Energy intelligent scheduling method and system for a flow battery

Abstract

The application discloses an energy intelligent scheduling method and system of a flow battery, and relates to the technical field of flow batteries.The method comprises the following steps: collecting flow battery energy operation parameters in real time, dynamically estimating electrolyte states to obtain a plurality of energy state information; setting a to-be-scheduled target according to the energy state information, and formulating a fuzzy control strategy through retrograde multi-target optimization; executing the fuzzy control strategy to generate a set of regulation and control signals and to schedule; and finally, performing reverse tracing optimization on the fuzzy control strategy according to the scheduling result, updating the set of regulation and control signals, and obtaining a plurality of regulation and control instructions to intelligently schedule the energy of the flow battery.The application solves the technical problems that the traditional flow battery scheduling method is difficult to accurately capture the energy state in real time, the scheduling strategy has low matching degree with actual demand, and there is large power loss and fast battery life attenuation, and achieves the technical effects of flow battery energy intelligent scheduling, reduction of power loss, prolongation of battery life, and improvement of the matching degree of scheduling and actual demand.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of liquid flow battery, and particularly relates to an energy intelligent scheduling method and system of liquid flow battery. BACKGROUND

[0002] Liquid flow battery which converts chemical energy into electrical energy is widely used in the field of energy storage, and its energy scheduling efficiency is crucial to system performance. In the prior art, the energy scheduling of liquid flow battery depends on fixed strategy or simple parameter feedback, which can play a certain role in stable working conditions, but as the application scenarios become more complex, the traditional scheduling method gradually shows limitations. Due to the dynamic changes of electrolyte state of liquid flow battery, involving concentration, temperature, flow and other multi-parameter coupling, the traditional method is difficult to accurately capture the energy state in real time, resulting in low matching degree of scheduling strategy and actual demand, and easy to cause large power loss, fast battery life decay and other problems, which is difficult to meet the efficient and long-term energy scheduling demand. SUMMARY

[0003] The present application provides an energy intelligent scheduling method and system of liquid flow battery, which is used to solve the technical problems that the traditional scheduling method of liquid flow battery is difficult to accurately capture the energy state in real time, resulting in low matching degree of scheduling strategy and actual demand, and large power loss and fast battery life decay.

[0004] In a first aspect, the present application provides an energy intelligent scheduling method of liquid flow battery, which comprises: collecting energy operation parameters of liquid flow battery in real time, dynamically estimating electrolyte state according to the energy operation parameters to obtain a plurality of energy state information; scheduling analysis of liquid flow battery based on the plurality of energy state information, setting a plurality of to-be-scheduled targets, multi-objective optimization according to the plurality of to-be-scheduled targets, and formulating a fuzzy control strategy according to the optimization result; dynamically updating the plurality of to-be-scheduled targets by executing the fuzzy control strategy, generating a set of control signals, bidirectional collaborative scheduling of liquid flow battery based on the set of control signals, generating a scheduling result; reverse tracing according to the scheduling result, incremental learning of the fuzzy control strategy according to the tracing path, generating a fuzzy control optimization strategy to replace the fuzzy control strategy, updating the set of control signals, obtaining a plurality of levels of control instructions to intelligently schedule the energy of liquid flow battery.

[0005] In a second aspect of the present application, an energy intelligent scheduling system for a flow battery is provided, which comprises: an energy state information acquisition module, configured to collect energy operation parameters of the flow battery in real time, dynamically estimate the state of electrolyte based on the energy operation parameters, and obtain a plurality of energy state information; a fuzzy control strategy acquisition module, configured to perform scheduling analysis on the flow battery based on the plurality of energy state information, set a plurality of to-be-scheduled targets, perform multi-objective optimization based on the plurality of to-be-scheduled targets, and formulate a fuzzy control strategy based on the optimization result; a scheduling result acquisition module, configured to perform dynamic updating on the plurality of to-be-scheduled targets by executing the fuzzy control strategy, generate a set of regulation and control signals, perform bidirectional collaborative scheduling on the flow battery based on the set of regulation and control signals, and generate a scheduling result; and an energy intelligent scheduling execution module, configured to perform reverse tracing according to the scheduling result, perform incremental learning on the fuzzy control strategy based on a tracing path, generate a fuzzy control optimization strategy to replace the fuzzy control strategy, update the set of regulation and control signals, and obtain a plurality of levels of regulation and control instructions to intelligently schedule the energy of the flow battery.

[0006] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0007] In the present application, the energy operation parameters of the flow battery are collected in real time, a plurality of energy state information is obtained through dynamic estimation, the to-be-scheduled targets are set based on the information, multi-objective optimization is performed, a fuzzy control strategy is formulated, the set of regulation and control signals is generated by executing the strategy, bidirectional collaborative scheduling is performed, the strategy is optimized through reverse tracing, the set of regulation and control signals is updated, and the plurality of levels of regulation and control instructions are obtained, so as to intelligently schedule the energy of the flow battery, make the scheduling more efficient and accurate, and achieve the technical effects of energy intelligent scheduling of the flow battery, reduction of power loss, prolongation of battery life, and improvement of the matching degree of scheduling and actual demand. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0009] Figure 1 is a flowchart of an energy intelligent scheduling method for a flow battery provided by an embodiment of the present application.

[0010] Figure 2 is a structural schematic diagram of an energy intelligent scheduling system for a flow battery provided by an embodiment of the present application.

[0011] Reference signs: energy state information acquisition module 1, fuzzy control strategy acquisition module 2, scheduling result acquisition module 3, energy intelligent scheduling execution module 4. DETAILED DESCRIPTION

[0012] The application provides an energy intelligent scheduling method and system for a flow battery, which is used to solve the technical problems that the traditional flow battery scheduling method is difficult to accurately capture the energy state in real time, the scheduling strategy and the actual demand have low matching degree, and there is large power loss and fast battery life attenuation.

[0013] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0014] It should be noted that the terms "first", "second" and the like in the specification and the above drawings of the application are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices.

[0015] Embodiment one, as shown in a kind of energy intelligent scheduling method for flow battery, wherein the method comprises: Figure 1

[0016] Step A100: real-time acquisition of energy operating parameters of flow battery, dynamic estimation of electrolyte state according to the energy operating parameters, to obtain a plurality of energy state information.

[0017] In the embodiments of the application, the flow battery is a battery device that realizes energy conversion and storage through the state change (such as the change of parameters such as flow, concentration and temperature) of electrolyte, which needs energy intelligent scheduling to optimize power loss and battery life. The energy operating parameters are the object of real-time acquisition and analysis, and are also the main body of the energy intelligent scheduling method. The energy state information includes state of charge (SOC) and state of health (SOH) of electrolyte.

[0018] ​Specifically, first, the energy operation parameters of the flow battery are collected in real time, the core of which is to obtain the key data reflecting the running state of the flow battery by deploying high-precision sensors to monitor in real time: for the electrolyte flow parameter, an electromagnetic flowmeter or a turbine flowmeter can be used, which is installed on the electrolyte circulating main pipeline and branch pipeline of the flow battery. By sensing the electromagnetic signal or the turbine rotation frequency generated when the electrolyte flows, the circulating flow of the electrolyte per unit time is monitored in real time; for the electrolyte concentration parameter, a concentration sensor based on ultraviolet-visible spectrum analysis can be used, the probe of which is inserted into the electrolyte flow channel of the electrolyte storage tank and the inlet and outlet of the battery stack. By detecting the absorption intensity of the electrolyte to the specific wavelength of light, the concentration data of the active substance in the electrolyte are obtained in real time; for the electrolyte temperature parameter, a platinum resistance temperature sensor or a thermocouple sensor can be used, which is closely attached to the inner wall of the pipeline through which the electrolyte flows or embedded in the electrolyte channel of the battery stack, directly contacting the electrolyte to sense the temperature change and collect the temperature information of the electrolyte in real time. The above sensors are connected to the system through data transmission lines to continuously output the collected parameter data at a sampling frequency of milliseconds or seconds, realizing real-time acquisition of the energy operation parameters.

[0019] Then, the energy operation parameters of the flow battery are analyzed, and an electrolyte state coupling model is constructed through concentration distribution calculation and extreme value correlation analysis, etc. The electrolyte charge and health state information are decoupled and calculated and added to the energy state information. The specific steps are described in detail in A110-A180.

[0020] Step A200: scheduling analysis of the flow battery based on the plurality of energy state information, setting a plurality of to-be-scheduled targets, multi-objective optimization according to the plurality of to-be-scheduled targets, and formulating a fuzzy control strategy according to the optimization result.

[0021] In the embodiments of the present application, the fuzzy control strategy is a strategy formulated based on the plurality of energy state information of the flow battery by setting a plurality of to-be-scheduled targets and performing multi-objective optimization.

[0022] Optionally, the power loss and battery life targets are set based on the energy state information, mapped to a unified decision space, combined with the running interval and decay rate constraints to calculate the target constraint satisfaction degree and sort, and then the fuzzy control strategy is optimized and formulated. The specific steps are described in detail in A210-A260.

[0023] Step A300: performing dynamic update of the plurality of to-be-scheduled targets according to the fuzzy control strategy, generating a set of control signals, and performing bidirectional collaborative scheduling of the flow battery based on the set of control signals to generate a scheduling result.

[0024] In the embodiments of the present application, the set of control signals includes real-time charging and discharging power signals of the flow battery and electrolyte flow control signals.

