5G base station power supply state intelligent monitoring system and method

Abstract

The invention discloses a 5G base station power supply state intelligent monitoring system and method, and belongs to the technical field of intelligent monitoring. According to the system, operation parameters of a 5G base station power supply are acquired in real time through the data acquisition module, and after the operation parameters are processed by the signal conditioning and conversion module, the master control processing unit analyzes digital signals in real time, detects abnormal states and generates alarm information. The communication module is responsible for data transmission and instruction receiving, and the power management and safety protection module provides overvoltage, undervoltage, overcurrent and short circuit protection and triggers emergency measures. The artificial intelligence analysis module performs fault prediction and performance optimization based on historical data and real-time data, and the remote monitoring terminal realizes data visual monitoring and management. According to the invention, the real-time performance and accuracy of monitoring are improved, unified management of multiple base stations is realized, the operation and maintenance cost is reduced, and stable operation of a 5G communication network is guaranteed.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent monitoring, and specifically relates to an intelligent monitoring system and method for the power supply status of 5G base stations. Background Art

[0002] The rapid development and wide application of 5G communication technology have promoted the rapid growth of the number of mobile communication base stations. To meet the increasing communication needs and coverage requirements, the number of 5G base stations far exceeds that of traditional 2G, 3G, and 4G base stations. However, with the increase in the number of 5G base stations, higher requirements are put forward for the operation and maintenance management of the base stations, especially the monitoring of the power supply status of the base stations is particularly important.

[0003] Traditional base station power supply monitoring mainly relies on manual inspection and regular maintenance. This method is inefficient, error-prone, and unable to grasp the operation status of the power supply in real time. Once a power supply failure occurs, it often takes a long time to discover and handle, resulting in the base station being unable to work properly for a long time, seriously affecting the communication quality and user experience.

[0004] To solve the above problems, the industry has proposed various base station power supply monitoring solutions. A common solution is to install various sensors in the base station, such as voltage, current, temperature and other sensors, and monitor the power supply status by collecting various parameters of the power supply. However, this solution has the following deficiencies: First, a large number of sensors need to be installed in each base station, resulting in high costs; Second, the sensors are easily affected by the environment, such as temperature, humidity and other factors, resulting in inaccurate measurement data; Third, the data collected by the sensors is large, and both transmission and storage face certain challenges.

[0005] Another solution is to use Internet of Things technology to connect the base station power supply to the Internet of Things platform and realize remote monitoring of the power supply through the Internet of Things platform. Although this solution can realize remote monitoring and reduce the workload of manual inspection, there are still some limitations. For example, the construction and maintenance costs of the Internet of Things platform are high, and there may be problems such as data transmission delay and packet loss, affecting the real-time performance and accuracy of monitoring.

[0006] With the development of artificial intelligence technology, applying artificial intelligence to base station power supply monitoring has become a new trend. By collecting historical data of the base station power supply and using machine learning algorithms to establish a prediction model of the power supply status, intelligent monitoring and fault warning of the power supply can be realized. This solution can not only improve the efficiency and accuracy of monitoring, but also through early warning, timely handle potential power supply failures and reduce the fault downtime of the base station.

[0007] However, existing power monitoring solutions based on artificial intelligence still have some deficiencies. First, most solutions can only achieve power monitoring for a single base station, lacking unified monitoring and management of the power supplies of multiple base stations within a region. Second, most existing solutions are based on specific communication protocols and interfaces, lacking generality and scalability, and are difficult to apply to base station equipment of different manufacturers and models. Third, existing solutions mainly focus on fault detection and early warning, paying insufficient attention to aspects such as power performance analysis and energy consumption optimization, and are unable to comprehensively evaluate the operating status of the power supply.

[0008] Therefore, there is an urgent need for a general and efficient intelligent 5G base station power status monitoring system and method to achieve unified monitoring and management of the power supplies of multiple base stations within a region, improve the intelligence level of power monitoring, reduce operation and maintenance costs, and ensure the stable operation of the 5G communication network. This not only has important significance for the development of the 5G communication network but also will provide useful reference and inspiration for equipment monitoring and fault diagnosis in other industries.

[0009] Technical Solution

[0010] To solve the above problems, the present invention adopts the following technical solutions.

[0011] An intelligent monitoring system for the power status of a 5G base station includes the following:

[0012] A data acquisition module for collecting the voltage, current, temperature, remaining battery power, and battery health of the 5G base station power supply;

[0013] A signal conditioning and conversion module electrically connected to the data acquisition module for filtering, amplifying, and converting the analog signals in the operating parameters into digital signals;

[0014] A main control processing unit electrically connected to the signal conditioning and conversion module for receiving the digital signals and performing real-time analysis on them, and generating an alarm message if an abnormal state is detected;

[0015] A communication module connected to the main control processing unit for sending the digital signals and alarm messages to a remote monitoring end and receiving control instructions issued by the remote monitoring end;

[0016] A power management and safety protection module electrically connected to the main control processing unit for protecting the power supply against overvoltage, undervoltage, overcurrent, and short circuit, and triggering emergency protection measures in case of an abnormal state;

[0017] An artificial intelligence analysis module deployed on the main control processing unit or the remote monitoring end for predicting power supply faults based on historical data and real-time monitoring data and outputting fault prediction results;

[0018] The remote monitoring terminal, which is communicatively connected to the communication module, is used to receive digital signals, alarm information, and fault prediction results, and perform visual monitoring and management of the power supply status based on the digital signals and fault prediction results.

[0019] Preferably, the signal conditioning and conversion module includes:

[0020] A filter circuit for filtering out noise from the analog signals output by the data acquisition module;

[0021] An amplifier circuit for amplifying the analog signals processed by the filter circuit;

[0022] An analog-to-digital conversion circuit for converting the analog signals amplified by the amplifier circuit into digital signals and transmitting the digital signals to the main control processing unit for analysis and processing.

[0023] Preferably, the main control processing unit includes:

[0024] A data reception and processing unit for receiving the digital signals transmitted by the signal conditioning and conversion module and performing preliminary processing and storage on them;

[0025] A real-time analysis unit for performing real-time analysis on the received digital signals using a preset algorithm and generating alarm information if an abnormal state is detected;

[0026] A control and regulation unit for remotely controlling or regulating the base station power supply equipment according to the analysis results and the instructions from the remote monitoring terminal;

[0027] A communication interface for connecting to the communication module, performing two-way communication with the remote monitoring terminal, receiving control instructions, and sending monitoring data and alarm information.

[0028] Preferably, the communication module includes:

[0029] A wireless communication unit for sending digital signals and alarm information to the remote monitoring terminal via a wireless network and receiving control instructions issued by the remote monitoring terminal;

[0030] A wired communication unit for communicating with the remote monitoring terminal via a wired network;

[0031] A data encryption and decryption unit for encrypting the transmitted data;

[0032] A communication protocol conversion unit for converting data with different communication protocols;

[0033] A communication status monitoring unit for real-time monitoring of the status of the communication link and sending an alarm signal in case of communication anomalies.

[0034] Preferably, the power management and security protection module includes:

[0035] An overvoltage protection unit for detecting whether the power supply voltage exceeds a preset safety threshold, and triggering protection measures if it exceeds to prevent damage to the power supply;

[0036] An undervoltage protection unit for detecting whether the power supply voltage is lower than a preset safety threshold, and triggering protection measures if it is lower to prevent power overload or battery over-discharge;

[0037] An overcurrent protection unit for detecting whether the power supply current exceeds a preset safety threshold, and triggering protection measures if it exceeds to prevent power overload or short circuit;

[0038] A short-circuit protection unit for detecting whether the power supply is short-circuited, and immediately cutting off the power supply if it occurs;

[0039] An emergency protection unit for taking corresponding emergency measures according to a preset emergency handling strategy when an abnormal state is detected;

[0040] A power supply status monitoring unit for real-time monitoring of the operating status of the power supply and transmitting the monitoring data to the main control processing unit for analysis;

[0041] A power supply optimization unit for optimizing and adjusting the power supply according to the analysis results and instructions from the remote monitoring terminal.

