station cannot be determined due to low harmonic detection efficiency of the charging station.

[0006] According to one aspect of an embodiment of the present invention, a method for determining the working mode of a charging station is provided. The method may include: obtaining a harmonic signal of a charging station to be detected; inputting the harmonic signal into a target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal, wherein the target detection model is a tree-shaped two-layer combination model obtained by training using amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as structural features and operating features of the charging station; based on the amplitude data and phase angle data, determining the working mode of the charging station.

[0007] Optionally, the method for determining the working mode of the charging station also includes: acquiring structural characteristics and operating characteristics of the charging station, the structural characteristics are used to indicate the structure of the charging station, and the operating characteristics are used to indicate the operating rules of the charging station; based on the structural characteristics and the operating characteristics, determining sampling parameters of the harmonic signal of the charging station; based on the sampling parameters, acquiring historical harmonic signals; and using the historical harmonic signals to train an initial detection model to obtain a target detection model.

[0008] Optionally, the initial detection model is trained using historical harmonic signals to obtain a target detection model, including: determining training nodes based on historical harmonic signals, wherein the training nodes are used to indicate the number of node layers for training the initial detection model; training the initial detection model based on amplitude data samples and phase angle data samples of historical harmonic signals and the training nodes to obtain initial detection results, wherein the initial detection results are used to indicate the results output by the initial detection model; and determining the target detection model based on the initial detection results.

[0009] Optionally, based on the initial detection result, determining the target detection model includes: comparing the initial detection result with an expected threshold to obtain a comparison result; in response to the comparison result being greater than or equal to the comparison result threshold, determining that the initial detection model corresponding to the initial detection result is the target detection model.

[0010] Optionally, the method for determining the working mode of the charging station further includes: in response to the comparison result being less than a comparison result threshold, determining an update parameter, wherein the update parameter is used to update the initial detection model; and using the update parameter to update the initial detection model to obtain a target detection model.

[0011] Optionally, determining the operating mode of the charging station based on the amplitude data and the phase angle data includes: determining a charging state of the charging station based on the amplitude data and the phase angle data; and determining the operating mode of the charging station based on the charging state.

[0012] According to another aspect of an embodiment of the present invention, a device for determining the working mode of a charging station is also provided. The device may include: an acquisition unit for acquiring a harmonic signal of a charging station to be detected; an analysis unit for inputting the harmonic signal into a target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal, wherein the target detection model is a tree-shaped two-layer combination model obtained by training using amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as structural features and operating features of the charging station; a determination unit for determining the working mode of the charging station based on the amplitude data and phase angle data.

[0013] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium including a stored program, wherein when the program is executed by a processor, the device where the storage medium is located is controlled to execute the method for determining the working mode of the charging station in an embodiment of the present invention.

[0014] According to another aspect of an embodiment of the present invention, a processor is further provided. The processor is used to run a program, wherein when the program is run, the method for determining the working mode of the charging station in the embodiment of the present invention is executed.

[0015] According to another aspect of an embodiment of the present invention, a computer program product is further provided, wherein the program product includes computer instructions, and when the computer instructions are executed by a processor, the method for determining the working mode of the charging station in the embodiment of the present invention is implemented.

[0016] In an embodiment of the present invention, a harmonic signal of a charging station to be detected is obtained; the harmonic signal is input into a target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal, wherein the target detection model is a tree-shaped two-layer combination model obtained by training with amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as the structural characteristics and operation characteristics of the charging station; based on the amplitude data and phase angle data, the working mode of the charging station is determined. That is, the present invention detects the harmonic signal of the charging station, inputs the harmonic signal into the target detection model, obtains the amplitude data and phase angle data of the harmonic signal, and thus determines the working mode of the charging station according to the amplitude data and phase angle data. Since the target detection model of the present invention is a tree-shaped two-layer combination model obtained by training with amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as the structural characteristics and operation characteristics of the charging station, the target detection model can accurately analyze the characteristics of the harmonic signal, solve the technical problem that the working mode of the charging station cannot be determined due to the low harmonic detection efficiency of the charging station, and achieves the technical effect of improving the harmonic detection efficiency of the charging station, thereby determining the working mode of the charging station. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 is a flow chart of a method for determining an operating mode of a charging station according to an embodiment of the present invention;

[0019] Figure 2 is a flow chart of a harmonic detection method based on a wavelet-LVQ neural network according to an embodiment of the present invention;

[0020] Figure 3 is a flow chart of a calculation method based on a wavelet-LVQ neural network according to an embodiment of the present invention;

[0021] Figure 4 is a schematic diagram of a network structure based on a wavelet-LVQ neural network according to an embodiment of the present invention;

[0022] Figure 5 is a schematic diagram of a device for determining an operating mode of a charging station according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, functional component or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, functional components or devices.

[0025] According to an embodiment of the present invention, an embodiment of a method for determining an operating mode of a charging station is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0026] Figure 1 is a flow chart of a method for determining an operating mode of a charging station according to an embodiment of the present invention. Figure 1 As shown, the method may include the following steps:

[0027] Step S101, obtaining a harmonic signal of a charging station to be detected.

[0028] In the technical solution provided in the above step S101 of the present invention, the charging station to be detected can at least be an S2G ​​charging station.