[0025] In an embodiment of the present application, a fuzzy control strategy is executed, a target priority weight coefficient is dynamically updated, a real-time charging and discharging power signal is generated, parameters are obtained through electric-hydraulic bidirectional timing collaborative optimization, and a scheduling result is dynamically scheduled accordingly. The specific steps are described in detail in A310-A350.

[0026] Step A400: According to the scheduling result, backtracking is performed, incremental learning of the fuzzy control strategy is performed according to the backtracking path, a fuzzy control optimization strategy is generated to replace the fuzzy control strategy, the set of control signals is updated, and multi-level control instructions are obtained to intelligently schedule the energy of the flow battery.

[0027] Specifically, first, a historical period is set according to the scheduling result and execution data is collected, a parameter is analyzed based on a deviation, a defect is located based on a backtracking path, an incremental learning of the fuzzy control strategy is performed based on the defect to generate a fuzzy control optimization strategy, and the specific steps are described in detail in A410-A440.

[0028] Then, after verifying whether the fuzzy control optimization strategy can eliminate the execution deviation, the original strategy is replaced and a signal cascade reconstruction space is constructed, a set of control signals is reconstructed, multi-level control instructions are generated to match the scheduling level, and energy is intelligently scheduled, and the specific steps are described in detail in A450-A480.

[0029] Further, the step A100 in the method provided by the embodiment of the present application comprises:

[0030] A110: The energy operation parameters are analyzed to obtain electrolyte flow parameters, electrolyte concentration parameters, and electrolyte temperature parameters of the flow battery.

[0031] A120: Based on the electrolyte concentration parameters, concentration distribution calculation is performed on the flow battery to generate electrolyte concentration distribution data, and extreme value correlation analysis is performed according to the electrolyte concentration distribution data to obtain a dynamic correlation relationship of concentration polarization.

[0032] A130: According to the dynamic correlation relationship of concentration polarization, the electrolyte flow parameters are mapped to an equivalent circuit to obtain a variable resistance parameter.

[0033] A140: Based on the variable resistance parameter, the open circuit voltage value of the flow battery is retrieved, the electrolyte concentration distribution data is associated with the open circuit voltage value, and a concentration-voltage compensation coefficient table is constructed.

[0034] A150: The electrolyte temperature parameters are nonlinearly corrected and fused with the concentration-voltage compensation coefficient table to construct an electrolyte state coupling model.

[0035] A160: performing dynamic decoupling calculation of electrolyte charging through the electrolyte state coupling model to obtain electrolyte charging state information.

[0036] A170: performing dynamic decoupling calculation of electrolyte health through the electrolyte state coupling model to obtain health state information.

[0037] A180: adding the electrolyte charging state information and the health state information to the plurality of energy state information.

[0038] In the embodiments of the present application, the electrolyte charging state information directly reflects the remaining proportion of active substances that can currently participate in energy conversion, and is a key basis for judging the real-time energy storage capacity of the battery and determining the charging and discharging demand. The health state information reflects the long-term performance of the electrolyte, including the degree of active substance loss and the stability of the concentration, and is an important reference for evaluating the battery life attenuation trend and formulating a long-term scheduling strategy.

[0039] Specifically, after collecting the energy operation parameters of the flow battery in real time, first, these parameters are analyzed to extract key indicators that can reflect the core state of the electrolyte, including electrolyte flow parameters, electrolyte concentration parameters and electrolyte temperature parameters. For example, assuming that the energy operation parameters collected contain the circulation amount of electrolyte in the battery circulation system per unit time, the electrolyte flow parameter is obtained after analysis; the active substance molar concentration obtained by the concentration sensor is analyzed to obtain the electrolyte concentration parameter; the electrolyte temperature parameter is obtained by real-time monitoring of the temperature sensor.

[0040] Next, when calculating the concentration distribution based on the analyzed electrolyte concentration parameter, the internal structure of the flow battery, such as the positive cavity, the negative cavity, the electrolyte flow channel, the liquid storage tank and other regions, needs to be combined with the computational fluid dynamics and mass transfer coupling model. First, the internal space of the battery is divided into grids, and the positive reaction area, the negative reaction area, the inlet and outlet flow channel and the like are divided into a plurality of calculation units; then, the initial electrolyte concentration parameter obtained by analysis is input, for example, the electrolyte concentration in the liquid storage tank is 2.8 mol / L, and the electrolyte flow rate, diffusion coefficient, electrode reaction rate and other parameters are introduced, the concentration change in each unit is calculated by numerical iteration, for example, the positive area consumes active substances due to oxidation reaction, the concentration gradually decreases with the reaction, while the negative area generates active substances due to reduction reaction, the concentration gradually increases, and finally the concentration distribution data covering all areas of the battery is generated, such as the concentration of the positive cavity center area is 2.6 mol / L, the edge area is 2.5 mol / L, the negative cavity center area is 3.0 mol / L, the edge area is 2.9 mol / L, and the inlet and outlet concentrations are 2.7 mol / L and 2.8 mol / L, respectively.

[0041] According to the concentration distribution data, the maximum and minimum values of the concentration at each time are extracted from the concentration distribution data: from the above distribution data, the maximum value of the concentration at a certain time is determined to be 3.0 mol / L in the center region of the negative electrode cavity, and the minimum value is 2.5 mol / L in the edge region of the positive electrode cavity. Subsequently, the maximum and minimum values and their difference are recorded by continuously tracking the change of the maximum and minimum values at different times, such as the difference at this time is 3.0-2.5=0.5 mol / L, and the polarization voltage of the battery at the corresponding time is monitored synchronously, which reflects the degree of concentration polarization. For example, statistical analysis of a plurality of data sets shows that when the difference between the maximum and minimum values is 0.4 mol / L, the polarization voltage increases by 0.02 V; when the difference increases to 0.6 mol / L, the polarization voltage jumps to 0.05 V; and when the difference exceeds 0.7 mol / L, the polarization voltage increases to 0.08 V. Thus, a dynamic correlation between the difference between the maximum and minimum values of the concentration and the increase in the polarization voltage is established, i.e. a dynamic correlation between the concentration polarization and the polarization voltage, and the greater the difference between the maximum and minimum values of the concentration, the more significant the voltage loss caused by the concentration polarization.

[0042] Then, according to the above dynamic correlation between the concentration polarization and the polarization voltage, i.e. the greater the difference between the maximum and minimum values of the concentration, the more significant the voltage loss caused by the concentration polarization, the change of the flow rate parameter of the electrolyte will directly affect the difference between the maximum and minimum values of the concentration: when the flow rate increases, the circulation speed of the electrolyte in the battery increases, the electrolyte in each region is more fully mixed, the concentration distribution is more uniform, the difference between the maximum and minimum values of the concentration decreases, and the degree of concentration polarization decreases; when the flow rate decreases, the circulation of the electrolyte slows down, the concentration difference in each region increases, the difference between the maximum and minimum values of the concentration increases, and the degree of concentration polarization increases. The voltage loss caused by the concentration polarization can be represented by a variable resistor in the equivalent circuit, and the resistance value of the resistor is positively correlated with the degree of concentration polarization.

[0043] When the change of the flow rate parameter is mapped to the equivalent circuit, the corresponding change of the difference between the maximum and minimum values of the concentration and the resulting increase in the polarization voltage loss are first determined according to the change in the flow rate and the dynamic correlation between the concentration polarization and the polarization voltage; then, based on the relationship between the polarization voltage loss and the variable resistance in the equivalent circuit model, the variable resistance parameter under the change of the flow rate is inversely deduced. For example, when the flow rate decreases from 5 L / min to 3 L / min, according to the correlation relationship, the difference between the maximum and minimum values of the concentration increases from 0.4 mol / L to 0.6 mol / L, and the polarization voltage loss increases from 0.02 V to 0.05 V. If the circuit current is 10 A at this time, the variable resistance parameter increases from 0.02 V / 10 A=0.002 to 0.05 V / 10 A=0.005 , realizing the mapping of the change of the flow rate to the variable resistance parameter of the equivalent circuit.

[0044] Based on the obtained variable resistance parameters, the open circuit voltage value of the flow battery is retrieved, assuming that the open circuit voltage value at this time is 1.4V. The previously generated electrolyte concentration distribution data is associated with the open circuit voltage value, such as the positive electrode concentration of 3.0mol / L corresponding to the voltage of 1.45V, and the negative electrode concentration of 2.5mol / L corresponding to the voltage of 1.35V. Through these associated data, a concentration-voltage compensation coefficient table is constructed, which records the compensation value required for the open circuit voltage at different concentrations. For example, for every 0.1mol / L decrease in concentration, the voltage compensation is -0.01V.

[0045] After the electrolyte temperature parameter is fused with the concentration-voltage compensation coefficient table through nonlinear correction, it is necessary to clarify that the concentration-voltage compensation coefficient table originally only reflects the association between different electrolyte concentrations and corresponding voltage compensation values, but temperature changes will change the physicochemical properties of the electrolyte, such as the diffusion ability of active substances, electrode reaction rate, etc., and then nonlinearly affect the corresponding relationship between concentration and voltage, i.e. the compensation coefficient is not changed by a fixed proportion with the increase or decrease of temperature, but the influence degree of different temperature intervals on the concentration-voltage association is different.

