Clustered orderly charging and discharging management system based on LoRa Internet of Things

Abstract

The application discloses a cluster type orderly charging and discharging management system based on LoRa Internet of Things, relates to the technical field of power system management, and is used for solving the problem of how to realize accurate state perception and optimal scheduling of the cluster charging and discharging system under the condition that LoRa Internet of Things communication is limited. The key state under the condition of missing report and time delay is kept continuously available by establishing a belief state of consistency checking and time sequence tracking. The power instruction sequence is formed in the rolling time domain based on the state, and the information value density is obtained through the sensitivity of the target function, so that the reporting set and the time stamp arrangement are organized according to the priority, the key information in the limited air interface is reached first, and the instruction is aligned in the cycle. When the difference between observation and prediction increases, local correction is triggered to limit the adjustment range and quickly converge the error, and the record and index are written back to the next cycle, so that the estimation and the instruction are kept trackable and consistent in the cross-cycle execution.

Description

Technical Field

[0001] This invention relates to the field of power system management technology, and more specifically, to a clustered orderly charging and discharging management system based on LoRa Internet of Things. Background Technology

[0002] With the increasing prevalence of electric vehicles and distributed energy storage devices, the scale of clustered charging and discharging systems in power distribution networks is constantly expanding. These systems typically contain dozens to hundreds of charging and discharging nodes, requiring coordination and control via communication networks. However, they face three main technical challenges in actual operation: First, wireless communication networks inevitably suffer from packet loss and transmission delays, especially in low-power wide-area IoT systems like LoRa, where the uncertainty of node-reported status can lead to deviations in the scheduling system's perception of the actual cluster status. Second, the actuators of the charging and discharging devices themselves exhibit response lag, and device parameters (such as internal resistance and efficiency) change with temperature and aging, further increasing the difficulty of status observation. Finally, traditional scheduling methods either assume perfect communication and are overly optimistic, or adopt extremely conservative strategies that sacrifice economic efficiency, making it difficult to achieve a balance between safety and economy under communication-constrained conditions.

[0003] In existing technologies, some solutions reduce uncertainty by increasing the reporting frequency, but this significantly increases communication burden and node energy consumption, and may even violate the duty cycle limit of LoRa networks. Some solutions use simple weighted averaging or filtering methods to process missing data, but fail to make full use of the physical conservation relationship, which may lead to the state estimation deviating from the actual feasible region. Other solutions use completely conservative robust optimization, but often result in overly conservative and uneconomical scheduling strategies due to the overly loose uncertainty set.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a clustered orderly charging and discharging management system based on LoRa Internet of Things to solve the problems mentioned in the background art.

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

[0007] A clustered, orderly charging and discharging management system based on LoRa IoT includes:

[0008] The belief state representation and update module 102 is used to obtain node reported and unreported records, as well as the previous cycle instruction and timestamp, calculate the state mean, covariance and interval and advance the boundary according to the missing report, maintain the time sequence index and missing report count, form the belief state set and output it.

[0009] The robust swarm optimization module 103 is used to obtain the belief state set and station-level power target and operating constraints, construct rolling time-domain optimization and obtain the power command sequence of each node, extract the sensitivity summary of the objective function to state and command disturbances, and generate the power command of the current period.

[0010] Information value assessment module 104 is used to obtain the sensitivity summary, communication quality index and link parameters, calculate the information value density of each node and form a priority ranking and reporting set definition, and generate a reporting timestamp requirement and node selection list for communication scheduling.

[0011] The communication compliance scheduling module 105 is used to obtain the sorting, the reporting set and the air interface time budget, allocate the node reporting timestamps according to the duty cycle and time-division access rules and generate a timestamp table, and package the timestamp table and power instructions into a delivery message and send it to the gateway.

[0012] In a preferred embodiment, the belief state representation and update module 102 updates the state mean and covariance of successfully reported nodes using unscented Kalman filtering, advances the upper and lower boundaries and increments the missing report count for unreported nodes using interval observation, maintains the timestamp index and writes it back to the belief state set, and establishes measurement source labels, quality markers and data source descriptions for subsequent reference.

[0013] In a preferred embodiment, when processing voltage, current and temperature samples, the belief state representation and update module 102 performs sliding window detrending and robust filtering, shrinks out-of-bounds interval boundaries according to historical boundaries, performs proportional expansion on the main diagonal elements of the covariance and updates the missing report count synchronously, and sets anomaly observation suppression flag and alternative observation source entries, while keeping them consistent with the time series index.