[0029] In this embodiment, the harmonic signal of the charging station to be detected is obtained. For example, the harmonic signal of the charging station to be detected is detected by using a harmonic analysis instrument. This is only an illustrative example and does not limit the specific method of detecting the harmonic signal of the charging station to be detected.

[0030] Step S102: input the harmonic signal into the target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal.

[0031] In the technical solution provided in the above step S102 of the present invention, the target detection model is a tree-shaped two-layer combination model trained using amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as structural features and operating features of the charging station.

[0032] In this embodiment, after the harmonic signal of the charging station to be detected is acquired in step S101, the harmonic signal is input into the target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal.

[0033] Optionally, structural characteristics and operating characteristics of the charging station are obtained, the structural characteristics are used to indicate the structure of the charging station, and the operating characteristics are used to indicate the operating rules of the charging station; based on the structural characteristics and operating characteristics of the charging station, sampling parameters of the harmonic signal of the charging station are determined, and the sampling parameters include at least a sampling position and a sampling value.

[0034] For example, assuming that the charging station to be tested is a six-pulse charging pile, the charging station rectifier uses a six-pulse rectifier bridge, and the AC side impedance is ignored. a (t), S b (t), S c(t) represents the diode switching function. The Fourier series expressions in one cycle are shown in the following formulas (1) and (2):

[0035]

[0036] If the current is continuously output, the output voltage expression of the rectifier bridge is as shown in the following formula (3):

[0037] u d (t) = s a (t)u a (t)+s b (t)u b (t)+s c (t)u c (t) = u d +Σu dn (t) (3)

[0038] Among them, the components of the nth harmonic are decomposed into u by Fourier series. dn (t) and DC component u d As shown in the following formula (4) and formula (5) respectively:

[0039]

[0040] DC side current i d The expression of (t) is shown in the following formula (6):

[0041]

[0042] Among them, Z d is the DC side resistance, Z d =R C , Z n is the nth harmonic impedance, Z n As shown in the following formula (7):

[0043] Z n =jnωL f +R C / / (1 / jnωC f )=|Z n | (7)

[0044] The DC side current i d Substituting (t) into the above formula (7), we can get the result Z n As shown in the following formula (8):

[0045]

[0046] Among them, the a-phase current can be expressed as the following formula (9):

[0047]

[0048] Based on the above calculations, it can be concluded that the S2G charging station mainly generates 3rd, 5th, and 7th harmonic currents, all of which are odd harmonics. The 3rd, 5th, and 7th harmonics of the charging station are sampled. Since the highest frequency in the signal is the 7th harmonic frequency, which is about 350Hz, according to the Nyquist theorem, the sampling frequency is set to 1000Hz; one cycle of the harmonic signal is about 0.02s. To ensure the detection accuracy, two cycles of 0.04s are taken, so the sampling point is set to 40 and two cycles are sampled.

[0049] Optionally, historical harmonic signals are obtained according to sampling parameters; training nodes are determined according to the historical harmonic signals, wherein the training nodes are used to indicate the number of node layers for training the initial detection model. The training nodes may also be referred to as the number of hidden layer nodes.

[0050] For example, in order to reduce the amount of calculation and improve the calculation efficiency, the number of hidden layer nodes is generally much smaller than the number of training samples. The number of training samples needs to be higher than the connection weight of the network. When performing harmonic detection of S2G charging stations, the detection accuracy increases with the increase of the number of hidden layer nodes. However, after the number of hidden layer nodes reaches a certain value, continuing to increase the number of hidden layer nodes has little effect on the detection accuracy. The network hidden layer nodes are determined by the mean of the number of input layer and output layer nodes. n is the number of input layer neurons, l is the number of output layer neurons, and m is the number of hidden layer neurons. [] means rounding down the number in the symbol. The expression of the number of hidden layer nodes is shown in the following formula (10):

[0051]

[0052] Optionally, the initial detection model is trained according to the amplitude data samples and phase angle data samples of the historical harmonic signal and the training nodes to obtain an initial detection result, wherein the initial detection result is used to indicate the result output by the initial detection model.

[0053] For example, the initial detection result is compared with the expected threshold to obtain a comparison result; when the comparison result is greater than or equal to the comparison threshold, the initial detection model corresponding to the initial detection result is determined to be the target detection model.

[0054] Step S103: determining the working mode of the charging station based on the amplitude data and the phase angle data.

[0055] In the technical solution provided in the above step S103 of the present invention, the working mode of the charging station at least includes a charging mode or a discharging mode.

[0056] In this embodiment, after the amplitude data and the phase angle data of the harmonic signal are determined in step S102, the working mode of the charging station is determined according to the amplitude data and the phase angle data.

[0057] Optionally, the charging state of the charging station is determined based on the amplitude data and the phase angle data; thereby, the operating mode of the charging station is determined based on the charging state.

[0058] For example, the working mode of the charging station can be: during the charging process, the rectifier generates harmonics; during the discharging process, the inverter generates harmonics. The harmonics of the rectifier are mainly concentrated in the low order, such as the 5th, 7th, 11th and 13th, and the harmonics generated by the inverter.

[0059] It should be noted that the above embodiment can be executed by a device for determining the working mode of a charging station.

