Industrial wide-temperature-range adaptive read / write methods, systems, and storage media

By sensing the temperature of the storage medium in real time and adaptively adjusting the flash memory parameters and data priority, the problem of abnormal read and write operations in the training industrial control terminal under wide temperature conditions was solved, and stable read and write operations of the storage medium were achieved in high and low temperature environments, ensuring the reliability of training data and the teaching effect.

CN123114850APending Publication Date: 2026-11-17SHENZHEN NEITWAY INFORMATION & TECH DEV CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611545539.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-10-08
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

The existing flash memory storage system of the training industrial control terminal cannot dynamically adjust the read and write strategy according to the temperature of the storage medium. This leads to read errors and a surge in ECC error correction pressure in high-temperature environments, an increase in read error rate in low-temperature environments, and a decrease in storage read and write performance under wide temperature conditions, which affects the reliability of training data and teaching effectiveness.

Method used

By collecting the temperature of the storage medium in real time, three temperature ranges are divided, and the flash memory hardware parameters and data read/write priority strategies are adaptively adjusted according to the performance degradation characteristic parameters of the temperature range. This includes adjusting the flash memory read/write reference voltage, ECC error correction bits, and read/write bandwidth, and dynamically generating data read/write priority strategies to achieve adaptive read/write of the storage medium.

Benefits of technology

It effectively suppressed storage errors and ECC error correction overload caused by high-temperature charge leakage and low-temperature threshold offset, ensuring the stability and reliability of reading and writing key training data, and improving the reproducibility and teaching effectiveness of training experiments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123114850A_ABST
    Figure CN123114850A_ABST
Patent Text Reader

Abstract

This application discloses a method, system, and storage medium for adaptive read / write under wide temperature conditions in industrial applications. The method includes: real-time acquisition of the operating temperature of the storage medium; dividing the storage medium into three temperature ranges based on a preset temperature threshold; and matching flash memory performance degradation characteristic parameters under the three temperature ranges. When the current temperature range is within the normal temperature range, the default read / write parameters are maintained, and no adjustment is initiated. When the current temperature range is within the extreme temperature range, the severity of performance degradation is determined based on the performance degradation characteristic parameters corresponding to the extreme temperature range. For the same performance degradation severity in the extreme temperature range, the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources are adaptively and synchronously adjusted. The integrity of each training data read / write result is verified. If the verification is abnormal, the reread / rewrite mechanism and read / write window are adaptively adjusted based on the current temperature range parameters and the severity of performance degradation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to an adaptive read / write method, system, and storage medium for storage under wide temperature conditions in industrial applications. Background Technology

[0002] In the industry-education integration intelligent manufacturing training scenario, the teaching and training industrial control terminals are deployed in training classrooms, workshop training workstations and other places, where the ambient temperature fluctuates greatly, forming a wide temperature working condition; the training industrial control terminals need to continuously collect and store a large number of industrial defect image samples, model weight files, training operation logs, evaluation index data and other types of business data.

[0003] The existing flash memory storage system in the training industrial control terminal has fixed read and write parameters and cannot dynamically adjust the read and write strategy according to the temperature of the storage medium. In high-temperature environments, flash memory charge leakage intensifies, easily leading to read errors and a surge in ECC error correction pressure. In low-temperature environments, the flash memory threshold voltage shifts, increasing the read error rate. At the same time, the existing storage read and write scheduling strategy does not distinguish the priority of training business data. When the wide temperature range causes the storage read and write performance to degrade, high-priority training image samples and model evaluation logs compete with low-priority temporary cache data for storage resources, causing critical training data read and write timeouts, packet loss, or even file corruption, directly affecting the reproduction of training experiments, teaching evaluation, and the generation of experimental reports. Summary of the Invention

[0004] In view of this, embodiments of this application provide a method, system, and storage medium for adaptive read / write under wide temperature conditions in industrial applications. This enables the training industrial control terminal to dynamically adapt and adjust its storage read / write strategy according to the real-time temperature of the storage medium. This adapts to the complex working conditions of wide temperature ranges in training classrooms and workshops, effectively avoiding abnormal reading / writing, loss, or damage of training data caused by wide temperature environments. It ensures the stable operation of industrial image defect detection training and significantly improves the reproducibility of training experiments and the overall teaching and training effect.

[0005] In a first aspect, embodiments of this application provide a method for adaptive read / write storage under wide-temperature industrial conditions. This method is applied to an industrial control terminal for intelligent manufacturing training in a production-education integration environment. The training terminal is deployed in training classrooms and workshops under wide-temperature conditions and is used to store industrial defect image samples, model weights, training operation logs, and evaluation index data. The method includes the following steps: The operating temperature of the storage medium is collected in real time, and three temperature ranges are divided according to preset temperature thresholds. The flash memory performance degradation characteristic parameters under the three temperature ranges are matched. When the current temperature range falls within the normal temperature range, the default read / write parameters are maintained, and no adjustment is initiated. When the current temperature range falls within the extreme temperature range, the severity of performance degradation is determined based on the performance degradation characteristic parameters corresponding to the extreme temperature range. For the same performance degradation severity in the extreme temperature range, the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources are adaptively and synchronously adjusted. Specifically, this includes: adaptively adjusting flash memory hardware parameters to compensate for the storage read / write deviation caused by the wide temperature environment based on the differences in flash memory degradation characteristics in different temperature ranges. The hardware parameters include flash memory read / write reference voltage and ECC error correction bits. The training data read / write priority strategy is dynamically generated synchronously based on the severity of performance degradation. High-value training business data is classified and graded, with industrial defect image samples and evaluation index data set as high priority, model weights set as medium priority, and temporary cache data set as low priority. Read / write bandwidth and retry resources are dynamically allocated based on priority. The integrity of each training data read and write result is checked. If the check is abnormal, the reread and rewrite mechanism and read and write window are adaptively adjusted based on the current temperature range parameters and the severity of performance degradation.

