5G base station optical storage integrated power supply control system

Abstract

The invention discloses a 5G base station optical storage integrated power supply control system, and relates to the field of 5G base station optical storage integrated power supply control, and the system comprises a detection module which detects the output voltage of a photovoltaic array, the SOC of an energy storage battery and the impedance parameter of a DC bus, and obtains a detection result; and the 5G wireless sensing module monitors reference signal receiving power and channel state information by initializing a 5GNR physical layer chip based on a detection result, transmits the reference signal receiving power to the micro-control unit through an SPI interface, and reads the maximum power point voltage and current of the photovoltaic module to obtain a time sequence and a photovoltaic array. According to the method, the Lyapunov exponent is solved through phase-space reconstruction and the Wolf algorithm, the chaos state flag bit and the load impedance are generated, the parameters of the LLC resonant converter are dynamically matched, the nonlinear dynamics theory is applied to the field of power supply control, the working modes can be adaptively switched through intelligent recognition and utilization of chaos features, and the power supply efficiency is improved. And optimal scheduling of energy is realized while the ZVS state is ensured.

Description

Technical Field

[0001] The present invention relates to the field of 5G base station integrated optical storage power supply control, and in particular to a 5G base station integrated optical storage power supply control system. Background Art

[0002] In the field of integrated photovoltaic and energy storage power supply control for 5G base stations, conventional technical solutions typically achieve energy scheduling by combining photovoltaic MPPT control, battery energy storage management, and load power regulation. Existing systems are mostly based on fixed-threshold PID control algorithms. By detecting parameters such as the photovoltaic array output voltage and battery SOC, they achieve power conversion in conjunction with DC-DC converters. Under stable operating conditions, they can meet basic power supply requirements and enable data exchange between local devices via the CAN bus or Ethernet. With the high-density deployment of 5G base stations and the growth of dynamic load demands, traditional solutions have further introduced wireless communication modules (such as 4G / LTE) to achieve remote monitoring and optimize energy allocation strategies. This technical approach has formed a relatively mature application system for the integration of standardized photovoltaic and energy storage systems.

[0003] However, conventional methods still have limitations when dealing with the sudden load fluctuations and complex environmental interference of 5G base stations. Traditional PID control relies on linearized models, making it difficult to adapt to the nonlinear coupling characteristics of photovoltaic output and base station load, resulting in limited dynamic response speed. Existing wireless communication modules have insufficient transmission latency and data throughput, restricting the ability to update high-precision energy scheduling strategies in real time. In particular, in multivariable collaborative optimization scenarios, rigid architectures with fixed control parameters make it difficult to achieve a globally optimal match between photovoltaic power generation, storage, and load. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a 5G base station integrated optical storage power supply control system to solve the technical bottleneck problems of traditional methods in terms of dynamic response speed and multi-variable collaborative optimization.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] The present invention provides a 5G base station photovoltaic storage integrated power supply control system, which includes a detection module for detecting the photovoltaic array output voltage, the energy storage battery SOC and the DC bus impedance parameters to obtain the detection results;

[0008] The 5G wireless sensing module monitors the reference signal received power and channel status information by initializing the 5GNR physical layer chip based on the detection results. The reference signal received power is transmitted to the microcontroller unit through the SPI interface, and the maximum power point voltage and current of the photovoltaic module are read to obtain the time series and photovoltaic array;

[0009] The photovoltaic characteristic analysis module uses a sliding time window to spatially reconstruct the time series and photovoltaic array, and uses the Wolf algorithm to obtain the Lyapunov exponent. The microcontroller unit generates the chaotic state flag and load impedance based on the time series and photovoltaic array.

[0010] The chaos control decision module inputs the chaos state flag and load impedance into the digital controller of the LLC resonant converter, dynamically matches the resonant cavity parameters according to the load impedance, makes the converter operate in the ZVS state, triggers the pre-discharge mode, and obtains the PWM control signal and battery current command;

[0011] LLC resonant control module, PWM control signal and battery current instruction are sent to the execution unit through the CAN bus to obtain operation data, which is uploaded to the cloud through the 5GSA module, and the control strategy update package is generated through the LSTM-PSO hybrid optimization model.

[0012] As a preferred solution of the 5G base station photovoltaic storage integrated power supply control system of the present invention, wherein: detecting the photovoltaic array output voltage, energy storage battery SOC and DC bus impedance parameters to obtain the detection results includes the following steps:

[0013] A high-precision ADC is used to sample the output voltage of the photovoltaic array. The remaining battery capacity is calculated using the Coulomb counting method combined with the open-circuit voltage method to obtain the percentage value SOC. The phase difference between the response current and voltage is analyzed through FFT to obtain the DC bus impedance parameters.

[0014] The photovoltaic array output voltage, percentage value SOC and DC bus impedance parameters are summarized to obtain the detection results.

[0015] As a preferred solution of the 5G base station optical storage integrated power supply control system of the present invention, based on the detection results, the reference signal receiving power and channel state information are monitored by initializing the 5GNR physical layer chip, including the following steps:

[0016] Based on the detection results, a hardware status verification signal is generated and the 5GNR physical layer chip initialization permission is unlocked. The low-power sensing mode of the baseband chip is activated through the verification signal, the cyclic prefix detection parameters of the CSI-RS and SSB signals are configured, and the physical layer ready flag is output;

[0017] The initialized physical layer chip is used to capture the reference signal received power and channel state information in real time, and the original signal data set is output through the SPI interface;

[0018] Extract subcarrier power distribution and multipath delay characteristics from the original signal data set, generate wireless channel feature vectors, dynamically adjust the power allocation ratio of 5G resource blocks, and obtain a wireless resource configuration table;

[0019] The wireless resource configuration table is temporally and spatially aligned with the MPPT parameters of the photovoltaic array and the battery SOC to generate a joint control feature matrix.

