Alkaline seawater metal-air battery hybrid power supply predicted energy management method and system

Abstract

The application discloses a kind of alkaline seawater metal-air battery hybrid energy supply predicted energy management method and system, belong to marine equipment power management technical field, the method first collects the relevant state and multiple data information of hybrid energy supply system;Estimate current maximum output power and sustained power;Based on the occurred load power, task phase information and historical prediction residual, construct predicted load mean sequence and predicted load confidence upper bound sequence;According to the part of predicted load confidence upper bound exceeding sustained power, calculate predicted reserve demand, and generate predicted energy supplement lower limit when current available energy is insufficient;In the candidate power set satisfying constraint, according to multi-objective cost function, the current power instruction is determined by rolling. The application can improve the peak task energy supply guarantee capability under the condition of load prediction deviation and pulse load, and reduce anode high rate use, direct current bus voltage fluctuation and auxiliary energy storage cycle pressure.

Description

Technical Field

[0001] This invention relates to the field of power management for marine equipment, and in particular to a predictive energy management method and system for alkaline seawater metal-air battery hybrid power supply. Background Technology

[0002] Alkaline seawater metal-air batteries use dissolved oxygen from seawater as the cathode reactant and active metals (such as aluminum and magnesium) as the anode, offering advantages such as high energy density and good safety, making them suitable for underwater power supply scenarios. However, the output power of this type of battery is significantly affected by the residual state of the anode, the state of the electrolyte (flow rate, conductivity, pH value), temperature, and polarization state, making it difficult to be equivalent to a constant voltage and constant power source. Furthermore, high-rate discharge accelerates polarization and anode consumption, resulting in a slower response speed and difficulty in independently coping with transient and drastic load fluctuations.

[0003] Therefore, existing technologies typically employ alkaline seawater metal-air batteries and auxiliary energy storage such as lithium batteries or supercapacitors to form a hybrid energy supply system. Its energy management strategies mainly include: (1) a fixed ratio allocation strategy, which allocates power according to preset rules, but lacks adaptability; (2) a reactive threshold strategy, which triggers energy replenishment or discharge based on the current voltage or state of charge, but due to the lack of prediction of future load, it is prone to insufficient reserves when power increases sharply; (3) an energy replenishment strategy based on load average prediction, which introduces a prediction link, but does not quantify the risks brought about by prediction deviation, and may not be able to achieve a reasonable balance between power supply reliability and anode consumption and energy storage life under extreme operating conditions.

[0004] Based on the foregoing, there is a need for an energy management method that can establish linkages between battery state estimation, mission phase prediction, and reserve demand assessment, and coordinate multiple objectives such as power supply reliability, anode consumption, and energy storage lifetime. Summary of the Invention

[0005] The purpose of this invention is to provide a predictive energy management method and system for hybrid energy supply using alkaline seawater metal-air batteries, which addresses the shortcomings of existing technologies and improves the power supply reliability, energy utilization rate, and lifespan of each energy storage unit in the hybrid energy supply system.

[0006] To achieve the above objectives, this invention provides a method for predictive energy management of hybrid power supply using alkaline seawater metal-air batteries. This method is applied to a hybrid power supply system comprising an alkaline seawater metal-air battery module, an auxiliary energy storage module, a power regulation module, a DC bus, and a load module. The method includes the following steps: Step 1: Collect current load power, alkaline seawater metal-air battery module status, auxiliary energy storage module state of charge, DC bus voltage, electrolyte status parameters, temperature parameters, and task stage information; Step 2: Based on the state of the alkaline seawater metal-air battery module, the electrolyte state parameters, and the temperature parameters, estimate the current maximum output power and continuous power of the alkaline seawater metal-air battery module, and determine the current available energy of the auxiliary energy storage module based on the state of charge of the auxiliary energy storage module. Step 3: Based on the load power that has occurred, task phase information, and historical prediction residuals, construct the predicted load mean sequence and the predicted load confidence upper bound sequence within the future prediction window; Step 4: Based on the portion of the predicted load confidence upper bound sequence that exceeds the continuously borne power, calculate the predicted reserve demand of the auxiliary energy storage module, and compare the predicted reserve demand with the current available energy to obtain the reserve gap; Step 5: When there is a reserve gap and the alkaline seawater metal-air battery module has a power margin, generate a predicted lower limit for energy replenishment based on the reserve gap and the available energy replenishment window within the future prediction window. Step 6: From the candidate power set that meets the predicted lower limit of energy replenishment, the current maximum output power, the state of charge range of the auxiliary energy storage module, the charging and discharging power range of the auxiliary energy storage module, and the DC bus voltage range, the current power command of the alkaline seawater metal-air battery module and the auxiliary energy storage module is determined in a rolling manner according to the multi-objective cost function. Step 7: Execute the current power command and update the status to enter the next control cycle.