[0060] In the above steps S101 to S103 of the present invention, by detecting the harmonic signal of the charging station, the harmonic signal is input into the target detection model to obtain the amplitude data and phase angle data of the harmonic signal, thereby determining the working mode of the charging station according to the amplitude data and phase angle data. Since the target detection model of the present invention is a tree-shaped two-layer combination model obtained by training with the amplitude data samples and phase angle data samples of the historical harmonic signal of the charging station, as well as the structural characteristics and operation characteristics of the charging station, the target detection model can accurately analyze the harmonic signal characteristics, solve the technical problem that the working mode of the charging station cannot be determined due to the low harmonic detection efficiency of the charging station, and achieve the technical effect of improving the harmonic detection efficiency of the charging station, thereby determining the working mode of the charging station.

[0061] The above method of this embodiment is further introduced below.

[0062] As an optional implementation manner, the method for determining the working mode of a charging station also includes: acquiring structural characteristics and operating characteristics of the charging station, the structural characteristics are used to indicate the structure of the charging station, and the operating characteristics are used to indicate the operating rules of the charging station; based on the structural characteristics and the operating characteristics, determining sampling parameters of the harmonic signal of the charging station; based on the sampling parameters, acquiring historical harmonic signals; and using the historical harmonic signals to train an initial detection model to obtain a target detection model.

[0063] In this embodiment, the structural characteristics and operating characteristics of the charging station are obtained; based on the structural characteristics and operating characteristics, the sampling parameters of the harmonic signal of the charging station are determined; based on the sampling parameters, the historical harmonic signal is obtained; and the initial detection model is trained using the historical harmonic signal to obtain a target detection model.

[0064] Optionally, the sampling parameters are specifically determined as shown in the above formulas (1) to (9), which will not be described again here.

[0065] As an optional implementation method, the initial detection model is trained using historical harmonic signals to obtain a target detection model, including: determining training nodes based on historical harmonic signals, wherein the training nodes are used to indicate the number of node layers for training the initial detection model; training the initial detection model based on amplitude data samples and phase angle data samples of historical harmonic signals and the training nodes to obtain initial detection results, wherein the initial detection results are used to indicate the results output by the initial detection model; and determining the target detection model based on the initial detection results.

[0066] In this embodiment, a training node is determined based on a historical harmonic signal, and an initial detection model is trained based on amplitude data samples and phase angle data samples of the historical harmonic signal and the training nodes to obtain an initial detection result, thereby determining a target detection model based on the initial detection result.

[0067] For example, assuming that the initial detection model is a wavelet-LVQ neural network, the input layer, hidden layer, and output layer of the wavelet neural network are designed to perform Fourier series expansion on the harmonic current signal of the S2G charging station, where n is the harmonic order, a 0 、a n 、b n is the Fourier coefficient i(t) is the harmonic current signal, and the expressions are shown in the following formula (11):

[0068]

[0069] The sampling value and sampling position of the harmonic current signal are used as the input of the wavelet neural network. According to the sampling period and sampling nodes, the number of input layer nodes of the wavelet neural network is designed to be 40, the number of hidden layer neurons is 24, and the output layer is the amplitude of different harmonic orders. Taking the highest harmonic as the 7th harmonic as an example, the output needs to include the 3rd harmonic, the 5th harmonic and the 7th harmonic. Each harmonic requires two outputs, so the number of output layer nodes is 6. The initial parameters are optimized, and the wavelet-LVQ neural network continuously adjusts the connection weights between networks and the scaling and translation factors of the wavelet function through training. The weights between each neuron are ω ij , each weight has a threshold ω 0i , the scaling and translation factors are a i , b i These parameters are arranged in sequence and encoded using real values. The optimization objective is to minimize the output error of the neural network. The expressions of the real number encoding form and the optimization objective are shown in the following formulas (15) and (16), respectively:

[0070]

[0071]

[0072] in, is the expected output, is the actual output of the network, and the fitness of the genetic algorithm is shown in the following formulas (17) and (18):

[0073]

[0074] For another example, in the harmonic detection of S2G charging stations, due to the presence of a large amount of pulse noise interference, there are a large number of spikes in the signal, and the required amount of calculation is large, which places higher requirements on the real-time performance of wavelet neural network calculation. Moret wavelet is widely used, has better locality in the time and frequency domain, and is relatively simple to calculate. Therefore, the Morle wavelet function is used as the activation function of the wavelet neural network, and the function expression is shown in the following formula (19) and formula (20):

[0075]

[0076] Coding and population initialization, the parameters of the wavelet neural network include the connection weights v ji and w kj , expansion factor a j and translation factor b j according to Perform real number coding. Generate an initial code randomly and initialize the population P. Select the operation, substitute the current population data into the wavelet neural network, calculate the fitness value of each individual, sort and select them, select the individual variables for the next evolution operation, and record the current optimal fitness and average fitness value. The fitness function f(n) of individual n is shown in the following formula (21):

[0077]

[0078] According to the crossover operator, the probability Pc of the two parents performing the crossover is calculated, and the two parents are crossovered according to the crossover probability to generate new individuals, and the individuals that have not been operated are kept unchanged and directly copied before proceeding to the next step. The uniform mutation method that is more suitable for floating-point coding is adopted to generate new individuals using the probability Pm. The newly generated individuals are added to the population, and the fitness value of the new individuals is calculated to achieve iterative evolution. Determine whether the number of iterations has reached the upper limit. If the stopping condition is met, that is, the number of iterations reaches the upper limit or the error is less than the threshold, the current optimal individual is recorded and the process ends. If the stopping criterion is met and the optimal individual is iterated, the optimal individual is reversely decoded according to the encoding method to obtain the initial network parameters, connection weights and scaling factors of the wavelet neural network, and then brought into the calculation of the wavelet neural network to calculate the network output value.

