Power factor improvement control method for charging system

Abstract

The invention relates to the technical field of charging systems, and particularly discloses a power factor improvement control method for a charging system, and the method comprises the steps: obtaining a time sequence of an input current and a time sequence of an input voltage; local time sequence correlation and time sequence feature significant aggregation of the input current and the input voltage are carried out by adopting a data processing and analysis algorithm based on artificial intelligence; therefore, a recommended duty ratio proportion value is intelligently obtained based on an interactive response relation between time sequence aggregation representation characteristics of the input current and the input voltage. Therefore, the control strategy can be adjusted in time according to the time sequence dynamic change of the input voltage and current, so that a high power factor can be kept even under the condition that the load condition changes, meanwhile, power oscillation and voltage fluctuation of a power grid caused by excessive compensation are prevented, and the stability, the reliability and the intelligent degree of the system are improved.

Description

Technical Field

[0001] The present application relates to the technical field of charging systems, and more specifically, to a power factor improvement control method for a charging system. Background Art

[0002] In the field of power electronics, power factor is an important indicator to measure the efficiency of electrical systems. Low power factor will lead to reduced grid efficiency, increased energy waste and unnecessary load on the grid. With the popularity of high-power loads such as electric vehicles, improving power factor has become particularly important. This is because low power factor not only increases the burden on the grid, but also leads to increased electricity bills.

[0003] However, the methods for improving power factor in traditional charging systems mainly rely on reactive power compensation devices, such as shunt capacitors or synchronous compensators. These methods reduce grid load and improve power factor by providing necessary reactive power. But in some cases, this method may over-compensate. That is, the over-compensation phenomenon in traditional power factor correction may cause grid power oscillation, increase voltage fluctuations, and even trigger grid protection mechanisms, affecting power supply reliability.

[0004] Therefore, an optimized power factor improvement control method for a charging system is desired. Summary of the invention

[0005] The present application provides a power factor improvement control method for a charging system, which timely adjusts the control strategy according to the timing dynamic changes of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, while preventing power oscillation and voltage fluctuation of the power grid caused by overcompensation, so as to improve the stability, reliability and intelligence of the system.

[0006] In a first aspect, a power factor improvement control method for a charging system is provided, comprising: obtaining a time series of input current and a time series of input voltage; inputting the time series of input current and the time series of input voltage into a sequence encoder to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors; inputting the sequence of input current local timing association feature vectors and the sequence of input voltage local timing association feature vectors into a feature aggregation network based on feature energy local significance gating to obtain an input current timing aggregation significant representation vector and an input voltage timing aggregation significant representation vector; inputting the input current timing aggregation significant representation vector and the input voltage timing aggregation significant representation vector into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupling timing reinforcement interaction global representation vector as an input current-input voltage decoupling timing reinforcement interaction global representation feature; and obtaining a control instruction based on the input current-input voltage decoupling timing reinforcement interaction global representation feature, wherein the control instruction is used to represent a recommended duty cycle ratio value.

[0007] In one possible implementation, the time series of the input current and the time series of the input voltage are input into a sequence encoder to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors, including: inputting the time series of the input current and the time series of the input voltage into a sequence encoder based on a 1D-CNN model to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors.

[0008] In a possible implementation, the sequence of the input current local timing association feature vectors and the sequence of the input voltage local timing association feature vectors are input into a feature aggregation network based on feature energy local significance gating to obtain an input current timing aggregation significant representation vector and an input voltage timing aggregation significant representation vector, including: based on the maximum value, minimum value, mean value and variance of each input current local timing association feature vector in the sequence of the input current local timing association feature vectors, determining the characteristic energy level coefficients of each input current local timing association feature vector to obtain a set of input current local timing association feature energy level coefficients; determining the scale of the local neighborhood, calculating the input current local The local neighborhood significance descriptors of each input current local timing correlation feature energy level coefficient in the set of local timing correlation feature energy level coefficients are obtained to obtain a set of input current local timing correlation feature energy level significance descriptors; the set of input current local timing correlation feature energy level significance descriptors is input into a mask module based on a gating function to obtain a set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors; the set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors is used as a sequence of weights, and the position-weighted sum of the sequence of input current local timing correlation feature vectors is calculated to obtain the input current timing aggregated significant representation vector.

[0009] In a possible implementation, based on the maximum value, minimum value, mean value and variance of each input current local timing association feature vector in the sequence of the input current local timing association feature vector, the characteristic energy level coefficients of each input current local timing association feature vector are determined to obtain a set of input current local timing association feature energy level coefficients, including: respectively extracting the maximum value and the minimum value of the input current local timing association feature vector to obtain the maximum value of the input current local timing association feature and the minimum value of the input current local timing association feature; respectively calculating the mean value and the variance of the input current local timing association feature vector to obtain the mean value of the input current local timing association feature and the variance of the input current local timing association feature; adding the input current local timing association feature variance and the preset hyperparameter to obtain the input The first energy level coefficient of the local timing correlation characteristic of the current; calculating the square of the difference between the maximum value of the local timing correlation characteristic of the input current and the mean value of the local timing correlation characteristic of the input current to obtain the maximum difference value of the local timing correlation of the input current; calculating the square of the difference between the minimum value of the local timing correlation characteristic of the input current and the mean value of the local timing correlation characteristic of the input current to obtain the minimum difference value of the local timing correlation of the input current; adding the maximum difference value of the local timing correlation of the input current, the minimum difference value of the local timing correlation of the input current and the preset hyperparameter to obtain the second energy level coefficient of the local timing correlation characteristic of the input current; dividing the first energy level coefficient of the local timing correlation characteristic of the input current by the second energy level coefficient of the local timing correlation characteristic of the input current to obtain the energy level coefficient of the local timing correlation characteristic of the input current.

[0010] In a possible implementation, the set of input current local timing correlation feature energy level significance descriptors is input into a mask module based on a gating function to obtain a set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors, including: performing normalization processing based on a Sigmoid function on each input current local timing correlation feature energy level significance descriptor in the set of input current local timing correlation feature energy level significance descriptors to obtain a set of normalized input current local timing correlation feature energy level significance descriptors; and inputting each normalized input current local timing correlation feature energy level significance descriptor in the set of normalized input current local timing correlation feature energy level significance descriptors into a gating function for masking processing to obtain the set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors.