[0007] The present invention also provides an alkaline seawater metal-air battery hybrid power supply predictive energy management system for implementing the above method, comprising: The status acquisition module is used to collect current load power, status of alkaline seawater metal-air battery module, status of charge of auxiliary energy storage module, DC bus voltage, electrolyte status parameters, temperature parameters, and task stage information; The state estimation module is used to estimate the current maximum output power and continuous power of the alkaline seawater metal-air battery module based on the state of the alkaline seawater metal-air battery module, electrolyte state parameters and temperature parameters, and to determine the current available energy of the auxiliary energy storage module based on the state of charge of the auxiliary energy storage module. The task phase prediction module is used to construct the mean sequence of predicted load and the upper bound sequence of predicted load confidence within the future prediction window based on the load power that has occurred, task phase information and historical prediction residuals. The reserve assessment module is used to calculate the predicted reserve demand of the auxiliary energy storage module based on the portion of the predicted load confidence upper bound sequence that exceeds the continuously carried power, and compare the predicted reserve demand with the current available energy to obtain the reserve gap; The energy replenishment planning module is used to generate a predicted energy replenishment lower limit based on the reserve gap and the available energy replenishment window within the future prediction window when there is a reserve gap and the alkaline seawater metal-air battery module has a power margin. The rolling power allocation module is used to determine the current power command of the alkaline seawater metal-air battery module and the auxiliary energy storage module from the candidate power set that meets the predicted lower limit of energy replenishment, the current maximum output power, the state of charge range of the auxiliary energy storage module, the charging and discharging power range of the auxiliary energy storage module, and the DC bus voltage range, based on a multi-objective cost function, and then output the execution command.

[0008] Therefore, the alkaline seawater metal-air battery hybrid power supply prediction and energy management method and system of the present invention have the following beneficial effects: (1) This invention incorporates the residual state of the anode, the state of the electrolyte (flow rate / conductivity / pH value), the operating temperature and the polarization state into the output capacity estimation model of the alkaline seawater metal-air battery module, and determines the current maximum output power and the continuous power to be carried, so that the power allocation can reflect the coupled influence of metal anode consumption, seawater environment fluctuations and electrochemical polarization on the actual energy supply capacity of the battery, and avoids the overestimation of capacity (causing system protective shutdown) or underestimation of capacity (causing overcharging and discharging of auxiliary energy storage) caused by the fixed power model.

[0009] (2) Based on task phase information and historical prediction residuals, this invention constructs a predicted load mean sequence and a predicted load confidence upper bound sequence. It does not directly use the future real load, and calculates the auxiliary energy storage prediction reserve demand by integrating the difference between the predicted load confidence upper bound and the continuously undertaken power. It quantifies the load prediction uncertainty into a reserve constraint that can be evaluated online, and overcomes the defect of insufficient reserves when encountering load pulses or prediction deviations in existing prediction strategies based on instantaneous load or mean.

[0010] (3) When the available energy of the current auxiliary energy storage is insufficient to cover the predicted reserve demand, the present invention identifies the available energy replenishment window from the future prediction window and generates the predicted energy replenishment lower limit according to the reserve gap and the window length, so that the system can actively build up reserves for auxiliary energy storage before the peak load arrives. This changes the existing strategy of passively replenishing energy only when the current SOC is lower than the threshold, and reduces the risk of load power loss caused by reaction lag.

[0011] (4) The present invention uses the predicted lower limit of energy replenishment, the output capability boundary of alkaline seawater metal-air battery, the state of charge range of auxiliary energy storage, the charge and discharge power range and the DC bus voltage range to form a candidate power feasible domain. In the multi-objective cost function, the load satisfaction, bus voltage stability, reserve adequacy, high anode output cost, auxiliary energy storage cycle pressure and power fluctuation impact are jointly considered, so that the system can perform rolling coordination between power supply robustness and anode consumption and energy storage life, taking into account the differentiated control requirements under different task stages.

[0012] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the hybrid energy supply system structure in Embodiment 1 of the present invention; Figure 2 This is a flowchart of the confidence reserve prediction energy management method in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of state observation and output capability estimation in Embodiment 1 of the present invention; Figure 4 This is a schematic diagram of task phase prediction and load confidence upper bound construction in Embodiment 2 of the present invention; Figure 5 This is a schematic diagram illustrating the calculation and prediction of the reserve gap and the generation of the lower limit for energy replenishment in Embodiment 2 of the present invention; Figure 6 This is a schematic diagram of the constraint rolling power allocation and execution in Embodiment 3 of the present invention; Figure 7 This is a diagram of causal prediction and confidence reserve energy replenishment response in Embodiment 4 of the present invention, where a is the upper bound of causal prediction and confidence driven by the task stage, b is the pre-power allocation of predicted energy replenishment under low load, and c is the energy storage gap constraint and pre-replenishment response. Figure 8 This is a heatmap showing the robustness changes in multiple scenarios in Embodiment 4 of the present invention; Figure 9 This is a comparison chart of the multi-task statistical distribution in Embodiment 4 of the present invention, where a represents power supply reliability, b represents anode cost, c represents bus fluctuation, and d represents energy storage lifetime pressure. Figure 10 This is a dual-target distribution diagram of anode consumption and auxiliary energy storage cycle pressure in Embodiment 4 of the present invention; Figure 11 This is a comparison diagram of the high-frequency pulse power supply effect and anode cost in Embodiment 4 of the present invention, where a represents the unmet energy in the high-frequency pulse scenario, b represents the anode consumption in the high-frequency pulse scenario, and c represents the anode consumption. Detailed Implementation