[0020] The triggered chaos control mode is judged based on the joint control characteristic matrix to obtain the reference signal receiving power and channel state information.

[0021] As a preferred solution of the 5G base station integrated photovoltaic storage power supply control system described in the present invention, the reference signal received power is transmitted to the microcontroller unit through the SPI interface, the maximum power point voltage and current of the photovoltaic module are read, and the time series and photovoltaic array are obtained, including the following steps:

[0022] Receive the reference signal received power from the 5G physical layer chip through the SPI interface to obtain a timestamp-aligned reference signal received power sampling sequence;

[0023] When the reference signal receiving power sampling sequence aligned with the timestamp arrives, the ADC is triggered to synchronously sample and read the maximum power point voltage and current of the photovoltaic module;

[0024] The maximum power point voltage and current of the photovoltaic module are used to construct the IV characteristic curve of the photovoltaic array, and the time series and photovoltaic array are obtained.

[0025] As a preferred solution of the 5G base station integrated photovoltaic storage power supply control system of the present invention, wherein: a sliding time window is used to reconstruct the phase space of the time series and the photovoltaic array, including the following steps:

[0026] Divide data subsets based on time series and photovoltaic arrays;

[0027] The voltage-current series in each data subset window is embedded with time delay, and the embedding dimension and time delay are obtained using Takens' theorem, and the phase space of the time series and photovoltaic array is reconstructed.

[0028] As a preferred solution of the 5G base station integrated optical storage power supply control system of the present invention, wherein: the Lyapunov index is obtained by the Wolf algorithm, including the following steps:

[0029] Based on the phase space of the reconstructed time series and photovoltaic array, the initial reference point is selected and the nearest neighbor point is calculated to obtain the initial trajectory pair set;

[0030] Track the evolution of time steps for each trajectory pair, record the distance change ratio of adjacent trajectories, and obtain the distance logarithm sequence;

[0031] A least squares linear fit is performed on the logarithmic series of distances, and the fitting slope is the maximum Lyapunov exponent.

[0032] As a preferred solution of the 5G base station integrated photovoltaic storage power supply control system of the present invention, the microcontroller generates a chaotic state flag and load impedance according to the time series and the photovoltaic array, including the following steps:

[0033] Based on the Lyapunov exponent value, a Lyapunov exponent value threshold is set;

[0034] When the Lyapunov exponent value is greater than the Lyapunov exponent value threshold, a chaotic state flag is generated;

[0035] Calculate the load impedance based on the real-time maximum power point voltage and current of the PV array.

[0036] As a preferred solution of the 5G base station integrated optical storage power supply control system described in the present invention, the chaotic state flag and load impedance are input into the digital controller of the LLC resonant converter, the resonant cavity parameters are dynamically matched according to the load impedance, the converter is operated in the ZVS state, the pre-discharge mode is triggered, and the PWM control signal and battery current instruction are obtained, including the following steps:

[0037] Receive the chaos state flag and load impedance sent by the microcontroller through the CAN bus, and output them as a control instruction parameter set after parsing;

[0038] According to the amplitude and phase angle of the load impedance, the resonant inductor and capacitor parameters of the LLC resonant converter are dynamically adjusted to obtain the resonant frequency;

[0039] Monitor the MOSFET's Vds voltage zero crossing point in real time at the resonant frequency to obtain the ZVS lock signal and switching frequency.

[0040] The switching frequency is increased to the reference frequency and a chaotic modulation wave is injected to generate a PWM signal with chaotic characteristics;

[0041] When a sudden load change is detected and the ZVS lock is valid, a PWM control signal and a battery current command are obtained.

[0042] As a preferred solution of the 5G base station integrated power supply control system of the present invention, the PWM control signal and the battery current instruction are sent to the execution unit through the CAN bus to obtain the operation data, including the following steps:

[0043] The PWM control signal and battery current command generated by chaotic modulation are encapsulated into data frames according to the CAN protocol format to obtain a standardized control command packet;

[0044] Sending standardized control instruction packets to the execution unit via the CAN bus to obtain a transmission status confirmation signal;

[0045] The execution unit receives and parses the standardized control instruction package, drives the power MOSFET to act according to the PWM signal, adjusts the battery discharge to the specified value, and obtains the operating parameters.

[0046] As a preferred solution of the 5G base station integrated optical storage power supply control system described in the present invention, the operation data is uploaded to the cloud through the 5GSA module, and the control strategy update package is generated through the LSTM-PSO hybrid optimization model, including the following steps:

[0047] The operating parameters are encrypted and packaged through the 5G independent networking module and uploaded to the cloud server via the UDP protocol to obtain a successful transmission confirmation signal;

[0048] Based on the transmission success confirmation signal, a multi-dimensional feature matrix is ​​generated;

[0049] The multidimensional feature matrix is ​​input into the LSTM-PSO hybrid model to generate a control strategy update package.

[0050] The beneficial effects of the present invention are: deep coupling of communication and energy is achieved through the 5G wireless sensing module. The module monitors the reference signal receiving power and channel status information in real time based on the 5GNR physical layer chip, and transmits the wireless signal characteristics to the microcontroller unit through the SPI interface. It not only realizes the accurate perception of the base station load status, but also builds a data foundation for the coordinated optimization of communication and energy supply through the spatiotemporal alignment and fusion of wireless channel characteristics and photovoltaic characteristics. It can predict load fluctuations in advance and respond to them. The chaos control decision module obtains the Lyapunov exponent through phase space reconstruction and Wolf algorithm, generates chaos state flags and load impedance, and dynamically matches LLC resonant converter parameters, applying nonlinear dynamics theory to the field of power supply control. Through the intelligent identification and utilization of chaos characteristics, it can adaptively switch the working mode and achieve optimal energy scheduling while ensuring the ZVS state. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 Schematic diagram of the integrated optical storage power supply control system for 5G base stations.

