A micro-grid hybrid energy storage optimization method

Abstract

The application belongs to the technical field of micro-grid, and specifically discloses a micro-grid hybrid energy storage optimization method, which comprises the following steps: collecting a direct-current bus power imbalance in real time, decomposing the direct-current bus power imbalance into high-frequency and low-frequency components through adaptive filtering, forming a reference power, collecting photovoltaic actual power data, counting the duration, total length and frequency of over-limit, determining the type and severity of photovoltaic fluctuation based on the above data and preset rules, querying a power distribution mapping table according to the fluctuation characteristics, determining the power distribution strategy of the battery and the super capacitor, controlling the energy storage to perform operation, collecting actual power response data to analyze the tracking accuracy and response timeliness, dynamically adjusting the strategy according to the analysis result, and forming a closed-loop optimization; the application realizes accurate perception of photovoltaic fluctuation characteristics and optimization adaptation of energy storage resources, effectively improves the renewable energy consumption capacity and system operation stability, and significantly prolongs the overall life of the hybrid energy storage device.

Description

Technical Field

[0001] This invention belongs to the field of microgrid technology and relates to an optimization method for hybrid energy storage in microgrids. Background Technology

[0002] A microgrid is a small-scale power generation and distribution system that integrates distributed generation units, energy storage devices, loads, and monitoring and protection devices. It can operate in grid-connected or islanded mode and plays a crucial role in improving the absorption capacity of renewable energy and enhancing power supply reliability and resilience. However, with the high proportion of renewable energy integration, the intermittent and random nature of distributed power sources such as photovoltaics and wind turbines leads to frequent power fluctuations within the microgrid, and the problem of DC bus power imbalance is becoming increasingly prominent. Since a single energy storage technology cannot simultaneously meet the system's requirements for high energy density and high power density, hybrid energy storage devices are currently commonly used. These devices combine energy-type energy storage components, such as lithium-ion battery energy storage devices, with power-type energy storage components, such as supercapacitor energy storage devices, to maintain system power stability through complementary characteristics.

[0003] For example, Chinese invention patent CN109390926B discloses an optimization method for hybrid energy storage equipment in DC microgrids. This method decomposes the power to be smoothed into low-frequency and high-frequency components through a low-pass filter, which are then smoothed by battery energy storage equipment and supercapacitor energy storage equipment respectively. The filtering time constant T is adjusted in real time according to the power deviation range between the generator end and the load end to optimize power distribution and improve system stability.

[0004] The existing technologies mentioned above have the following shortcomings: 1. The current main method of selecting the time constant is based on a preset power range, which is essentially a rule-based, open-loop adjustment method. It cannot identify and respond to the inherent characteristics of photovoltaic fluctuations in real time, and therefore it is difficult to make optimal decisions when facing complex, variable, and atypical fluctuation scenarios. The filtering accuracy and adaptability are limited.

[0005] 2. The current optimization goal focuses on obtaining the stable domain and initial allocation, without constructing a closed loop for online monitoring and evaluation of the actual response performance of energy storage equipment and dynamic adjustment of power allocation strategy based on the evaluation results. This leads to a mismatch between theoretical allocation and actual equipment response capability, which will accelerate the aging of energy storage equipment or affect the actual effect of power smoothing under long-term operation. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background technology, a microgrid hybrid energy storage optimization method is proposed.

[0007] The objective of this invention can be achieved through the following technical solution: This invention provides a microgrid hybrid energy storage optimization method, including: S1, real-time acquisition of the power imbalance of the microgrid DC bus, and decomposition of the power imbalance into low-frequency components and high-frequency components through a preset adaptive filtering method, thereby forming a reference power sequence for battery energy storage devices and supercapacitor energy storage devices respectively.

[0008] S2. Within the preset observation period, collect the actual power data of the photovoltaic power generation system, and count the cases where it exceeds the reference power sequence to obtain the over-limit duration, total over-limit duration and number of over-limit occurrences.

[0009] S3. Based on the over-limit duration, total over-limit duration and number of over-limit occurrences, and in conjunction with the preset photovoltaic fluctuation judgment rules, determine the type and severity level of photovoltaic fluctuations.

[0010] S4. Based on the type and severity level of photovoltaic fluctuations, determine the power allocation strategy for battery energy storage devices and supercapacitor energy storage devices according to the preset power allocation mapping table.

[0011] S5. Control the energy storage device to perform power distribution operations and collect the power response data of the energy storage device within a preset period, and perform power tracking accuracy analysis and response timeliness analysis accordingly.