[0006] An adaptive read / write system for industrial wide-temperature conditions is applied to an industrial control terminal for intelligent manufacturing training in an industry-education integration context. The training terminal is deployed in training classrooms and workshop workstations under wide-temperature conditions and is used to store industrial defect image samples, model weights, training operation logs, and evaluation index data. The system includes: The acquisition and matching module is used to acquire the operating temperature of the storage medium in real time, divide it into three temperature ranges according to a preset temperature threshold, and match the flash memory performance degradation characteristic parameters under the three temperature ranges. The control module is used to maintain default read / write parameters and not initiate adjustment when the current temperature range is within the normal temperature range; when the current temperature range is within the extreme temperature range, it determines the severity of performance degradation based on the performance degradation characteristic parameters corresponding to the extreme temperature range, and adaptively and synchronously adjusts the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources for the same performance degradation severity in the extreme temperature range. Specifically, it includes: adaptively adjusting flash memory hardware parameters to compensate for storage read / write deviation caused by the wide temperature environment based on the differences in flash memory degradation characteristics in different temperature ranges, wherein the hardware parameters include flash memory read / write reference voltage and ECC error correction bits; synchronously and dynamically generating a training data read / write priority strategy based on the severity of performance degradation, classifying high-value training business data, setting industrial defect image samples and evaluation index data as high priority, model weights as medium priority, and temporary cache data as low priority, and dynamically allocating read / write bandwidth and retry resources according to priority; The verification module is used to verify the integrity of each training data read and write result. If the verification is abnormal, the reread and rewrite mechanism and read and write window are adaptively adjusted based on the current temperature range parameters and the severity of performance degradation.

[0007] Thirdly, embodiments of this application provide a computer-readable storage medium storing a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps of the method described in the first aspect of embodiments of this application.

[0008] As can be seen, the technical solution of this application targets the industrial control terminal for intelligent manufacturing training in industry-education integration deployed in training classrooms and workshops under wide-temperature conditions. It achieves this by real-time sensing of storage medium temperature, partitioning and matching flash memory performance degradation characteristics, and synchronously and adaptively adjusting the flash memory read / write reference voltage and ECC error correction bits according to the degree of temperature-range degradation. This precisely compensates for flash memory read / write deviations caused by high and low temperatures at the hardware level, effectively suppressing storage errors, read / write anomalies, and ECC error correction overload problems caused by high-temperature charge leakage and low-temperature threshold shifts. Furthermore, the above technical solution deeply links the degree of storage temperature-range degradation with the scheduling of training business data read / write, dynamically generating data priority strategies based on the severity of wide-temperature performance degradation, and prioritizing industrial defect image samples and training evaluation indicators. This application prioritizes resources for high-value training data, ensuring the read and write of core training data during performance degradation in extreme temperature ranges. It also suppresses resource contention for low-priority temporary cache data, completely resolving the shortcomings of traditional fixed scheduling mechanisms that lead to timeouts, loss, and corruption of critical teaching data. Furthermore, this application incorporates a reread and rewrite mechanism that adaptively adjusts read and write anomalies based on real-time temperature parameters. This achieves closed-loop self-repair of storage read and write operations under wide temperature conditions, significantly improving the storage stability and data reliability of the training industrial control terminal under complex conditions with large temperature differences. It effectively ensures the reproducibility of industrial image defect detection training experiments, stably supports teaching evaluation and training report generation, and significantly enhances the overall teaching effectiveness and equipment operational robustness of industry-education integrated intelligent manufacturing training. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A schematic diagram of the structure of a teaching and training industrial control system provided in this application; Figure 2 A flowchart illustrating an adaptive read / write method for storage under wide temperature conditions in industrial applications provided in this application; Figure 3 A flowchart illustrating an adaptive read / write method for storage under wide temperature conditions in industrial applications, provided in Embodiment 1 of this application; Figure 4This application provides a schematic diagram of the structure of an industrial wide-temperature adaptive read / write storage system. Detailed Implementation

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

[0012] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0013] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article indicates that the preceding and following related objects have an "or" relationship.

[0014] In this application's embodiments, "multiple" refers to two or more. In this application's embodiments, "connection" refers to various connection methods, such as direct or indirect connections, to achieve communication between devices; this application's embodiments do not impose any limitations on this.

[0015] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0016] The following describes the relevant content, concepts, meanings, technical issues, technical solutions, and beneficial effects involved in the embodiments of this application.

[0017] ECC (Error Correction Code) is a built-in data error correction mechanism in flash memory controllers.

[0018] See Figure 1 , Figure 1 A schematic diagram of the structure of a teaching and training industrial control system provided in this application is shown below. Figure 1 As shown, the teaching and training industrial control system is divided into a hardware layer, an adaptive control layer, and a result output layer. The hardware layer includes: a main control chip, NAND flash memory units, a temperature sensor, an interface module, and a display unit. The main control chip is electrically connected to the NAND flash memory units, the temperature sensor, the interface module, and the display unit. The temperature sensor is used to collect the medium temperature of the NAND flash memory units. The NAND flash memory units are used to store training business data such as industrial defect image samples, model weights, training logs, and evaluation indicators. The interface module is used for training data interaction. The display unit is used to display the storage operation status and wide-temperature anomaly information.

[0019] The adaptive control layer runs on the main control chip and sequentially executes storage medium temperature acquisition, temperature range determination, hardware parameter adjustment, error correction intensity adjustment, and task priority adjustment. The adaptive control layer determines the temperature range based on the acquired storage medium temperature, adjusts the hardware read / write parameters and error correction intensity of the NAND storage unit based on the temperature range result, and dynamically adjusts the priority of training data read / write tasks in combination with the storage performance degradation degree corresponding to the temperature range.

[0020] The output layer is used to output the results of storage, reading, writing and execution, realize log saving, model weight writing, exception information recording, sample information reading and writing, and complete the reliable storage and reading of training business data under wide temperature conditions.