[0014] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0016] Example 1: System Composition and State Estimation This embodiment provides an alkaline seawater metal-air battery hybrid power supply predictive energy management system, such as... Figure 1 and Figure 3 As shown. The system specifically includes: The status acquisition module is used to collect current load power, status of alkaline seawater metal-air battery module, status of charge of auxiliary energy storage module, DC bus voltage, electrolyte status parameters, temperature parameters, and task stage information; The state estimation module is used to estimate the current maximum output power and continuous power of the alkaline seawater metal-air battery module based on the state of the alkaline seawater metal-air battery module, electrolyte state parameters and temperature parameters, and to determine the current available energy of the auxiliary energy storage module based on the state of charge of the auxiliary energy storage module. The task phase prediction module is used to construct the mean sequence of predicted load and the upper bound sequence of predicted load confidence within the future prediction window based on the load power that has occurred, task phase information and historical prediction residuals. The reserve assessment module is used to calculate the predicted reserve demand of the auxiliary energy storage module based on the portion of the predicted load confidence upper bound sequence that exceeds the continuously carried power, and compare the predicted reserve demand with the current available energy to obtain the reserve gap; The energy replenishment planning module is used to generate a predicted energy replenishment lower limit based on the reserve gap and the available energy replenishment window within the future prediction window when there is a reserve gap and the alkaline seawater metal-air battery module has a power margin. The rolling power allocation module is used to determine the current power command of the alkaline seawater metal-air battery module and the auxiliary energy storage module from the candidate power set that meets the predicted lower limit of energy replenishment, the current maximum output power, the state of charge range of the auxiliary energy storage module, the charging and discharging power range of the auxiliary energy storage module, and the DC bus voltage range, based on a multi-objective cost function, and then output the execution command.

[0017] The alkaline seawater metal-air battery module is connected to the DC bus via a DC / DC converter to serve as a basic energy source for continuous basic load. The auxiliary energy storage module is directly connected to the DC bus or connected via a bidirectional DC / DC converter to provide transient peak power and absorb load power fluctuations. The power regulation module is used to achieve voltage matching and power control between each unit and the DC bus. The load module includes underwater electrical equipment such as thrusters, communication equipment, and sensors.

[0018] Figure 2 The complete control logic framework for the predictive energy management method using this system is presented, such as... Figure 2 As shown, within each control cycle, the method is executed in the following order: Step 1: Collect the current load power, alkaline seawater metal-air battery module status (including voltage, current, anode residual status and polarization status), auxiliary energy storage module state of charge, DC bus voltage, electrolyte state parameters (including flow rate, conductivity, pH value), temperature parameters and task stage information through the status acquisition module; Step 2: The state estimation module estimates the current maximum output power and continuous power of the alkaline seawater metal-air battery module based on the parameters collected in Step 1, and determines the current available energy based on the state of charge of the auxiliary energy storage module. Step 3: Based on the load power that has occurred, the task phase prediction module constructs the mean sequence of predicted load and the upper bound sequence of predicted load confidence within the future prediction window, according to the load power that has occurred, the task phase information, and the historical prediction residuals of the same task phase. Step 4: The reserve gap calculation module calculates the predicted reserve demand of the auxiliary energy storage module based on the portion of the predicted load confidence upper bound sequence that exceeds the continuously borne power, and compares the predicted reserve demand with the current available energy to obtain the reserve gap. Step 5: When there is a reserve gap and the alkaline seawater metal-air battery module has a power margin, the predictive energy replenishment module generates a predicted energy replenishment lower limit based on the reserve gap and the available energy replenishment window within the future prediction window. Step 6: The rolling power allocation module determines the current power command of the alkaline seawater metal-air battery module and the auxiliary energy storage module from the candidate power set that meets the predicted lower limit of energy replenishment, the current maximum output power, the state of charge range of the auxiliary energy storage module, the charging and discharging power range of the auxiliary energy storage module, and the DC bus voltage range, based on the multi-objective cost function. Step 7: Execute the current power command, update the system status, and return to step 1 in the next control cycle.

[0019] In this method, steps 3 to 5 constitute a "causal prediction and reserve planning" link. This link does not rely on the prediction of the actual future load, but forms a load confidence upper bound based on task stage information and historical prediction residuals, and then transforms it into a quantifiable reserve gap constraint. Step 6 further incorporates this constraint together with the source-side output boundary, energy storage operation boundary and bus voltage boundary into a rolling optimization framework, so that the power allocation result satisfies the timing logic of "first ensure reserves, then adjust allocation".

