Control method of emergency mobile power supply and emergency mobile power supply

Abstract

The invention relates to the technical field of emergency mobile power supplies, and discloses a control method of an emergency mobile power supply and the emergency mobile power supply, and the method comprises the steps: carrying out the operation parameter data collection and adaptive wavelet transform noise reduction of the emergency mobile power supply, and obtaining a standardized feature matrix; performing hierarchical feature extraction through an extended convolutional capsule network to obtain global power supply features and local power supply features; establishing a deep deterministic strategy network for power supply state scoring to obtain a state scoring result; inputting historical data in the experience playback pool into an action generation module, performing optimization iteration on the control action to obtain a first control strategy, and recording state transition data in a control process; and carrying out power loss compensation control analysis to obtain compensation control quantity, superposing the compensation control quantity to obtain a second control strategy, carrying out PWM (Pulse Width Modulation), and generating a charging and discharging control signal. According to the invention, accurate power regulation and energy storage unit control are realized, and stable operation of the system is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency mobile power supplies, and in particular to a control method of an emergency mobile power supply and an emergency mobile power supply. Background Art

[0002] As a key portable power supply device, emergency mobile power supplies play an important role in emergency rescue, disaster response, and temporary power supply scenarios. However, the current emergency mobile power supply control system generally has problems such as insufficient control accuracy, slow dynamic response, and poor power supply stability. Traditional control methods mainly rely on local power supply feature optimization, which is difficult to effectively handle dynamic load changes in complex power supply scenarios, resulting in slow system response and large power supply fluctuations.

[0003] In the existing emergency mobile power control methods, the power loss compensation problem has not received sufficient attention. In practical applications, line loss, switch loss and energy storage unit loss will significantly affect the power supply efficiency and stability of the system. Especially when the power output is high and the working state is frequently switched, the superposition effect of various losses will cause the system performance to drop sharply, affecting the reliability of emergency power supply. In addition, the existing control strategy has insufficient perception of the global state of the system, and it is difficult to accurately capture the key characteristic information of the power supply process. Summary of the invention

[0004] The present invention provides a control method for an emergency mobile power source and an emergency mobile power source, and the present invention realizes accurate power regulation and energy storage unit control, thereby ensuring stable operation of the system.

[0005] In a first aspect, the present invention provides a control method for an emergency mobile power supply, the control method for an emergency mobile power supply comprising: The operating parameter data of the emergency mobile power supply is collected and denoised by adaptive wavelet transform to obtain a standardized feature matrix; Using the standardized feature matrix to construct a multidimensional feature vector, and inputting the multidimensional feature vector into an extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features; A deep deterministic strategy network is established based on the global power supply characteristics and the local power supply characteristics, wherein the deep deterministic strategy network includes a value assessment module and an action generation module; the value assessment module is used to score the power supply status to obtain a status scoring result, and the status scoring result and the corresponding control action are stored in an experience replay pool; Inputting the historical data in the experience replay pool into the action generation module, optimizing and iterating the control action to obtain a first control strategy; performing online control using the first control strategy, and recording state transition data during the control process; The state transfer data is input into a power loss compensation model for power loss compensation control analysis to obtain a compensation control amount, and the compensation control amount is superimposed on the first control strategy to obtain a second control strategy. PWM modulation is performed based on the second control strategy to generate corresponding charge and discharge control signals.

[0006] In a second aspect, the present invention provides an emergency mobile power supply, the emergency mobile power supply comprising: The acquisition module is used to collect the operating parameter data of the emergency mobile power supply and perform adaptive wavelet transform noise reduction to obtain a standardized feature matrix; A feature extraction module, used to construct a multidimensional feature vector using the standardized feature matrix, and input the multidimensional feature vector into an extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features; An establishment module is used to establish a deep deterministic strategy network based on the global power supply characteristics and the local power supply characteristics, wherein the deep deterministic strategy network includes a value assessment module and an action generation module; the value assessment module is used to score the power supply status to obtain a status scoring result, and the status scoring result and the corresponding control action are stored in an experience replay pool; An iteration module, used for inputting the historical data in the experience replay pool into the action generation module, optimizing and iterating the control action, and obtaining a first control strategy; performing online control using the first control strategy, and recording state transition data during the control process; A generation module is used to input the state transfer data into a power loss compensation model to perform power loss compensation control analysis to obtain a compensation control amount, and superimpose the compensation control amount with the first control strategy to obtain a second control strategy, perform PWM modulation based on the second control strategy, and generate corresponding charge and discharge control signals.

[0007] In the technical solution provided by the present invention, by introducing adaptive wavelet transform for data noise reduction and feature extraction, combined with maximum and minimum value normalization calculation and time series analysis, a complete data preprocessing process is established, which effectively improves the quality and availability of the original data, and adopts an extended convolutional capsule network for hierarchical feature extraction. Through multi-level feature mapping and dynamic routing mechanism, the effective separation of global power supply characteristics and local power supply characteristics is achieved, and the system's perception of the power supply status is enhanced; a control architecture based on a deep deterministic strategy network is designed, and a complete state evaluation and action generation mechanism is established through the synergy of the value evaluation module and the action generation module. The system improves the accuracy and adaptability of the control strategy; an experience replay pool is built to store historical data, and random sampling and time difference calculation methods are used to optimize the control strategy. Through multiple rounds of iterative training and soft update mechanism, the continuous optimization and stable convergence of the control strategy are guaranteed; a multi-dimensional power loss compensation scheme is proposed, which significantly reduces the power loss of the system and improves the power supply efficiency through accurate modeling and compensation of line loss, switching loss and energy storage loss; a charge and discharge control mechanism based on PWM modulation is designed, and accurate power regulation and energy storage unit control are achieved through dead time compensation and triangle wave carrier modulation, ensuring the stable operation of the system.

[0008] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0009] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A schematic diagram of an embodiment of a control method for an emergency mobile power supply in an embodiment of the present invention; Figure 2 FIG. 1 is a schematic diagram of an embodiment of an emergency mobile power supply in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 are within the scope of protection of the present invention.

[0012] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device end including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or device ends.