[0021] See Figure 2 , Figure 2 This application provides a flowchart illustrating an adaptive read / write method for storage under wide-temperature conditions in industrial applications. The method described above... Figure 1 The above method is executed under the structure of a teaching and training industrial control system shown in the figure. It is applied to an industry-education integration intelligent manufacturing training industrial control terminal, which is deployed in training classrooms and workshops under wide-temperature operating conditions. The terminal is used to store industrial defect image samples, model weights, training operation logs, and evaluation index data. The method includes the following steps: Step S201: Real-time acquisition of the operating temperature of the storage medium, division into three temperature ranges based on preset temperature thresholds, and matching of flash memory performance degradation characteristic parameters under the three temperature ranges. The specific implementation methods of the above steps may include: The system collects real-time temperature samples from the on-chip temperature sensor of the flash memory storage medium and pre-configures temperature thresholds for three temperature ranges: low temperature, normal temperature, and high temperature. For example, it sets the following parameters: low temperature range T < 0℃, normal temperature range 0℃ ≤ T ≤ 60℃, and high temperature range T > 60℃. The collected real-time temperature values ​​are compared with the preset temperature thresholds to determine the current temperature range of the storage medium. The system pre-calibrates and stores the flash memory performance degradation characteristic parameter sets corresponding to each of the three temperature ranges. These parameters include: flash memory cell charge leakage rate, threshold voltage offset, basic raw bit error rate, and ECC basic error correction load. These degradation characteristic parameters are obtained by batch stress testing of flash memory chips in the corresponding temperature range and are pre-installed in the configuration table of the training industrial control terminal. After determining the current temperature range, the system retrieves the configuration table, matches and obtains the complete set of flash memory performance degradation characteristic parameters corresponding to that temperature range, and outputs this set of degradation characteristic parameters to the subsequent parameter adjustment module as input for adaptive adjustment of flash memory hardware parameters and calculation of training data read / write priority.

[0022] The above temperatures are continuously collected at a fixed sampling period (e.g., 200ms or 300ms); when the temperature fluctuates slightly at the threshold, temperature hysteresis anti-jitter is added to avoid frequent switching between temperature ranges and prevent repeated jittering of stored parameters.

[0023] Refer to Table 1, which is a matching table of temperature range-flash memory performance degradation characteristic parameters.

[0024] Table 1:

[0025] For the low-temperature temperature range matching method in Table 1, flash memory core failure in low-temperature environments is a positive shift in the threshold voltage, resulting in inaccurate read discrimination levels and increased read errors; however, charge leakage is low and write operations are relatively stable. The system is matched with a second-level attenuation level, corresponding to a hardware compensation strategy that increases the read reference voltage and improves the ECC error correction bits, while appropriately tightening high-priority data read and write resources.

[0026] For the room temperature range matching method in Table 1, the temperature is within the optimal operating range of the flash memory, and the various attenuation parameters are close to the factory standard, with no obvious performance loss. There is no attenuation level in the matching, and the default read / write voltage and basic ECC error correction parameters are directly used without the need for additional resource scheduling intervention, ensuring maximum read / write efficiency.

[0027] For the high-temperature range matching method in Table 1, high temperature is the core scenario for storage failure in the training terminal. Rapid flash memory charge leakage, significant negative shift in threshold voltage, a surge in bit error rate, and ECC error correction overload constitute severe performance degradation. The system is matched to a level 2 severe degradation, which not only strengthens ECC error correction and reduces write speed to avoid errors, but also significantly tightens bandwidth resources, prioritizing the reading and writing of high-value data such as industrial defect images and training evaluation indicators.

[0028] The aforementioned industrial defect image samples can specifically include: real-life photos of parts collected by students during training; such as images of scratches on the outer ring of bearings, photos of chipped edges on gears, images of porosity in castings, and images of cracks in sheet metal; including original images and images of cropped ROI defect areas; the aforementioned model weights can be: weight files for defect detection deep learning models, such as weight files trained with YOLO or CNN; the aforementioned training operation logs can be: automatically recorded logs of the entire training process; including: student account, training start and end times, image acquisition timestamps, parameter setting records, student operation steps, and records of each inference operation; structured logs; the aforementioned evaluation index data can include: AI model evaluation and quantitative indicators of training effectiveness; such as defect recognition accuracy, recall rate, false negative rate, inference time, model loss value, and confusion matrix data.

[0029] Step S202: When the current temperature range is within the normal temperature range, maintain the default read / write parameters and do not initiate adjustment; when the current temperature range is within the extreme temperature range, determine the severity of performance degradation based on the performance degradation characteristic parameters corresponding to the extreme temperature range, and adaptively and synchronously adjust the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources for the same performance degradation severity in the extreme temperature range; adaptively adjust the flash memory hardware parameters to compensate for the storage read / write deviation caused by the wide temperature environment based on the differences in flash memory degradation characteristics in different temperature ranges, the hardware parameters including flash memory read / write reference voltage and ECC error correction bits; determine the severity of performance degradation based on the performance degradation characteristic parameters corresponding to the current temperature range, and synchronously and dynamically generate a training data read / write priority strategy based on the severity of performance degradation, classify and categorize high-value training business data, and dynamically allocate read / write bandwidth and retry resources; For example, the aforementioned adaptive adjustment of flash memory hardware parameters to compensate for storage read / write deviations caused by wide-temperature environments, based on the differences in flash memory degradation characteristics across different temperature ranges, may specifically include: For the low-temperature range, the flash memory read reference voltage is adaptively increased to compensate for the read discrimination deviation caused by the positive offset of the threshold voltage at low temperatures, thereby improving data read accuracy; the standard write voltage and write rate are maintained; the ECC error correction bit depth is increased to the next higher level (e.g., medium-high level) to strengthen error correction for random read errors caused by low temperatures, avoiding abnormal reading of training logs and evaluation indicators due to single read failures; for the normal temperature range, the chip's default read / write reference voltage, basic ECC error correction bit depth, and working read / write bandwidth and rate are adopted; for the high-temperature range, the write reference voltage is adaptively decreased to match the negative offset of the threshold voltage at high temperatures, suppressing write level distortion caused by excessive charge leakage; the flash memory write rate is reduced to reduce single write charge accumulation errors and improve high-temperature write stability; the highest level of ECC error correction bit depth is enabled to adapt to the extremely high original bit error rate at high temperatures, alleviating the pressure of full-load ECC error correction; in conjunction with the aforementioned dynamic priority mechanism, high-priority data resources are allocated while compensating for hardware parameters.