[0046] In the fusion process, the concentration-voltage compensation coefficient at the reference temperature needs to be used as the basis, combined with the actual collected temperature parameters, to analyze the specific influence law of the deviation of the temperature from the reference temperature on the concentration-voltage relationship, and through the introduction of a nonlinear correction function, such as a nonlinear mapping algorithm based on improved BP neural network, the values in the compensation coefficient table are dynamically adjusted. Specifically, during model construction, the historically collected electrolyte concentration parameters, temperature parameters and corresponding actual voltage compensation values are used as training data sets, wherein the input layer is set to the concentration and temperature parameters, and the output layer is set to the corrected voltage compensation coefficient; the neural network algorithm is used to iteratively optimize the hidden layer weights and thresholds, and to learn the nonlinear association law between concentration and voltage compensation coefficient at different temperature intervals, such as the logarithmic change of the influence of temperature on the compensation coefficient at low temperature and the exponential change at high temperature, so that the model can accurately capture the nonlinear coupling relationship between the two. After multiple iterations and training until the prediction error is lower than the set threshold, the model can receive the current concentration and temperature parameters in real time, output the voltage compensation coefficient corrected by temperature, which reflects the voltage compensation demand brought by the change of concentration and accurately reflects the nonlinear interference of temperature fluctuation on this compensation relationship. Through such a fusion process, an electrolyte state coupling model that can reflect the influence of concentration and temperature on the electrolyte state is finally constructed.

[0047] Next, when the electrolyte charge state is calculated by the electrolyte state coupling model, the core is to separate the cross interference of concentration, temperature and other factors, and to focus on the quantification of the state of charge. The state of charge of the electrolyte essentially reflects the remaining proportion of active substances that can participate in energy conversion, while concentration changes and temperature fluctuations indirectly interfere with the calculation of this proportion by affecting the voltage signal. The coupling model has integrated the concentration-voltage compensation relationship and the nonlinear correction law of temperature on the compensation coefficient. When decoupling calculation, the model will first eliminate the interference of temperature fluctuations on the voltage signal, and lock the true correlation between concentration and voltage based on the voltage compensation value corrected by temperature; then exclude the local signal deviation caused by uneven concentration distribution, and combine the overall conversion proportion of active substances in the concentration distribution data, such as the completion degree of active substance oxidation-reduction reaction in the charging and discharging process, to strip the cross interference through the model algorithm, and finally calculate the state of charge information reflecting only the remaining amount of active substances.

[0048] When the electrolyte health is calculated by the electrolyte state coupling model, the interference of short-term operating parameters (such as instantaneous concentration fluctuations and temperature changes) on long-term performance evaluation needs to be separated, and the focus is on the health degradation characteristics of the electrolyte. The health state mainly reflects the long-term efficiency of the electrolyte, and is directly related to long-term indicators such as the loss degree of active substances and the concentration stability. Based on historical operation data and real-time monitoring results, the coupling model will first exclude the influence of short-term concentration fluctuations (such as concentration changes caused by single charging and discharging) and temperature fluctuations, and focus on long-term trend analysis: for example, by comparing the initial total amount of active substances with the current total amount of active substances, the cumulative loss rate of active substances is calculated; by analyzing whether there are phenomena such as continuous stratification and precipitation in long-term concentration distribution data, the concentration stability is evaluated. After the short-term interference is decoupled and separated by the model, the health state information reflecting the overall health degree of the electrolyte is obtained by comprehensively considering these long-term indicators.

[0049] Finally, the obtained electrolyte state of charge information and health state information are added to the multiple energy state information, and the addition of the two makes the multiple energy state information not only cover the real-time energy available state of the battery, but also include the long-term health performance state of the battery.

[0050] By analyzing the key parameters, constructing the correlation relationship and the coupling model, and performing dynamic decoupling calculation, the accurate acquisition of the electrolyte state of charge and health state is realized, and a reliable state basis is provided for the energy scheduling of the flow battery.

[0051] Further, the step A200 in the method provided in the embodiments of the present application comprises:

[0052] A210: performing temperature analysis based on the electrolyte state of charge information, performing charging and discharging analysis of the flow battery according to the operating temperature parameter, and setting a power loss target.

[0053] A220: perform a degradation analysis based on the health status information, perform a battery life analysis on the flow battery according to a health degradation rate, and set a battery life target.

[0054] A230: map the power loss target and the battery life target to a unified decision space, and set a plurality of to-be-scheduled targets.

[0055] A240: set a running interval constraint based on the electrolyte state of charge information, and perform a degradation rate constraint based on the health status information.

[0056] A250: perform a calculation on the plurality of to-be-scheduled targets based on the running interval constraint and the degradation rate constraint, generate a target constraint satisfaction degree, sort the target constraint satisfaction degrees in descending order, and generate a target dominance order.

[0057] A260: perform a multi-objective optimization on the plurality of to-be-scheduled targets based on the target dominance order, and formulate the fuzzy control strategy according to an optimization result.

[0058] In the embodiments of the present application, the to-be-scheduled targets include a power loss target set based on electrolyte state of charge information through temperature analysis and charge-discharge analysis, and a battery life target set based on health status information through degradation analysis and battery life analysis.

[0059] Optionally, when performing temperature analysis based on electrolyte state of charge information, first, accurate electrolyte state of charge information is obtained through dynamic decoupling calculation of an electrolyte state coupling model, for example, the state of charge is 72% obtained through decoupling calculation, which is derived from dynamic estimation and model calculation of electrolyte flow, concentration, temperature and other parameters. Subsequently, combined with the real-time collected running temperature parameters, such as continuous monitoring through a temperature sensor, the current battery working temperature is stabilized at 30°C, and a person skilled in the art can carry out charge-discharge simulation and experiment under different temperature conditions: for example, based on the state of charge of 72%, the charge-discharge process at 25°C, 30°C, 35°C and 40°C is tested respectively, and the internal resistance change and power output data at each temperature are recorded. Analysis shows that when the temperature rises from 30°C to 36°C, due to the change of electrolyte active material diffusion rate, the charge-discharge internal resistance increases from 0.03Ω to 0.05Ω, and the power loss rate increases from 1.8% to 4.2%; while the temperature is lower than 28°C, although the loss rate is low, the charge-discharge efficiency decreases. Comprehensive efficiency and loss balance, based on the analysis result, the power loss target is set to not more than 3% of the total output power, to ensure efficient and low-loss operation within a reasonable temperature interval.

[0060] Based on the health state information, the current electrolyte health state information is obtained through dynamic decoupling calculation of the electrolyte state coupling model, for example, the calculated health state is 85%, which integrates long-term monitoring indicators such as active material loss and concentration stability. Then, the health decay rate is calculated based on the historical operation data: for example, by recording the change of health state in the past 6 months from 88% to the current 85%, it is concluded that the average monthly decay is 0.5%. Further analysis is carried out in combination with battery life decay: if the decay rate is maintained at 0.5% per month, combined with the battery design life parameters such as the theoretical cycle number of active material, it is calculated that the battery can maintain a service life of 10 years; if the decay rate increases to 0.7% per month, the service life will be shortened to less than 8 years. In order to balance use and life, according to this analysis, the battery life target is set to control the health decay rate to be less than 0.5% per month, so as to ensure that the actual service life of the battery is not less than 90% of the designed life.

[0061] Next, the power loss target ≤ 3% and the battery life target, i.e. the decay rate ≤ 0.5% / month, are mapped to a unified decision space, which has power loss rate and health decay rate as two dimensions, and the two targets are converted into specific coordinate points in this space, for example, the power loss target corresponds to the coordinate (3%, 0), and the battery life target corresponds to the coordinate (0, 0.5% / month). Through this mapping, the specific position and range of multiple to-be-scheduled targets are clearly defined.

[0062] Then, based on the electrolyte state of charge information, the operation interval constraint is set, for example, according to the characteristics of the state of charge, the operation interval is set to 20%-90%, to avoid performance degradation caused by excessive discharge when the state of charge is below 20%, or overcharging risk when the state of charge is above 90%. Based on the health state information, the decay rate constraint is set, for example, when the health state is below 85%, the upper limit of the decay rate is further tightened from 0.5% / month to 0.4% / month, to prevent rapid deterioration of the health state.

[0063] Then, based on the operation interval constraint and the decay rate constraint, the target constraint satisfaction degree of multiple to-be-scheduled targets is calculated, first the specific requirements of each constraint are defined: the operation interval constraint requires the electrolyte state of charge to be maintained within the set interval, the decay rate constraint requires the health decay rate to be less than the set threshold, and the to-be-scheduled target itself requires the power loss rate to be less than the set target. When calculating, for each to-be-scheduled target, check whether it meets the above constraints one by one: first judge whether its corresponding state of charge is within the operation interval of 20%-90%, then check whether the health decay rate is less than 0.5% / month, and finally confirm whether the power loss rate is controlled within 3%, the target domination ordering rule is shown in Table 1.

[0064] For example, the power loss rate of a certain target to be scheduled is 2.8%, which meets ≤3%, the health degradation rate is 0.45% / month, which meets ≤0.5% / month, and the state of charge is maintained at 50%, which is in the interval of 20%-90%, so the constraint satisfaction degree of the target is 100%; the power loss rate of another target to be scheduled is 3.2%, which does not meet ≤3%, and the health degradation rate is 0.55% / month, which does not meet ≤0.5% / month, so the constraint satisfaction degree is 0%. According to the descending order of the target constraint satisfaction degree from high to low, a target dominance ordering is generated, in which the target with a satisfaction degree of 100% is ranked first as the object of priority optimization.