[0013] To facilitate understanding of this embodiment, a control method for an emergency mobile power supply disclosed in an embodiment of the present invention is first described in detail. Figure 1 As shown, the method comprises the following steps: 101. Collect operating parameter data of the emergency mobile power supply and perform adaptive wavelet transform noise reduction to obtain a standardized feature matrix; It is understandable that the execution subject of the present invention may be an emergency mobile power supply, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0014] Specifically, data collection is performed to collect key operating parameters such as the input voltage, output current, load power and battery status of the emergency mobile power supply to reflect the overall load of the power supply and the health of the battery. The original operating parameter data is processed in segments. The continuous original data is cut and divided into several small time windows or data segments, and each feature is analyzed and processed within a more accurate time range. Wavelet transform is performed on each segment of data to extract the frequency domain characteristics of the signal. Wavelet transform is a signal analysis tool that can perform localized analysis in two dimensions: time and frequency. By calculating the wavelet coefficients of each segment of data, a preliminary wavelet decomposition result is obtained, which reflects the contribution of different frequency components to the original data and helps identify the details of noise and signal. The wavelet decomposition result is optimized by setting an adaptive threshold function. The adaptive threshold function dynamically adjusts the threshold according to the local characteristics of the data to distinguish between signals and noise. By calculating the size of the wavelet coefficient of each segment of data, a suitable threshold is determined. When the absolute value of the coefficient is lower than the threshold, it is considered to be noise and filtered out; otherwise, important signal components are retained. In this way, noise is effectively removed and the signal-to-noise ratio of the data is improved. Substitute the optimized threshold parameters into the wavelet inverse transform equation, reconstruct the initial wavelet decomposition results, and obtain the denoised operating parameter data. Perform maximum and minimum value normalization calculations on the denoised operating parameter data, unify the value range of each data item into a fixed interval, usually [0,1], eliminate the differences in dimensions of different parameters, make the data analyzed at the same scale, and avoid some features from having unfair effects on subsequent analysis due to excessively large or small value ranges. Through normalization processing, a set of standardized data is obtained. Based on the standardized data, the static characteristic parameters of the power supply system are extracted. Static characteristics reflect the behavior of the power supply at a certain moment or in a fixed state, such as the amplitude of the input voltage, the average value of the output current, and the remaining battery power. Static characteristics can provide the basic working conditions of the power supply, such as whether there are abnormal voltage fluctuations or battery status problems. After extracting the static features, consider the dynamic characteristics of the power supply. These features mainly reflect the changes in the power supply in the time series, helping to capture the volatility and operating trends of the system. For example, by calculating indicators such as voltage fluctuation rate, power change rate and load change trend, we can identify the response of the power supply when the load changes, and further understand whether the power supply can adapt to the load changes in time. Perform matrix concatenation operation on the static feature parameter set and the dynamic feature parameter set to obtain a standardized feature matrix.

[0015] 102. Use the standardized feature matrix to construct a multidimensional feature vector, and input the multidimensional feature vector into the extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features; Specifically, the standardized feature matrix is ​​reconstructed and feature combination operations are performed to fuse the static feature data and dynamic feature data in the matrix into a unified multidimensional feature vector. The multidimensional feature vector is input into the extended convolutional capsule network for hierarchical feature extraction. The feature vector enters the first convolutional layer of the network for feature mapping. In this layer, 32 3×3 convolution kernels are used to process the input features through convolution operations. At the same time, the ReLU activation function is used to introduce nonlinear factors, and the convolution results are standardized through batch normalization to accelerate the training process and reduce the risk of overfitting to obtain the primary feature map. The primary feature map is subjected to the maximum pooling operation. The role of the pooling operation is to reduce the dimension of the feature map. Through the 2×2 pooling kernel and the step size is set to 2, the pooling operation reduces the size of the feature map by half and retains important feature information. The reduced dimension feature map obtained after pooling is input into the main capsule layer of the extended convolutional capsule network for processing. More complex feature information is captured through tensor transformation and directional voting calculation of the main capsule layer. In this process, the main capsule layer contains 8 capsule units, each of which outputs a 16-dimensional vector, representing different aspects of the features captured by the capsule unit, effectively describing the detailed information of the power system, and obtaining the main capsule features. The main capsule features are input into the digital capsule layer of the extended convolutional capsule network. The digital capsule layer performs feature routing through a dynamic routing algorithm to extract more abstract features. Dynamic routing is an iterative optimization process that weights the outputs of different capsule units according to similarity to obtain more representative features. To ensure the stability of the routing process, the number of routing iterations is set to 3. Through multiple iterations, the routing algorithm can form reasonable connections between capsule units to ensure that the feature transfer is more accurate and efficient. Perform global and local feature extraction. In global feature extraction, the feature vector obtained by routing is globally averaged and pooled to obtain global power supply features. The global average pooling process is to perform a weighted average of all features to obtain a feature that represents the working state of the entire power system, reflecting the operating performance of the power supply at the overall level, such as the overall change trend of voltage, current and load. In order to further refine the state of the power supply, information is extracted from the local level. In the process of extracting local features, the routing feature vector is subjected to local attention weighted calculation. By introducing the attention mechanism, we focus on those features that have a greater impact on the performance of the power supply, such as a sudden change in current at a certain moment or an abnormal fluctuation in a power supply unit. Local features can describe the specific working status of each subsystem or power supply unit in the power supply, and help identify potential faults or unstable factors. Through the above steps, we can obtain global power supply features and local power supply features.

[0016] 103. Establish a deep deterministic strategy network based on global power supply characteristics and local power supply characteristics, wherein the deep deterministic strategy network includes a value assessment module and an action generation module; use the value assessment module to score the power supply status, obtain the status scoring result, and store the status scoring result and the corresponding control action in the experience replay pool; Specifically, the global power supply features and local power supply features are concatenated to obtain a combined feature vector, which includes the overall operating status of the power supply system and the specific performance of each power supply unit, reflecting the operating status of the power supply system under different conditions. The structure of the deep deterministic strategy network is designed. In the value assessment module, the combined feature vector is dimensionally mapped. In order to ensure the effective expression of the input features in the network, it is processed by a three-layer fully connected neural network. The fully connected layer can extract and integrate the input feature information layer by layer, and map the high-dimensional input data to a feature space suitable for evaluating the current power supply status. Through dimensional mapping, the network captures the key status information of the power supply more accurately. After processing, the value assessment module evaluates the operating status of the current power supply system according to the input feature vector, and calculates the status score result, including the power supply stability score, the power balance score and the response speed score. At the same time, the combined feature vector is input into the action generation module. This module generates the corresponding control action according to the operating status of the current power supply. In order to achieve this goal, the combined feature vector is dimensionally reduced through a four-layer fully connected neural network to ensure that the feature expression is more concise without losing important information. The reduced feature vector is input into the action generation module to generate the corresponding control action parameters. The control action parameters include power regulation instructions and charge and discharge instructions of the energy storage unit, ensuring that the power supply can adjust its power output and battery charge and discharge strategy according to the current demand to meet the stable operation requirements of the system. A reward function is set for the state scoring result to measure the effect of the control strategy. The reward function comprehensively considers multiple factors such as power supply stability, power balance and response speed, and calculates a reward value based on these indicators. The reward value reflects the effect of the current strategy execution and can guide the deep reinforcement learning algorithm to further optimize the strategy so that the power supply system can achieve the optimal power supply performance under different working conditions. In order to achieve continuous optimization of the strategy, an experience replay pool is constructed. The role of the experience replay pool is to store historical state, action and reward information for sampling during the training process to update the model parameters. The storage capacity of the experience replay pool is set to 10,000. The system can store a large amount of historical experience and provide sufficient data support for the training process. Whenever the network is trained, the data in the experience replay pool will be stored in time sequence, including the current state scoring result, control action parameters and state value evaluation results calculated according to the reward function. Through the accumulation of these historical data, we continuously learn and improve the control strategy, gradually improving the power supply stability and responsiveness of the power supply system. Through the above steps, the deep deterministic strategy network continuously optimizes based on historical experience and the current system status, allowing the power supply to adjust its working strategy more intelligently to meet the power supply needs in emergency situations.