[0079] As an optional implementation method, based on the initial detection result, a target detection model is determined, including: comparing the initial detection result with an expected threshold to obtain a comparison result; in response to the comparison result being greater than or equal to the comparison result threshold, determining that the initial detection model corresponding to the initial detection result is the target detection model.

[0080] In this embodiment, the initial detection result is compared with the expected threshold to obtain a comparison result; when the comparison result is greater than or equal to the comparison result threshold, it means that the initial detection result is qualified. Based on this, it can be determined that the initial detection model corresponding to the initial detection result is the target detection model.

[0081] As an optional implementation manner, the method for determining the working mode of the charging station also includes: in response to the comparison result being less than the comparison result threshold, determining an update parameter, wherein the update parameter is used to update the initial detection model; using the update parameter, updating the initial detection model to obtain a target detection model.

[0082] In this embodiment, if the comparison result is less than the comparison result threshold, an update parameter is determined; and the update parameter is used to update the initial detection model to obtain a target detection model.

[0083] For example, the gradient is used to perform reverse operations to update the network connection weights and the scaling and translation coefficients of the wavelet.

[0084] As an optional implementation manner, determining the working mode of the charging station based on the amplitude data and the phase angle data includes: determining the charging state of the charging station based on the amplitude data and the phase angle data; and determining the working mode of the charging station based on the charging state.

[0085] In this embodiment, the charging state of the charging station is determined according to the amplitude data and the phase angle data; and the working mode of the charging station is determined according to the charging state.

[0086] For example, the working mode of the charging station can be determined based on the detected harmonic current content and the location where the harmonics are generated. During the charging process, the rectifier generates harmonics; during the discharging process, the inverter generates harmonics. The harmonics of the rectifier are mainly concentrated in the low order, such as the 5th, 7th, 11th and 13th, and the harmonics generated by the inverter.

[0087] It should be noted that the above embodiment can be executed by a device for determining the working mode of a charging station.

[0088] In this embodiment, by detecting the harmonic signal of the charging station, the harmonic signal is input into the target detection model to obtain the amplitude data and phase angle data of the harmonic signal, thereby determining the working mode of the charging station according to the amplitude data and phase angle data. Since the target detection model of the present invention is a tree-shaped two-layer combination model obtained by training with the amplitude data samples and phase angle data samples of the historical harmonic signal of the charging station, as well as the structural characteristics and operation characteristics of the charging station, the target detection model can accurately analyze the harmonic signal characteristics, solve the technical problem that the working mode of the charging station cannot be determined due to the low harmonic detection efficiency of the charging station, and achieve the technical effect of improving the harmonic detection efficiency of the charging station, thereby determining the working mode of the charging station.

[0089] The technical solution of the embodiment of the present invention is illustrated below in conjunction with preferred implementation modes.

[0090] At present, S2G charging stations are an advanced charging infrastructure that allows charging stations to interact with the power grid and optimize the distribution and use of power resources through intelligent management systems. However, the operation of S2G charging stations may also have harmonic effects on the power grid. As a typical nonlinear load during the charging process, electric vehicle charging stations generate a large amount of harmonics, which flow into the distribution network through the public connection point, reducing the power factor of the charger. When the power department measures the meter, part of the reactive current in the value is calculated as active current, causing a certain error. Therefore, it is of utmost importance to detect harmonics at the charging station to determine the working mode of the charging station to ensure the normal operation of the charging station.

[0091] In the related technology, commonly used harmonic detection methods include instantaneous reactive power theory and Hilbert-Huang transform analysis. The instantaneous reactive power theory initially defined the pq theory based on instantaneous real power p and instantaneous imaginary power q, and then gradually improved this theory, thus generating i p -i q Detection method is used to detect fundamental and harmonic components in three-phase current. This method has good real-time performance and simple principle. It can detect fundamental, harmonic and other frequency components and compensate reactive power, but it is only applicable to harmonic detection and fundamental reactive current detection of three-phase circuits with three-phase sinusoidal symmetry; Fourier transform analysis method can map time domain signals to frequency domain to complete accurate detection of each harmonic component, but there is loss of time domain information, and it is impossible to achieve coexistence of time and frequency domain information. Transient interference, non-stationary signals, and asynchronous sampling will lead to spectrum leakage and fence effect. Therefore, there is a technical problem that the working mode of the charging station cannot be determined due to the low efficiency of harmonic detection of the charging station. In view of the above-mentioned technical problem that the working mode of the charging station cannot be determined due to the low efficiency of harmonic detection of the charging station, no effective solution has been proposed.

[0092] However, an embodiment of the present invention proposes a harmonic detection method for an S2G ​​charging station based on wavelet-LVQ neural network fusion, uses a genetic algorithm to initialize network parameters, and combines the wavelet neural network with the LVQ neural network based on the structure of the S2G charging station, thereby solving the technical problem of being unable to determine the working mode of the charging station due to the low harmonic detection efficiency of the charging station, and achieving the technical effect of improving the harmonic detection efficiency of the charging station and determining the working mode of the charging station.

[0093] The embodiments of the present invention are further described below.

