Home energy storage power supply remote management system based on internet of things and management method thereof

Abstract

The application belongs to the technical field of household power intelligent scheduling, and particularly discloses a household energy storage power remote management system based on an Internet of Things and a management method thereof, which comprises a cloud prediction module, an edge intelligent gateway, a battery management module and a multi-source data acquisition module; the battery state, real-time electricity price, environmental data and user preferences are acquired through the multi-source data acquisition module; the load and power generation prediction values and prediction confidence are acquired through the cloud prediction module; based on the acquired and predicted data, the multi-parameter dynamic coupling algorithm is executed through the edge intelligent gateway to dynamically adjust and optimize the weight, safety boundary and stress threshold; according to the coupling result, the optimization problem is solved and the battery scheduling instruction is generated; the battery charging and discharging behavior is controlled by executing the battery scheduling instruction; the whole life cycle adaptive management of the battery can be realized, the decision is accurate, the adaptive capacity is strong, the balance of economic benefits can be realized, and the system robustness is enhanced.

Description

Technical Field

[0001] This invention belongs to the field of home power intelligent dispatching technology, specifically relating to a remote management system and management method for home energy storage power supply based on the Internet of Things. Background Technology

[0002] With the increasing adoption of renewable energy in homes and the large-scale deployment of energy storage systems, the complexity and demand for intelligent home energy management are rising. Existing home energy storage management systems, such as the Chinese invention patent with publication number CN110784011A, can perform charge and discharge scheduling based on time-of-use pricing to achieve a certain degree of economic optimization, but their optimization objectives are singular and fail to comprehensively consider multi-dimensional factors such as battery health status, predictive uncertainties, and real-time grid conditions.

[0003] The existing Chinese invention patent with publication number CN112260327A improves dispatch accuracy by introducing artificial intelligence algorithms for power generation and load forecasting. However, its decision-making process remains within an open-loop framework of "prediction-optimization," lacking dynamic feedback and coupling mechanisms between parameters. For example, battery aging is a slow process, and the existing system cannot adaptively adjust its optimization strategy according to the battery's health status; when prediction uncertainty is high, the system still uses fixed safety boundaries, which may lead to decision-making risks; under extreme electricity price scenarios, the system cannot intelligently balance economic benefits with battery degradation. Summary of the Invention

[0004] To address the problems of existing technologies, this invention proposes an IoT-based remote management system and method for home energy storage power supplies. This system can achieve complex trade-offs among multiple objectives such as economy, safety, and battery life in home energy storage systems, thereby improving overall system performance and adaptability. The invention provides the following technical solution: A remote management system for home energy storage power based on the Internet of Things includes: The cloud-based forecasting module is used to predict load demand and photovoltaic power generation based on historical and real-time data, and output the prediction confidence level. An edge intelligent gateway, connected to the cloud prediction module and battery management module, is used to execute a multi-parameter dynamic coupling algorithm and generate battery scheduling instructions; The battery management module is used to monitor the battery's state of charge, health status, temperature, and current. A multi-source data acquisition module, connected to the edge smart gateway, is used to collect real-time electricity prices, environmental data, and user preferences; The edge smart gateway is configured to execute a multi-parameter dynamic coupling algorithm based on the data from the multi-source data acquisition module and the cloud prediction module, and output battery power control commands.

[0005] Preferably, the multi-parameter dynamic coupling algorithm includes at least two of the following coupling relationships: Dynamically adjust and optimize target weights based on battery health status; Dynamically adjust the prediction confidence level to optimize the step size and safety margin; The battery stress threshold is dynamically adjusted based on real-time electricity prices.

[0006] Preferably, the coupling relationship of the target weights for dynamically adjusting the battery health status is as follows: Define the health decay coefficient ; Dynamically adjust and optimize the objective function Weighting coefficients in: , , in, , ν represents the initial weights, and μ and ν are configurable coupling coefficients.

[0007] Preferably, the coupling relationship between the dynamic adjustment optimization step size of the prediction confidence and the safety boundary is as follows: Dynamically adjust the battery SOC operating boundary based on the predicted confidence level (Conf): The maximum allowable charging SOC is , The minimum permissible discharge state of charge (SOC) is , Where δ is a configurable boundary adjustment coefficient. and This is the default SOC boundary.

[0008] Preferably, the coupling relationship of the real-time electricity price dynamically adjusting the battery stress threshold is specifically as follows: Define electricity price influencing factors , Dynamically adjust battery stress threshold , in is the initial stress threshold, and k is a configurable adjustment coefficient.