[0053] Figure 2 Schematic diagram for finding the maximum Lyapunov exponent. DETAILED DESCRIPTION

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0055] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0056] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0057] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a 5G base station integrated optical storage power supply control system, including the following steps:

[0058] The detection module detects the output voltage of the photovoltaic array, the SOC of the energy storage battery and the DC bus impedance parameters to obtain the detection results.

[0059] Detect the photovoltaic array output voltage, energy storage battery SOC and DC bus impedance parameters to obtain the detection results.

[0060] A high-precision ADC is used to sample the output voltage of the photovoltaic array. The remaining battery capacity is calculated by combining the coulomb counting method with the open-circuit voltage method to obtain the percentage value SOC. The phase difference between the response current and voltage is analyzed by FFT to obtain the DC bus impedance parameters.

[0061] Furthermore, a high-precision ADC samples the output voltage of the photovoltaic array in real time to obtain the DC voltage signal of the photovoltaic array. Based on the sampled photovoltaic array output voltage, the coulomb counting method is used to accumulate the battery charge and discharge capacity. Combined with the open-circuit voltage method, the battery's remaining capacity percentage SOC is calculated through the mapping relationship between the battery's static voltage and the SOC calibration curve. When the battery is in working condition, a small 1kHz AC disturbance signal is injected into the DC bus. The amplitude ratio and phase difference of the disturbance response current and voltage are analyzed through fast Fourier transform, and the fundamental component is extracted to obtain the DC bus impedance parameters.

[0062] The photovoltaic array output voltage, percentage value SOC and DC bus impedance parameters are summarized to obtain the detection results.

[0063] Furthermore, the PV array output voltage sampled by the high-precision ADC, the percentage SOC calculated by the Coulomb counting method and the open-circuit voltage method, and the amplitude and phase angle of the DC bus impedance parameter obtained by fast Fourier transform analysis are aligned and formatted in a standardized manner. The PV array output voltage, percentage SOC, and DC bus impedance parameter are then packaged into a structured data packet according to a predefined communication protocol. After data verification, the final test result is output.

[0064] Based on the detection results, the 5G wireless sensing module monitors the reference signal receiving power and channel status information by initializing the 5GNR physical layer chip. The reference signal receiving power is transmitted to the microcontroller unit through the SPI interface, and the maximum power point voltage and current of the photovoltaic module are read to obtain the time series and photovoltaic array.

[0065] Initialize the 5GNR physical layer chip to monitor the reference signal receiving power and channel status information.

[0066] Based on the detection results, a hardware status verification signal is generated and the 5GNR physical layer chip initialization permission is unlocked. The low-power perception mode of the baseband chip is activated through the verification signal, the cyclic prefix detection parameters of the CSI-RS and SSB signals are configured, and the physical layer ready flag is output.

[0067] Furthermore, a logical AND operation is performed based on the photovoltaic array output voltage, percentage value SOC and DC bus impedance parameters in the detection results. When the photovoltaic array output voltage is within the normal operating range, a high-level hardware status verification signal is generated. The hardware status verification signal is connected to the enable pin of the 5GNR physical layer chip to release the chip initialization lock state. During the initialization process, the baseband chip is configured to operate in low-power perception mode, the period of the channel state information reference signal CSI-RS is set to 20ms, and the cyclic prefix type of the synchronization signal block SSB is an extended cyclic prefix. After completing the parameter configuration, the 5GNR physical layer chip outputs a high-level physical layer ready flag signal.

[0068] The initialized physical layer chip is used to capture the reference signal received power and channel state information in real time, and the original signal data set is output through the SPI interface.

[0069] Furthermore, the initialized 5GNR physical layer chip starts the downlink reference signal received power measurement function when the physical layer ready flag signal is valid, performs power detection on the channel state information reference signal CSI-RS in the 3.5GHz frequency band, and demodulates the primary synchronization signal and secondary synchronization signal in the synchronization signal block SSB. The baseband processing unit calculates the reference signal received power value and channel impulse response on each resource block, and encapsulates the original data containing the subcarrier index, reference signal received power value, and channel matrix element according to the data frame format of the SPI communication protocol. The encapsulated original signal data set is transmitted to the micro control unit through the MOSI signal line at a SPI clock frequency of 1MHz, and the confirmation response signal from the micro control unit is received through the MISO signal line to complete the data transmission process.

[0070] The subcarrier power distribution and multipath delay characteristics are extracted from the original signal data set to generate the wireless channel feature vector. The power allocation ratio of the 5G resource block is dynamically adjusted to obtain the wireless resource configuration table.

[0071] Furthermore, the original signal data set received through the SPI interface is parsed, and the power spectral density distribution characteristics of each subcarrier are extracted from the channel state information reference signal CSI-RS. The delay spread parameters of the multipath channel are calculated by inverse fast Fourier transform, and the subcarrier power distribution characteristics and multipath delay characteristics are combined in the frequency domain-time domain dimension to construct a wireless channel feature vector containing 120-dimensional features. Based on the channel quality indication parameters in the wireless channel feature vector, the proportional fairness algorithm is used to dynamically calculate the power allocation weight coefficient of each 5G resource block, and generate a wireless resource configuration table containing resource block index, transmit power, and modulation and coding scheme.

[0072] The wireless resource configuration table is spatially and temporally aligned with the MPPT parameters of the photovoltaic array and the battery SOC to generate a joint control feature matrix.