[0094] Figure 2 1 is a flow chart of a harmonic detection method based on a wavelet-LVQ neural network according to an embodiment of the present invention, the harmonic detection method comprising the following steps:

[0095] Step S201: data preprocessing.

[0096] In this embodiment, the harmonic signal of the S2G charging station is sampled and analyzed and a neural network structure is designed.

[0097] Optionally, a harmonic sampling design is used to analyze the harmonic signals of the charging station according to the structural characteristics and operating characteristics of the S2G charging station.

[0098] Optionally, the network propagation error is designed to introduce the scaling and translation factors of the wavelet function into the wavelet-LVQ network, and the squared error gradient of the output is used to update the parameters.

[0099] Optionally, the number of hidden layer nodes is determined by determining the network hidden layer nodes by the average of the number of input layer nodes and the number of output layer nodes.

[0100] Step S202: Use the wavelet-LVQ neural network to train and analyze the harmonic samples of the S2G charging station to obtain the amplitude and phase angle data of each harmonic.

[0101] In this embodiment, the wavelet-LVQ neural network is used to train and analyze the harmonic samples of the S2G charging station to obtain the amplitude and phase angle data of each harmonic.

[0102] Optionally, the input layer, hidden layer and output layer of the wavelet neural network are designed.

[0103] Optionally, initial parameter optimization uses a genetic algorithm to optimize the initial parameters of the network.

[0104] Optionally, the activation function is selected. The real-time requirement of wavelet neural network calculation is higher. Morle wavelet is widely used, has better locality in time and frequency domain, and is simpler to calculate. Morle wavelet function is used as the activation function of wavelet neural network.

[0105] Optionally, the optimized initial network parameters, connection weights and scaling translation factors of the wavelet neural network are brought into the calculation of the wavelet neural network.

[0106] Optionally, the difference between the network output value and the expected value is calculated to determine whether the error is less than a preset threshold or the number of iterations exceeds an upper limit.

[0107] Optionally, a reverse operation is performed based on the output mean square error using the gradient to update the network connection weights and the wavelet scaling and translation coefficients so that the network output gradually approaches the expected value.

[0108] Optionally, wavelet transform has excellent time-frequency analysis capabilities and can accurately analyze the local characteristics of signals, especially for non-stationary and nonlinear harmonic signals. Combined with the classification ability of LVQ network, the accuracy of harmonic detection can be improved. Wavelet network can process real-time signals and quickly respond to changes in grid status, which is especially important for occasions such as electric vehicle charging stations that require rapid response. As a self-organizing neural network, LVQ network has the ability of automatic learning and adaptive adjustment, and can work stably in the presence of noise and interference.

[0109] Optionally, preprocessing of the signal with wavelet transform can enhance the robustness of the network and maintain stable detection performance even when the signal characteristics change.

[0110] Step S203, combining the LVQ network with the wavelet neural network to construct a tree-shaped two-layer combination model.

[0111] In this embodiment, during the learning phase, the output layer of the wavelet neural network is used as the input layer of the LVQ network, and the initial value is selected as a random value to perform calculations; the amplitude and phase angle of each harmonic are obtained.

[0112] Optionally, the LVQ network is good at pattern recognition and classification, and can identify and classify different harmonic sources, which is of great significance for the identification and management of harmonic sources. Compared with the traditional Fourier transform, the wavelet transform usually has higher computational efficiency when processing real signals, especially when processing non-stationary signals.

[0113] Step S204: Detect harmonics using a tree-shaped two-layer combination model.

[0114] In this embodiment, the tree-shaped two-layer combination model constructed in step S203 is used to detect harmonics.

[0115] The method of the present invention is further described below by means of specific examples.

[0116] Taking the six-pulse charging pile as an example, considering the harmonic characteristics generated by the S2G charging station and the characteristics of the S2G charging station itself, the harmonics in the working process of the charging station are analyzed.

[0117] In this embodiment, the harmonic sampling design analyzes the harmonic signals of the charging station according to the structural characteristics and operating characteristics of the S2G charging station. The S2G charging process is divided into a constant current and voltage limiting stage and a constant voltage and current limiting stage. In the constant current and voltage limiting stage: the current is large at the beginning, and the power supply can be continuously and stably increased. The power increases with the voltage, and the maximum voltage and output power appear at the moment of charging completion; the constant voltage and current limiting stage: the output voltage is maintained at the highest value at the beginning, and then the output power decreases as the output current decreases.

[0118] Optionally, the charging station rectifier adopts a six-pulse rectifier bridge, ignoring the AC side impedance, S a (t), S b (t), S c (t) represents the diode switching function. The Fourier series expressions in one cycle are shown in the above formulas (1) and (2). If the current is output continuously, the output voltage expression of the rectifier bridge is shown in the above formula (3). The components of the nth harmonic are decomposed into u by Fourier series. dn (t) and DC component u d As shown in the above formulas (4) and (5), the DC side current i d The expression of (t) is shown in the above formula (6), Z d is the DC side resistance, Z d =R C , Z n is the nth harmonic impedance, Z n As shown in the above formula (7), the DC side current i d Substituting (t) into the above formula (7), we can get the result Z n As shown in the above formula (8): The a-phase current can be expressed as the above formula (9), which will not be repeated here.