[0020] The status acquisition module collects the following parameters in real time: voltage and current of the alkaline seawater metal-air battery module; state of charge (SOC) of the auxiliary energy storage module; DC bus voltage; electrolyte flow rate, conductivity, and pH value; battery operating temperature; and current task stage information.

[0021] The state estimation module estimates the anode residual state based on the collected parameters, using power integration and state correction methods. and polarization state .

[0022] Anode residual state The remaining state is calculated based on the current integral and the Faraday reaction equivalent, combined with the total charge corresponding to the initial rated reaction equivalent at the anode. The calculation formula is as follows: ; ; in, The cumulative electricity consumption of the anode up to the current moment (unit: Ah). for The module output current at any given time (unit: A). The rated charge (unit: Ah) is the equivalent of the initial total reaction at the anode. The value ranges from 0 to 1, with values ​​closer to 1 indicating more sufficient remaining anodic reaction capacity. During the calculation, the integral results can be corrected by incorporating electrolyte conditions and temperature parameters to compensate for the effects of side reactions (hydrogen evolution) and temperature on the anodic consumption rate, thereby improving estimation accuracy.

[0023] polarization state The estimation method is as follows: an online identification method based on voltage deviation is adopted, with the equivalent polarization internal resistance characterizing the degree of polarization. During module operation, when the current changes... Greater than the preset threshold (In this embodiment, 0.5A is used) and the duration is greater than the preset stabilization time. (In this embodiment, 0.5s is used) The steady-state voltage deviation and steady-state current change in this interval are collected, and the equivalent polarization resistance is calculated using the following formula: ; in, The steady-state voltage deviation before and after the load change (unit: V). This represents the corresponding steady-state current change (unit: A). The equivalent polarization resistance is... Relative to the reference polarization internal resistance The ratio is normalized to obtain the polarization state. : ; in, The reference polarization resistance (unit: Ω) under rated operating conditions is obtained through offline calibration experiments. A larger value indicates more severe battery polarization and a more significant decrease in output capability. In this embodiment, This represents nonpolarization (ideal state). The closer to 0, the more severe the polarization. The identification process uses a sliding time window for recursive updates, and the window length can be set to 5 to 10 control cycles depending on the control cycle.

[0024] The current maximum output power of the alkaline seawater metal-air battery module was then calculated using the following formula. : ; in, Rated power, To estimate the effective state of the electrolyte, For temperature measurement, various correction functions , , , These functions are used to correct for the effects of residual anode, electrolyte state, temperature, and polarization on output capability. Each correction function outputs 1 when the state parameters are at their rated reference values, and outputs less than 1 when the state deteriorates.

[0025] Each correction function is implemented using a piecewise linear interpolation table: the discrete points of the correction coefficients corresponding to different state parameters are obtained in advance through calibration experiments, and the real-time correction values ​​are obtained through interpolation calculations during online operation. The anode residual state correction function is used as an example. For example: when hour, This indicates that there is no limit to the output power when there is sufficient anode remaining; when hour, Follow A linear decrease in anode consumption has a gradual impact on output capability; when hour, The accelerated decline reflects the sharp decrease in output capacity as the anode approaches depletion.

[0026] Temperature correction function Piecewise linear interpolation is used, with calibration points including low-temperature zones (e.g., 0℃, 10℃), rated temperature zones (e.g., 20~30℃), and high-temperature zones (e.g., 40℃, 50℃). The output is 1 within the rated temperature range and monotonically decreases when deviating from the rated range. Electrolyte state correction function. and polarization state correction function All are implemented using the same calibration method, with a rated value of 1, and a linear or piecewise linear decrease to 0 when the state deteriorates.

[0027] The state estimation module further determines the continuous power consumption using the following formula. : ; in, The preset continuous bearing coefficient is taken in this embodiment. This means that the alkaline seawater metal-air battery module should be considered to have a safe continuous output limit of 76% of its maximum output power during long-term operation.

[0028] Current available energy of the auxiliary energy storage module Determined by the following formula: ; in, The rated energy of the auxiliary energy storage module (85Wh in this embodiment) is used. The current state of charge, This is the limit of the state of charge (0.18 in this embodiment).

[0029] Example 2: Mission Phase Confidence Prediction and Reserve Gap Assessment like Figure 4 and Figure 5 As shown, the task phase prediction module constructs a load prediction sequence within the future prediction window based on the already occurred load power, the pre-stored task phase plan, and the historical prediction residuals of the same task phase. In this embodiment, the prediction window length is 20 minutes, and the control period is 20 seconds.

[0030] The predicted load mean sequence is constructed using the following formula: ; in, For the first The base load for each predicted time point corresponds to the task phase (obtained by looking up the task phase plan table). This represents the correction value for historical forecast residuals during the same task phase.