[0065] Finally, the target dominance relationship graph is generated based on the target dominance ordering rule, the weight distribution is performed after multi-objective conflict analysis and optimization resolution, the dominant target strength is generated and added to the fuzzy control strategy, and the specific steps are described in detail in A261-A264.

[0066] By setting specific targets based on energy state information, mapping to a unified space, setting constraints and ordering, precise planning of flow battery scheduling targets is realized, which provides a clear basis for subsequent multi-objective optimization and fuzzy control strategy formulation.

[0067] Table 1: Target dominance ordering rule table

[0068]

[0069] Further, the method provided in the embodiments of the application comprises the following steps A260:

[0070] A261: Hierarchically prioritizing the plurality of targets to be scheduled based on the target dominance ordering rule to generate a target dominance relationship graph.

[0071] A262: Performing multi-objective conflict analysis according to the target dominance relationship graph to generate a conflict path, and performing optimization resolution on the plurality of targets to be scheduled according to the conflict path to generate an optimization result.

[0072] A263: Based on the optimization result, weight distribution is performed to obtain a plurality of priority weight coefficients, and the optimization result is analyzed according to the plurality of priority weight coefficients to generate a dominant target strength.

[0073] A264: Adding the dominant target strength to the fuzzy control strategy.

[0074] Specifically, when the priority of the plurality of to-be-scheduled targets is layered based on the target dominance ordering rule, the targets are divided into four layers according to the constraint satisfaction degrees of 100%, 67%, 33% and 0% of the targets. For example, there are four to-be-scheduled targets: the state of charge of target 1 is 60%, which is in the interval of 20%-90%; the health attenuation rate is 0.4% / month≤0.5% / month; the power loss rate is 2.8%≤3%, all of which meet the constraints, and the constraint satisfaction degree is 100%, which is classified into the first layer; the second, third and fourth layers are classified into the second, third and fourth layers according to the target dominance ordering rule of step A250. Through this layering, a target dominance relationship graph is generated, in which the first layer of targets dominates the second, third and fourth layers of targets, the second layer dominates the third and fourth layers, and so on, so that the priority level relationship of each target is clearly presented.

[0075] Then, when the multi-target conflict analysis is performed according to the target dominance relationship graph, the conflict points between targets of different levels need to be identified. For example, target 1 (power loss rate 2.8%, life attenuation rate 0.4%) in the first layer of the graph and target 5 (power loss 2.5%, life attenuation 0.52%) in the second layer have a conflict: the power loss of target 5 is lower, but the life attenuation rate exceeds the constraint of 0.5% / month, which conflicts with the life target of target 1. By tracking the associated path of the target in the graph, a conflict path is generated: power loss reduction→charge and discharge power increase→electrolyte circulation speed increase→active material loss increase→life attenuation rate increase. For this conflict path, optimization and resolution are performed, for example, the charge and discharge power is adjusted to a moderate level, so that the power loss is controlled at 2.6% and the life attenuation rate is reduced to 0.48%, which meets the power loss target≤3% and the life attenuation constraint≤0.5% / month, and an optimized result is generated.

[0076] When the weight is allocated based on the optimized result, the priority weight coefficients of the targets need to be determined in combination with the power grid operation scenario. Assuming that the optimized result contains two feasible schemes: scheme one, power loss 2.6%, life attenuation 0.48%; scheme two, power loss 2.9%, life attenuation 0.42%. If the power grid is currently in a power consumption peak, more attention is paid to energy output efficiency, and the weight of the power loss target is set to 0.6 and the weight of the battery life target is set to 0.4. The comprehensive score of scheme one is calculated: 2.6×0.6+0.48×0.4=1.56+0.192=1.752; the comprehensive score of scheme two is calculated: 2.9×0.6+0.42×0.4=1.74+0.168=1.908. Scheme one has a lower score and is more optimal, and the priority weight coefficients are determined as power loss 0.6 and life 0.4. According to the weight, the optimized result is analyzed to generate a dominant target strength: scheme one is superior to scheme two in the power loss index, and has a higher weight, so the dominant target strength of scheme one to scheme two is 0.75.

[0077] Finally, the dominance target strength is added to the fuzzy control strategy, so that the strategy can respond to better targets according to the strength. For example, scheme one is preferentially executed during scheduling, ensuring that the battery life is considered while meeting the efficiency requirements, so that the fuzzy control strategy is more in line with the actual operation requirements.

[0078] By layering the constraint satisfaction degree, determining the target priority, analyzing the conflict path, optimizing the conflict resolution, combining the scene to assign weights to generate the dominance strength, and integrating the strength into the control strategy, the fuzzy control strategy can finally accurately respond to the multi-objective optimization result, and improve the scientificity and effectiveness of the flow battery energy scheduling.

[0079] Further, the step A300 in the method provided by the embodiment of the present application comprises:

[0080] A310: executing the fuzzy control strategy, dynamically updating the priority weight coefficients of the plurality of to-be-scheduled targets according to the grid operating condition change parameters, and generating a plurality of target weight coefficients.

[0081] A320: performing power analysis on the charge and discharge of the flow battery based on the plurality of target weight coefficients, and generating a real-time charge and discharge power signal.

[0082] A330: performing flow compensation analysis on the temperature distribution of the flow battery based on the plurality of target weight coefficients, and generating an electrolyte flow control signal.

[0083] A340: performing electric-liquid bidirectional time sequence collaborative optimization on the flow battery according to the real-time charge and discharge power signal and the electrolyte flow control signal, and obtaining time sequence collaborative optimization parameters.

[0084] A350: performing dynamic scheduling on the flow battery according to the time sequence collaborative optimization parameters, and generating the scheduling result.

[0085] Specifically, first, based on the flow battery energy operation parameter record operating condition change, the fuzzy control strategy is executed to detect the power loss and / or battery life mutation result, and then the corresponding target weight coefficient is determined, and the specific steps are described in detail in A311-A315. In addition, the control signal set is jointly constructed by the real-time charge and discharge power signal and the electrolyte flow control signal of the flow battery.

[0086] Then, after generating multiple target weight coefficients, when power analysis is performed on the charge and discharge of the flow battery based on these coefficients, the optimal power range needs to be determined in combination with the weight distribution result. For example, if the target weight coefficient is 0.7 for the power loss target and 0.3 for the battery life target, it indicates that the current focus is on reducing energy loss. Through the power analysis model, the real-time load demand of the power grid is input, for example, 120 kW, and the weight is calculated to obtain that the charge and discharge power needs to be controlled within the range of 90-100 kW, which can meet most of the load demand and avoid the increase of internal resistance caused by high power. Accordingly, a real-time charge and discharge power signal is generated, such as 95 kW.

[0087] Further, the construction of the power analysis model needs to be oriented by the target weight coefficient, and the load characteristics of the power grid and the operation constraints of the battery are integrated. First, the input parameters of the model are determined, including the real-time load demand of the power grid, the target weight coefficient, the characteristic curve of the battery internal resistance with power change, such as the nonlinear relationship data of power and internal resistance, and the charge and discharge efficiency threshold. Then, a double-layer analysis structure is built: the first layer is a load matching unit, based on the high priority of the power loss in the target weight, the power range that can meet more than 80% of the load demand is first selected, such as 80-120 kW under the condition of 120 kW load; the second layer is a loss evaluation unit, which calculates the power loss rate corresponding to different powers in the range in combination with the internal resistance characteristic curve, and refers to the battery life weight to eliminate the range where the health decay rate suddenly rises due to high power. Finally, through iterative calibration, the target weight coefficient is converted into the balanced weight of loss and life, and the preliminary range is contracted to determine the optimal power range that can match the load and meet the weight orientation, such as 90-100 kW, and the real-time response algorithm is embedded to ensure that the specific power value is dynamically output according to the load fluctuation, and the model construction is completed.

[0088] Then, when the flow compensation analysis of the temperature distribution of the flow battery is performed based on multiple target weight coefficients, the influence of temperature on different targets needs to be balanced according to the weight priority. Assuming that the current temperature distribution of the battery is 36℃ for the positive electrode area and 32℃ for the negative electrode area, with a temperature difference of 4℃, and the target weight is 0.3 for the battery life target, it can be known that a large temperature difference will accelerate the decay of active substances, for example, when the temperature difference exceeds 5℃, the health decay rate increases from 0.4% / month to 0.6% / month. Through the flow compensation model, it is calculated that the electrolyte flow needs to be adjusted from the original 4L / min to 5L / min to enhance the circulation heat exchange and reduce the temperature difference to within 2℃. Accordingly, an electrolyte flow control signal is generated, such as 5L / min.

[0089] Further, the construction of the flow compensation model needs to take the temperature distribution data and target weight as the core input, and combine the correlation characteristics of electrolyte flow to temperature adjustment to realize dynamic compensation. First, determine the model input parameters, including real-time temperature distribution of the battery, target weight coefficient, historical correlation data of electrolyte flow and temperature field change, such as the change rate of temperature difference under different flow and the corresponding relationship between health degradation rate and temperature difference. Second, build the temperature difference-flow correlation module, obtain the temperature difference adjustment effect under different flow through experiment or simulation, for example, record the corresponding relationship that the temperature difference is maintained at 4°C under the flow of 4L / min, the temperature difference is reduced to 3°C under the flow of 4.5L / min, and the temperature difference is reduced to 2°C under the flow of 5L / min, and establish the nonlinear mapping curve of flow and temperature difference.