[0012] S6. Based on the analysis results of power tracking accuracy and response timeliness, dynamically adjust the power allocation strategy.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention achieves accurate perception and classification of fluctuation characteristics by measuring the duration of the over-limit, the total duration of the over-limit, and the number of times the over-limit occurs, and by determining the type and severity level of the fluctuation according to preset rules. Thus, the power allocation strategy is adaptively adjusted according to the fluctuation characteristics, which significantly enhances the adaptability and allocation accuracy of complex fluctuation scenarios.

[0014] (2) This invention collects power response data of energy storage devices, performs tracking accuracy and response timeliness analysis, and dynamically adjusts filtering parameters or power allocation strategies based on the analysis results to form a closed-loop optimization mechanism, which significantly improves the control accuracy of the system and realizes online monitoring of system performance and dynamic self-healing of strategies.

[0015] (3) By accurately matching the fluctuation characteristics with the response characteristics of the energy storage element and adjusting them based on actual performance, this invention avoids the frequent response of battery energy storage devices to high-frequency fluctuations and the overcharging and over-discharging of supercapacitor energy storage devices, and makes reasonable use of the advantages of both, significantly extending the life of energy storage devices, effectively extending the overall life of hybrid energy storage devices, and fundamentally reducing the operation and maintenance costs of the system.

[0016] (4) This invention uses an adaptive optimization method to quickly and accurately suppress power imbalance caused by renewable energy fluctuations and load changes, maintain bus voltage stability, effectively improve the power quality and operational stability of the microgrid, and ensure the safe, reliable and efficient operation of the system. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram showing the connections between the steps of the method of the present invention.

[0019] Figure 2 This is a schematic diagram showing the connection steps of the adaptive filtering method of the present invention.

[0020] Figure 3 This is a schematic diagram showing the connection steps of the dynamic adjustment of the power allocation strategy of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 As shown, the present invention provides a microgrid hybrid energy storage optimization method, which includes: S1, real-time acquisition of the power imbalance of the microgrid DC bus, and decomposition of the power imbalance into low-frequency components and high-frequency components through a preset adaptive filtering method, thereby forming a reference power sequence for battery energy storage devices and supercapacitor energy storage devices respectively.

[0023] The low-frequency component refers to a power component that changes relatively slowly and lasts for a long time, mainly caused by factors such as slow load changes and the diurnal cycle trend of photovoltaic power. The timescale of this component's change is typically from several seconds to tens of minutes, which matches the high energy density of battery energy storage devices and their suitability for medium- and long-term, large-capacity charge and discharge response characteristics.

[0024] The high-frequency component refers to the rapidly fluctuating power component, mainly caused by impulsive factors such as instantaneous changes in solar radiation and sudden load switching. The timescale of this component's change is typically between milliseconds and several seconds, which matches the high power density, fast response speed, and suitability for instantaneous power throughput of supercapacitor energy storage devices.

[0025] The adaptive filtering method is used to distribute low-frequency power components to battery energy storage devices for smoothing and high-frequency power components to supercapacitor energy storage devices for smoothing, thereby achieving a reasonable and efficient power distribution within the hybrid energy storage system based on response characteristics.

[0026] Please see Figure 2 As shown, exemplarily, the preset adaptive filtering method includes: inputting the DC bus power imbalance into a low-pass filter based on the least mean square adaptive algorithm, outputting a low-frequency power component, and subtracting the DC bus power imbalance from the low-frequency power component to obtain a high-frequency power component.

[0027] It should be added that the low-pass filter based on the least mean square adaptive algorithm can automatically decompose the DC bus power imbalance into frequency components suitable for different energy storage devices. The specific process is as follows: First, the DC bus power imbalance signal is used as the target signal and input into an adaptive digital filter model. The filtering characteristics of this model are determined by internally adjustable weighting coefficients. Simultaneously, the target signal is passed through a low-pass filter with a fixed cutoff frequency to generate an auxiliary reference signal containing the main low-frequency trend.

[0028] The auxiliary reference signal is input into the adaptive filter, and the output of the adaptive filter is the decomposed low-frequency power component. This component changes slowly and is suitable for battery energy storage devices. The difference between the output of the adaptive filter and the original target signal is the high-frequency power component. This component changes rapidly and drastically and is suitable for supercapacitor energy storage devices.