[0014] In a preferred embodiment, the robust population optimization module 103 constructs a rolling time-domain optimization within the control period and the prediction time domain, applies upper and lower bounds and rate of change constraints on node power, obtains a power command sequence and issues only the current period command, uses the remaining sequence for the initialization of the next period, and records the constraint active set, solution status code and running constraint list for subsequent reference, and establishes a correspondence with the time series index.

[0015] In a preferred embodiment, the robust population optimization module 103 calculates the objective function increment by reading the dual variables or applying small perturbations to the state and instructions to form a sensitivity vector arranged by node index, and compresses it to generate the sensitivity summary containing node identifier and period number fields for use by the information value assessment module. It is also encapsulated into a structured description according to the module interface agreement and a version tag is added to support cross-period calls.

[0016] In a preferred embodiment, the information value assessment module 104 calculates the information value density using the sensitivity summary and link parameters. The information value density is a normalized amount of the objective function increment relative to the single packet air interface time. It generates a priority ranking and reporting set, and outputs the timestamp requirement and packet number requirement corresponding to the node. At the same time, it generates a reporting subset of channel resource groups and persists a snapshot of the priority ranking and reporting set for the communication compliance scheduling module to read.

[0017] In a preferred embodiment, the communication compliance scheduling module 105, under the constraints of duty cycle and regional parameters, allocates non-overlapping reporting timestamps to the reporting set, forming a timestamp table containing node identifier, timestamp, sequence number and validity period, and packages the timestamp table and power command into a delivery message and sends it to the gateway. At the same time, it generates a synchronization offset and check value aligned with the gateway clock and writes them into the message header according to the interface agreement to ensure that the mapping relationship between timestamp and node address is consistent.

[0018] In a preferred embodiment, a fast local correction module 106 is also included. The fast local correction module 106 is activated when the residual exceeds a threshold. It constructs a small-scale quadratic programming in the current period and subsequent finite periods, constrains the power adjustment range of a single node and the rate of change of adjacent periods, selects a small number of nodes to generate correction instructions according to a preset upper limit and submits them to the gateway, and records the correction version label and the basis for the selection of the target set for the next period's audit record and the belief state representation and update module to read.

[0019] In a preferred embodiment, the power command message includes a node address field, a power command field, a timestamp field, and a verification field; the uplink reporting message includes a measurement value, a missing report count, and a version number field; the order of the message fields is uniformly parsed by the gateway and corresponds to the link parameters, and the message header carries area parameters and spreading factor indications for joint parsing and use by the communication compliance scheduling module and the information value assessment module.

[0020] In a preferred embodiment, the communication compliance scheduling module 105 maintains a delivery and retransmission queue. The retransmission interval is determined by the gateway policy. Delivery items that have expired and have not been acknowledged are transferred to the retransmission queue. At the end of the control cycle, expired timestamps are cleared, and the retransmission sequence number and acknowledgment status are recorded for scheduling in the next cycle. At the same time, the timestamp table snapshot and selection results are updated and output to the belief status representation and update module.

[0021] The technical effects and advantages of the LoRa IoT-based clustered ordered charging and discharging management system of this invention are as follows:

[0022] This invention addresses IoT scenarios with narrowband bandwidth and random packet loss, as well as ordered charging and discharging scenarios with large node scales. Focusing on two core constraints—limited duty cycle and asynchronous reporting—it establishes a belief state with consistency verification and time-series tracking, ensuring the continuous availability of critical states under conditions of missing reports and delays. Based on this state, a power command sequence is formed in the rolling time domain, and the information value density is obtained through the sensitivity of the objective function. This allows for the prioritization of reporting sets and timestamp arrangements, ensuring that critical information in the limited air interface arrives first and aligns with commands within the cycle. When the difference between observation and prediction increases, local correction is triggered to limit the adjustment range and quickly converge the error. Simultaneously, records and indexes are written back to the next cycle, ensuring that estimates and commands remain traceable and consistently executed across cycles. This enables ordered charging and discharging and stable deployment at the cluster level in weak network and uncertain environments. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of a clustered orderly charging and discharging management system based on LoRa Internet of Things according to the present invention.

