Low-power-consumption storage access control method based on mobile power supply board

Abstract

The invention discloses a low-power-consumption storage access control method based on a power panel of a mobile power supply, and relates to the technical field of energy management and power scheduling, and the method comprises the steps: S1, obtaining the access frequency resolution of each storage region, determining the time sequence distribution of an access request, obtaining the trend collection window data of each region, and storing the trend collection window data in a database; s2, pre-establishing a mapping relationship between time difference characteristics and power distribution, and judging a dynamic adjustment demand of recording an electric quantity change value once per second at an electric quantity sampling interval, S3, obtaining the dynamic adjustment demand from the mapping relationship, adjusting different request time periods displayed by the access frequency resolution, and determining an adjusted power distribution time sequence; the low-power-consumption storage access control method based on the mobile power supply board does not need extra complex hardware support, is suitable for a mobile power supply platform, an embedded system or a multi-task Internet of Things terminal, can be widely applied to energy sensitive electronic equipment, and has good engineering realizability and popularization value.

Description

Technical Field

[0001] The present invention relates to the technical field of energy management and power supply scheduling, and particularly to a low-power storage access control method based on a mobile power supply board. Background Art

[0002] Energy management technology plays a crucial role in modern electronic devices and intelligent systems, directly affecting the operating efficiency, service life, and environmental adaptability of the devices. With the widespread application of the Internet of Things, mobile devices, and renewable energy systems, how to achieve efficient and long-lasting operation under limited energy conditions has become a key issue that cannot be ignored in this field. The design of power management strategies not only needs to meet performance requirements but also find a balance between energy consumption and functionality, which forces researchers to explore innovative technical paths to cope with increasingly complex application scenarios.

[0003] Currently, many power management methods mainly rely on static allocation or simple dynamic adjustment strategies, such as controlling the sleep and wake-up of devices through fixed thresholds. However, these methods often ignore the dynamic change characteristics of the access frequencies of different storage areas in the system, resulting in low energy utilization efficiency. Especially in high-load or long-term operation scenarios, the power consumption is too fast to meet the demand for continuous operation. In addition, the existing solutions have limited resource scheduling capabilities when dealing with multi-task parallelism, and fail to fully utilize the time difference characteristics of the hardware, further exacerbating the energy consumption problem. In this context, the core challenges faced in the research field have gradually emerged, and the most critical technical factors include the time difference characteristics of the access frequencies of storage areas, the implementation difficulty of time division multiplexing, and the stability of long-term power operation. Due to the time difference characteristics of the access frequencies not being fully exploited, the system cannot dynamically adjust power allocation according to the actual load, resulting in resource waste or performance bottlenecks. Although the introduction of time division multiplexing can theoretically improve efficiency, its complexity in hardware implementation and compatibility issues with existing architectures have become technical problems that need to be solved urgently. In addition, the power management for long-term operation needs to balance dynamic load and energy reserve, and this contradiction further deepens the research difficulty. Summary of the Invention

[0004] The purpose of the present invention is to provide a low-power storage access control method based on a mobile power supply board, overcoming the limitations of hardware complexity and compatibility to ensure long-term stable operation of the power.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A low-power storage access control method based on a mobile power supply board, including:

[0006] S1. Obtain the access frequency resolution of each storage area, determine the time series distribution of access requests, and obtain the trend acquisition window data of each area;

[0007] S2. Pre - establish the mapping relationship between the time - difference characteristic and power distribution, and judge the dynamic adjustment requirement of recording the power change value every second for the power sampling interval;

[0008] S3. Obtain the dynamic adjustment requirement from the mapping relationship, adjust the access - frequency resolution for different request time periods, and determine the adjusted power - distribution time series;

[0009] S4. Use time - division multiplexing technology to divide time slices for the power supply of multiple storage areas, and obtain the stability - detection period for the power distribution of each area;

[0010] S5. Obtain the power - switching frequency. If the switching frequency exceeds the hardware response threshold, then determine the optimized time - division multiplexing scheduling scheme by extending or shortening the time slice;

[0011] S6. According to the optimized time - division multiplexing scheduling scheme, obtain the real - time readings of the data - consumption source from the hardware power - consumption monitoring interface, and get the scheduling time series after compatibility adjustment;

[0012] S7. According to the scheduling time series after compatibility adjustment, judge whether the continuous operation for 72 hours meets the minimum power requirement for the operation - duration threshold, and get the optimized time series for power management;

[0013] S8. According to the time series of power management, obtain the synchronous acquisition of the environmental variables (recording temperature and load impact on power) from the hardware interface. If the data - storage format shows that the abnormal - trigger condition continuously exceeds the change - threshold range for 5 minutes in the power - log file marked with time stamps, then adjust the time - slice allocation ratio to determine the final power - management time series.

[0014] Preferably, S1 includes:

[0015] Obtain the access - request data of each storage area, determine the access frequency through time - series analysis, extract the distribution characteristics from the time series to get the access - frequency resolution of each area. For the distribution characteristics, use the sliding - window method to determine the trend - acquisition window. Through the trend - acquisition window, obtain the area data covering the past 24 hours. If the access frequency exceeds the preset threshold, then perform frequency analysis on the storage area to get the change trend. According to the change trend, adjust the area division to determine the optimized access - request distribution, and use the K - means algorithm to cluster the adjusted distribution to obtain the trend characteristics of each area.

[0016] Preferably, S2 includes:

[0017] By analyzing the time difference characteristics, data of the sampling interval is obtained from the power sampling, the recording frequency per second is determined, according to the change of the sampling interval, the adjustment of the recording frequency is adopted, and the real-time value of the power change is obtained. For the real-time value of the power change, through the pre-constructed mapping relationship, the corresponding mode of power distribution is judged. If the mode of power distribution exceeds the preset threshold of the distribution model, the sampling interval is dynamically adjusted to obtain a new change value. According to the new change value, the key points of the adjustment requirements are extracted from the judgment basis, the direction of dynamic adjustment is determined, and through the direction of dynamic adjustment, the support vector machine algorithm is used to obtain the optimized power distribution scheme. For the change trend of the optimized scheme, the logistic regression algorithm is used to judge the final result of the adjustment requirements.

[0018] Preferably, the S3 includes:

[0019] Obtain the data in the mapping relationship, analyze the dynamic adjustment requirements, obtain the change trend of the access frequency. Through the change trend of the access frequency, the resolution display method is adopted to distinguish different request time periods, and the boundary of time period adjustment is determined. For the boundary of the request time period, the corresponding load data is extracted from the storage area, and the adjustment amplitude of the power supply is judged. If the adjustment amplitude of the power supply exceeds the preset threshold, the power distribution scheme after time period adjustment is calculated through the frequency resolution algorithm, and the adjusted distribution time series is obtained. According to the adjusted distribution time series, the real-time state of the storage area is obtained, the new configuration of the power supply is determined, and through the newly configured power supply, the feedback data in the mapping relationship is analyzed to judge the execution effect of the dynamic adjustment, and the updated adjustment requirements are obtained. The updated adjustment requirements are used to combine with the change of the access frequency to determine the power distribution time series of the next cycle.