[0011] In a possible implementation, the input current time series aggregation significant representation vector and the input voltage time series aggregation significant representation vector are input into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupled time series reinforcement interaction global representation vector, including: performing feature decoupling on the input current time series aggregation significant representation vector and the input voltage time series aggregation significant representation vector to obtain a sequence of input current local time series aggregation significant representation vectors and a sequence of input voltage local time series aggregation significant representation vectors; calculating the local semantic association interaction vector between each group of corresponding input current local time series aggregation significant representation vectors and input voltage local time series aggregation significant representation vectors in the sequence of input current local time series aggregation significant representation vectors and the sequence of input voltage local time series aggregation significant representation vectors to obtain A sequence of input current-input voltage local temporal semantic association interaction vectors; calculating the fine-grained attention weights of each input current-input voltage local temporal semantic association interaction vector in the sequence of input current-input voltage local temporal semantic association interaction vectors to obtain a sequence of input current-input voltage local temporal fine-grained attention weights; inputting the sequence of input current-input voltage local temporal fine-grained attention weights into a random deactivation module to obtain a sequence of input current-input voltage local temporal fine-grained attention weights after deactivation; using the sequence of input current-input voltage local temporal fine-grained attention weights after deactivation as a weight distribution, calculating the weighted sum of the sequence of input current-input voltage local temporal semantic association interaction vectors to obtain the input current-input voltage decoupled temporal reinforcement interaction global representation vector.

[0012] In a possible implementation, the local semantic association interaction vector between each group of corresponding input current local timing aggregation significant representation vectors and input voltage local timing aggregation significant representation vectors in the sequence of the input current local timing aggregation significant representation vectors and the sequence of the input voltage local timing aggregation significant representation vectors is calculated to obtain a sequence of input current-input voltage local timing semantic association interaction vectors, including: respectively calculating the position difference, position dot multiplication and position addition between the input current local timing aggregation significant representation vector and the input voltage local timing aggregation significant representation vector to obtain the input current-input voltage local timing aggregation significant difference vector, Input current-input voltage local timing aggregation significant dot product vector and input current-input voltage local timing aggregation significant sum vector; cascade the input current-input voltage local timing aggregation significant difference vector, the input current-input voltage local timing aggregation significant dot product vector and the input current-input voltage local timing aggregation significant sum vector, and then perform one-dimensional convolution encoding to obtain an input current-input voltage local timing aggregation significant multi-dimensional fusion vector; perform local window-based maximum pooling processing on the input current-input voltage local timing aggregation significant multi-dimensional fusion vector to obtain the input current-input voltage local timing semantic association interaction vector.

[0013] In a possible implementation, the fine-grained attention weights of each input current-input voltage local timing semantic association interaction vector in the sequence of input current-input voltage local timing semantic association interaction vectors are calculated to obtain a sequence of input current-input voltage local timing fine-grained attention weights, including: multiplying each input current-input voltage local timing semantic association interaction vector in the sequence of input current-input voltage local timing semantic association interaction vectors with a weight matrix, and then adding the obtained vector with a bias vector by position to obtain a sequence of input current-input voltage local timing semantic association interaction modulation vectors; inputting the sequence of input current-input voltage local timing semantic association interaction modulation vectors into a tanh function for activation processing to obtain a sequence of input current-input voltage local timing semantic association interaction activation vectors; multiplying each input current-input voltage local timing semantic association interaction activation vector in the sequence of input current-input voltage local timing semantic association interaction activation vectors with a weight coefficient vector to obtain a sequence of input current-input voltage local timing fine-grained attention weights.

[0014] In a possible implementation, a control instruction is obtained based on the input current-input voltage decoupling timing enhancement interaction global representation feature, including: inputting the current-input voltage decoupling timing enhancement interaction global representation vector into a decoder-based power factor improvement control module to obtain the control instruction.

[0015] In a possible implementation, the current-input voltage decoupling timing enhancement interaction global representation vector is input into a decoder-based power factor improvement control module to obtain the control instruction, including: using a decoder to decode and regress the current-input voltage decoupling timing enhancement interaction global representation vector according to the following formula to obtain a control instruction for representing a recommended duty cycle ratio value;

[0016] Wherein, the formula is: Wherein, X is the global representation vector of the current-input voltage decoupling timing enhancement interaction, Y is the control instruction for representing the recommended duty cycle ratio value, W is the weight matrix, Represents matrix multiplication.

[0017] The present application provides a power factor improvement control method for a charging system, which obtains the time series of input current and the time series of input voltage, and uses artificial intelligence-based data processing and analysis algorithms to perform local timing correlation and significant aggregation of timing characteristics of the input current and the input voltage, so as to intelligently obtain the recommended duty ratio value based on the interactive response relationship between the timing aggregation representation characteristics of the input current and the input voltage. In this way, the control strategy can be adjusted in time according to the dynamic changes in the timing of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, while preventing power grid power oscillation and voltage fluctuations caused by overcompensation, so as to improve the stability, reliability and intelligence of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic flow chart of a power factor improvement control method for a charging system according to an embodiment of the present application.

[0019] Figure 2 Schematic diagram of data flow of a power factor improvement control method for a charging system according to an embodiment of the present application.

[0020] Figure 3 A circuit diagram of a power factor improvement control method for a charging system according to another embodiment of the present application.

[0021] Figure 4 It is a schematic block diagram of a power factor improvement control system for a charging system according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0023] The present application proposes an optimized power factor improvement control method for a charging system, which mainly adds a PFC (power factor correction) circuit after the rectifier bridge to improve the power factor during the rectification process, so as to realize the impact of the charging system on the power grid during the charging process. The PFC circuit can adjust the working state of the inductor and capacitor according to the input voltage and input current, so that the phase of the input voltage and the input current are consistent, thereby improving the power factor of the power supply, achieving energy saving and reducing the impact of electrical equipment on the power grid. In detail, the controller in the PFC circuit monitors the phase and amplitude of the input voltage and current in real time, and by adjusting the switching frequency and duty cycle of the power electronic device (switching device), it can dynamically adjust the working state of the inductor and capacitor, thereby controlling the waveform of the output current. Specifically, the duty cycle is the ratio of the conduction time of the power electronic device to the entire cycle time. By adjusting the duty cycle, the controller can change the charge and discharge rate of the current in the inductor, thereby affecting the output voltage and current, thereby ensuring that the circuit can quickly adapt to changes in the input voltage and load to maintain a high power factor. In particular, the power factor is the ratio of active power to apparent power in an AC circuit, reflecting how effectively the circuit utilizes electrical energy, while a low power factor means low grid transmission efficiency, which may lead to energy waste and increased grid load.

