5G base station micro-photovoltaic multi-source data dynamic charging method and system

Abstract

The invention discloses a 5G base station micro-photovoltaic multi-source data dynamic charging method and system, and belongs to the technical field of 5G base station charging. Multi-source data (including a base station operation state, weather forecast and a regional power consumption task) are integrated, an LSTM and XGBoost fused fault prediction model, time convolutional network power consumption demand prediction and a multi-target optimization model (optimizing power grid power purchase cost, energy storage loss and photovoltaic utilization rate) are adopted, a charging strategy is dynamically generated in combination with a cloud, and the power consumption of the power grid is optimized. And the strategy is adjusted in real time based on the fault prediction result. According to the method, the optimized strategy is executed by remotely controlling the micro-photovoltaic equipment, global energy scheduling driven by multi-dimensional data is realized, the photovoltaic utilization rate is remarkably improved, meanwhile, through a fault feedforward mechanism and dynamic weight adjustment, the operation cost is reduced on the basis of guaranteeing the power supply stability, and the power supply efficiency is improved. The problems of data island, response lag, fault vulnerability and the like in a traditional scheme are solved.

Description

Technical Field

[0001] This invention relates to the field of 5G base station charging technology, and more specifically to a method and system for dynamic charging of 5G base station micro-photovoltaic multi-source data. Background Technology

[0002] Currently, the large-scale deployment of 5G communication base stations has led to a sharp increase in energy consumption. Traditional power supply mainly relies on grid access and diesel generator backup power. However, grid power supply is geographically limited and has carbon emission issues, while diesel generators have high operating costs, are noisy, and do not conform to the trend of green energy development. To reduce base station operating costs and improve energy sustainability, some base stations are trying to introduce photovoltaic power generation systems.

[0003] However, existing photovoltaic charging solutions have some shortcomings in practical applications. First, they fail to effectively integrate heterogeneous data from multiple sources, such as base station operating status, weather forecasts, and regional electricity consumption tasks, resulting in a lack of global perspective in charging strategy formulation. Second, they do not fully consider the impact of equipment failure prediction on energy dispatch, easily leading to power redundancy or insufficiency under abnormal operating conditions of base station equipment. Third, they lack a cloud-based collaborative remote control mechanism, making it difficult to respond promptly to emergencies such as sudden weather changes and fluctuations in electricity load, resulting in low photovoltaic utilization rates.

[0004] Therefore, how to provide a dynamic charging method and system for multi-source data of micro-photovoltaic 5G base stations to improve charging efficiency and system stability is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a method and system for dynamic charging of 5G base station micro-photovoltaics based on multi-source data. The method dynamically adjusts the charging strategy of 5G base station micro-photovoltaics in real time based on multi-source data, optimizes power supply by combining regional power consumption tasks, and adjusts the charging strategy in advance based on base station fault prediction results to improve energy utilization efficiency and ensure system stability.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] On one hand, this invention discloses a method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations, comprising:

[0008] Collect multi-source data and transmit it to the cloud. The multi-source data includes base station data and meteorological data.

[0009] Based on the power consumption task input area in the cloud and combined with the multi-source data, a 5G base station charging strategy is obtained, including:

[0010] Analyze the base station data to obtain fault prediction results;

[0011] Based on the regional power consumption task, the base station power demand is initially determined, and a preliminary charging strategy is obtained using the base station power demand and the multi-source data.

[0012] The initial charging strategy is adjusted based on the fault prediction results to obtain the 5G base station charging strategy;

[0013] The cloud-based remote control unit on the base station side executes the 5G base station charging strategy.

[0014] Preferably, the fault prediction result is obtained by analyzing the base station-side data, including:

[0015] A fault prediction model is trained using historically collected base station data. The fault prediction model is constructed by fusing LSTM neural network and XGBoost algorithm.

[0016] The real-time collected base station data is input into the trained fault prediction model, and the fault type and corresponding fault probability of the base station equipment are output as the fault prediction result.

[0017] Preferably, based on the regional power consumption task, the base station power demand is initially determined, and a preliminary charging strategy is obtained using the base station power demand and the multi-source data, including:

[0018] If there is a pre-input regional electricity consumption task, a dynamic electricity demand curve is generated based on the electricity consumption time, peak electricity load, and geographical coverage of the regional electricity consumption task.

[0019] If there is no regional power consumption task, the baseline of power demand is predicted by using a temporal convolutional network based on historical base station data.