[0009] Preferably, the edge intelligent gateway is configured to: Solving the constrained optimization problem based on the dynamic coupling results generates battery power reference commands. ; The constraints include dynamic SOC boundary constraints and dynamic stress threshold constraints.

[0010] Preferably, it also includes a user interaction module for receiving user-defined work mode preferences; The working modes include at least an economy-first mode, a lifespan-first mode, and a balanced mode. Different working modes correspond to different initial weight coefficient configurations.

[0011] Preferably, the cloud prediction module uses an LSTM neural network model to predict load and power generation, and calculates the prediction confidence level Conf through Dropout layer variance or ensemble learning variance.

[0012] A method for remote management of home energy storage power based on the Internet of Things includes the following steps: The system collects battery status, real-time electricity price, environmental data, and user preferences through a multi-source data acquisition module. Load and power generation forecasts and forecast confidence levels are obtained through a cloud-based forecasting module; Based on the collected and predicted data, a multi-parameter dynamic coupling algorithm is executed through an edge smart gateway to dynamically adjust and optimize weights, safety boundaries, and stress thresholds. Based on the coupling results, solve the optimization problem and generate battery scheduling instructions; The battery scheduling command is executed to control the charging and discharging behavior of the battery.

[0013] Preferably, the step of executing the multi-parameter dynamic coupling algorithm includes at least two of the following coupling relationships: Optimize target weights based on dynamic adjustment of battery health status; Dynamically adjust the optimization step size and safety boundary based on prediction confidence; The battery stress threshold is dynamically adjusted based on real-time electricity prices.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. By dynamically adjusting and optimizing the weights based on battery health status, the system strategy intelligently evolves as the battery ages, shifting from an economic priority to a lifespan priority, thereby enhancing the value of the battery throughout its entire lifecycle. 2. By dynamically adjusting the safety boundary and optimization step size based on the prediction confidence level, the system automatically switches to a conservative mode when the prediction uncertainty is high, thereby enhancing the robustness of the system. 3. By dynamically adjusting the stress threshold through real-time electricity prices, moderate battery stress is intelligently allowed in extreme electricity price scenarios in exchange for high returns, thus achieving global optimization; 4. Through deep coupling and feedback between parameters, the system possesses adaptability and self-optimization capabilities, and its overall performance far exceeds the effect of simply superimposing various technical features. 5. By combining big data prediction in the cloud with real-time optimization at the edge, both the accuracy of decision-making and the real-time nature of control are ensured. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall system structure of the present invention; Figure 2 This is a schematic diagram of the multi-parameter dynamic coupling algorithm of the present invention; Figure 3 This is a schematic diagram illustrating the coupling relationship between battery health status and optimization weights in this invention; Figure 4 This is a schematic diagram illustrating the coupling relationship between prediction confidence and safety boundary in this invention; Figure 5 This is a schematic diagram illustrating the coupling relationship between real-time electricity price and stress threshold in this invention; Figure 6 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0016] To enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. The directional terms mentioned in the following embodiments, such as "up", "down", "left", "right", etc., are only for reference to the directions in the accompanying drawings. Therefore, the directional terms used are for illustration and not for limiting the invention.

[0017] like Figure 1-6 As shown, a remote management system for home energy storage power based on the Internet of Things (IoT) mainly includes the following hardware components: The cloud-based forecasting module is used to predict load demand and photovoltaic power generation based on historical and real-time data, and output the prediction confidence level. An edge intelligent gateway, connected to the cloud prediction module and battery management module, is used to execute a multi-parameter dynamic coupling algorithm and generate battery scheduling instructions; The battery management module is used to monitor the battery's state of charge, health status, temperature, and current. A multi-source data acquisition module, connected to the edge smart gateway, is used to collect real-time electricity prices, environmental data, and user preferences; The edge smart gateway is configured to execute a multi-parameter dynamic coupling algorithm based on the data from the multi-source data acquisition module and the cloud prediction module, and output battery power control commands.

[0018] The cloud-based prediction module, edge smart gateway, battery management module, and multi-source data acquisition module work together to form an organic whole, jointly realizing the intelligent management and optimized operation of home energy storage power supply. This breaks the limitations of traditional single-module independent operation and maximizes the overall performance of the system through information interaction and data sharing between modules.

[0019] Traditional forecasting models often only provide a single-point prediction, which is far from sufficient for energy decisions requiring risk assessment. The cloud-based forecasting module, deployed on a cloud server, relies on its uncertainty quantification capabilities. Employing an LSTM neural network model, it predicts household power load for the next 24 hours based on historical load data, meteorological information, and temporal characteristics. and photovoltaic power generation It outputs the prediction confidence level Conf; the prediction confidence level Conf is a key indicator for measuring the reliability of the prediction results and directly affects the degree of radical or conservative decision-making in subsequent decisions.