[0024] However, since the duty cycle control of traditional controllers is usually based on static parameters and predefined control rules, they are slow to respond to rapid changes in grid and load conditions and are usually unable to analyze and adapt to the dynamic characteristics of the grid in real time, such as sudden increases in load or rapid fluctuations in voltage, which affects grid stability.

[0025] Therefore, in response to the above technical problems, the technical concept of the present application is to obtain the time series of the input current and the time series of the input voltage, and use artificial intelligence-based data processing and analysis algorithms to perform local timing correlation and significant aggregation of timing characteristics of the input current and the input voltage, so as to intelligently obtain the recommended duty ratio value based on the interactive response relationship between the timing aggregation representation characteristics of the input current and the input voltage. In this way, the control strategy can be adjusted in time according to the dynamic changes in the timing of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, while preventing power oscillations and voltage fluctuations caused by overcompensation in the power grid, so as to improve the stability, reliability and intelligence of the system.

[0026] Figure 1 It is a schematic flow chart of a power factor improvement control method for a charging system according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the power factor improvement control method for a charging system according to an embodiment of the present application. Figure 1 and Figure 2 As shown, the power factor improvement control method for the charging system includes: S1, obtaining the time series of input current and the time series of input voltage; S2, inputting the time series of input current and the time series of input voltage into a sequence encoder to obtain a sequence of input current local time series associated feature vectors and a sequence of input voltage local time series associated feature vectors; S3, inputting the sequence of input current local time series associated feature vectors and the sequence of input voltage local time series associated feature vectors into a feature aggregation network based on feature energy local significance gating to obtain an input current time series aggregate significant representation vector and an input voltage time series aggregate significant representation vector; S4, inputting the input current time series aggregate significant representation vector and the input voltage time series aggregate significant representation vector into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupling time series enhanced interaction global representation vector as an input current-input voltage decoupling time series enhanced interaction global representation feature; S5, based on the input current-input voltage decoupling time series enhanced interaction global representation feature, obtaining a control instruction, the control instruction is used to represent a recommended duty cycle ratio value.

[0027] In the above optimized power factor improvement control method for a charging system, the step S1 is to obtain a time series of an input current and a time series of an input voltage.

[0028] In the above-mentioned optimized power factor improvement control method for a charging system, S2, the time series of the input current and the time series of the input voltage are input into a sequence encoder to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors. It should be understood that, considering that the time series of the input current and the time series of the input voltage have a local correlation relationship within a local time segment, and the 1D-CNN model is good at extracting local features and local dependency patterns from one-dimensional sequence data, which helps to identify the association between the local timing of current and voltage. Based on this, in the technical solution of the present application, the time series of the input current and the time series of the input voltage are input into a sequence encoder based on a 1D-CNN model to capture and extract the timing association dependency within each local time period, and obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors. In particular, the essence of the sequence encoder based on the 1D-CNN model is a 1D-CNN model.

[0029] Optionally, in one embodiment of the present application, the time series of the input current and the time series of the input voltage are input into a sequence encoder to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors, including: inputting the time series of the input current and the time series of the input voltage into a sequence encoder based on a 1D-CNN model to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors.

[0030] In the above-mentioned optimized power factor improvement control method for a charging system, S3, the sequence of the input current local time series associated feature vectors and the sequence of the input voltage local time series associated feature vectors are input into a feature aggregation network based on local significant gating of feature energy to obtain an input current time series aggregate significant representation vector and an input voltage time series aggregate significant representation vector. It should be understood that, considering that the sequence of the input current local time series associated feature vectors and the sequence of the input voltage local time series associated feature vectors have different degrees of influence and importance on the entire time series in the local time period. In order to selectively enhance the features of different local time points, so as to obtain comprehensive characterization features of the entire time domain, in the technical solution of the present application, the sequence of the input current local time series associated feature vectors and the sequence of the input voltage local time series associated feature vectors are input into a feature aggregation network based on local significant gating of feature energy to obtain an input current time series aggregate significant representation vector and an input voltage time series aggregate significant representation vector. That is, the feature aggregation network based on the local significant gating of feature energy is a kind of key feature that focuses on each local time series feature and makes these important features significant, thereby improving the expression and processing capabilities of the entire time series. Here, the sequence of the input current local time series associated feature vector is taken as an example for explanation. In detail, the feature energy level coefficient of each current local time series associated feature vector is first calculated based on the maximum value, minimum value, mean and variance of each current local time series associated feature vector to quantify the energy distribution of the feature. Next, the range of the local neighborhood is determined, and the feature significance descriptor in its local neighborhood is calculated for each feature energy level coefficient, with the aim of identifying the energy distribution pattern of each current local time series associated feature within the local range. Then, these significant descriptors are sent to the mask module based on the gating mechanism, the gating function is used for nonlinear probability conversion, and the mask function is applied for adaptive feature selection to dynamically adjust the significance of the current local time series associated feature. Afterwards, these converted probabilistic saliency descriptors are used as weights to perform weighted summation on the sequence of input current local temporal correlation feature vectors to obtain a comprehensive input current temporal aggregate saliency representation vector, thereby enhancing the local saliency features of each input current.

[0031] Optionally, in one embodiment of the present application, the sequence of the input current local timing association feature vectors and the sequence of the input voltage local timing association feature vectors are input into a feature aggregation network based on feature energy local significance gating to obtain an input current timing aggregation significant representation vector and an input voltage timing aggregation significant representation vector, including: based on the maximum value, minimum value, mean value and variance of each input current local timing association feature vector in the sequence of the input current local timing association feature vectors, determining the characteristic energy level coefficients of each input current local timing association feature vector to obtain a set of input current local timing association feature energy level coefficients; determining the scale of the local neighborhood, calculating the input current The local neighborhood significance descriptors of each input current local timing correlation feature energy level coefficient in the set of flow local timing correlation feature energy level coefficients are obtained to obtain a set of input current local timing correlation feature energy level significance descriptors; the set of input current local timing correlation feature energy level significance descriptors is input into a mask module based on a gating function to obtain a set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors; the set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors is used as a sequence of weights, and the position-weighted sum of the sequence of input current local timing correlation feature vectors is calculated to obtain the input current timing aggregated significant representation vector.