[0020] The dynamic electricity demand curve or electricity demand baseline is combined with the multi-source data for multi-objective optimization to generate a preliminary charging strategy.

[0021] Preferably, a multi-objective function is constructed for multi-objective optimization, and the multi-objective function is as follows:

[0022] min F=ω1·C grid +ω2·L batt +ω3·E pv_util ;

[0023] In the formula, C grid The objective is to minimize the cost of electricity purchased from the power grid; L batt The goal is to minimize losses in the energy storage system; E pv_util The objective is to maximize the utilization rate of photovoltaic energy; ω1, ω2, and ω3 are the weights of grid electricity purchase cost, energy storage system loss, and photovoltaic energy utilization rate, respectively.

[0024] Preferably, the goal C is to minimize the power grid purchase cost. grid The calculation formula is as follows:

[0025]

[0026] In the formula, T represents the total number of scheduling periods, and P represents the total number of scheduling periods. grid ρ(t) represents the power purchased from the grid during time period t; ρ(t) represents the grid electricity price during time period t.

[0027] Preferably, the energy storage system minimizes losses to target L. batt The calculation formula is as follows:

[0028]

[0029] In the formula, ΔSOC(t) is the change in the state of charge of the energy storage from time t to time t+1; This is an indicator function used to trigger SOC out-of-bounds penalties; SOC(t) is the state of charge of the energy storage battery at time t; SOC nom δ represents the midpoint of the rated operating range of the energy storage battery; δ is the allowable threshold for SOC deviation from the rated range; λ and μ are the SOC fluctuation and over-limit penalty coefficients, respectively.

[0030] Preferably, the photovoltaic energy utilization maximization target E pv_util The calculation formula is as follows:

[0031]

[0032] In the formula, P pv (t) represents the predicted power generation of the photovoltaic system during time period t, obtained through SVM model combined with meteorological data; P pv_dump (t) represents the amount of photovoltaic power curtailed during time period t.

[0033] Preferably, the initial charging strategy is adjusted based on the fault prediction result to obtain the 5G base station charging strategy, including:

[0034] Establish a mapping rule base between fault prediction results and adjustment strategies. The mapping rule base includes the power adjustment range, power supply mode switching logic and equipment self-test trigger conditions corresponding to the fault prediction results.

[0035] Based on the fault prediction results, a corresponding adjustment strategy is matched to revise the initial charging strategy, thereby generating a 5G base station charging strategy that includes a fault response mechanism.

[0036] On the other hand, the present invention also proposes a 5G base station micro-photovoltaic multi-source data dynamic charging system for implementing the above-mentioned 5G base station micro-photovoltaic charging method, including a sensor data acquisition terminal, a cloud platform, and a communication module; and the sensor data acquisition terminal interacts with the communication module.

[0037] The sensor data acquisition terminal is deployed on the 5G base station and its micro photovoltaic equipment to collect multi-source data, including base station side data and meteorological data.

[0038] The cloud platform is used to input regional power consumption tasks and, combined with the multi-source data, to obtain 5G base station charging strategies, including:

[0039] The fault analysis module is used to analyze the base station-side data to obtain fault prediction results;

[0040] The preliminary strategy acquisition module determines the base station power demand based on the regional power consumption task, and acquires a preliminary charging strategy using the base station power demand and the multi-source data.

[0041] The charging strategy output module adjusts the preliminary charging strategy based on the fault prediction result to obtain the 5G base station charging strategy.

[0042] The cloud platform sends control commands through the communication module to remotely control the micro-photovoltaic equipment on the base station side to execute the 5G base station charging strategy.

[0043] As can be seen from the above technical solution, compared with the prior art, this invention discloses a dynamic charging method and system for 5G base station micro-photovoltaics using multi-source data. It collects base station-side data and meteorological data and transmits them to the cloud. A fault prediction model constructed using a fusion of LSTM neural network and XGBoost algorithm analyzes the base station-side data to obtain fault prediction results. Dynamic electricity demand curves or electricity demand baselines are generated for areas with and without electricity consumption tasks. A multi-objective function, including grid power purchase cost, energy storage system loss, and photovoltaic energy utilization rate, is constructed and optimized using multi-source data. The initial charging strategy is then corrected based on a mapping rule library between the fault prediction results and adjustment strategies. Finally, the micro-photovoltaic equipment is remotely controlled by the cloud to execute the charging strategy. This invention achieves dynamic optimization of charging strategies driven by multi-source data, effectively improving energy utilization efficiency, reducing grid power purchase cost, enhancing system stability and fault response capabilities, and solving problems such as lack of globality in strategy formulation, insufficient fault response, and low photovoltaic utilization rate in existing solutions. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0045] Figure 1 A flowchart of the method provided by the present invention;

[0046] Figure 2 The system architecture diagram provided for this invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] On the one hand, such as Figure 1 As shown in the figure, this invention discloses a method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations, comprising:

[0049] S1. Collect multi-source data and transmit it to the cloud. The multi-source data includes base station data and meteorological data.