[0119] Optionally, the above calculations show that the S2G charging station mainly generates 3rd, 5th, and 7th harmonic currents, all of which are odd harmonics. The 3rd, 5th, and 7th harmonics of the charging station are sampled. Since the highest frequency in the signal is the 7th harmonic frequency, which is about 350Hz, the sampling frequency is set to 1000Hz according to the Nyquist theorem; one cycle of the harmonic signal is about 0.02s. To ensure the detection accuracy, two cycles of 0.04s are taken, so the sampling point is set to 40 and two cycles are sampled.

[0120] Optionally, the activation function of the wavelet-LVQ neural network is a wavelet function, a scaling and translation factor is introduced, and the square error gradient of the output is used to update the parameters.kl is the kth value of the input sample connected to the Lth layer, y il is the i-th dimension value output by the l-th neural network, w ij is the weight coefficient of the connection from the hidden layer to the output layer, r jk is the weight coefficient of the connection from the input layer to the hidden layer, a j is the expansion coefficient of the wavelet function, b j is the translation coefficient of the wavelet function, d il is the output value of the i-th neuron in the l-th dimension, E is the square error between the output and the expectation, u i is the threshold of the neural network, which is responsible for controlling whether the activation function works. m is the number of input layers, L is the number of samples, N is the number of output layers, and n is the number of hidden layers. The network error propagation process expression is shown in the following formula (22) and formula (23):

[0121]

[0122] Among them, if net jl As shown in the following formula (24), y il As shown in the following formula (25):

[0123]

[0124] Where η is the learning coefficient, and the parameter adjustment process is shown in the following formulas (26) to (29):

[0125]

[0126]

[0127] Optionally, the number of hidden layer nodes is determined. In order to reduce the amount of calculation and improve the calculation efficiency, the number of hidden layer nodes is generally much smaller than the number of training samples. The number of training samples needs to be higher than the connection weight of the network. When performing harmonic detection of S2G charging stations, the detection accuracy increases with the increase of the number of hidden layer nodes. However, after the number of hidden layer nodes reaches a certain value, continuing to increase the number of hidden layer nodes has little effect on the detection accuracy. The hidden layer nodes of the network are determined by the mean of the number of input layer and output layer nodes. n is the number of input layer neurons, l is the number of output layer neurons, m is the number of hidden layer neurons, [] indicates rounding down the number in the symbol, and the expression of the number of hidden layer nodes is as shown in the above formula (10), which will not be repeated here.

[0128] Figure 3 FIG. 1 is a flow chart of a calculation method based on a wavelet-LVQ neural network according to an embodiment of the present invention. Figure 3As shown in the figure, the wavelet-LVQ neural network is used to train and analyze the harmonic samples of the S2G charging station to obtain the amplitude and phase angle data of each harmonic.

[0129] Step S301, generating training samples.

[0130] In this embodiment, Figure 4 is a schematic diagram of a network structure based on a wavelet-LVQ neural network according to an embodiment of the present invention. Figure 4 As shown in Figure 1, the network structure includes input layer, hidden layer, output layer and competition layer. The input layer, hidden layer and output layer of the wavelet neural network are designed to perform Fourier series expansion on the harmonic current signal of the S2G charging station. n is the harmonic order, a 0 、a n 、b n is the Fourier coefficient i(t) is the harmonic current signal, and the expressions are respectively shown in the above formulas (11) to (14), which will not be repeated here.

[0131] Optionally, the sampling value and sampling position of the harmonic current signal are used as the input of the wavelet neural network. According to the sampling period and sampling nodes, the number of input layer nodes of the wavelet neural network is designed to be 40, the number of hidden layer neurons is 24, and the output layer is the amplitude of different harmonic orders. Taking the highest harmonic as the 7th harmonic as an example, the output needs to include the 3rd harmonic, the 5th harmonic and the 7th harmonic. Each harmonic requires two outputs, so the number of output layer nodes is 6.

[0132] Step 302: initial parameter optimization.

[0133] In this embodiment, the wavelet-LVQ neural network continuously adjusts the connection weights between networks and the scaling and translation factors of the wavelet function through training. The weights between each neuron are ω ij , each weight has a threshold ω 0i , the scaling and translation factors are a i , b i These parameters are arranged in sequence and encoded using real values. The optimization objective is to minimize the output error of the neural network. The expressions of the real number encoding form and the optimization objective are shown in the above formulas (15) and (16), respectively. is the expected output, is the actual output of the network. The fitness of the genetic algorithm is shown in the above formulas (17) and (18), which will not be repeated here.

[0134] Step 303, performing forward calculation on the neural network.

[0135] In this embodiment, the activation function is determined. In the harmonic detection of S2G charging stations, due to the presence of a large amount of pulse noise interference, there are a large number of spikes in the signal, and the required amount of calculation is large, so the real-time requirements of the wavelet neural network calculation are higher. Moret wavelet is widely used, has good locality in the time and frequency domain, and is relatively simple to calculate. Therefore, the Morle wavelet function is used as the wavelet neural network, and the function expression is shown in the above formula (19) and formula (20), which will not be repeated here.

[0136] Optionally, encoding and population initialization, the parameters of the wavelet neural network include the connection weights v ji and w kj , expansion factor a j and translation factor b j according to Perform real number coding. Generate an initial code randomly and initialize the population P.

[0137] Optionally, select an operation. Substitute the current population data into the wavelet neural network, calculate the fitness value of each individual, sort and select them, select the individual variables for the next evolutionary operation, and record the current optimal fitness and average fitness value. The fitness function f(n) of individual n is shown in the above formula (21), which will not be repeated here.