[0031] The upper bound sequence of predicted load confidence is constructed by the following formula: ; in, The historical forecast residual scale The confidence coefficient is preset. In this embodiment, we take... This corresponds to a 95% confidence level. The controller does not read future actual loads throughout the forecasting process, relying solely on task phase information and historical statistical characteristics.

[0032] The reserve assessment module calculates the predicted reserve requirement of the auxiliary energy storage module based on the portion of the predicted load confidence upper bound that exceeds the continuously carried power. ; in, For the first Upper bound of the prediction load confidence at each prediction time; For the first The continuous power handling capacity of the alkaline seawater metal-air battery module at each predicted moment; To control the cycle; This is a reserve for uncertainty, used to cover the additional energy gap caused by random fluctuations in the forecast residuals. Its specific calculation formula is as follows: ; In the formula, The confidence coefficient is set to 2.05 in this embodiment. For the first The historical prediction residual scale corresponding to each prediction time; This represents the total number of control cycles contained within the future prediction window. It is determined based on the confidence coefficient, historical forecast residual scale, forecast window length, and control period, and is matched with the relevant duration of the forecast residual.

[0033] The reserve assessment module further compares the projected reserve demand with the current available energy to obtain the reserve gap: ; ; when When this occurs, it indicates that the currently available auxiliary energy storage is insufficient to cover the peak load gap within the future forecast window.

[0034] Example 3: Predictive Energy Supplement Lower Limit Generation and Rolling Power Allocation when Furthermore, when the current maximum output power of the alkaline seawater metal-air battery module is greater than the current load power, the energy replenishment planning module identifies a usable energy replenishment window from the future prediction window. The usable energy replenishment window refers to a continuous period of time in which the predicted average load is lower than the continuously borne power.

[0035] The energy replenishment planning module is based on the reserve gap. and available energy replenishment window length Determine the power replenishment And the maximum charging power of the auxiliary energy storage module. Limit the power of the replenishment and generate a predicted lower limit for the replenishment. .in, This represents the upper limit of the replenishment power after the limit is set, which is the predicted lower limit constraint that the auxiliary energy storage charging power must meet in subsequent rolling optimization.

[0036] like Figure 6 As shown, the rolling power allocation module constructs a candidate power set within each control cycle. Each candidate power set includes the candidate output power of the alkaline seawater metal-air battery module. Candidate charge / discharge power of auxiliary energy storage modules And it satisfies all of the following constraints: (1) Lower limit constraint for predicted energy replenishment: ; when At that time, this constraint requires that auxiliary energy storage operate at a rate not lower than [a certain value] within the replenishment window. Charging is performed using the power available. When there is no reserve gap (i.e....) )hour, This constraint does not impose any practical limitations.

[0037] (2) Output capacity constraints of alkaline seawater metal-air battery module: ; (3) Auxiliary energy storage state of charge constraints: ; in, For based on Predicted state of charge for the next control cycle.

[0038] (4) Auxiliary energy storage charging and discharging power constraints: ; in, Indicates charging. This indicates a discharge.

[0039] (5) DC bus voltage constraint: ; The rolling power allocation module selects the candidate power combination that minimizes the cost function value from the candidate power set that satisfies all the above constraints, based on the multi-objective cost function, as the current power command: ; in, The candidate power feasible region is defined for all the above constraints.

[0040] In this embodiment, the multi-objective cost function is expressed in a weighted summation form: ; in, The weighting coefficients for the corresponding penalty items can be adjusted according to the control priority of the task scenario to meet the requirements. and .

[0041] The specific calculation methods for each penalty item are as follows: (1) The load does not meet the power penalty term : The formula for calculating the penalty for situations where the load power demand cannot be fully met is as follows: ; in, This represents the current load power. The absolute value of the candidate discharge power (only when...) (Valid for a period of time) This is an indicator function (this term is effective only when the energy storage is discharging; it is 0 when the energy storage is charging). This penalty term amplifies the impact of the power deficit in a squared form, guiding the controller to prioritize meeting load demands. It should be noted that this term reflects the steady-state supply capacity assessment; transient coordination is constrained by the DC bus voltage. ensure.

[0042] (2) DC bus voltage deviation penalty item : The formula used to penalize deviations of the DC bus voltage from its rated reference value is as follows: : in, For the predicted DC bus voltage of the next control cycle, This is the rated reference voltage for the DC bus. The square form of the relative deviation is used to ensure comparability between systems of different voltage levels.

[0043] (3) Penalty for insufficient auxiliary energy storage reserves : The formula for calculating the penalty for auxiliary energy storage conditions where available energy is lower than predicted storage demand is as follows: ; in, To predict reserve demand, The available energy for the auxiliary energy storage module is predicted for the next control cycle. A squared form of the relative gap is used to avoid difficulties in uniformly adjusting weights due to differences in energy storage capacity.

[0044] (4) High-rate output penalty for alkaline seawater metal-air batteries : This is used to penalize batteries for prolonged operation in high-rate ranges, reducing anode consumption and exacerbating polarization. The calculation formula is as follows: ; in, This is the continuous power handling capability of the alkaline seawater metal-air battery module. This penalty only applies when the candidate output power is higher than the continuous power handling capability; it is zero when the output power is lower than or equal to the continuous power handling capability.