[0020] Preferably, the S4 includes:

[0021] The time slices of the power supply are divided by the time division multiplexing technology, the power supply time series data of each storage area is obtained, and the preliminary area division result is obtained. The distribution characteristics of the power supply are extracted from the area division result, and the preset threshold is used to judge the stability of the supply control, and the power supply balance state of each area is determined. For the power supply balance state, the statistical value of the power fluctuation within the detection period is calculated to obtain the distribution range of the fluctuation amplitude. According to the distribution range of the fluctuation amplitude, the time series characteristics of the change situation are analyzed to judge whether there is an abnormal fluctuation mode. If the abnormal fluctuation mode exists, by adjusting the time slot ratio of the time slice division, the optimized supply control parameters are obtained, and the new distribution stable state is obtained. The periodic analysis data is extracted from the new distribution stable state, and the support vector machine algorithm is used to classify the power fluctuation to determine the long-term trend of the fluctuation amplitude. Through the long-term trend analysis result, the boundary conditions of the area division are updated to obtain the final optimized stability detection period scheme.

[0022] Preferably, S5 includes:

[0023] Obtain power switching frequency data, get the switching frequency by measuring the number of switches within a unit time. If the switching frequency exceeds the preset hardware response threshold, then judge the exceeding state by comparing the frequency value with the threshold. According to the exceeding state, obtain the adjustment length by calculating the ratio of the time slice length to the switching frequency. Adopt the adjustment length to determine the new time slice configuration by extending or shortening the time slice length. Through the new time slice configuration, combine the time division multiplexing technology to obtain an optimized scheduling scheme. For the optimized scheduling scheme, judge whether the frequency still exceeds the threshold by simulating the running data. If the frequency still exceeds the threshold, then obtain the final scheduling scheme by iteratively adjusting the time slice length.

[0024] Preferably, S6 includes:

[0025] Collect real-time data through the hardware power consumption monitoring interface, generate an initial scheduling scheme based on the time division multiplexing technology to obtain a preliminary time series. Extract real-time readings from the preliminary time series. Adopt the compatibility adjustment rule to judge whether the data source meets the requirements of the optimization scheme. If it meets, then retain the current series. If it does not meet, then adjust the series parameters to obtain an adjusted time series. For the adjusted time series, obtain the power consumption monitoring data, analyze the change trend of the hardware power consumption to determine the power consumption distribution characteristics. According to the power consumption distribution characteristics, adopt the support vector machine algorithm to classify the high-power consumption interval and the low-power consumption interval in the time series to obtain the classification result. Through the classification result, adjust the parameters of the time division multiplexing scheduling scheme to generate an optimized scheduling sequence. Obtain the optimized scheduling sequence, combine the real-time data to judge whether the sequence meets the hardware power consumption constraint conditions. If it meets, then output the final time series. If it does not meet, then return to the step of adjusting the series parameters to reprocess. Extract the adjusted sequence from the final time series, and combine the data source to verify the consistency of the real-time readings to obtain a scheduling time series after compatibility verification.

[0026] Preferably, S7 includes:

[0027] Obtain the data after compatibility adjustment from the scheduling time series, determine the adjustment result through time series analysis. Adopt the adjustment result to extract the running duration data, judge whether the continuous running reaches the requirement of 72 hours, and obtain the threshold matching status by comparing the running duration with the threshold setting.

[0028] Preferably, S7 further includes:

[0029] If the threshold matching status indicates insufficiency, extract the power management parameters from the lowest power data, determine the optimized time series, adjust the management series according to the optimized time series to obtain the preliminary series of power management, perform a matching judgment between the preliminary series and the lowest power requirement to determine the final optimized time series, and use the final optimized time series to update the time series data to complete the generation of the power management series.

[0030] Preferably, the S8 includes:

[0031] Collect environmental variable data through the hardware interface, record the temperature and load impacts, generate a time series, process the time series using synchronous acquisition technology, mark it with timestamps and store it in the log file. If the abnormal trigger condition in the log file exceeds the change threshold continuously for 5 minutes, determine that the time series is abnormal. According to the abnormal trigger result, calculate the adjustment value of the time slice allocation ratio to obtain the optimized allocation plan, update the time series through the optimized allocation plan to determine the preliminary management series, obtain the comparison data between the preliminary management series and the original time series, judge the adjustment consistency, and analyze the consistency data using the regression algorithm to determine the final power management time series.

[0032] As can be seen from the above technical solutions, the present invention has the following beneficial effects:

[0033] The low-power storage access control method based on the power board of the mobile power supply can accurately reflect the access characteristics of different regions at different time periods by obtaining the access frequency resolution and access request time series of each storage area and combining with the trend acquisition window, realizing the refined control of power distribution and improving the energy utilization efficiency. By pre-establishing the mapping relationship between the access frequency and power distribution and combining the power sampling data to dynamically judge the adjustment requirements, it realizes the adaptive power control based on the actual power consumption trend and reduces the system overload risk. By dividing the power supply of multiple storage areas into time slices and combining the power switch frequency detection and hardware response threshold judgment to dynamically adjust the time slice length, it optimizes the scheduling efficiency and system compatibility of the time division multiplexing scheme. By constructing the scheduling time series and real-time monitoring the power consumption and environmental variables, the system can achieve the adaptive optimization of the continuous operation state of the power (such as 72 hours), improving the continuous operation ability of the system and the robustness of power scheduling. Combining the synchronous acquisition of environmental temperature, load changes and power logs, it can automatically adjust the time slice allocation ratio when an abnormal condition is triggered (such as the power fluctuation exceeding the threshold within 5 minutes), enhancing the intelligent response ability and environmental adaptability of the system. This method does not require additional complex hardware support, is applicable to mobile power platforms, embedded systems or multi-task IoT terminals, and can be widely used in energy-sensitive electronic devices, having good engineering feasibility and promotion value. Description of the Drawings

[0034] Figure 1This is the flowchart of the method of the present invention. Detailed implementation manners

[0035] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0036] As Figure 1 shown, the present invention provides a technical solution: a low-power storage access control method based on a power board of a mobile power source, including:

[0037] S1. Obtain the access frequency resolution of each storage area, determine the time series distribution of access requests, and obtain the trend acquisition window data of each area;

[0038] S2. Establish a mapping relationship between the time difference characteristic and the power supply allocation in advance, and judge the dynamic adjustment requirement for recording the power change value every second at the power sampling interval;

[0039] S3. Obtain the dynamic adjustment requirement from the mapping relationship, adjust the access frequency resolution to display different request time periods, and determine the adjusted power supply allocation time series;

[0040] S4. Divide the time slices of the power supply to multiple storage areas through time-division multiplexing technology to obtain the stability detection period of the power supply allocation for each area;

[0041] S5. Obtain the power supply switching frequency. If the switching frequency exceeds the hardware response threshold, determine the optimized time-division multiplexing scheduling scheme by extending or shortening the time slice;

[0042] S6. According to the optimized time-division multiplexing scheduling scheme, obtain the real-time readings of the consumption data source from the hardware power consumption monitoring interface to obtain the scheduling time series after compatibility adjustment;

[0043] S7. According to the scheduling time series after compatibility adjustment, judge whether the continuous operation for 72 hours of the operation duration threshold meets the minimum power requirement to obtain the optimized time series of power management;

[0044] S8. According to the time series of power management, obtain the synchronous acquisition of the environmental variables recorded by the temperature and load on the power from the hardware interface. If the data storage format shows that the abnormal trigger condition continuously exceeds the change threshold range for 5 minutes in the power log file marked with a timestamp, adjust the time slice allocation ratio to determine the final power management time series.

