Sodium-ion battery energy storage and charging cooperative operation adjusting method and device and medium

Abstract

The embodiment of the invention discloses an adjusting method and device for energy storage and charging cooperative operation of a sodium-ion battery and a medium, and relates to the technical field of sodium-ion batteries, and the method comprises the steps: obtaining a real-time battery state parameter of a sodium-ion battery energy storage system, a real-time external energy parameter of an external energy source and a real-time charging demand parameter of a charging facility in real time, grading the current cooperative operation condition through a preset fuzzy logic model, and determining the priority of the condition; based on the working condition priority, a cooperative operation strategy is generated and executed, and the cooperative operation strategy comprises an energy flow distribution strategy, a power grid interaction strategy and a vehicle network interaction strategy; and dynamically adjusting strategy parameters in the cooperative operation strategy through a preset fusion control algorithm according to the real-time state parameters of the sodium ion battery energy storage system and the dynamic change of the cooperative operation working condition. The comprehensive energy efficiency of the light storage and charging cooperative system is remarkably improved, and the service life of the battery energy storage body is prolonged.

Description

Technical Field

[0001] This specification relates to the field of sodium-ion battery technology, and in particular to an adjustment method, device and medium for the coordinated operation of sodium-ion battery energy storage and charging. Background Technology

[0002] Sodium-ion batteries, with their abundant resources, relatively low cost, and high safety, have shown promising application prospects in large-scale electrochemical energy storage, smart charging stations, and vehicle-to-grid (V2G) scenarios. To achieve efficient collaborative operation between energy storage systems, charging facilities, and external energy sources (such as photovoltaics and the power grid), existing technical solutions mostly adopt static control strategies based on fixed rules.

[0003] Existing technologies typically pre-determine static charging and discharging parameters (such as fixed current and voltage thresholds) and operating logic for sodium-ion battery energy storage and charging systems. These strategies may be effective when dealing with single or stable operating conditions, but in actual operation, the system faces a multivariate, strongly coupled dynamic environment. The internal state of the battery itself (such as state of charge, temperature, and internal resistance that increases with aging) is constantly changing; the external energy supply (such as photovoltaic output) is intermittent and fluctuating; and user charging demand also exhibits randomness and peak-valley characteristics. Static strategies cannot perceive and respond to these multi-dimensional dynamic changes in real time. When the battery is under non-ideal conditions (such as low temperature, high charge state, or increased internal resistance after aging), using fixed parameters for charging and discharging will exacerbate irreversible side reactions inside the battery, leading to a significant reduction in cycle life and potentially causing safety risks. Furthermore, in the face of the dynamic imbalance between photovoltaic output and charging demand, static strategies are unable to achieve real-time optimal allocation of energy flow, resulting in low photovoltaic absorption rate or high grid dependence and poor overall system energy efficiency. In addition, the lack of coordinated consideration of grid load status and battery status makes it impossible to achieve a balance between peak shaving and valley filling and battery life maintenance, and it is even more difficult to support the fine-grained scheduling of advanced functions such as vehicle-to-grid interaction.

[0004] Therefore, in the process of sodium-ion battery energy storage and charging facilities operating in tandem, the existing control strategies are mostly statically set and cannot adapt to the dynamic internal state of the battery, fluctuating external energy and random charging needs in real time. As a result, the system operating efficiency, economy and battery life cannot be optimized in a coordinated manner, and the needs of large-scale reliable application of sodium-ion batteries in complex dynamic scenarios cannot be met. Summary of the Invention

[0005] This specification provides one or more embodiments of an adjustment method, device, and medium for the coordinated operation of sodium-ion battery energy storage and charging, which is used to solve the following technical problem: During the coordinated operation of sodium-ion battery energy storage and charging facilities, because the existing control strategies are mostly statically set, they cannot adapt to the dynamic internal state of the battery, fluctuating external energy, and random charging needs in real time. As a result, the system operating efficiency, economy, and battery life cannot be optimized in a coordinated manner, and the requirements for large-scale reliable application of sodium-ion batteries in complex dynamic scenarios cannot be met.

[0006] One or more embodiments of this specification employ the following technical solutions: This specification provides one or more embodiments of a method for adjusting the coordinated operation of sodium-ion battery energy storage and charging. The method includes: acquiring real-time battery state parameters of the sodium-ion battery energy storage system, real-time external energy parameters of the external energy source, and real-time charging demand parameters of the charging facility in real time; classifying the current coordinated operation conditions according to the real-time battery state parameters, the real-time charging demand parameters, and the real-time external energy parameters using a preset fuzzy logic model to determine the operation condition priority; generating and executing a coordinated operation strategy for the sodium-ion battery energy storage system, the charging facility, and the external energy source based on the operation condition priority, wherein the coordinated operation strategy includes an energy flow allocation strategy, a grid interaction strategy, and a vehicle-grid interaction strategy; and dynamically adjusting the strategy parameters in the coordinated operation strategy according to the dynamic changes of the real-time state parameters of the sodium-ion battery energy storage system and the coordinated operation conditions using a preset fusion control algorithm.

[0007] This specification provides one or more embodiments of an adjustment device for the coordinated operation of sodium-ion battery energy storage and charging, comprising: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform the above-described method.

[0008] This specification provides one or more embodiments of a non-volatile computer storage medium storing computer-executable instructions configured to perform the above-described method.

[0009] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects: At the perception layer, battery state parameters, external energy parameters, and charging demand parameters are collected simultaneously to form a multi-dimensional state space covering all elements. This breaks through the limitations of traditional control that relies on only a single variable (such as only SOC or only grid electricity price), and can understand the comprehensive characteristics of the current operating environment from a global perspective. At the decision-making layer, a preset fuzzy logic model is introduced to classify and prioritize the current collaborative operating conditions, which is in line with the real-world scenarios of strong fluctuations in new energy output and high randomness in user charging behavior. Different priorities correspond to different control objectives, thereby avoiding strategy confusion under multiple conflicting objectives. At the execution layer, energy flow allocation strategies, grid interaction strategies, and vehicle-grid interaction strategies generated based on operating condition priorities achieve multi-level and multi-dimensional collaborative control. The energy flow allocation strategy ensures that local renewable energy (such as photovoltaics) is preferentially consumed locally, reducing the impact on the grid. The grid interaction strategy utilizes the fast response characteristics of sodium-ion batteries to discharge and support loads during peak grid periods and charge and store energy during off-peak periods, playing a role in peak shaving and valley filling. The vehicle-grid interaction strategy further treats electric vehicles as mobile energy storage units, allowing them to supply power to the grid or energy storage system in reverse under specific operating conditions, greatly improving system flexibility. At the optimization layer, the strategy parameters are dynamically adjusted through a pre-defined fusion control algorithm, giving the system continuous learning and adaptive capabilities. This achieves deep collaboration between the sodium-ion battery energy storage system, charging facilities, and external energy sources, significantly improving energy utilization efficiency, system operational safety, and grid support capabilities, thus meeting the requirements for large-scale reliable application of sodium-ion batteries in complex dynamic scenarios. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings: Figure 1 A flowchart illustrating an adjustment method for coordinated operation of sodium-ion battery energy storage and charging, provided in an embodiment of this specification. Figure 2 A charge-discharge efficiency comparison experimental curve provided for an embodiment of this specification; Figure 3 A cycle life comparison experimental curve provided for an embodiment of this specification; Figure 4 This is a schematic diagram of the structure of an adjustment device for coordinated operation of sodium-ion battery energy storage and charging, provided as an embodiment of this specification. Detailed Implementation

[0011] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0012] This specification provides an adjustment method for the coordinated operation of sodium-ion battery energy storage and charging. It should be noted that the execution subject in this specification embodiment can be a server or any device with data processing capabilities. Figure 1 This is a flowchart illustrating an adjustment method for the coordinated operation of sodium-ion battery energy storage and charging, as provided in the embodiments of this specification. Figure 1 As shown, the main steps include the following: Step S101: Real-time battery status parameters of the sodium-ion battery energy storage system, real-time external energy parameters of the external energy source, and real-time charging demand parameters of the charging facility are acquired in real time. Based on the real-time battery status parameters, real-time charging demand parameters, and real-time external energy parameters, the current collaborative operation conditions are classified and the priority of the operation conditions is determined through a preset fuzzy logic model.