[0073] Furthermore, the resource block transmission power parameters in the wireless resource configuration table are timestamp-aligned with the output voltage and output current in the maximum power point tracking (MPPT) parameters of the photovoltaic array. At the same time, the percentage value SOC is interpolated to the same time series according to the sampling period. The Kalman filter algorithm is used to eliminate the time synchronization error between the wireless resource configuration table and the photovoltaic array MPPT parameters. A four-dimensional joint control feature matrix containing the resource block transmission power, photovoltaic array output voltage, photovoltaic array output current and percentage value SOC is established. The rows of the matrix correspond to the sampling time and the columns correspond to the feature dimensions, completing the spatiotemporal alignment and feature fusion of multi-source data.

[0074] The triggered chaos control mode is judged based on the joint control characteristic matrix to obtain the reference signal receiving power and channel state information.

[0075] Furthermore, principal component analysis and dimensionality reduction processing is performed on the resource block transmission power, photovoltaic array output voltage, photovoltaic array output current and percentage value SOC in the joint control feature matrix, and the first three principal components are extracted to form a feature vector. The feature vector is input into a pre-trained support vector machine classifier. When the classifier output result is 1, it is determined that the chaos control mode needs to be triggered. When the chaos control mode is triggered, the reference signal receiving power value and channel state information matrix at the current moment are re-collected, the corresponding parameters in the wireless channel feature vector are updated, and the adjusted reference signal receiving power and channel state information are output for subsequent control decisions.

[0076] The reference signal received power is transmitted to the microcontroller unit through the SPI interface, and the maximum power point voltage and current of the photovoltaic module are read to obtain the time series and photovoltaic array.

[0077] The reference signal received power is received from the 5G physical layer chip through the SPI interface to obtain a reference signal received power sampling sequence with timestamp alignment.

[0078] Furthermore, the 5G physical layer chip transmits the reference signal received power data frame through the MOSI signal line of the SPI interface at a clock frequency of 1MHz. The data frame contains a 16-bit timestamp and a 24-bit reference signal received power value. The microcontroller unit samples the MOSI signal line data at the rising edge of the SPI clock, stores the received data frame in the buffer, parses the continuous data frames in the buffer, extracts the timestamp field and the reference signal received power value field, arranges them in the order of timestamps, and generates a timestamp-aligned reference signal received power sampling sequence. Each data point in the sequence contains a 32-bit timestamp and a corresponding 24-bit reference signal received power value.

[0079] When the reference signal received power sampling sequence aligned with the timestamp arrives, the ADC is triggered to synchronously sample and read the maximum power point voltage and current of the photovoltaic module.

[0080] Furthermore, when a new data point arrives in the timestamp-aligned reference signal received power sampling sequence, the microcontroller sends a sampling trigger pulse to the ADC through the GPIO pin. After receiving the trigger pulse, the ADC immediately performs synchronous sampling of the output voltage and output current of the PV module. The output voltage sampling channel measures the DC side voltage of the PV array, and the output current sampling channel measures the output signal of the Hall sensor connected in series in the PV circuit. The ADC transmits the sampled 12-bit PV module output voltage value and 12-bit output current value to the microcontroller via a parallel data bus. The microcontroller binds the voltage and current values ​​to the timestamp of the reference signal received power sampling sequence at the trigger moment to form a synchronous data record containing the timestamp, reference signal received power, PV module output voltage and output current.

[0081] The maximum power point voltage and current of the photovoltaic module are used to construct the IV characteristic curve of the photovoltaic array, and the time series and photovoltaic array are obtained.

[0082] Furthermore, based on the output voltage and output current values ​​of the photovoltaic modules in the synchronous data records, discrete data points are plotted in the voltage-current two-dimensional coordinate system, and the cubic spline interpolation algorithm is used to fit and generate a continuous photovoltaic array IV characteristic curve. The photovoltaic array IV characteristic curves at different sampling times are arranged in timestamp order to construct three-dimensional time series data containing time dimension, voltage dimension and current dimension; the time series data is subjected to sliding average filtering to eliminate random measurement noise, and finally a smooth photovoltaic array IV characteristic curve time series and a set of photovoltaic array electrical parameters under the current working state are output.

[0083] The photovoltaic characteristic analysis module uses a sliding time window to spatially reconstruct the time series and photovoltaic array, and obtains the Lyapunov exponent through the Wolf algorithm. The microcontroller unit generates the chaotic state flag and load impedance based on the time series and photovoltaic array.

[0084] A sliding time window is used to spatially reconstruct the time series and photovoltaic array.

[0085] Data subsets were divided based on time series and photovoltaic arrays.

[0086] Furthermore, the three-dimensional data in the time series of the IV characteristic curve of the photovoltaic array is segmented into sliding windows according to the time dimension, with a window length of 100 consecutive sampling points and a sliding step of 10 sampling points. The data in each window constitutes a data subset, which contains 100 sets of timestamps, the corresponding relationship between the output voltage of the photovoltaic module and the output current of the photovoltaic module. Each data subset is normalized, and the values ​​of the output voltage and output current of the photovoltaic module are linearly transformed to the interval [0, 1], while the original timestamps are kept unchanged. The standardized data subset is generated for subsequent phase space reconstruction.

[0087] The voltage-current series in each data subset window is embedded with time delay, and the embedding dimension and time delay are obtained using Takens' theorem, and the phase space of the time series and photovoltaic array is reconstructed.

[0088] Furthermore, the autocorrelation functions of the PV module output voltage and PV module output current time series in the standardized data subset are calculated respectively, and the time delay parameter is determined by the first zero crossing point of the autocorrelation function. The false neighbor ratio of the PV module output voltage and PV module output current time series under different embedding dimensions is calculated by the false neighbor method. When the false neighbor ratio is less than 5%, the corresponding minimum dimension is used as the embedding dimension. According to Takens' theorem, the PV module output voltage and PV module output current time series are reconstructed in phase space according to the time delay parameter and embedding dimension, and a set of trajectory points in the dimensional phase space is generated. Each trajectory point contains the state value of the PV module output voltage and PV module output current at the time delay point, thereby completing the phase space reconstruction of the dynamic characteristics of the photovoltaic array.