[0138] Optionally, the initial network parameters, connection weights and scaling factors of the wavelet neural network are calculated. The probability Pc of two parents performing crossover is calculated according to the crossover operator, and the two parents are crossovered according to the crossover probability to generate new individuals, and the individuals that have not been operated are kept unchanged and directly copied before proceeding to the next step. The uniform mutation method that is more suitable for floating-point coding is used to generate new individuals using the probability Pm.

[0139] Optionally, the newly generated individual is added to the population, and the fitness value of the new individual is calculated to achieve iterative evolution. Determine whether the number of iterations reaches the upper limit. If the stopping condition is met, that is, the number of iterations reaches the upper limit or the error is less than the threshold, the current optimal individual is recorded and ended. If the stopping criterion is met and the optimal individual is iterated, the optimal individual is reversely decoded according to the encoding method to obtain the initial network parameters, connection weights and scaling factors of the wavelet neural network, and then brought into the calculation of the wavelet neural network.

[0140] Step S304, calculating the error.

[0141] In this embodiment, the difference between the network output value and the expected value is calculated.

[0142] Step S305: whether the stop condition is met.

[0143] In this embodiment, it is determined whether the error is less than a preset threshold or the number of iterations exceeds an upper limit. If so, the training process ends and the structural parameters of the network are output; otherwise, step S306 is performed.

[0144] Step S306, performing reverse operation on the neural network.

[0145] In this embodiment, according to the output mean square error, the gradient is used to perform reverse operation to update the network connection weights and the wavelet scaling and translation coefficients, so that the network output gradually approaches the expected value.

[0146] Step S307, adjusting the weights of each layer.

[0147] In this embodiment, based on the wavelet neural network, the LVQ network is combined with the wavelet neural network to construct a tree-shaped two-layer combination model. There are several wavelet neural networks as classifiers at the bottom layer, and then the LVQ neural network is used as the upper-layer classification combiner to solve the problems of slow convergence speed of wavelet neural networks, easy to fall into local optimal solutions, difficult to determine the structure, and random initial parameters having a large impact on network performance.

[0148] Optionally, in the learning phase, the output layer of the wavelet neural network is used as the input layer of the LVQ network, and the input sequence is specified as: {x 1 ,x 2 ,x 3 ,…,x N}, the initial value is selected as a random value, the learning coefficient α(k) is selected, and the iteration starts from k = 0. In the test phase, the maximum value is learned through the MAXNET network competition, and its corresponding Voronoi quantized vector attribute is the category of x. The iterative expression is shown in the following formula (30):

[0149]

[0150] The amplitude and phase angle of each harmonic are shown in the following formula (31):

[0151]

[0152] Optionally, based on the different mechanisms and characteristics of harmonics generated by the S2G charging station during charging and discharging, the working mode of the charging station is determined according to the detected harmonic current content and the location where the harmonics are generated. During the charging process, the rectifier generates harmonics; during the discharging process, the inverter generates harmonics. The harmonics of the rectifier are mainly concentrated in low orders, such as the 5th, 7th, 11th and 13th, and the harmonics generated by the inverter.

[0153] In this embodiment, a genetic algorithm is used to initialize network parameters, and based on the structure of the S2G charging station, a wavelet neural network is combined with a LVQ neural network, thereby solving the technical problem of being unable to determine the working mode of the charging station due to the low harmonic detection efficiency of the charging station, and achieving the technical effect of improving the harmonic detection efficiency of the charging station and determining the working mode of the charging station.

[0154] According to an embodiment of the present invention, a device for determining the working mode of a charging station is also provided. It should be noted that the device for determining the working mode of a charging station can be used to execute the method for determining the working mode of a charging station in the method embodiment.

[0155] Figure 5 is a schematic diagram of a device for determining the working mode of a charging station according to an embodiment of the present invention. Figure 5 As shown, the device 500 for determining the working mode of the charging station may include: an acquisition unit 501 , an analysis unit 502 and a determination unit 503 .

[0156] The acquisition unit 501 is used to acquire the harmonic signal of the charging station to be detected.

[0157] The analysis unit 502 is used to input the harmonic signal into the target detection model for analysis to obtain the amplitude data and phase angle data of the harmonic signal, wherein the target detection model is a tree-shaped two-layer combination model trained using the amplitude data samples and phase angle data samples of the historical harmonic signals of the charging station, as well as the structural characteristics and operation characteristics of the charging station.

[0158] The determination unit 503 is used to determine the working mode of the charging station based on the amplitude data and the phase angle data.

[0159] Optionally, the device 500 for determining the working mode of the charging station may include: a first acquisition module, used to acquire structural characteristics and operating characteristics of the charging station, the structural characteristics are used to indicate the structure of the charging station, and the operating characteristics are used to indicate the operating rules of the charging station; a first determination module, used to determine the sampling parameters of the harmonic signal of the charging station based on the structural characteristics and the operating characteristics; a second acquisition module, used to acquire historical harmonic signals based on the sampling parameters; a training module, used to train the initial detection model using the historical harmonic signals to obtain a target detection model.