[0045] This method evaluates the request activity of different storage areas by using access frequency resolution, and analyzes the time difference characteristics by sampling the power change value to construct the mapping relationship between power distribution and access behavior. By dynamically adjusting the access frequency and power time series, the adaptability to power consumption is improved. The time-division multiplexing technology is applied to the power supply control of multiple regions, making the power supply behavior predictable and stable. The hardware response threshold limits the power switching frequency. If it exceeds the standard, the time slice allocation is optimized to ensure the stable operation of the system. The power consumption monitoring interface provides real-time data sources to support the compatibility correction of the scheduling scheme. In the continuous operation state, the system judges whether the power can support the operation time of more than 72 hours, so as to optimize the power distribution time series. Finally, the impact of power is analyzed in combination with environmental variables (such as temperature and load). If the abnormal situation exceeding the fluctuation threshold is continuously detected by the log monitoring, the time slice ratio is adjusted to finally determine a more stable and energy-saving power management strategy.

[0046] For example, in practical applications, an embedded device is equipped with a 128GB eMMC and a 512MB RAM module. During the operation of the device, the storage control unit finds that the access frequency in the eMMC area is significantly higher than other periods from 8:00 to 10:00 and from 19:00 to 21:00 every day. According to this rule, the system increases the power supply frequency during this period, and increases the power output from an average of 30% to 60% through the mapping algorithm. At the same time, the access of the RAM shows periodic fluctuations. According to the current change value recorded by the power sampling module (such as INA219 current sensor) per second, the power supply cycle is adjusted to avoid data errors caused by insufficient power supply during the peak period.

[0047] The method of the present invention realizes the precise control of the power supply behavior of the mobile power supply board through multi-dimensional data fusion analysis. On the premise of ensuring the data access requirements, it effectively reduces the overall power consumption of the system and extends the battery life of the device. Its dynamic mapping mechanism and access frequency adaptation adjustment significantly improve the power utilization efficiency. The time-division multiplexing is used to reduce the pressure of frequent power supply switching and enhance the system response stability. At the same time, through the linkage analysis of logs and environmental variables, the adaptability and intelligence of the scheduling strategy are improved, and it has good engineering application value.

[0048] S1 includes obtaining the access request data of each storage area, determining the access frequency through time series analysis, extracting the distribution characteristics from the time series to obtain the access frequency resolution of each area. For the distribution characteristics, the sliding window method is used to determine the trend acquisition window. Through the trend acquisition window, the area data covering the past 24 hours is obtained. If the access frequency exceeds the preset threshold, frequency analysis is performed on the storage area to obtain the change trend. According to the change trend, the area division is adjusted to determine the optimized access request distribution. The K-means algorithm is used to cluster the adjusted distribution to obtain the trend characteristics of each area.

[0049] In this implementation, an intelligent perception of the access patterns of different storage areas and a behavior pattern modeling are achieved by constructing a time series analysis mechanism based on access behavior data. Specifically, the system first collects the original access request data from each storage partition. These data are indexed by timestamps to form a continuous sequence of access records. By introducing time series analysis methods, such as extracting frequency domain features through Fourier transform or fitting short-term trends using the ARIMA model, the system can effectively identify the periodic, sudden, and stable characteristics of access behaviors. On this basis, to ensure the timeliness of trend recognition, the system constructs a sliding window analysis model, typically set to cover a 24-hour window with a step size of 10 minutes, to achieve dynamic sliding monitoring and trend capture of access data. The data within each sliding window will be further calculated for access frequency resolution, to judge the access intensity level of the current area, and compared with the frequency threshold set by the system. When it is detected that the access frequency of a certain area has increased significantly (such as the access frequency exceeds the set threshold in multiple consecutive windows), the system triggers the trend detection module to extract the trend changes of the access behavior in this area, such as the rising / falling trend of the access frequency, the duration of the peak value, etc. Then, the system combines the trend data to re-evaluate the boundaries and distributions of the current storage areas. By adjusting the logical area division, the storage blocks that were originally processed uniformly are subdivided into multiple sub-areas to achieve finer-grained resource management. Finally, after the area reconstruction is completed, the system uses the K-means clustering algorithm to perform clustering analysis on the newly divided access data. By setting an appropriate value of K (for example, determined by evaluating the silhouette coefficient), the system can identify sub-area clusters with different access patterns, such as high-frequency stable clusters, low-frequency fluctuation clusters, discrete burst clusters, etc. These clustering results serve as important bases for subsequent power scheduling and resource priority allocation, helping to achieve targeted power supply and on-demand access, thereby optimizing the overall energy consumption control and response performance.

[0050] For example, in a vehicle-mounted terminal equipped with an embedded Linux system, its internal storage is divided into multiple areas such as a system area, a navigation area, and a video buffer area. The system continuously collects the access logs of these three areas and performs sliding time series analysis with a 10-minute step size and a 1-hour window width. It is found that the access frequency in the navigation area significantly increases during the morning rush hour (7:30–9:00) and the evening rush hour (17:00–19:00) on weekdays, reaching more than 25 times per minute, far exceeding the system-set threshold of 15 times / minute. The access frequency of the video buffer area shows a periodic pattern of increase on Wednesday and Friday nights every week. After extracting the access characteristics, the system uses the K-means algorithm to cluster the data distribution of each area in the past 24 hours. Finally, the navigation area is divided into three sub-areas: "route calculation", "real-time traffic synchronization", and "historical track". Among them, the "route calculation" sub-block shows high-frequency short-time peak characteristics, so it is given a higher power supply priority in the subsequent power distribution strategy; while the "historical track" is set to a delayed power supply mode due to its low access frequency and slow change. Through this dynamically optimized access division strategy, while ensuring the navigation experience, the overall power consumption of the device is reduced by about 13%, and at the same time, the system cache delay is reduced by 22%, achieving a dual optimization of performance and energy consumption.

[0051] This method realizes the dynamic optimization and reconstruction of the storage area based on real-time access behavior. On the premise of ensuring the data response ability, it effectively reduces the redundancy of power distribution and improves the overall energy consumption control level of the system, and is particularly suitable for scenarios with strict energy efficiency requirements under mobile power supply.

[0052] S2 includes obtaining the data of the sampling interval from the power sampling by analyzing the time difference characteristics, determining the frequency recorded per second, adjusting the recording frequency according to the change of the sampling interval to obtain the real-time value of the power change. For the real-time value of the power change, judging the corresponding mode of power distribution through a pre-constructed mapping relationship. If the mode of power distribution exceeds the preset threshold of the distribution model, dynamically adjust the sampling interval to obtain a new change value. According to the new change value, extract the key points of the adjustment requirement from the judgment basis to determine the direction of dynamic adjustment. Through the direction of dynamic adjustment, use the support vector machine algorithm to obtain the optimized power distribution plan. For the change trend of the optimized plan, judge the final result of the adjustment requirement through the logistic regression algorithm.

