Intelligent power grid data transmission system and method of narrowband Internet of Things

Abstract

The invention relates to the technical field of power systems and communication, and discloses an intelligent power grid data transmission system and method of a narrowband Internet of Things, and the system comprises a network coverage enhancement module, a low-delay data processing module, a low-power-consumption management module, and an intelligent load management and data analysis module. The method and the device are used for realizing the stability of data transmission in a signal-limited environment. By introducing the multi-network hybrid transmission architecture and deploying the intelligent antenna and the small base station, the signal coverage range and the transmission stability can be remarkably improved in a signal limited or complex environment, and the multi-network hybrid transmission architecture can determine an optimal network according to real-time signal quality, delay and packet loss probability, so that the stability and the reliability of data transmission are ensured, and the data transmission efficiency is improved. The beam forming of the intelligent antenna and the optimized arrangement of the small base station effectively solve the problem that the signal of the traditional NB-IoT is weak in a building dense area or a remote area by improving the signal gain and the coverage depth.

Description

Technical Field

[0001] The present invention relates to the field of power systems and communication technologies, and in particular to a narrowband Internet of Things (NBIOT) smart grid data transmission system and method. Background Art

[0002] In the development of smart grids, the power system has high requirements for the real-time, reliability and coverage of data transmission. As a product of the combination of power systems and information and communication technologies, smart grids rely on sensors and control equipment to collect, transmit and analyze data to achieve intelligent management of power grids. However, the widespread deployment of smart grids faces many challenges, especially in terms of data transmission technology. Smart grid systems usually cover a wide geographical area, including cities, suburbs and remote rural areas. Traditional communication technologies have the problem of low signal coverage in densely built urban areas or remote mountainous areas, underground facilities and other areas with weak signals, resulting in unstable data transmission and affecting the real-time monitoring and management of the system. The data transmission of smart grids needs to meet real-time requirements, especially in fault detection and rapid response scenarios. However, NB-IoT has high latency and is usually suitable for non-real-time data transmission. It cannot support the low-latency requirements of some key applications in smart grids. Therefore, it is difficult to achieve real-time response by relying on NB-IoT for data transmission in smart grids, and the system's rapid response capability is limited. Many sensor devices in the smart grid system are widely distributed, and some devices are located in remote areas and cannot be powered continuously, which limits the battery life of the devices. In addition, the grid system covers a wide area, making daily maintenance of the devices difficult. Therefore, it is necessary to further optimize the power consumption management of the devices to extend the battery life of the devices and reduce the maintenance frequency. Smart grids require accurate prediction and intelligent dispatching of power loads to achieve efficient use of power resources and stable power supply. However, traditional load forecasting methods are difficult to adapt to the dynamic changes in power demand, resulting in low load forecasting accuracy. In addition, abnormal detection of equipment status mainly relies on preset threshold triggers, which cannot achieve predictive maintenance, easily leading to sudden equipment failures and delayed maintenance, affecting the stable operation of the power grid. The transmission data of the smart grid contains user electricity consumption information. The security and privacy protection of its data are important issues in the design. The existing NB-IoT data transmission technology has defects in security and lacks a complete encryption and authentication mechanism, which leads to the leakage of sensitive information and the system being attacked by network attacks.

[0003] In summary, when applied to smart grids, traditional NB-IoT is limited by issues such as network coverage, transmission delay, power consumption and data management, and it is difficult to fully meet the needs of smart grids.

[0004] Therefore, those skilled in the art provide a narrowband Internet of Things (NB-IoT) smart grid data transmission system and method to solve the problems raised in the above background technology. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a narrowband Internet of Things (NB-IoT) smart grid data transmission system and method to solve the problems raised in the above background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a narrowband Internet of Things smart grid data transmission system, including a network coverage enhancement module, a low-latency data processing module, a low-power management module and an intelligent load management and data analysis module: The network coverage enhancement module is used to achieve stability of data transmission in a signal-limited environment; The low-latency data processing module is used to process data in real time at the edge node; The low power consumption management module is used to reduce the power consumption of the device when there is no transmission demand, thereby extending the battery life of the device; The intelligent load management and data analysis module is used to predict and analyze power grid load and support power dispatching and equipment maintenance.