[0032] Optionally, in one embodiment of the present application, based on the maximum value, minimum value, mean value and variance of each input current local timing association feature vector in the sequence of the input current local timing association feature vector, the characteristic energy level coefficients of each input current local timing association feature vector are determined to obtain a set of input current local timing association feature energy level coefficients, including: respectively extracting the maximum value and the minimum value of the input current local timing association feature vector to obtain the maximum value of the input current local timing association feature and the minimum value of the input current local timing association feature; respectively calculating the mean value and the variance of the input current local timing association feature vector to obtain the mean value of the input current local timing association feature and the variance of the input current local timing association feature; adding the input current local timing association feature variance and a preset hyperparameter to obtain Input current local timing correlation characteristic first energy level coefficient; calculate the square of the difference between the maximum value of the input current local timing correlation characteristic and the mean value of the input current local timing correlation characteristic to obtain the input current local timing correlation maximum difference value; calculate the square of the difference between the minimum value of the input current local timing correlation characteristic and the mean value of the input current local timing correlation characteristic to obtain the input current local timing correlation minimum difference value; add the input current local timing correlation maximum difference value, the input current local timing correlation minimum difference value and the preset hyperparameter to obtain the input current local timing correlation characteristic second energy level coefficient; divide the input current local timing correlation characteristic first energy level coefficient by the input current local timing correlation characteristic second energy level coefficient to obtain the input current local timing correlation characteristic energy level coefficient.

[0033] Optionally, in one embodiment of the present application, the set of input current local timing association feature energy level significance descriptors is input into a mask module based on a gating function to obtain a set of mask-gated probabilistic input current local timing association feature energy level significance descriptors, including: performing normalization processing based on a Sigmoid function on each input current local timing association feature energy level significance descriptor in the set of input current local timing association feature energy level significance descriptors to obtain a set of normalized input current local timing association feature energy level significance descriptors; and inputting each normalized input current local timing association feature energy level significance descriptor in the set of normalized input current local timing association feature energy level significance descriptors into a gating function for masking processing to obtain the set of mask-gated probabilistic input current local timing association feature energy level significance descriptors.

[0034] Specifically, in one embodiment of the present application, the sequence of the input current local time series associated feature vectors is input into a feature aggregation network based on feature energy local significance gating, and processed using the following feature aggregation formula to obtain the input current time series aggregated significant representation vector; wherein the feature aggregation formula is:

[0035] X={x1,x2,...,x k ,...,x n}

[0036]

[0037] w i =sigmoid(e ci )

[0038] w si =mask(w i )

[0039]

[0040] Where X is the sequence of the local time series correlation characteristic vectors of the input current, x1, x2, x k 、x n represents the first, second, kth, and nth input current local time series correlation feature vectors in the sequence of the input current local time series correlation feature vectors, respectively, n represents the number of feature vectors in the sequence of the input current local time series correlation feature vectors, μ i and σ 2 i are the mean and variance of the local time series correlation feature vector of the i-th input current, max(x i ) and min(x i ) are the maximum and minimum values ​​in the local time series correlation feature vector of the i-th input current, ε is the preset hyperparameter, e i is the characteristic energy level coefficient of the local time series correlation characteristic vector of the i-th input current, e q is the characteristic energy level coefficient of the qth local time series correlation characteristic vector of the input current, Num r Indicates that the i The number of characteristic energy coefficients in the local neighborhood centered at ci is the input current local time series correlation feature energy level significance descriptor corresponding to the i-th input current local time series correlation feature vector, sigmoid(·) is the sigmoid function, w iis the ith normalized input current local time series correlation feature energy level significant descriptor in the set of normalized input current local time series correlation feature energy level significant descriptors, mask(·) is the masking process, w si is the ith mask-gated probabilistic input current local time series correlation feature energy level significant descriptor in the set of mask-gated probabilistic input current local time series correlation feature energy level significant descriptors, θ is a predetermined threshold, and f1 is the input current time series aggregate significant representation vector. Similarly, the encoding method of the sequence of the input voltage local time series correlation feature vector is as shown in the above encoding.

[0041] In the above-mentioned optimized power factor improvement control method for charging system, S4, the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector are input into the local fine-grained sequence interaction response analysis network based on feature decoupling to obtain the input current-input voltage decoupling time series enhanced interaction global representation vector as the input current-input voltage decoupling time series enhanced interaction global representation feature. It should be understood that, considering that there is a fine-grained dynamic relationship and interaction between the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector in the local time series, the subtle interactive correlation relationship between the two plays an important role in the control of the duty cycle in the PFC circuit. Therefore, in the technical solution of the present application, the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector are input into the local fine-grained sequence interaction response analysis network based on feature decoupling to obtain the input current-input voltage decoupling time series enhanced interaction global representation vector. That is, the local fine-grained sequence interaction response analysis network based on feature decoupling improves the global interaction representation by the local fine-grained interaction in the feature. Specifically, firstly, the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector are feature decoupled to divide an aggregated representation feature into a sequence of features containing multiple local time series, which provides a basis for the subsequent more refined interaction between the input current and the input voltage in the local time series. Then, each group of features after feature decoupling is subjected to local semantic interaction processing to capture and refine the interaction dependency between the two in the local time series, so as to provide a quantitative perspective to observe how the features interact with each other in the local range to obtain a sequence of input current-input voltage local time series semantic association interaction vectors. Then, fine-grained attention weights are calculated for each vector in the sequence of these local semantic association interaction vectors to make the model more focused on the key time series interaction part. Then, the sequence of these fine-grained attention weights is subjected to weight sparse processing through a random dropout module. In particular, the random dropout reduces the overfitting of the model to the training data by randomly "closing" some weights, while ensuring that the information of key feature interactions is maintained. Finally, the sequence of local semantic association interaction vectors is weightedly summed using the inactivated fine-grained attention weight sequence to combine the local interaction information and attention weight values ​​of the input current and voltage to obtain the global representation vector of the input current-input voltage decoupled temporal reinforcement interaction.