[0045] (5) Auxiliary energy storage cycle pressure penalty item : The formula used to penalize the charging / discharging switching and large power fluctuations of the auxiliary energy storage module, thereby extending its cycle life, is as follows: ; in, This represents the actual output power of the auxiliary energy storage module in the previous control cycle. This penalty term is used to supplement the rated power of the energy storage module. It generates a large penalty value when the energy storage power changes significantly or the charging / discharging direction switches, thereby suppressing unnecessary cycle depth and extending the energy storage life.

[0046] (6) Power change impact penalty item : Used to penalize drastic jumps in total power command between adjacent control cycles, thus improving system stability, the calculation formula is as follows: ; in, and These represent the actual output power of the alkaline seawater metal-air battery module and the auxiliary energy storage module, respectively, in the previous control cycle. This is the rated power of the load. This penalty term, in the form of the square of the relative rate of change of total power, suppresses sudden changes in the total output power of the system and improves the smoothness of the DC bus power balance.

[0047] In each control cycle, the rolling power allocation module traverses all candidate power combinations in the candidate power set and calculates the corresponding multi-objective cost function value. Select to make Minimum candidate power combination Power command as the current control cycle The command is then issued and executed. The status is re-acquired in the next control cycle, and the above process is repeated to achieve rolling optimization control.

[0048] It should be noted that the above This is the multi-objective cost function. When Time indicates charging, The time indicates discharge, and further rolling optimization is underway. As a hard constraint, the remaining penalty terms are coordinated by soft constraints through weight coefficients, and together they complete the rolling optimization of power commands in the candidate power set.

[0049] Example 4: Simulation Verification This embodiment verifies the technical effectiveness of the method of the present invention through Monte Carlo simulation.

[0050] (1) Simulation model and operating environment: This simulation was conducted based on a digital twin model of an alkaline seawater metal-air battery hybrid power supply system. The model includes an electrochemical module for the alkaline seawater metal-air battery, an auxiliary energy storage charging / discharging module, a DC bus circuit module, a load timing generation module, and four types of energy management strategy control modules. The simulation was also conducted using a discrete-time model of the alkaline seawater metal-air battery hybrid power supply system built in Python. This model includes a metal-air battery output capacity module, an auxiliary energy storage charging / discharging module, a DC bus module, a load timing generation module, and four types of energy management strategy control modules. The simulation was performed discretely with a fixed control period of 20 seconds and a prediction window of 60 control periods.

[0051] (2) System simulation parameters: The alkaline seawater metal-air battery module has a rated power of 104W, an equivalent energy of 520Wh, a maximum power variation limit of 36W per control cycle, and a rated DC bus voltage of 24V. The current maximum output power is determined by a combination of factors including the remaining anode state, the effective electrolyte state, temperature, and polarization state. The continuous power output is set at 76% of the current maximum output power.

[0052] The auxiliary energy storage module has a rated energy of 85Wh, an initial state of charge (SOC) of 70%, an allowable SOC range of 18% to 94%, a maximum normal discharge power of 132W, and a maximum charging power of 62W. Controlled short-term rate discharge is permitted when the SOC exceeds the safety lower limit and there is a short-term peak demand.

[0053] (3) Task scenario setting: Four mission scenarios were constructed: a conventional hybrid mission, a high-frequency communication pulse mission, an electrolyte fluctuation mission, and a low anode margin mission, each lasting 4 hours. The mission load consisted of the basic load of the mission phase, periodic fluctuations, time-dependent random residuals, and random pulses superimposed on each scenario. The conventional hybrid mission was used to characterize the combined operation of the cruise, maneuver, inspection, and communication phases; the high-frequency communication pulse scenario increased the probability and amplitude of random pulses; the electrolyte fluctuation scenario reduced the effective state of the electrolyte and superimposed time-varying disturbances; and the low anode margin mission set the initial anode residual state to 0.66 to simulate the decrease in source-side output capability in the later stages of the mission.

[0054] (4) Comparison strategy: To verify the technical effectiveness of the method of the present invention, the method of the present invention (confidence reserve energy replenishment strategy) is compared with the following three strategies: Fixed ratio allocation strategy: The metal-air battery outputs a fixed 60% of its rated power, and any shortfall or excess is balanced by auxiliary energy storage without prediction or optimization.

[0055] Reactive threshold strategy: Control is performed based on the current load and the current state of charge of the auxiliary energy storage; when the state of charge is below 0.42, the output of the metal-air battery is increased; when the state of charge is between 0.42 and 0.55 and the current load is low, energy replenishment is performed; when the current load is close to the upper limit of the output capacity, the output of the source side is increased, and the future prediction window is not used.