[0042] Preferably, the artificial intelligence analysis module includes:

[0043] A data preprocessing unit for preprocessing the collected power supply operation parameters;

[0044] A machine learning algorithm unit for using a variety of machine learning algorithms to model the preprocessed data and establish a prediction model of the power supply status; among which the variety of machine learning algorithms include the random forest supervised learning algorithm and the isolation forest unsupervised learning algorithm; the model of the random forest supervised learning algorithm is as follows: in the input layer, the feature vector: contains 6 types of core parameters X = [V in , V out , I load , T amb , SOC, SOH] T Where: V in : input voltage, range 200 - 240V, V out : output voltage, volatility ≤ 5%, I load : load current, sampling rate 1kHz, I amb: Ambient temperature, ranging from -40°C to +85°C, SOC: State of Charge of the battery, with an accuracy of ±1%, SOH: State of Health of the battery. The decision tree is constructed as follows. Each decision tree is generated in the following way: Feature subset selection: Randomly select 3 features for each split. The node splitting criterion adopts the principle of minimizing the Gini index. where p k is the sample proportion of fault type k. The ensemble strategy is as follows. For classification tasks, the majority voting mechanism is adopted. For regression tasks, the average value is taken. where the model of the Isolation Forest unsupervised learning algorithm is as follows: Input features, multi-dimensional time series parameters: Time series containing 6 core parameters. where: ΔV out : Output voltage volatility, standard deviation of a 1-minute window, Slope of the load current change, linear regression coefficient of a 5-second window, σ(T): Standard deviation of temperature fluctuation, 15-minute window, Gradient of the power change, power decay rate per unit time, ΔR int : Rate of change of battery internal resistance, ρ(V, I): Voltage-current correlation coefficient. The isolation tree is constructed as follows. Each isolation tree is generated in the following way: Random feature selection: Randomly select 2 features for each split. Split value generation: Randomly select a split threshold within the feature value range. Recursive splitting: Until the following conditions are met: Data points are completely isolated and the maximum tree depth h max = 15;

[0045] The fault prediction unit predicts power supply faults using the prediction model based on historical data and real-time monitoring data, and outputs the fault prediction results;

[0046] The performance optimization unit optimizes the power supply performance according to the analysis results and the instructions from the remote monitoring terminal;

[0047] The anomaly detection unit is used to monitor the power supply operation parameters in real time. If an abnormal state is detected, it generates an alarm message and provides an analysis of the cause of the anomaly;

[0048] The self-learning unit continuously learns and updates the model parameters;

[0049] The visualization analysis unit presents the analysis results in a visual way.

[0050] Preferably, the remote monitoring terminal includes:

[0051] The data reception and storage unit is used to receive the digital signals, alarm messages, and fault prediction results transmitted by the communication module, and store and manage them;

[0052] The visualization monitoring unit performs real-time visualization monitoring of the power supply status based on the received digital signals and fault prediction results;

[0053] The alarm and notification unit sends alarm messages to the operation and maintenance personnel via text messages, emails, and push notifications when an abnormal state is detected;

[0054] The control and regulation unit remotely controls or regulates the base station power supply equipment according to the analysis results and the instructions of the operation and maintenance personnel;

[0055] The historical data analysis unit uses historical data for trend analysis and performance evaluation, and provides an evaluation report on the long-term operation status;

[0056] The user interface unit provides a friendly user interface;

[0057] The security management unit is used to manage user permissions, data encryption, and transmission security, ensuring the security of the system and the confidentiality of data.

[0058] A method for intelligent monitoring of the power supply status of a 5G base station includes the following steps:

[0059] S1. Collect the operating parameters of the voltage, current, temperature, remaining battery power, and battery health of the 5G base station power supply through the data acquisition module;

[0060] S2. The signal conditioning and conversion module filters, amplifies, and converts the analog signals in the operating parameters into digital signals;

[0061] S3. The main control processing unit receives the digital signals and performs real-time analysis on them. If an abnormal state is detected, alarm messages are generated;

[0062] S4. The communication module sends the digital signals and alarm messages to the remote monitoring end and receives the control instructions issued by the remote monitoring end;

[0063] S5. The power management and security protection module protects the power supply against overvoltage, undervoltage, overcurrent, and short circuit, and triggers emergency protection measures in case of an abnormal state;

[0064] S6. The artificial intelligence analysis module predicts power supply faults based on historical data and real-time monitoring data, and outputs fault prediction results;

[0065] S7. The remote monitoring end receives the digital signals, alarm messages, and fault prediction results, and performs visualization monitoring and management of the power supply status based on the digital signals and fault prediction results.

[0066] Beneficial effects

[0067] Compared with the prior art, the beneficial effects of the present invention are:

[0068] Improved Real-time Performance and Accuracy: Through the collaborative work of the data acquisition module and the signal conditioning and conversion module, real-time acquisition and high-precision conversion of the operating parameters of the 5G base station power supply (such as voltage, current, temperature, remaining battery power, and health status) are achieved, ensuring the accuracy of the monitoring data. The main control processing unit combines preset algorithms to perform real-time analysis on digital signals, enabling rapid detection of abnormal states and generation of alarm information, significantly shortening the fault response time.

[0069] Enhanced Intelligence Level: The artificial intelligence analysis module utilizes historical data and real-time monitoring data to achieve fault prediction and performance optimization through machine learning algorithms, providing early warnings for potential problems and avoiding base station downtime caused by power failures. The self-learning function can continuously optimize the algorithm model, improving the system's adaptability and prediction accuracy.

[0070] Unified Management of Multiple Base Stations: The system supports centralized monitoring and management of the power supplies of multiple base stations in a region, overcoming the limitation of single monitoring in traditional solutions and improving the operation and maintenance efficiency.

[0071] Enhanced Security and Reliability: The power management and security protection module provides multiple protection measures such as overvoltage, undervoltage, overcurrent, and short circuit, and triggers an emergency handling mechanism in abnormal states to ensure the security and stability of power operation. The communication module ensures the security and compatibility of data transmission through data encryption and protocol conversion functions, and simultaneously monitors the communication status in real time to avoid monitoring failures caused by communication interruptions.

[0072] Optimization of Cost and Energy Consumption: Through the performance optimization unit to analyze and adjust the power operation status, energy consumption optimization is achieved, reducing the operating cost. The need for manual inspections is reduced, significantly lowering the labor maintenance cost while improving the operation and maintenance efficiency.

[0073] Strong Versatility and Scalability: The system is compatible with 5G base station equipment from different manufacturers and models, with good versatility, and can be widely applied to various 5G base station scenarios. The modular design facilitates the expansion of functions or the upgrade of hardware to meet the needs of future technological development.

[0074] Convenient Visual Management: The remote monitoring terminal provides an intuitive visual interface, displaying the power status, alarm information, and fault prediction results in the form of charts, facilitating operation and maintenance personnel to quickly grasp the operation status of the base station and make decisions.

[0075] In summary, the present invention significantly improves the intelligence, real-time performance, and reliability of the 5G base station power status monitoring, while reducing the operation and maintenance cost, providing a strong guarantee for the stable operation of the 5G communication network, and also providing important reference value for equipment monitoring in other fields. Brief Description of the Drawings

[0076] Figure 1It is the operation flow chart of the intelligent monitoring system for the power supply status of 5G base stations;

[0077] Figure 2 It is the schematic diagram of the Isolation Forest algorithm in the intelligent monitoring method for the power supply status of 5G base stations;

[0078] Figure 3 It is the prediction result graph of the Isolation Forest algorithm in the intelligent monitoring method for the power supply status of 5G base stations.

[0079] Figure 4 It is the operation method diagram of the intelligent monitoring system for the power supply status of 5G base stations. Specific implementation mode

[0080] The present invention will be further described below in conjunction with specific embodiments.

[0081] Unless otherwise defined, the technical and scientific terms used in the following embodiments have the same meanings as commonly understood by those skilled in the art to which the present invention belongs.

[0082] Embodiment

[0083] As Figure 1 shown, the intelligent monitoring system for the power supply status of 5G base stations includes the following:

[0084] A data acquisition module, which is used to collect the voltage, current, temperature, remaining battery power, and battery health of the 5G base station power supply;

[0085] The data acquisition module is a key component of the intelligent monitoring system for the power supply status of 5G base stations, responsible for real-time acquisition of the core operation parameters of the base station power supply, and providing basic data for subsequent signal processing and analysis. The functions and technical details of this module include the following:

[0086] Voltage: Real-time measurement of the input and output voltages of the base station power supply, used to determine whether the power supply is within the normal working range.

[0087] Current: Monitoring the load current of the base station power supply to evaluate the power consumption and load status of the equipment.

[0088] Temperature: Obtaining the temperature inside or around the power supply equipment to prevent equipment damage caused by overheating.

[0089] Remaining battery power: Evaluating the available time of the backup battery by measuring the discharge capacity of the battery.

[0090] Battery health: Judging the aging degree and health status of the battery based on parameters such as the number of charge and discharge cycles and the internal resistance of the battery.

[0091] Hardware composition:

[0092] Sensor Unit: It includes high-precision voltage sensors, current sensors, temperature sensors, etc., to ensure the accuracy and reliability of the collected data.

[0093] Interface Module: It supports multiple signal input interfaces (such as analog signals, digital signals) to adapt to different types of sensors.

[0094] Anti-interference Design: It adopts shielding measures and anti-interference filtering technologies to reduce the influence of the external environment on the collected signals.

[0095] Collection Method:

[0096] It supports continuous collection mode for real-time monitoring of dynamic change parameters.

[0097] It provides a timing collection function to obtain key parameters according to a preset interval, reducing the system power consumption.

[0098] It has an event-triggered collection function, automatically starting high-frequency collection when detecting abnormalities or specific conditions.