[0029] The system continuously calculates the magnitude of the above difference and adjusts the weighting coefficients inside the adaptive filter in real time based on the goal of minimizing the average energy of the difference. This enables the filter to dynamically track low-frequency changes in the power signal and achieve adaptive separation of high- and low-frequency components.

[0030] Finally, the stable filter output signal is used as the reference power command for the battery energy storage device, and the difference between the output signal and the original target signal is used as the reference power command for the supercapacitor energy storage device.

[0031] The absolute value of the high-frequency power component is continuously monitored. If the absolute value does not exceed the filtering convergence error threshold within a preset continuous observation period, the filtering is determined to be converged. From the moment the filtering convergence is determined, the low-frequency power component sequence and the high-frequency power component sequence obtained by decomposition are used as the reference power sequences of the battery energy storage device and the supercapacitor energy storage device, respectively.

[0032] It should be added that the preset continuous observation period refers to the time interval set based on the time scale characteristics of the DC bus power fluctuation of the microgrid, used to continuously verify the stability of the high-frequency power components, while avoiding misjudgment caused by the instantaneous filtering error meeting the standard, and ensuring that the output reference power sequence can stably match the DC bus power fluctuation characteristics. This period consists of several filtering sampling periods.

[0033] The preset continuous observation period is determined based on the power fluctuation characteristics of the microgrid's DC bus, the characteristics of the filtering algorithm, and the response characteristics of the energy storage device. The specific steps are as follows: Collect historical operating data of the microgrid, analyze the time-scale characteristics of the DC bus power imbalance fluctuations, and distinguish the typical duration patterns of high-frequency and low-frequency fluctuations. Since high-frequency fluctuations have stricter real-time requirements for filter convergence, the preset continuous observation period must prioritize covering the minimum fluctuation stabilization time, meaning it must be able to completely capture the entire process of a high-frequency fluctuation from its occurrence to its stabilization. Based on this, the period is set to several typical high-frequency fluctuation durations to ensure that the period length covers the stabilization process of the high-frequency fluctuation and avoid misjudgments due to missing the stabilization stage caused by an excessively short period.

[0034] The filter convergence error threshold is a pre-set allowable power error value used to determine whether the adaptive filtering algorithm has reached a stable convergence state. It represents the maximum acceptable deviation between the theoretical power and the actual power of the filtered output. When the filter error remains below this threshold, it indicates that the filter's tracking of the current power fluctuation characteristics is accurate enough, and its output is reliable and usable.

[0035] The method for obtaining the filter convergence error threshold includes the following steps: collecting historical operating data and analyzing the long-term statistical distribution of the adaptive filter error signal during stable system operation. The threshold is set above the high percentile of this statistical distribution to ensure successful convergence under most stable operating conditions.

[0036] If the filtering convergence condition is not met, the least mean square adaptive algorithm continues to run, adjusting the internal parameters of the filter according to the high-frequency power component, and repeating the signal decomposition and convergence determination process until the filtering convergence condition is met.

[0037] S2. Within the preset observation period, collect the actual power data of the photovoltaic power generation system, and count the cases where it exceeds the reference power sequence to obtain the over-limit duration, total over-limit duration and number of over-limit occurrences.

[0038] It should be added that the preset observation period is a fixed time window set for accurately statistically analyzing the fluctuation characteristics of the actual power of the photovoltaic power generation system exceeding the benchmark reference power sequence. Its core function is to provide a statistically sufficient and reliable data basis for subsequent determination of the type and severity of photovoltaic fluctuations. This period needs to be long enough to capture the complete fluctuation process and avoid feature omissions or misjudgments due to excessively short duration. At the same time, it should not be too long to prevent data redundancy and ensure the real-time power control of the hybrid energy storage system.

[0039] The specific value of the preset observation period needs to be determined by comprehensively considering the fluctuation characteristics of photovoltaic power in the microgrid, the update frequency of the benchmark reference power sequence, and the control response speed of the energy storage system. The specific setting steps are as follows: First, analyze the fluctuation characteristics of photovoltaic power in historical operating data to identify the main time scales of its short-term and long-term fluctuations. Then, set the observation period to cover the duration of a typical short-term fluctuation process, usually including several complete power over-limit events to ensure statistical validity. Finally, combine this with fine-tuning of the energy storage system control cycle to ensure that the statistical results are applied in a timely manner to the generation and adjustment of power allocation strategies.

[0040] For example, the statistical analysis of cases where the actual power data exceeds the reference power sequence includes: performing time axis calibration on the actual power data and the reference power sequence to ensure that the sampling frequency, observation period start and end points of the two are completely consistent, so as to avoid statistical deviations caused by time misalignment.