[0089] The Lyapunov exponent is obtained using the Wolf algorithm.

[0090] Based on the phase space of the reconstructed time series and photovoltaic array, the initial reference points are selected and the nearest neighbor points are calculated to obtain the initial trajectory pair set.

[0091] Furthermore, in the phase space of the reconstructed time series and the photovoltaic array, 100 points are randomly selected from the trajectory point set of the photovoltaic module output voltage and the photovoltaic module output current as initial reference points; for each initial reference point, the Euclidean distance of all other trajectory points in the phase space is calculated, and the non-adjacent trajectory point with the smallest Euclidean distance is selected as the nearest neighbor point. The initial reference point is paired with the corresponding nearest neighbor point to form an initial trajectory pair set containing 100 pairs of trajectory points.

[0092] The evolution of the time step of each trajectory pair is tracked, and the distance change ratio of adjacent trajectories is recorded to obtain a distance logarithmic sequence.

[0093] Furthermore, for each pair of trajectory points in the initial trajectory pair set, the phase space is evolved forward for 10 time steps, and each step corresponds to the sampling interval of the original time series. After each evolution step, the Euclidean distance between the evolved reference point trajectory and the nearest neighbor point trajectory is recorded, and the natural logarithm of the ratio of the current distance to the initial distance is recorded. The natural logarithm of the distance change ratio of 100 pairs of trajectory points within 10 time steps is arranged in chronological order to generate a distance logarithm sequence containing 1000 data points. Each data point in the sequence represents the logarithm of the distance change rate of a specific trajectory pair under a specific evolution step.

[0094] A least squares linear fit is performed on the logarithmic series of distances, and the fitting slope is the maximum Lyapunov exponent.

[0095] Furthermore, a least-squares linear regression equation was constructed using the 1,000 data points in the distance logarithm sequence according to the evolution time step as the independent variable and the natural logarithm of the distance change rate as the dependent variable. The slope parameter of the regression line was calculated by solving the normal equation system. This slope parameter is the maximum Lyapunov exponent that characterizes the dynamic characteristics of the photovoltaic array. The regression results were tested for significance. When the determination coefficient was greater than 0.85, the fit was confirmed to be valid. The maximum Lyapunov exponent was output as a quantitative indicator of chaotic characteristics.

[0096] Specifically, the expression is,

[0097]

[0098] Where λ is the maximum Lyapunov exponent, y i is the dependent variable at the i-th time step, t i is the independent variable at the i-th time step, N is the total number of data points in the distance logarithm sequence, and i is the index of the time step.

[0099] The micro control unit generates a chaotic state flag and load impedance according to the time series and the photovoltaic array.

[0100] Based on the Lyapunov exponent value, a Lyapunov exponent threshold is set.

[0101] Furthermore, the maximum Lyapunov exponent value distribution calculated in the photovoltaic array under stable working state and chaotic state is divided into two categories using the K-means clustering algorithm. The middle value of the center points of the two categories is taken as the Lyapunov index value threshold. When the maximum Lyapunov exponent value calculated in real time exceeds the threshold, it is determined that the photovoltaic array enters a chaotic state. The Lyapunov index value threshold is stored in the non-volatile memory of the microcontroller unit.

[0102] Specifically, the expression is,

[0103]

[0104] Among them, λ th is the Lyapunov index value threshold, μ0 is the cluster center of the stable state samples, and μ1 is the cluster center of the chaotic state samples.

[0105] When the Lyapunov exponent value is greater than the Lyapunov exponent value threshold, a chaotic state flag is generated.

[0106] Furthermore, the maximum Lyapunov exponent value is compared with the Lyapunov exponent value threshold stored in the non-volatile memory. When the maximum Lyapunov exponent value is greater than the Lyapunov exponent value threshold, the microcontroller unit sets the chaos state flag register to 1; when the maximum Lyapunov exponent value is less than or equal to the Lyapunov exponent value threshold, the chaos state flag register remains at 0; the value in the chaos state flag register is transmitted to the digital signal processor via the data bus for subsequent control decision-making.

[0107] Calculate the load impedance based on the real-time maximum power point voltage and current of the PV array.

[0108] Furthermore, the real-time maximum power point voltage and current values ​​of the photovoltaic array are obtained by synchronous sampling. The maximum power point voltage of the photovoltaic array is divided by the maximum power point current to obtain the amplitude of the load impedance. The phase difference between the output voltage and output current of the photovoltaic array is analyzed by fast Fourier transform, and the phase angle of the load impedance is calculated. The load impedance amplitude and phase angle are converted into complex form and expressed as the sum of the real part and the imaginary part. The load impedance parameters containing impedance amplitude and phase information are generated for the subsequent parameter matching control of the LLC resonant converter.

[0109] Specifically, the expression is,

[0110]

[0111] Among them, Z load is the load impedance, V MPP is the voltage, I MPP For current.

[0112] The chaos control decision module inputs the chaos state flag and load impedance into the digital controller of the LLC resonant converter, dynamically matches the resonant cavity parameters according to the load impedance, makes the converter operate in the ZVS state, triggers the pre-discharge mode, and obtains the PWM control signal and battery current instruction.

[0113] The chaotic state flag and load impedance are input into the digital controller of the LLC resonant converter. The resonant cavity parameters are dynamically matched according to the load impedance, so that the converter operates in the ZVS state, triggering the pre-discharge mode and obtaining the PWM control signal and battery current instruction.

[0114] The chaotic state flag and load impedance sent by the microcontroller unit are received via the CAN bus and are output as a control instruction parameter set after parsing.