[0013] In one embodiment of this specification, the state parameters (temperature T, state of charge SOC, DC internal resistance DCIR, terminal voltage U, charging and discharging current I), external energy parameters (photovoltaic output P_PV, real-time grid load P_grid), and charging demand parameters (number of charging guns N_gun, average charging power P_avg) of the sodium-ion battery energy storage system are collected in real time, with a sampling frequency ≥10Hz.

[0014] Specifically, real-time battery state parameters refer to the internal operating state quantities of a sodium-ion battery energy storage system, mainly including: the battery temperature (T) obtained by sampling at a frequency not lower than a preset frequency using an NTC temperature sensor (such as the MF52-10K type) installed at the tab of the battery cell; the state of charge (SOC) obtained by real-time acquisition of charging and discharging current (I) and terminal voltage (U) by a current sensor (such as the ACS712) and a voltage sensor, and by fusion calculation using the ampere-hour integration method combined with the periodic open-circuit voltage calibration method; and the DC internal resistance (DCIR) obtained by measurement using an AC injection method internal resistance tester (such as using a 1kHz test frequency).

[0015] Real-time external energy parameters refer to the instantaneous status of the external energy sources connected to the system, including the photovoltaic array output power (P_PV) obtained by a photovoltaic power sensor at a specific sampling frequency, and the real-time grid load (P_grid) obtained by the smart grid load monitoring module. Real-time charging demand parameters refer to the instantaneous load status of the charging facilities, which is obtained in real time by the charging controller. The core parameters include the number of charging guns currently in operation (N_gun) and the average charging power per gun (P_avg) calculated based on the total charging power and the number of operating guns. All the above parameters are converted from analog to digital by a high-precision ADC chip (such as ADS1256) and synchronized and preprocessed by the main control MCU (such as STM32H750) to ensure data time consistency and availability.

[0016] After obtaining the above parameters, based on the real-time battery status parameters, real-time charging demand parameters, and real-time external energy parameters, the current collaborative operation conditions are classified and the priority of the operation conditions is determined through a preset fuzzy logic model. This is specifically achieved in the following ways: First, the first ratio of photovoltaic output power in the real-time external energy parameters to the average charging power in the real-time charging demand parameters is calculated, i.e., the ratio of photovoltaic output power (P_PV) to average charging power (P_avg). This ratio directly reflects the real-time supply and demand relationship between photovoltaic power generation and charging load. Second, the second ratio of the real-time grid load to the preset grid rated capacity in the real-time external energy parameters is calculated. This second ratio is the ratio of the real-time grid load (P_grid) to the preset grid rated capacity, and it is used to quantify the load pressure level on the grid side. Simultaneously, the real-time calculated state of charge (SOC) value is used as the third core input.

[0017] The first ratio, the second ratio, and the state of charge value from the real-time battery state parameters are input into a pre-built fuzzy logic model stored in the processor. The model includes three core units: fuzzification processing, fuzzy inference, and defuzzification. Table 1 below provides an example of the input parameters and level division of a fuzzy logic model provided in an embodiment of this specification: Table 1. Examples of input parameters and hierarchical classification for a fuzzy logic model provided in the embodiments of this specification. ; The fuzzification processing unit in the preset fuzzy logic model determines the first membership value set of the first ratio, the state of charge (SOC), and the second ratio relative to multiple preset membership functions. Within the fuzzification processing unit, multiple semantic fuzzy sets, or linguistic variables, are defined for each input variable (first ratio, SOC, second ratio). These sets directly correspond to the level description intervals under each parameter in Table 1. For example, for the first ratio, five fuzzy sets—very small, small, medium, large, and very large—can be defined, based on the numerical ranges given in Table 1 (e.g., <0.5, 0.5~0.8, 0.8~1.2, 1.2~1.5, >1.5). A corresponding membership function (e.g., triangular or trapezoidal function) is configured for each linguistic variable. The membership degree is a value between 0 and 1, precisely quantifying the degree to which a specific input value belongs to a certain fuzzy set. Taking the first ratio as an example, let its input value... According to Table 1, the ratio falls within the range... It belongs to a lower level, in The content belongs to the medium level. Fuzziness is achieved using overlapping triangular membership functions, with the vertices of the lower-level functions located at... left boundary right boundary The vertex of an intermediate-order function is at... left boundary right boundary The calculation process is as follows: Calculate the membership degree to the smaller level: Located at the rising edge of the small function According to the linear interpolation formula: Calculate the membership degree for the intermediate level: Less than the left boundary of the medium function ,therefore In this example, The membership degree of the smaller level is approximately The membership degree of the intermediate level is .

[0018] The system performs this operation on each of the three input variables, thereby generating a set of membership values ​​for each input value to its respective fuzzy set. This set of values ​​constitutes the first membership value set. The function of this unit is to convert the precise numerical values ​​of the input into membership values ​​corresponding to each linguistic variable, thus forming the first membership value set.

[0019] The fuzzy inference unit is the core decision-making part of the model, and it has a built-in fuzzy rule base predefined based on domain knowledge. Each rule is usually in the form of "IF (first ratio is A) AND (SOC is B) AND (second ratio is C) THEN (operating condition level is D)", where A, B, C, and D are linguistic variables, and the source of the rule is directly mapped from Table 1. Through the fuzzy inference unit in the preset fuzzy logic model, inference calculation is performed based on the preset fuzzy rule base and the first membership value set to obtain multiple second membership values ​​representing that the current collaborative operating condition belongs to multiple preset operating condition levels; the fuzzy inference unit applies all relevant fuzzy rules for parallel inference based on the first membership value set output by the fuzzification processing unit. It receives the first membership value set from the fuzzification processing unit and performs parallel activation and calculation on all applicable rules in the rule base. Fuzzy logic operations (such as taking the min operator for "AND" connection) are used to determine the activation strength of the precondition of each rule and this strength is passed to the conclusion. Finally, the conclusions of all activated rules regarding the same output level (such as "overload") are aggregated (e.g., by taking the maximum value), thus obtaining the probability that the current collaborative operation condition belongs to one of the five preset operating condition levels: "overload", "high load", "balanced", "low load", and "idle". This results in a set of second membership values. Each value (e.g., the probability of belonging to the overload level is 0.8) represents a fuzzy judgment of the operating condition after the system integrates all inputs.

[0020] The defuzzification processing unit in the pre-defined fuzzy logic model processes the multiple second membership values ​​to determine the target operating condition level to which the current collaborative operating condition belongs. The defuzzification processing unit is responsible for converting the set of second membership values ​​representing different operating condition level possibilities output by fuzzy inference back into a unique operating condition level determination. Commonly used methods include the centroid method and the maximum membership method. Through this unit's processing, the target operating condition level to which the current collaborative operating condition belongs is finally determined. Then, according to the pre-defined mapping relationship between operating condition levels and priorities, the target operating condition level is mapped to the corresponding operating condition priority. Based on the pre-defined mapping relationship table (e.g., mapping "overload" to the highest priority level 1 and "idle" to the lowest priority level 5), the target operating condition level is mapped to the operating condition priority value.

[0021] Leveraging the advantages of fuzzy logic models in handling uncertainties and nonlinear problems, precise numerical inputs are transformed into qualitative evaluations with linguistic meaning. Reasoning is then performed using a fuzzy rule base that simulates human expert experience, enabling more realistic and flexible intelligent classification of complex collaborative operating conditions. This overcomes the problems of frequent strategy switching or sluggish response that can result from the absolutism of boundary demarcation in conventional methods, making system state assessment more continuous and reasonable. Based on the operating condition priorities generated by this intelligent classification, more targeted collaborative operating strategies (such as energy flow allocation, grid interaction, and vehicle-grid interaction) can be triggered.