[0020] Specifically, confidence scores can be estimated by the variance of the model's Dropout layer or the differences between different models in ensemble learning. Both methods essentially introduce "perturbations" to observe the stability of the model's output. When the input data has clear patterns and strong historical regularity (such as photovoltaic power generation prediction under clear weather conditions), the model's predictions are highly consistent, resulting in a high confidence score (Conf). However, when facing uncertain scenarios such as sudden weather changes or holiday electricity consumption, the model's predictions exhibit greater dispersion, leading to a lower Conf score. This sends a signal to the margins that "the signal-to-noise ratio is decreasing, and decisions should be made with caution."

[0021] The edge intelligent gateway is the core of this system's decision-making, its value lying in the fusion and trade-off between complex cloud-based predictions and rapidly changing local real-time states. Specifically, the edge intelligent gateway uses an embedded processor based on the ARM Cortex-A72 architecture, runs a Linux operating system, and has a built-in optimization solver. This not only ensures sufficient computing performance to handle complex coupled algorithms, but its open Linux environment also enables continuous algorithm iteration and the integration of third-party functions.

[0022] Specifically, the optimization solver can employ open-source tools such as IPOPT or CVXOPT. These solvers are specifically designed for solving constrained nonlinear optimization problems, enabling them to quickly and accurately find the optimal battery power command under multiple dynamic constraints. The gateway maintains a lightweight local database to cache short-term historical data and system states, ensuring basic system stability even during brief disconnections from the cloud, based on the last known valid predictions and real-time data.

[0023] Specifically, the battery management module monitors the battery pack's SOC, SOH, and temperature in real time via the BMS. and current Among them, SOH (State of Health) is a key parameter reflecting the degree of battery aging, and its accurate estimation is crucial.

[0024] Specifically, State of Harm (SOH) estimation can employ Kalman filtering or coulomb counting combined with impedance analysis. Kalman filtering is an optimal estimation algorithm that can dynamically and accurately estimate the battery's internal parameters (such as internal resistance and capacity) by processing noisy voltage and current sequences, thereby deducing the SOH. Coulomb counting combined with impedance analysis, on the other hand, tracks changes in battery capacity during full charge and discharge cycles and the evolution of the AC impedance spectrum to comprehensively assess capacity decay and internal resistance growth. These advanced algorithms enable the system to "sense" subtle battery aging, providing a basis for subsequent strategy adjustments.

[0025] Specifically, the multi-source data acquisition module includes: Electricity price acquisition unit: Connects to the power company's API via Wi-Fi to obtain real-time time-of-use electricity prices; this enables the system to accurately capture price signals in the electricity market, providing a data foundation for "low-charge, high-discharge" arbitrage strategies. Future expansion will support demand response signals, allowing home energy storage to become a flexible unit within a virtual power plant (VPP).

[0026] Environmental sensing unit: Deploys temperature and humidity sensors to collect indoor and outdoor environmental data; this not only assists the cloud in load forecasting, but more importantly, it can be used to assess the battery's operating environment. For example, in high-temperature environments, the system can proactively reduce charging and discharging power or trigger active cooling to slow down battery aging.

[0027] User interaction unit: Provides a mobile app to receive user-defined preferences and strategies. Users are no longer bystanders but active participants in decision-making. For example, users can set personalized instructions such as "There's a party tonight, make sure the battery has enough power to handle peak hours."

[0028] A multi-parameter dynamic coupling algorithm can achieve a leap from static rules to dynamic adaptive strategies. By identifying and quantifying the intrinsic relationships between key parameters and dynamically embedding them into optimization objectives and constraints in conjunction with mathematical relationships, the system strategy gains context awareness and self-regulation capabilities.

[0029] Specifically, the edge intelligent gateway is configured to execute a multi-parameter dynamic coupling algorithm, which includes the following steps: Step S1: System initialization and parameter configuration.

[0030] Download initial parameters from the cloud: Optimize initial weight values. =0.6, =0.4, initial stress threshold =0.8, coupling coefficients μ=0.3, ν=0.2, δ=5, κ=0.5, and electricity price benchmark. =0.6 yuan / kWh. The default SOC boundary is set to... =90%, =20%. These initial parameters constitute the "baseline character" of the system's behavior, while subsequent dynamic coupling will allow its behavior to "flexibly adapt" to the actual situation.

[0031] Step S2: Real-time acquisition and prediction of multi-source data.