[0053] This implementation mode realizes the intelligent optimization of the power dynamic distribution strategy by analyzing the time difference characteristics between the power supply sampling interval and the power change. First, the system continuously samples the output power of the power supply board, initially setting the sampling frequency to 1 time per second and recording the actual interval between each sampling. By detecting the minute fluctuations in the sampling interval, the system captures characteristics such as the power supply response delay or advance caused by the load change, and then obtains the real-time gradient value of the power change. Subsequently, the system compares this real-time change value with a preset mapping model to determine whether the current power distribution conforms to the established power supply strategy. For example, when the power change rate in a certain stage is much higher than the load increment corresponding to the typical model, it may indicate an overloading tendency in the power supply. Based on this, the system determines that the power distribution mode has exceeded the acceptable range. To address the above abnormal situation, the system automatically adjusts the sampling frequency to improve the time resolution of the power change response. The adjusted sampling results are used to re-evaluate the distribution requirements and extract key change points, such as slope mutation points or continuous growth segments, as the criteria for the optimization direction. Through the classification training of these change points using the support vector machine, the optimal power distribution model under the current load environment can be obtained. Then, the system uses the logistic regression algorithm to fit and judge the change trend of the optimized model to determine whether further adjustment is necessary, so as to achieve a dynamic adaptive power management mechanism.

[0054] For example, in a handheld multi-sensor acquisition terminal, the initial sampling frequency of the system is set to 1 Hz. However, after connecting the Bluetooth module and enabling GNSS positioning, the instantaneous power consumption of the device increases significantly, resulting in the power consumption rate exceeding 25% of the predicted value of the original mapping model, triggering the dynamic adjustment mechanism of the sampling interval. The system increases the sampling frequency to 5 Hz and identifies 3 power change inflection points by recording the change curve, and conducts support vector classification training on them to determine the current "high-speed load boost" mode. On this basis, the system uses logistic regression fitting to obtain that the average power supply upper limit needs to be increased by 12% to operate stably. The optimized solution successfully controls the upward trend of power consumption, making the overall power consumption curve tend to be stable. The test results show that the continuous operation time of the device is extended by about 9%, and the sensor data synchronization accuracy is maintained in complex scenarios, demonstrating good energy efficiency adaptability and load dynamic response capabilities.

[0055] This method constructs a power distribution regulation mechanism with self-learning ability by integrating sampling frequency control, power change trend analysis, and machine learning modeling. Compared with the traditional fixed sampling and static mapping strategies, this method can improve the power consumption response accuracy without increasing the hardware burden, and is especially suitable for device scenarios with high requirements for energy management and large fluctuations in the power supply environment. At the same time, the introduction of the SVM and logistic regression algorithms enables the model to have stronger generalization ability and decision-making flexibility, and can realize the real-time recognition and adaptation of various load modes, improving the energy efficiency and operation stability of the system.

[0056] S3 includes obtaining data in the mapping relationship, parsing the dynamic adjustment requirements, obtaining the change trend of the access frequency. Based on the change trend of the access frequency, using the resolution display method to distinguish different request time periods, determining the boundaries of period adjustment. For the boundaries of the request time periods, extracting the corresponding load data from the storage area, judging the adjustment range of the power supply. If the adjustment range of the power supply exceeds the preset threshold, then through the frequency resolution algorithm, calculating the power distribution plan after period adjustment, and obtaining the adjusted allocation time series. According to the adjusted allocation time series, obtaining the real-time state of the storage area, determining the new configuration of the power supply. Through the newly configured power supply, parsing the feedback data in the mapping relationship, judging the execution effect of the dynamic adjustment, obtaining the updated adjustment requirements, and using the updated adjustment requirements, combined with the change of the access frequency, determining the power distribution time series for the next cycle.

[0057] This embodiment takes the mapping relationship as the core input parameter and drives the dynamic evolution of the power distribution strategy through the change trend of the access frequency. First, the system reads historical and real-time data from the established access frequency - power configuration mapping table and parses the power load response relationship reflected therein. By performing trend analysis on the access frequency change curve (such as first derivative, moving average fitting, etc.), the system can accurately divide the time periods of high-frequency and low-frequency requests and use the time resolution enhancement method (such as refining to a 5-minute granularity) to accurately define the boundaries of the request time periods. After identifying the key change points, the system enters the load analysis stage, extracting load characteristics such as the amount of data read and written per unit time and the current response from the corresponding storage area to determine whether there is a situation exceeding the preset power supply fluctuation threshold (for example, ±10%). When the power supply adjustment requirement exceeds the threshold, the system starts the frequency resolution algorithm and calculates the adjusted power distribution time series according to the access frequency density function to form a time period power supply plan with fine-grained control ability. Subsequently, the system obtains the real-time operating state of the storage area (including temperature, voltage, current, etc.) after the adjustment plan is executed to correct the feedback link in the mapping relationship. If the feedback data shows that there is a deviation in the current allocation, the feedback result is injected into the dynamic adjustment model, the updated demand parameters are recalculated, and combined with the latest access trend, a power distribution sequence for the next cycle is formed to ensure that the entire power supply strategy has a closed-loop adjustment ability.

[0058] For example, in an edge computing device equipped with a multitasking operating system, the system maintenance module analyzes the trend of access frequency changes within the most recent 48 hours and discovers that the peak access hours are from 10 am to 12 pm and from 3 pm to 5 pm on weekdays. During these hours, the I / O requests in the storage area are intensive and fluctuate violently. Based on this, the system refines the originally set 30-minute fixed power supply cycle into a 10-minute dynamically adjusted period and temporarily increases the power supply frequency from 1 Hz to 3 Hz through a frequency resolution algorithm to meet the response requirements in the high-load area. After the implementation of this strategy, the real-time monitoring data of the system shows that the temperature rise rate in the video buffer slows down and the voltage volatility drops by 7%. The feedback link determines that the adjustment effect is good, so this new allocation strategy is updated as the benchmark configuration for the next cycle. The overall test results show that this solution can save more than 10% of energy on average under the same hardware conditions and improve the response stability and task completion rate of the device under complex load conditions.

[0059] Through this method, the system can achieve a refined dynamic matching between access behavior and power supply, thereby reducing redundant power supply and improving power utilization efficiency. The combined use of resolution display and the frequency resolution algorithm enhances the scheduling granularity and the response speed of energy regulation. Combining the feedback mechanism and the cycle update strategy endows the power control with the ability of adaptive evolution. Even in an environment with frequent load changes or high uncertainty in data access behavior, the stability and intelligence of the system power supply can still be maintained.