[0050] Real-time acquisition and preprocessing of multi-source data are achieved by deploying high-precision sensors and edge computing devices. For base station-side data, a current sensor with an accuracy of ±0.1% is used to collect CPU and RF module power consumption (once per minute), and the battery management system (BMS) is used to obtain battery SOC, SOH, and temperature (once per minute). The base station's built-in diagnostic module captures equipment fault codes and other status information in real time (triggered acquisition). Meteorological data is comprehensively acquired through a photovoltaic-specific irradiance sensor (response time <1s, once per 10 seconds), a six-element weather station (once per minute), and the meteorological bureau's API (once per hour). The edge computing device performs 3σ rule outlier filtering and linear normalization on the raw data, and caches 72 hours of data in the event of a network outage. After recovery, the data is retransmitted through a TLS 1.3 encrypted channel to ensure data integrity and security.

[0051] S2. Input the regional power consumption task in the cloud, combine it with multi-source data, and obtain the 5G base station charging strategy, including:

[0052] S21. Analyze the base station data to obtain fault prediction results, including:

[0053] A fault prediction model is trained using historically collected base station data. The fault prediction model is constructed by fusing LSTM neural network and XGBoost algorithm.

[0054] The training set was selected from 12 months of historical base station data, and the features included time-series features and structured features. The time-series features included base station load rate (moving average over the past 24 hours), battery temperature (hourly average), and number of abnormal fan speeds (daily cumulative). The structured features included fault code frequency, equipment runtime, and ambient humidity.

[0055] The fault prediction model uses two stacked LSTM layers with 128 neurons each to extract long-term dependency features from time-series data. The XGBoost layer processes structured features with a learning rate of 0.1, a tree depth of 6, and a subsampling rate of 0.8. The fusion layer integrates the outputs of LSTM and XGBoost through stacking, and outputs a three-class classification result (normal / warning / fault) through a Softmax layer.

[0056] The Adam optimizer was used during the specific training process, with cross-entropy as the loss function, a validation set ratio of 20%, 500 iterations, and a convergence threshold of 0.001.

[0057] The real-time collected base station data is input into the trained fault prediction model, and the fault type and corresponding fault probability of the base station equipment are output as the fault prediction result.

[0058] Fault types include battery overheating, fan failure, inverter malfunction, etc. A level one warning is triggered when the probability is greater than 0.8, and a level two warning is triggered when the probability is between 0.7 and 0.8.

[0059] S22. Based on the regional power consumption task, the power demand of the base station is initially determined, and a preliminary charging strategy is obtained using the power demand of the base station and multi-source data, including:

[0060] S221. If there is a pre-input regional electricity consumption task, then a dynamic electricity demand curve is generated based on the electricity consumption time, peak electricity load, and geographical coverage of the regional electricity consumption task.

[0061] Task parameters that receive Excel templates or API input in task scenarios include:

[0062] Basic parameters: task name, time interval, list of target base station latitude and longitude;

[0063] Load parameters: additional power consumption requirements, power supply reliability requirements;

[0064] Priority: Level 1 task (emergency support), Level 2 task (large-scale events), Level 3 task (routine scheduling).

[0065] Based on the load increase ratio during the task period, the dynamic power demand curve P is obtained by adding the increment to the historical baseline power consumption. task (t) is as follows:

[0066] P task (t)=P hist (t)×(1+q);

[0067] In the formula, P task (t) represents the dynamic power demand during the t-th scheduling cycle within the task period. This parameter is a real-time adjusted power value generated after combining the regional power demand, reflecting the actual power consumption of the base station after considering additional load demand, and it changes dynamically with time t.

[0068] P hist (t) represents the historical baseline power consumption when there is no regional power consumption task. The average power consumption during normal periods, obtained by analyzing the historical operation data of the base station and removing outliers, serves as the benchmark value for dynamic adjustment, reflecting the stable power consumption requirements of the base station under normal operating conditions.