[0160] Optionally, the training module may include: a first determination submodule, used to determine the training nodes based on the historical harmonic signals, wherein the training nodes are used to indicate the number of node layers for training the initial detection model; a training submodule, used to train the initial detection model based on the amplitude data samples and phase angle data samples of the historical harmonic signals and the training nodes to obtain the initial detection results, wherein the initial detection results are used to indicate the results output by the initial detection model; and a second determination submodule, used to determine the target detection model based on the initial detection results.

[0161] Optionally, the second determination submodule can also be used to compare the initial detection result with the expected threshold to obtain a comparison result; in response to the comparison result being greater than or equal to the comparison result threshold, determine that the initial detection model corresponding to the initial detection result is the target detection model.

[0162] Optionally, the second determination submodule can also be used to determine an update parameter in response to the comparison result being less than a comparison result threshold, wherein the update parameter is used to update the initial detection model; and the initial detection model is updated using the update parameter to obtain a target detection model.

[0163] Optionally, the determination unit 503 may include: a second determination module, used to determine the charging state of the charging station based on the amplitude data and the phase angle data; and a third determination module, used to determine the working mode of the charging station based on the charging state.

[0164] In this embodiment, by detecting the harmonic signal of the charging station, the harmonic signal is input into the target detection model to obtain the amplitude data and phase angle data of the harmonic signal, thereby determining the working mode of the charging station according to the amplitude data and phase angle data. Since the target detection model of the present invention is a tree-shaped two-layer combination model obtained by training with the amplitude data samples and phase angle data samples of the historical harmonic signal of the charging station, as well as the structural characteristics and operation characteristics of the charging station, the target detection model can accurately analyze the harmonic signal characteristics, solve the technical problem that the working mode of the charging station cannot be determined due to the low harmonic detection efficiency of the charging station, and achieve the technical effect of improving the harmonic detection efficiency of the charging station, thereby determining the working mode of the charging station.

[0165] According to an embodiment of the present invention, a computer-readable storage medium is further provided. The storage medium includes a stored program, wherein the program executes the method for determining the working mode of the charging station in the method embodiment.

[0166] According to an embodiment of the present invention, a processor is further provided. The processor is used to run a program, wherein when the program is run, the method for determining the working mode of the charging station in the method embodiment is executed.

[0167] According to an embodiment of the present invention, a computer program product is further provided. The computer program product includes computer instructions. When the computer instructions are executed by a processor, the method for determining the working mode of the charging station in the method embodiment is implemented.

[0168] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0169] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0170] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0171] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0172] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0173] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent functional component, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software functional component, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc., which can store program codes.

[0174] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining the working mode of a charging station, characterized in that: include: Obtaining harmonic signals of the charging station to be detected; The harmonic signal is input into a target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal, wherein the target detection model is a tree-shaped two-layer combination model trained using amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as structural characteristics and operation characteristics of the charging station; An operating mode of the charging station is determined based on the amplitude data and the phase angle data.

2. The method according to claim 1, characterized in that The method further comprises: Acquire structural features and operating features of the charging station, wherein the structural features are used to indicate the structure of the charging station, and the operating features are used to indicate operating rules of the charging station; Determining sampling parameters of the harmonic signal of the charging station based on the structural characteristics and the operating characteristics; Based on the sampling parameters, acquiring the historical harmonic signal; The historical harmonic signal is used to train an initial detection model to obtain the target detection model.

3. The method according to claim 2, characterized in that The initial detection model is trained using the historical harmonic signal to obtain the target detection model, including: Based on the historical harmonic signal, determining a training node, wherein the training node is used to indicate the number of node layers for training the initial detection model; Based on the amplitude data samples and the phase angle data samples of the historical harmonic signal and the training nodes, the initial detection model is trained to obtain an initial detection result, wherein the initial detection result is used to indicate a result output by the initial detection model; Based on the initial detection result, the target detection model is determined.

4. The method according to claim 3, characterized in that Determining the target detection model based on the initial detection result includes: Comparing the initial detection result with the expected threshold value to obtain a comparison result; In response to the comparison result being greater than or equal to a comparison result threshold, determining that the initial detection model corresponding to the initial detection result is the target detection model.

5. The method according to claim 4, characterized in that The method further comprises: In response to the comparison result being less than the comparison result threshold, determining an update parameter, wherein the update parameter is used to update the initial detection model; The initial detection model is updated using the update parameters to obtain the target detection model.

6. The method according to any one of claims 1 to 5, characterized in that: Determining an operating mode of the charging station based on the amplitude data and the phase angle data includes: determining a charging state of the charging station based on the amplitude data and the phase angle data; Based on the charging status, the operating mode of the charging station is determined.

7. A device for determining the working mode of a charging station, characterized in that: include: An acquisition unit, used to acquire a harmonic signal of a charging station to be detected; an analysis unit, configured to input the harmonic signal into a target detection model for analysis to obtain amplitude data and phase angle data of the harmonic signal, wherein the target detection model is a tree-shaped two-layer combination model trained using amplitude data samples and phase angle data samples of historical harmonic signals of the charging station, as well as structural features and operation features of the charging station; A determination unit is used to determine an operating mode of the charging station based on the amplitude data and the phase angle data.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed by a processor, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

9. A processor, characterized in that: The processor is used to run a program, wherein the program executes the method according to any one of claims 1 to 6 when running.

10. A computer program product, characterized in that The computer program product comprises computer instructions, which implement the method according to any one of claims 1 to 6 when executed by a processor.

Images