[0017] 104. Input the historical data in the experience replay pool into the action generation module, optimize and iterate the control action, and obtain the first control strategy; use the first control strategy to perform online control, and record the state transition data in the control process; Specifically, historical data in the experience replay pool are randomly sampled to obtain training data batches. The training data batch contains historical state, action and reward information, representing the performance and decision-making process of the power system under different operating environments. The training data batch is subjected to time difference calculation, and the target Q value is calculated through the current state transition sequence. The Q value is a standard for evaluating the value of a state-action pair, indicating the long-term return that can be brought by taking a certain action in a certain state. The target Q value is used as the optimization target of the value evaluation module to guide the subsequent learning process. By calculating the difference between the current state and the target state, the model is continuously updated so that the strategy is gradually optimized. The training data batch is input into the action generation module, and the network parameters of the action generation module are gradient updated. The Adam optimizer is used to optimize the network parameters. The Adam optimizer is an adaptive learning rate optimization algorithm that dynamically adjusts the learning rate according to the gradient of each parameter, thereby improving the efficiency and stability of training. In order to ensure the smooth progress of the training process, the learning rate is set to 0.001 to ensure that the amplitude of each parameter update is moderate, which not only avoids the local optimum caused by too fast convergence, but also prevents the problem of too slow convergence. Through this step, the network parameters of the action generation module are gradually updated, so that the generated control action is more in line with the needs of the power supply system. At the same time, the value assessment module is soft updated. The soft update strategy updates the parameters of the value assessment module through a certain step coefficient, so that the parameters of the target network are close to the parameters of the current network, but it will not be updated significantly at one time to avoid parameter instability. The update step coefficient is set to 0.01. By gradually approaching the target network parameters, the parameters of the value assessment module are gradually optimized, improving the accuracy of its power supply status assessment. After updating the action generation module and the value assessment module, the entire deep deterministic strategy network undergoes multiple rounds of iterative training. Each iteration is trained based on the latest network parameters, so that the system can continuously optimize the control strategy and gradually obtain the first control strategy. After multiple rounds of training, the deep deterministic strategy network can make accurate decisions under the ever-changing power demand and environmental conditions, providing a more efficient and stable control strategy for the power supply system. The first control strategy is deployed to the control system of the emergency mobile power supply. In the control system, the input voltage, output current and load power are adjusted in real time according to the first control strategy to ensure that the power supply responds quickly to factors such as load changes and environmental changes in actual use, and to ensure the stability and reliability of power supply. During the control process, the system response data during the adjustment process is recorded. The system response data is sampled in segments. The system state is sampled according to the time step, and the power supply state parameters and control parameters at each sampling moment are recorded to form complete state transition data. The state transition data within each period of time contains the current power supply state of the system, the control actions taken, and the corresponding reward information. By analyzing these data, the effect of the current control strategy is evaluated, and the necessary data support is provided for the next round of strategy optimization.

[0018] 105. Input the state transfer data into the power loss compensation model to perform power loss compensation control analysis to obtain a compensation control amount, and superimpose the compensation control amount with the first control strategy to obtain a second control strategy, perform PWM modulation based on the second control strategy, and generate corresponding charge and discharge control signals.

[0019] Specifically, the state transition data is parsed. By extracting the data of line loss, switch loss and energy storage unit loss, multi-dimensional loss data is formed to represent the energy loss of each component of the power system during operation. A power loss compensation model is constructed based on the multi-dimensional loss data. The model models the power loss of the power system based on the loss data and calculates the distribution matrix of the power loss. This matrix reflects the spatial distribution and time variation trend of various types of losses, and reflects the performance of each loss type under different loads and operating conditions. Through this step, it is identified which losses have the greatest impact on the overall power supply efficiency, so as to compensate for them in a targeted manner. A weighted sum operation is performed on the power loss distribution matrix to calculate the compensation weight coefficient. Each loss item is assigned a suitable compensation coefficient according to its impact on the system stability. After weighted summation, a compensation parameter matrix is ​​obtained, in which each element represents the compensation amount of a specific loss type. Based on the compensation parameter matrix, various compensation control quantities are calculated, including line compensation quantity, switch compensation quantity and energy storage compensation quantity. These compensation control quantities specifically reduce the negative impact of various types of losses on the overall performance of the system, ensuring that the power supply can operate at a higher efficiency. The compensation control quantity is linearly superimposed with the first control strategy to obtain the second control strategy. Based on the second control strategy, PWM modulation is performed to generate corresponding charge and discharge control signals. In order to realize PWM modulation, a triangular wave carrier signal is used to modulate and generate a PWM duty cycle sequence. In each cycle of the duty cycle sequence, the rising edge and the falling edge correspond to the intersection of the triangular wave carrier, respectively. This modulation method can accurately control the power output of the power supply. Dead time compensation is performed on the power regulation pulse signal and the charge and discharge control signal of the energy storage unit. Dead time refers to the response lag area of ​​the signal when the switching element changes from the closed state to the open state during the change of the switching signal. By compensating for the dead time, the instability caused by signal lag is avoided, ensuring that each switching operation can be accurately executed, and finally obtaining a stable charge and discharge control signal.

[0020] The line loss data in the power loss distribution matrix are normalized to ensure that each data type has the same dimension and comparability. Based on the input voltage and output current, the impedance of the line is calculated to determine the weight coefficient of the line loss, which reflects the relationship between the energy loss in the line and the input current and voltage. The switching loss data in the power loss distribution matrix are analyzed in the frequency domain. The switching loss is mainly affected by the switching frequency and current stress. The frequency domain analysis method is used to identify the switching loss at different frequencies through frequency domain analysis. According to the switching frequency and current stress, the conduction loss and turn-off loss in the switching process are calculated to obtain the weight coefficient of the switching loss to measure the energy loss in the switching process. Similarly, the energy storage unit loss data is analyzed in a time series manner. The loss of the energy storage unit is affected by the charge and discharge state and the change of the internal resistance. The change of the internal resistance of the battery is monitored during different charge and discharge processes, and the energy storage loss is calculated based on these changes. The change of the internal resistance directly affects the energy loss of the battery. Through a comprehensive analysis of the charge and discharge state and the change of the internal resistance, the weight coefficient of the energy storage loss is calculated to reflect the loss degree of the energy storage unit under different working conditions. The line loss weight coefficient, switch loss weight coefficient and energy storage loss weight coefficient are combined to construct a weight matrix. The weight matrix is ​​normalized to obtain a compensation parameter matrix. Based on the compensation parameter matrix, various compensation control quantities are calculated. Based on the compensation parameter matrix, the line power compensation value is calculated. The calculation formula of the line compensation quantity shows that the line compensation quantity is proportional to the product of the input voltage and the output current. According to the real-time data of the input voltage and the output current, the corresponding line power compensation value is calculated to effectively compensate for the energy loss caused by the line impedance, thereby optimizing the efficiency of the power supply system. Similarly, based on the compensation parameter matrix, the switch power compensation value is calculated. The calculation basis of the switch compensation quantity is the product of the switching frequency and the switching current. According to the real-time frequency and current data during the switching process, the corresponding switch power compensation value is calculated. This compensation quantity is used to compensate for the loss caused by frequent switching operations, thereby avoiding the negative impact of high-frequency switching on the stability of the power supply system. The energy storage power compensation value is calculated based on the compensation parameter matrix. The energy storage compensation amount is proportional to the product of the charge and discharge current and the equivalent internal resistance. The energy storage power compensation value is calculated based on the current data and the change in battery internal resistance during the charge and discharge process, which effectively reduces the energy loss during the battery charge and discharge process and improves the battery efficiency. Line power compensation value, switch power compensation value and energy storage power compensation value are linearly superimposed to obtain the compensation control amount.