[0030] For example, the ECC error correction bit levels mentioned above include, but are not limited to: Basic level: supports 8-bit error correction per 512B / 1KB data block; Medium-high level: supports 16-bit error correction per 512B / 1KB data block; Highest level: supports 24-bit error correction per 512B / 1KB data block.

[0031] For example, the above-mentioned write speeds may specifically include: standard write rate: 180MB / s, mid-range write rate: 140MB / s, and low-range write rate: 90MB / s; the above-mentioned write voltages may specifically include: standard write voltage: the chip's factory default programming voltage Vpp (e.g., 20V), lowering the write reference voltage can be lowered by -0.4V, and raising the write reference voltage can be raised by +0.3V; the read voltages may specifically include: standard read voltage: the chip's factory default read voltage Vref, raising the read voltage can be raised by +0.2V, and lowering the read voltage can be lowered by -0.25V.

[0032] For example, the strategy of dynamically generating read / write priorities for training data based on the severity of performance degradation can specifically include: The performance degradation characteristic parameters are weighted and determined, and a unique performance degradation level is output based on the weighted determination result. The corresponding read and write priority strategy is dynamically matched based on the unique performance degradation level.

[0033] The aforementioned weighted determination can be calculated by taking the flash memory cell charge leakage rate, threshold voltage offset, basic raw bit error rate, ECC basic error correction load, and weighting coefficients for each parameter to obtain a weighted value. A unique performance degradation level is determined based on the interval corresponding to the weighted value, and the sum of the weighting coefficients for each parameter equals 1. These performance degradation levels include no degradation level, Level 1 mild degradation level, and Level 2 severe degradation level. The names of these three levels can be adjusted according to actual application scenarios.

[0034] The above-mentioned dynamic matching of read / write priority strategies based on the unique performance degradation level, classifying high-value training business data and dynamically allocating read / write bandwidth and retry resources, can specifically include: if there is no degradation level, a balanced scheduling strategy is adopted, specifically: using the manufacturer's default parameters; if it is a first-level mild degradation level, a critical guarantee scheduling strategy is adopted, specifically: 70% (first threshold) of the total IO bandwidth quota is allocated to high-priority data, the remaining 30% is shared by the remaining priority data, and high-priority read / write requests are inserted into the queue with priority; medium-priority requests are queued in normal order; low-priority requests are postponed; the number of read / write retries for high-priority data is increased by 2 (or 1 or 3, etc.) on the default basis, and the single read / write timeout window is extended; low-priority requests are temporarily... Cached data does not preempt high-priority resources. When the IO load exceeds the threshold, low-priority requests are delayed and not directly discarded. For the second level of severe degradation, an extreme priority protection strategy is adopted, specifically: high-priority data occupies 80% (second threshold) or more of the IO bandwidth; medium-priority data only uses the remaining bandwidth when there are no read / write tasks for high-priority data; high-priority requests have the highest priority to jump the queue; medium-priority requests yield resources; low-priority read / write requests are temporarily suspended; high-priority data is simultaneously enabled with the highest level of ECC error correction and corresponding voltage and write speed hardware compensation parameters; if the storage load continues to exceed the limit, low-priority temporary cached data can be directly discarded and regenerated after the temperature returns to normal, ensuring the read / write stability of high-value data such as defect images and training evaluation indicators.

[0035] Step S203: Perform integrity verification on each training data read / write result. If the verification is abnormal, adaptively adjust the reread / rewrite mechanism and read / write window based on the current temperature range parameters and the severity of performance degradation.

[0036] For example, the above-mentioned adaptive adjustment mechanism for rereading and rewriting based on current temperature range parameters and the severity of performance degradation, and the read / write window, may specifically include: Re-collect the operating temperature of the storage medium. If the operating temperature is determined to be in an extreme temperature range, extract the severity of performance degradation in the extreme temperature range. If it is a low temperature level 1 mild degradation level, increase the number of retries and widen the read / write window. If it is a high temperature level 2 severe degradation level, set the maximum number of retries and the longest read / write window.

[0037] The above technical solution is based on a two-dimensional adaptive matching fault tolerance strategy of temperature range parameters and the severity of performance degradation, while combining differentiated backup for the priority of training business data: it significantly improves the success rate of reading and writing error correction of high-value training data under wide temperature and harsh working conditions, and avoids damage and reading / writing failure of training data in the large temperature difference environment of workshops and classrooms; the efficient mechanism of automatic recovery under normal working conditions does not sacrifice normal reading and writing performance, and accurately balances reliability and throughput efficiency.

[0038] The technical solution of this application targets the industrial control terminal for intelligent manufacturing training in industry-education integration deployed in training classrooms and workshops under wide-temperature conditions. It achieves this by real-time sensing of storage medium temperature, partitioning and matching flash memory performance degradation characteristics, and synchronously and adaptively adjusting the flash memory read / write reference voltage and ECC error correction bits based on the degree of temperature-dependent degradation. This precisely compensates for flash memory read / write deviations caused by high and low temperatures at the hardware level, effectively suppressing storage errors, read / write anomalies, and ECC error correction overload problems caused by high-temperature charge leakage and low-temperature threshold shifts. Furthermore, the technical solution deeply links the degree of storage temperature-dependent degradation with the scheduling of training business data read / write, dynamically generating data priority strategies based on the severity of wide-temperature performance degradation, and prioritizing industrial defect image samples, training evaluation indicators, etc. High-value training data is prioritized for resource allocation, ensuring the read and write of core training data during performance degradation in extreme temperature ranges, and suppressing resource contention for low-priority temporary cache data. This completely solves the defects of traditional fixed scheduling mechanisms, such as the easy timeout, loss, and damage of critical teaching data. In addition, this application combines a reread and rewrite mechanism that adaptively adjusts read and write anomalies based on real-time temperature parameters, realizing closed-loop self-repair of storage read and write under wide temperature conditions. This significantly improves the storage stability and data reliability of the training industrial control terminal under complex conditions with large temperature differences, effectively ensures the reproducibility of industrial image defect detection training experiments, stably supports teaching evaluation and training report generation, and significantly improves the overall teaching effect and equipment operation robustness of industry-education integration intelligent manufacturing training.