[0042] Optionally, in one embodiment of the present application, the input current time series aggregation significant representation vector and the input voltage time series aggregation significant representation vector are input into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupled time series reinforcement interaction global representation vector, including: performing feature decoupling on the input current time series aggregation significant representation vector and the input voltage time series aggregation significant representation vector to obtain a sequence of input current local time series aggregation significant representation vectors and a sequence of input voltage local time series aggregation significant representation vectors; calculating a local semantic association interaction vector between each group of corresponding input current local time series aggregation significant representation vectors and input voltage local time series aggregation significant representation vectors in the sequence of input current local time series aggregation significant representation vectors and the sequence of input voltage local time series aggregation significant representation vectors. Obtain a sequence of input current-input voltage local temporal semantic association interaction vectors; calculate the fine-grained attention weights of each input current-input voltage local temporal semantic association interaction vector in the sequence of input current-input voltage local temporal semantic association interaction vectors to obtain a sequence of input current-input voltage local temporal fine-grained attention weights; input the sequence of input current-input voltage local temporal fine-grained attention weights into a random deactivation module to obtain a sequence of input current-input voltage local temporal fine-grained attention weights after deactivation; use the sequence of input current-input voltage local temporal fine-grained attention weights after deactivation as a weight distribution, calculate the weighted sum of the sequence of input current-input voltage local temporal semantic association interaction vectors to obtain the input current-input voltage decoupled temporal reinforcement interaction global representation vector.

[0043] Optionally, in one embodiment of the present application, the local semantic association interaction vector between each group of corresponding input current local timing aggregation significant representation vectors and input voltage local timing aggregation significant representation vectors in the sequence of the input current local timing aggregation significant representation vectors and the sequence of the input voltage local timing aggregation significant representation vectors is calculated to obtain a sequence of input current-input voltage local timing semantic association interaction vectors, including: respectively calculating the position difference, position dot multiplication and position addition between the input current local timing aggregation significant representation vector and the input voltage local timing aggregation significant representation vector to obtain the input current-input voltage local timing aggregation significant difference vector. The input current-input voltage local timing aggregation significant dot product vector and the input current-input voltage local timing aggregation significant sum vector are cascaded and then one-dimensional convolutional coding is performed to obtain an input current-input voltage local timing aggregation significant multi-dimensional fusion vector; the input current-input voltage local timing aggregation significant multi-dimensional fusion vector is subjected to local window-based maximum pooling processing to obtain the input current-input voltage local timing semantic association interaction vector.

[0044] Optionally, in one embodiment of the present application, the fine-grained attention weights of each input current-input voltage local timing semantic association interaction vector in the sequence of input current-input voltage local timing semantic association interaction vectors are calculated to obtain a sequence of input current-input voltage local timing fine-grained attention weights, including: multiplying each input current-input voltage local timing semantic association interaction vector in the sequence of input current-input voltage local timing semantic association interaction vectors with a weight matrix, and then adding the obtained vector to the bias vector by position to obtain a sequence of input current-input voltage local timing semantic association interaction modulation vectors; inputting the sequence of input current-input voltage local timing semantic association interaction modulation vectors into a tanh function for activation processing to obtain a sequence of input current-input voltage local timing semantic association interaction activation vectors; multiplying each input current-input voltage local timing semantic association interaction activation vector in the sequence of input current-input voltage local timing semantic association interaction activation vectors with a weight coefficient vector to obtain a sequence of input current-input voltage local timing fine-grained attention weights.

[0045] Specifically, in one embodiment of the present application, the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector are input into a local fine-grained sequence interaction response analysis network based on feature decoupling, and processed with the following interaction response to obtain the input current-input voltage decoupling time series reinforcement interaction global representation vector; wherein the interaction response formula is:

[0046] Decouple(f1)={f 11 ,f 12 ,...,f 1i ,...,f 1n}

[0047] Decouple(f2)={f 21 ,f 22 ,...,f 2i ,...,f 2n}

[0048]

[0049] E i =v T tanh(W r f int (f 1i ,f 2i )+b r )

[0050] W i = dropout(E i )

[0051] V int =∑ i f int (f 1i ,f 2i )·W i

[0052] Wherein, f1 and f2 are respectively the input current time series aggregate significant representation vector and the input voltage time series aggregate significant representation vector, Decouple(·) is the feature decoupling operation, and f 1n and f 2n are the nth input current local time series aggregation significant representation vector and the nth input voltage local time series aggregation significant representation vector in the sequence of the input current local time series aggregation significant representation vector and the sequence of the input voltage local time series aggregation significant representation vector, respectively, n represents the number of feature vectors in the sequence of the input current local time series aggregation significant representation vector and the sequence of the input voltage local time series aggregation significant representation vector, f 1i and f 2iare respectively the Ith local time series aggregated significant representation vector of the input current and the Ith local time series aggregated significant representation vector of the input voltage, ⊙ and are position difference, position dot multiplication and position addition respectively, [·,·,·] is cascade processing, conv1D(·) is a one-dimensional convolutional encoding operation, MaxPool(·) is a maximum pooling operation, and f int (f 1i ,f 2i ) is the i-th input current-input voltage local time sequence semantic association interaction vector in the sequence of the input current-input voltage local time sequence semantic association interaction vector, W r is the weight matrix, b r is the bias vector, tanh(·) is the hyperbolic tangent function, v T is the weight coefficient vector, E i is the i-th input current-input voltage local time series fine-grained attention weight in the sequence of the input current-input voltage local time series fine-grained attention weights, dropout(·) represents the deactivation process, W i is the i-th input current-input voltage local time series fine-grained attention weight after deactivation processing in the sequence of the input current-input voltage local time series fine-grained attention weight after deactivation processing, V int is the global representation vector of the input current-input voltage decoupling timing enhancement interaction.