[0115] Furthermore, the digital controller receives data frames from the microcontroller unit at a rate of 1 Mbps via the CAN bus. The data frames include a 1-bit chaos state flag, a 16-bit real part of the load impedance, and a 16-bit imaginary part of the load impedance. A CRC check is performed on the received data frame. After the check passes, the payload data is extracted, the chaos state flag is stored in a status register, the real part and the imaginary part of the load impedance are converted into floating-point format and stored in a parameter buffer, and the control instruction parameter set including the chaos state flag, the real part of the load impedance, and the imaginary part of the load impedance is generated by combining them.

[0116] According to the amplitude and phase angle of the load impedance, the resonant inductor and capacitor parameters of the LLC resonant converter are dynamically adjusted to obtain the resonant frequency.

[0117] Furthermore, the resonant parameter calculation unit reads the real and imaginary parts of the load impedance in the control instruction parameter set, calculates the load impedance amplitude through square and square root operations, and calculates the load impedance phase angle through the inverse tangent function; queries the pre-stored resonant parameter mapping table according to the load impedance amplitude and phase angle, and matches the optimal resonant inductance reference value and resonant capacitance reference value; the digital signal processor adjusts the control current of the variable inductor through the PWM signal, adjusts the resonant inductance to the reference value, and adjusts the resonant capacitance to the reference value through the digital potentiometer; based on the adjusted resonant inductance and resonant capacitance parameters, the real-time resonant frequency is calculated and output to the switching frequency control unit.

[0118] Specifically, the expression is,

[0119]

[0120] Where f is the resonant frequency, C is the resonant capacitance, and L is the resonant inductance.

[0121] The Vds voltage zero crossing point of the MOSFET is monitored in real time at the resonant frequency to obtain the ZVS locking signal and switching frequency.

[0122] Furthermore, the switching frequency control unit outputs the resonant frequency as the initial switching frequency to the gate drive circuit to drive the MOSFET. The high-speed comparator monitors the drain-source voltage Vds of the MOSFET in real time. When the Vds voltage is detected to cross zero, a pulse signal is generated. The digital phase-locked loop captures the phase relationship between the pulse signal and the drive signal. When the Vds zero crossing is detected to be synchronized with the falling edge of the drive signal for five consecutive cycles, a high-level ZVS lock signal is output. At the same time, the phase-locked loop feeds back the adjusted switching frequency to the switching frequency control unit to form a closed-loop regulation and output a stable ZVS lock signal and switching frequency.

[0123] The switching frequency is increased to the reference frequency and a chaotic modulation wave is injected to generate a PWM signal with chaotic characteristics.

[0124] Furthermore, when the ZVS locking signal is valid, the digital signal processor increases the switching frequency to 120% of the reference frequency as the carrier frequency. Based on the Lyapunov exponent value in the chaos state flag register, a chaotic sequence is iteratively generated through logistic mapping, and normalized as the modulation wave. A bipolar modulation method is used to compare the chaotic modulation wave with the triangular carrier. When the instantaneous value of the modulation wave is greater than the carrier, a high level is output, otherwise a low level is output, generating a PWM signal with chaotic characteristics and a duty cycle that randomly fluctuates in the range of 45% to 65%. The frequency of the PWM signal is the increased reference frequency, and the pulse width is dynamically controlled by the chaotic modulation wave.

[0125] When a sudden load change is detected and the ZVS lock is valid, a PWM control signal and a battery current command are obtained.

[0126] Furthermore, the digital signal processor continuously monitors the real-time change rate of the load impedance. When the load impedance change rate exceeds 10% / ms and the ZVS lock signal remains at a high level, the control instruction update is triggered, and the PWM signal with chaotic characteristics is output to the gate drive circuit as a PWM control signal after being isolated by an optocoupler. At the same time, based on the load impedance mutation amplitude and the chaotic state flag, the battery charge and discharge current command value is calculated according to a preset lookup table method, and a joint control instruction set including the PWM control signal and the battery current command is output, wherein the PWM control signal frequency maintains the increased base frequency, and the battery current command value is a signed 16-bit integer value, where a positive value indicates charging and a negative value indicates discharging.

[0127] LLC resonant control module, PWM control signal and battery current instruction are sent to the execution unit through the CAN bus to obtain operation data, which is uploaded to the cloud through the 5GSA module, and the control strategy update package is generated through the LSTM-PSO hybrid optimization model.

[0128] The PWM control signal and battery current command are sent to the execution unit via the CAN bus to obtain the operating data.

[0129] The PWM control signal and battery current command generated by chaotic modulation are encapsulated into data frames according to the CAN protocol format to obtain a standardized control command packet.

[0130] Furthermore, the digital signal processor converts the PWM control signal generated by chaotic modulation into an 8-bit duty cycle value, which is combined with the 16-bit battery current instruction into 24 bits of valid data. The data is encapsulated according to the CAN2.0B extended frame format, and the 29-bit identifier field is set to 0x18FFA001. The data field contains the duty cycle value and the battery current instruction, which is padded to 8 bytes. A 15-bit CRC checksum and a 1-bit ACK confirmation bit are added to generate a complete standardized control instruction packet, which is then sent to the CAN controller buffer for transmission.

[0131] The standardized control instruction packet is sent to the execution unit via the CAN bus to obtain a transmission status confirmation signal.

[0132] Furthermore, the CAN controller sends standardized control instruction packets to the CAN bus at a rate of 1Mbps, and the data frame transmits the SOF identifier, control segment, data segment, CRC sequence and ACK slot in sequence; after successfully receiving the data frame, the execution unit replies with a dominant bit confirmation signal within the ACK slot period. The CAN controller monitors the bus status. When a complete ACK confirmation signal is detected and no error frame is generated, the successful transmission status register is set and a high-level transmission status confirmation signal is output. If an error frame is detected or the ACK times out, the failure transmission counter is incremented by one and the retransmission mechanism is triggered until the maximum number of retransmissions is reached and a low-level transmission status confirmation signal is output.