[0007] Preferably, the multi-network hybrid transmission architecture submodule in the network coverage enhancement module adopts a network method based on signal quality, and the determination function of the optimal network is:

[0008] in, , , is the weight coefficient, Represents the network with the best signal quality calculated by the system. represents the optimal network, It is the received signal strength indicator of the current network. is the signal strength threshold, is the current network transmission delay, is the current network packet loss rate.

[0009] Preferably, the smart antenna and small base station deployment submodule in the network coverage enhancement module includes: Small cell optimization deployment, used to achieve optimal signal coverage by minimizing the average distance between small cells and devices; The small base station location optimization satisfies the following formula: , in, is the coordinate position of the small base station, is the coordinate position of the jth device to be covered, The number of devices to be covered; The beamforming control of smart antennas is used to dynamically adjust the antenna direction to improve the signal reception gain. Its gain model is: , in, is the antenna unit gain coefficient, is the beam direction, The antenna is in the beam direction The signal gain on is the number of antenna elements, is a complex exponential term.

[0010] Preferably, the edge computing processing submodule in the low-latency data processing module includes: A fault detection model based on support vector machines is used to monitor device status in real time at edge nodes; The discriminant function of the fault detection model is: , in, is the Lagrange multiplier, is the training sample label, is the kernel function, is the output of the discriminant function, Input data to be classified or predicted. is the number of training samples, is the data feature vector of the i-th training sample, is the bias term.

[0011] Preferably, the optimized transmission protocol and dynamic window submodule in the low-latency data processing module includes: Use the MQTT-SN protocol to achieve low-overhead event-triggered messaging and reduce the transmission of handshake and heartbeat packets; The control model for dynamically adjusting the transmission window size W is: , in, is the base window size, is the average signal strength, The received signal strength indicator of the current signal, Current packet loss rate.

[0012] Preferably, the intelligent sleep and wake-up control submodule in the low power management module includes: Based on the event-triggered sleep mechanism, it enters deep sleep when there is no event trigger; When the trigger condition is met, the system wakes up through the following conditions: , in, Indicates the output value of the trigger signal, Represents the event value, Indicates the threshold value; like hour, 1, indicating that the trigger condition is met and the system will perform the specified operation; like hour, 0, indicating that the trigger condition is not met and the system remains in the original state; The environmental energy collection and power management submodule in the low power consumption management module includes: configuring energy harvesting devices to harvest energy from the environment; The input energy power satisfy: , in, is the area of ​​the energy harvesting device, is the ambient energy density, is the energy conversion efficiency.

[0013] Preferably, the load forecasting submodule based on machine learning in the intelligent load management and data analysis module includes: Use the LSTM model to predict the load time series data. The load prediction formula is: , in, For the future The load value at the moment, is the prediction function, is the load data at time n+1; The model minimizes the loss function Optimize prediction accuracy, in, is the loss function, It is The actual load value at the moment, is the first The load value at the moment.

[0014] Preferably, the load scheduling optimization algorithm in the intelligent load management and data analysis module includes: Optimize load dispatch based on load forecast results. The goal is to minimize load peak and dispatch cost. The optimization objective function is: , in, is the load value, is the weight coefficient, The scheduling cost.

[0015] Preferably, the anomaly detection and predictive maintenance submodule in the intelligent load management and data analysis module includes: Anomaly detection model based on random forests enables real-time monitoring of equipment through feature extraction; The probability of equipment abnormality satisfies the following formula: , in, is the output result of the i-th tree, is the number of trees, is the input feature, is the input feature The corresponding anomaly probability vector.

[0016] A method for transmitting data of a smart grid of a narrowband Internet of Things comprises the following steps: Step 1: When the system is initialized, the network coverage enhancement module is activated to ensure stable data transmission in signal-limited or complex environments, and the system detects the signal strength, delay and packet loss rate of the candidate networks, and determines the network with the best signal quality through the optimal network determination function; Step 2: Deploy small base stations in key areas where the signal is weak, and use the beamforming control technology of smart antennas to dynamically adjust the antenna direction through the gain model; Step 3: Start the low-latency data processing module on the edge node, and use the support vector machine model for fault detection. The device status is calculated through the discriminant function. If it is abnormal, an alarm is triggered immediately. At the same time, the MQTT-SN protocol is used to realize event-triggered transmission through the transmission window dynamic control model; Step 4: If the system detects that there is no transmission demand, the low power management module switches the device to sleep mode and wakes up the system using an event trigger mechanism. In the sleep mode, the environmental energy collection and power management submodule uses the energy collection formula to collect energy from the environment to power the device and extend the battery life. Step 5: In the intelligent load management and data analysis module, use the LSTM model to predict the load time series data, and optimize the load dispatch based on the load prediction results; Step 6: Use the anomaly detection model in the intelligent load management and data analysis module to monitor the equipment status in real time through random forest, and calculate the anomaly probability based on the random forest model. If the anomaly probability exceeds the threshold, the system generates a maintenance task.