[0021] In the embodiment of the present invention, an adaptive wavelet transform is introduced to perform data noise reduction and feature extraction, and a complete data preprocessing process is established in combination with maximum and minimum value normalization calculation and time series analysis, which effectively improves the quality and availability of the original data. An extended convolutional capsule network is used for hierarchical feature extraction, and the global power supply features and local power supply features are effectively separated through multi-level feature mapping and dynamic routing mechanisms, thereby enhancing the system's ability to perceive the power supply status. A control architecture based on a deep deterministic strategy network is designed, and a complete state evaluation and action generation mechanism is established through the synergy of a value evaluation module and an action generation module. The accuracy and adaptability of the control strategy have been improved; an experience replay pool has been built to store historical data, and the control strategy has been optimized using random sampling and time difference calculation methods. Through multiple rounds of iterative training and soft update mechanisms, the continuous optimization and stable convergence of the control strategy are guaranteed; a multi-dimensional power loss compensation scheme has been proposed, which significantly reduces the power loss of the system and improves the power supply efficiency through accurate modeling and compensation of line loss, switching loss and energy storage loss; a charging and discharging control mechanism based on PWM modulation has been designed, and accurate power regulation and energy storage unit control have been achieved through dead time compensation and triangular wave carrier modulation, ensuring the stable operation of the system.

[0022] In a specific embodiment, the process of executing step 101 may specifically include the following steps: Collect data on the input voltage, output current, load power and battery status of the emergency mobile power supply to obtain original operating parameter data; The original operating parameter data is processed in segments, and the wavelet coefficient of each segment of data is calculated to obtain the initial wavelet decomposition result; Based on the initial wavelet decomposition result, an adaptive threshold function is set to obtain an optimized threshold parameter, and the optimized threshold parameter is substituted into the wavelet inverse transform equation, and the initial wavelet decomposition result is reconstructed to obtain the noise-reduced operating parameter data; Performing maximum and minimum normalization calculations on the noise-reduced operating parameter data to obtain standardized data, and extracting static features from the standardized data according to the inherent parameters of the power system to obtain a static feature parameter set; The standardized data is analyzed in time series, the voltage fluctuation rate, power change rate and load change trend are calculated to obtain a dynamic feature parameter set, and the static feature parameter set and the dynamic feature parameter set are concatenated to obtain a standardized feature matrix.

[0023] Specifically, the input voltage of the power supply is collected from the sensor in real time , output current , Load power and battery status , respectively represent the system's input power, output current flow, system load power consumption, and the battery's current working state (such as remaining power or temperature). Assume that the collected time series data is ,in Indicates time. These data are segmented and the entire time series Divide into The length is Each data segment is represented by ,in is a segment index, Then, wavelet transform is performed on each data segment to calculate the wavelet coefficients. Wavelet transform is a time-frequency analysis method that selects a wavelet basis function. , for the input data Decompose. The continuous wavelet transform formula is: ; in, are the wavelet coefficients, is the scale parameter (controls the frequency resolution), is the translation parameter (controls the temporal resolution), is the complex conjugate of the wavelet basis function. By calculating , and get the initial wavelet decomposition result. Set the adaptive threshold function based on the initial wavelet decomposition result The threshold function is set adaptively according to the amplitude of the wavelet coefficient, such as using the soft threshold method: ; in, is the standard deviation of the noise, is the data length. For each wavelet coefficient ,like , it is considered as noise and set to zero; if , then reduce it. Substitute the processed wavelet coefficients into the inverse wavelet transform equation: ; in, is the data after noise reduction, is the regularization constant of the wavelet basis function. After denoising, the denoised data is normalized, and the maximum and minimum value normalization method is used to map all data to the [0,1] interval. The normalization formula is: ; in, and are the minimum and maximum values ​​of the data, respectively. It is the normalized data. The normalized data is subjected to static feature extraction according to the inherent parameters of the power system. Static features are parameters that reflect the overall state of the data, such as the average value of the voltage. , output current variance , Maximum value of load power Etc. The static feature set is expressed as: ; Perform time series analysis on normalized data to extract dynamic features. Dynamic features describe the trend of data changes over time, such as voltage fluctuation rate. , power change rate , Load change trend Etc. The dynamic feature set is expressed as: ; Static feature collection With dynamic feature collection Perform matrix concatenation to form a standardized feature matrix : ; The standardized characteristic matrix describes the operating status of the emergency mobile power supply, including static overall characteristics and dynamic change trends.

[0024] In a specific embodiment, the process of executing step 102 may specifically include the following steps: Performing dimension reconstruction and feature combination operations on the standardized feature matrix to obtain a multi-dimensional feature vector, which includes static feature data and dynamic feature data; The multidimensional feature vector is input into the first convolutional layer of the extended convolutional capsule network for feature mapping. The first convolutional layer contains 32 3×3 convolution kernels, uses the ReLU activation function and batch normalization to obtain the primary feature map; Perform a maximum pooling operation on the primary feature map with a pooling kernel size of 2×2 and a step size of 2 to obtain a reduced-dimensional feature map, and input the reduced-dimensional feature map into the main capsule layer of the extended convolutional capsule network; In the main capsule layer, the dimension reduction feature map is transformed into a tensor and directional voting is calculated. The main capsule layer contains 8 capsule units, each of which outputs a 16-dimensional vector to obtain the main capsule feature. The main capsule feature is input into the digital capsule layer of the extended convolutional capsule network. The digital capsule layer contains 10 capsule units. The dynamic routing algorithm is used for feature routing. The number of routing iterations is 3, and the routing feature vector is obtained. A global average pooling operation is performed on the routing feature vector to obtain the global power supply feature, and a local attention weighted calculation is performed on the routing feature vector to obtain the local power supply feature. The global power supply feature reflects the overall power supply status of the system, and the local power supply feature represents the working status of each power supply unit.

[0025] Specifically, the input standardized feature matrix Reconstruct. Assume is a The matrix of represents the number of features, Represents the time step. Through the feature combination operation, the static feature data of each column and dynamic feature data Concatenate into a set of multidimensional feature vectors The multidimensional feature vector is represented as: ; in, Indicates Static characteristics of the column, such as average voltage or load power; Indicates In this way, the static and dynamic information of the feature matrix are fused to obtain a multidimensional feature vector set containing two parts of features. The multidimensional feature vector is input into the first convolutional layer of the extended convolutional capsule network for feature mapping. The first convolutional layer contains 32 The convolution kernel of the convolution operation is expressed by the following formula: ; in, It is The convolution kernels are located at The output features of is the weight of the convolution kernel, is the bias, is the ReLU activation function, which is used to introduce nonlinearity. Through this convolution operation, each convolution kernel extracts local patterns in the input data, such as the local law of voltage fluctuation or the instantaneous change trend of load power. At the same time, the convolution layer uses batch normalization technology to normalize each convolution feature. The formula is: ; in, is the mean of the mini-batch data, is the variance, is a very small value to prevent the denominator from being zero. After ReLU and batch normalization, the primary feature map is obtained. , the size of each feature map is obtained by subtracting the convolution kernel size from the input matrix, and the formula is: ; in, and are the height and width of the input matrix, is the size of the convolution kernel. The primary feature map passes through the convolution layer and is then input into the maximum pooling layer for dimensionality reduction. The maximum pooling operation is performed through a The pooling kernel and sliding window with a step size of 2 are implemented. The formula is: ; in, is the value of the feature map after pooling. By selecting the maximum value in the sliding window, the size of the feature map is effectively reduced while retaining the most significant features. The reduced-dimensional feature map is input into the main capsule layer of the extended convolutional capsule network. In the main capsule layer, the reduced-dimensional feature map is processed by tensor transformation and directional voting mechanism. The main capsule layer contains 8 capsule units, each of which outputs a 16-dimensional vector , the formula is: ; in, is the input feature vector, is the weight matrix, is the directional voting score, is a scaling function used to normalize the output amplitude. In this way, each capsule unit captures a specific power supply mode, such as global voltage fluctuations or power anomalies in a local unit. The main capsule features are input into the digital capsule layer for processing. The digital capsule layer contains 10 capsule units and uses a dynamic routing algorithm for feature routing. The formula for dynamic routing is: ; in, is the routing score, and are the feature vectors of the main capsule layer and the digital capsule layer respectively. After the routing is iterated 3 times, the output routing feature vector Contains comprehensive information of multiple layers of features. Perform global average pooling on the routing feature vector, the formula is: ; Global power supply characteristics It reflects the overall power supply status of the power system, such as total power stability and average load fluctuation rate. At the same time, the routing feature vector is weighted by local attention, and the formula is: ; Among them, the attention weight Calculated through a weighting function to focus on the description of certain key features. Local power supply characteristics Reflects the working status of each power supply unit, such as the charge and discharge characteristics of a specific battery unit or the abnormal changes in a current path.