[0041] When the instantaneous power value at a certain moment in the actual power data of the photovoltaic power generation system after time axis calibration is greater than the instantaneous power value at the corresponding moment in the reference power sequence, it is determined to be a single over-limit event triggered, and monitoring is initiated.

[0042] If the actual power values ​​of the subsequent M-1 consecutive sampling points are all greater than the reference power value, then the time of the first out-of-limit sampling point is determined as the out-of-limit start time; otherwise, the trigger is determined to be an instantaneous interference, and the monitoring status is reset.

[0043] It should be added that parameter M is the effective trigger threshold for over-limit events. Its core function is to suppress transient interference caused by sensor noise or instantaneous irradiance fluctuations, ensuring that an over-limit event is only determined and monitoring is initiated when all M subsequent consecutive sampling points exceed the limit, thus avoiding false judgments. The value of M needs to be compatible with the photovoltaic power sampling frequency and determined based on the typical duration of transient interference in the microgrid's historical operating data. It is usually set to an integer value slightly larger than the number of sampling points corresponding to this typical duration.

[0044] When a single over-limit event is triggered, if the actual power value of a certain sampling point is detected to be less than or equal to the reference power value, and the subsequent N-1 consecutive sampling points all meet this condition, then the time of the first sampling point that meets this condition is determined as the over-limit termination time.

[0045] It should be added that parameter N is the threshold for determining the termination of an over-limit event. Its core function is to confirm that the photovoltaic power has stably recovered to a non-over-limit state, preventing premature termination of the over-limit event due to brief fluctuations in power near the critical value, thereby accurately counting the duration of a single over-limit event. The value of N needs to be determined comprehensively based on the photovoltaic power sampling interval and the typical stabilization time after the power recovers from the over-limit state. It is usually set to an integer value that matches the number of sampling points corresponding to the typical stabilization time, and generally N≥M.

[0046] The duration of a single over-limit event is obtained by subtracting the timestamps corresponding to the over-limit termination time and the over-limit start time.

[0047] The total duration of each individual exceeding the limit within the preset observation period is calculated by summing up the durations.

[0048] If there are at least P consecutive sampling points in a non-over-limit state between two over-limit events, they are determined to be two independent over-limit events, corresponding to the number of over-limit occurrences for the two events. Otherwise, they are determined to be continuous fluctuations of the same over-limit event, recorded as 1 over-limit occurrence.

[0049] It should be added that parameter P is the threshold for classifying independent over-limit events. Its core function is to distinguish between two independent over-limit events and continuous fluctuations within the same event. If there are at least P consecutive sampling points in a non-over-limit state between two over-limit events, they are judged as two independent events; otherwise, they are recorded as the same event, thus ensuring the accuracy of the over-limit occurrence statistics. The value of P is determined based on the typical intermittent characteristics of photovoltaic fluctuations. It is set by analyzing the minimum interval between two independent fluctuations in historical operating data and converting it into the corresponding number of sampling points.

[0050] The number of all independent out-of-limit events within the preset observation period is counted and used as the number of out-of-limit occurrences.

[0051] S3. Based on the over-limit duration, total over-limit duration and number of over-limit occurrences, and in conjunction with the preset photovoltaic fluctuation judgment rules, determine the type and severity level of photovoltaic fluctuations.

[0052] For example, determining the type of photovoltaic fluctuation includes comparing the over-limit duration, total over-limit duration, and number of over-limit occurrences with the corresponding thresholds in the preset photovoltaic fluctuation determination rules.

[0053] If the number of times the limit is exceeded is greater than the first threshold, the duration is less than the first time threshold, and the total duration of the limit is less than the first total duration threshold, then it is determined to be a high-frequency drastic fluctuation.

[0054] If the number of times the limit is exceeded is less than the second threshold, the duration is greater than the second time threshold, and the total duration of the limit exceeds the second total duration threshold, then it is determined to be a low-frequency continuous fluctuation; otherwise, it is determined to be a mixed fluctuation. In this case, the first number threshold is greater than the second number threshold, and the first time threshold is less than the second time threshold.

[0055] It should be added that the various thresholds in the preset photovoltaic fluctuation judgment rules, including but not limited to the first count threshold, the first time threshold, the first total duration threshold, the second count threshold, the second time threshold, and the second total duration threshold, are preset based on historical operating data, system characteristics, and target optimization requirements. The acquisition methods include, but are not limited to, the following steps: collecting historical photovoltaic power data and corresponding DC bus power data of the microgrid under typical operating scenarios to form a training dataset.