[0022] Step S102: Based on operating condition priorities, generate and execute a coordinated operation strategy for the sodium-ion battery energy storage system, charging facilities, and external energy sources.

[0023] The collaborative operation strategy includes energy flow allocation strategy, grid interaction strategy and vehicle-grid interaction strategy; Based on this operating condition priority, a coordinated operation strategy for the sodium-ion battery energy storage system, the charging facility, and the external energy source is generated and executed, specifically through the following methods: Higher priority (e.g., level 1) indicates a more stressful or intervention-critical system condition (e.g., overload), while lower priority (e.g., level 5) indicates a more relaxed system condition (e.g., idle). Based on the preset range of the condition's priority, the corresponding sub-strategy module is activated and executed. In response to a condition's priority falling within the preset first priority range, the system determines the current condition as an energy supply-demand imbalance, generating and executing the energy flow allocation strategy. In other words, when the system determines the condition's priority to be within the preset first priority range—for example, corresponding to overload or high load levels in fuzzy logic hierarchy—it means the most prominent issue is an energy supply-demand imbalance, and the energy flow allocation strategy will be generated and executed first.

[0024] The energy flow allocation strategy includes comparing the photovoltaic output power with the average charging power. If the photovoltaic output power is greater than the average charging power, a first charging command is generated to control the sodium-ion battery energy storage system to charge. If the photovoltaic output power is less than the average charging power, a first discharging command is generated to control the sodium-ion battery energy storage system to discharge. The core decision-making basis of this strategy is to compare the photovoltaic output power (P_PV, continuously collected by a photovoltaic power sensor) with the average charging power (P_avg, calculated by the charging controller) in real time. If the photovoltaic output power is greater than the average charging power, it is determined that the photovoltaic output is excessive, and a first charging command is generated. The first charging command is a set of control commands containing target parameters (such as charging current rate and target voltage), which is sent to the battery management system (BMS) of the sodium-ion battery energy storage system to control the energy storage system to store the excess electrical energy generated by the photovoltaic system at an appropriate power (such as a certain charging current rate). Conversely, if the photovoltaic output power is less than the average charging power, it is determined that the photovoltaic output is insufficient, and the first discharge command is generated to control the energy storage system to release electrical energy to make up for the power gap of the charging facilities, thereby maintaining the continuity of charging services and maximizing the absorption of local photovoltaic power.

[0025] In response to the operating condition priority indicator being within a preset second priority range, the system determines that the current operating condition is a grid load adjustment opportunity, and generates and executes the grid interaction strategy. When the operating condition priority indicator is currently within the preset second priority range, for example, corresponding to a balanced or specific grid-related condition requiring attention, the core task shifts to grid load adjustment, and a grid interaction strategy is then generated and executed. The grid interaction strategy includes: determining whether the current time period is a preset grid peak period; if so, generating a second discharge command to prioritize the sodium-ion battery energy storage system supplying power to the charging facility; if not, generating a second charging command to control the sodium-ion battery energy storage system to charge from the grid. The execution of the strategy depends on obtaining the current time information and determining whether it belongs to a preset grid peak period (such as certain daytime hourly segments set based on historical load curves or electricity price signals) or a grid off-peak period (such as at night). If it is determined that the current time is a grid peak period, a second discharge command is generated. This command has a higher priority than direct photovoltaic supply, aiming to control the sodium-ion battery energy storage system to prioritize discharging to the charging facility, thereby reducing the purchase of electricity from the high-priced grid and playing a peak-shaving role. If it is during a low-voltage period of the power grid, a second charging command is generated to control the energy storage system to charge from the power grid, utilizing the low-voltage electricity price to store energy, thereby filling the valley and reducing the overall cost of electricity.

[0026] In response to the operating condition priority indication being within the preset third priority range and the state of charge (SOC) in the real-time battery state parameters reaching a preset high threshold, the conditions for vehicle-to-grid (V2G) interaction are determined to be met, and the V2G interaction strategy is generated and executed. When the operating condition priority indication is within the preset third priority range, it typically corresponds to an operating condition where the system's own resources are relatively abundant and the grid has peak-shaving needs, such as when the energy storage SOC is high. Simultaneously, when the system detects that the SOC in the real-time battery state parameters reaches or exceeds a preset high threshold (e.g., 80%), the conditions for starting vehicle-to-grid (V2G) interaction are determined to be met. At this time, the system will generate and execute the V2G interaction strategy. This V2G interaction strategy includes: generating a discharge enable command to allow new energy vehicles connected to the charging facility to discharge to the grid according to a preset discharge power limit. The discharge enable command is sent to the V2G function module of the charging pile, unlocking the permission for connected new energy vehicles to discharge back to the grid. At the same time, the strategy will set a preset discharge power limit based on the real-time peak-shaving demand of the power grid and battery safety limitations, and inform the vehicles through communication protocols (such as ISO 15118, CHAdeMO, etc.) to control them to feed power to the power grid at a power not exceeding the discharge power limit, so that the electric vehicle cluster becomes an adjustable distributed energy storage resource and participates in the peak-shaving of the virtual power plant.

[0027] Through a unified operational condition priority assessment, external energy conditions, grid load demand, and battery status are coupled for decision-making, intelligently selecting the most pressing issue for intervention based on priority. In cases of severe energy imbalance, priority is given to ensuring local photovoltaic (PV) consumption and charging services; when supply and demand are relatively balanced, a smooth switch to economically oriented grid interaction occurs; advanced vehicle-to-grid (V2G) functionality is only safely and controllably activated when the energy storage is healthy and the grid has a genuine demand. This hierarchical, conditionally triggered logic avoids policy clashes and ensures the system can find the optimal operating point under any conditions. Intelligent policy selection based on operational conditions maximizes energy self-sufficiency when PV is abundant, reducing operating costs; all policy execution is predicated on real-time monitoring of battery status (e.g., V2G activation requires detecting a high SOC threshold), preventing operations that damage battery life, such as high-power discharge, when batteries are in unfavorable conditions (e.g., low SOC, high temperature). This ensures the core health and long-term value of energy storage assets while pursuing economic benefits.

[0028] Step S103: Using a preset fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions.

[0029] The preset fusion control algorithm includes a rule base control module, a proportional-integral-derivative adaptive control module, and a reinforcement learning model module. The rule base control module is configured to have the highest execution priority, triggering and executing the corresponding emergency control strategy when the real-time state parameters meet preset extreme conditions. The proportional-integral-derivative adaptive control module is configured to perform closed-loop regulation of the charging voltage of the sodium-ion battery energy storage system. The reinforcement learning model module is configured to output optimized charging and discharging strategy parameters based on the dynamic changes of the real-time state parameters and the cooperative operating conditions.

[0030] In one embodiment of this specification, the rule base preset, proportional-integral-derivative adaptive control, and reinforcement learning model are not independent or simply connected in series, but rather constitute a layered, integrated, prioritized, and synergistic intelligent control architecture. Specifically, the rule base preset is established based on defined electrochemical knowledge and safety red lines (such as absolute thresholds for temperature and SOC). Its triggering method is event-driven. Once the real-time parameters reach preset extreme conditions (e.g., temperature exceeding 60°C), other control logic is immediately interrupted with the highest priority, and a preset emergency strategy (such as stopping charging) is forcibly executed, thereby providing a basic and insurmountable safety guarantee layer for the system.

[0031] The proportional-integral-derivative adaptive control operates in a continuously enabled mode, constantly making rapid, closed-loop fine adjustments to the key physical quantity of charging voltage to ensure that the voltage remains stable near the target value. When the adjustment accuracy decreases due to system changes (voltage deviation exceeds the threshold), its internal parameter self-tuning mechanism is triggered to dynamically optimize its own control parameters, thereby ensuring the real-time performance and robustness of the underlying control.