[0039] For example, the above method may also include: when the operating temperature of the acquired storage medium is within the set range of preset thresholds for each temperature range, starting the temperature anti-shake hysteresis processing flow, performing sliding mean filtering on multiple consecutive temperature sample values ​​to obtain a filtered temperature that eliminates instantaneous interference, and when the filtered temperature is within the set range, maintaining the current temperature range unchanged and not performing temperature range switching.

[0040] This technology features a sliding average filter applied to the temperature sampling data of the storage medium to eliminate instantaneous temperature spikes and noise. It also sets a hysteresis interval near the temperature threshold, ensuring that the current temperature range remains unchanged when the filtered temperature falls within this interval. This solves the problem of frequent temperature switching caused by slight temperature fluctuations and sampling noise in wide-temperature environments such as industry-education integration training workshops and classrooms. It prevents parameters such as flash memory read / write reference voltage, ECC error correction level, business read / write priority, and retry resources from repeatedly changing with temperature fluctuations, thus preventing oscillations in storage system regulation. It ensures that the flash memory hardware compensation strategy and the training business data scheduling strategy remain stable under scenarios with slight temperature fluctuations, improving the stability of storage read / write operations in training industrial control terminals. This prevents read / write anomalies in high-priority critical training data such as industrial defect image samples and training evaluation indicators, ensuring the normal conduct of training experiment reproduction and teaching evaluation.

[0041] For example, the above method also includes: when the filtration temperature exceeds the preset threshold range of each temperature range and the temperature range switching condition is met, temperature range switching is performed; after the temperature range switching is completed, repeated switching of temperature ranges is prohibited within the locking time.

[0042] The above technical solution only executes a legitimate temperature range switch when the filtered temperature stably exceeds the temperature hysteresis range and meets the actual operating temperature change conditions, ensuring the accuracy of temperature range adjustment triggering. Simultaneously, a fixed lockout duration is set after each temperature range switch, prohibiting repeated temperature range switching during the lockout period. This effectively avoids the problem of short-term, multiple strategy switches caused by temperature critical oscillations and repeated threshold jumps under wide temperature fluctuation conditions. It also avoids frequent updates and changes to flash memory hardware parameters, ECC error correction levels, IO bandwidth allocation strategies, and reread / rewrite fault tolerance mechanisms within a short period, ensuring the stability and continuous execution of the storage adaptive control strategy. This prevents training data read / write disorder, parameter adaptation failure, and abnormal bit error fluctuations caused by frequent strategy switches, further improving the storage reliability and data read / write consistency of the industrial control terminal in production and education training under alternating high and low temperature conditions.

[0043] For example, the above method further includes: when the filtering temperature exceeds the preset threshold range of each temperature range and meets the condition of switching the temperature range back to the normal temperature range, the read / write priority strategy, read / write bandwidth, retry resources, reread / rewrite mechanism, and read / write window are restored to default parameters in multiple steps according to a preset recovery order. The preset recovery order is as follows: read / write bandwidth quota and read / write priority strategy, number of retry times for reread resources and read / write window duration, flash memory write rate, write reference voltage, read reference voltage, and ECC error correction level. Each recovery step is executed according to a preset step size. After each step is executed, an observation period is set. The read / write error rate is collected and verified within the observation period. After the verification is passed, the next recovery step is executed.

[0044] The above technical solution, when the storage medium temperature drops and meets the conditions for normal temperature recovery, abandons the crude recovery method of instantaneous overall reset and adopts a progressive parameter rollback mechanism that is modular, orderly, step-by-step, and verified. It gradually restores the default parameters in a specific order according to the dedicated sequence of business scheduling resources, fault tolerance retry mechanism, flash memory write performance, hardware voltage parameters, and error correction level. Each parameter reset adopts small step adjustment and sets an independent observation period and real-time verification of read and write error rate. The next recovery step is only entered after the verification is correct. This effectively solves the problems of instantaneous read and write parameter mismatch, hardware adaptation disorder, sudden increase in error rate, and abnormal data read and write caused by a one-time global reset of multiple storage parameters after the end of wide temperature operation. By first restoring the upper-layer business scheduling resources and then restoring the underlying hardware parameters of the flash memory, the stability of the read and write scheduling of the training business data is prioritized. Then, the basic parameters of the flash memory hardware are gradually restored to achieve a stable, safe, and oscillating parameter reset after the temperature range returns to normal. This maximizes the protection of the read and write integrity of core training data such as industrial defect image samples, model weights, and evaluation indicators, and further improves the storage operation stability of the training industrial control terminal under wide temperature alternation conditions.

[0045] Example 1 This application embodiment applies to an industrial control terminal for intelligent manufacturing training in an industry-education integration context. The terminal is deployed in training classrooms and workshop workstations, carrying out intelligent manufacturing visual defect detection training. The terminal's storage medium is industrial-grade NAND flash memory, and it faces long-term wide temperature fluctuations from -20℃ to +85℃. The terminal daily stores four types of core training data: industrial defect image samples, model weight files, training operation logs, and evaluation index data. This embodiment pre-divides three temperature ranges: low temperature range: -20℃ to 0℃; normal temperature range: 0℃ to 60℃; high temperature range: 60℃ to 85℃. The system collects the flash memory operating temperature every 10ms, performs sliding average filtering on five consecutive frames of temperature data, and outputs the filtered temperature. Each temperature range threshold is set with a ±0.5℃ hysteresis range; when the temperature fluctuates slightly near the threshold, the current temperature range remains unchanged, preventing frequent erroneous switching. Figure 3 As shown, Figure 3 This is a flowchart illustrating an adaptive read / write method for storage under wide temperature conditions in industrial applications, as provided in Embodiment 1 of this application. Figure 3 As shown, it includes the following steps: Step S301: Collect the operating temperature of the storage medium, perform sliding mean filtering on 5 consecutive frames of temperature data, and output the filtered temperature. If the filtration temperature is in the normal temperature range, proceed to step S302; if the filtration temperature is in the low temperature range, proceed to steps S303 and S305; if the filtration temperature is in the high temperature range, proceed to steps S304 and S305.