[0053] In the above-mentioned optimized power factor improvement control method for a charging system, S5, based on the input current-input voltage decoupling timing enhancement interaction global representation feature, a control instruction is obtained, and the control instruction is used to represent the recommended duty cycle ratio value. Optionally, in one embodiment of the present application, based on the input current-input voltage decoupling timing enhancement interaction global representation feature, a control instruction is obtained, including: inputting the current-input voltage decoupling timing enhancement interaction global representation vector into a decoder-based power factor improvement control module to obtain the control instruction. That is, the input current-input voltage decoupling timing enhancement interaction global representation feature obtained by fine-grained interaction of the input current timing aggregation significant representation vector and the input voltage timing aggregation significant representation vector is decoded and processed, so as to intelligently obtain the recommended duty cycle ratio value. It should be understood that the current-input voltage decoupling timing enhancement interaction global representation vector contains important significant interaction information of the input current and input voltage at the local and global levels, and the switching frequency and duty cycle of the power electronic device can be adjusted by inputting the decoder-based power factor improvement control module. That is to say, by adjusting the duty cycle, the control module can optimize the phase relationship between the input current and the input voltage, increase the proportion of active power, and thus improve the power factor. In this way, the control strategy can be dynamically adjusted according to the timing changes of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, while preventing power oscillations and voltage fluctuations caused by overcompensation in the power grid, so as to improve the stability, reliability and intelligence of the system. In this way, the control strategy can be timely adjusted according to the timing dynamic changes of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, while preventing power oscillations and voltage fluctuations caused by overcompensation in the power grid, so as to improve the stability, reliability and intelligence of the system.

[0054] Optionally, in one embodiment of the present application, the current-input voltage decoupling timing enhancement interaction global representation vector is input into a decoder-based power factor improvement control module to obtain the control instruction, including: using a decoder to decode and regress the current-input voltage decoupling timing enhancement interaction global representation vector using the following formula to obtain a control instruction for representing a recommended duty cycle ratio value;

[0055] Wherein, the formula is: Wherein, X is the global representation vector of the current-input voltage decoupling timing enhancement interaction, Y is the control instruction for representing the recommended duty cycle ratio value, W is the weight matrix, Represents matrix multiplication.

[0056] In a preferred embodiment of the present application, inputting the current-input voltage decoupling timing enhancement interaction global representation vector into a decoder-based power factor improvement control module to obtain a control instruction comprises:

[0057] Subtract the zero norm of the current-input voltage decoupling timing reinforcement interaction global representation vector from the length of the current-input voltage decoupling timing reinforcement interaction global representation vector to obtain a current-input voltage decoupling timing reinforcement interaction global isolated representation value;

[0058] Calculate the logarithm of the square of the current-input voltage decoupling timing enhancement interaction global isolated representation value and the sum of the current-input voltage decoupling timing enhancement interaction global isolated representation value with base 2 to obtain the current-input voltage decoupling timing enhancement interaction global information order value;

[0059] Calculate a power function with each eigenvalue of the current-input voltage decoupling timing enhancement interaction global representation vector as the base, the difference of the current-input voltage decoupling timing enhancement interaction global isolated representation value minus one as the exponent, and perform point multiplication with the current-input voltage decoupling timing enhancement interaction global information order value to obtain the current-input voltage decoupling timing enhancement interaction global leading vector;

[0060] The current-input voltage decoupling timing enhancement interaction global representation vector is multiplied by the difference between the current-input voltage decoupling timing enhancement interaction global isolated representation value minus one, and then multiplied by the reciprocal of the current-input voltage decoupling timing enhancement interaction global isolated representation value to obtain a current-input voltage decoupling timing enhancement interaction global field constraint vector;

[0061] Calculate an exponential function with a natural constant as the base and each eigenvalue of the current-input voltage decoupling timing enhancement interaction global field constraint vector as the exponent to obtain a current-input voltage decoupling timing enhancement interaction global field bias vector;

[0062] Performing point addition of the current-input voltage decoupling timing enhancement interaction global pioneer vector and the current-input voltage decoupling timing enhancement interaction global field bias vector to obtain an optimized current-input voltage decoupling timing enhancement interaction global representation vector;

[0063] The optimized current-input voltage decoupling timing enhancement interaction global representation vector is input into a decoder-based power factor improvement control module to obtain a control instruction.

[0064] Specifically, the optimization process of the current-input voltage decoupling timing enhancement interaction global representation vector, for example, denoted as V, is expressed as:

[0065]

[0066] n=L-||V||0

[0067] Wherein, V represents the global representation vector of the current-input voltage decoupling timing enhancement interaction, ||·||0 represents the zero norm of the vector, L represents the length of the global representation vector of the current-input voltage decoupling timing enhancement interaction, n represents the global isolated representation value of the current-input voltage decoupling timing enhancement interaction, ⊙ is the point multiplication by position, is added by position, (·) ⊙(n-1) A power function with the difference of the current-input voltage decoupling timing enhancement interaction global isolated representation value minus one as the exponent of the characteristic value of each position in the calculated characteristic vector is represented, and V' represents the optimized current-input voltage decoupling timing enhancement interaction global representation vector.

[0068] Here, the input current timing aggregate significant representation vector and the input voltage timing aggregate significant representation vector respectively represent the local time domain one-dimensional timing correlation characteristics of the input current and input voltage based on the global time domain aggregation characteristics of the local time domain timing distribution energy. Therefore, when performing local fine-grained sequence interaction response analysis based on feature decoupling, the timing sequence response analysis offset caused by the local fine-grained interaction difference of the timing feature distribution caused by the source timing distribution is taken into account. It is expected to further improve the full time domain interaction response aggregation regression comprehensibility of the current-input voltage decoupling timing reinforcement interaction global feature vector, thereby improving the accuracy of the control instructions obtained by the power factor improvement control module based on the decoder.

[0069] Therefore, for the high-dimensional feature manifold of the current-input voltage decoupling timing reinforcement interaction global representation vector, the eigenvalues ​​of the feature set are used as the vector field representation of the aggregation dimension, and the superposition value of the vector field of the current-input voltage decoupling timing reinforcement interaction global representation vector at the isolated zero position is used as the order information to fix the local position of the eigenvalues ​​of its feature set, and a bias for the reversibility of the feature regression distribution field of the current-input voltage decoupling timing reinforcement interaction global representation vector is added as a reward to achieve the mapping target tracking of the regression distribution of the current-input voltage decoupling timing reinforcement interaction global representation vector for the eigenvalue position, so that the feature set of the current-input voltage decoupling timing reinforcement interaction global representation vector perceives the mapping migration to the aggregate distribution, thereby improving the accuracy of the control instructions obtained by the current-input voltage decoupling timing reinforcement interaction global feature vector input to the decoder-based power factor improvement control module by improving the full-time domain interactive response aggregate regression comprehensibility of the current-input voltage decoupling timing reinforcement interaction global representation vector. In this way, the control strategy can be adjusted in time according to the dynamic changes in the timing of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, while preventing grid power oscillations and voltage fluctuations caused by overcompensation, thereby improving the stability, reliability and intelligence of the system.