[0032] The reinforcement learning model combines periodic decision-making triggering with event-based model optimization triggering. It periodically evaluates the battery health status (temperature, SOC, and level of internal resistance change rate) and outputs long-term optimal charge and discharge strategy parameters (such as current and voltage setpoints) that balance efficiency and lifespan. At the same time, when the system operation data accumulates to a certain extent or the performance indicators show a trend deviation, it will trigger the model to be retrained, so that its strategy can continue to evolve.

[0033] The relationship among these three components can be summarized as follows: the reinforcement learning model formulates long-term optimization strategies within the safety boundaries predefined by the rule base; proportional-integral-derivative (PID) control is responsible for accurately and quickly executing specific voltage commands under this strategy; and the rule base constantly monitors the overall situation, possessing the highest authority to intervene forcefully when dangerous signs appear. This design achieves a seamless integration of strategic intelligence, tactical precision, and safety safeguards, jointly ensuring the efficient, safe, and long-life operation of sodium-ion battery energy storage systems under complex dynamic conditions.

[0034] Using a pre-defined fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation condition. The specific implementation process of proportional-integral-derivative is as follows: The voltage deviation between the real-time charging voltage and the preset target charging voltage of the sodium-ion battery energy storage system's charging circuit is obtained. The real-time charging voltage refers to the instantaneous value obtained by synchronously sampling the battery terminal voltage using a high-precision analog-to-digital converter chip (such as a 24-bit ADS1256). The preset target charging voltage is the expected stable voltage value determined by the upper-level strategy (such as the output of a reinforcement learning model or a nominal value preset according to the battery chemistry system, such as 3.8V). The voltage deviation is the instantaneous algebraic difference between the two (real-time value minus target value). This sampling and calculation process is continuously executed at a frequency of not less than 10Hz to provide a real-time error signal for subsequent control.

[0035] The voltage deviation value is input to the proportional-integral-derivative (PI-DE) control module in the fusion control algorithm. Based on the current proportional, integral, and derivative coefficients, this module calculates the voltage deviation value to generate a first voltage adjustment command, thereby adjusting the output of the charging power supply. The PI-DE control module stores the proportional coefficient (Kp), integral coefficient (Ki), and derivative coefficient (Kd) determined after initialization or the previous adaptive adjustment. Three parallel calculations are performed on the input voltage deviation value: the proportional calculation directly multiplies the current deviation by Kp, generating a fast response component proportional to the instantaneous error; the integral calculation multiplies the cumulative sum of historical deviations by Ki to eliminate steady-state error; and the derivative calculation multiplies the rate of change of the current deviation by Kd to predict error trends and suppress overshoot. These three calculation results are then summed to generate a comprehensive first voltage adjustment command, which is typically a control signal adjusting the pulse width or amplitude. The first voltage adjustment command is sent to the drive circuit of the charging power supply (such as a bidirectional DC / DC converter). By changing the duty cycle of the switching transistor or the reference voltage, the output voltage or current is dynamically adjusted, thereby achieving precise closed-loop control of the battery charging voltage and making it closely follow the preset target.

[0036] To address the performance degradation caused by time-varying system parameters such as battery internal resistance changes and contact resistance fluctuations, an adaptive adjustment mechanism is introduced. When the absolute value of the voltage deviation is detected to continuously exceed a preset voltage deviation threshold, the adaptive adjustment submodule of the proportional-integral-derivative (PID) control module is triggered. This submodule dynamically updates the proportional, integral, and derivative coefficients based on the gradient descent method. The preset voltage deviation threshold is a small value (e.g., 5mV) set according to the control accuracy requirements. The absolute value of the voltage deviation is continuously monitored. Once it exceeds the threshold and persists for a certain period, it indicates that the static error cannot be effectively eliminated or the system response is unsatisfactory. This means the current PID parameters are no longer optimal, triggering the adaptive adjustment submodule. The adaptive adjustment submodule uses the gradient descent method. The principle is to construct a performance index function with Kp, Ki, and Kd as independent variables, typically related to the square of the voltage deviation, and calculate the gradient of this function relative to each coefficient, identifying the direction of the fastest change. Then, along the reverse direction of the gradient, Kp, Ki, and Kd are finely adjusted with a preset learning step size, for example, optimizing Kp within the range of 0.5 ± 0.1. Through multiple iterative adjustments, a new set of control parameters can be automatically found to minimize the sum of squares of voltage deviations under the current system dynamic characteristics. This restores or improves the dynamic performance and steady-state accuracy of the control loop, ensuring that the voltage remains stable within a very small range near the target value (e.g., 3.8V ± 5mV). Thus, a complete adaptive control cycle is achieved, from voltage error sensing and classic PID fast adjustment to online parameter self-tuning.

[0037] In conventional methods, controller parameters (Kp, Ki, Kd) are typically fixed after initial tuning based on typical operating conditions during the system design phase. However, as the internal resistance of the sodium-ion battery gradually increases with the number of cycles, or as the contact resistance of connecting components increases due to aging, the dynamic characteristics of the entire charging circuit change. The original fixed-parameter controller may then respond more slowly, experience increased overshoot, or fail to completely eliminate steady-state errors, leading to charging voltage fluctuations. This not only affects charging efficiency but may also damage battery life due to voltage deviations from the optimal window. This solution introduces an online adaptive adjustment mechanism triggered by voltage deviation exceeding a threshold, endowing the control system with continuous self-learning and self-optimization capabilities. Regardless of the battery system's stage in its life cycle or the slow drift of external circuit parameters, the adaptive mechanism actively identifies control performance degradation and automatically retunes the optimal PID parameters to suit the current system characteristics. This ensures that the charging voltage is precisely controlled within the electrochemically optimal narrow window over a long period, guaranteeing efficient and safe charging and avoiding the risks of overcharging or undercharging due to inaccurate voltage control. Secondly, it reduces the stringent requirements for initial parameter tuning and the dependence on system modeling accuracy, greatly improving the convenience of engineering deployment and the universality of the system in different application scenarios.

[0038] Using a pre-defined fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation condition. The specific implementation process of the reinforcement learning model is as follows: The temperature, state of charge (SOC), and rate of change of internal resistance (RCR) values ​​from the real-time battery state parameters are normalized to obtain the first state feature vector. The temperature value refers to the individual cell temperature collected in real-time by an NTC temperature sensor (such as the MF52-10K type); the SOC value refers to the percentage of remaining battery charge estimated in real-time using a fusion algorithm combining ampere-hour integration and open-circuit voltage calibration; and the RCR value refers to the percentage change (ΔR / R0) of the current DC internal resistance relative to the battery's initial healthy state internal resistance, measured using the AC injection method. Since these three parameters have different physical meanings and dimensions, directly inputting them into the neural network would lead to model training difficulties; therefore, normalization is necessary.

[0039] In one example, based on preset normal ranges for parameters (such as temperature range, SOC range of 0%-100%, and internal resistance change rate range), the original values ​​of each parameter are linearly or non-linearly scaled to a uniform interval (such as [0, 1]). For example, the actual temperature value is mapped to a level representation based on the preset temperature range, the SOC value is divided by 100, and the internal resistance change rate value is divided by the preset maximum change rate threshold. After this processing, the three scalars are combined into a regular, machine-readable first state feature vector, such as a three-dimensional vector [T_norm, SOC_norm, ΔR_R0_norm], which quantitatively characterizes the current health status and operating conditions of the battery.