[0032] A decision cycle lasts 15 minutes. The edge smart gateway receives: Real-time data from BMS: SOC=60%, SOH=88%. =25℃, =10A; Data from the cloud-based prediction module: =2.5kW, =3.0kW, Conf=0.9; Data from the data acquisition module: Real-time electricity price = 1.2 yuan / kWh.

[0033] Step S3: Multi-parameter dynamic coupling calculation.

[0034] Coupling Relationship 1: Dynamically adjust and optimize weights based on battery health status.

[0035] Specifically, calculate the health decay coefficient: ; Dynamically adjust weights: =0.4 + 0.3 × 0.12 = 0.436; = (0.1, 0.6 - 0.2 × 0.12) = 0.576.

[0036] As the battery ages (SOH drops from 100% to 88%), the system automatically and gradually adjusts its regulatory tendencies, prioritizing battery care, increasing the lifespan weight β, and correspondingly decreasing the economic weight α. This means that for a brand-new battery, the system will more actively participate in high-yield but potentially stressful charge-discharge cycles; while for an aged battery, the system will shift to a gentler, more protective strategy, thereby maximizing benefits throughout the entire battery lifecycle.

[0037] Coupling Relationship 2: Dynamically Adjusting Safety Boundaries Based on Prediction Confidence.

[0038] Specifically, dynamically adjust the SOC operating boundaries: =90% - 5 × (1 - 0.9) = 89.5%; =20%+5×(1-0.9)=20.5%.

[0039] When predictions are inaccurate (low Conf), the system acknowledges its "inability to see the future" and therefore adopts a conservative strategy, tightening the operating boundaries of the State of Charge (SOC). This is equivalent to leaving more "safety buffer capacity" for the batteries to cope with unexpected load surges or sudden drops in photovoltaic power generation, greatly improving the system's robustness and reliability. Conversely, when predictions are highly reliable, the system is more willing to utilize a wider SOC range for scheduling to capture more benefits.

[0040] Coupling relationship 3: Real-time electricity price dynamic adjustment stress threshold.

[0041] Specifically, calculate the factors affecting electricity prices: = =2.0; Dynamically adjust the stress threshold: =0.8× ≈1.131.

[0042] During periods of extremely high electricity prices, the economic value generated by deep battery discharge increases dramatically. At this time, the system intelligently and temporarily relaxes the "stress limit" on the battery (Smax increases from 0.8 to 1.131), allowing the battery to discharge under slightly higher stress within a controllable range. This allows for the exchange of minimal, recoverable battery loss for substantial, certain economic gains. This reflects the system's evolution from mechanically enforcing protection rules to making intelligent decisions based on economic value.

[0043] Step S4: Constrained optimization solution.

[0044] Objective function: ; Constraints: (Dynamic SOC boundary); (Dynamic stress threshold); Power balance constraints: ; Here, all dynamic parameters (α, β, dynamic SOC boundary, ...) obtained from the aforementioned coupled calculations are... These factors all become the "inputs" for this optimization solution. The task of the optimization solver is to find an optimal battery power Pbat_ref under these time-varying multiple constraints, minimizing the dynamically weighted overall objective. The optimal battery power command is obtained through this solution. =-1.2kW (discharge). This result represents a global trade-off, responding to market signals of high electricity prices while also taking into account battery aging and forecast uncertainties.

[0045] Step S5: Instruction execution and continuous monitoring.

[0046] Edge smart gateway will The data is sent to the BMS for execution. The system continuously monitors data changes and re-executes the coupling and optimization process in the next cycle. This forms a continuous closed loop of "perception-coupling-decision-execution-re-perception," enabling the system to always synchronize with environmental changes and maintain an optimal or suboptimal operating state.

[0047] Specifically, the system also includes a user interaction module, allowing users to select different working modes via a mobile app: Economy-first mode: Settings =0.8, =0.2, focusing on reducing electricity costs, suitable for users who are sensitive to electricity bills.

[0048] Lifespan Priority Mode: Settings =0.3, =0.7, focusing on extending battery life, suitable for users who want their devices to serve them for longer.

[0049] Balanced Mode: Settings =0.6, =0.4, balancing economy and battery life, can be used as the default recommended mode for the system.

[0050] Specifically, when the system detects a power grid anomaly (such as a sudden voltage drop), it automatically forces the prediction confidence level Conf to be set to 0.5 and triggers coupling recalculation to quickly tighten the safety boundary and ensure that the system operates in a conservative mode. It can utilize existing coupling logic and achieve instantaneous switching of operating mode without additional complex rules, thus ensuring the safety of equipment and power grid.