[0056] Point-based predictive power replenishment strategy: based on load mean prediction (excluding confidence upper bound) and uncertainty reserves ( ) When the predicted average load is lower than the continuous power consumption, supplemental power is provided.

[0057] Under the same scenario and the same random seed, the four strategies use the same load sequence, electrolyte change sequence, temperature change sequence and initial scenario conditions, only the control rules are different.

[0058] (5) Evaluation indicators and statistical methods: The calculation methods for each evaluation indicator are as follows: Load fulfillment rate: The proportion of the actual load energy fulfilled to the total load energy demand of the task; Anode consumption: The proportion of the cumulative reaction consumption at the anode at the end of the simulation to the initial total reaction equivalent; Bus voltage standard deviation: The standard deviation of the DC bus voltage sampling values ​​throughout the simulation; Equivalent Cycle Pressure (EFC): A normalized cycle index calculated by combining auxiliary energy storage charge and discharge flux, high-rate operation, and low-charge state pressure.

[0059] For each scenario, 24 random seeds are set to generate differentiated load fluctuations, electrolyte disturbances, and initial state deviations. The final result is the statistical average of all samples in the same scenario.

[0060] (6) Simulation results: The statistical results of the 96 tasks (4 types of scenarios × 24 random seeds) are shown in Table 1 below: Table 1: Simulation Results

[0061] Combination Figures 7 to 11 As shown in Table 1, among the overall average results of 96 tasks in four task scenarios, the load satisfaction rate of the confidence reserve replenishment strategy is 97.20%, the anode consumption is 53.86%, the standard deviation of bus voltage is 0.1221V, and the equivalent cycle pressure of auxiliary energy storage is 0.2659EFC.

[0062] Compared to the reactive threshold strategy, the method of this invention improves the load fulfillment rate by 4.05 percentage points, reduces anode consumption by 17.82 percentage points, reduces the bus voltage standard deviation by 0.1960V, and reduces the equivalent cycle pressure of auxiliary energy storage by 0.1869EFC. This is because the reactive threshold strategy only triggers energy replenishment after the auxiliary energy storage's state of charge decreases, while the method of this invention pre-establishes a reserve gap based on the predicted upper bound of the load confidence and performs pre-replenishment using a low load window.

[0063] Compared to the point-prediction energy replenishment strategy, the method of this invention improves the load fulfillment rate by 3.39 percentage points, reduces the overall average unmet energy from 16.23Wh to 7.27Wh, reduces anode consumption by 13.94 percentage points, and reduces the standard deviation of bus voltage by 0.1561V. These results demonstrate that the confidence upper bound and uncertainty reserve can convert historical forecast residuals into additional reserve requirements, reducing the risk of insufficient reserves caused by underestimating point forecasts.

[0064] In high-frequency communication pulse scenarios, the unmet energy of the method of this invention is 7.06Wh, the standard deviation of the bus voltage is 0.15V, and the anode consumption is 56.79%; the corresponding values ​​for the fixed ratio strategy are 17.96Wh, 0.41V, and 66.03%, respectively. This indicates that controlled short-time rate discharge, predicted energy replenishment, and bus voltage compensation can jointly reduce peak power shortage and bus fluctuations.

[0065] In tasks with low anode margin, the current maximum output power and continuous power handling capacity of the metal-air battery module still constitute the main constraints, and the method of this invention cannot break through the physical output boundary of the source side.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A predictive energy management method for hybrid power supply using alkaline seawater metal-air batteries, applied to a hybrid power supply system comprising an alkaline seawater metal-air battery module, an auxiliary energy storage module, a power regulation module, a DC bus, and a load module, characterized in that, Includes the following steps: Step 1: Collect current load power, alkaline seawater metal-air battery module status, auxiliary energy storage module state of charge, DC bus voltage, electrolyte status parameters, temperature parameters, and task stage information; Step 2: Based on the state of the alkaline seawater metal-air battery module, the electrolyte state parameters, and the temperature parameters, estimate the current maximum output power and continuous power of the alkaline seawater metal-air battery module, and determine the current available energy of the auxiliary energy storage module based on the state of charge of the auxiliary energy storage module. Step 3: Based on the load power that has occurred, task phase information, and historical prediction residuals, construct the predicted load mean sequence and the predicted load confidence upper bound sequence within the future prediction window; Step 4: Based on the portion of the predicted load confidence upper bound sequence that exceeds the continuously borne power, calculate the predicted reserve demand of the auxiliary energy storage module, and compare the predicted reserve demand with the current available energy to obtain the reserve gap; Step 5: When there is a reserve gap and the alkaline seawater metal-air battery module has a power margin, generate a predicted lower limit for energy replenishment based on the reserve gap and the available energy replenishment window within the future prediction window. Step 6: From the candidate power set that meets the predicted lower limit of energy replenishment, the current maximum output power, the state of charge range of the auxiliary energy storage module, the charging and discharging power range of the auxiliary energy storage module, and the DC bus voltage range, the current power command of the alkaline seawater metal-air battery module and the auxiliary energy storage module is determined in a rolling manner according to the multi-objective cost function. Step 7: Execute the current power command and update the status to enter the next control cycle.

2. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 1, characterized in that: In step 1, the state of the alkaline seawater metal-air battery module includes at least the anodic residual state and the polarized state.

3. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 2, characterized in that, In step 2, the current maximum output power of the alkaline seawater metal-air battery module is estimated. This can be achieved through the following formula: ； in, Rated power; To estimate the remaining state of the anode; To estimate the effective state of the electrolyte; For measuring temperature; To estimate the polarization state; , , and These are the correction functions for the anode residual state, electrolyte state, temperature, and polarization state, respectively. Continuous power satisfy: ； Where α is the preset continuous bearing coefficient.

4. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 3, characterized in that, Step 3 involves constructing the mean forecast load sequence and the upper bound confidence sequence for the forecast load within the future forecast window, including: The predicted load mean sequence satisfies: ； The predicted load confidence upper bound sequence satisfies: ； in, For the first The predicted average load at each time point; For the first time within the future prediction window The base load for each predicted time point corresponds to the task phase. These are the historical forecast residual correction values ​​for the same task phase; For the first The upper bound of the predicted load confidence at any given moment; The historical forecast residual scale, For pre-set confidence coefficients.

5. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 4, characterized in that, In step 4, the predicted storage demand of the auxiliary energy storage module is calculated. , represented as: ； in, For the first Upper bound of the prediction load confidence at each prediction time; For the first The continuous power handling capacity of the alkaline seawater metal-air battery module at each predicted time point; To control the cycle; This is to reserve the uncertainty based on the historical forecast residual scale, confidence coefficient, and future forecast window length.

6. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 5, characterized in that, In step 4, the currently available energy satisfy: ； in, To supplement the rated energy of the auxiliary energy storage module, To assist in the current state of charge of the energy storage module, This is the limit of the state of charge; Reserve gap Represented as: 。 7. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 6, characterized in that, Step 5 specifically involves: when Furthermore, when the current maximum output power of the alkaline seawater metal-air battery module is greater than the current load power, the replenishment power is determined based on the reserve gap and the length of the available replenishment window within the future prediction window, and the replenishment power is limited by the maximum charging power of the auxiliary energy storage module to generate the predicted replenishment lower limit.

8. The method for predictive energy management of alkaline seawater metal-air battery hybrid power supply according to claim 7, characterized in that: In step 6, the multi-objective cost function includes at least two of the following: load power unmet penalty, DC bus voltage deviation penalty, auxiliary energy storage insufficient reserve penalty, alkaline seawater metal-air battery high-rate output penalty, auxiliary energy storage cycle pressure penalty, and power change impact penalty.

9. An alkaline seawater metal-air battery hybrid power supply predictive energy management system, used to implement the method according to any one of claims 1-8, characterized in that, include: The status acquisition module is used to collect current load power, status of alkaline seawater metal-air battery module, status of charge of auxiliary energy storage module, DC bus voltage, electrolyte status parameters, temperature parameters, and task stage information; The state estimation module is used to estimate the current maximum output power and continuous power of the alkaline seawater metal-air battery module based on the state of the alkaline seawater metal-air battery module, electrolyte state parameters and temperature parameters, and to determine the current available energy of the auxiliary energy storage module based on the state of charge of the auxiliary energy storage module. The task phase prediction module is used to construct the mean sequence of predicted load and the upper bound sequence of predicted load confidence within the future prediction window based on the load power that has occurred, task phase information and historical prediction residuals. The reserve assessment module is used to calculate the predicted reserve demand of the auxiliary energy storage module based on the portion of the predicted load confidence upper bound sequence that exceeds the continuously carried power, and compare the predicted reserve demand with the current available energy to obtain the reserve gap; The energy replenishment planning module is used to generate a predicted energy replenishment lower limit based on the reserve gap and the available energy replenishment window within the future prediction window when there is a reserve gap and the alkaline seawater metal-air battery module has a power margin. The rolling power allocation module is used to determine the current power command of the alkaline seawater metal-air battery module and the auxiliary energy storage module from the candidate power set that meets the predicted lower limit of energy replenishment, the current maximum output power, the state of charge range of the auxiliary energy storage module, the charging and discharging power range of the auxiliary energy storage module, and the DC bus voltage range, based on a multi-objective cost function, and then output the execution command.

10. The alkaline seawater metal-air battery hybrid power supply predictive energy management system according to claim 9, characterized in that: The state of the alkaline seawater metal-air battery module includes at least the anodic residual state and the polarization state; the state estimation module estimates the current maximum output power using the following formula: ； and through Determine the continuous power to be borne, among which, Rated power; To estimate the remaining state of the anode; To estimate the effective state of the electrolyte; For measuring temperature; To estimate the polarization state; , , and These are the anode residual state correction function, electrolyte state correction function, temperature correction function, and polarization state correction function, respectively, with α being the preset continuous bearing coefficient.

Images