[0040] In another example, each parameter is mapped to a temperature level, a state of charge level, and an internal resistance change level according to a preset mapping relationship. That is, the first state feature vector input here is [T level, SOC level, ΔR / R0 level]. Table 2 below is an example of a state level division provided by the embodiments of this specification: Table 2. An example of a state level classification provided in the embodiments of this specification. ; The first state feature vector is input into the reinforcement learning decision module of the fusion control algorithm. The deep Q-network model within this module processes the first state feature vector and outputs Q-values ​​corresponding to multiple selectable charging / discharging actions. The reinforcement learning decision module is a deep Q-network model that has undergone extensive offline training and online fine-tuning. It is a deep neural network whose input layer receives the aforementioned first state feature vector. The network contains multiple hidden layers to abstract and mine the complex nonlinear relationships between states and actions layer by layer. The model's output layer corresponds to an action space, which defines all possible charging / discharging actions the system can take in the current state. Each action is typically a set of discrete or parameterized operation instructions, such as {"charging current rate": 1.0C, "discharging current rate": 1.2C, "charging cutoff voltage adjustment": 0mV} or {"charging current rate": 0.8C, "discharging current rate": 1.0C, "charging cutoff voltage adjustment": -5mV}, etc. For each possible action A, the deep Q-network calculates and outputs the corresponding Q-value. The Q-value is an evaluation metric that predicts the expected long-term cumulative reward of performing action A in the current state and subsequently following the optimal strategy. This reward is defined by a predefined reward function that combines immediate gains (such as high charge / discharge efficiency) with long-term costs (such as battery life loss), and its form is, for example, Reward = w1 * (charge / discharge efficiency - baseline efficiency) + w2 * (1 - battery life loss rate), where w1 and w2 are weighting coefficients. The model learns this mapping relationship between state, action, and long-term value through training.

[0041] The target charge / discharge action corresponding to the highest Q value is selected. A second adjustment command is generated based on this target charge / discharge action to update the charge / discharge strategy parameters in the cooperative operation strategy. This target charge / discharge action includes the target charging current rate, the target discharging current rate, and the target cutoff voltage. The learning decision module compares the output Q values ​​corresponding to all available actions and executes the selection logic, choosing the action with the highest Q value as the target charge / discharge action. This selection process reflects an intelligent decision-making process that considers which action is most beneficial (most efficient and best for battery protection) in the long run under the current battery state. The specific parameters included in the target action, such as the target charging current rate, the target discharging current rate, and the target cutoff voltage, are the optimal settings derived from the intelligent decision-making perspective. Subsequently, a second adjustment command is generated based on these settings. This instruction is a series of specific, executable control commands, such as adjusting the current setpoint of the constant current charging stage to X amperes or correcting the charging cutoff voltage to Y volts, in order to update the charging and discharging strategy parameters in the collaborative operation strategy. That is, it covers or corrects the specific execution parameters in the relatively macroscopic power command issued by the upper-level collaborative operation strategy (such as the energy flow allocation strategy), thereby realizing the fine-tuning, adaptive and long-term optimal intelligent fine-tuning of the underlying charging and discharging process within the macroscopic strategy framework.

[0042] Using a pre-defined fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions. The specific implementation method of the pre-defined rules is as follows: The temperature and state of charge (SOC) values ​​from the real-time battery status parameters are compared with multiple extreme condition thresholds stored in the preset rule base of the fusion control algorithm. The temperature value in the real-time battery status parameters refers to the current battery temperature obtained at a high sampling frequency by an NTC temperature sensor (such as the MF52-10K type) installed at the battery cell tabs. The SOC value refers to the percentage of remaining battery charge calculated in real-time by the fusion algorithm of ampere-hour integration and open-circuit voltage calibration. The preset rule base is a data structure stored in the processor's non-volatile memory. Internally, it stores multiple extreme condition thresholds and corresponding emergency strategies predefined by domain knowledge (electrochemical safety boundaries) in the form of condition-action pairs. For example, the first preset temperature threshold corresponds to the critical temperature for battery thermal runaway risk (e.g., 60°C), and the first preset SOC threshold may correspond to the critical charge level for battery over-discharge damage (e.g., 5%). The comparison process is executed cyclically by the processor, logically judging the real-time collected temperature and SOC values ​​against these preset absolute value thresholds in the rule base.

[0043] If the temperature value is not less than a first preset temperature threshold, a first emergency rule is called from the rule base to generate a first emergency command to cut off the charging circuit and activate the cooling system. When the comparison logic determines that the temperature value is not less than the first preset temperature threshold, i.e., the real-time temperature reaches or exceeds a safety upper limit of, for example, 60°C, any other optimization or control algorithms are immediately interrupted, and the first emergency rule bound to this condition is called from the rule base. This rule clearly defines the set of emergency operations that must be performed. Based on this rule, a first emergency command containing specific control commands is generated. This command typically contains two atomic operations: first, cutting off the charging circuit, i.e., physically disconnecting the electrical connection between the battery and the charging power supply by driving switching devices such as relays or power MOSFETs, fundamentally stopping any energy input that may exacerbate the temperature rise; second, activating the cooling system, i.e., sending a start signal to the controller of the cooling fan or liquid pump to force the cooling device to start and accelerate the dissipation of heat from the battery pack. The generation and issuance of this command have the highest priority, aiming to curb the risk of thermal runaway within milliseconds.

[0044] If the state of charge (SOC) value is not greater than the first preset SOC threshold, a second emergency rule is invoked from the rule base to generate a second emergency instruction that limits the discharge power to a percentage of the rated power; the first emergency instruction and / or the second emergency instruction are then executed. When the comparison logic determines that the SOC value is not greater than the first preset SOC threshold, i.e., the real-time SOC reaches or falls below, for example, a deep discharge warning line of 5%, the rule base is also immediately triggered. The corresponding second emergency rule is invoked. Based on this rule, a second emergency instruction designed to protect the battery from deep discharge damage is generated. The core operation of this instruction is to limit the discharge power to a percentage of the rated power; for example, limiting the maximum continuous discharge power allowed by the battery management system (BMS) to a lower percentage (e.g., 30%) of its rated power. This does not completely prohibit discharge (to ensure necessary emergency power supply or safe system shutdown), but rather slows down the voltage drop rate by significantly reducing the discharge current, preventing irreversible capacity loss and structural damage to the battery due to over-discharge.

[0045] Through a pre-defined fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions. This also includes a deterministic empirical strategy execution layer. It is not one of the three methods mentioned above (highest priority rule-based emergency control, real-time fine-tuning PID control, and long-term optimization reinforcement learning), but rather a basic strategy module that runs parallel to and collaborates with them. Typically built based on extensive experimental data and engineering experience, it provides benchmark strategies or recommended parameters for safe and efficient charging and discharging under different battery states. Its decision-making logic is a deterministic lookup table mapping. The connection to the aforementioned three methods is that the reinforcement learning model can reference or optimize the contents of these mapping tables during training and decision-making; the PID controller receives the voltage and current setpoints output by this module or reinforcement learning for precise tracking; and the rule base defines the safety boundaries for the execution of this module (for example, even if the mapping table recommends a certain current, it will be rejected if a high-temperature rule is triggered).

[0046] The temperature and state of charge (SOC) values ​​from the real-time battery status parameters are combined as a joint query index. This index is then used to query a first preset collaborative mapping table to determine the corresponding target charging current, target discharging current, and target charging mode. A third adjustment instruction is then generated to set the charging / discharging current and charging mode. The first preset collaborative mapping table is a strategy table defined based on the intersection of different temperature ranges and different SOC ranges. The real-time temperature and SOC values ​​are converted into corresponding range coordinates according to preset range division rules (e.g., temperature ranges such as T<0℃, 0≤T<25℃, and SOC ranges such as SOC<20%, 20≤SOC<80%). The combination involves pairing temperature range identifiers with SOC range identifiers to form a joint query index such as (low temperature range, low SOC range). This index is used to retrieve a first preset cooperative mapping table stored in memory. The first preset cooperative mapping table is a two-dimensional matrix strategy table, with its rows and columns corresponding to different temperature ranges and state of charge ranges, respectively. Each cell stores the optimal strategy parameter set pre-set for that specific "temperature-SOC" combination. Table 3 shows the first preset cooperative mapping table provided in the embodiments of this specification.