[0060] S4 includes dividing the time slices of the power supply through time-division multiplexing technology, obtaining the power supply time series data of each storage area, getting the preliminary area division result, extracting the allocation characteristics of the power supply from the area division result, using a preset threshold to judge the stability of the supply control, determining the power supply balance state of each area, calculating the statistical value of the power fluctuation within the detection cycle for the power supply balance state, obtaining the distribution range of the fluctuation amplitude, analyzing the time series characteristics of the change situation according to the distribution range of the fluctuation amplitude, judging whether there is an abnormal fluctuation pattern. If an abnormal fluctuation pattern exists, then by adjusting the time slot ratio of the time slice division, obtaining the optimized supply control parameters, getting the new allocation stable state, extracting the cycle analysis data from the new allocation stable state, classifying the power fluctuation using the support vector machine algorithm, determining the long-term trend of the fluctuation amplitude, and updating the boundary conditions of the area division through the long-term trend analysis result to obtain the final optimized scheme for the stability detection cycle.

[0061] This implementation realizes fine regulation of the power supply through time-division multiplexing technology, and distributes limited power resources among different storage areas in a time-slot manner. The system first monitors the access behaviors of each storage area and establishes a corresponding power supply time series model. Each time slot represents the power supply window of the area within a scheduling cycle. The preliminary division result is generated based on the access frequency and historical power consumption data to delimit the length of the time slots occupied by each area. Then, the system extracts characteristic parameters from the division result, such as time slot density, average current, duration, etc., and combines the set stability threshold (e.g., fluctuations less than ±5% are considered stable) to judge the power supply balance of each area. During the detection period, the system calculates the power fluctuation value and plots the fluctuation distribution map to evaluate the adaptability and stability of different area power supply strategies during actual operation. If the fluctuation of a certain area exceeds the expected value or shows a sudden peak, the system determines this as an abnormal fluctuation, and then dynamically adjusts its time slot ratio (e.g., shortening the time slot during peak periods and extending the time slot during stable periods) to obtain a more balanced power supply strategy. This adjusted configuration is the new allocation stable state, and the system inputs it into the support vector machine model for training and classification analysis to identify the long-term trend of power fluctuations in each area (such as a slightly increasing trend or periodic peak characteristics in multiple consecutive cycles), and then updates the area boundary division according to the trend analysis result, such as merging low-frequency areas or splitting high-frequency areas, to achieve the final optimization of the stability detection cycle.

[0062] For example, in an industrial monitoring gateway with a partitioned storage structure, the storage areas are divided into three categories: real-time data buffer area, long-term archive area, and configuration file area. Initially, the system supplies power in a time-division manner according to a ratio of 6:3:1. After running for a period of time, through the time-division multiplexing log, it is found that the current fluctuation in the real-time data area exceeds ±8%, which is higher than the system-set stability threshold of ±5%, and the fluctuations are mainly concentrated at the whole hour and half-hour moments of each hour. The system classifies this abnormal fluctuation as a periodic peak type, dynamically adjusts the time slot, and temporarily increases the power supply ratio during peak hours to 70% and decreases it to 40% during the remaining hours. Subsequently, the support vector machine model labels the optimization result as "load peak type" and designates this area as an independent scheduling sub-block in the next cycle. After three rounds of optimization iterations, the power supply stability of the system is significantly improved, the fluctuation is reduced to ±3%, and the average power utilization rate of the device is increased by about 14%, ensuring the continuity of data collection and reducing the overheating risk.

[0063] By introducing the time - slice scheduling strategy into the mobile - power - supplied system, this method significantly improves the utilization efficiency of power resources and the accuracy of power - supply regulation. The introduced stability - discrimination model and support - vector - machine trend - classification mechanism enable the system to adaptively identify and optimize the boundary of regional division during long - term operation, preventing abnormal local loads or power - supply interruptions caused by unbalanced power supply, and enhancing the overall system stability and operation continuity. It is particularly suitable for complex power - management environments such as embedded devices with multi - module collaborative work and cache control systems.

[0064] S5 includes obtaining power - switching frequency data. The switching frequency is obtained by measuring the number of switches within a unit time. If the switching frequency exceeds the preset hardware - response threshold, then the exceeding state is judged by comparing the frequency value with the threshold. According to the exceeding state, the adjustment length is obtained by calculating the ratio of the time - slice length to the switching frequency. Using the adjustment length, the new time - slice configuration is determined by extending or shortening the time - slice length. Through the new time - slice configuration, combined with the time - division multiplexing technology, an optimized scheduling scheme is obtained. For the optimized scheduling scheme, it is judged whether the frequency still exceeds the threshold by simulating the operation data. If the frequency still exceeds the threshold, the final scheduling scheme is obtained by iteratively adjusting the time - slice length.

[0065] This embodiment aims to solve the problems of system response lag, overheating of power modules, or decreased stability caused by too high power - switching frequency. First, the system counts the switching events of the power supply within a unit time (such as 1 second) to obtain the current switching - frequency value. If this frequency exceeds the response threshold set by the hardware module (such as 20 times / second), it is judged that there is a high - frequency - switching problem in the current power - supply strategy. Subsequently, the system calculates the recommended adjustment length (such as adjustment coefficient = actual frequency / allowed frequency) by performing a ratio operation on the current time - slice length and the switching frequency, and accordingly extends or shortens the time - slice to make the interval of power - switching actions more reasonable. The new time - slice configuration will be used as input again and combined with the time - division multiplexing scheduling mechanism to form a new power - supply scheduling scheme, which is executed in a simulated operation environment to detect the adjustment effect. If the simulation operation result shows that the switching frequency still does not drop below the threshold, the system adjusts the time - slice length again and repeats the above process until the frequency is controlled within an acceptable range, forming a final stable power - supply scheduling scheme. The whole process has the ability of closed - loop feedback regulation, ensuring that the system can continuously and adaptively optimize the power - control parameters according to the operation dynamics.

[0066] For example, in an intelligent camera integrating a Wi-Fi module and NAND Flash, the original scheduling policy sets the time slice length to 100 ms. After the system runs, it is measured that the power supply switching frequency between Wi-Fi and the storage unit reaches 27 times per second, exceeding the stable threshold of 20 times per second of the internal MOS transistors of the device. Based on this, the system calculates and adjusts the length, initially extending the time slice to 135 ms and generating a new scheduling configuration. The simulated operation data shows that the frequency drops to 21 times per second but still slightly exceeds the limit. Subsequently, the system performs the second round of iterative adjustment, extending the time slice to 150 ms. Finally, the switching frequency stabilizes within 18 times per second. At the same time, the system power consumption does not increase significantly, and the temperature in the heat concentration area drops by about 4°C, verifying that this method can maintain the access scheduling efficiency while ensuring hardware security.

[0067] This method significantly reduces the hardware load and response pressure caused by high-frequency power supply switching, and improves the service life and operation reliability of the power supply board. Its dynamic adjustment mechanism ensures the coordinated control of the power supply switching frequency and time slice scheduling, and realizes the smoothness of the power supply operation on the premise of maintaining the access efficiency. Compared with the static time slice configuration method, this solution is more flexible and adaptable, especially suitable for embedded systems or low-power edge devices with frequent fluctuations in multiple loads.