[0099] Data Processing and Output:

[0100] It performs preliminary processing on the original data (such as denoising, smoothing) to improve the data quality.

[0101] It outputs a standardized signal to the signal conditioning and conversion module for subsequent processing.

[0102] Reliability and Adaptability:

[0103] The module design adapts to harsh environments (such as high temperature, high humidity or strong interference scenarios) to ensure stable operation under various conditions.

[0104] It supports remote configuration and upgrade, facilitating maintenance personnel to adjust the collection strategy or update functions.

[0105] Through the above design, the data collection module can comprehensively and accurately obtain the operating status of the 5G base station power supply, providing reliable data support for intelligent analysis and fault prediction, while enhancing the overall monitoring ability and stability of the system.

[0106] A signal conditioning and conversion module, electrically connected to the data collection module, is used to filter, amplify and convert the analog signal in the operating parameters into a digital signal;

[0107] The signal conditioning and conversion module includes:

[0108] A filtering circuit for filtering the noise of the analog signal output by the data collection module;

[0109] An amplification circuit for amplifying the analog signal processed by the filtering circuit;

[0110] An analog-to-digital conversion circuit is used to convert the amplified analog signal from the amplifier circuit into a digital signal and transmit the digital signal to the main control processing unit for analysis and processing.

[0111] The signal conditioning and conversion module is one of the core components of the 5G base station power status intelligent monitoring system. It is responsible for processing and converting the analog signal output by the data acquisition module and providing a high-quality digital signal input to the main control processing unit. The detailed functions and design of this module are as follows:

[0112] Filter circuit:

[0113] Function: Filter the noise of the analog signal output by the data acquisition module, remove high-frequency interference and environmental noise, and ensure the purity of the signal.

[0114] Type:

[0115] Low-pass filter: Remove high-frequency noise.

[0116] High-pass filter: Remove low-frequency interference.

[0117] Band-pass filter: Retain signals in a specific frequency band.

[0118] Implementation method: Use passive filters (resistors, capacitors, inductors) or active filters (operational amplifiers).

[0119] Amplifier circuit:

[0120] Function: Amplify the filtered analog signal, increase the signal amplitude, and ensure that the signal can be accurately recognized by the subsequent analog-to-digital conversion circuit.

[0121] Amplification type:

[0122] Differential amplification: Enhance differential signals and suppress common-mode noise.

[0123] Variable gain amplification: Dynamically adjust the gain according to the amplitude of the input signal to avoid overload or distortion.

[0124] Implementation method: Design a high-precision amplifier circuit using operational amplifiers (such as OPA series).

[0125] Analog-to-digital conversion circuit (ADC):

[0126] Function: Convert the amplified analog signal into a digital signal and transmit the digital signal to the main control processing unit for analysis and processing.

[0127] Parameters:

[0128] Resolution (such as 12 bits, 16 bits): Determines the accuracy of the digital signal.

[0129] Sampling rate (such as 1kHz, 10kHz): Determines the time accuracy of signal capture.

[0130] Implementation method: Use high-precision ADC chips (such as ADS1115, ADS1256).

[0131] Generally speaking, multi-channel support: Supports multiple input channels and can simultaneously process analog signals from different sensors such as voltage, current, and temperature. Auto-calibration: Provides auto-zero calibration and gain calibration functions to ensure stability and accuracy during long-term operation. Anti-interference ability: Enhances the anti-electromagnetic interference ability and reduces the influence of external noise through shielding design and grounding optimization.

[0132] The working process is as follows: The analog signal output by the data acquisition module enters the filtering circuit to remove interference components and retain valid information. The filtered signal enters the amplification circuit to adjust the signal amplitude to a suitable range for ADC input as needed. The amplified analog signal enters the analog-to-digital conversion circuit to be converted into a digital signal. The digital signal is transmitted to the main control processing unit through the interface for subsequent analysis and processing.

[0133] The above operations improve the quality of the acquired data and provide a reliable basis for subsequent intelligent analysis. Support high-precision and high-speed sampling to meet the data requirements in the complex operating environment of 5G base stations. The modular design is convenient for maintenance and upgrade and can adapt to different types of sensors and application scenarios. Through the above design, the signal conditioning and conversion module can effectively improve the system performance and provide strong technical support for the power status monitoring of 5G base stations.

[0134] The main control processing unit, electrically connected to the signal conditioning and conversion module, is used to receive the digital signal and perform real-time analysis on it, and generate an alarm message if an abnormal state is detected;

[0135] The main control processing unit includes: a data receiving and processing unit, which is used to receive the digital signals transmitted by the signal conditioning and conversion module, and perform preliminary processing and storage on them; a real-time analysis unit, which uses a preset algorithm to perform real-time analysis on the received digital signals, and generates an alarm message if an abnormal state is detected; a control and regulation unit, which remotely controls or regulates the base station power supply equipment according to the analysis results and the instructions of the remote monitoring terminal; a communication interface, which is used to connect with the communication module, perform two-way communication with the remote monitoring terminal, receive control instructions, and send monitoring data and alarm messages. The preset algorithms in the real-time analysis unit include: a threshold comparison algorithm, which is used to compare the received digital signals with the preset normal range thresholds to determine whether they exceed the normal range; a trend analysis algorithm, which is used to analyze the change trend of the digital signals and predict the change situation in a future period of time; a correlation analysis algorithm, which is used to analyze the correlation between different operating parameters and discover the association rules between the parameters; an anomaly detection algorithm, which is used to establish a normal model of the power supply state through machine learning methods and real-time judge whether the power supply operation state is abnormal; a root cause analysis algorithm, which is used to analyze the cause of the anomaly and locate the fault point when an abnormal state is detected; a prediction algorithm, which is used to establish a prediction model based on historical data and real-time data and predict the future power supply state.

[0136] Among them, the threshold comparison algorithm: Function: Perform real-time monitoring on the received digital signals, compare each operating parameter (such as voltage, current, temperature, etc.) with the preset normal range thresholds, and determine whether it exceeds the normal range. Implementation details: Set the upper and lower limit thresholds for each parameter (such as the voltage range is 220V ± 10V). If a parameter value exceeds the set range, immediately trigger an alarm and record the abnormal time and parameter value. Application scenario: Quickly detect obvious anomalies, such as overvoltage, undervoltage, or high temperature situations.

[0137] Among them, the trend analysis algorithm: Function: Analyze the change trend of the digital signals, predict the possible change situation of the parameters in a future period of time, and discover potential problems in advance. Implementation details: Use the moving average method or weighted moving average method to smooth the real-time data and extract the change trend. Based on linear regression or time series models (such as ARIMA) to predict the future data change. If the prediction result shows that the parameter is about to exceed the safe range, issue a warning in advance. Application scenario: Monitor the decreasing trend of the remaining battery power or the gradually increasing temperature situation.

[0138] Among them, the correlation analysis algorithm: Function: Analyze the correlation between different operating parameters, discover the association rules between parameters, and assist in judging the cause of anomalies. Implementation details: Use statistical methods (such as Pearson correlation coefficient, Spearman correlation coefficient) to calculate the strength of the correlation between parameters. Establish a multivariate model (such as principal component analysis PCA) to identify key influencing factors. If it is found that the correlation between certain parameters changes abnormally, it is prompted that there may be a systematic failure. Application scenario: Analyze the relationship between voltage, current, and temperature to determine whether there is an overload or poor heat dissipation problem.

[0139] Among them, the anomaly detection algorithm: Function: Establish a normal model of the power supply state through machine learning methods, and judge in real time whether the power supply operation state is abnormal. Implementation details: Use unsupervised learning algorithms (such as Isolation Forest, One-Class SVM) to model historical data and identify the normal operating state. Input real-time data into the model. If it deviates from the normal state, it is determined as abnormal and an alarm is triggered. Support dynamic model update to improve adaptability and accuracy. Application scenario: Detect complex anomaly patterns, such as multiple parameter anomalies occurring simultaneously but not exceeding a single threshold.

[0140] Here, taking the Isolation Forest algorithm as an example, Function: Establish a normal model of the power supply state through machine learning methods, and judge in real time whether the power supply operation state is abnormal.

[0141] Implementation details:

[0142] Data preprocessing:

[0143] Clean and standardize the collected power supply operation parameters to ensure data quality.

[0144] Extract key features (such as voltage volatility, current peak value, etc.) to reduce redundant information and improve analysis efficiency.

[0145] Modeling process:

[0146] Isolation Forest algorithm:

[0147] Random splitting: The Isolation Forest algorithm constructs a series of isolation trees by randomly selecting features and randomly selecting splitting values. The construction process of each tree is as follows:

[0148] Randomly select a feature.

[0149] Randomly select a splitting value within the range of this feature.

[0150] Divide the data points into two parts according to the splitting value, and repeat this process until each data point is isolated or reaches the preset tree depth.

[0151] Path length: The path length of each data point in the tree (i.e., the number of edges from the root node to the leaf node) is recorded. Due to their rarity and distinctiveness, outliers usually require fewer splits to be isolated and thus have shorter path lengths.