[0056] Statistical distribution analysis was conducted on the duration, number of occurrences, and total duration of all out-of-limit events in the dataset.

[0057] Based on the statistical distribution results and combined with the response characteristics of batteries and supercapacitors, a quantitative boundary for distinguishing between high-frequency fluctuations, low-frequency fluctuations, and mixed fluctuations was determined.

[0058] Based on the quantitative boundary, a microgrid model is constructed in the simulation platform for testing and verification. The quantitative boundary is then optimized and calibrated based on the simulation results. Finally, the determined threshold group is written into the photovoltaic fluctuation judgment rule to obtain various thresholds.

[0059] For example, determining the severity level of photovoltaic fluctuations includes: calculating the over-limit magnitude for each over-limit event and comparing it with the corresponding benchmark reference power value to obtain the over-limit intensity ratio of each over-limit event.

[0060] The over-limit intensity ratio is matched with the over-limit intensity ratio range corresponding to each level to obtain the severity level of each over-limit event. The highest severity level among all over-limit events is taken as the severity level of photovoltaic fluctuation.

[0061] It should be added that the over-limit intensity ratio ranges corresponding to each level are pre-defined ranges of over-limit intensity ratios to quantify the impact of photovoltaic over-limit events on microgrids and hybrid energy storage systems. These ranges are based on microgrid operational stability requirements, the safety tolerance of energy storage equipment, and power control accuracy targets. They are the core quantitative standard for determining the severity of a single over-limit event as mild, moderate, or severe. This range follows the logic that the higher the over-limit intensity ratio, the more significant the impact on the system: the mild range corresponds to minimal over-limit impact, which can be mitigated by conventional strategies with no system fluctuation risk; the moderate range corresponds to a moderate over-limit impact, requiring adjustments to the energy storage power allocation ratio for mitigation, with a slight risk of system fluctuation; and the severe range corresponds to a significant over-limit impact, requiring the activation of emergency coordination strategies, otherwise, equipment protection or system instability may be triggered.

[0062] The steps for obtaining the over-limit intensity ratio range corresponding to each level are as follows: Collect historical operating data of the microgrid for at least one complete operating cycle, covering data from the photovoltaic side, the system side, and the energy storage side. Specifically, the photovoltaic side includes the time of occurrence of each over-limit event, the instantaneous over-limit amplitude, and the corresponding reference power value at that time; the system side includes the DC bus voltage fluctuation value and frequency deviation value at and after the occurrence of the over-limit event; and the energy storage side includes the battery or supercapacitor operating status data for the period corresponding to the over-limit event.

[0063] Multi-dimensional statistics were conducted on historical operating data to establish the correspondence between the over-limit intensity ratio and the system response and energy storage response. Specifically, this included the compliance rate of microgrid system stability indicators, the safe operation rate of energy storage equipment, and the success rate of conventional energy storage strategies in mitigating the impact of different over-limit intensity ratios.

[0064] Based on the core criteria of system risk-free operation, equipment safety, and effective strategies, the boundaries of each level range are defined. The mild range is selected from the range of over-limit intensity ratios where the system stability compliance rate, energy storage equipment safety operation rate, and conventional strategy mitigation success rate are all high. In this range, the impact of over-limit events on the system is negligible. The severe range is selected from the range of over-limit intensity ratios corresponding to low system stability compliance rate, low energy storage equipment safety operation rate, and low conventional strategy mitigation success rate. In this range, over-limit events are likely to cause system risks. The moderate range is between the mild and severe ranges, corresponding to the range of over-limit intensity ratios where the system stability compliance rate, energy storage equipment safety operation rate, and conventional strategy mitigation success rate are all at an intermediate level. In this range, strategies need to be adjusted to avoid risks.

[0065] It should be added that the core function of the severity level determination for photovoltaic fluctuations is to combine the statistical analysis of the quantity of over-limit events with the qualitative assessment, providing a more refined and adaptive decision-making basis for subsequent power allocation strategies. The severity level directly reflects the stress experienced by the energy storage device. High-frequency and high-severity fluctuations mean that the battery is undergoing frequent, high-current charge-discharge cycles, which will significantly accelerate its aging. By identifying high-severity events and recording their frequency and intensity, the system provides crucial data input for assessing battery health and optimizing charge-discharge strategies to extend battery life.