[0047] Table 3 First Preset Cooperative Mapping Table ; Through the aforementioned joint query, the target charging current value (e.g., a value expressed as a multiplier C), the target discharging current value, and the target charging mode (e.g., constant current charging, constant current constant voltage charging, constant voltage charging) are read from the corresponding cell in the mapping table. Subsequently, a third adjustment instruction is generated. For example, the target charging current value (e.g., "1.0C") is converted into a specific ampere value and sent to the charger, the discharging current limit is set to the BMS, and an instruction is sent to the charger to switch to the specified charging mode. This instruction is then executed, thereby setting the charging and discharging current and charging mode, achieving the first fine-tuning of the charging and discharging strategy with the real-time battery status.

[0048] The preset rate of change interval to which the internal resistance change rate value in the real-time battery state parameters belongs is used as a query index to query the second preset compensation mapping table, determine the corresponding target charging cut-off voltage value and target discharging cut-off voltage value, and generate a fourth adjustment command to set the charging and discharging cut-off voltage. The second preset compensation mapping table defines the direct correspondence between the internal resistance change rate interval and the target charging cut-off voltage value and target discharging cut-off voltage value. The internal resistance change rate value (ΔR / R0) is calculated by measuring the current DC internal resistance using the 1kHz AC injection method and comparing it with the internal resistance value at the battery's factory or initial health state. This value is categorized into a preset rate of change interval (e.g., <5%, 5%~10%, etc.). This interval identifier serves as a query index for retrieving independent second preset compensation mapping tables. Table 4 shows a second preset compensation mapping table provided in the embodiments of this specification, which defines the voltage parameter adjustment strategies required for the battery at different aging stages (characterized by the rate of internal resistance increase).

[0049] Table 4 Second Preset Compensation Mapping Table ; The mapping table is consulted to obtain the target charging cut-off voltage and target discharging cut-off voltage values ​​bound to the current internal resistance change rate range. For example, for a change rate range of 5% to 10%, the target values ​​might be 3.795V and 2.005V. These values ​​are absolute voltage points pre-calculated based on aging characteristics. Next, a fourth adjustment command is generated, containing the two target voltage values, and sent to the charging power supply (setting its upper charging voltage limit) and the battery management system's discharge protection module (setting its lower discharging voltage limit), respectively. This sets the charging and discharging cut-off voltages, completing software compensation for the ohmic voltage drop caused by increased internal resistance due to battery aging, ensuring the battery can be effectively charged and over-discharged throughout its entire lifespan, maintaining its usable capacity. The entire process, through two efficient table lookup operations, achieves rapid, reliable, and experience-driven dynamic adaptation of the three core strategy parameters: current, mode, and voltage.

[0050] In addition to the dynamic adaptation methods mentioned above, it can also be achieved in the following ways: Based on the output of the decision-making layer, deterministic adjustment rules based on multi-dimensional battery state parameters are executed. These rules respond quickly through table lookup or logical judgment, transforming macroscopic collaborative operation strategies into microscopic control commands that precisely match the real-time electrochemical characteristics of the battery, thereby ensuring the safety and efficiency of the charging and discharging process and protecting battery life. This dynamic adjustment mainly revolves around three key dimensions: temperature, state of charge (SOC), and rate of change of internal resistance. The specific rules are as follows: The temperature adaptation rules are designed to address the significant impact of temperature on the electrochemical activity, safety margins, and lifespan of sodium-ion batteries. The system performs graded current control and low-temperature protection based on real-time battery temperature (T). When the battery is detected to be in a low-temperature environment (T < 0°C), to suppress the risk of sodium dendrite growth, the upper limit of the charging current is set to a relatively low range of 0.6C to 0.8C, and the battery preheating function is automatically activated simultaneously, providing gentle heating with a heating power of no more than 5W until the battery temperature rises above 5°C. In the normal temperature range (0 ≤ T < 25°C), battery activity recovers, and the charging current is set to 0.8C and the discharging current to 1.0C to achieve balanced performance output. In the optimal electrochemical temperature range (25 ≤ T < 45°C), a more aggressive strategy is adopted, setting the charging current in the range of 1.0C to 1.2C and the discharging current to 1.2C to maximize charge / discharge efficiency and power capability. When the temperature rises to a higher range (45≤T<60℃), in order to prevent accelerated degradation and thermal risks due to high temperature, the charging current is actively limited to 0.7C and the discharging current is limited to 0.9C.

[0051] The SOC adaptation rule optimizes the charging stage and depth of discharge protection based on the battery's real-time remaining capacity (SOC). In the low capacity range (SOC < 20%), the system adopts a simple constant current charging mode, using a 1.0C current for rapid energy replenishment, while strictly setting the discharge cutoff voltage at 2.0V ± 5mV to avoid over-discharge. In the main capacity range (20 ≤ SOC < 80%), a more refined two-stage constant current and constant voltage charging strategy is adopted: first, constant current charging at 1.0C is used until the voltage inflection point, then it automatically switches to 0.5C constant voltage charging until the current drops to the cutoff condition. The discharge cutoff voltage in this stage is also maintained at 2.0V ± 5mV. When entering the high capacity range (SOC ≥ 80%), to alleviate the structural stress of the positive electrode material at high potential, the system switches to a constant voltage charging mode, stabilizing the charging voltage at 3.8V ± 5mV. At the same time, to protect the battery, the upper limit of the discharge current in this range is limited to 1.0C.

[0052] Internal resistance matching rules are used to compensate for changes in internal parameters caused by aging during the battery's lifespan. The rate of change (ΔR / R0) of the DC internal resistance relative to the initial healthy state is monitored in real time using an AC injection method. Based on this, the charge and discharge cut-off voltages are dynamically fine-tuned to offset the ohmic voltage drop caused by increased internal resistance, ensuring the actual charge and discharge capacity of the battery. This is executed through a preset lookup table. When the rate of change in internal resistance is less than 5%, the battery is considered to be in good condition, and no voltage adjustment is performed. When the rate of change is between 5% and 10%, the system lowers the charging cut-off voltage by 5mV and simultaneously raises the discharging cut-off voltage by 5mV. When the rate of change reaches 10% to 15%, the reduction in charging cut-off voltage increases to 10mV, and the increase in discharging cut-off voltage increases by 10mV simultaneously. When the rate of change in internal resistance reaches or exceeds 15%, indicating accelerated battery aging, the system performs the maximum compensation, lowering the charging cut-off voltage by 15mV and raising the discharging cut-off voltage by 15mV.

[0053] The aforementioned temperature adaptation, SOC adaptation, and internal resistance adaptation rules together constitute a multi-dimensional, dynamically complementary strategy parameter adjustment system. Based on the real-time collected temperature, SOC, and internal resistance change rate data, the system comprehensively utilizes these data to collaboratively generate the final current, voltage, and mode control commands. This enables the charging and discharging strategy to respond to the battery's transient state and long-term aging trend in real time and accurately, fundamentally overcoming the shortcomings of traditional static strategies. This provides a concrete and effective technical guarantee for achieving efficient, safe, and long-life operation of sodium-ion battery energy storage systems across all operating conditions and throughout their entire life cycle.