[0026] In a specific embodiment, the process of executing step 103 may specifically include the following steps: The global power supply feature and the local power supply feature are concatenated to obtain a combined feature vector; The combined feature vector is dimensionally mapped, and a value assessment module of a deep deterministic strategy network is constructed through a three-layer fully connected neural network; The combined feature vector is processed with dimension reduction, and the action generation module of the deep deterministic policy network is constructed through a four-layer fully connected neural network; The combined feature vector is input into the value assessment module to score the current power supply status and obtain the status scoring result, which includes the power supply stability score, power balance score and response speed score; The combined feature vector is input into the action generation module to generate corresponding control action parameters, which include power adjustment instructions and energy storage unit charging and discharging instructions; A reward function is set for the status scoring result, and the reward value is calculated according to the power supply stability, power balance and system response speed to obtain the status value evaluation result; Construct an experience replay pool with a storage capacity of 10,000. Store the state scoring results, control action parameters, and state value assessment results in the experience replay pool in chronological order.

[0027] Specifically, the global power supply feature and local power supply characteristics Perform feature stitching. Global power supply feature Represents the overall power supply status of the power system, such as the average stability of system voltage and the fluctuation of power output, while local power supply characteristics The state of each power supply unit is refined, including the charging and discharging status of the energy storage unit and the power supply efficiency of the local load. Feature concatenation generates a combined feature vector by merging the two feature vectors by dimension. , whose expression is: ; Among them, the combined feature vector Contains the global and local feature information of the system, and can fully describe the operating status of the emergency mobile power supply. Input value assessment module and action generation module. Dimension mapping is performed to reduce the complexity of features and extract meaningful features. In the value assessment module, a three-layer fully connected neural network is constructed, and the output of each layer represents the step-by-step mapping and feature extraction of the input features. Output of the layer , expressed as: ; in, It is The weight matrix of the layer, is the bias term, is the activation function (e.g. ReLU). The first layer takes the input features Mapped to the latent space, the second layer extracts more abstract features, and the third layer outputs the score of the current power supply status. The scoring results include the power supply stability score , Power Balance Score and responsiveness score The state scoring result is expressed as: ; in, Measures the voltage and power stability of the system, Reflects the degree of power supply balance of the load. Indicates the response speed of the power supply to load changes. At the same time, the combined eigenvector The input action generation module performs dimensionality reduction processing to generate the corresponding control action parameters. In the action generation module, a four-layer fully connected neural network is used to Perform gradual dimensionality reduction and output the control action parameters The output calculation formula of the action generation module is similar to that of the value assessment module and is expressed as: ; in, Contains two main control instructions: power adjustment instructions and energy storage unit charge and discharge instructions . Power regulation instructions Controls the overall power output of the power system, while the charge and discharge instructions Determine the charging or discharging strategy of the energy storage unit at the current moment. Scoring the status results Define a reward function , which is used to measure the pros and cons of the current control strategy. The reward function comprehensively considers power supply stability, power balance and system response speed, and the formula is expressed as: ; in, , and is the weight parameter of the reward function, which is used to adjust the influence of different scoring items in the total reward. , evaluate the control strategy according to the state scoring results, and obtain the state value evaluation results , whose value is: ; in, is the discount factor, is the value estimate of the next state, which is used to balance short-term benefits and long-term benefits. In order to optimize the control strategy, an experience replay pool is constructed to store historical states, actions, and reward information. The storage capacity of the experience replay pool is set to 10,000, which can save enough historical samples to support subsequent training optimization. Each execution of the control strategy generates a new sample, including the state scoring result S, the control action parameter A, and the state value evaluation result Q. These samples are stored in the experience replay pool in time sequence to provide training data for the deep reinforcement learning algorithm.

[0028] In a specific embodiment, the process of executing step 104 may specifically include the following steps: Randomly sample historical data in the experience replay pool to obtain training data batches; Perform time series difference calculation on the training data batches, calculate the target Q value based on the state transition sequence, and use the target Q value as the optimization target of the value assessment module; The training data batches are input into the action generation module, and the network parameters of the action generation module are gradient updated. The Adam optimizer is used for parameter optimization, and the learning rate is set to 0.001; The value assessment module is soft-updated, the update step coefficient is set to 0.01, and the parameters of the value assessment module are brought closer to the target network parameters to obtain the updated value assessment module; Based on the updated value assessment module and action generation module, the deep deterministic policy network is trained for multiple rounds to obtain the first control strategy; Deploy the first control strategy to the control system of the emergency mobile power supply, adjust the input voltage, output current and load power in real time, and record the system response data during the adjustment process; The system response data is sampled in sections, and the power supply state parameters and control parameters at each sampling moment are recorded to obtain the state transfer data.

[0029] Specifically, historical data is randomly sampled from the experience replay pool to generate a training data batch. This data contains the states recorded during the previous system operation. ,action ,award and the next state . Assume that the size of the experience replay pool is , the size of the training data batch is , the random sampling process is represented by drawing from the playback pool without replacement samples, expressed as: ; in, is the training data set of the current batch, Indicates The state of the samples, is the corresponding control action, is the reward obtained after performing the action. is the next state after the state transition. A time difference calculation is performed on each sample in the training data batch to estimate the target Q value. The time difference method defines the target Q value according to the Bellman equation , the formula is: ; in, is a discount factor that balances short-term rewards with long-term rewards, The next state The maximum Q value under . Through time difference calculation, combined with the current reward and future potential benefits, and estimate the target Q value for each sample. This target value is used as the optimization target of the value assessment module to guide the update of network parameters. Input to the action generation module, and update its network parameters by gradient. The action generation module is fitted Function predicts the best action ,in is the policy function, which means that in the state In order to optimize the parameters of the action generation module, the Adam optimizer is used. Adam is an adaptive optimization algorithm based on the first-order gradient, and its update formula is: ; in, It is The parameters for the iteration, is the learning rate (set to 0.001), and are the first-order momentum and second-order momentum of the gradient, is a small value that avoids the denominator being zero. Through the Adam optimizer, the network parameters of the action generation module can be updated in an efficient and stable manner. At the same time, the value assessment module is soft-updated to improve its prediction ability. The principle of soft update is to gradually move the parameters of the current network closer to the parameters of the target network, thereby smoothly updating the model and avoiding instability caused by large parameter adjustments. The formula for soft update is: ; in, is the update step size coefficient (set to 0.01), are the current network parameters, is the target network parameter. By gradually approaching, the updated value assessment module can predict the Q value more accurately and provide reliable support for policy optimization. After completing the optimization of the action generation module and the value assessment module, the deep deterministic policy network is trained for multiple rounds of iterations based on the updated modules. Each round of iteration uses a cycle of random sampling, time difference calculation, and parameter update to enable the network to gradually learn the optimal control strategy for the emergency mobile power supply. After multiple rounds of training, the first control strategy obtained can output the optimal control action in real time according to the input state of the system. After deploying the first control strategy to the control system of the emergency mobile power supply, the input voltage , output current and load power Make real-time adjustments. Each adjustment is based on the current status. and the actions generated by the strategy Perform power distribution and charge and discharge control of energy storage units. During the control process, record the system response data at each adjustment moment in real time, including state parameters and control parameters. Sampling the system response data in segments, taking each segment of data as a state transition sample, and recording the current state, execution action, and state information after the transition. Optimize the control strategy through the continuous accumulation of these data.