[0070] In another embodiment of the present application, a power factor improvement control method for a charging system is also provided, which mainly adds a PFC circuit after the rectifier bridge to improve the power factor during the rectification process, so as to realize the impact of the charging system on the power grid during the charging process. The PFC circuit can adjust the working state of the inductor and the capacitor according to the input voltage and the input current to make the phase of the input voltage and the input current consistent, thereby improving the power factor of the power supply, achieving the effect of energy saving and reducing the impact of electrical equipment on the power grid. Figure 3 10 is a rectifier bridge, 20 is a PFC circuit, 30 is a DC conversion circuit, 40 is the control method, 401 is a PFC driver, 402 is a current detection unit, 403 is a comparison unit, and 404 is a controller.

[0071] In summary, the power factor improvement control method for the charging system according to the embodiment of the present application is explained, which obtains the time series of the input current and the time series of the input voltage, and adopts the data processing and analysis algorithm based on artificial intelligence to perform the local timing correlation and timing feature significant aggregation of the input current and the input voltage, so as to intelligently obtain the recommended duty ratio value based on the interactive response relationship between the timing aggregation representation characteristics of the input current and the input voltage. In this way, the control strategy can be adjusted in time according to the dynamic changes of the timing of the input voltage and current, so that a high power factor can be maintained even when the load conditions change, and the power oscillation and voltage fluctuation of the power grid caused by overcompensation can be prevented, so as to improve the stability, reliability and intelligence of the system.

[0072] Figure 4 Schematic block diagram of a power factor improvement control system for a charging system according to an embodiment of the present application. Figure 4 As shown, a power factor improvement control system 110 for a charging system includes: an input current-input voltage acquisition module 110, which is used to acquire a time series of an input current and a time series of an input voltage; an input current-input voltage timing encoding module 120, which is used to input the time series of the input current and the time series of the input voltage into a sequence encoder to obtain a sequence of input current local timing correlation feature vectors and a sequence of input voltage local timing correlation feature vectors; an input current-input voltage feature aggregation module 130, which is used to input the sequence of input current local timing correlation feature vectors and the sequence of input voltage local timing correlation feature vectors into a feature energy local significance gating-based ... A feature aggregation network is used to obtain an input current timing aggregation significant representation vector and an input voltage timing aggregation significant representation vector; an input current-input voltage feature decoupling module 140 is used to input the input current timing aggregation significant representation vector and the input voltage timing aggregation significant representation vector into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupling timing reinforcement interaction global representation vector as an input current-input voltage decoupling timing reinforcement interaction global representation feature; a control instruction generation module 150 is used to obtain a control instruction based on the input current-input voltage decoupling timing reinforcement interaction global representation feature, and the control instruction is used to represent a recommended duty cycle ratio value.

[0073] Here, those skilled in the art will appreciate that the specific operations of the various modules in the power factor improvement control system for the charging system have been described in the above reference. Figure 1 to Figure 2 The power factor improvement control method for a charging system has been described in detail, and therefore, its repeated description will be omitted.

[0074] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0075] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the module division is only a logical function division, and there may be other division methods in actual implementation. The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0076] In addition, each functional module 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 hardware plus software functional modules.

[0077] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0078] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units stated in the device claim can also be implemented by one unit through software or hardware.

[0079] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A power factor improvement control method for a charging system, characterized in that: include: Obtain the time series of input current and input voltage; Inputting the time series of the input current and the time series of the input voltage into a sequence encoder to obtain a sequence of input current local time series associated feature vectors and a sequence of input voltage local time series associated feature vectors; Inputting the sequence of the input current local time series associated feature vectors and the sequence of the input voltage local time series associated feature vectors into a feature aggregation network based on feature energy local significance gating to obtain an input current time series aggregated significant representation vector and an input voltage time series aggregated significant representation vector; Inputting the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupling time series reinforcement interaction global representation vector as an input current-input voltage decoupling time series reinforcement interaction global representation feature; Based on the input current-input voltage decoupling timing enhancement interaction global representation feature, a control instruction is obtained, and the control instruction is used to represent a recommended duty cycle ratio value.

2. The power factor improvement control method for a charging system according to claim 1, characterized in that: Inputting the time series of the input current and the time series of the input voltage into a sequence encoder to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors, including: inputting the time series of the input current and the time series of the input voltage into a sequence encoder based on a 1D-CNN model to obtain a sequence of input current local timing association feature vectors and a sequence of input voltage local timing association feature vectors.

3. The power factor improvement control method for a charging system according to claim 2, characterized in that: Inputting the sequence of the input current local time series associated feature vectors and the sequence of the input voltage local time series associated feature vectors into a feature aggregation network based on feature energy local significance gating to obtain an input current time series aggregated significant representation vector and an input voltage time series aggregated significant representation vector, including: Determine the characteristic energy level coefficients of each input current local time series association characteristic vector in the sequence of the input current local time series association characteristic vectors to obtain a set of input current local time series association characteristic energy level coefficients based on the maximum value, minimum value, mean value and variance of each input current local time series association characteristic vector in the sequence of the input current local time series association characteristic vectors; Determine the scale of the local neighborhood, calculate the local neighborhood significance descriptor of each input current local timing correlation feature energy level coefficient in the set of input current local timing correlation feature energy level coefficients to obtain a set of input current local timing correlation feature energy level significance descriptors; Inputting the set of the input current local temporal correlation feature energy level significance descriptors into a mask module based on a gating function to obtain a set of mask-gated probabilistic input current local temporal correlation feature energy level significance descriptors; The set of the mask-gated probabilistic input current local temporal correlation feature energy level significance descriptors is used as a sequence of weights, and the position-weighted sum of the sequence of the input current local temporal correlation feature vectors is calculated to obtain the input current temporal aggregated significant representation vector.