[0068] S6 includes collecting real-time data through the hardware power consumption monitoring interface, generating an initial scheduling plan according to the time-division multiplexing technology to obtain a preliminary time series, extracting real-time readings from the preliminary time series, using the compatibility adjustment rule to judge whether the data source meets the requirements of the optimization plan. If it meets, the current sequence is retained; if not, the sequence parameters are adjusted to obtain an adjusted time series. For the adjusted time series, power consumption monitoring data is obtained, the change trend of the hardware power consumption is analyzed to determine the power consumption distribution characteristics. According to the power consumption distribution characteristics, the support vector machine algorithm is used to classify the high-power consumption interval and the low-power consumption interval in the time series to obtain a classification result. Through the classification result, the parameters of the time-division multiplexing scheduling plan are adjusted to generate an optimized scheduling sequence. After obtaining the optimized scheduling sequence, combined with the real-time data, it is judged whether the sequence meets the hardware power consumption constraint conditions. If it meets, the final time series is output; if not, it returns to the step of adjusting the sequence parameters for reprocessing. The adjusted sequence is extracted from the final time series, and the consistency of the real-time readings is verified in combination with the data source to obtain a scheduling time series after compatibility verification.

[0069] Based on the real-time acquisition and intelligent analysis of power consumption data, this embodiment realizes the compatibility dynamic optimization and verification of the power supply scheduling scheme. First, the system collects real-time data such as voltage, current, and power through a hardware power consumption monitoring interface (such as multi-channel sensors like INA226 and MAX34407), and establishes a preliminary power supply time series based on the time slice scheduling scheme generated by the initial time division multiplexing strategy. The system extracts key reading points from the time series and makes a matching judgment with the established optimization goals. If the current reading is within the allowable fluctuation range, it indicates that the scheduling scheme has good compatibility and the sequence is retained; if there is a deviation, a parameter fine-tuning mechanism is activated, such as adjusting the time slice length, task execution interval, or load priority. The adjusted time series is used again to collect power consumption behavior data, and by trend modeling the data, the power consumption response patterns of the system in different time periods are identified. The system extracts power consumption distribution characteristics (such as average power consumption, maximum power consumption, change rate, etc.), and uses a support vector machine to perform binary classification on high-power consumption intervals and low-power consumption intervals. The classification results are used to guide the further adjustment of the scheduling strategy parameters, such as smoothing, peak shaving, or peak shifting scheduling for high-power consumption areas, to generate an optimized scheduling sequence. Finally, the system compares and analyzes the optimized scheduling sequence with the real-time power consumption data to determine whether it meets the power consumption limit conditions set by the system (for example, the average power consumption within a single cycle < 200 mW and the peak value not exceeding 500 mW). If it meets the conditions, the final time series is output; if it does not meet the conditions, it returns to the parameter adjustment step to form an iterative closed loop. Finally, to ensure system consistency, the adjustment factors are retrospectively analyzed from the final time series, and real-time reading consistency verification is performed in combination with the original data source to form a scheduling time series with compatibility confirmation.

[0070] For example, in an intelligent monitoring host for edge video analysis, it has four built-in storage areas for caching, encoding, encryption, and uploading respectively. At the initial stage of system operation, an average power supply cycle sequence was generated according to the default time division multiplexing strategy. By collecting power consumption data through INA226, it was found that the encoding module had intermittent power consumption peaks during high-resolution video processing, with a maximum up to 620 mW, exceeding its safety threshold of 500 mW. The system immediately adjusted the time slice configuration of the encoding area to reduce the power supply during peak periods, and through SVM, it was identified that the encoding area was a high-power consumption area from 18:00 to 22:00 every day, and a power supply scheme of "preheating in advance + peak period buffering" was formulated. After the scheduling adjustment, the system operated stably during the power consumption peak period, the maximum power consumption dropped to 480 mW, the system temperature decreased by 5 °C, and the total operating power consumption decreased by 11%. The final scheduling time series was verified to be completely matched with the actual sampling data, proving that this method has remarkable effects in dynamic regulation, high-precision identification, and system compatibility.

[0071] This method upgrades the traditional static power supply scheduling strategy to a data-driven adaptive optimization process. By introducing power consumption trend analysis and high and low power consumption interval classification mechanisms, it enhances the dynamic adaptability of the scheduling scheme to the system operating state. The support vector machine algorithm improves the pattern recognition accuracy for complex fluctuating data, enabling the scheduling optimization to not only avoid energy consumption peaks but also ensure the smoothness and security of the system during high-load periods. The compatibility verification mechanism ensures that the scheme is consistent with the actual hardware state, effectively enhancing the overall system stability, power consumption control ability, and operation reliability.

[0072] S7 includes obtaining compatibility-adjusted data from the scheduling time series, determining the adjustment result through time series analysis, extracting operation duration data using the adjustment result, judging whether the continuous operation meets the 72-hour requirement, and obtaining the threshold matching status by comparing the operation duration with the threshold setting.

[0073] This embodiment analyzes and verifies the continuous operation ability of the system, aiming to ensure that the device can continuously operate for no less than 72 hours under the optimized scheduling strategy. First, the system extracts relevant data from the compatibility-adjusted scheduling time series generated in the previous stage, including power supply cycle, task execution records, power consumption, and other information. Through time series analysis methods (such as trend fitting, moving average), the system evaluates the stability and consistency of the scheduling scheme during long-term operation and identifies factors that may affect continuous power supply, such as sudden load changes, sudden resource occupations, etc. Subsequently, the system statistically analyzes the actual operation duration under the current scheme, combines multi-dimensional data such as task execution logs, battery life information, and operation interval time to form a continuous operation time record. The system sets a threshold (such as 72 hours) as the basic reliability target, compares the current recorded value with this threshold, and judges whether the minimum standard for continuous operation is met. If the continuous operation time is equal to or exceeds 72 hours, it indicates that the current scheduling strategy has good operation persistence; otherwise, it is necessary to trace back to the previous stage parameters and trigger the optimization module to adjust the scheduling structure again.

[0074] For example, in a solar-powered edge node for ecological monitoring in a remote environment, after the system completes the scheduling optimization, it is necessary to confirm that it can still maintain data collection and upload under continuous cloudy conditions. Through scheduling time series analysis, the system obtains the scheduling record and power consumption information after the latest update, and combines the battery capacity and power consumption curve to calculate its operable time as 74.5 hours, exceeding the set 72-hour threshold. At the same time, the system also extracts the operation status within the recent 72 hours from the log, and no task loss or interruption occurs. The continuous operation judgment result is "match successful". Based on this, the system confirms that the current scheduling configuration can meet the requirements of long-term low-power continuous operation.

[0075] This method can effectively evaluate the stability and endurance guarantee ability of the scheduling optimization scheme, and improve the adaptability of the system to long-term operation scenarios. By operating the duration matching judgment mechanism, it enhances the system's early recognition ability for risks such as abnormal power consumption and unstable scheduling, and ensures the continuous execution requirement of tasks. It is especially applicable to unattended or low-maintenance-frequency application environments such as edge computing devices and remote sensing terminals.

[0076] S7 also includes that if the threshold matching status shows insufficiency, extract the power management parameters from the lowest power data, determine the optimized time series, adjust the management sequence according to the optimized time series, obtain the preliminary sequence of power management, perform a matching judgment between the preliminary sequence and the lowest power requirement, determine the final optimized time series, and update the time series data with the final optimized time series to complete the generation of the power management sequence.