[0024] Figure 2 This is a schematic diagram of the workflow of the belief state representation and update module of the present invention;

[0025] Figure 3 This is a schematic diagram of the rolling time-domain optimization process of the robust population optimization module of the present invention;

[0026] Figure 4 This is a schematic diagram of the orchestration process of the information value assessment module and the communication compliance scheduling module of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0028] This application provides a clustered orderly charging and discharging management system based on LoRa IoT, referring to... Figure 1 , Figure 1 This is a system structure block diagram provided in an embodiment of this application. The system is deployed in the central controller of the cluster charging and discharging system and includes:

[0029] The communication statistics and physical conservation joint module 101 is used to periodically perform the following operations: For each node, it counts the arrival of data packets in the most recent hour and calculates the arrival probability pi. For example, if node 23 successfully reported data in all cycles within the most recent hour (12 control cycles), then pi = 12 / 12 = 1.0; if node 47 failed to report in 3 cycles, then pi = 9 / 12 = 0.75. Simultaneously, it calculates the average arrival time interval as the typical communication delay τi. This is based on the readings of the station-level smart meters. The total power command value issued to all nodes in the previous cycle Calculate the physical conservation residual .when A value greater than the corresponding standard kilowatt-hour is considered to indicate significant inconsistency, and this flag will be used for subsequent pruning of the belief state set. The output of this module is the communication quality index {pi, τi} for each node and the system-level conserved residual r(t).

[0030] The belief state representation and update module 102 is used to maintain a belief state for each node i. .in, It is a state mean vector, which usually includes parameters such as SOC and internal resistance; The covariance matrix is ​​used to estimate the error; For the state interval ([ , (etc.). For nodes that successfully report data within period t, an unscented Kalman filter (UKF) is used to absorb new measurements, such as voltage and current, and update the data. and For nodes that have not reported for three consecutive cycles, the state boundary is advanced using an interval observer. Taking SOC as an example, its interval advancement formula is:

[0031] ;

[0032] ;

[0033] Where ηc and ηd are the charging and discharging efficiencies, respectively; Pmin and Pmax are the minimum and maximum node power, respectively; Cmax is the nominal battery capacity; and Δt is the control period (5 minutes). Meanwhile, the covariance matrix... The main diagonal elements (variance) expand at a rate of 0.5% per period to reflect the increase in uncertainty in state estimation.

[0034] Robust population optimization module 103 employs the Receding Horizon Optimization framework, setting the prediction time domain H to 12 periods (i.e., 1 hour). The optimization problem is in the form of:

[0035] ;

[0036] ;

[0037] ;

[0038] ;

[0039] in, The station-level power target at time τ (derived from grid dispatch instructions or local optimization). Let τ be the electricity price at time τ. , , , where is the weighting coefficient used to adjust the importance of tracking accuracy, economy, and smoothness. X is the uncertainty set consisting of the belief states of all nodes. This min-max robust optimization problem is transformed into a deterministic second-order cone programming (SOCP) or linear programming (LP) problem by using dual variables or constrained robust optimization methods. The solution is the optimal power command sequence for each node over the next H cycles, but only the command at the current time t is implemented. .

[0040] Information value assessment module 104 calculates the information value density of each node i in each control period t. First, based on the dual variables obtained from solving the robust optimization problem or through perturbation analysis, estimate the expected improvement ΔJi of the objective function value J caused by the state uncertainty change of each node (e.g., halving the width of its SOC estimation interval). For example, for node 15, its SOC estimate is 50%, and the uncertainty interval is [45%, 55%). The impact on the objective function value when the SOC is 45% and 55% can be calculated, and the larger absolute value is taken as the estimate of ΔJi. Then, calculate... ,in The air interface time required to send a data packet (20 bytes) to node i is approximately 0.2 seconds. Finally, all nodes are sorted according to... The values ​​are sorted from highest to lowest.

[0041] The communication compliance scheduling module 105 calculates the total available air interface time budget for the system based on the 1% duty cycle limit of the LoRa network in the 470MHz band. For a system with 100 nodes, the total available time per hour is 3600 seconds * 1% = 36 seconds. Module 105 prioritizes allocating available time to information value density. The highest-ranked node is prioritized until the time budget is exhausted. Simultaneously, this module employs Time Division Multiple Access (TDMA) to assign a specific timestamp to each node requiring reporting, thus avoiding message conflicts and ensuring communication conforms to the LoRaWAN protocol specifications.