4. The power factor improvement control method for a charging system according to claim 3, characterized in that: Based on the maximum value, minimum value, mean value and variance of each input current local time series association feature vector in the sequence of the input current local time series association feature vector, determining the characteristic energy level coefficient of each input current local time series association feature vector to obtain a set of input current local time series association feature energy level coefficients, including: Respectively extracting the maximum value and the minimum value in the input current local time series correlation feature vector to obtain the input current local time series correlation feature maximum value and the input current local time series correlation feature minimum value; Respectively calculating the mean and variance of the input current local time series correlation feature vector to obtain the input current local time series correlation feature mean and the input current local time series correlation feature variance; Adding the input current local time series correlation feature variance and a preset hyperparameter to obtain a first energy level coefficient of the input current local time series correlation feature; Calculate the square of the difference between the maximum value of the input current local timing correlation feature and the mean value of the input current local timing correlation feature to obtain the input current local timing correlation maximum difference value; Calculate the square of the difference between the minimum value of the input current local timing correlation feature and the mean value of the input current local timing correlation feature to obtain the input current local timing correlation minimum difference value; Adding the maximum difference value of the local timing correlation of the input current, the minimum difference value of the local timing correlation of the input current and the preset hyperparameter to obtain the second energy level coefficient of the local timing correlation feature of the input current; The first energy level coefficient of the input current local timing correlation characteristic is divided by the second energy level coefficient of the input current local timing correlation characteristic to obtain the energy level coefficient of the input current local timing correlation characteristic.

5. The power factor improvement control method for a charging system according to claim 4, characterized in that: Inputting the set of the input current local time series correlation feature energy level significance descriptors into a mask module based on a gating function to obtain a set of mask-gated probabilistic input current local time series correlation feature energy level significance descriptors, including: Performing normalization processing based on a Sigmoid function on each input current local timing correlation feature energy level significance descriptor in the set of input current local timing correlation feature energy level significance descriptors to obtain a set of normalized input current local timing correlation feature energy level significance descriptors; Each normalized input current local timing correlation feature energy level significance descriptor in the set of normalized input current local timing correlation feature energy level significance descriptors is input into a gating function for masking processing to obtain the set of mask-gated probabilistic input current local timing correlation feature energy level significance descriptors.

6. The power factor improvement control method for a charging system according to claim 5, characterized in that: Inputting the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector into a local fine-grained sequence interaction response analysis network based on feature decoupling to obtain an input current-input voltage decoupling time series enhanced interaction global representation vector, including: Performing feature decoupling on the input current time series aggregated significant representation vector and the input voltage time series aggregated significant representation vector to obtain a sequence of input current local time series aggregated significant representation vectors and a sequence of input voltage local time series aggregated significant representation vectors; Calculating the local semantic association interaction vectors between each group of corresponding input current local time series aggregation significant representation vectors and input voltage local time series aggregation significant representation vectors in the sequence of the input current local time series aggregation significant representation vectors and the sequence of the input voltage local time series aggregation significant representation vectors to obtain a sequence of input current-input voltage local time series semantic association interaction vectors; Calculating the fine-grained attention weight of each input current-input voltage local time series semantic association interaction vector in the sequence of input current-input voltage local time series semantic association interaction vectors to obtain a sequence of input current-input voltage local time series fine-grained attention weights; Inputting the sequence of input current-input voltage local time series fine-grained attention weights into a random dropout module to obtain a sequence of input current-input voltage local time series fine-grained attention weights after dropout processing; Taking the sequence of input current-input voltage local temporal fine-grained attention weights after the deactivation process as the weight distribution, the weighted sum of the sequence of input current-input voltage local temporal semantic association interaction vectors is calculated to obtain the input current-input voltage decoupled temporal reinforcement interaction global representation vector.

7. The power factor improvement control method for a charging system according to claim 6, characterized in that: Calculating the local semantic association interaction vector between each group of corresponding input current local time series aggregation significant representation vectors and input voltage local time series aggregation significant representation vectors in the sequence of the input current local time series aggregation significant representation vectors and the sequence of the input voltage local time series aggregation significant representation vectors to obtain a sequence of input current-input voltage local time series semantic association interaction vectors, including: Respectively calculating the position difference, position dot product and position addition between the input current local time series aggregation significant representation vector and the input voltage local time series aggregation significant representation vector to obtain an input current-input voltage local time series aggregation significant difference vector, an input current-input voltage local time series aggregation significant dot product vector and an input current-input voltage local time series aggregation significant sum vector; The input current-input voltage local time series aggregation significant difference vector, the input current-input voltage local time series aggregation significant dot product vector and the input current-input voltage local time series aggregation significant sum vector are cascaded and then one-dimensional convolutional encoded to obtain an input current-input voltage local time series aggregation significant multi-dimensional fusion vector; The input current-input voltage local time series aggregation significant multi-dimensional fusion vector is subjected to a local window-based maximum pooling process to obtain the input current-input voltage local time series semantic association interaction vector.

8. The power factor improvement control method for a charging system according to claim 7, characterized in that: Calculating the fine-grained attention weight of each input current-input voltage local time series semantic association interaction vector in the sequence of input current-input voltage local time series semantic association interaction vectors to obtain a sequence of input current-input voltage local time series fine-grained attention weights, including: After multiplying each input current-input voltage local timing semantic association interaction vector in the sequence of input current-input voltage local timing semantic association interaction vectors by a weight matrix, the obtained vector is added to the bias vector by position to obtain a sequence of input current-input voltage local timing semantic association interaction modulation vectors; Inputting the sequence of input current-input voltage local time sequence semantic association interaction modulation vectors into a tanh function for activation processing to obtain a sequence of input current-input voltage local time sequence semantic association interaction activation vectors; Each input current-input voltage local temporal semantic association interaction activation vector in the sequence of input current-input voltage local temporal semantic association interaction activation vectors is multiplied by a weight coefficient vector to obtain a sequence of input current-input voltage local temporal fine-grained attention weights.

9. The power factor improvement control method for a charging system according to claim 8, characterized in that: Based on the input current-input voltage decoupling timing enhancement interaction global representation feature, a control instruction is obtained, including: inputting the current-input voltage decoupling timing enhancement interaction global representation vector into a decoder-based power factor improvement control module to obtain the control instruction.

10. The power factor improvement control method for a charging system according to claim 9, characterized in that: Inputting the current-input voltage decoupling timing enhancement interaction global representation vector into a decoder-based power factor improvement control module to obtain the control instruction, including: using a decoder to decode and regress the current-input voltage decoupling timing enhancement interaction global representation vector according to the following formula to obtain a control instruction for representing a recommended duty cycle ratio value; Wherein, the formula is: Wherein, X is the global representation vector of the current-input voltage decoupling timing enhancement interaction, Y is the control instruction for representing the recommended duty cycle ratio value, W is the weight matrix, Represents matrix multiplication.

Images