[0066] S4. Based on the type and severity level of photovoltaic fluctuations, determine the power allocation strategy for battery energy storage devices and supercapacitor energy storage devices according to the preset power allocation mapping table.

[0067] For example, the power allocation mapping table is obtained through the following steps: collecting historical operating data, and extracting the fluctuation type, severity level, and corresponding power commands of battery energy storage devices and supercapacitor energy storage devices for each historical operating period to form a training sample set.

[0068] Inductive learning is performed on the training sample set to establish a mapping relationship from fluctuation type and severity level to power allocation strategy, which serves as a power allocation mapping table.

[0069] It should be added that the power allocation mapping table constructed in this invention is the core structured decision-making carrier connecting photovoltaic fluctuation characteristics and hybrid energy storage control strategies. Its core lies in establishing a mapping relationship between photovoltaic fluctuation type, severity level and power allocation strategy, thereby encapsulating the optimal control experience verified in historical operation into a structured scheme that can be called in real time. When the microgrid is running, the system can quickly match and call the appropriate power command from the power allocation mapping table according to the current photovoltaic fluctuation type and severity level, without the need to perform complex algorithm iteration and parameter optimization in real time. It can achieve a precise response of the energy storage strategy at the moment when photovoltaic power changes cause source-load imbalance, effectively maintain the stability of the microgrid DC bus voltage and frequency balance, and avoid problems such as voltage over-limit and frequency oscillation caused by control delay.

[0070] For example, the power allocation strategy for determining the battery energy storage device and the supercapacitor energy storage device includes: inputting the type and severity level of photovoltaic fluctuations into a preset power allocation mapping table.

[0071] Based on the mapping relationship between fluctuation type, severity level and power allocation strategy, power allocation instructions for battery energy storage devices and supercapacitor energy storage devices are generated.

[0072] S5. Control the energy storage device to perform power distribution operations and collect the power response data of the energy storage device within a preset period, and perform power tracking accuracy analysis and response timeliness analysis accordingly.

[0073] For example, the power point tracking accuracy analysis includes: extracting the actual power of each monitoring point from the power response data and subtracting it from the theoretical power corresponding to the power allocation command to obtain the power deviation value of each monitoring point.

[0074] The power deviation value is compared with the preset allowable deviation range, and the number of monitoring points and the cumulative duration of the power deviation value exceeding the preset allowable deviation range are counted.

[0075] If the number of monitoring points whose power deviation exceeds the preset allowable deviation range is greater than the threshold for the number of monitoring points exceeding the limit or the cumulative duration is greater than the threshold for the duration exceeding the limit, then the power tracking accuracy is determined to be unqualified; otherwise, the power tracking accuracy is determined to be qualified.

[0076] For example, the timely response analysis includes: measuring the response delay time of the energy storage device and comparing it with a preset response delay threshold, wherein the response delay time refers to the time required for the actual power response to rise from zero to a preset proportion of the power control command rating.

[0077] If the response delay time is greater than the preset response delay threshold, the response timeliness is deemed unqualified; otherwise, the response timeliness is deemed qualified.

[0078] S6. Based on the analysis results of power tracking accuracy and response timeliness, dynamically adjust the power allocation strategy.

[0079] Please see Figure 3 As shown, exemplarily, the dynamic adjustment of the power allocation strategy includes: determining whether the power allocation strategy needs to be adjusted based on the analysis results of the power tracking accuracy and response timeliness; if it is determined that no adjustment is needed, then the current power allocation strategy is maintained and continues to operate.

[0080] It should be added that the criteria for determining whether the power allocation strategy needs to be adjusted are as follows: if both power tracking accuracy and response timeliness are satisfactory, then no adjustment is required; if either is unsatisfactory, then adjustment is required. The unsatisfactory items are those with unsatisfactory power tracking accuracy and those with unsatisfactory response timeliness.

[0081] If adjustment is deemed necessary, the source of the power distribution deviation is located based on the type and characteristics of the non-conformance. Specifically, if the power tracking accuracy is non-conforming, the source of the deviation is a mismatch in the power distribution ratio between energy storage devices. If the response timeliness is non-conforming, the source of the deviation is a lag in high-frequency separation of the filter or insufficient response of the supercapacitor energy storage device.