[0046] Step S302: Do not initiate any adaptive adjustment; Step S303: Obtain performance degradation characteristic parameters (e.g., threshold voltage offset, increased read error rate), determine it as a level 1 mild degradation, increase the flash memory read reference voltage, maintain the write voltage and write rate unchanged, and increase the ECC error correction bit depth by one level; adopt a critical guarantee scheduling strategy, allocate more than 60% of the total IO bandwidth to high-priority data (defect image samples, evaluation index data, training operation logs), and share the remaining bandwidth with medium-priority model weights and low-priority temporary caches; prioritize high-priority requests, queue medium-priority requests normally, and postpone low-priority requests; increase the number of data reread and rewrite operations and widen the read and write window; after the low temperature range switch is completed, lock it for 2 seconds to prevent switching again, ensuring that the parameters are stable and effective; The above-mentioned increase in the number of data rereads and rewrites and the expansion of the read / write window can be specifically defined as follows: maximum number of rereads and rewrites: 4; read / write window: 25ms; default maximum number of rereads and rewrites: 2; read / write window: 15ms.

[0047] Step S304: Obtain performance degradation characteristic parameters (e.g., severe threshold voltage deviation, extremely high read error rate), determine it as a level 2 severe degradation, lower the write reference voltage to 2.05V, reduce the flash write rate by 100MB / s, and enable the highest level of ECC error correction; adopt an extreme priority protection strategy, set the second threshold to 80%, and dedicate more than 80% of the IO bandwidth to high-priority training data; the weight of the medium-priority model only occupies the remaining bandwidth when there are no high-priority tasks, medium-priority resources actively yield, and low-priority read and write requests are temporarily suspended; enable the maximum number of retries and the longest read and write window to fully ensure the integrity of high-value training data writing; lock for 2 seconds after high-temperature switching to prevent repeated fluctuations of the high-temperature critical temperature from causing strategy disorder.

[0048] The above-mentioned increases in data reread and rewrite counts and the widening of the read / write window are specifically as follows: maximum retry count for read / write is 4; read / write window is 25ms; default maximum retry count for read / write is 2; read / write window is 15ms; maximum retry count is 6, and the longest read / write window is 40ms; default write reference voltage: 2.2V; flash memory write rate: 180MB / s (normal temperature range, low temperature range); lower write reference voltage to reduce flash memory write rate: write reference voltage: 2.05V, flash memory write rate: 100MB / s (high temperature range).

[0049] Step S305: When the subsequently collected filter temperature stabilizes and falls back to the normal temperature range and remains stable for 3 seconds without fluctuation, the system must not perform a one-time global reset. Instead, it must strictly follow the preset sequence to restore the default parameters step by step with verification.

[0050] This embodiment achieves precise differential compensation of flash memory hardware parameters under wide temperature conditions, completely solving the problem of flash memory read / write reliability degradation under high and low temperature environments. This invention divides the flash memory into three temperature ranges: low temperature, normal temperature, and high temperature. It employs targeted parameter tuning strategies to address the physical degradation characteristics of flash memory under different temperature conditions: Under low temperature conditions, flash memory read offset is significant; by increasing the read reference voltage, raising the ECC error correction level, appropriately increasing the number of retry cycles, and widening the read / write window, the problem of increased read errors caused by low temperature is effectively offset. Under high temperature conditions, flash memory charge leakage is severe and the risk of overprogramming is high; by lowering the write reference voltage, reducing the write rate, enabling the highest level ECC error correction mechanism, configuring the maximum number of retries, and the longest read / write window, the problems of programming anomalies and data loss under high temperature environments are significantly suppressed. Compared to the traditional, crude management model that uses fixed read / write voltages, fixed error correction levels, and fixed fault tolerance parameters, this application matches corresponding hardware compensation parameters based on the degree of temperature degradation, significantly reducing the storage read / write error rate over a wide temperature range and significantly improving the operational stability of flash memory in industrial training terminals. This application also implements tiered and differentiated resource scheduling for training business data, matching the data value characteristics of industry-education training scenarios and ensuring the absolute security of core teaching data. Furthermore, this application innovatively divides training data into high, medium, and low priorities. For high-value core training data such as industrial defect image samples, training operation logs, and evaluation index data, over 60% of the IO bandwidth is allocated under mild degradation conditions, and over 80% of the IO bandwidth is allocated under severe degradation conditions. Simultaneously, it ensures that high-priority tasks are not forcibly interrupted, only yielding resources to medium and low priority tasks. The mechanism addresses several issues: First, it temporarily suspends processing. Second, it uses medium-priority scheduling for repeatedly downloadable model weights and low-priority scheduling for temporary cached data. Third, it completely solves the problems of read / write lag, loss, and verification anomalies in critical training data caused by the traditional storage's equal bandwidth allocation and indiscriminate scheduling. This effectively ensures the integrity and accuracy of traceability data, model evaluation quantification data, and defect sample teaching datasets during the training process, guaranteeing the reproducibility of intelligent manufacturing training experiments and the traceability of teaching assessments. Fourth, it completely eliminates system control oscillations caused by temperature range critical fluctuations through a multi-layer temperature anti-jitter and state latching mechanism. Fifth, the workshop and classroom training environments are characterized by small-amplitude high-frequency temperature fluctuations, instantaneous sampling noise, and repeated jumps in critical temperatures, which can easily cause traditional temperature-controlled storage systems to frequently switch temperature range parameters, leading to frequent refreshes and oscillations in ECC settings, read / write voltages, bandwidth strategies, and fault tolerance parameters. This invention eliminates instantaneous interference by using continuous multi-frame temperature sliding average filtering, combined with a temperature range threshold of ±0.5℃ hysteresis anti-shake range, to ensure that the original temperature range remains unchanged when the temperature fluctuates slightly; at the same time, a lock duration is set after each temperature range switch to prevent repeated switching of temperature range in a short period of time.The dual anti-jitter latching mechanism effectively avoids the defects of parameter disorder, frequent read / write strategy changes, and system instability caused by ping-pong switching of critical temperatures, significantly improving the operational stability and parameter adaptation continuity of the storage system under wide temperature fluctuation conditions. The temperature range fallback step-by-step verification parameter recovery mechanism of this application solves the technical defects of traditional solutions that cause read / write adaptation disorder and instantaneous bit error spikes due to global instantaneous reset after temperature fallback. Conventional wide-temperature storage solutions directly restore all default parameters at once after the temperature returns to normal, which can easily cause instantaneous mismatch between underlying hardware voltage parameters, upper-layer business scheduling parameters, and fault tolerance retry parameters, leading to short-term read / write anomalies and data corruption. This invention adopts a gradient recovery logic of "first business scheduling, then fault tolerance mechanism, then performance parameters, and finally hardware voltage and error correction level". Each step of parameter reset adopts small-step adjustment and is configured with an independent observation period to verify the read / write bit error rate in real time. The next step of recovery is only executed after the verification is qualified. At the same time, it supports an abnormal rollback mechanism. If the temperature enters the abnormal temperature range again during the recovery process, the reset is immediately terminated and the current temperature range adaptive parameters are rolled back. This orderly, controllable, verifiable, and rollback-capable recovery method achieves a smooth transition from extreme temperature ranges to normal operating conditions, completely eliminating system oscillations caused by parameter mutations and ensuring a smooth, safe, and controllable process for switching storage system parameters. This application constructs a closed-loop adaptive control system of "temperature acquisition - attenuation determination - hardware compensation - service scheduling - fault-tolerant adjustment - anti-jitter steady-state - step-by-step recovery," achieving dynamic and precise matching of storage parameters according to the degree of operating condition degradation. This invention outputs a unique performance degradation level through multi-parameter weighted determination, ensuring that hardware parameter tuning, bandwidth allocation, retry windows, and ECC settings are all controlled from the same source by the actual physical degradation degree of the flash memory, rather than simply mechanically switching temperature thresholds. This achieves a dynamic balance between storage performance and reliability under wide temperature conditions. It avoids the waste of resources and excessive power consumption caused by excessive fault tolerance and high-specification error correction under normal operating conditions, and also avoids data anomalies caused by insufficient fault tolerance and inadequate compensation under high and low temperature adverse operating conditions. While ensuring the high reliability of training data, it also takes into account the storage and operation efficiency of industrial control terminals, and is perfectly adapted to the long-term, continuous, wide-temperature, and multi-business parallel operation scenarios of industry-education integration intelligent manufacturing training equipment.