[0069] q represents the additional load rate, which typically ranges from 0.2 to 0.5, or 20% to 50%. This parameter is explicitly input by the regional power consumption task and is used to quantify the proportion of additional power consumption brought about by scenarios such as large-scale events and emergency support.

[0070] S222. If there is no regional power consumption task, then predict the baseline of power demand based on historical base station data through a temporal convolutional network.

[0071] In this embodiment, the Temporal Convolutional Network (TCN) structure includes 3 causal convolutional layers, with a receptive field of 48 hours. It takes load data from the past 7 days as input and outputs hourly baseline power consumption for the next 24 hours, with a prediction error rate of ≤10%.

[0072] S223. Perform multi-objective optimization on the dynamic electricity demand curve or electricity demand baseline with multi-source data to generate a preliminary charging strategy.

[0073] A multi-objective function is constructed for multi-objective optimization, as follows:

[0074] min F=ω1·C grid +ω2·L batt +ω3·E pv_util ;

[0075] In the formula, C grid The objective is to minimize the cost of electricity purchased from the power grid; L batt The goal is to minimize losses in the energy storage system; E pv_utilThe objective is to maximize the utilization rate of photovoltaic energy; ω1, ω2, and ω3 are the weights of grid electricity purchase cost, energy storage system loss, and photovoltaic energy utilization rate, respectively.

[0076] Objective C: Minimize grid power purchase costs grid The calculation formula is as follows:

[0077]

[0078] In the formula, T represents the total number of scheduling periods, and P represents the total number of scheduling periods. grid ρ(t) represents the power purchased from the grid during time period t; ρ(t) represents the grid electricity price during time period t.

[0079] The goal of minimizing energy storage system losses is L. batt The calculation formula is as follows:

[0080]

[0081] In the formula, ΔSOC(t) is the change in the state of charge of the energy storage from time t to time t+1; This is an indicator function used to trigger SOC out-of-bounds penalties; SOC(t) is the state of charge of the energy storage battery at time t; SOC nom δ represents the midpoint of the rated operating range of the energy storage battery; δ is the allowable threshold for SOC deviation from the rated range; λ and μ are the SOC fluctuation and over-limit penalty coefficients, respectively.

[0082] Target E for maximizing photovoltaic energy utilization pv_util The calculation formula is as follows:

[0083]

[0084] In the formula, P pv (t) represents the predicted power generation of the photovoltaic system during time period t, obtained through SVM model combined with meteorological data; P pv_dump (t) represents the amount of photovoltaic power curtailed during time period t.

[0085] The Particle Swarm Optimization (PSO) algorithm is used in conjunction with constraints to solve the multi-objective function, dynamically adjust the inertia weight, and output the initial charging strategy.

[0086] Constraints include power balance, energy storage capacity, and equipment limitations, specifically:

[0087] Power balance: P pv (t)+P grid (t)+P batt_dis (t)=P load (t)+P ch (t);

[0088] Energy storage capacity: 0.2≤SOC(t)≤0.8 (normal scenario), with a minimum SOC increase to 0.9 during Level 1 tasks;

[0089] Device limitations: Charging power P ch ≤5kW, discharge power P batt_dis ≤10kW.

[0090] The preliminary charging strategy obtained from the solution includes:

[0091] Photovoltaic inverter power regulation command (0-100% rated power);

[0092] Battery charging and discharging plan (e.g., charging to 80% SOC during off-peak hours and discharging to 30% SOC during peak hours);

[0093] Power supply mode switching logic (such as photovoltaic priority mode, mains power supplement mode).

[0094] S23. Adjust the initial charging strategy based on the fault prediction results to obtain the 5G base station charging strategy, including:

[0095] S231. Establish a mapping rule base between fault prediction results and adjustment strategies. The mapping rule base includes the power adjustment range, power supply mode switching logic and equipment self-test trigger conditions corresponding to the fault prediction results.

[0096] For fault prediction of excessively high battery temperature (e.g., temperature > 60℃), when the prediction probability is ≥ 0.8, a first-level adjustment strategy is triggered: immediately suspend photovoltaic charging, start the battery air cooling system, and force the power supply mode to switch to grid priority mode to ensure that the base station load is directly powered by the grid; simultaneously adjust the energy storage system constraints, temporarily setting the lower limit of battery state of charge (SOC) to 0.5 to avoid charging in high-temperature environments from aggravating equipment damage. This strategy must be executed within 10 seconds after the fault prediction result is detected.