[0030] In a specific embodiment, the process of executing step 105 may specifically include the following steps: Analyze and process the state transfer data, extract line loss data, switch loss data and energy storage unit loss data, and obtain multi-dimensional loss data; Building a power loss compensation model based on multi-dimensional loss data, and inputting the multi-dimensional loss data into the power loss compensation model to obtain a power loss distribution matrix; Perform weighted sum calculation on the power loss distribution matrix, determine the compensation weight coefficients of various types of losses, obtain the compensation parameter matrix, and calculate the compensation control amount based on the compensation parameter matrix. The compensation control amount includes line compensation amount, switch compensation amount and energy storage compensation amount; Performing a linear superposition operation on the compensation control amount and the first control strategy to obtain a second control strategy, and performing triangular wave carrier modulation based on the second control strategy to obtain a PWM duty cycle sequence; Generate a power regulation pulse signal and a charge and discharge control signal of the energy storage unit based on the PWM duty cycle sequence, wherein the rising edge and the falling edge of the power regulation pulse signal correspond to the intersection of the triangular wave carrier; The dead time compensation is performed on the power regulation pulse signal and the energy storage unit charge and discharge control signal to obtain the final charge and discharge control signal.

[0031] Specifically, key parameters related to various types of losses are identified from the state transition data. Assume that the state transition data includes input voltage , output current , switching frequency , charge and discharge current And the equivalent line impedance and the internal resistance of the energy storage unit .in and Indicates the input power status of the power supply system. Reflects the operating frequency of the switch, Reflects the working intensity of the energy storage unit, and are the main factors of line loss and energy storage unit loss respectively. For line loss, the line loss power Calculated by the following formula: ; in, is the output current, is the line impedance. This formula indicates that the line loss is proportional to the square of the output current and the line impedance. For switching loss, the switching loss is divided into conduction loss and turn-off loss. The calculation formula is: ; Turn-off loss The calculation formula is: ; in, is the switching voltage, and They are the on-time and off-time of the switch, is the switching frequency. The total switching loss is the sum of conduction loss and turn-off loss: ; The loss of energy storage unit is mainly caused by the charge and discharge current. and the internal resistance of the energy storage unit Determine. Energy storage loss power It is expressed as: ; The above three formulas respectively obtain line loss, switch loss and energy storage loss data, which constitute a multi-dimensional loss data set. Based on multidimensional loss data , build a power loss compensation model. After the loss data is input into the model, the power loss distribution matrix is ​​calculated , whose rows represent time series and columns represent different loss types. Each element Indicates The time point Class loss value. Power loss distribution matrix Perform weighted sum calculation to determine the compensation weight coefficients for various types of losses The weight coefficient is calculated by normalization: ; in, It is The weight of the class loss, is the length of the time series. The larger the weight, the more significant the impact of this type of loss on the system. Construct a compensation parameter matrix based on the compensation weight coefficient , and calculate the compensation control amount. Line compensation amount It is proportional to the product of input voltage and output current, and the formula is: ; Switch compensation It is proportional to the product of the switching frequency and the switching current, and the formula is: ; Energy storage compensation It is proportional to the product of the charge and discharge current and the internal resistance. The formula is: ; The compensation control quantity and the first control strategy Perform linear superposition to obtain the second control strategy : ; Based on the second control strategy, the PWM duty cycle sequence is generated by triangular wave carrier modulation. The duty cycle sequence is determined by the intersection of the control signal and the triangular wave: ; in, is a triangle wave signal, It is the output signal of PWM modulation. It compensates the dead time of the power regulation pulse signal and the energy storage unit charge and discharge signal, adjusts the switching timing, and ensures the non-response time between the rising edge and the falling edge. The compensated signal is used as the final charge and discharge control signal, which is applied to the control of the emergency mobile power supply and optimizes the system operation efficiency.

[0032] In a specific embodiment, the execution step performs weighted sum calculation on the power loss distribution matrix, determines the compensation weight coefficients of various types of losses, obtains the compensation parameter matrix, and calculates the compensation control amount according to the compensation parameter matrix. The compensation control amount includes the line compensation amount, the switch compensation amount and the energy storage compensation amount. The process can specifically include the following steps: Normalizing the line loss data in the power loss distribution matrix, calculating the line impedance based on the input voltage and output current, and obtaining the line loss weight coefficient; Perform frequency domain analysis on the switching loss data in the power loss distribution matrix, calculate the turn-on loss and turn-off loss based on the switching frequency and current stress, and obtain the switching loss weight coefficient; Perform time series analysis on the energy storage unit loss data in the power loss distribution matrix, calculate the energy storage loss based on the charge and discharge state and the internal resistance change, and obtain the energy storage loss weight coefficient; The line loss weight coefficient, the switch loss weight coefficient and the energy storage loss weight coefficient are combined to construct a weight matrix, and the weight matrix is ​​normalized to obtain a compensation parameter matrix; Calculating a line power compensation value based on the compensation parameter matrix, where the line compensation amount is proportional to the product of the input voltage and the output current, and calculating a switch power compensation value based on the compensation parameter matrix, where the switch compensation amount is proportional to the product of the switching frequency and the switching current; The energy storage power compensation value is calculated based on the compensation parameter matrix. The energy storage compensation amount is proportional to the product of the charging and discharging current and the equivalent internal resistance. The line power compensation value, the switch power compensation value and the energy storage power compensation value are linearly superimposed to obtain the compensation control amount.