[0077] This embodiment is used to improve the operation endurance ability by optimizing the power management strategy when the system cannot meet the 72-hour continuous operation threshold. When the judgment mechanism described in claim 8 finds that the operation duration is lower than the set threshold (such as 65 hours < 72 hours), the system will enter the power compensation optimization mode. First, the system extracts the minimum value of the current power and its corresponding operating conditions (i.e., the lowest power data) from the power consumption log and the battery sensor record, and analyzes the key parameters that cause the lower power limit, such as the frequency, single power consumption, and execution period of a certain high-load task. Then, the system constructs an optimized time series model based on this, that is, without affecting the key tasks, by compressing the scheduling window of low-priority tasks, extending some idle intervals, or introducing a delayed trigger mechanism, to achieve an overall downward adjustment of the power consumption curve. The optimized time series is used to adjust the original management sequence to obtain the preliminary sequence of power management, and then perform a matching calculation with the lowest power support condition set by the system again. If the preliminary sequence can already cover the 72-hour operation requirement, confirm it as the final optimized time series; if it is still insufficient, execute the process of compression and adjustment again until the matching is successful. Finally, the system uses this optimized sequence to replace the original management strategy to complete the update of the time series data and the generation of the power management sequence.

[0078] For example, in a forest fire monitoring node, after continuous rainy weather, the system's power consumption drops significantly. The scheduling system determines that the current minimum remaining power can only support about 68 hours of operation, which does not reach the set safety threshold of 72 hours. After analyzing the lowest power data, the system finds that the power consumption peak of the night image upload task is obvious, and there are redundancies in some log collection tasks. Therefore, an optimized time series is constructed. The interval of the log collection task is extended from 5 minutes to 15 minutes, and at the same time, the image upload task is adjusted from once per hour to once every 2 hours. The generated power management sequence after optimization shows that the estimated sustainable operation duration is increased to 74 hours, and it is finally confirmed to meet the continuous operation requirements through matching. The system then updates the scheduling configuration and pushes it to the device side to ensure automatic switching to this management sequence when the energy is insufficient.

[0079] This solution has good dynamic compensation capabilities. In the case of limited system operation resources and insufficient power, it can achieve battery life guarantee through task-level adjustment. The combination of its optimized time series and the lowest power model can accurately locate the energy consumption bottleneck and improve the overall power utilization efficiency. By dynamically adjusting the management sequence in an incremental manner, it avoids large-scale disturbances to the original scheduling logic of the system, improves the system stability and maintainability, and is applicable to intelligent terminals and edge computing devices with strict battery life requirements.

[0080] S8 includes collecting environmental variable data through a hardware interface, recording the influence of temperature and load, generating a time series, processing the time series using synchronous acquisition technology, storing it in a log file marked with a timestamp. If the abnormal trigger condition in the log file exceeds the change threshold continuously for 5 minutes, it is determined that the time series is abnormal. According to the abnormal trigger result, the adjustment value of the time slice allocation ratio is calculated to obtain an optimized allocation scheme. The time series is updated through the optimized allocation scheme to determine a preliminary management sequence. The comparison data between the preliminary management sequence and the original time series is obtained to judge the adjustment consistency. A regression algorithm is used to analyze the consistency data to determine the final power management time series.

[0081] This embodiment is mainly used to dynamically correct the power scheduling strategy when affected by external environmental interferences (such as temperature and load changes), ensuring the safety of system power supply and the stability of performance. The system first collects environmental variable data through hardware interfaces (such as temperature sensors and load monitoring modules), and correlates indicators such as temperature, current, and voltage fluctuations with the load status in the storage area to construct multi-dimensional time series data. Through a synchronous acquisition mechanism (such as triggered by the I2C bus or GPIO interrupt), the system ensures that all variable data is recorded in real time in a unified timestamp format and written into a log file. When it is detected that the abnormal trigger condition of a certain variable in the log file (such as the temperature continuously rising beyond the ±2°C threshold) persists for more than 5 minutes, the system marks the current time series as an abnormal state. Subsequently, based on the power consumption fluctuations and access behavior patterns in the abnormal interval, the adjustment value of the time slice allocation ratio is calculated (such as extending the time slice of low-frequency tasks and compressing the high power consumption period), thereby generating an optimized power distribution plan. This optimized plan is used to replace part of the content of the current time series to form a new preliminary management sequence. The system further compares this preliminary management sequence with the original scheduling time series, and the comparison indicators include task execution delay, average voltage fluctuation, power dissipation gradient, etc., to evaluate the scheduling consistency. If the consistency meets the requirements of operation stability, fitting analysis is performed on the comparison data through regression algorithms (such as linear regression and polynomial regression), and the trend and deviation correction coefficients are extracted, and finally a corrected power management time series is generated to ensure that the adjustment plan achieves the optimal balance between compatibility and operation performance.

[0082] For example, in an edge AI camera terminal deployed in a southern forest area, the system frequently encountered the situation that the temperature in the image processing area soared above 70°C during the high-temperature (above 35°C) operation during the day. Through the environmental variable acquisition module, the system found that the temperature highly coincided with the peak load area, and the temperature fluctuation amplitude in the log file exceeded the threshold for 5 consecutive minutes, triggering the abnormal judgment mechanism. Based on this, the system shortened the time slice of high-power consumption tasks by 20%, and scheduled low-priority tasks to be executed during the low-temperature period at night. The comparison between the adjusted time series and the original series showed that the maximum temperature dropped by 6.8°C, and the power fluctuation range narrowed by 12%. After regression analysis and correction, the system generated a final power management time series with excellent consistency and energy-saving effects, and maintained a stable operation state during the subsequent multi-day high-temperature operation.

[0083] This solution enhances the sensitivity and self-regulating ability of the scheduling strategy to external environmental changes. It can dynamically optimize the power resource allocation in the case of environmental fluctuations (such as high temperature and sudden load), reduce the high power consumption risk, and improve the operation stability of the device under extreme working conditions. The synchronous acquisition mechanism and timestamp log record enhance the data traceability ability and problem diagnosis accuracy, while the strategy correction method based on the regression algorithm improves the consistency and control accuracy of the time series.

[0084] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A low-power storage access control method based on a power bank power board, characterized in that Including: S1. Obtain the access frequency resolution of each storage area, determine the time series distribution of access requests, and obtain the trend acquisition window data of each area; S2. Establish a mapping relationship between the time difference characteristic and the power distribution in advance, and judge the dynamic adjustment requirement for recording the power change value once per second for the power sampling interval; S3. Obtain the dynamic adjustment requirement from the mapping relationship, adjust the display of different request time periods for the access frequency resolution, and determine the adjusted power distribution time series; S4. Divide the time slices for the power supply of multiple storage areas through the time division multiplexing technology to obtain the stability detection period of the power distribution for each area; S5. Obtain the power supply switching frequency. If the switching frequency exceeds the hardware response threshold, determine the optimized time division multiplexing scheduling scheme by extending or shortening the time slice; S6. According to the optimized time division multiplexing scheduling scheme, obtain the real-time readings of the consumption data source from the hardware power consumption monitoring interface to obtain the scheduling time series after compatibility adjustment; S7. According to the scheduling time series after compatibility adjustment, judge whether the continuous operation for 72 hours of the running duration threshold meets the minimum power requirement to obtain the optimized time series of power management; S8. According to the time series of power management, obtain the synchronous acquisition of the environmental variables recorded temperature and load's influence on the power from the hardware interface. If the data storage format shows that the abnormal trigger condition in the power log file marked with the timestamp exceeds the change threshold range for 5 consecutive minutes, adjust the time slice allocation ratio to determine the final power management time series.