[0054] In one embodiment of this specification, a feedback layer closed-loop optimization process is also included, which is an intelligent optimization based on performance deviation detection and trigger-based re-optimization. Specifically, the stability of core state parameters is continuously monitored. When the deviation between the real-time estimated value of state of charge (SOC) and the target value based on the model or historical data continuously exceeds a preset range, or the battery temperature fluctuates significantly beyond the normal range in a short period of time, or the measured DC internal resistance value deviates significantly from the predicted value based on the aging model, it is determined that the current operating strategy is mismatched with the actual battery state or external environment, thereby automatically triggering the closed-loop optimization process. For example, the triggering conditions can be set as SOC deviation > 3%, temperature fluctuation > 2℃, and internal resistance deviation > 5mΩ. In addition, to ensure long-term adaptability, a fixed-cycle optimization trigger point based on running time is also provided. For example, after completing a certain number of complete charge-discharge cycles, regardless of whether the parameter deviation exceeds the limit, a comprehensive strategy review and optimization will be forcibly started. This can be 1 to 3 charge-discharge cycles, and the time for a single optimization is ≤100ms. Once optimization is triggered, the system will reassemble the latest multi-source parameters and drive the fuzzy logic model and fusion algorithm (including reinforcement learning model) of the decision layer to perform collaborative calculations again. This secondary optimization process does not start from scratch, but rather fine-tunes and optimizes the previous round of strategy parameters by incorporating new state information, thereby quickly generating new strategy parameters that are more suitable for the current and next stage of operating conditions. The optimization goal is to ensure a charge-discharge efficiency of ≥90% and a cycle life of ≥1700 cycles (1C / 1C cycle, capacity retention ≥80%). The entire feedback optimization process is carefully designed, and the time consumed by each calculation is strictly controlled to ensure the real-time response of the system and avoid the impact of delays introduced by optimization calculations on control performance under dynamic operating conditions. Through this low-latency closed-loop feedback triggered by both events (deviation exceeding limits) and time (cycle arrival), it possesses powerful online learning and self-correction capabilities, effectively coping with battery aging, environmental temperature changes caused by seasonal changes, and the long-term evolution of photovoltaic output and charging load patterns, ensuring that the photovoltaic-storage-charging collaborative system maintains efficient, safe, and robust operation throughout its entire life cycle.

[0055] The above-described at least one technical solution adopted in the embodiments of this specification can achieve the following beneficial effects: At the perception layer, battery state parameters, external energy parameters, and charging demand parameters are collected simultaneously to form a multi-dimensional state space covering all elements. This breaks through the limitations of traditional control that relies on only a single variable (such as only SOC or only grid electricity price), and can understand the comprehensive characteristics of the current operating environment from a global perspective. At the decision-making layer, a preset fuzzy logic model is introduced to classify and prioritize the current collaborative operating conditions, which is in line with the real-world scenarios of strong fluctuations in new energy output and high randomness in user charging behavior. Different priorities correspond to different control objectives, thereby avoiding strategy confusion under multiple conflicting objectives. At the execution layer, energy flow allocation strategies, grid interaction strategies, and vehicle-grid interaction strategies generated based on operating condition priorities achieve multi-level and multi-dimensional collaborative control. The energy flow allocation strategy ensures that local renewable energy (such as photovoltaics) is preferentially consumed locally, reducing the impact on the grid. The grid interaction strategy utilizes the fast response characteristics of sodium-ion batteries to discharge and support loads during peak grid periods and charge and store energy during off-peak periods, playing a role in peak shaving and valley filling. The vehicle-grid interaction strategy further treats electric vehicles as mobile energy storage units, allowing them to supply power to the grid or energy storage system in reverse under specific operating conditions, greatly improving system flexibility. At the optimization layer, the strategy parameters are dynamically adjusted through a pre-defined fusion control algorithm, giving the system continuous learning and adaptive capabilities. This achieves deep collaboration between the sodium-ion battery energy storage system, charging facilities, and external energy sources, significantly improving energy utilization efficiency, system operational safety, and grid support capabilities, thus meeting the requirements for large-scale reliable application of sodium-ion batteries in complex dynamic scenarios.

[0056] To verify the technical effects of the embodiments in this specification, a control experiment was designed for verification. The experiment used sodium-ion battery cells with clearly defined specifications and standardized testing equipment to ensure the comparability of experimental conditions and the reliability of results. The core of the experiment lies in setting up a control group and an experimental group for parallel testing. The control group adopts the traditional static strategy described in the background art, where parameters such as charging and discharging current and voltage remain fixed throughout the test process, without adjustment based on changes in battery state or external simulated operating conditions. The experimental group, on the other hand, applies the adaptive dynamic adjustment method proposed in this invention, fully deploying a complete closed-loop process including multi-source parameter sensing, fuzzy logic operating condition classification, collaborative strategy formulation, and dynamic adjustment of the fusion algorithm.

[0057] Specifically, the experiment focused on the Na3V2(PO4)3 / C sodium-ion battery cell series, with a nominal capacity of 200Ah and a rated voltage of 3.8V. This material system is one of the mainstream technologies for sodium-ion batteries and is representative. The testing hardware platform included a Neware CT-4008 charge-discharge tester for accurately applying current-voltage curves, a high-low temperature chamber capable of simulating any static or dynamic temperature environment within the range of -40℃ to 150℃, and an internal resistance tester for accurately measuring changes in the battery's internal state. To ensure fair comparison, the experiment established two groups: the control group adopted a traditional static strategy, with a fixed current rate of 1.0C during both charging and discharging, a fixed charging cutoff voltage of 3.8V, and a fixed discharging cutoff voltage of 2.0V; the experimental group fully applied the adaptive dynamic adjustment method of this invention, including multi-source parameter sensing, fuzzy logic operating condition classification, collaborative operation strategy formulation, and dynamic adjustment of the fusion control algorithm.

[0058] Figure 2 A charge-discharge efficiency comparison experimental curve is provided for the embodiments of this specification, such as... Figure 2 As shown, regarding charge and discharge efficiency, by conducting multiple complete cycle tests on both groups of batteries and calculating the average efficiency, the experimental group achieved an average charge and discharge efficiency of 91.2%, while the control group's average efficiency was 85.0%, representing a significant improvement of 6.2 percentage points. This result directly proves that the embodiments in this specification effectively reduce energy loss during charging and discharging through a dynamic adaptation strategy. For example, it actively reduces the current to reduce ohmic heat loss when the battery's internal resistance increases due to temperature rise, or switches to a constant voltage mode at the end of charging to improve charging efficiency.

[0059] Figure 3 This is a cycle life comparison test curve provided for an embodiment of this specification. For example... Figure 3 As shown, in the cycle life test, with the capacity retention rate dropping to 80% of the initial capacity as the end point, the experimental group battery only reached 80% capacity retention after 1725 standard cycles (1C charge / 1C discharge), while the control group battery reached the end point after 1500 cycles, extending the cycle life of the experimental group by 15%. This data strongly confirms the superior effect of the adaptive adjustment method in mitigating battery degradation and extending service life. This effect stems from the dynamic adjustment method, which limits the current and enables preheating at low temperatures to avoid the risk of dendrite growth; uses gentle constant voltage charging in the high SOC range to reduce the structural stress of the cathode material; and dynamically compensates the cutoff voltage according to the change in internal resistance to ensure that the battery can still be reasonably charged and discharged during aging, slowing down the rate of capacity degradation.

[0060] Furthermore, regarding low-temperature performance, a full charge test from 0% to 100% SOC was conducted in a harsh -10℃ environment. The experimental group completed charging in 180 minutes, while the control group required 240 minutes, representing a 25% reduction in charging time. This highlights the effective improvement in charging speed under low-temperature conditions achieved by the embodiments in this specification through strategy adjustments (such as current management and preheating). In summary, the above experimental data demonstrates that, through algorithmic innovation, without increasing hardware costs, the overall performance of the sodium-ion battery energy storage charging system in terms of efficiency, lifespan, and environmental adaptability has been significantly improved, validating its remarkable technological advancement and engineering application value.