[0051] See Figure 4 , Figure 4 This application provides a schematic diagram of a storage adaptive read / write system for industrial wide-temperature conditions, applied to an industrial control terminal for intelligent manufacturing training in industry-education integration. The training terminal is deployed in training classrooms and workshop workstations under wide-temperature conditions, and is used to store industrial defect image samples, model weights, training operation logs, and evaluation index data. The system includes: The acquisition and matching module 401 is used to acquire the operating temperature of the storage medium in real time, divide it into three temperature ranges according to a preset temperature threshold, and match the flash memory performance degradation characteristic parameters under the three temperature ranges. The control module 402 is used to maintain default read / write parameters and not initiate adjustment when the current temperature range is within the normal temperature range; when the current temperature range is within the extreme temperature range, it determines the severity of performance degradation based on the performance degradation characteristic parameters corresponding to the extreme temperature range, and adaptively and synchronously adjusts the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources for the same performance degradation severity in the extreme temperature range; specifically, it includes: adaptively adjusting flash memory hardware parameters to compensate for the storage read / write deviation caused by the wide temperature environment based on the differences in flash memory degradation characteristics in different temperature ranges, wherein the hardware parameters include flash memory read / write reference voltage and ECC error correction bits; synchronously and dynamically generating a training data read / write priority strategy based on the severity of performance degradation, classifying high-value training business data, setting industrial defect image samples and evaluation index data as high priority, model weights as medium priority, and temporary cache data as low priority, and dynamically allocating read / write bandwidth and retry resources according to priority; The verification module 403 is used to verify the integrity of each training data read and write result. If the verification is abnormal, the reread and rewrite mechanism and read and write window are adaptively adjusted based on the current temperature range parameters and the severity of performance degradation.

[0052] For example, the control module 402 is specifically used to adaptively increase the flash memory read reference voltage for low temperature ranges, maintain the standard write voltage and write rate, and increase the ECC error correction bit depth to the next level; for high temperature ranges, it adaptively decreases the write reference voltage, reduces the flash memory write rate, and enables the highest level of ECC error correction bit depth.

[0053] For example, the control module 402 performs a weighted judgment on the performance degradation characteristic parameters, outputs a unique performance degradation level based on the weighted judgment result, and dynamically matches the corresponding read / write priority strategy based on the unique performance degradation level.

[0054] For example, the performance degradation levels include: no degradation level, Level 1 mild degradation level, and Level 2 severe degradation level; the control module 402 is specifically used to: if it is Level 1 mild degradation level, adopt a critical guarantee scheduling strategy, specifically: allocate the quota of total IO bandwidth greater than the first threshold to high-priority data, the remaining total bandwidth quota is shared by the remaining priority data, high-priority read and write requests are inserted into the queue with priority, medium-priority requests are queued in normal order, and low-priority requests are postponed; if it is Level 2 severe degradation level, adopt an extreme priority protection strategy, specifically: high-priority data occupies IO bandwidth greater than the second threshold; medium-priority data only uses the remaining IO bandwidth when there are no read and write tasks for high-priority data; high-priority requests have the highest queue-jumping privilege; medium-priority requests yield resources; low-priority read and write requests are temporarily suspended; wherein, the second threshold > the first threshold > 50%.

[0055] For example, the control module 402 re-collects the operating temperature of the storage medium. When the operating temperature is determined to be in an extreme temperature range, it extracts the severity of performance degradation in the extreme temperature range. If it is a low temperature level 1 mild degradation level, it increases the number of retries and widens the read / write window. If it is a high temperature level 2 severe degradation level, it sets the maximum number of retries and the longest read / write window.

[0056] For example, the control module 402 is also used to start the temperature anti-shake hysteresis processing flow when the operating temperature of the acquisition storage medium is within the set range of the preset threshold of each temperature range, to perform sliding average filtering on the temperature sampling values ​​of multiple consecutive frames to obtain the filtered temperature that eliminates instantaneous interference, and when the filtered temperature is within the set range, to maintain the current temperature range unchanged and not to perform temperature range switching.