[0017] The present invention provides a narrowband Internet of Things smart grid data transmission system and method. It has the following beneficial effects: 1. The present invention can significantly improve signal coverage and transmission stability in signal-limited or complex environments by introducing a multi-network hybrid transmission architecture and the deployment of smart antennas and small base stations. The multi-network hybrid transmission architecture can determine the optimal network based on real-time signal quality, delay and packet loss rate to ensure the stability and reliability of data transmission. The beamforming of smart antennas and the optimized deployment of small base stations can effectively solve the problem of weak signals in traditional NB-IoT in densely built-up areas or remote areas by improving signal gain and coverage depth, so that the system can maintain efficient data transmission capabilities in a wider range of environments.

[0018] 2. The present invention adopts edge computing processing and optimization of transmission protocol methods to share the burden of cloud computing on the data transmission path, and reduces communication delays through an event-triggered dynamic window adjustment mechanism. The edge computing submodule performs real-time fault detection and preprocessing at the edge node of the data acquisition terminal, reduces the frequency of data upload, and reduces the overall transmission delay of the system. At the same time, through the MQTT-SN protocol and the adaptive transmission window control model, the system can maintain a low transmission delay when the network environment is unstable, realize the strict requirements of the smart grid for real-time data transmission, and improve the response speed and reliability of the system.

[0019] 3. The present invention greatly reduces the power consumption of the device and significantly extends the battery life of the device through the intelligent sleep and event wake-up mechanism and environmental energy collection technology. The intelligent sleep mechanism ensures that the device automatically enters the low-power mode when there is no data transmission demand and wakes up when an event is triggered, which greatly reduces unnecessary power consumption. At the same time, the environmental energy collection module provides auxiliary power for the device through solar energy or vibration energy, so that the device can maintain long-term operation in a remote or non-continuous power supply environment. Its power consumption management strategy effectively improves the overall energy efficiency of the system, making the device suitable for remote or difficult-to-maintain environments where it is unattended for a long time.

[0020] 4. The present invention uses intelligent load prediction and anomaly detection modules and machine learning algorithms to accurately predict the grid load, and conducts real-time monitoring and predictive maintenance of the equipment's operating status. The LSTM-based load prediction model can dynamically adjust the prediction period according to historical data, and maintain high prediction accuracy in the case of load fluctuations, optimize power dispatching, and reduce peak load pressure. The random forest anomaly detection model performs probabilistic analysis on the abnormal state of the equipment, identifies potential failure risks in advance, supports predictive maintenance, reduces power outages and losses caused by sudden equipment failures, effectively improves the operating stability and management efficiency of the smart grid, and realizes intelligent dispatching of grid resources and automation of equipment maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a system framework diagram of the present invention; Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0022] In order to make the technical personnel in the technical field understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in combination with the drawings in the embodiment of the present invention. Obviously, the described embodiment is a partial embodiment of the present invention, not a complete embodiment. Based on the embodiment of the present invention, other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the present invention.

[0023] The present invention is described in detail below in conjunction with the accompanying drawings: Example: Please see attached Figure 1 and attached Figure 2 , an embodiment of the present invention provides a narrowband Internet of Things smart grid data transmission system, including a network coverage enhancement module, a low-latency data processing module, a low power consumption management module and an intelligent load management and data analysis module: Network coverage enhancement module, used to achieve stability of data transmission in signal-limited environments; Low-latency data processing module for real-time data processing at edge nodes; Low power management module, used to reduce device power consumption and extend device battery life when there is no transmission demand; Intelligent load management and data analysis module, used to predict and analyze grid load, and support power dispatching and equipment maintenance.