[0090] At the same time, embed the target weight oriented mechanism, when the battery life target weight is high, the model prioritizes to inhibit health degradation as the goal, sets the temperature difference safety threshold, such as ≤2°C, and based on the mapping curve, reversely calculates the minimum flow adjustment value that meets the threshold, such as from 4L / min to 5L / min. Finally, add a constraint verification link to ensure that the calculated flow value will not cause excessive increase of power loss, such as excessive flow will increase pump consumption, by balancing the temperature adjustment effect and additional loss, output the final electrolyte flow regulation signal, complete the construction of the flow compensation model.

[0091] Afterwards, according to the real-time charging and discharging power signal 95kW and the electrolyte flow regulation signal 5L / min, the power-liquid bidirectional time sequence collaborative optimization is carried out, which needs to realize the dynamic matching of power change and flow adjustment. For example, when the grid load suddenly increases to 130kW, the real-time charging and discharging power signal needs to rise to 105kW, at this time the flow regulation signal needs to be increased to 5.5L / min. If the flow is not adjusted in time, the temperature may rise to 38°C, causing the power loss rate to break through 3%; if the flow is adjusted to 6L / min, although the temperature can be controlled, the pump consumption will be increased, which will in turn increase the total loss. Through the time sequence collaborative algorithm, the corresponding relationship between power and flow is determined, for example, the power 95-105kW corresponds to the flow 5-5.5L / min, and the adjustment delay is not more than 2 seconds, and finally the time sequence collaborative optimization parameters are obtained.

[0092] Finally, according to the time sequence collaborative optimization parameters, the flow battery is dynamically scheduled, and the system will automatically execute the adjustment according to the real-time working condition. For example, in the stable load stage, maintain the power of 95kW and the flow of 5L / min, the power loss rate is stabilized at 2.2%, and the health degradation rate is 0.38% / month; in the load fluctuation stage, the parameters are quickly responded according to the parameters, the power and flow are changed synchronously, and it is ensured that the loss and degradation are controlled within the target range, and finally the scheduling results including time period, power, flow, loss rate and degradation rate are generated.

[0093] Through power resolution guided by target weight, flow compensation, and electric-hydraulic timing coordination optimization, precise linkage of the liquid flow battery electrical system and the liquid system is realized, the power loss and battery life are balanced while meeting the demand of the power grid, and the stability and economy of energy scheduling are improved.

[0094] Further, the step A310 in the method provided by the embodiment of the application includes:

[0095] A311: Record the environmental conditions based on the energy operation parameters of the liquid flow battery to obtain battery condition change parameters.

[0096] A312: Execute the fuzzy control strategy, detect the plurality of to-be-scheduled targets according to the battery condition change parameters, and generate a detection result, which is a power loss mutation result and / or a battery life mutation result.

[0097] A313: When the detection result is the power loss mutation result, the priority weight of the power loss target is raised to the highest level, and a first target weight coefficient is determined.

[0098] A314: When the detection result is the battery life mutation result, the priority weight of the battery life target is raised to the highest level, and a second target weight coefficient is determined.

[0099] A315: When the detection result is the power loss mutation result and the battery life mutation result, then the power loss mutation result and the battery life mutation result are analyzed for scheduling influence, and a third target weight coefficient is determined according to an influence factor.

[0100] Specifically, when recording the environmental conditions based on the energy operation parameters of the liquid flow battery, key operation data such as electrolyte flow parameters, concentration parameters, and temperature parameters need to be continuously collected, and the change trend thereof over time is recorded. For example, the electrolyte flow, concentration, and temperature are recorded once every 5 minutes, and information such as load fluctuation on the power grid side is associated. Through integration of these data, battery condition change parameters are formed, reflecting the real-time changes of the battery operation environment and its own state.

[0101] When executing the fuzzy control strategy, the battery operating condition change parameters described above are used to detect multiple targets to be dispatched, i.e., power loss targets and battery life targets, in real time. Specifically, the power loss mutation threshold is set to 2%, i.e., a power loss rate that increases by more than 2% within 1 hour is considered to be a mutation; the battery life mutation threshold is 0.3% / month, i.e., a health decay rate that increases by more than 0.3% / month within 1 hour is considered to be a mutation. If the power loss rate is detected to increase from 1.8% to 4.2% exceeding the 2% threshold, a power loss mutation result is generated; if the health decay rate is detected to increase from 0.4% / month to 0.8% / month exceeding the 0.3% threshold, a battery life mutation result is generated; if both exceed the corresponding threshold, both mutation results are generated.

[0102] When the detection result is the power loss mutation result, it indicates that the current energy output stability is severely affected, and the power transmission efficiency needs to be prioritized, so the priority weight of the power loss target is raised to the highest level. For example, originally the power loss target weight is 0.5 and the battery life target weight is 0.5, which is adjusted to a power loss target weight of 0.8 and a battery life target weight of 0.2, and determined as the first target weight coefficient.

[0103] When the detection result is the battery life mutation result, it indicates that the long-term performance of the battery is at risk of rapid decay, and the battery life needs to be prioritized, so the priority weight of the battery life target is raised to the highest level. For example, it is adjusted to a battery life target weight of 0.8 and a power loss target weight of 0.2, and determined as the second target weight coefficient.

[0104] When the detection result contains both power loss mutation and battery life mutation, the scheduling impact analysis needs to quantify the influence degree of both on scheduling from the two dimensions of real-time operation and long-term performance. First, for the power loss mutation, the influence on the current energy output is analyzed: by comparing the output power data before and after the mutation, the interference degree of a 15% power reduction on the power grid load balance and power supply stability is determined, for example, whether this reduction causes a local power supply gap, triggers the start of backup power, etc., to evaluate its urgency and influence range on real-time scheduling; for the battery life mutation, the influence on long-term operation is analyzed: combined with the designed life of the battery, which is assumed to be 10 years, the increase in total life cycle cost and maintenance frequency due to a 3-year reduction in life is calculated, for example, a shortened life will cause the replacement cycle to be advanced and the total maintenance cost to increase by 20%, etc., to evaluate its persistent influence on long-term scheduling strategy. On this basis, by combining the real-time power supply priority and long-term economic weight, the real-time influence of the power loss mutation is quantified as an influence factor of 0.6, and the long-term influence of the battery life mutation is quantified as an influence factor of 0.4, the sum of which is 1, and meets the principle that the greater the influence, the higher the weight. Finally, the weight proportion of the power loss target and the battery life target in the third target weight coefficient is determined according to the factor.

[0105] By recording the working condition changes, detecting the mutation results, and dynamically adjusting the priority weight coefficients in different scenes, the importance of the to-be-scheduled targets under different working conditions is accurately adapted, and it is ensured that the energy scheduling of the flow battery can cope with the emergency and balance the short-term efficiency and long-term performance.

[0106] Further, step A400 in the method provided by the embodiment of the application comprises:

[0107] A410: setting a historical scheduling period according to the scheduling result, and collecting execution result data of the historical scheduling period.

[0108] A420: performing deviation analysis based on the execution result data and the fuzzy control strategy, and generating an execution deviation parameter.

[0109] A430: constructing a reverse tracing path according to the execution deviation parameter, performing deviation positioning according to the reverse tracing path, and obtaining a tracing defect positioning point.

[0110] A440: performing incremental learning on the fuzzy control strategy based on the tracing defect positioning point, and generating the fuzzy control optimization strategy.

[0111] In one embodiment, when the historical scheduling period is set according to the scheduling result, the period length needs to be determined in combination with the time span and fluctuation characteristics of the scheduling result. For example, if the scheduling result shows that there are 3 obvious cases of power loss exceeding the standard in the past 30 days, and the interval is about 10 days, 10 days is set as a historical scheduling period. In each period, the execution result data is collected, including the actual charging and discharging power, the power loss rate, the health decay rate, the electrolyte flow, the temperature distribution and other key indicators. For example, the actual average power loss rate in a certain period is 3.2%, and the expected value of the fuzzy control strategy is 3%; the actual health decay rate is 0.52% / month, and the expected value of the strategy is 0.5% / month. These data need to be recorded in detail and associated with the strategy parameters of the corresponding period.

[0112] Then, when the deviation analysis is performed based on the execution result data and the fuzzy control strategy, the actual indicators and the target values preset by the strategy need to be compared one by one. For example, for the power loss indicator, the absolute deviation between the actual value and the expected value is 3.2%-3%=0.2%, and the relative deviation is 0.2% / 3%≈6.7%; for the health decay rate, the deviation is 0.52%-0.5%=0.02% / month, and the relative deviation is 0.02% / 0.5%=4%. At the same time, the associated factors of the deviation are analyzed, such as whether the power loss deviation is related to the too high charging and discharging power signal, whether the health decay deviation is related to the insufficient electrolyte flow regulation, and the like. Based on the deviation values and the associated analysis results, the execution deviation parameters are generated, such as the power loss deviation 0.2%, the health decay deviation 0.02% / month and the corresponding associated factor labels.