[0061] To ensure the feasibility of complex algorithms in practical engineering systems, the execution time of the core algorithms was rigorously tested. Under typical operating loads, the processing time of the fuzzy logic collaborative operating condition analysis module did not exceed 20 milliseconds; the core calculation time of the strategy optimization of the fusion control algorithm (including PID adaptive, reinforcement learning decision-making, etc.) did not exceed 50 milliseconds; and the total time of the entire closed-loop feedback process (including trigger judgment, data preparation, optimization execution, and instruction update) did not exceed 30 milliseconds. In summary, the total time for completing a complete closed-loop action from perception to adjustment—including perception, decision-making, execution, and feedback—was strictly controlled within 100 milliseconds. This fully meets the real-time control requirements of high-dynamic energy systems, proving that the technical solution of this invention not only has superior performance but also possesses engineering practicality for real-time operation on actual hardware platforms (such as the STM32H750 main control MCU).

[0062] This specification also provides an adjustment device for the coordinated operation of sodium-ion battery energy storage and charging, such as... Figure 4 As shown, the device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described method.

[0063] This specification also provides a non-volatile computer storage medium storing computer-executable instructions configured to perform the above-described method.

[0064] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the embodiments of apparatus, devices, and non-volatile computer storage media are basically similar to the method embodiments, so the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0065] The above description is merely one or more embodiments of this specification and is not intended to limit this specification. Various modifications and variations can be made to the one or more embodiments of this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of one or more embodiments of this specification should be included within the scope of the claims of this specification.

Claims

1. A method for adjusting the coordinated operation of sodium-ion battery energy storage and charging, characterized in that, The method includes: Real-time battery status parameters of the sodium-ion battery energy storage system, real-time external energy parameters of the external energy source, and real-time charging demand parameters of the charging facility are acquired. Based on the real-time battery status parameters, the real-time charging demand parameters, and the real-time external energy parameters, the current cooperative operating conditions are classified and the operating conditions are determined by a preset fuzzy logic model. Based on the operating condition priority, a coordinated operation strategy is generated and executed for the sodium-ion battery energy storage system, the charging facility and the external energy source, wherein the coordinated operation strategy includes an energy flow allocation strategy, a grid interaction strategy and a vehicle-grid interaction strategy. By using a preset fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions.

2. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 1, characterized in that, Based on the real-time battery status parameters, the real-time charging demand parameters, and the real-time external energy parameters, the current collaborative operating conditions are classified and the operating conditions are prioritized using a preset fuzzy logic model. Specifically, this includes: Calculate the first ratio of the photovoltaic output power in the real-time external energy parameters to the average charging power in the real-time charging demand parameters, and calculate the second ratio of the real-time grid load to the preset grid rated capacity in the real-time external energy parameters; The first ratio, the second ratio, and the state of charge value in the real-time battery state parameters are input into the preset fuzzy logic model, so that the fuzzification processing unit in the preset fuzzy logic model can determine the first membership value set of the first ratio, the state of charge value, and the second ratio relative to a preset plurality of membership functions. Through the fuzzy inference unit in the preset fuzzy logic model, inference calculation is performed based on the preset fuzzy rule base and the first membership value set to obtain multiple second membership values ​​that indicate that the current collaborative operation condition belongs to multiple preset operation condition levels. The defuzzification processing unit in the preset fuzzy logic model processes the multiple second membership values ​​to determine the target operating condition level to which the current collaborative operating condition belongs. According to the preset mapping relationship between working condition level and priority, the target working condition level is mapped to the corresponding working condition priority.

3. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 1, characterized in that, Based on the aforementioned operating condition priorities, a coordinated operation strategy is generated and executed for the sodium-ion battery energy storage system, the charging facility, and the external energy source, specifically including: In response to the operating condition priority being a preset first priority range, the current operating condition is determined to be an energy supply and demand imbalance condition, and the energy flow allocation strategy is generated and executed. In response to the operating condition priority indication being within a preset second priority range, the current operating condition is determined to be a grid load regulation timing condition, and the grid interaction strategy is generated and executed. In response to the operating condition priority indication being within the preset third priority range and the state of charge in the real-time battery status parameters reaching a preset high threshold, it is determined that the vehicle-to-grid interaction conditions are met, and the vehicle-to-grid interaction strategy is generated and executed.

4. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 1, characterized in that, The preset fusion control algorithm includes a rule base control module, a proportional-integral-derivative adaptive control module, and a reinforcement learning model module. The rule base control module is configured to have the highest execution priority and is used to trigger and execute the corresponding emergency control strategy when the real-time status parameters meet preset extreme conditions. The proportional-integral-derivative adaptive control module is configured to perform closed-loop regulation of the charging voltage of the sodium-ion battery energy storage system. The reinforcement learning model module is configured to output optimized charging and discharging strategy parameters based on the dynamic changes of the real-time state parameters and the cooperative operating conditions.

5. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 4, characterized in that, Through a preset fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions. Specifically, this includes: The temperature value, state of charge value, and rate of change of internal resistance value in the real-time battery state parameters are normalized to obtain the first state feature vector; The first state feature vector is input to the reinforcement learning decision module in the fusion control algorithm, so that the first state feature vector is processed by the deep Q network model in the reinforcement learning decision module, and the Q value corresponding to multiple optional charging and discharging actions is output. Select the target charge / discharge action corresponding to the highest Q value, and generate a second adjustment instruction based on the target charge / discharge action to update the charge / discharge strategy parameters in the cooperative operation strategy. The target charge / discharge action includes the target charging current rate, the target discharging current rate, and the target cutoff voltage.

6. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 4, characterized in that, Through a preset fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions. Specifically, this includes: The temperature value and state of charge value in the real-time battery state parameters are compared with multiple extreme operating condition thresholds stored in the preset rule base of the fusion control algorithm. If the temperature value is not less than the first preset temperature threshold, then the first emergency rule is called from the rule base to generate a first emergency command to cut off the charging circuit and start the heat dissipation system. If the state of charge value is not greater than the first preset state of charge threshold, then the second emergency rule is called from the rule base to generate a second emergency instruction that limits the discharge power to a ratio of the rated power. Execute the first emergency command and / or the second emergency command.

7. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 1, characterized in that, Through a preset fusion control algorithm, the strategy parameters in the collaborative operation strategy are dynamically adjusted based on the real-time state parameters of the sodium-ion battery energy storage system and the dynamic changes in the collaborative operation conditions. Specifically, this includes: Obtain the voltage deviation between the real-time charging voltage and the preset target charging voltage of the charging circuit of the sodium-ion battery energy storage system; The voltage deviation value is input to the proportional-integral-derivative (PI-DE) control module in the fusion control algorithm. The PI-DE control module calculates the voltage deviation value based on the current proportional coefficient, integral coefficient, and derivative coefficient, and generates a first voltage adjustment command to adjust the output of the charging power supply. When the absolute value of the voltage deviation is detected to continuously exceed the preset voltage deviation threshold, the adaptive adjustment submodule of the proportional-integral-derivative control module is triggered to dynamically update the proportional coefficient, integral coefficient, and derivative coefficient based on the gradient descent method.

8. The adjustment method for coordinated operation of sodium-ion battery energy storage and charging according to claim 1, characterized in that, The energy flow allocation strategy includes: comparing the photovoltaic output power with the average charging power; if the photovoltaic output power is greater than the average charging power, generating a first charging command to control the sodium-ion battery energy storage system to charge; if the photovoltaic output power is less than the average charging power, generating a first discharging command to control the sodium-ion battery energy storage system to discharge. The grid interaction strategy includes: determining whether the current time period is a preset grid peak period; if so, generating a second discharge command to prioritize controlling the sodium-ion battery energy storage system to supply power to the charging facility; if not, generating a second charging command to control the sodium-ion battery energy storage system to charge from the grid. The vehicle-to-grid interaction strategy includes generating a discharge enable command to allow new energy vehicles connected to the charging facility to discharge to the grid according to a preset discharge power limit.

9. An adjustment device for coordinated operation of sodium-ion battery energy storage and charging, characterized in that, The device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method as described in any one of claims 1-8.

10. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer-executable instructions are configured to perform the method as described in any one of claims 1-8.

Images