[0024] Benefits of the network coverage enhancement module: The system significantly improves the data transmission stability in signal-restricted environments. It can determine the best network based on indicators such as signal strength, delay and packet loss rate, ensuring that smart grid equipment can achieve reliable data transmission in remote, underground or densely built areas, reducing signal blind spots and improving the overall connectivity and data transmission success rate of the system; Benefits of low-latency data processing modules: Reduce the delay of data transmission. The application of edge computing shares the burden of cloud computing, allowing the system to quickly detect equipment anomalies and respond in a timely manner. This is very critical in fault alarm and rapid response scenarios. It can significantly improve the real-time performance of the system, reduce the impact of delays on system decisions and operations, and ensure the efficient operation of smart grids at critical moments. Benefits of the low power management module: It reduces the power consumption of the device when it is not in operation. It can decide whether the device should enter the sleep state according to the actual transmission demand. At the same time, it can wake up the device immediately when an event is detected. In addition, it uses environmental energy collection to supplement the power of the device, further extending the battery life of the device. Low power management reduces the maintenance frequency of the device and reduces the long-term operating cost, making the system more suitable for remote and unattended scenarios. The benefits of the intelligent load management and data analysis module include supporting the rational scheduling of grid loads through load forecasting, reducing load peaks and power waste. At the same time, the anomaly detection function can detect potential equipment failures in advance, support predictive maintenance, and avoid sudden failures, so as to improve the utilization efficiency of grid resources, reduce the risk of equipment failures, and contribute to the stable and efficient operation of smart grids.

[0025] The multi-network hybrid transmission architecture submodule in the network coverage enhancement module adopts a network method based on signal quality, and the determination function of the optimal network is:

[0026] in, , , is the weight coefficient, Represents the network with the best signal quality calculated by the system. represents the optimal network, It is the received signal strength indicator of the current network. is the signal strength threshold, is the current network transmission delay, is the packet loss rate of the current network; The smart antenna and small base station deployment submodules in the network coverage enhancement module include: Small cell optimization deployment, used to achieve optimal signal coverage by minimizing the average distance between small cells and devices; Small base station location optimization satisfies the following formula: , in, is the coordinate position of the small base station, is the coordinate position of the jth device to be covered, The number of devices to be covered; The beamforming control of smart antennas is used to dynamically adjust the antenna direction to improve the signal reception gain. Its gain model is: , in, is the antenna unit gain coefficient, is the beam direction, The antenna is in the beam direction The signal gain on is the number of antenna elements, is a complex exponential term.

[0027] Benefits of the multi-network hybrid transmission architecture submodule: Dynamic network switching is achieved to ensure that the system can automatically switch to the optimal network in a complex environment to ensure the stability and reliability of data transmission. The multi-network hybrid transmission architecture can reduce the problem of single network coverage and ensure that the system can achieve stable data transmission under various signal conditions, thereby improving the overall network coverage and system adaptability. The benefits of smart antennas and small base station deployment submodules are that they can provide better coverage in weak signal areas or key areas, ensure signal strength and stability, and significantly improve the success rate of data transmission, so as to effectively reduce signal blind spots and dead angles, ensure comprehensive network coverage, and are suitable for data transmission needs in remote areas, underground facilities and other special places; The benefits of beamforming control of smart antennas are to dynamically adjust the directivity of the antenna, improve the signal strength of the antenna in the target direction, effectively improve the signal coverage and transmission efficiency, improve signal quality, reduce signal interference in irrelevant directions, and improve the overall signal quality and anti-interference ability of the system. This technology is very suitable for scenarios that require precise control of signal direction, and improves the system's adaptability and signal transmission quality in a variety of environments.

[0028] The edge computing processing submodules in the low-latency data processing module include: A fault detection model based on support vector machines is used to monitor device status in real time at edge nodes; The discriminant function of the fault detection model is: , in, is the Lagrange multiplier, is the training sample label, is the kernel function, is the output of the discriminant function, Input data to be classified or predicted. is the number of training samples, is the data feature vector of the i-th training sample, is the bias term; The optimized transmission protocol and dynamic window submodules in the low-latency data processing module include: Use the MQTT-SN protocol to achieve low-overhead event-triggered messaging and reduce the transmission of handshake and heartbeat packets; The control model for dynamically adjusting the transmission window size W is: , in, is the base window size, is the average signal strength, The received signal strength indicator of the current signal, Current packet loss rate.