[0042] The fast local error correction module 106 is triggered when a new measurement value arrives. When a measurement value arrives at a node i, module 106 compares it with the predicted belief state value. If the deviation exceeds a preset threshold, for example, if the SOC deviation is greater than 5%, a short time window is formed within the current period t and the next two periods (t+1, t+2), and a fast quadratic programming (QP) problem is solved within this window. The optimization objectives of this QP problem include fast tracking of the station-level target, minimizing the power adjustment amplitude, and maintaining command smoothness. Because the problem is small in scale, involving only a few nodes and a short time domain, its solution time can be controlled within 100 milliseconds, thus achieving fast error correction.

[0043] The instruction shaping and execution compensation module 107, based on the device response characteristics identified in advance through experiments or historical data, adjusts the power commands generated by the optimization module 103 or 106. Preprocessing is performed. For example, for devices with a known 30-second response hysteresis, instructions are issued 30 seconds earlier; alternatively, a first-order inertial filtering algorithm is used to smooth and shape the instruction curve, avoiding sharp changes in instructions that could cause significant execution errors. Preprocessed instructions. Then it is distributed to each charging and discharging node.

[0044] The audit and security switchover module 108 continuously monitors three types of key residuals:

[0045] Tracking residuals: ;

[0046] Physical conservation residuals: ;

[0047] Predicting innovation residuals: For each node, This refers to the difference between the actual measured value and the predicted measured value.

[0048] When any residual sequence exceeds its preset warning threshold for three consecutive periods, module 108 triggers an enhanced monitoring mode. In this mode, the weighting coefficients in the information value assessment module 104 are increased, and the system attempts to collect more data to eliminate uncertainty. When any residual exceeds a higher danger threshold, module 108 forces the system to switch to a fully robust mode. In this mode, the robust optimization module 103 adopts the most conservative uncertainty set (such as the largest state interval), and all safety constraints are tightened to ensure that the system remains safe under any possible circumstances, at which point economic objectives are secondary.

[0049] The online learning and adaptive update module 109 is activated during periods of low system load, such as at night. This module utilizes historical operating data to re-identify and update the response characteristics of the equipment (such as hysteresis time constant and filtering parameters). Simultaneously, based on a large amount of historical data, it establishes a mapping model between the information value density ΔJi and various state variables (such as SOC, SOH, electricity price, power command, etc.) through regression analysis or machine learning methods (such as random forest and gradient boosting tree), thereby obtaining a fast estimation model for ΔJi. This model can significantly reduce the online computation load of the information value assessment module 104 and improve the system response speed.

[0050] Through the coordinated work of the above modules, the system can intelligently allocate limited communication resources, accurately estimate node status, robustly optimize scheduling strategies, and promptly correct deviations and adapt to changes under the strict constraints of LoRa communication, ultimately achieving safe, economical, and reliable cluster charging and discharging management.

[0051] In an optional embodiment, such as Figure 2 As shown, the specific workflow of the belief state representation and update module 102 is as follows:

[0052] Step S201: At the beginning of each control cycle, check the data reporting status of each node.

[0053] Step S202: For nodes that successfully reported, proceed to S203; for nodes that did not report, proceed to S206.

[0054] Step S203: Extract the measured values ​​from the reported data, such as voltage, current, and temperature.

[0055] Step S204: The unscented Kalman filter (UKF) algorithm is used to incorporate the measured values ​​into the state estimate. UKF approximates the state distribution by generating Sigma points, eliminating the need to calculate the Jacobian matrix, making it particularly suitable for nonlinear systems such as batteries. Updated state mean. Covariance It is closer to the true value.

[0056] Step S205: Tighten the state interval based on the updated state estimate. For example, the SOC interval can be set to [μSOC-3σSOC, μSOC+3σSOC], where σSOC is the standard deviation of the SOC estimate.

[0057] Step S206: For nodes that have not reported, confirm the number of consecutive periods K in which they have not reported.

[0058] Step S207: Using an interval observer, the boundary of the state interval is advanced based on the battery model (such as the equivalent circuit model) and the known upper and lower limits of the power command. During the advancement process, process noise is considered, and the interval will gradually expand.

[0059] Step S208: Increase the covariance matrix according to a preset rule (e.g., 0.5% expansion per cycle). The main diagonal elements reflect the increase in uncertainty over time.

[0060] Step S209: Output the updated belief state of all nodes { }, and pass it to the robust optimization module 103.