[0051] Through the aforementioned multi-parameter dynamic coupling mechanism, this system possesses battery age perception capabilities, with its strategy intelligently evolving as the battery ages, shifting from an initial aggressive strategy to a later stable one, resulting in a significant improvement in global returns. The system also possesses uncertainty perception capabilities, automatically switching to a robust mode when information is incomplete, greatly enhancing system reliability. Furthermore, the system can intelligently and controllably overdraw at critical moments, exchanging minimal and recoverable battery losses for substantial and certain economic benefits. The deep interaction between these technical features transforms the system from a mechanically rule-following automated system into an adaptive, self-learning, and self-optimizing system with preliminary intelligent agent characteristics.

Claims

1. A remote management system for home energy storage power supply based on the Internet of Things, characterized in that, include: The cloud-based forecasting module is used to predict load demand and photovoltaic power generation based on historical and real-time data, and output the prediction confidence level. An edge intelligent gateway, connected to the cloud prediction module and battery management module, is used to execute a multi-parameter dynamic coupling algorithm and generate battery scheduling instructions; The battery management module is used to monitor the battery's state of charge, health status, temperature, and current. A multi-source data acquisition module, connected to the edge smart gateway, is used to collect real-time electricity prices, environmental data, and user preferences; The edge smart gateway is configured to execute a multi-parameter dynamic coupling algorithm based on the data from the multi-source data acquisition module and the cloud prediction module, and output battery power control commands.

2. The IoT-based remote management system for home energy storage power supply according to claim 1, characterized in that, The multi-parameter dynamic coupling algorithm includes at least two of the following coupling relationships: Dynamically adjust and optimize target weights based on battery health status; Dynamically adjust the prediction confidence level to optimize the step size and safety margin; The battery stress threshold is dynamically adjusted based on real-time electricity prices.

3. The IoT-based remote management system for home energy storage power supply according to claim 2, characterized in that, The coupling relationship of the target weights for dynamically adjusting the battery health status is as follows: Define the health decay coefficient ; Dynamically adjust and optimize the objective function Weighting coefficients in: ， ， in, , ν represents the initial weights, and μ and ν are configurable coupling coefficients.

4. The IoT-based remote management system for home energy storage power supply according to claim 2, characterized in that, The coupling relationship between the dynamic adjustment optimization step size of the prediction confidence and the safety boundary is as follows: Dynamically adjust the battery SOC operating boundary based on the predicted confidence level (Conf): The maximum allowable charging SOC is , The minimum permissible discharge state of charge (SOC) is , Where δ is a configurable boundary adjustment coefficient. and This is the default SOC boundary.

5. The IoT-based remote management system for home energy storage power supply according to claim 2, characterized in that, The coupling relationship of the real-time electricity price dynamic adjustment of the battery stress threshold is specifically as follows: Define electricity price influencing factors , Dynamically adjust battery stress threshold , in is the initial stress threshold, and k is a configurable adjustment coefficient.

6. The IoT-based remote management system for home energy storage power supply according to claim 1, characterized in that, The edge smart gateway is configured to: Solving the constrained optimization problem based on the dynamic coupling results generates battery power reference commands. ; The constraints include dynamic SOC boundary constraints and dynamic stress threshold constraints.

7. The IoT-based remote management system for home energy storage power supply according to claim 1, characterized in that, It also includes a user interaction module for receiving user-defined work mode preferences; The working modes include at least an economy-first mode, a lifespan-first mode, and a balanced mode. Different working modes correspond to different initial weight coefficient configurations.

8. The IoT-based remote management system for home energy storage power supply according to claim 1, characterized in that, The cloud-based prediction module uses an LSTM neural network model to predict load and power generation, and calculates the prediction confidence level Conf through Dropout layer variance or ensemble learning variance.

9. A method for remote management of home energy storage power supply based on the Internet of Things, characterized in that, Includes the following steps: The system collects battery status, real-time electricity price, environmental data, and user preferences through a multi-source data acquisition module. Load and power generation forecasts and forecast confidence levels are obtained through a cloud-based forecasting module; Based on the collected and predicted data, a multi-parameter dynamic coupling algorithm is executed through an edge smart gateway to dynamically adjust and optimize weights, safety boundaries, and stress thresholds. Based on the coupling results, solve the optimization problem and generate battery scheduling instructions; The battery scheduling command is executed to control the charging and discharging behavior of the battery.

10. The method for remote management of home energy storage power supply based on the Internet of Things according to claim 9, characterized in that, The steps of executing the multi-parameter dynamic coupling algorithm include at least two of the following coupling relationships: Optimize target weights based on dynamic adjustment of battery health status; Dynamically adjust the optimization step size and safety boundary based on prediction confidence; The battery stress threshold is dynamically adjusted based on real-time electricity prices.

Images