[0082] Based on the source of the deviation, an appropriate adjustment method is selected to perform targeted correction. After the correction is completed, the power response data of the energy storage device is re-acquired, and power tracking accuracy and response timeliness analysis are performed again. The adjustment effect is verified based on the pass / fail judgment results. The adjustment methods include: adjusting the time constant or cutoff frequency of the adaptive filtering method to optimize the allocation benchmark of high and low frequency power components; and correcting the power allocation ratio or power command limit between battery energy storage devices and supercapacitor energy storage devices in the power allocation rules.

[0083] If the verification result fails again, the above adjustment, data collection and verification process will be repeated until the power allocation strategy meets the preset qualification standard or the system safety operation strategy is triggered.

[0084] It should be added that the preset qualification standards are a set of quantitative criteria used to determine whether the dynamic adjustment process of the power allocation strategy has terminated, including requirements for both power tracking accuracy and response timeliness. Power tracking accuracy must simultaneously meet the following requirements: the number of out-of-limit monitoring points does not exceed a threshold for the number of out-of-limit monitoring points; and the cumulative out-of-limit duration does not exceed a threshold for out-of-limit duration. Response timeliness must meet the following requirements: the average response delay time does not exceed a delay threshold. These thresholds are determined by those skilled in the art based on the specific system.

[0085] The aforementioned safe operation strategy refers to a systematic protection mechanism activated when the power allocation strategy of a microgrid hybrid energy storage system, after multiple dynamic adjustments, still fails to simultaneously meet the requirements of power point tracking accuracy and response timeliness. This strategy, through a closed-loop logic comprised of emergency intervention, load grading protection, and fault tracing optimization, terminates ineffective adjustment cycles, avoids system risks caused by persistent power imbalances, and prioritizes ensuring the power supply stability of critical loads. Its core objective is to maintain basic system operational stability while ensuring equipment safety; that is, to temporarily abandon refined power optimization and instead adopt a fixed, low-risk operation mode to effectively prevent equipment failures and system oscillations.

[0086] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0087] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0088] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0089] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0090] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A microgrid hybrid energy storage optimization method, characterized in that: The method includes: S1. Real-time acquisition of power imbalance of the DC bus of the microgrid, and decomposition of the power imbalance into low-frequency and high-frequency components through a preset adaptive filtering method, thereby forming a reference power sequence for battery energy storage equipment and supercapacitor energy storage equipment respectively. S2. Within the preset observation period, collect the actual power data of the photovoltaic power generation system, count the cases where it exceeds the reference power sequence, and obtain the over-limit duration, total over-limit duration and number of over-limit occurrences; The statistical analysis of situations where the power exceeds the reference power sequence includes: performing time-axis calibration on the actual power data and the reference power sequence; when the instantaneous power value at a certain moment in the actual power data of the photovoltaic power generation system after time-axis calibration is greater than the instantaneous power value at the corresponding moment in the reference power sequence, it is determined as a single over-limit event trigger, and monitoring is initiated; if the actual power values ​​of the subsequent M-1 consecutive sampling points are all greater than the reference power value, the moment of the first over-limit sampling point is determined as the over-limit start moment; otherwise, this trigger is determined as an instantaneous interference, and the monitoring state is reset; after a single over-limit event is triggered, if the actual power value of a certain sampling point is detected to be less than or equal to the reference power... If the rate value is found to be true, and the subsequent N-1 consecutive sampling points all satisfy this condition, then the time of the first sampling point that satisfies this condition is determined as the over-limit termination time. The difference between the timestamps corresponding to the over-limit termination time and the over-limit start time is calculated to obtain the duration of a single over-limit event. The durations of all single over-limit events within the preset observation period are accumulated to obtain the total over-limit duration. If there are at least P consecutive sampling points in a non-over-limit state between two over-limit events, they are determined to be two independent over-limit events, corresponding to the number of over-limit events. Otherwise, they are determined to be a continuous fluctuation of the same over-limit event, recorded as 1 over-limit event. The number of all independent over-limit events within the preset observation period is counted and used as the over-limit event count. S3. Based on the over-limit duration, total over-limit duration, and number of over-limit occurrences, and in conjunction with the preset photovoltaic fluctuation judgment rules, determine the type and severity level of photovoltaic fluctuations; S4. Based on the type and severity level of photovoltaic fluctuations, determine the power allocation strategy for battery energy storage devices and supercapacitor energy storage devices according to the preset power allocation mapping table. S5. Control the energy storage device to perform power distribution operation and collect the power response data of the energy storage device within a preset period, and perform power tracking accuracy analysis and response timeliness analysis accordingly. S6. Based on the analysis results of power tracking accuracy and response timeliness, dynamically adjust the power allocation strategy.

2. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The preset adaptive filtering method includes: The DC bus power imbalance is input into a low-pass filter based on the least mean square adaptive algorithm, and the low-frequency power component is output. The DC bus power imbalance is then subtracted from the low-frequency power component to obtain the high-frequency power component. The absolute value of the high-frequency power component is continuously monitored. If the absolute value does not exceed the filtering convergence error threshold within the preset continuous observation period, the filtering is determined to be converged. From the moment of the filtering convergence determination, the low-frequency power component sequence and the high-frequency power component sequence obtained by decomposition are used as the reference power sequences of the battery energy storage device and the supercapacitor energy storage device, respectively. If the filtering convergence condition is not met, the least mean square adaptive algorithm continues to run, adjusting the internal parameters of the filter according to the high-frequency power component, and repeating the signal decomposition and convergence determination process until the filtering convergence condition is met.

3. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The types of photovoltaic fluctuations are defined as follows: The duration of the over-limit, the total duration of the over-limit, and the number of times the over-limit occurred are compared with the corresponding thresholds in the preset photovoltaic fluctuation judgment rules. If the number of times the limit is exceeded is greater than the first threshold, the duration is less than the first time threshold, and the total duration of the limit is less than the first total duration threshold, then it is determined to be a high-frequency and severe fluctuation. If the number of times the limit is exceeded is less than the second threshold, the duration is greater than the second time threshold, and the total duration of the limit exceeds the second total duration threshold, then it is determined to be a low-frequency continuous fluctuation; otherwise, it is determined to be a mixed fluctuation.

4. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The severity levels for determining photovoltaic fluctuations include: For each over-limit event, its over-limit amplitude is calculated and compared with the corresponding reference power value to obtain the over-limit intensity ratio of each over-limit event; The over-limit intensity ratio is matched with the over-limit intensity ratio range corresponding to each level to obtain the severity level of each over-limit event. The highest severity level among all over-limit events is taken as the severity level of photovoltaic fluctuation.

5. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The power allocation mapping table is obtained through the following steps: Collect historical operating data and extract the fluctuation type, severity level, and corresponding power commands of battery energy storage devices and supercapacitor energy storage devices for each historical operating period to form a training sample set; Inductive learning is performed on the training sample set to establish a mapping relationship from fluctuation type and severity level to power allocation strategy, which serves as a power allocation mapping table.

6. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The power allocation strategy for determining battery energy storage devices and supercapacitor energy storage devices includes: Input the type and severity level of photovoltaic fluctuations into the preset power allocation mapping table; Based on the mapping relationship between fluctuation type, severity level and power allocation strategy, power allocation instructions for battery energy storage devices and supercapacitor energy storage devices are generated.

7. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The power point tracking accuracy analysis includes: The actual power of each monitoring point is extracted from the power response data and subtracted from the theoretical power corresponding to the power allocation command to obtain the power deviation value of each monitoring point. The power deviation value is compared with the preset allowable deviation range, and the number of monitoring points and the cumulative duration of the power deviation value exceeding the preset allowable deviation range are counted. If the number of monitoring points whose power deviation exceeds the preset allowable deviation range is greater than the threshold for the number of monitoring points exceeding the limit or the cumulative duration is greater than the threshold for the duration exceeding the limit, then the power tracking accuracy is determined to be unqualified; otherwise, the power tracking accuracy is determined to be qualified.

8. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: Response timeliness analysis includes: Measure the response delay time of the energy storage device and compare it with a preset response delay threshold; If the response delay time is greater than the preset response delay threshold, the response timeliness is deemed unqualified; otherwise, the response timeliness is deemed qualified.

9. The microgrid hybrid energy storage optimization method according to claim 1, characterized in that: The dynamic power allocation adjustment strategy includes: Based on the analysis results of the power tracking accuracy and response timeliness, it is determined whether the power allocation strategy needs to be adjusted. If it is determined that no adjustment is needed, the current power allocation strategy will continue to operate. If adjustment is deemed necessary, the source of the power distribution deviation will be located based on the type and characteristics of the non-conformance. Based on the source of the deviation, select the appropriate adjustment method to perform targeted correction, and after the correction is completed, re-collect the power response data of the energy storage device, and perform power tracking accuracy analysis and response timeliness analysis again. Verify the adjustment effect based on the pass / fail judgment results. If the verification result fails again, the above adjustment, data collection and verification process will be repeated until the power allocation strategy meets the preset qualification standard or the system safety operation strategy is triggered.

Images