[0152] Outlier score: The outlier score of each data point is obtained by calculating the average path length of the data point in all trees. The lower the outlier score, the more likely the data point is an outlier.

[0153] Model training:

[0154] Use historical data to train multiple isolation trees to form an isolation forest.

[0155] Evaluate the model performance through cross-validation to ensure the generalization ability of the model.

[0156] Real-time detection:

[0157] Data input: Input the power operation parameters collected in real time into the trained isolation forest model.

[0158] Outlier judgment: If the path length of the real-time data is significantly shorter than the average path length of the normal data, it is determined as an outlier and an alarm is triggered.

[0159] Dynamic update: To improve the adaptability and accuracy of the model, the system supports updating the model parameters regularly or under specific conditions (such as a significant change in data distribution).

[0160] Application scenarios:

[0161] Complex outlier pattern detection: The isolation forest algorithm is particularly suitable for detecting complex outlier patterns, such as the situation where multiple parameter outliers occur simultaneously but do not exceed a single threshold.

[0162] High-dimensional data processing: Due to its ability to handle high-dimensional data, it is applicable to scenarios involving a large number of parameters in the power status monitoring of 5G base stations.

[0163] Real-time monitoring: In the intelligent monitoring system for the power status of 5G base stations, the isolation forest algorithm can analyze the data stream in real time, quickly identify abnormal states, and reduce the fault downtime.

[0164] Data noise tolerance: The isolation forest algorithm has good tolerance for noise in the data, can ignore noise data to a certain extent, and improve the accuracy of outlier detection.

[0165] The Python implementation code is as follows:

[0166] import numpy as np

[0167] from skleam.ensemble import IsolationForest

[0168] from sklearn.preprocessing import StandardScaler

[0169] # Assume we have the following dataset, containing parameters such as voltage, current, temperature, etc.

[0170] data = np.array(

[0171] [220, 10, 30], # Normal data

[0172] [225, 12, 32], # Normal data

[0173] [215, 8, 28], # Normal data

[0174] [250, 20, 40], # Abnormal data

[0175] [210, 7, 25], # Normal data

[0176] [230, 15, 35], # Normal data

[0177] [240, 18, 38], # Abnormal data )

[0179] # Data preprocessing

[0180] scaler = StandardScaler()

[0181] data_scaled = scaler.fit_transform(data)

[0182] # Train the Isolation Forest model

[0183] iso_forest = IsolationForest(contamination = 0.1, random_state = 42)

[0184] iso_forest.fit(data_scaled)

[0185] # Real-time data input

[0186] real_time_data = np.array([[245, 19, 39]]) # Assume this is the real-time collected data

[0187] real_time_data_scaled = scaler.transform(real_time_data)

[0188] # Anomaly detection

[0189] anomaly_score = iso_forest.decision_function(real_time_data_scaled)

[0190] anomaly_label = iso_forest.predict(real_time_data_scaled)

[0191] # Output result

[0192] if anomaly_label[0] == -1:

[0193] print(f'Detected abnormal state, anomaly score: {anomaly_score[0]}')

[0194] else:

[0195] print(f'Power supply is operating normally, anomaly score: {anomaly_score[0]}')

[0196] # Dynamically update the model (assuming to update once every 1000 new data points)

[0197] def update_model(new_data):

[0198] global iso_forest

[0199] new_data_scaled = scaler.transform(new_data)

[0200] iso_forest = IsolationForest(contamination = 0.1, random_state = 42)

[0201] iso_forest.fit(np.vstack((data_scaled, new_data_scaled)))

[0202] # Example: Update the model

[0203] new_data = np.array([[222, 11, 31], [218, 9, 29]])

[0204] update_model(new_data)

[0205] Through the above code example, the Isolation Forest algorithm can play a role in the intelligent monitoring system of 5G base station power status, providing efficient and accurate anomaly detection capabilities to ensure the stable operation of the system.

[0206] Meanwhile, as Figure 2 shown, this figure intuitively describes the working principle of the Isolation Forest algorithm: Isolation tree structure: Each tree randomly selects features and split points, and recursively divides the data into subsets. Outliers (black nodes) are usually isolated at shallower levels in the tree due to their sparse distribution. Inliers (gray nodes) require more splits to be isolated individually due to their dense distribution. Anomaly score: The color bar on the left shows the anomaly score range: Scores close to 1: Highly suspicious of being outliers. Scores close to 0.5: Likely to be inliers. The anomaly score is calculated as the average of the path lengths in all isolation trees. The shorter the path, the more likely it is to be an outlier. Multiple isolation trees (1, 2,..., n): The Isolation Forest consists of multiple independent isolation trees, which improves the detection stability through ensemble learning. Each tree randomly cuts the samples, and finally comprehensively evaluates the anomaly of the samples.

[0207] For Figure 3 , this figure shows the anomaly detection results of the Isolation Forest algorithm on two-dimensional data. The specific explanations are as follows: Black dots (data points): Represent all sample points in the dataset. Circles (outlier range): Potential outliers identified by the Isolation Forest and their influence range. Outliers are usually located in areas with low data density and sparse distribution. Prediction error score: 8: Indicates that the model detected 8 outliers. These points may deviate from the normal data distribution. The Isolation Forest distinguishes outliers from inliers by randomly constructing multiple isolation trees. Outliers are more likely to be isolated, so the path length is shorter.

[0208] Among them, the root cause analysis algorithm: Function: When an abnormal state is detected, analyze the cause of the anomaly, locate the fault point, and provide a decision-making basis for maintenance personnel. Implementation details: Analyze the causal relationship between abnormal parameters and possible faults based on a causal inference model (such as a Bayesian network). Use historical data and a rule base to quickly locate possible fault points and generate handling suggestions. Application scenarios: When overcurrent is detected, locate whether it is caused by excessive load or equipment short circuit.

[0209] Among them, the prediction algorithm: Function: Establish a prediction model based on historical data and real-time data to predict the future power supply status and prevent potential problems from occurring. Implementation details: Use machine learning regression models (such as LSTM, random forest regression) to predict and model key parameters. Input the latest data in real time to update the model and output the prediction results (such as the remaining battery life) within a certain period of time in the future. Adjust the operation strategy or send out warning signals according to the prediction results. Application scenarios: Predict the decline trend of battery health or the impact of load increase on system stability.

[0210] Through the collaborative work of the above-mentioned multiple algorithms, the real-time analysis unit can achieve comprehensive monitoring and intelligent management of the power supply status of 5G base stations, including quickly detecting obvious faults, early warning of potential risks, analyzing complex associated problems, and providing accurate fault location and solutions. This greatly improves the stability and reliability of system operation while reducing the operation and maintenance costs.

[0211] The communication module, connected to the main control processing unit, is used to send digital signals and alarm information to the remote monitoring end and receive control instructions issued by the remote monitoring end;

[0212] The communication module is a key part of the 5G base station power supply status intelligent monitoring system to realize data transmission, instruction reception and communication link management. Its functions and components are detailed as follows:

[0213] Wireless communication unit function: Real-time send digital signals and alarm information to the remote monitoring end through wireless networks (such as 4G / 5G, Wi-Fi, LoRa, etc.). Receive control instructions issued by the remote monitoring end and transfer them to the main control processing unit for execution. Design features: Support multiple wireless communication protocols (such as MQTT, HTTP, CoAP) to adapt to different network environments. Provide signal strength detection function to ensure the stability of data transmission. Built-in redundancy mechanism to switch to the backup network when the main network fails (such as switching from 5G to Wi-Fi).

[0214] Wired communication unit function: Achieve high-speed and stable communication with the remote monitoring end through wired networks (such as Ethernet, fiber optic). Design features: Provide high bandwidth support, suitable for large-scale data transmission scenarios. Equipped with anti-interference design to ensure reliable data transmission in complex electromagnetic environments. Support PoE (Power over Ethernet) function to reduce the need for independent power supply.

[0215] Data encryption and decryption unit function: Encrypt the transmitted data to ensure the security of data during transmission and prevent it from being stolen or tampered with. Design features: Support multiple encryption algorithms (such as AES, RSA) to meet different security level requirements. Provide end-to-end encryption function to protect data security throughout the link from data collection to the remote monitoring end. Integrate digital signature technology to verify data integrity and source credibility.

[0216] Function of the communication protocol conversion unit: Convert data of different communication protocols to achieve compatibility between the system and multiple devices or platforms. Design features: Support common industrial protocols (such as Modbus, CAN, OPC-UA) and Internet of Things protocols (such as MQTT, HTTP). Provide automatic protocol recognition function to complete conversion without manual configuration. Support data bridging between heterogeneous networks, for example, forward data of wireless networks to wired networks.

[0217] Function of the communication status monitoring unit: Real-time monitor the status of the communication link to ensure the stability and reliability of data transmission. Send an alarm signal in case of communication anomalies and attempt to automatically repair the connection. Design features: Provide link quality evaluation metrics (such as latency, packet loss rate, bandwidth utilization). Integrate a fault recovery mechanism, such as automatic reconnection and network switching. Support logging and remote diagnosis functions for quickly locating and solving communication problems.