[0029] The benefits of the edge computing processing submodule are that it can realize real-time monitoring of device status, timely detect potential faults, and effectively distinguish abnormal and normal states in complex data patterns. It performs calculations at the edge node without frequent uploading to the cloud, reducing transmission delays. This module can respond quickly when the device status is abnormal, ensuring the reliability and security of the smart grid and reducing the burden of data transmission and cloud processing. Benefits of optimized transmission protocol and dynamic window submodules Simplified messaging protocol, designed for low power and low bandwidth environments, reduces the number of handshakes and heartbeat packets in traditional protocols. Through event trigger mode, MQTT-SN communicates when an event occurs, significantly reducing the communication overhead during transmission and optimizing transmission efficiency. This protocol is particularly suitable for data transmission scenarios in power grids that require low latency and low power consumption, improving the timeliness of data upload and resource utilization; The benefits of the dynamic transmission window size control model are that when the signal strength is strong and the packet loss rate is low, the window will speed up the transmission rate. When the signal is weak or the packet loss rate is high, the window will ensure the stability of the transmission. Its adaptive adjustment mechanism can effectively reduce the transmission delay when the network conditions fluctuate, ensuring the real-time and reliability of data transmission. It is particularly suitable for application scenarios in smart grids that have high requirements for real-time transmission.

[0030] The intelligent sleep and wake-up control submodules in the low power management module include: Based on the event-triggered sleep mechanism, it enters deep sleep when there is no event trigger; When the trigger condition is met, the system wakes up through the following conditions: , in, Indicates the output value of the trigger signal, Represents the event value, Indicates the threshold value; like hour, 1, indicating that the trigger condition is met and the system will perform the specified operation; like hour, 0, indicating that the trigger condition is not met and the system remains in the original state; The environmental energy collection and power management submodules in the low power management module include: configuring energy harvesting devices to harvest energy from the environment; Input energy power satisfy: , in, is the area of ​​the energy harvesting device, is the ambient energy density, is the energy conversion efficiency.

[0031] The benefits of the intelligent sleep and wake-up control submodule are that it can put the system into low-power mode when there is no data demand or low activity, wake up the device when needed, and determine whether to perform an operation based on the state of the trigger signal. When the event value exceeds the set threshold, the system will wake up immediately and perform necessary operations to ensure the response speed of key events. This mechanism can reduce the energy consumption of the device in an inactive state, and can flexibly wake up the system according to actual event requirements, significantly extending the battery life of the device, which is very suitable for long-term unattended devices in smart grids; Benefits of the environmental energy harvesting and power management submodule: The system can use the energy in the environment to provide auxiliary power for the device when the external power supply is insufficient or unstable, to ensure that the device can obtain sufficient energy. At the same time, it improves energy utilization by improving conversion efficiency. Its energy harvesting method extends the battery life of the device and reduces dependence on battery power, making the device suitable for remote, unpowered or unattended scenarios, further reducing the system's maintenance costs and replacement frequency.

[0032] The machine learning-based load forecasting submodule in the intelligent load management and data analysis module includes: Use the LSTM model to predict the load time series data. The load prediction formula is: , in, For the future The load value at the moment, is the prediction function, is the load data at time n+1; The model minimizes the loss function Optimize prediction accuracy, in, is the loss function, It is The actual load value at the moment, is the first The load value at the moment; The load dispatch optimization algorithms in the intelligent load management and data analysis module include: Optimize load dispatch based on load forecast results. The goal is to minimize load peak and dispatch cost. The optimization objective function is: , in, is the load value, is the weight coefficient, is the scheduling cost; The anomaly detection and predictive maintenance submodules in the intelligent load management and data analysis module include: Anomaly detection model based on random forests enables real-time monitoring of equipment through feature extraction; The probability of equipment abnormality satisfies the following formula: , in, is the output result of the i-th tree, is the number of trees, is the input feature, is the input feature The corresponding anomaly probability vector.