[0113] Then, when constructing the reverse tracing path according to the execution deviation parameter, it is necessary to trace back from the deviation index to each link of the strategy layer by layer. For example, for the power loss deviation of 0.2%, first trace back to the real-time charging and discharging power signal, and find that the power signal is 105kW at a certain period, which is higher than the recommended interval of 95-100kW of the strategy model; then trace back to the target weight coefficient, and find that the power loss target weight is 0.6 and the battery life target is 0.4 at this period, and the internal resistance growth under high power is not fully considered during power analysis; continue to trace back to the target dominance ordering, and find that there is a mistake in the target constraint satisfaction calculation at this period, and the influence of high power on loss is not fully considered. Through such hierarchical tracing, a reverse tracing path of power loss deviation→power signal→target weight coefficient→target constraint satisfaction calculation is constructed, and the defect positioning point is located according to the path, such as determining that the defect point is the target constraint satisfaction calculation link which does not fully consider the nonlinear relationship between power and loss.

[0114] Finally, based on the defect positioning point of the tracing, the defect points of the electro-hydraulic coordination sequence and the life target are determined, and after corresponding testing and confidence analysis, the fuzzy control strategy is incrementally learned to generate an optimized fuzzy control strategy, which is specifically described in A441-A444.

[0115] By setting periodic data collection, analyzing deviation, and reverse tracing defect positioning, the key links in the fuzzy control strategy that cause execution deviation are accurately identified, providing a clear improvement direction for subsequent incremental learning of the strategy and generation of an optimized strategy, and improving the adaptability and accuracy of the strategy.

[0116] Further, the step A440 in the method provided by the embodiment of the application includes:

[0117] A441: based on the defect positioning point of the tracing, defect analysis is performed to determine the defect points of the electro-hydraulic coordination sequence and the life target.

[0118] A442: based on the defect point of the electro-hydraulic coordination sequence, power reduction testing is performed to generate a first test result for confidence analysis, and a first confidence degree is obtained.

[0119] A443: based on the defect point of the life target, attenuation increment testing is performed to generate a second test result for confidence analysis, and a second confidence degree is obtained.

[0120] A444: according to the first test result combined with the first confidence degree and the second test result combined with the second confidence degree, the fuzzy control strategy is incrementally learned to generate an optimized fuzzy control strategy.

[0121] Optionally, when analyzing the defect based on the defect positioning point, first combine the specific link of the positioning point. For example, if the defect positioning point of the target constraint satisfaction degree calculation does not fully consider the nonlinear relationship between power and loss, further analysis will find that: in the electric-liquid coordinated scheduling, when the charge and discharge power signal suddenly increases from 95kW to 105kW, the electrolyte flow control signal does not follow up synchronously, causing the battery internal temperature to rise from 32℃ to 36℃ in a short time, and the power loss rate increases from 2.2% to 3.2%, which is an electric-liquid coordination timing defect point; at the same time, in the life target setting, the original strategy does not take into account the accelerated influence of temperature above 35℃ on health decay, when the temperature is continuously higher than 35℃, the health decay rate increases from 0.5% / month to 0.65% / month, which exceeds the expected target, which is a life target defect point.

[0122] When testing the power reduction based on the electric-liquid coordination timing defect point, the timing coordination effect under different powers needs to be simulated. The specific steps are: starting from 105kW, gradually decreasing the charge and discharge power to 85kW by 5kW as a gradient, keeping at each power point for 10 minutes, and recording the response delay time of the electrolyte flow control signal and the corresponding power loss rate. For example, the test found that: when the power is 105kW, the flow response delay is 2 seconds, and the loss rate is 3.2%; when the power decreases to 95kW, the flow response delay shortens to 0.5 seconds, and the loss rate is 2.3%; when the power decreases to 90kW, the flow has no delay, and the loss rate is 2.1%. Select the 95kW power point with the optimal balance of loss rate and response delay as the key test result, i.e. the first test result. Confidence analysis is performed on the result, and the test is repeated 10 times. If the loss rate is stable at 2.2%-2.4% and the response delay is ≤0.5 seconds in 9 of the 10 times, the first confidence is 90%.

[0123] When testing the decay increment based on the life target defect point, the health decay trend under different temperatures needs to be simulated. The specific steps are: starting from 30℃, gradually increasing the battery temperature to 40℃ by 2℃ as a gradient, continuously running for 24 hours at each temperature point, and recording the change of health decay rate. For example, the test found that: when the temperature is 30℃, the decay rate is 0.5% / month; when the temperature is 32℃, the decay rate is 0.52% / month; when the temperature is 35℃, the decay rate is 0.58% / month; when the temperature is 38℃, the decay rate is 0.68% / month. When the electrolyte flow is increased from 5L / min to 5.5L / min at 35℃, the decay rate decreases to 0.53% / month, which is the second test result. Confidence analysis is performed on the result, and the test is repeated 8 times. The decay rate is stable at 0.52%-0.54% in 7 of the 8 times, so the second confidence is 87.5%.

[0124] Finally, according to the first test result combined with 90% confidence, the power and flow coordination timing parameters in the fuzzy control strategy are adjusted to shorten the response threshold; according to the second test result combined with 87.5% confidence, the correlation coefficient of temperature and flow in the strategy is corrected. Through the incremental learning of these two aspects, the fuzzy control optimization strategy that can optimize the electric-hydraulic coordination efficiency and life attenuation control at the same time is finally generated.

[0125] Through targeted defect analysis, step-by-step testing and confidence verification, precise iteration of the fuzzy control strategy is realized, so that the optimized strategy can more efficiently balance power loss and battery life, and improve the stability and long-term effectiveness of flow battery energy scheduling.

[0126] Further, the step A400 in the method provided by the embodiment of the application comprises:

[0127] A450: verifying whether the fuzzy control optimization strategy can eliminate the execution deviation parameter of the fuzzy control strategy, and generating a verification result.

[0128] A460: when the verification result is that the fuzzy control optimization strategy can eliminate the execution deviation parameter of the fuzzy control strategy, generating a replacement instruction, replacing the fuzzy control optimization strategy with the fuzzy control strategy through the replacement instruction, and constructing a signal cascade reconstruction space.

[0129] A470: mapping the set of control signals to the signal cascade reconstruction space for reconstruction compensation, and generating a set of control reconstruction signals.

[0130] A480: performing scheduling analysis based on the set of control reconstruction signals, dividing a plurality of scheduling levels, matching the plurality of scheduling levels with the set of control reconstruction signals, and generating the multi-level control instruction.

[0131] In one embodiment, first, simulation verification is performed based on historical execution data. For example, the original fuzzy control strategy has an execution deviation parameter of power loss deviation 0.2% and health attenuation deviation 0.02% / month, the fuzzy control optimization strategy is applied to the same historical scheduling scene, and the corresponding power loss rate and health attenuation rate are simulated and calculated. If the simulation result shows that the power loss deviation is reduced to within 0.05% and the health attenuation deviation is reduced to within 0.005% / month, and the deviation is stable in this range for 10 consecutive scheduling periods, a verification result that the optimization strategy can eliminate the execution deviation is generated; if there is still a deviation that exceeds the acceptable range, a verification result that the optimization strategy cannot eliminate the execution deviation is generated.

[0132] Then, when the verification result is that the fuzzy control optimization strategy can eliminate the execution deviation parameter, the system generates a replacement instruction by which the optimization strategy formally replaces the original fuzzy control strategy. At the same time, a signal cascade reconstruction space is constructed, which is used to integrate the correlation of each signal in the regulation signal set, for example, a dynamic mapping model of the real-time charging and discharging power signal and the electrolyte flow regulation signal is established to ensure that the logic consistency of the replaced strategy can be maintained when each signal is called.

[0133] Then, when the verification result is that the fuzzy control optimization strategy can eliminate the execution deviation parameter, the system generates a replacement instruction by which the optimization strategy formally replaces the original fuzzy control strategy. At the same time, a signal cascade reconstruction space is constructed, which is used to integrate the correlation of each signal in the regulation signal set, for example, a dynamic mapping model of the real-time charging and discharging power signal and the electrolyte flow regulation signal is established to ensure that the logic consistency of the replaced strategy can be maintained when each signal is called.

[0134] Finally, based on the regulation reconstruction signal set, scheduling analysis is performed, and multiple scheduling levels are determined according to the load fluctuation characteristics of the power grid. For example, the power grid conditions are divided into three levels: peak load > 120kW, flat peak 80-120kW, and valley load < 80kW, which correspond to the high-power-high-flow, medium-power-medium-flow, and low-power-low-flow signal intervals in the regulation reconstruction signal set, respectively. The levels are matched with the corresponding reconstruction signals to generate multi-level regulation instructions, such as the peak level instruction of 110kW charging and discharging power and 5.5L / min electrolyte flow, the flat peak level of 90kW and 5L / min, and the valley level of 70kW and 4.5L / min.

[0135] By verifying the effectiveness of the optimization strategy, replacing the strategy, reconstructing the regulation signal, and matching the scheduling level to generate instructions, the precise coordination and hierarchical response of the regulation signal are achieved, and the adaptability and control accuracy of the energy intelligent scheduling of the flow battery are improved.