2. The low-power storage access control method based on a power bank power board according to claim 1, wherein: The S1 includes: Obtain the access request data of each storage area, determine the access frequency through time series analysis, extract the distribution characteristics from the time series to obtain the access frequency resolution of each area. For the distribution characteristics, use the sliding window method to determine the trend acquisition window. Through the trend acquisition window, obtain the area data covering the past 24 hours. If the access frequency exceeds the preset threshold, perform frequency analysis on the storage area to obtain the change trend. According to the change trend, adjust the area division to determine the optimized access request distribution, and use the K-means algorithm to cluster the adjusted distribution to obtain the trend characteristics of each area.

3. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S2 includes: Analyze the time difference characteristic, obtain the data of the sampling interval from the power sampling, determine the frequency recorded per second, adjust the recording frequency according to the change of the sampling interval to obtain the real-time value of the power change. For the real-time value of the power change, judge the corresponding mode of the power distribution through the pre-constructed mapping relationship. If the mode of the power distribution exceeds the preset threshold of the distribution model, dynamically adjust the sampling interval to obtain a new change value. According to the new change value, extract the key points of the adjustment requirement from the judgment basis to determine the direction of dynamic adjustment. Through the direction of dynamic adjustment, use the support vector machine algorithm to obtain the optimized power distribution scheme. For the change trend of the optimized scheme, use the logistic regression algorithm to judge the final result of the adjustment requirement.

4. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S3 includes: Obtain the data in the mapping relationship, analyze the dynamic adjustment requirements, obtain the change trend of the access frequency. Through the change trend of the access frequency, adopt the resolution display method to distinguish different request time periods, determine the boundary of time period adjustment. For the boundary of the request time period, extract the corresponding load data from the storage area, judge the adjustment range of the power supply. If the adjustment range of the power supply exceeds the preset threshold, then through the frequency resolution algorithm, calculate the power distribution plan after time period adjustment to obtain the adjusted allocation time series. According to the adjusted allocation time series, obtain the real-time state of the storage area, determine the new configuration of the power supply. Through the newly configured power supply, analyze the feedback data in the mapping relationship, judge the execution effect of the dynamic adjustment, obtain the updated adjustment requirements. Adopt the updated adjustment requirements, combined with the change of the access frequency, determine the power distribution time series of the next cycle.

5. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S4 includes: Divide the time slices of the power supply through the time division multiplexing technology, obtain the power supply time series data of each storage area, get the preliminary area division result. Extract the allocation characteristics of the power supply from the area division result, use the preset threshold to judge the stability of the supply control, determine the power supply balance state of each area. For the power supply balance state, calculate the statistical value of the power fluctuation within the detection period to obtain the distribution range of the fluctuation amplitude. According to the distribution range of the fluctuation amplitude, analyze the time series characteristics of the change situation, judge whether there is an abnormal fluctuation mode. If the abnormal fluctuation mode exists, then by adjusting the time slot ratio of the time slice division, obtain the optimized supply control parameters to get the new allocation stable state. Extract the periodic analysis data from the new allocation stable state, use the support vector machine algorithm to classify the power fluctuation, determine the long-term trend of the fluctuation amplitude. Through the long-term trend analysis result, update the boundary conditions of the area division to obtain the final optimized scheme of the stability detection period.

6. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S5 includes: Obtain the power switching frequency data, get the switching frequency by measuring the number of switches per unit time. If the switching frequency exceeds the preset hardware response threshold, then judge the exceeding state by comparing the frequency value with the threshold. According to the exceeding state, calculate the adjustment length by dividing the time slice length by the switching frequency. Adopt the adjustment length, determine the new time slice configuration by extending or shortening the time slice length. Through the new time slice configuration, combined with the time division multiplexing technology to get the optimized scheduling scheme. For the optimized scheduling scheme, judge whether the frequency still exceeds the threshold by simulating the running data. If the frequency still exceeds the threshold, then obtain the final scheduling scheme by iteratively adjusting the time slice length.

7. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S6 includes: Collect real-time data through the hardware power consumption monitoring interface, generate an initial scheduling plan based on time-division multiplexing technology to obtain a preliminary time series, extract real-time readings from the preliminary time series, and use the compatibility adjustment rule to determine whether the data source meets the requirements of the optimization plan. If it meets, retain the current series; if not, adjust the series parameters to obtain an adjusted time series. For the adjusted time series, obtain the power consumption monitoring data, analyze the change trend of the hardware power consumption, determine the power consumption distribution characteristics, and according to the power consumption distribution characteristics, use the support vector machine algorithm to classify the high-power consumption interval and the low-power consumption interval in the time series to obtain the classification result. Through the classification result, adjust the parameters of the time-division multiplexing scheduling plan to generate an optimized scheduling sequence. Obtain the optimized scheduling sequence, combine it with the real-time data, and determine whether the sequence meets the hardware power consumption constraint conditions. If it meets, output the final time series; if not, return to the step of adjusting the series parameters to reprocess, extract the adjusted sequence from the final time series, and combine the data source to verify the consistency of the real-time readings to obtain the scheduling time series after compatibility verification.

8. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S7 includes: Obtain the data after compatibility adjustment from the scheduling time series, determine the adjustment result through time series analysis, extract the running duration data using the adjustment result, judge whether the continuous running reaches the requirement of 72 hours, and obtain the threshold matching status by comparing the running duration with the threshold setting.

9. A low-power storage access control method based on a power board of a mobile power supply according to claim 8, characterized in that: The S7 also includes: If the threshold matching status shows insufficiency, extract the power management parameters from the lowest power data, determine the optimized time series, adjust the management sequence according to the optimized time series to obtain the preliminary sequence of power management, determine the final optimized time series through the matching judgment between the preliminary sequence and the lowest power requirement, and use the final optimized time series to update the time series data to complete the generation of the power management sequence.

10. A low-power storage access control method based on a power board of a mobile power supply according to claim 1, characterized in that: The S8 includes: Collect environmental variable data through the hardware interface, record the influence of temperature and load, generate a time series, process the time series using synchronous acquisition technology, mark it with a timestamp and store it in the log file. If the abnormal trigger condition in the log file exceeds the change threshold continuously for 5 minutes, judge that the time series is abnormal. According to the abnormal trigger result, calculate the adjustment value of the time slice allocation ratio to obtain an optimized allocation plan, update the time series through the optimized allocation plan to determine the preliminary management sequence, obtain the comparison data between the preliminary management sequence and the original time series, judge the adjustment consistency, and use the regression algorithm to analyze the consistency data to determine the final power management time series.

Images