[0061] In an optional embodiment, such as Figure 3 As shown, the rolling temporal optimization process of the robust population optimization module 103 is as follows:

[0062] Step S301: At time t, obtain the latest belief state of all nodes { }, forming an uncertain set X.

[0063] Step S302: Obtain the station-level power target for the next H cycles And electricity price information {cτ}.

[0064] Step S303: Construct a min-max robust optimization problem, with the objective function including tracking error, electricity cost, and smoothness penalty.

[0065] Step S304: Using the duality principle or robust optimization techniques (such as converting interval uncertainty into linear constraints), the original problem is transformed into a deterministic convex optimization problem (such as LP or SOCP).

[0066] Step S305: Call a convex optimization solver (such as CVXPY, Gurobi) to solve the transformed deterministic problem.

[0067] Step S306: Obtain the optimized solution, which is the power command sequence {ui,τ*} for each node in the future time domain.

[0068] Step S307: Extract only the instructions at the current time t Used to issue and discard instructions at subsequent times (but the cache is used for initialization of the next cycle).

[0069] Step S308: Wait for the next cycle, and repeat S301-S307.

[0070] In an optional embodiment, such as Figure 4As shown, the system manages 50 V2G (vehicle-to-grid) nodes. The main differences compared to the first implementation are: the nodes have bidirectional power flow capability, with a power range of -10kW (powering the grid) to +10kW (charging from the grid); the battery capacities vary significantly (20kWh to 100kWh); and the communication module uses a LoRa+BLE dual-mode connection, directly connecting via BLE when the vehicle is present (high-speed, reliable), and communicating via LoRa when the vehicle is away (low-speed, limited).

[0071] In this configuration, the belief state representation and update module 102 needs to handle vehicle departure situations specially (S401). When a vehicle leaves the site, its last reported SOC (e.g., 60%) and departure timestamp are recorded. During departure, module 102 uses an interval observer to advance its SOC boundary. Because the vehicle may be using low-power electricity or self-discharging, SOCmin will decrease slowly, while SOCmax will remain basically unchanged or decrease slightly (considering self-discharge), and the state interval [ The uncertainty of the SOC estimate gradually increases (S402), reflecting the increase in uncertainty of the SOC estimate with the departure time. When the vehicle re-enters the site (S403), it quickly reports its actual SOC (e.g., 55%), voltage, current, and other status information via BLE connection (S404). After receiving this information, module 102 resets the state uncertainty of the node, and... Set as the SOC to be reported. Reset to a smaller initial value, state range [ Tighten (S405).

[0072] The information value assessment module 104 needs to additionally consider the vehicle's schedulable time window. For vehicles about to leave, their state information value density will increase significantly. A time window factor is introduced into the value density calculation: Where Tremain is the estimated remaining dispatchable time of the vehicle (estimated based on user settings or historical data), and λ is the decay coefficient (e.g., 0.1). This means that the shorter the Tremain, the more efficient the vehicle dispatching. The larger, The higher the priority, the higher the priority for the node to obtain communication resources, so as to ensure that a reasonable scheduling strategy is formulated before it leaves the site.

[0073] The robust swarm optimization module 103 needs to consider the user's travel constraints. A constraint is added to the optimization problem: the State of Charge (SOC) of vehicle i when it leaves the site must not be lower than the minimum value set by the user. For example, 50%. This ensures that the vehicle has enough battery power to complete the user's expected journey after leaving the site, thus improving user satisfaction.

[0074] The specific embodiments described above demonstrate that this system can adapt to different types of charging and discharging nodes (unidirectional charging piles, bidirectional V2G), achieving safe, economical, and reliable clustered charging and discharging management under LoRa communication constraints. Through a combined communication statistics and physical conservation information processing approach, value-driven communication resource allocation, and a multi-timescale optimization framework, the system effectively solves the cluster control challenges in LoRa IoT scenarios.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A clustered, orderly charging and discharging management system based on LoRa IoT, characterized in that, include: The belief state representation and update module 102 is used to obtain node reported and unreported records, as well as the previous cycle instruction and timestamp, calculate the state mean, covariance and interval and advance the boundary according to the missing report, maintain the time sequence index and missing report count, form the belief state set and output it. The robust swarm optimization module 103 is used to obtain the belief state set and station-level power target and operating constraints, construct rolling time-domain optimization and obtain the power command sequence of each node, extract the sensitivity summary of the objective function to state and command disturbances, and generate the power command of the current period. Information value assessment module 104 is used to obtain the sensitivity summary, communication quality index and link parameters, calculate the information value density of each node and form a priority ranking and reporting set definition, and generate a reporting timestamp requirement and node selection list for communication scheduling. The communication compliance scheduling module 105 is used to obtain the sorting, the reporting set and the air interface time budget, allocate the node reporting timestamps according to the duty cycle and time-division access rules and generate a timestamp table, and package the timestamp table and power instructions into a delivery message and send it to the gateway.