[0136] In summary, the energy intelligent scheduling method for flow batteries provided by the embodiments of the present application has the following technical effects:

[0137] The present application collects the energy operation parameters of the flow battery in real time, obtains the energy state information and scheduling related data through dynamic estimation, multi-objective optimization, fuzzy control strategy formulation, etc., calculates the optimization results and regulation signals of the to-be-scheduled target, and adjusts them in combination with the incremental learning and strategy optimization of the reverse tracing, thereby accurately implementing the energy intelligent scheduling of the flow battery, making the energy scheduling results of the flow battery more accurate and reliable, and achieving the technical effects of energy intelligent scheduling of the flow battery, reducing power loss, prolonging battery life, and improving the matching degree of scheduling and actual demand.

[0138] In some embodiments, the energy intelligent scheduling system for flow battery is provided based on the same inventive concept as the foregoing embodiment one. Figure 2 In some embodiments, the energy intelligent scheduling system for flow battery is provided based on the same inventive concept as the foregoing embodiment one.

[0139] An energy state information acquisition module 1 is configured to collect energy operation parameters of the flow battery in real time, dynamically estimate the electrolyte state according to the energy operation parameters, and obtain a plurality of energy state information.

[0140] A fuzzy control strategy acquisition module 2 is configured to perform scheduling analysis on the flow battery based on the plurality of energy state information, set a plurality of to-be-scheduled targets, perform multi-objective optimization according to the plurality of to-be-scheduled targets, formulate a fuzzy control strategy according to the optimization result.

[0141] A scheduling result acquisition module 3 is configured to perform dynamic update on the plurality of to-be-scheduled targets according to the fuzzy control strategy, generate a set of regulation and control signals, perform bidirectional collaborative scheduling on the flow battery based on the set of regulation and control signals, and generate a scheduling result.

[0142] An energy intelligent scheduling execution module 4 is configured to perform reverse tracing according to the scheduling result, perform incremental learning on the fuzzy control strategy according to the tracing path, generate a fuzzy control optimization strategy to replace the fuzzy control strategy, update the set of regulation and control signals, obtain a plurality of regulation and control instructions, and intelligently schedule the energy of the flow battery.

[0143] Further, the energy state information acquisition module 1 is configured to perform the following steps:

[0144] The energy operation parameters are analyzed to obtain electrolyte flow parameters, electrolyte concentration parameters and electrolyte temperature parameters of the flow battery. The concentration distribution of the flow battery is calculated based on the electrolyte concentration parameters to generate electrolyte concentration distribution data. Extreme value correlation analysis is performed according to the electrolyte concentration distribution data to obtain a dynamic correlation relationship of concentration polarization. The electrolyte flow parameters are mapped to an equivalent circuit according to the dynamic correlation relationship of concentration polarization to obtain a variable resistance parameter. The open-circuit voltage value of the flow battery is retrieved based on the variable resistance parameter. The electrolyte concentration distribution data is correlated with the open-circuit voltage value to construct a concentration-voltage compensation coefficient table. The electrolyte temperature parameters are nonlinearly corrected and fused with the concentration-voltage compensation coefficient table to construct an electrolyte state coupling model. Dynamic decoupling calculation of electrolyte charge is performed through the electrolyte state coupling model to obtain electrolyte state of charge information. Dynamic decoupling calculation of electrolyte health is performed through the electrolyte state coupling model to obtain health state information. The electrolyte state of charge information and the health state information are added to the plurality of energy state information.

[0145] Further, the fuzzy control strategy acquisition module 2 is configured to perform the following steps:

[0146] Temperature analysis is performed based on the electrolyte state of charge information. Charge and discharge analysis of the flow battery is performed according to the operating temperature parameters to set a power loss target. Decay analysis is performed based on the health state information. Battery life analysis of the flow battery is performed according to the health decay rate to set a battery life target. The power loss target and the battery life target are mapped to a unified decision space to set a plurality of to-be-scheduled targets. Operating interval constraints are set based on the electrolyte state of charge information. Decay rate constraints are performed based on the health state information. The plurality of to-be-scheduled targets are calculated based on the operating interval constraints and the decay rate constraints to generate a target constraint satisfaction degree. The target constraint satisfaction degree is sorted in descending order to generate a target dominance order. The plurality of to-be-scheduled targets are multi-objective optimized based on the target dominance order. The fuzzy control strategy is formulated according to the optimization result.

[0147] Further, the fuzzy control strategy acquisition module 2 is configured to perform the following steps:

[0148] The target domination ordering rule is used to prioritize the plurality of to-be-scheduled targets, a target domination relationship graph is generated, multi-target conflict analysis is performed according to the target domination relationship graph, a conflict path is generated, the plurality of to-be-scheduled targets are optimized and eliminated according to the conflict path, and an optimization result is generated; based on the optimization result, a plurality of priority weight coefficients are obtained, the optimization result is analyzed according to the plurality of priority weight coefficients, and a domination target strength is generated; and the domination target strength is added to the fuzzy control strategy.

[0149] Further, the scheduling result acquisition module 3 is configured to perform the following steps:

[0150] The fuzzy control strategy is executed, the priority weight coefficients of the plurality of to-be-scheduled targets are dynamically updated according to the grid operating condition change parameters, a plurality of target weight coefficients are generated, the power of the charge and discharge of the flow battery is analyzed based on the plurality of target weight coefficients, a real-time charge and discharge power signal is generated, the flow compensation analysis of the temperature distribution of the flow battery is performed based on the plurality of target weight coefficients, an electrolyte flow control signal is generated, the flow battery is optimized in an electric-liquid bidirectional time sequence according to the real-time charge and discharge power signal and the electrolyte flow control signal, and a time sequence cooperative optimization parameter is obtained; the flow battery is dynamically scheduled according to the time sequence cooperative optimization parameter, and the scheduling result is generated.

[0151] Further, the scheduling result acquisition module 3 is configured to perform the following steps:

[0152] The environment operating condition record based on the energy operating parameters of the flow battery is performed, and a battery operating condition change parameter is obtained; the fuzzy control strategy is executed, the plurality of to-be-scheduled targets are detected according to the battery operating condition change parameter, and a detection result is generated, the detection result being a power loss mutation result and / or a battery life mutation result; when the detection result is the power loss mutation result, the priority weight of the power loss target is raised to the highest level, and a first target weight coefficient is determined; when the detection result is the battery life mutation result, the priority weight of the battery life target is raised to the highest level, and a second target weight coefficient is determined; when the detection result is the power loss mutation result and the battery life mutation result, the power loss mutation result and the battery life mutation result are analyzed, and a third target weight coefficient is determined according to the influence factor.

[0153] Further, the energy intelligent scheduling execution module 4 is configured to perform the following steps:

[0154] According to the scheduling result, a historical scheduling period is set, execution result data of the historical scheduling period is collected, deviation analysis is performed based on the execution result data in combination with the fuzzy control strategy, an execution deviation parameter is generated, a reverse tracing path is constructed according to the execution deviation parameter, deviation positioning is performed according to the reverse tracing path, a traced defect positioning point is obtained, the fuzzy control strategy is subjected to incremental learning based on the traced defect positioning point, and the fuzzy control optimization strategy is generated.

[0155] Further, the energy intelligent scheduling execution module 4 is used to execute the following steps:

[0156] Based on the traced defect positioning point, defect analysis is performed to determine an electro-hydraulic coordination time sequence defect point and a life target defect point, power reduction testing is performed based on the electro-hydraulic coordination time sequence defect point, a first test result is generated for confidence analysis, a first confidence degree is obtained, attenuation increment testing is performed based on the life target defect point, a second test result is generated for confidence analysis, a second confidence degree is obtained, and the fuzzy control strategy is subjected to incremental learning according to the first test result in combination with the first confidence degree and the second test result in combination with the second confidence degree, and the fuzzy control optimization strategy is generated.

[0157] Further, the energy intelligent scheduling execution module 4 is used to execute the following steps:

[0158] Whether the fuzzy control optimization strategy can eliminate the execution deviation parameter of the fuzzy control strategy is verified, a verification result is generated, when the verification result is that the fuzzy control optimization strategy can eliminate the execution deviation parameter of the fuzzy control strategy, a replacement instruction is generated, the fuzzy control optimization strategy is used to replace the fuzzy control strategy through the replacement instruction, a signal cascade reconstruction space is constructed, the control signal set is mapped to the signal cascade reconstruction space for reconstruction compensation, a control reconstruction signal set is generated, scheduling analysis is performed based on the control reconstruction signal set, a plurality of scheduling levels are demarcated, the plurality of scheduling levels are matched with the control reconstruction signal set, and the multi-level control instruction is generated.

[0159] The energy intelligent scheduling system of the flow battery provided in the embodiment can execute the energy intelligent scheduling method of the flow battery provided in any embodiment of the application, has the corresponding function modules and beneficial effects of the execution method.

[0160] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, the various units and modules are only divided according to the functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific name of each functional unit is only for the convenience of mutual differentiation, and does not serve to limit the protection scope of the present application.