[0033] Specifically, the line loss data in the power loss distribution matrix is ​​normalized. Assume that the power loss distribution matrix Elements Indicated in Time point Class loss data, where represents the line loss, represents the switching loss, Represents the energy storage unit loss. When normalizing the line loss data, the maximum and minimum value normalization method is used, and the formula is as follows: ; in, is the normalized line loss data, and Respectively represent the minimum and maximum values ​​of line loss data. Normalization eliminates the influence of different data dimensions, making subsequent analysis more unified. After obtaining the normalized line loss data, based on the input voltage and output current Calculating Line Impedance , the formula is: ; Line loss weight coefficient The proportion of line loss in total loss is calculated as follows: ; in, is the time step, is the line loss weight coefficient, which indicates the importance of line loss in multi-dimensional loss. The switching loss data in the power loss distribution matrix is ​​analyzed in the frequency domain to calculate the conduction loss and turn-off loss. The two main components of switching loss are conduction loss and turn-off loss. and turn-off losses The calculation formula for conduction loss is: ; The calculation formula for turn-off loss is: ; in, is the switching voltage, and They are the on-time and off-time, is the switching frequency. Indicates the switching loss weight coefficient, the switching loss weight coefficient By normalization calculation, the formula is: ; The switching loss weight coefficient reflects the influence of the switching frequency on the total loss. At the same time, the loss data of the energy storage unit is analyzed in time series. The energy storage loss is mainly caused by the charge and discharge current. And the equivalent internal resistance of the energy storage unit Determine. Energy storage loss power The calculation formula is: ; In timing analysis, the charge and discharge status of the energy storage unit is measured by The positive and negative values ​​represent charging and the negative values ​​represent discharging. Represents the energy storage loss weight coefficient, energy storage loss weight coefficient The calculation formula is: ; Weight coefficient Reflects the proportion of energy storage unit loss in total loss. , Switching loss weight coefficient and energy storage loss weight coefficient Combined into a weight matrix After normalizing the weight matrix, the compensation parameter matrix is ​​obtained : ; in, is the normalized compensation parameter. Based on the compensation parameter matrix, the line power compensation value is calculated. The formula is: ; Calculate the switch power compensation value : ; Calculate the energy storage power compensation value : ; in, Indicates the line compensation amount, Indicates the switch compensation amount, Represents the energy storage compensation amount. Line power compensation value, switch power compensation value and energy storage power compensation value are linearly superimposed to obtain the compensation control amount : .

[0034] The control method of the emergency mobile power supply in the embodiment of the present invention is described above. The emergency mobile power supply in the embodiment of the present invention is described below. Figure 2 In one embodiment of the present invention, an emergency mobile power supply includes: The acquisition module 201 is used to collect operating parameter data of the emergency mobile power supply and perform adaptive wavelet transform noise reduction to obtain a standardized feature matrix; A feature extraction module 202 is used to construct a multidimensional feature vector using a standardized feature matrix, and input the multidimensional feature vector into an extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features; Establishing module 203, used to establish a deep deterministic strategy network based on global power supply characteristics and local power supply characteristics, the deep deterministic strategy network includes a value evaluation module and an action generation module; using the value evaluation module to score the power supply state, obtain the state scoring result, and store the state scoring result and the corresponding control action in the experience replay pool; The iteration module 204 is used to input the historical data in the experience playback pool into the action generation module, optimize and iterate the control action, and obtain the first control strategy; use the first control strategy to perform online control, and record the state transition data during the control process; The generation module 205 is used to input the state transfer data into the power loss compensation model to perform power loss compensation control analysis, obtain the compensation control amount, and superimpose the compensation control amount with the first control strategy to obtain the second control strategy, perform PWM modulation based on the second control strategy, and generate corresponding charge and discharge control signals.

[0035] Through the synergy of the above components, by introducing adaptive wavelet transform for data denoising and feature extraction, combined with maximum and minimum value normalization calculation and time series analysis, a complete data preprocessing process was established, which effectively improved the quality and availability of the original data. The extended convolution capsule network was used for hierarchical feature extraction. Through multi-level feature mapping and dynamic routing mechanism, the effective separation of global power supply characteristics and local power supply characteristics was achieved, and the system's ability to perceive the power supply status was enhanced. A control architecture based on a deep deterministic strategy network was designed. Through the synergy of the value evaluation module and the action generation module, a complete state evaluation and action generation system was established. The system adopts a comprehensive mechanism to improve the accuracy and adaptability of the control strategy; builds an experience replay pool to store historical data, adopts random sampling and time difference calculation methods to optimize the control strategy, and ensures continuous optimization and stable convergence of the control strategy through multiple rounds of iterative training and soft update mechanism; proposes a multi-dimensional power loss compensation scheme, which significantly reduces the power loss of the system and improves the power supply efficiency through accurate modeling and compensation of line loss, switching loss and energy storage loss; designs a charge and discharge control mechanism based on PWM modulation, realizes accurate power regulation and energy storage unit control through dead time compensation and triangle wave carrier modulation, and ensures the stable operation of the system.

[0036] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0037] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0038] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A control method for an emergency mobile power supply, characterized in that: The method comprises: The operating parameter data of the emergency mobile power supply is collected and denoised by adaptive wavelet transform to obtain a standardized feature matrix; Using the standardized feature matrix to construct a multidimensional feature vector, and inputting the multidimensional feature vector into an extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features; A deep deterministic strategy network is established based on the global power supply characteristics and the local power supply characteristics, wherein the deep deterministic strategy network includes a value assessment module and an action generation module; the value assessment module is used to score the power supply status to obtain a status scoring result, and the status scoring result and the corresponding control action are stored in an experience replay pool; Inputting the historical data in the experience replay pool into the action generation module, optimizing and iterating the control action to obtain a first control strategy; performing online control using the first control strategy, and recording state transition data during the control process; The state transfer data is input into a power loss compensation model for power loss compensation control analysis to obtain a compensation control amount, and the compensation control amount is superimposed on the first control strategy to obtain a second control strategy. PWM modulation is performed based on the second control strategy to generate corresponding charge and discharge control signals.

2. The control method of the emergency mobile power supply according to claim 1, characterized in that: The operation parameter data of the emergency mobile power supply is collected and the adaptive wavelet transform is used for noise reduction to obtain a standardized feature matrix, including: Collect data on the input voltage, output current, load power and battery status of the emergency mobile power supply to obtain original operating parameter data; Performing data segmentation processing on the original operating parameter data, and calculating the wavelet coefficient of each segment of data to obtain an initial wavelet decomposition result; An adaptive threshold function is set based on the initial wavelet decomposition result to obtain an optimized threshold parameter, and the optimized threshold parameter is substituted into an inverse wavelet transform equation to perform a reconstruction operation on the initial wavelet decomposition result to obtain noise-reduced operating parameter data; Performing maximum and minimum value normalization calculation on the noise-reduced operating parameter data to obtain standardized data, and performing static feature extraction on the standardized data according to inherent parameters of the power supply system to obtain a static feature parameter set; The standardized data is subjected to time series analysis, the voltage fluctuation rate, the power change rate and the load change trend are calculated to obtain a dynamic characteristic parameter set, and the static characteristic parameter set is subjected to matrix concatenation operation with the dynamic characteristic parameter set to obtain the standardized characteristic matrix.

3. The control method of the emergency mobile power supply according to claim 1, characterized in that: The method of constructing a multidimensional feature vector by using the standardized feature matrix and inputting the multidimensional feature vector into an extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features includes: Performing dimension reconstruction and feature combination operations on the standardized feature matrix to obtain a multidimensional feature vector, wherein the multidimensional feature vector includes static feature data and dynamic feature data; Inputting the multidimensional feature vector into the first convolutional layer of the extended convolutional capsule network for feature mapping, wherein the first convolutional layer includes 32 3×3 convolution kernels, uses a ReLU activation function and batch normalization processing, and obtains a primary feature map; Performing a maximum pooling operation on the primary feature map, with a pooling kernel size of 2×2 and a step size of 2, to obtain a reduced-dimensional feature map, and inputting the reduced-dimensional feature map into the main capsule layer of the extended convolutional capsule network; Performing tensor transformation and direction voting calculation on the dimension reduction feature map in the main capsule layer, the main capsule layer includes 8 capsule units, each capsule unit outputs a 16-dimensional vector, and obtaining a main capsule feature; Input the main capsule feature into the digital capsule layer of the extended convolutional capsule network, the digital capsule layer includes 10 capsule units, use a dynamic routing algorithm to perform feature routing, the number of routing iterations is 3, and obtain a routing feature vector; A global average pooling operation is performed on the routing feature vector to obtain a global power supply feature, and a local attention weighted calculation is performed on the routing feature vector to obtain a local power supply feature, wherein the global power supply feature reflects the overall power supply status of the system, and the local power supply feature characterizes the working status of each power supply unit.