[0133] The execution unit receives and parses the standardized control instruction package, drives the power MOSFET to act according to the PWM signal, adjusts the battery discharge to the specified value, and obtains the operating parameters.

[0134] Furthermore, the CAN transceiver of the execution unit receives a standardized control instruction packet, extracts the 8-bit PWM duty cycle value and the 16-bit battery current instruction after CRC check, and the PWM generator generates a corresponding drive signal according to the duty cycle value, controls the on and off of the power MOSFET through the gate driver, and the current closed-loop controller collects the battery discharge current in real time, generates an adjustment signal after comparing it with the battery current instruction, and adjusts the PWM duty cycle to make the discharge current track the instruction value, and synchronously monitors the Vds voltage, battery terminal voltage and temperature parameters of the power MOSFET. After ADC conversion, it is stored as an operating parameter data packet containing voltage, current and temperature.

[0135] The operating data is uploaded to the cloud through the 5GSA module, and the control strategy update package is generated through the LSTM-PSO hybrid optimization model.

[0136] The operating parameters are encrypted and packaged through the 5G independent networking module, uploaded to the cloud server via the UDP protocol, and a confirmation signal of successful transmission is obtained.

[0137] Furthermore, the operating parameter data packet is encrypted using the AES-256 algorithm and encapsulated as a UDP datagram. The source port is set to a random high-order port and the destination port is fixed to 5683. The 5G independent networking module establishes an NR-Uu interface connection, maps the encrypted UDP datagram to the GBR service flow of 5GQoSFlow, and sends it to the gNodeB after encryption through the PDCP layer. After the cloud server successfully receives the data, it decrypts and verifies the data integrity and returns an ACK confirmation message containing a receiving timestamp. After receiving the ACK message, the 5G independent networking module parses the timestamp and generates a transmission success confirmation signal containing latency indicators and transmission status.

[0138] Based on the transmission success confirmation signal, a multi-dimensional feature matrix is ​​generated.

[0139] Furthermore, the delay index and transmission status parameters in the transmission success confirmation signal are analyzed and time-aligned with the voltage, current, and temperature parameters in the locally stored operating parameter data packet. The parameters of the five dimensions of delay index, transmission status, voltage, current, and temperature are arranged in a time series. Each time point corresponds to a set of five-dimensional feature vectors. The Z-score normalization method is used to normalize the data of each dimension. After eliminating the dimensional difference, a multidimensional feature matrix is ​​constructed in which rows represent time points and columns represent feature dimensions. The matrix elements are the normalized parameter values.

[0140] The multidimensional feature matrix is ​​input into the LSTM-PSO hybrid model to generate a control strategy update package.

[0141] Furthermore, the LSTM neural network performs time series modeling on the multi-dimensional feature matrix, extracts the timing features through three-layer LSTM units and outputs a state vector. The particle swarm optimization algorithm uses the state vector output by LSTM as the initial population and searches for the optimal control parameter combination in the solution space. After 50 generations of iteration, the particle position parameters with the highest fitness are selected as the optimization results, including the PWM duty cycle adjustment amount, the battery current command correction value and the impedance matching coefficient. The optimized parameters are packaged into a control strategy update package in a predefined format, which contains a 16-bit header identifier, 48-bit parameter data and an 8-bit check code, and sent to the local execution unit through the 5G independent networking module.

[0142] In summary, the present invention realizes the deep coupling of communication and energy through the 5G wireless sensing module. The module monitors the reference signal receiving power and channel status information in real time based on the 5GNR physical layer chip, and transmits the wireless signal characteristics to the microcontroller unit through the SPI interface. It not only realizes the accurate perception of the base station load status, but also builds a data foundation for the coordinated optimization of communication and energy supply through the spatiotemporal alignment and fusion of wireless channel characteristics and photovoltaic characteristics. It can predict load fluctuations in advance and respond to them. The chaos control decision module obtains the Lyapunov exponent through phase space reconstruction and Wolf algorithm, generates chaos state flag and load impedance, and dynamically matches LLC resonant converter parameters, applying nonlinear dynamics theory to the field of power supply control. Through the intelligent identification and utilization of chaos characteristics, it can adaptively switch the working mode and achieve optimal energy scheduling while ensuring the ZVS state.

[0143] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A 5G base station integrated optical storage power supply control system, characterized by: include, The detection module detects the output voltage of the photovoltaic array, the SOC of the energy storage battery and the DC bus impedance parameters to obtain the detection results; The 5G wireless sensing module monitors the reference signal received power and channel status information by initializing the 5GNR physical layer chip based on the detection results. The reference signal received power is transmitted to the microcontroller unit through the SPI interface, and the maximum power point voltage and current of the photovoltaic module are read to obtain the time series and photovoltaic array; The photovoltaic characteristic analysis module uses a sliding time window to spatially reconstruct the time series and photovoltaic array, and uses the Wolf algorithm to obtain the Lyapunov exponent. The microcontroller unit generates the chaotic state flag and load impedance based on the time series and photovoltaic array. The chaos control decision module inputs the chaos state flag and load impedance into the digital controller of the LLC resonant converter, dynamically matches the resonant cavity parameters according to the load impedance, makes the converter operate in the ZVS state, triggers the pre-discharge mode, and obtains the PWM control signal and battery current command; LLC resonant control module, PWM control signal and battery current instruction are sent to the execution unit through the CAN bus to obtain operation data, which is uploaded to the cloud through the 5GSA module, and the control strategy update package is generated through the LSTM-PSO hybrid optimization model.