[0218] Through the collaborative work of the above modules, the communication module can achieve the following goals: Efficient data transmission: Support both wireless and wired communication methods to ensure the real-time and reliability of data transmission. Strong security: Protect data security through the encryption and decryption unit to effectively prevent information leakage or tampering. Wide compatibility: Adapt to multiple devices and platforms through the protocol conversion unit without additional hardware modification. Intelligent management: Real-time evaluate the link status through the communication status monitoring unit and automatically repair anomalies to improve the operational stability of the system. This module provides a reliable data interaction channel for the 5G base station power status intelligent monitoring system and is an important part to ensure the efficient operation of the system.

[0219] The power management and security protection module, electrically connected to the main control processing unit, is used to protect the power supply against overvoltage, undervoltage, overcurrent and short circuit, and trigger emergency protection measures in abnormal states;

[0220] The power management and security protection module includes:

[0221] The overvoltage protection unit is used to detect whether the power supply voltage exceeds the preset safety threshold. If it exceeds, trigger protection measures to prevent damage to the power supply;

[0222] The undervoltage protection unit is used to detect whether the power supply voltage is lower than the preset safety threshold. If it is lower, trigger protection measures to prevent power supply overload or battery over-discharge;

[0223] The overcurrent protection unit is used to detect whether the power supply current exceeds the preset safety threshold. If it exceeds, trigger protection measures to prevent power supply overload or short circuit;

[0224] The short circuit protection unit is used to detect whether the power supply has a short circuit. If so, immediately cut off the power supply;

[0225] An emergency protection unit, when detecting an abnormal state, takes corresponding emergency measures according to a preset emergency handling strategy;

[0226] A power status monitoring unit, which is used to monitor the operating status of the power supply in real time and transmit the monitoring data to the main control processing unit for analysis;

[0227] A power optimization unit optimizes and adjusts the power supply according to the analysis results and the instructions from the remote monitoring terminal.

[0228] The power management and security protection module is an important part of the 5G base station power status intelligent monitoring system, responsible for protecting and optimizing the operating status of the power supply in real time to ensure the stability and security of the system. The following are the detailed functions and designs of this module:

[0229] Overvoltage protection unit function: Detect whether the power supply voltage exceeds the preset safety threshold. If it exceeds, trigger protection measures to prevent the power supply equipment from being damaged due to overvoltage. Implementation details: Use a high-precision voltage sensor to monitor the input and output voltages in real time. Set the overvoltage threshold (such as 110% of the rated voltage). When the detected voltage exceeds the range, immediately cut off the power supply or reduce the load. Provide a multi-level response mechanism: Adjust the operating parameters in case of slight overvoltage, and directly disconnect the power supply in case of severe overvoltage. Application scenario: Prevent overvoltage problems caused by unstable external power supply or equipment failures.

[0230] Undervoltage protection unit function: Detect whether the power supply voltage is lower than the preset safety threshold. If it is lower, trigger protection measures to avoid abnormal operation or damage of the equipment due to undervoltage. Implementation details: Monitor the voltages at the input and output ends in real time. When the voltage is lower than the threshold (such as 90% of the rated voltage), trigger an alarm or protection action. In case of undervoltage, maintain the normal operation of the system by starting the backup power supply or adjusting the load. Provide an undervoltage recording function for analyzing the long-term stability of the power supply system. Application scenario: Prevent equipment shutdown or performance degradation caused by insufficient power supply or line problems.

[0231] Overcurrent protection unit function: Detect whether the power supply current exceeds the preset safety threshold. If it exceeds, trigger protection measures to avoid equipment damage due to overload or short circuit. Implementation details: Equip with a high-precision Hall effect sensor or shunt to monitor the current in real time. When the detected current exceeds the rated load (such as 120%), gradually reduce the load or cut off the power supply. Support fast response (millisecond level) to handle sudden overcurrent situations. Application scenario: Prevent damage caused by equipment overloading or line short circuit.

[0232] Short-circuit protection unit function: Detect whether the power supply is short-circuited. If so, immediately cut off the power supply to avoid further damage to the equipment or causing a fire. Implementation details: Monitor the impedance change at the output end in real time. When the impedance drops sharply to near zero ohms, it is judged as a short circuit. After a short circuit occurs, quickly cut off the power supply through a circuit breaker or relay and record the event information. Provide an automatic reset function to resume normal operation after the short circuit is removed. Application scenarios: Cope with emergencies such as line faults and internal short circuits of equipment.

[0233] Emergency protection unit function: When detecting an abnormal state (such as overvoltage, undervoltage, overcurrent, short circuit), execute corresponding measures according to the preset emergency handling strategies to ensure system safety. Implementation details: Support multiple emergency strategies, such as switching to a backup power supply, reducing the load power, sending alarm information, etc. Link with the main control processing unit and dynamically adjust the emergency measures according to the real-time analysis results. Provide two modes: manual and automatic. Maintenance personnel can intervene in the operation according to needs. Application scenarios: Maintain the system operation and minimize losses in case of complex faults.

[0234] Power status monitoring unit function: Monitor the operating status of the power supply in real time, including parameters such as input / output voltage, current, temperature, etc., and transmit the data to the main control processing unit for analysis. Implementation details: Equipped with multiple sensors (such as temperature sensors, current sensors, voltage sensors) to ensure full coverage of key parameters. The data acquisition frequency can be adjusted to adapt to different scenario requirements (such as high-frequency sampling for fault diagnosis). Support historical data storage and trend analysis for long-term performance evaluation and optimization. Application scenarios: Provide accurate data support for the main control processing unit to achieve intelligent analysis and decision-making.

[0235] Power optimization unit function: According to the analysis results of the main control processing unit and the instructions issued by the remote monitoring terminal, optimize and adjust the power supply to improve efficiency and reduce energy consumption. Implementation details: Dynamically adjust the output power and optimize the energy consumption distribution according to the actual load demand. Enter the low-power mode during off-peak hours to extend the equipment life and reduce the operating cost. Provide intelligent charge and discharge management function to reasonably schedule the backup battery to extend its service life. Application scenarios: Improve the energy use efficiency while ensuring the system stability.

[0236] Through the collaborative work of the above units, the power management and safety protection module can: Achieve rapid response and effective protection against various abnormal situations (such as overvoltage, undervoltage, overcurrent, short circuit); Provide comprehensive status monitoring and data support as the basis for intelligent analysis and optimization; Dynamically adjust the operation strategy to improve the system efficiency and reduce the operating cost; Ensure the stability and reliability of the 5G base station in complex environments and provide strong guarantee for the communication network.

[0237] An artificial intelligence analysis module, deployed in the main control processing unit or the remote monitoring end, is used to predict power failures based on historical data and real-time monitoring data, and output fault prediction results;

[0238] The artificial intelligence analysis module includes:

[0239] A data preprocessing unit, used to preprocess the collected power operation parameters;

[0240] A machine learning algorithm unit, using a variety of machine learning algorithms to model the preprocessed data and establish a prediction model for the power status;

[0241] A fault prediction unit, based on historical data and real-time monitoring data, uses the prediction model to predict power failures and output fault prediction results;

[0242] A performance optimization unit, according to the analysis results and the instructions of the remote monitoring end, optimizes the power performance;

[0243] An anomaly detection unit, used to monitor the power operation parameters in real time. If an abnormal state is detected, an alarm message is generated and an analysis of the cause of the anomaly is provided;

[0244] A self-learning unit, by continuously learning and updating model parameters;

[0245] A visualization analysis unit, presenting the analysis results in a visual way.

[0246] The artificial intelligence analysis module is one of the core components of the 5G base station power status intelligent monitoring system, responsible for intelligent processing, fault prediction, performance optimization and visual display of the collected power operation data. The following are the detailed functions and components of this module: Functions of the data preprocessing unit: Clean and standardize the collected power operation parameters (such as voltage, current, temperature, remaining battery power and health status) to provide high-quality data for subsequent analysis. Implementation details: Data cleaning: Remove noise data, outliers and missing values to ensure data integrity. Standardization processing: Normalize data with different dimensions to a unified range (such as 0-1) for algorithm modeling. Feature extraction: Extract key features (such as voltage volatility, current peak, etc.) to reduce redundant information and improve analysis efficiency.