[0097] When the predicted probability of fan failure warning is in the range of 0.7-0.8, a secondary adjustment strategy is triggered: first, the current charging power is reduced by 30% to reduce the heat dissipation burden on the fan; at the same time, the real-time monitoring mechanism of the backup fan is activated, collecting fan speed data every 10 minutes. If abnormal speed is detected twice in a row, the backup fan will be automatically switched to operation. During the execution of this strategy, the photovoltaic priority power supply mode is maintained, but the stability of the heat dissipation system is ensured through power regulation and equipment redundancy mechanisms, and the response time is controlled within 30 seconds.

[0098] For fault prediction of photovoltaic inverter malfunctions (probability ≥ 0.8), the adjustment strategy is as follows: immediately shut down the faulty inverter, switch to a hybrid power supply mode of mains power and battery, prioritize battery energy storage power supply (when SOC ≥ 0.3), and supplement the insufficient part by the grid; simultaneously trigger the inverter self-test process, execute equipment reset and parameter verification through remote command, and generate a manual maintenance work order if the self-test fails; this strategy must complete the power supply mode switch within 5 seconds to ensure that the base station power supply interruption time is < 10ms, meeting the high reliability requirements.

[0099] In addition, the rule base also covers other potential fault types, such as power module failure and photovoltaic module short circuit. Each fault corresponds to a specific power adjustment range (e.g., reducing charging power by 20%-50%), power supply priority switching logic (priority adjustment of mains power / photovoltaic / battery), and equipment control parameters (e.g., starting backup power and triggering periodic self-tests). All adjustment strategies are assigned execution priority based on the severity of the fault. Level 1 fault strategies have the highest execution authority, can interrupt the current charging task and take effect immediately, ensuring the stability and fault tolerance of the system under abnormal operating conditions.

[0100] S232. Match the corresponding adjustment strategy according to the fault prediction result, modify the initial charging strategy, and generate a 5G base station charging strategy that includes a fault response mechanism.

[0101] Based on the fault prediction results and matching rule base, adjust the power parameters and power supply mode in the initial strategy. For example, if the battery is predicted to overheat, set the charging power to 0 and increase the proportion of mains power purchase to ensure the stability of the base station load.

[0102] S3. The cloud-based remote control of the micro-photovoltaic equipment on the base station side executes the 5G base station charging strategy.

[0103] Sending 0-100% power regulation signals to the smart inverter via the Modbus protocol executes power regulation commands with a response time of <500ms;

[0104] After receiving instructions from the cloud, the BMS controls the battery charging and discharging current and supports soft start to prevent impact.

[0105] The intelligent distribution box executes load switching commands for switching between mains power, photovoltaic power, and battery power, with a switching time of <200ms and supports seamless switching (interruption time <10ms).

[0106] On the other hand, this invention also proposes a 5G base station micro-photovoltaic multi-source data dynamic charging system to implement the above-mentioned 5G base station micro-photovoltaic charging method, referring to... Figure 2 It includes a sensor data acquisition terminal, a cloud platform, and a communication module; the sensor data acquisition terminal interacts with the communication module.

[0107] The sensor data acquisition terminal is deployed on the 5G base station and its micro photovoltaic equipment to collect multi-source data, including base station side data and meteorological data.

[0108] The cloud is used to input regional power consumption tasks and combine multi-source data to obtain 5G base station charging strategies, including:

[0109] The fault analysis module is used to analyze base station-side data and obtain fault prediction results;

[0110] The preliminary strategy acquisition module determines the base station's power demand based on the regional power consumption task, and acquires a preliminary charging strategy using the base station's power demand and multi-source data.

[0111] The charging strategy output module adjusts the initial charging strategy based on the fault prediction results to obtain the 5G base station charging strategy.

[0112] The cloud sends control commands through the communication module to remotely control the micro-photovoltaic equipment on the base station side to execute the 5G base station charging strategy.

[0113] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0114] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations, characterized in that, include: Collect multi-source data and transmit it to the cloud. The multi-source data includes base station data and meteorological data. Based on the power consumption task input area in the cloud and combined with the multi-source data, a 5G base station charging strategy is obtained, including: Analyze the base station data to obtain fault prediction results; Based on the regional power consumption task, the base station power demand is initially determined, and a preliminary charging strategy is obtained using the base station power demand and the multi-source data. The initial charging strategy is adjusted based on the fault prediction results to obtain the 5G base station charging strategy; The cloud-based remote control unit on the base station side executes the 5G base station charging strategy.