2. The 5G base station integrated optical-storage power supply control system according to claim 1, characterized in that: Detecting the photovoltaic array output voltage, energy storage battery SOC and DC bus impedance parameters to obtain the detection results includes the following steps: A high-precision ADC is used to sample the output voltage of the photovoltaic array. The remaining battery capacity is calculated using the Coulomb counting method combined with the open-circuit voltage method to obtain the percentage value SOC. The phase difference between the response current and voltage is analyzed through FFT to obtain the DC bus impedance parameters. The photovoltaic array output voltage, percentage value SOC and DC bus impedance parameters are summarized to obtain the detection results.

3. The 5G base station integrated optical storage power supply control system according to claim 2, characterized in that: Based on the detection results, the reference signal received power and channel state information are monitored by initializing the 5GNR physical layer chip, including the following steps: Based on the detection results, a hardware status verification signal is generated and the 5GNR physical layer chip initialization permission is unlocked. The low-power sensing mode of the baseband chip is activated through the verification signal, the cyclic prefix detection parameters of the CSI-RS and SSB signals are configured, and the physical layer ready flag is output; The initialized physical layer chip is used to capture the reference signal received power and channel state information in real time, and the original signal data set is output through the SPI interface; Extract subcarrier power distribution and multipath delay characteristics from the original signal data set, generate wireless channel feature vectors, dynamically adjust the power allocation ratio of 5G resource blocks, and obtain a wireless resource configuration table; The wireless resource configuration table is temporally and spatially aligned with the MPPT parameters of the photovoltaic array and the battery SOC to generate a joint control feature matrix. The triggered chaos control mode is judged based on the joint control characteristic matrix to obtain the reference signal receiving power and channel state information.

4. The 5G base station integrated optical-storage power supply control system according to claim 3, characterized in that: The reference signal receiving power is transmitted to the microcontroller unit through the SPI interface, and the maximum power point voltage and current of the photovoltaic module are read to obtain the time series and photovoltaic array, including the following steps: Receive the reference signal received power from the 5G physical layer chip through the SPI interface to obtain a timestamp-aligned reference signal received power sampling sequence; When the reference signal receiving power sampling sequence aligned with the timestamp arrives, the ADC is triggered to synchronously sample and read the maximum power point voltage and current of the photovoltaic module; The maximum power point voltage and current of the photovoltaic module are used to construct the IV characteristic curve of the photovoltaic array, and the time series and photovoltaic array are obtained.

5. The 5G base station integrated optical-storage power supply control system according to claim 4, characterized in that: Using a sliding time window, the time series and photovoltaic array are spatially reconstructed, including the following steps: Divide data subsets based on time series and photovoltaic arrays; The voltage-current series in each data subset window is embedded with time delay, and the embedding dimension and time delay are obtained using Takens' theorem, and the phase space of the time series and photovoltaic array is reconstructed.

6. The 5G base station integrated optical-storage power supply control system according to claim 5, characterized in that: The Wolf algorithm is used to obtain the Lyapunov exponent, which includes the following steps: Based on the phase space of the reconstructed time series and photovoltaic array, the initial reference point is selected and the nearest neighbor point is calculated to obtain the initial trajectory pair set; Track the evolution of time steps for each trajectory pair, record the distance change ratio of adjacent trajectories, and obtain the distance logarithm sequence; A least squares linear fit is performed on the logarithmic series of distances, and the fitting slope is the maximum Lyapunov exponent.

7. The 5G base station integrated optical-storage power supply control system according to claim 6, characterized in that: The microcontroller generates the chaotic state flag and load impedance according to the time series and the photovoltaic array, including the following steps: Based on the Lyapunov exponent value, a Lyapunov exponent value threshold is set; When the Lyapunov exponent value is greater than the Lyapunov exponent value threshold, a chaotic state flag is generated; Calculate the load impedance based on the real-time maximum power point voltage and current of the PV array.

8. The 5G base station integrated optical-storage power supply control system according to claim 7, characterized in that: The chaotic state flag and load impedance are input into the digital controller of the LLC resonant converter, and the resonant cavity parameters are dynamically matched according to the load impedance to make the converter work in the ZVS state, trigger the pre-discharge mode, and obtain the PWM control signal and battery current instruction, including the following steps: Receive the chaos state flag and load impedance sent by the microcontroller through the CAN bus, and output them as a control instruction parameter set after parsing; According to the amplitude and phase angle of the load impedance, the resonant inductor and capacitor parameters of the LLC resonant converter are dynamically adjusted to obtain the resonant frequency; Monitor the MOSFET's Vds voltage zero crossing point in real time at the resonant frequency to obtain the ZVS lock signal and switching frequency. The switching frequency is increased to the reference frequency and a chaotic modulation wave is injected to generate a PWM signal with chaotic characteristics; When a sudden load change is detected and the ZVS lock is valid, a PWM control signal and a battery current command are obtained.

9. The 5G base station integrated optical-storage power supply control system according to claim 8, characterized in that: The PWM control signal and battery current command are sent to the execution unit via the CAN bus to obtain the operation data, including the following steps: The PWM control signal and battery current command generated by chaotic modulation are encapsulated into data frames according to the CAN protocol format to obtain a standardized control command packet; Sending standardized control instruction packets to the execution unit via the CAN bus to obtain a transmission status confirmation signal; The execution unit receives and parses the standardized control instruction package, drives the power MOSFET to act according to the PWM signal, adjusts the battery discharge to the specified value, and obtains the operating parameters.

10. The 5G base station integrated optical-storage power supply control system according to claim 9, characterized in that: Upload the operating data to the cloud through the 5GSA module, and generate the control strategy update package through the LSTM-PSO hybrid optimization model, including the following steps: The operating parameters are encrypted and packaged through the 5G independent networking module and uploaded to the cloud server via the UDP protocol to obtain a successful transmission confirmation signal; Based on the transmission success confirmation signal, a multi-dimensional feature matrix is ​​generated; The multidimensional feature matrix is ​​input into the LSTM-PSO hybrid model to generate a control strategy update package.

Images