[0161] The above detailed description does not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application. In some cases, the actions or steps described in the present application can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

Claims

1. A method of energy smart dispatching of a flow battery, characterized in that, The method comprises: Real-time acquisition of energy operation parameters of the flow battery, dynamic estimation of electrolyte state according to the energy operation parameters, and obtaining of multiple energy state information; The method for obtaining multiple energy state information comprises: Analyzing the energy operation parameters to obtain electrolyte flow parameters, electrolyte concentration parameters, and electrolyte temperature parameters of the flow battery; Concentration distribution calculation of the flow battery based on the electrolyte concentration parameters, generation of electrolyte concentration distribution data, extreme value correlation analysis according to the electrolyte concentration distribution data, and obtaining of a dynamic correlation relationship of concentration polarization; Mapping the electrolyte flow parameters to an equivalent circuit according to the dynamic correlation relationship of concentration polarization, and obtaining of a variable resistance parameter; Retrieving an open-circuit voltage value of the flow battery based on the variable resistance parameter, correlating the electrolyte concentration distribution data with the open-circuit voltage value, and constructing a concentration-voltage compensation coefficient table; Nonlinear correction fusion of the electrolyte temperature parameters and the concentration-voltage compensation coefficient table, and constructing an electrolyte state coupling model; Dynamic decoupling calculation of electrolyte charge through the electrolyte state coupling model, and obtaining of electrolyte state of charge information; Dynamic decoupling calculation of electrolyte health through the electrolyte state coupling model, and obtaining of health state information; Adding the electrolyte state of charge information and the health state information to the multiple energy state information; Scheduling analysis of the flow battery based on the multiple energy state information, setting multiple to-be-scheduled targets, multi-objective optimization according to the multiple to-be-scheduled targets, formulating a fuzzy control strategy according to the optimization result; Dynamic updating of the multiple to-be-scheduled targets by executing the fuzzy control strategy, generation of a set of regulation and control signals, bidirectional collaborative scheduling of the flow battery based on the set of regulation and control signals, and generation of a scheduling result; Reverse tracing according to the scheduling result, incremental learning of the fuzzy control strategy according to a tracing path, generation of a fuzzy control optimization strategy, replacement of the fuzzy control strategy, updating of the set of regulation and control signals, and obtaining of multi-level regulation and control instructions for intelligent scheduling of the energy of the flow battery; The method for generating a fuzzy control optimization strategy comprises: Setting a historical scheduling period according to the scheduling result, and collecting execution result data of the historical scheduling period; Bias analysis based on the execution result data in combination with the fuzzy control strategy, generation of an execution bias parameter; Constructing a reverse tracing path according to the execution bias parameter, bias positioning according to the reverse tracing path, and obtaining of a tracing defect positioning point; Incremental learning of the fuzzy control strategy based on the tracing defect positioning point, and generation of the fuzzy control optimization strategy.

2. The energy intelligent scheduling method of flow battery of claim 1, wherein, The method for scheduling analysis of the flow battery based on the multiple energy state information, setting multiple to-be-scheduled targets, multi-objective optimization according to the multiple to-be-scheduled targets, formulating a fuzzy control strategy according to the optimization result, comprises: Temperature analysis based on the electrolyte state of charge information, charge and discharge analysis of the flow battery according to an operating temperature parameter, and setting of a power loss target; Performing attenuation analysis based on the health state information, performing battery life analysis on the flow battery according to the health attenuation rate, and setting a battery life target; Mapping the power loss target and the battery life target to a unified decision space, and setting a plurality of to-be-scheduled targets; Setting an operating interval constraint based on the electrolyte state of charge information, and performing attenuation rate constraint based on the health state information; Performing calculation on the plurality of to-be-scheduled targets based on the operating interval constraint and the attenuation rate constraint, generating a target constraint satisfaction degree, and performing descending order sorting according to the target constraint satisfaction degree, and generating a target dominance order; Performing multi-objective optimization on the plurality of to-be-scheduled targets based on the target dominance order, and formulating the fuzzy control strategy according to the optimization result.

3. The energy intelligent scheduling method of flow battery of claim 2, wherein, Performing multi-objective optimization on the plurality of to-be-scheduled targets based on the target dominance order, and formulating the fuzzy control strategy according to the optimization result, the method comprising: Performing priority layering on the plurality of to-be-scheduled targets based on the target dominance order rule, and generating a target dominance relationship graph; Performing multi-objective conflict analysis according to the target dominance relationship graph, generating a conflict path, and performing optimization resolution on the plurality of to-be-scheduled targets according to the conflict path, and generating an optimization result; Performing weight distribution based on the optimization result, obtaining a plurality of priority weight coefficients, and performing analysis on the optimization result according to the plurality of priority weight coefficients, and generating a dominant target intensity; Adding the dominant target intensity to the fuzzy control strategy.

4. The energy intelligent scheduling method of flow battery of claim 3, wherein, Performing dynamic update on the plurality of to-be-scheduled targets by executing the fuzzy control strategy, generating a set of regulation and control signals, performing bidirectional collaborative scheduling on the flow battery based on the set of regulation and control signals, and generating a scheduling result, the method comprising: Performing dynamic update on the priority weight coefficients of the plurality of to-be-scheduled targets according to the grid operating condition change parameters by executing the fuzzy control strategy, and generating a plurality of target weight coefficients; Performing power analysis on the charging and discharging of the flow battery based on the plurality of target weight coefficients, and generating real-time charging and discharging power signals; Performing flow compensation analysis on the temperature distribution of the flow battery based on the plurality of target weight coefficients, and generating electrolyte flow regulation signals; Performing electro-liquid bidirectional time sequence collaborative optimization on the flow battery according to the real-time charging and discharging power signals and the electrolyte flow regulation signals, and obtaining time sequence collaborative optimization parameters; Performing dynamic scheduling on the flow battery according to the time sequence collaborative optimization parameters, and generating the scheduling result.

5. The energy intelligent scheduling method of flow battery of claim 4, wherein, Performing dynamic update on the priority weight coefficients of the plurality of to-be-scheduled targets according to the grid operating condition change parameters by executing the fuzzy control strategy, and generating a plurality of target weight coefficients, the method comprising: Recording environmental conditions based on the energy operating parameters of the flow battery, and obtaining battery operating condition change parameters; Performing detection on the plurality of to-be-scheduled targets according to the battery operating condition change parameters by executing the fuzzy control strategy, and generating a detection result, the detection result being a power loss mutation result and / or a battery life mutation result; When the detection result is the power loss mutation result, the priority weight of the power loss target is raised to the highest level, and a first target weight coefficient is determined; When the detection result is the battery life mutation result, the priority weight of the battery life target is raised to the highest level, and a second target weight coefficient is determined. When the detection result is the battery life mutation result, the priority weight of the battery life target is raised to the highest level, and a second target weight coefficient is determined; When the detection result is the power loss mutation result and the battery life mutation result, then the power loss mutation result and the battery life mutation result are analyzed for scheduling influence, and a third target weight coefficient is determined according to the influence factor.

6. The energy intelligent scheduling method of flow battery of claim 1, wherein, Based on the traceability defect positioning point, the fuzzy control strategy is incrementally learned to generate the fuzzy control optimization strategy, the method comprising: Based on the traceability defect positioning point, defect analysis is performed to determine the electro-hydraulic coordination timing defect point and the life target defect point; Based on the electro-hydraulic coordination timing defect point, power reduction testing is performed to generate a first test result for confidence analysis, obtaining a first confidence degree; Based on the life target defect point, attenuation increment testing is performed to generate a second test result for confidence analysis, obtaining a second confidence degree; According to the first test result combined with the first confidence degree, the second test result combined with the second confidence degree, the fuzzy control strategy is incrementally learned to generate the fuzzy control optimization strategy.

7. The energy intelligent scheduling method of flow battery of claim 1, wherein, The fuzzy control optimization strategy replaces the fuzzy control strategy, updates the set of control signals, and obtains multi-level control instructions for intelligent scheduling of the energy of the flow battery, the method comprising: Verifying whether the fuzzy control optimization strategy can eliminate the execution bias parameter of the fuzzy control strategy, generating a verification result; When the verification result is that the fuzzy control optimization strategy can eliminate the execution bias parameter of the fuzzy control strategy, a replacement instruction is generated, the fuzzy control optimization strategy replaces the fuzzy control strategy through the replacement instruction, and a signal cascade reconstruction space is constructed; The set of control signals is mapped to the signal cascade reconstruction space for reconstruction compensation to generate a set of control reconstruction signals; Based on the set of control reconstruction signals, scheduling analysis is performed to divide a plurality of scheduling levels, the plurality of scheduling levels are matched with the set of control reconstruction signals, and the multi-level control instructions are generated.

8. An energy intelligent scheduling system for a flow battery, characterized in that, A method for implementing the energy intelligent scheduling of the flow battery according to any one of claims 1-7, the system comprising: An energy state information acquisition module for real-time acquisition of energy operation parameters of the flow battery, dynamic estimation of electrolyte state according to the energy operation parameters, and obtaining a plurality of energy state information; A fuzzy control strategy acquisition module for scheduling analysis of the flow battery based on the plurality of energy state information, setting a plurality of to-be-scheduled targets, multi-objective optimization according to the plurality of to-be-scheduled targets, and formulating a fuzzy control strategy according to the optimization result; A scheduling result acquisition module for dynamic updating of the plurality of to-be-scheduled targets by executing the fuzzy control strategy, generating a set of control signals, and bidirectional collaborative scheduling of the flow battery based on the set of control signals to generate a scheduling result. The energy intelligent scheduling execution module is configured to perform reverse tracing according to the scheduling result, perform incremental learning on the fuzzy control strategy according to a tracing path, generate a fuzzy control optimization strategy to replace the fuzzy control strategy, update the set of control signals, and obtain multi-stage control instructions to intelligently schedule energy of the flow battery.

Images