[0033] The advantage of the load forecasting submodule based on machine learning is that it has a high learning ability on time series data, improves the prediction accuracy through long-term dependencies, and optimizes the prediction accuracy of the model by minimizing the loss function, ensuring that the load forecast results are closer to the actual situation. The application of this module helps the power system to prepare load dispatch in advance during peak loads, reduce power waste and overload risks, and help achieve more efficient power resource allocation; Benefits of load dispatch optimization algorithm: Reduce the peak load and unnecessary dispatching cost of the power system. Its optimization can effectively reduce the risk of overload of the power grid, improve the stability of the system, and optimize the efficiency of power resource utilization. At the same time, through the trade-off between dispatching cost and load smoothing, the algorithm can automatically adjust the load distribution strategy to ensure that the power grid can operate stably under the condition of imbalance between supply and demand, and avoid the power supply pressure caused by peak load. The benefits of the anomaly detection and predictive maintenance sub-module are that the results of multiple decision trees are used to make integrated judgments on the equipment status, effectively improving the accuracy of equipment anomaly detection. The random forest model performs adaptive sampling and diversified decision-making on input features, has strong anomaly detection capabilities, and can resist the interference of noise data, reduce the false alarm rate, and then monitor the equipment status in real time. When the probability of anomaly is detected to exceed the set threshold, a maintenance task is generated to support predictive maintenance of the power grid, help discover potential equipment failures in advance, reduce the risk of sudden equipment failures, and improve the stability and security of the power grid.

[0034] A method for transmitting data of a smart grid of a narrowband Internet of Things comprises the following steps: Step 1: When the system is initialized, the network coverage enhancement module is activated to ensure stable data transmission in signal-limited or complex environments, and the system detects the signal strength, delay and packet loss rate of the candidate networks, and determines the network with the best signal quality through the optimal network determination function; Step 2: Deploy small base stations in key areas where the signal is weak, and use the beamforming control technology of smart antennas to dynamically adjust the antenna direction through the gain model; Step 3: Start the low-latency data processing module on the edge node, and use the support vector machine model for fault detection. The device status is calculated through the discriminant function. If it is abnormal, an alarm is triggered immediately. At the same time, the MQTT-SN protocol is used to realize event-triggered transmission through the transmission window dynamic control model; Step 4: If the system detects that there is no transmission demand, the low power management module switches the device to sleep mode and wakes up the system using an event trigger mechanism. In the sleep mode, the environmental energy collection and power management submodule uses the energy collection formula to collect energy from the environment to power the device and extend the battery life. Step 5: In the intelligent load management and data analysis module, use the LSTM model to predict the load time series data, and optimize the load dispatch based on the load prediction results; Step 6: Use the anomaly detection model in the intelligent load management and data analysis module to monitor the equipment status in real time through random forest, and calculate the anomaly probability based on the random forest model. If the anomaly probability exceeds the threshold, the system generates a maintenance task.

[0035] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A narrowband Internet of Things smart grid data transmission system, characterized in that: It includes network coverage enhancement module, low-latency data processing module, low-power management module and intelligent load management and data analysis module: The network coverage enhancement module is used to achieve stability of data transmission in a signal-limited environment; The low-latency data processing module is used to process data in real time at the edge node; The low power consumption management module is used to reduce the power consumption of the device when there is no transmission demand, thereby extending the battery life of the device; The intelligent load management and data analysis module is used to predict and analyze power grid load and support power dispatching and equipment maintenance.

2. The smart grid data transmission system of narrowband Internet of Things according to claim 1, characterized in that: The multi-network hybrid transmission architecture submodule in the network coverage enhancement module adopts a network method based on signal quality, and the determination function of the optimal network is: in, , , is the weight coefficient, Represents the network with the best signal quality calculated by the system. represents the optimal network, It is the received signal strength indicator of the current network. is the signal strength threshold, is the current network transmission delay, is the current network packet loss rate.

3. A narrowband Internet of Things smart grid data transmission system according to claim 1 or 2, characterized in that: The smart antenna and small base station deployment submodule in the network coverage enhancement module includes: Small cell optimization deployment, used to achieve optimal signal coverage by minimizing the average distance between small cells and devices; The small base station location optimization satisfies the following formula: ， in, is the coordinate position of the small base station, is the coordinate position of the jth device to be covered, is the number of devices to be covered; The beamforming control of smart antennas is used to dynamically adjust the antenna direction to improve the signal reception gain. Its gain model is: ， in, is the antenna unit gain coefficient, is the beam direction, The antenna is in the beam direction The signal gain on is the number of antenna elements, is a complex exponential term.