[0247] Functions of the machine learning algorithm unit: Use a variety of machine learning algorithms to model the preprocessed data and establish a power status prediction model. The various machine learning algorithms include the random forest supervised learning algorithm and the isolation forest unsupervised learning algorithm; The model of the random forest supervised learning algorithm is as follows: Feature vector in the input layer: Contains 6 core parameters X = [V in , V out , I load , Tamb , SOC, SOH T Among them: V in : Input voltage, range 200 - 240V, V out : Output voltage, voltage fluctuation rate ≤ 5%, I load : Load current, sampling rate 1kHz, T amb : Ambient temperature, range -40°C to +85°C, SOC: State of Charge of the battery, accuracy ±1%, SOH: State of Health of the battery. The decision tree is constructed as follows. Each decision tree is generated in the following way: Feature subset selection: Randomly select 3 features for each split. The node splitting criterion adopts the principle of minimizing the Gini index Among them p k is the sample proportion of fault type k. The ensemble strategy is as follows. For classification tasks, the majority voting mechanism is adopted For regression tasks, take the average value Feature engineering is shown in Table 1 below

[0248] Table 1

[0249]

[0250] The model training process is as follows: Data preprocessing, sliding window processing: Extract statistical features (mean, variance, extreme values) from a 10 - minute window, normalization processing: The parameter configuration is as follows:

[0251]

[0252] The dynamic update mechanism is as follows: Perform incremental learning every 24 hours: Among them, η = 0.1 is the learning rate, and L is the cross - entropy loss function. The system integration application is as follows: Fault classification scenario: Input: Real - time voltage / current waveform features, Output: Probability distribution of 17 types of faults, Response time: <10ms (optimized through feature pre - calculation); Remaining useful life prediction scenario: Input: Battery charge - discharge cycle data, Output: Remaining useful life (RUL), Error control: ±3% (12 - month cycle)

[0253] The feature importance analysis is shown in Table 2 below:

[0254] Table 2

[0255]

[0256] The model realizes edge computing deployment through the main control processing unit, supporting: a processing throughput of 2000+ features per second, a fault identification accuracy rate of 98.7% (compared with 65% of the traditional threshold method), dynamic feature weight adjustment (the difference degree per tree > 0.85), and forming a complementarity with the isolation forest algorithm in the system: the random forest processes clearly labeled data, and the isolation forest detects unknown abnormal patterns. The overall detection accuracy is improved to 99.2% through Stacking integration.

[0257] The model of the unsupervised learning algorithm of the isolation forest is as follows: input features, multi-dimensional time series parameters: time series containing 6 core parameters Among them: ΔV out : Output voltage volatility, standard deviation of a 1-minute window, Slope of load current change, linear regression coefficient of a 5-second window, σ(T): standard deviation of temperature fluctuation, 15-minute window, Gradient of power change, power decay rate per unit time, ΔR int : Battery internal resistance change rate, ρ(V, I): voltage-current correlation coefficient. The isolation tree is constructed as follows. Each isolation tree is generated in the following way: Feature random selection: Randomly select 2 features for each split. Split value generation: Randomly select a split threshold within the feature value range. Recursive splitting: Until it meets: The data points are completely isolated and reach the maximum tree depth h max = 15. The anomaly scoring mechanism is as follows: Path length calculation: Path length of abnormal data: Anomaly scoring formula: Among them is the path length correction term.

[0258] Feature engineering is as shown in Table 3 below:

[0259] Table 3

[0260]

[0261] The model training process is as follows: Data preprocessing: Sliding window processing: Construct a 30-minute time series window, standardization processing: Parameter configuration is as follows:

[0262]

[0263] The dynamic update mechanism is as follows: Perform incremental training every 6 hours: T new = T old ∪{x new} Reconstruct 10% of the isolation trees.

[0264] The anomaly detection performance is shown in Table 4 below:

[0265] Table 4

[0266]

[0267]

[0268] System integration applications are as follows: Multi-parameter collaborative anomaly detection: Input: Combined voltage / current / temperature fluctuation patterns, Output: Probability distributions of 17 types of composite anomalies, Typical scenario: Detecting the hidden association between voltage fluctuations and temperature increases caused by capacitor aging; Discovery of unknown anomaly patterns: Identifying new anomalies through changes in path length distribution and automatically triggering the main control unit to save the original waveforms during the anomaly period; Collaboration with the supervised model: Outputting anomaly scores as feature inputs for the random forest model and triggering the retraining of the supervised model in real time. This model is deployed through the edge computing unit and supports: Real-time anomaly scoring of 500+ features / second, Dynamic adjustment of feature weights (inter-tree difference > 0.82), and Formation of a complementary detection mechanism with the supervised learning model (accuracy improved to 99.1%)

[0269] Functions of the fault prediction unit: Based on historical data and real-time monitoring data, use the prediction model to predict power supply faults and output fault prediction results. Implementation details: Input real-time monitoring data into the prediction model to calculate possible future abnormal situations. Provide the probability of fault occurrence and the estimated time to provide early warnings for maintenance personnel. Support the prediction of multiple fault types (such as overvoltage, undervoltage, overcurrent, short circuit, etc.).

[0270] Functions of the performance optimization unit: Optimize the power supply performance according to the analysis results and instructions from the remote monitoring terminal, improve the operation efficiency and reduce energy consumption. Implementation details: Dynamically adjust the power output of the power supply and optimize the energy consumption distribution according to the load demand. Optimize the charging and discharging strategies of backup batteries to extend the battery life and improve the backup capacity. Provide optimization suggestions (such as replacing aging equipment or adjusting operation parameters) to improve the overall performance.

[0271] Functions of the anomaly detection unit: Real-time monitor the operation parameters of the power supply. If an abnormal state is detected, generate an alarm message and provide an analysis of the cause of the anomaly. Implementation details: Use unsupervised learning algorithms (such as Isolation Forest, One-Class SVM) to detect anomaly patterns. Provide anomaly classification results (such as overvoltage, undervoltage, etc.) and possible cause analysis (such as equipment aging or too high environmental temperature). Support a multi-level alarm mechanism and send alarm signals at different levels according to the severity of the anomaly.

[0272] Functions of the self-learning unit: Continuously learn and update model parameters to improve prediction accuracy and adaptability. Implementation details: Implement online learning functions to update the model in real time to adapt to environmental changes. Integrate reinforcement learning algorithms to optimize the decision-making process through feedback mechanisms. Automatically identify new patterns or new fault types and expand the model's capabilities.

[0273] Visual analysis unit function: Present the analysis results in a visual way for easy understanding and operation by the operation and maintenance personnel. Implementation details: Provide various visualization tools (such as line charts, bar charts, heat maps) to display the change trends of key parameters and prediction results. Real-time display of the fault location and cause analysis results, supporting interactive operations (such as zooming the time axis or filtering specific events). Output report function, including historical operation status summary and future trend prediction report.

[0274] Through the collaborative work of the above units, the artificial intelligence analysis module can: Provide high-quality data support to achieve accurate fault prediction; Dynamically optimize system performance, improve operation efficiency and reduce energy consumption; Implement real-time anomaly detection and cause analysis to reduce fault downtime; Provide an intuitive visual interface to help operation and maintenance personnel make quick decisions; The continuous self-learning ability ensures that the system adapts to complex and changing environments, providing strong guarantee for the stable operation of 5G base stations.

[0275] The remote monitoring terminal, communicatively connected to the communication module, is used to receive digital signals, alarm information and fault prediction results, and visually monitor and manage the power supply status based on the digital signals and fault prediction results.

[0276] The remote monitoring terminal includes:

[0277] The data reception and storage unit is used to receive the digital signals, alarm information and fault prediction results transmitted by the communication module, and store and manage them;

[0278] The visual monitoring unit visually monitors the power supply status in real time based on the received digital signals and fault prediction results;

[0279] The alarm and notification unit sends alarm information to the operation and maintenance personnel by means of text messages, emails, push notifications when an abnormal state is detected;

[0280] The control and regulation unit remotely controls or regulates the base station power supply equipment according to the analysis results and the instructions of the operation and maintenance personnel;

[0281] The historical data analysis unit uses historical data for trend analysis and performance evaluation, and provides an evaluation report on the long-term operation status;

[0282] The user interface unit provides a friendly user interface;

[0283] The security management unit is used to manage user permissions, data encryption and transmission security to ensure the security of the system and the confidentiality of data.

[0284] The remote monitoring terminal is the core management platform of the 5G base station power status intelligent monitoring system, responsible for data reception, storage, analysis, visual display and remote control operations. The following are its detailed functions and components:

[0285] Function of Data Reception and Storage Unit: Receive digital signals, alarm information, and fault prediction results transmitted by the communication module. Classify, store, and manage the received data to provide support for subsequent analysis. Implementation details: Support multiple data transmission methods (such as TCP / IP, MQTT). Configure an efficient database (such as SQL or NoSQL) to store historical data and real-time data. Provide data compression and backup functions to ensure storage efficiency and security for long-term operation. Application scenario: Used to record historical information on the power supply operation status for trend analysis and problem tracing.

[0286] Function of Visualization Monitoring Unit: Based on the received digital signals and fault prediction results, perform real-time visualization monitoring of the power supply status. Implementation details: Provide various chart forms (such as line charts, pie charts, heat maps) to display the change trends of key parameters (such as voltage, current, temperature, etc.). Support real-time dynamic refresh to ensure that operation and maintenance personnel can quickly grasp the current status. Integrate the Geographic Information System (GIS) to display the operation status of multiple base stations in the form of a map. Application scenario: Help operation and maintenance personnel quickly locate abnormal base stations and take measures.