2. The method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations according to claim 1, characterized in that, Analyzing the base station-side data yields fault prediction results, including: A fault prediction model is trained using historically collected base station data. The fault prediction model is constructed by fusing LSTM neural network and XGBoost algorithm. The real-time collected base station data is input into the trained fault prediction model, and the fault type and corresponding fault probability of the base station equipment are output as the fault prediction result.

3. The method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations according to claim 1, characterized in that, Based on the regional power consumption task, the base station power demand is initially determined, and a preliminary charging strategy is obtained using the base station power demand and the multi-source data, including: If there is a pre-input regional electricity consumption task, a dynamic electricity demand curve is generated based on the electricity consumption time, peak electricity load, and geographical coverage of the regional electricity consumption task. If there is no regional power consumption task, the baseline of power demand is predicted by using a temporal convolutional network based on historical base station data. The dynamic electricity demand curve or electricity demand baseline is combined with the multi-source data for multi-objective optimization to generate a preliminary charging strategy.

4. The method for dynamic charging of multi-source data in a 5G base station micro-photovoltaic system according to claim 3, characterized in that, A multi-objective function is constructed for multi-objective optimization, and the multi-objective function is as follows: min F=ω1·C grid +ω2·L batt +ω3·E pv_util ; In the formula, C grid The objective is to minimize the cost of electricity purchased from the power grid; L batt The goal is to minimize losses in the energy storage system; E pv_util The objective is to maximize the utilization rate of photovoltaic energy; ω1, ω2, and ω3 are the weights of grid electricity purchase cost, energy storage system loss, and photovoltaic energy utilization rate, respectively.

5. A method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations according to claim 4, characterized in that, The objective of minimizing the power grid purchase cost is C. grid The calculation formula is as follows: In the formula, T represents the total number of scheduling periods, and P represents the total number of scheduling periods. grid ρ(t) represents the power purchased from the grid during time period t; ρ(t) represents the grid electricity price during time period t.

6. The method for dynamic charging of multi-source data in a 5G base station micro-photovoltaic system according to claim 4, characterized in that, The energy storage system loss minimization objective L batt The calculation formula is as follows: In the formula, ΔSOC(t) is the change in the state of charge of the energy storage from time t to time t+1; This is an indicator function used to trigger SOC out-of-bounds penalties; SOC(t) is the state of charge of the energy storage battery at time t; SOC nom δ represents the midpoint of the rated operating range of the energy storage battery; δ is the allowable threshold for SOC deviation from the rated range; λ and μ are the SOC fluctuation and over-limit penalty coefficients, respectively.

7. A method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations according to claim 4, characterized in that, The photovoltaic energy utilization maximization target E pv_util The calculation formula is as follows: In the formula, P pv (t) represents the predicted power generation of the photovoltaic system during time period t, obtained through SVM model combined with meteorological data; P pv_dump (t) represents the amount of photovoltaic power curtailed during time period t.

8. A method for dynamic charging of multi-source data from micro-photovoltaics in 5G base stations according to claim 1, characterized in that, The initial charging strategy is adjusted based on the fault prediction results to obtain the 5G base station charging strategy, including: Establish a mapping rule base between fault prediction results and adjustment strategies. The mapping rule base includes the power adjustment range, power supply mode switching logic and equipment self-test trigger conditions corresponding to the fault prediction results. Based on the fault prediction results, a corresponding adjustment strategy is matched to revise the initial charging strategy, thereby generating a 5G base station charging strategy that includes a fault response mechanism.

9. A 5G base station micro-photovoltaic multi-source dynamic data charging system, characterized in that, It includes a sensor data acquisition terminal, a cloud platform, and a communication module; and the sensor data acquisition terminal interacts with the communication module. The sensor data acquisition terminal is deployed on the 5G base station and its micro photovoltaic equipment to collect multi-source data, including base station side data and meteorological data. The cloud platform is used to input regional power consumption tasks and, combined with the multi-source data, to obtain 5G base station charging strategies, including: The fault analysis module is used to analyze the base station-side data to obtain fault prediction results; The preliminary strategy acquisition module determines the base station power demand based on the regional power consumption task, and acquires a preliminary charging strategy using the base station power demand and the multi-source data. The charging strategy output module adjusts the preliminary charging strategy based on the fault prediction result to obtain the 5G base station charging strategy. The cloud platform sends control commands through the communication module to remotely control the micro-photovoltaic equipment on the base station side to execute the 5G base station charging strategy.

Images