4. The control method of the emergency mobile power supply according to claim 1, characterized in that: The method includes establishing a deep deterministic strategy network based on the global power supply characteristics and the local power supply characteristics, wherein the deep deterministic strategy network includes a value assessment module and an action generation module; using the value assessment module to score the power supply status, obtaining a status scoring result, and storing the status scoring result and the corresponding control action in an experience replay pool, including: Perform feature concatenation on the global power supply feature and the local power supply feature to obtain a combined feature vector; Performing dimension mapping on the combined feature vector, and constructing a value assessment module of a deep deterministic strategy network through a three-layer fully connected neural network; Performing dimensionality reduction processing on the combined feature vector, and constructing an action generation module of the deep deterministic strategy network through a four-layer fully connected neural network; The combined feature vector is input into the value assessment module to score the current power supply status to obtain a status scoring result, wherein the status scoring result includes a power supply stability score, a power balance score and a response speed score; Inputting the combined feature vector into the action generation module to generate corresponding control action parameters, wherein the control action parameters include power adjustment instructions and energy storage unit charge and discharge instructions; A reward function is set for the state scoring result, and a reward value is calculated according to power supply stability, power balance and system response speed to obtain a state value evaluation result; An experience replay pool is constructed, the storage capacity of the experience replay pool is 10000, and the state scoring result, the control action parameter and the state value evaluation result are stored in the experience replay pool in chronological order.

5. The control method of the emergency mobile power supply according to claim 4, characterized in that: The historical data in the experience replay pool is input into the action generation module, and the control action is optimized and iterated to obtain a first control strategy; The first control strategy is used to perform online control and record state transition data during the control process, including: Randomly sampling historical data in the experience replay pool to obtain training data batches; Performing time series difference calculation on the training data batch, calculating the target Q value based on the state transition sequence, and the target Q value is used as the optimization target of the value evaluation module; Input the training data batches into the action generation module, perform gradient update on the network parameters of the action generation module, use the Adam optimizer to perform parameter optimization, and set the learning rate to 0.001; The value assessment module is soft-updated, the update step coefficient is set to 0.01, and the parameters of the value assessment module are brought closer to the target network parameters to obtain an updated value assessment module; Based on the updated value assessment module and the action generation module, performing multiple rounds of iterative training on the deep deterministic strategy network to obtain a first control strategy; Deploy the first control strategy to the control system of the emergency mobile power supply, adjust the input voltage, output current and load power in real time, and record the system response data during the adjustment process; The system response data is sampled in sections, and the power supply state parameters and control parameters at each sampling moment are recorded to obtain state transfer data.

6. The control method of the emergency mobile power supply according to claim 1, characterized in that: The step of inputting the state transition data into a power loss compensation model to perform power loss compensation control analysis to obtain a compensation control amount, superimposing the compensation control amount with the first control strategy to obtain a second control strategy, and performing PWM modulation based on the second control strategy to generate a corresponding charge and discharge control signal includes: Analyze and process the state transition data, extract line loss data, switch loss data and energy storage unit loss data, and obtain multi-dimensional loss data; Constructing a power loss compensation model based on the multi-dimensional loss data, and inputting the multi-dimensional loss data into the power loss compensation model to obtain a power loss distribution matrix; Performing weighted sum calculation on the power loss distribution matrix, determining compensation weight coefficients of various types of losses, obtaining a compensation parameter matrix, and calculating a compensation control amount according to the compensation parameter matrix, wherein the compensation control amount includes a line compensation amount, a switch compensation amount, and an energy storage compensation amount; Performing a linear superposition operation on the compensation control amount and the first control strategy to obtain a second control strategy, and performing triangular wave carrier modulation based on the second control strategy to obtain a PWM duty cycle sequence; Generate a power regulation pulse signal and an energy storage unit charge and discharge control signal based on the PWM duty cycle sequence, wherein the rising edge and the falling edge of the power regulation pulse signal correspond to the intersection of the triangular wave carrier; Dead time compensation is performed on the power regulation pulse signal and the energy storage unit charge and discharge control signal to obtain a final charge and discharge control signal.

7. The control method of the emergency mobile power supply according to claim 6, characterized in that: The power loss distribution matrix is ​​weighted and summed to determine the compensation weight coefficients of various types of losses, to obtain a compensation parameter matrix, and a compensation control amount is calculated according to the compensation parameter matrix, wherein the compensation control amount includes a line compensation amount, a switch compensation amount and an energy storage compensation amount, including: Normalizing the line loss data in the power loss distribution matrix, calculating the line impedance based on the input voltage and the output current, and obtaining a line loss weight coefficient; Performing frequency domain analysis on the switching loss data in the power loss distribution matrix, calculating the turn-on loss and turn-off loss based on the switching frequency and the current stress, and obtaining a switching loss weight coefficient; Performing a time series analysis on the energy storage unit loss data in the power loss distribution matrix, calculating the energy storage loss based on the charge and discharge state and the internal resistance change, and obtaining the energy storage loss weight coefficient; The line loss weight coefficient, the switch loss weight coefficient and the energy storage loss weight coefficient are combined to construct a weight matrix, and the weight matrix is ​​normalized to obtain a compensation parameter matrix; Calculating a line power compensation value based on the compensation parameter matrix, wherein the line compensation amount is proportional to the product of the input voltage and the output current, and calculating a switch power compensation value based on the compensation parameter matrix, wherein the switch compensation amount is proportional to the product of the switching frequency and the switching current; The energy storage power compensation value is calculated based on the compensation parameter matrix, the energy storage compensation amount is proportional to the product of the charging and discharging current and the equivalent internal resistance, and the line power compensation value, the switch power compensation value and the energy storage power compensation value are linearly superimposed to obtain the compensation control amount.

8. An emergency mobile power supply, characterized in that: A control method for executing an emergency mobile power supply according to any one of claims 1 to 7, comprising: The acquisition module is used to collect the operating parameter data of the emergency mobile power supply and perform adaptive wavelet transform noise reduction to obtain a standardized feature matrix; A feature extraction module, used to construct a multidimensional feature vector using the standardized feature matrix, and input the multidimensional feature vector into an extended convolutional capsule network for hierarchical feature extraction to obtain global power supply features and local power supply features; An establishment module is used to establish a deep deterministic strategy network based on the global power supply characteristics and the local power supply characteristics, wherein the deep deterministic strategy network includes a value assessment module and an action generation module; the value assessment module is used to score the power supply status to obtain a status scoring result, and the status scoring result and the corresponding control action are stored in an experience replay pool; An iteration module, used for inputting the historical data in the experience replay pool into the action generation module, optimizing and iterating the control action, and obtaining a first control strategy; performing online control using the first control strategy, and recording state transition data during the control process; A generation module is used to input the state transfer data into a power loss compensation model to perform power loss compensation control analysis to obtain a compensation control amount, and superimpose the compensation control amount with the first control strategy to obtain a second control strategy, perform PWM modulation based on the second control strategy, and generate corresponding charge and discharge control signals.

Images