[0287] Function of Alarm and Notification Unit: When an abnormal status is detected, send alarm information to operation and maintenance personnel through multiple methods. Implementation details: Support multiple notification methods such as SMS, email, and mobile application push. Provide a hierarchical alarm mechanism to send notifications with different priorities according to the severity of the abnormality. Integrate the alarm log function to record all alarm events and their handling situations. Application scenario: Ensure that operation and maintenance personnel can be informed and handle abnormal situations in a timely manner.

[0288] Function of Control and Regulation Unit: According to the analysis results and the instructions of operation and maintenance personnel, remotely control or regulate the base station power supply equipment. Implementation details: Support dynamic adjustment of power parameters (such as output power, voltage, etc.). Provide emergency operation options such as starting the backup power supply and cutting off faulty equipment. Integrate permission management to ensure that only authorized users can perform control operations. Application scenario: Remotely adjust the equipment operation status in case of emergency to avoid further losses.

[0289] Function of Historical Data Analysis Unit: Use historical data for trend analysis and performance evaluation, and provide an evaluation report on the long-term operation status. Implementation details: Apply time series analysis algorithms (such as ARIMA) to predict future operation trends. Automatically generate regular reports, including the change situations of key indicators (such as energy consumption, battery health) and optimization suggestions. Support screening and export of historical data for further research or external audits. Application scenario: Provide a scientific basis for system optimization and operation and maintenance decisions.

[0290] User Interface Unit Function: Provide a friendly user interface to facilitate operation and management of the system by operation and maintenance personnel. Implementation Details: The interface design is simple and intuitive, supporting multi-language switching and custom layout. Provide interactive operations such as zooming the timeline, filtering specific base stations or abnormal events, etc. Support mobile access, enabling operation and maintenance personnel to view the system status and perform operations anytime, anywhere. Application Scenario: Improve operation and maintenance efficiency and user experience.

[0291] Security Management Unit Function: Used to manage user permissions, data encryption, and transmission security to ensure the security of the system and data confidentiality. Implementation Details: User Permission Management: Support multi-level permission settings (such as administrator, ordinary user), restricting access to sensitive operations. Data Encryption: Use AES or RSA algorithms to encrypt transmitted data to prevent information leakage or tampering. Security Audit: Record all user operation logs for tracking potential security issues or violations. Application Scenario: Protect the system from malicious attacks or unauthorized access.

[0292] Through the collaborative work of the above units, the remote monitoring terminal can: achieve comprehensive monitoring and management of the power supply status of 5G base stations; provide real-time abnormal alarm and rapid response capabilities; support remote regulation and optimization to improve the system operation efficiency; provide intuitive data display and in-depth analysis to provide a scientific basis for decision-making; ensure system security and data confidentiality to provide guarantee for the stable operation of 5G networks.

[0293] As Figure 4 shown, the method for intelligent monitoring of the power supply status of 5G base stations includes the following steps:

[0294] S1. Collect the operating parameters of the voltage, current, temperature, remaining battery power, and battery health of the 5G base station power supply through the data acquisition module;

[0295] S2. The signal conditioning and conversion module filters, amplifies, and converts the analog signals in the operating parameters into digital signals;

[0296] S3. The main control processing unit receives the digital signals and performs real-time analysis on them. If an abnormal state is detected, an alarm message is generated;

[0297] S4. The communication module sends the digital signals and alarm messages to the remote monitoring terminal and receives the control instructions issued by the remote monitoring terminal;

[0298] S5. The power management and security protection module protects the power supply against overvoltage, undervoltage, overcurrent, and short circuit, and triggers emergency protection measures in case of abnormal states;

[0299] S6. The artificial intelligence analysis module predicts power supply faults based on historical data and real-time monitoring data and outputs the fault prediction results;

[0300] S7. The remote monitoring terminal receives digital signals, alarm information, and fault prediction results, and visually monitors and manages the power supply status based on the digital signals and fault prediction results.

[0301] Generally speaking, (1) Data acquisition: Acquisition content: Voltage: Measure the input and output voltages of the base station power supply in real time to ensure that the voltage is within a safe range. Current: Monitor the load current of the base station power supply to evaluate the power consumption and load status of the equipment. Temperature: Obtain the temperature inside or around the power supply equipment to prevent equipment damage caused by overheating. Remaining battery power: Evaluate the available time of the backup battery by measuring the discharge capacity of the battery. Battery health: Based on parameters such as the number of charge-discharge cycles and battery internal resistance, judge the aging degree and health status of the battery. Acquisition method: Continuous acquisition mode: Real-time monitor dynamic change parameters. Timed acquisition: Obtain key parameters at preset intervals to reduce system power consumption. Event-triggered acquisition: Automatically start high-frequency acquisition when an anomaly or specific condition is detected. (2) Signal conditioning and conversion: Filtering: Use a low-pass filter to remove high-frequency noise and ensure a pure signal. Use a high-pass filter to remove low-frequency interference and retain valid information. Use a band-pass filter to retain signals in a specific frequency band and reduce the influence of environmental noise. Amplification: Differential amplification: Enhance differential signals and suppress common-mode noise. Variable-gain amplification: Dynamically adjust the gain according to the amplitude of the input signal to avoid overload or distortion. Analog-to-digital conversion: Use high-precision ADC chips (such as ADS1115, ADS1256) to convert the amplified analog signal into a digital signal. Support multi-channel input and process analog signals from different sensors.

[0302] (3) Real-time analysis: Data reception and processing: The main control processing unit receives the digital signals transmitted by the signal conditioning and conversion module, and performs preliminary processing and storage. Real-time analysis algorithms: Threshold comparison: Compare the received digital signals with the preset normal range thresholds to determine whether they exceed the normal range. Trend analysis: Analyze the change trend of digital signals and predict the change situation in the next period of time. Correlation analysis: Analyze the correlation between different operating parameters and discover the correlation rules between parameters. Anomaly detection: Establish a normal model of the power supply status through machine learning methods and judge in real time whether the power supply operation status is abnormal. Root cause analysis: When an abnormal state is detected, analyze the cause of the anomaly and locate the fault point. Prediction: Based on historical data and real-time data, establish a prediction model to predict the future power supply status. Alarm generation: When an abnormal state is detected, generate corresponding alarm information according to the type and severity of the anomaly.

[0303] (4) Communication and Control: Data Transmission: The wireless communication unit sends digital signals and alarm information to the remote monitoring end through wireless networks (such as 4G / 5G, Wi-Fi, LoRa). The wired communication unit realizes communication with the remote monitoring end through wired networks (such as Ethernet, fiber optic). Data Encryption and Decryption: Encrypt the transmitted data to ensure the security of data during transmission. Communication Protocol Conversion: Convert data with different communication protocols to ensure the system is compatible with different types of remote monitoring ends. Communication Status Monitoring: Real-time monitor the status of the communication link, and send an alarm signal if the communication is abnormal. Control Instruction Reception: Receive the control instructions issued by the remote monitoring end and transfer them to the main control processing unit for execution.

[0304] (5) Power Management and Safety Protection: Overvoltage Protection: Detect whether the power supply voltage exceeds the preset safety threshold. If it exceeds, trigger protection measures to prevent damage to the power supply. Undervoltage Protection: Detect whether the power supply voltage is lower than the preset safety threshold. If it is lower, trigger protection measures to prevent power overload or battery over-discharge. Overcurrent Protection: Detect whether the power supply current exceeds the preset safety threshold. If it exceeds, trigger protection measures to prevent power overload or short circuit. Short Circuit Protection: Detect whether a short circuit occurs in the power supply. If it occurs, immediately cut off the power supply. Emergency Protection: When an abnormal state is detected, take corresponding emergency measures according to the preset emergency handling strategy. Power Supply Status Monitoring: Real-time monitor the operating status of the power supply and transmit the monitoring data to the main control processing unit for analysis. Power Optimization: Optimally adjust the power supply according to the analysis results and the instructions from the remote monitoring end.

[0305] (6) Artificial Intelligence Analysis: Data Preprocessing: Perform preprocessing such as cleaning, standardization, and feature extraction on the collected power supply operating parameters. Machine Learning Modeling: Use various machine learning algorithms to model the preprocessed data and establish a prediction model for the power supply status. Fault Prediction: Based on historical data and real-time monitoring data, use the prediction model to predict power supply faults and output fault prediction results. Performance Optimization: Optimize the power supply performance according to the analysis results and the instructions from the remote monitoring end. Anomaly Detection: Real-time monitor the power supply operating parameters. If an abnormal state is detected, generate an alarm message and provide an analysis of the cause of the anomaly. Self-Learning: Continuously learn and update the model parameters to improve the accuracy and adaptability of the prediction model. Visualization Analysis: Present the analysis results in a visual way such as charts and curves to facilitate monitoring and management by operation and maintenance personnel.