2. The clustered orderly charging and discharging management system based on LoRa IoT as described in claim 1, characterized in that: The belief state representation and update module 102 updates the state mean and covariance of successfully reported nodes using unscented Kalman filtering, advances the upper and lower boundaries and increments the missing report count for unreported nodes using interval observation, maintains the timestamp index and writes it back to the belief state set, and establishes measurement source labels, quality markers and data source descriptions for subsequent reference.

3. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 2, characterized in that: When processing voltage, current and temperature samples, the belief state representation and update module 102 performs sliding window detrending and robust filtering, shrinks out-of-bounds interval boundaries according to historical boundaries, performs proportional expansion on the main diagonal elements of the covariance and updates the missing report count synchronously, and sets anomaly observation suppression flag and alternative observation source entries, while keeping them consistent with the time series index.

4. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 1. Its characteristics are: The robust population optimization module 103 constructs rolling time-domain optimization within the control period and prediction time domain, applies upper and lower bounds and rate of change constraints on node power, obtains the power command sequence and issues only the current period command, uses the remaining sequence for the initialization of the next period, and records the constraint active set, solution status code and running constraint list for subsequent reference, and establishes a corresponding relationship with the time series index.

5. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 4, characterized in that: The robust population optimization module 103 calculates the objective function increment by reading dual variables or applying small perturbations to the state and instructions to form a sensitivity vector arranged by node index, and compresses it to generate the sensitivity summary containing node identifier and period number fields for use by the information value assessment module. It is also encapsulated into a structured description according to the module interface agreement and a version tag is added to support cross-period calls.

6. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 1, characterized in that, The information value assessment module 104 calculates the information value density using the sensitivity summary and link parameters. The information value density is a normalized amount of the objective function increment relative to the single packet air interface time. It generates a priority ranking and reporting set, and outputs the timestamp requirement and packet number requirement corresponding to the node. At the same time, it generates a reporting subset of channel resource groups and persists a snapshot of the priority ranking and reporting set for the communication compliance scheduling module to read.

7. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 6, characterized in that, Under the constraints of duty cycle and regional parameters, the communication compliance scheduling module 105 allocates non-overlapping reporting timestamps to the reporting set, forming a timestamp table containing node identifier, timestamp, sequence number and validity period. The timestamp table and power command are packaged into a delivery message and sent to the gateway. At the same time, a synchronization offset and check value aligned with the gateway clock are generated and written into the message header according to the interface agreement to ensure that the mapping relationship between timestamp and node address is consistent.

8. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 1, characterized in that, It also includes a fast local correction module 106, which is activated when the residual exceeds a threshold. It constructs a small-scale quadratic programming within the current period and subsequent finite periods, constrains the power adjustment range of a single node and the rate of change of adjacent periods, selects a small number of nodes to generate correction instructions according to a preset upper limit and submits them to the gateway, and records the correction version label and the basis for the selection of the target set for the next period's audit record and the belief state representation and update module to read.

9. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 4, characterized in that, The power command message includes a node address field, a power command field, a timestamp field, and a verification field. The uplink reporting message includes fields for measurement value, missing report count, and version number. The order of these message fields is uniformly parsed by the gateway and corresponds to the link parameters. The message header carries regional parameters and spreading factor indications for joint parsing and use by the communication compliance scheduling module and the information value assessment module.

10. A clustered orderly charging and discharging management system based on LoRa IoT as described in claim 7, characterized in that, The communication compliance scheduling module 105 maintains the delivery and retransmission queues. The retransmission interval is determined by the gateway policy. Delivery items that have expired and have not been acknowledged are transferred to the retransmission queue. At the end of the control cycle, expired timestamps are cleared, and the retransmission sequence number and acknowledgment status are recorded for scheduling in the next cycle. At the same time, the timestamp table snapshot and selection results are updated and output to the belief status representation and update module.

Images