4. The smart grid data transmission system of narrowband Internet of Things according to claim 1, characterized in that: The edge computing processing submodule in the low-latency data processing module includes: A fault detection model based on support vector machines is used to monitor device status in real time at edge nodes; The discriminant function of the fault detection model is: ， in, is the Lagrange multiplier, is the training sample label, is the kernel function, is the output of the discriminant function, Input data to be classified or predicted. is the number of training samples, is the data feature vector of the i-th training sample, is the bias term.

5. A narrowband Internet of Things smart grid data transmission system according to claim 1 or 4, characterized in that: The optimized transmission protocol and dynamic window submodule in the low-latency data processing module includes: Use the MQTT-SN protocol to achieve low-overhead event-triggered messaging and reduce the transmission of handshake and heartbeat packets; The control model for dynamically adjusting the transmission window size W is: ， in, is the base window size, is the average signal strength, It is the received signal strength indicator of the current signal. Current packet loss rate.

6. The smart grid data transmission system of narrowband Internet of Things according to claim 1, characterized in that: The intelligent sleep and wake-up control submodule in the low power management module includes: Based on the event-triggered sleep mechanism, it enters deep sleep when there is no event trigger; When the trigger condition is met, the system wakes up through the following conditions: ， in, Indicates the output value of the trigger signal, Represents the event value, Indicates the threshold value; like hour, 1, indicating that the trigger condition is met and the system will perform the specified operation; like hour, 0, indicating that the trigger condition is not met and the system remains in the original state; The environmental energy collection and power management submodule in the low power consumption management module includes: configuring energy harvesting devices to harvest energy from the environment; The input energy power satisfy: ， in, is the area of ​​the energy harvesting device, is the ambient energy density, is the energy conversion efficiency.

7. The smart grid data transmission system of narrowband Internet of Things according to claim 1, characterized in that: The load forecasting submodule based on machine learning in the intelligent load management and data analysis module includes: Use the LSTM model to predict the load time series data. The load prediction formula is: ， in, For the future The load value at the moment, is the prediction function, is the load data at time n+1; The model minimizes the loss function Optimize prediction accuracy, in, is the loss function, It is The actual load value at the moment, is the first The load value at the moment.

8. The smart grid data transmission system of narrowband Internet of Things according to claim 1, characterized in that: The load dispatch optimization algorithm in the intelligent load management and data analysis module includes: Optimize load dispatch based on load forecast results. The goal is to minimize load peak and dispatch cost. The optimization objective function is: , in, is the load value, is the weight coefficient, For scheduling cost.

9. The smart grid data transmission system of narrowband Internet of Things according to claim 1, characterized in that: The abnormality detection and predictive maintenance submodule in the intelligent load management and data analysis module includes: Anomaly detection model based on random forests enables real-time monitoring of equipment through feature extraction; The probability of equipment abnormality satisfies the following formula: , in, is the output result of the i-th tree, is the number of trees, is the input feature, is the input feature The corresponding anomaly probability vector.

10. A method for transmitting data in a smart grid of a narrowband Internet of Things, according to a smart grid data transmission system of a narrowband Internet of Things according to any one of claims 1 to 9, characterized in that: The steps include: Step 1: When the system is initialized, the network coverage enhancement module is activated to ensure stable data transmission in signal-limited or complex environments, and the system detects the signal strength, delay and packet loss rate of the candidate networks, and determines the network with the best signal quality through the optimal network determination function; Step 2: Deploy small base stations in key areas where the signal is weak, and use the beamforming control technology of smart antennas to dynamically adjust the antenna direction through the gain model; Step 3: Start the low-latency data processing module on the edge node, and use the support vector machine model for fault detection. The device status is calculated through the discriminant function. If it is abnormal, an alarm is triggered immediately. At the same time, the MQTT-SN protocol is used to realize event-triggered transmission through the transmission window dynamic control model; Step 4: If the system detects that there is no transmission demand, the low power management module switches the device to sleep mode and wakes up the system using an event trigger mechanism. In the sleep mode, the environmental energy collection and power management submodule uses the energy collection formula to collect energy from the environment to power the device and extend the battery life. Step 5: In the intelligent load management and data analysis module, use the LSTM model to predict the load time series data, and optimize the load dispatch based on the load prediction results; Step 6: Use the anomaly detection model in the intelligent load management and data analysis module to monitor the equipment status in real time through random forest, and calculate the anomaly probability based on the random forest model. If the anomaly probability exceeds the threshold, the system generates a maintenance task.

Images