[0306] (7) Remote monitoring and management: Data reception and storage: Receive the digital signals, alarm information, and fault prediction results transmitted by the communication module, and store and manage them. Visual monitoring: Based on the received digital signals and fault prediction results, conduct real-time visual monitoring of the power supply status. Alarm and notification: When an abnormal status is detected, send alarm information to the operation and maintenance personnel via text messages, emails, and push notifications. Control and adjustment: Remotely control or adjust the base station power supply equipment according to the analysis results and the instructions of the operation and maintenance personnel. Historical data analysis: Use historical data for trend analysis and performance evaluation, and provide an evaluation report on the long-term operating status. User interface: Provide a friendly user interface to facilitate the operation and maintenance personnel to perform system configuration, monitoring, and management operations. Security management: Manage user permissions, data encryption, and transmission security to ensure the security of the system and the confidentiality of data. Through the above steps, the intelligent monitoring system for the power supply status of 5G base stations can achieve comprehensive monitoring, analysis, and management of the power supply operation status, improve the stability and reliability of the system, and reduce the operation and maintenance costs at the same time.

[0307] The above content further elaborates on the present invention in combination with specific implementation manners. It cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the medical hydrogel technology field to which the present invention pertains, without departing from the concept of the present invention, several simple deductions or substitutions can still be made, and all should be regarded as falling within the protection scope determined by the claims submitted for the present invention.

Claims

1. An intelligent monitoring system for the power supply status of a 5G base station, the system comprising a data acquisition module, a signal conditioning and conversion module, a main control processing unit, a communication module, and a power management and safety protection module; It is characterized in that It further includes: An artificial intelligence analysis module, deployed in the main control processing unit or the remote monitoring end, for predicting power supply failures based on historical data and real-time monitoring data, and outputting failure prediction results; A remote monitoring end, communicatively connected to the communication module, for receiving digital signals, alarm information, and failure prediction results, and visually monitoring and managing the power supply status based on the digital signals and failure prediction results; The main control processing unit includes: A data reception and processing unit, for receiving the digital signals transmitted by the signal conditioning and conversion module, and performing preliminary processing and storage on them; A real-time analysis unit, using a preset algorithm to perform real-time analysis on the received digital signals, and generating alarm information if an abnormal state is detected; A control and regulation unit, according to the analysis results and the instructions of the remote monitoring end, remotely controls or regulates the base station power supply equipment; A communication interface, for connecting to the communication module, performing two-way communication with the remote monitoring end, receiving control instructions, and sending monitoring data and alarm information.

2. The intelligent monitoring system for the power supply status of a 5G base station according to claim 1, wherein: The signal conditioning and conversion module includes: A filter circuit, for filtering out noise from the analog signals output by the data acquisition module; An amplification circuit, for amplifying the analog signals processed by the filter circuit; An analog-to-digital conversion circuit, for converting the amplified analog signals by the amplification circuit into digital signals, and transmitting the digital signals to the main control processing unit for analysis and processing.

3. The intelligent monitoring system for the power supply status of a 5G base station according to claim 1, wherein The communication module includes: A wireless communication unit, for sending digital signals and alarm information to the remote monitoring end through a wireless network, and receiving control instructions issued by the remote monitoring end; A wired communication unit, for realizing communication with the remote monitoring end through a wired network; A data encryption and decryption unit, for encrypting the transmitted data; A communication protocol conversion unit, for converting data of different communication protocols; A communication status monitoring unit, for real-time monitoring of the status of the communication link, and sending an alarm signal in case of communication anomalies.

4. The intelligent monitoring system for the power supply status of a 5G base station according to claim 1, characterized in that The power management and safety protection module includes: An overvoltage protection unit, for detecting whether the power supply voltage exceeds a preset safety threshold, and triggering protection measures if it exceeds, to prevent damage to the power supply; An undervoltage protection unit, for detecting whether the power supply voltage is lower than a preset safety threshold, and triggering protection measures if it is lower, to prevent power supply overload or battery over-discharge; An overcurrent protection unit, for detecting whether the power supply current exceeds a preset safety threshold, and triggering protection measures if it exceeds, to prevent power supply overload or short circuit; A short-circuit protection unit, for detecting whether a short circuit occurs in the power supply, and immediately cutting off the power supply if it occurs; An emergency protection unit, when an abnormal state is detected, performing corresponding emergency measures according to a preset emergency treatment strategy; A power supply status monitoring unit, for real-time monitoring of the operating status of the power supply, and transmitting the monitoring data to the main control processing unit for analysis; The power optimization unit optimizes and adjusts the power supply according to the analysis results and the instructions from the remote monitoring terminal.

5. The intelligent monitoring system for the power supply status of a 5G base station according to claim 1, wherein, The artificial intelligence analysis module includes: The data preprocessing unit is used to preprocess the collected power operation parameters; The machine learning algorithm unit uses a variety of machine learning algorithms to model the preprocessed data and establish a prediction model for the power supply status; among them, the variety of machine learning algorithms include the random forest supervised learning algorithm and the isolation forest unsupervised learning algorithm; Among them, the model of the random forest supervised learning algorithm is as follows: Feature vectors in the input layer: include six types of core parameters X = [V in , V out , I load , T amb , SOC, SOH] T ; Where: V in : Input voltage, range 200 - 240V; V out : Output voltage, volatility ≤ 5%; I load : Load current, sampling rate 1 kHz; T amb : Ambient temperature, ranging from -40°C to +85°C; SOC: Remaining battery power, accuracy ±1%; SOH: Battery health; The decision tree is constructed as follows, and each decision tree is generated in the following way: Feature subset selection: Each split randomly selects 3 features, and the node splitting criterion adopts the principle of minimizing the Gini index. where p k is the sample proportion of fault type k; The integration strategy is as follows. For the classification task, the majority voting mechanism is adopted. For the regression task, the average value is taken. Among them, the model of the isolation forest unsupervised learning algorithm is as follows: Input features, multi-dimensional time series parameters: Time series containing 6 types of core parameters where: ΔV out : output voltage volatility, standard deviation of a 1-minute window; Load current change slope, linear regression coefficient of 5-second window; σ(T): Standard deviation of temperature fluctuation, 15-minute window; Power change gradient, power decay rate per unit time; ΔR int : Rate of change of battery internal resistance; ρ(V, I); Voltage-current correlation coefficient; The isolation tree is constructed as follows, and each isolation tree is generated in the following way: Feature random selection: Randomly select 2 features for each split; Split value generation: Randomly select a splitting threshold within the feature value range and recursively split until the following conditions are met: the data points are completely isolated and the maximum tree depth h max = 15; The fault prediction unit predicts the power supply fault based on historical data and real-time monitoring data, and outputs the fault prediction result; The performance optimization unit optimizes the power supply performance according to the analysis results and the instructions from the remote monitoring terminal; The anomaly detection unit is used to monitor the power operation parameters in real time. If an abnormal state is detected, it generates an alarm message and provides an analysis of the cause of the anomaly; The self-learning unit continuously learns and updates the model parameters; The visualization analysis unit presents the analysis results in a visual way.

6. The intelligent monitoring system for the power supply status of a 5G base station according to claim 1, wherein The remote monitoring terminal includes: The data reception and storage unit is used to receive the digital signals, alarm messages and fault prediction results transmitted by the communication module, and store and manage them; The visualization monitoring unit performs real-time visualization monitoring of the power supply status based on the received digital signals and fault prediction results; The alarm and notification unit sends alarm messages to the operation and maintenance personnel by means of text messages, emails, and push notifications when an abnormal state is detected; The control and adjustment unit remotely controls or adjusts the base station power supply equipment according to the analysis results and the instructions from the operation and maintenance personnel; The historical data analysis unit uses historical data for trend analysis and performance evaluation, and provides an evaluation report on the long-term operation status; The user interface unit provides a friendly user interface; The security management unit is used to manage user permissions, data encryption and transmission security to ensure the security of the system and the confidentiality of data.

7. A method for intelligent monitoring of the power status of a 5G base station, characterized in that, The method includes: S1. Collect the operation parameters of the voltage, current, temperature, remaining battery power and battery health of the 5G base station power supply through the data collection module; S2. The signal conditioning and conversion module filters, amplifies and converts the analog signals in the operation parameters into digital signals; S3. The main control processing unit receives the digital signals and performs real-time analysis on them. If an abnormal state is detected, it generates an alarm message; S4. The communication module sends the digital signals and alarm messages to the remote monitoring terminal and receives the control instructions issued by the remote monitoring terminal; S5. The power management and security protection module protects the power supply against overvoltage, undervoltage, overcurrent and short circuit, and triggers emergency protection measures in case of an abnormal state; S6. The artificial intelligence analysis module predicts power supply failures based on historical data and real-time monitoring data, and outputs the failure prediction results; S7. The remote monitoring terminal receives digital signals, alarm information and failure prediction results, and visually monitors and manages the power supply status based on the digital signals and failure prediction results.

Images