[0057] For example, the control module 402 is also used to perform temperature range switching when the filtration temperature exceeds the preset threshold range of each temperature range and the temperature range switching conditions are met. After the temperature range switching is completed, repeated switching of the temperature range is prohibited within the locking time.

[0058] For example, the control module 402 is also used to restore the read / write priority strategy, read / write bandwidth, retry resources, reread / rewrite mechanism, and read / write window to default parameters in multiple steps according to a preset recovery order when the filtered temperature exceeds the preset threshold range of each temperature range and meets the condition of switching the temperature range back to the normal temperature range. The preset recovery order is as follows: read / write bandwidth quota and read / write priority strategy, number of retry times for reread resources and read / write window duration, flash memory write rate, write reference voltage, read reference voltage, and ECC error correction level. Each recovery step is executed according to a preset step size. After each step is executed, an observation period is set. The read / write error rate is collected and verified within the observation period. After the verification is passed, the next recovery step is executed.

[0059] This application also provides a computer storage medium storing a computer program / instructions thereon, which, when executed by a processor, implements some or all of the steps of any of the methods described in the above method embodiments.

[0060] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments.

[0061] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0062] In the several embodiments provided in this application, it should be understood that the disclosed methods, apparatuses, and systems can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for example, the division of units is merely a logical functional division, and there may be other division methods in actual implementation; for example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0063] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0064] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can be physically included separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0065] While this application discloses the above information, it is not limited thereto. Any person skilled in the art can easily conceive of variations or substitutions without departing from the spirit and scope of this application, and can make various alterations and modifications, including combinations of the different functions and implementation steps described above, as well as software and hardware implementation methods, all of which are within the protection scope of this application.

Claims

1. A method for adaptive read / write storage under wide temperature conditions in industrial applications, characterized in that, The method is applied to an industrial control terminal for intelligent manufacturing training in an industry-education integration context. This training terminal is deployed in training classrooms and workshops under wide-temperature operating conditions, and is used to store industrial defect image samples, model weights, training operation logs, and evaluation index data. The method includes the following steps: The operating temperature of the storage medium is collected in real time, and three temperature ranges are divided according to preset temperature thresholds. The flash memory performance degradation characteristic parameters under the three temperature ranges are matched. When the current temperature range falls within the normal temperature range, the default read / write parameters are maintained, and no adjustment is initiated. When the current temperature range falls within the extreme temperature range, the severity of performance degradation is determined based on the performance degradation characteristic parameters corresponding to the extreme temperature range. For the same performance degradation severity in the extreme temperature range, the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources are adaptively and synchronously adjusted. Specifically, this includes: adaptively adjusting flash memory hardware parameters to compensate for the storage read / write deviation caused by the wide temperature environment based on the differences in flash memory degradation characteristics in different temperature ranges. The hardware parameters include flash memory read / write reference voltage and ECC error correction bits. The training data read / write priority strategy is dynamically generated synchronously based on the severity of performance degradation. High-value training business data is classified and graded, with industrial defect image samples and evaluation index data set as high priority, model weights set as medium priority, and temporary cache data set as low priority. Read / write bandwidth and retry resources are dynamically allocated based on priority. The integrity of each training data read and write result is checked. If the check is abnormal, the reread and rewrite mechanism and read and write window are adaptively adjusted based on the current temperature range parameters and the severity of performance degradation.

2. The method according to claim 1, characterized in that, The adaptive adjustment of flash memory hardware parameters to compensate for storage read / write deviations caused by wide-temperature environments, based on the different flash memory decay characteristics in different temperature ranges, specifically includes: For low-temperature ranges, the flash memory read reference voltage is adaptively increased to maintain standard write voltage and write speed; the ECC error correction bit depth is increased to the next level. For high-temperature ranges, the write reference voltage is adaptively lowered to reduce flash memory write speed, and the highest level of ECC error correction bits is enabled.

3. The method according to claim 1, characterized in that, The strategy for dynamically generating read / write priorities for training data based on the severity of performance degradation specifically includes: The performance degradation characteristic parameters are weighted and determined, and a unique performance degradation level is output based on the weighted determination result. The corresponding read and write priority strategy is dynamically matched based on the unique performance degradation level.

4. The method according to claim 3, characterized in that, The performance degradation levels include: no degradation level, Level 1 mild degradation level, and Level 2 severe degradation level; the dynamic matching of read / write priority strategies based on this unique performance degradation level specifically includes: If it is a Level 1 mild attenuation level, a critical guarantee scheduling strategy is adopted, specifically: the quota of total IO bandwidth greater than the first threshold is allocated to high-priority data, the remaining total bandwidth quota is shared by the remaining priority data, high-priority read and write requests are inserted into the priority position of the queue, medium-priority requests are queued in normal order, and low-priority requests are placed at the end. If it is a level 2 severe degradation level, the extreme priority protection strategy is adopted, specifically: high-priority data uses more than the second threshold of IO bandwidth; medium-priority data only uses the remaining IO bandwidth when there are no high-priority read / write tasks; high-priority requests have the highest priority to jump the queue; medium-priority requests yield resources; low-priority read / write requests are temporarily suspended. Among them, the second threshold is greater than the first threshold, and the first threshold is greater than 50%.

5. The method according to claim 4, characterized in that, The adaptive adjustment mechanism for rereading and rewriting based on current temperature parameters and the severity of performance degradation, and the read / write window specifically include: Re-collect the operating temperature of the storage medium. If the operating temperature is determined to be in an extreme temperature range, extract the severity of performance degradation in the extreme temperature range. If it is a level 1 mild degradation in the low temperature range, increase the number of retries and widen the read / write window. If it is a level 2 severe degradation in the high temperature range, set the maximum number of retries and the longest read / write window.

6. The method according to claim 1, characterized in that, The method further includes: when the operating temperature of the acquired storage medium is within the set range of preset thresholds for each temperature domain, starting the temperature anti-shake hysteresis processing flow, performing sliding mean filtering on multiple consecutive temperature sampling values ​​to obtain a filtered temperature that eliminates instantaneous interference, and when the filtered temperature is within the set range, maintaining the current temperature domain range unchanged and not performing temperature domain switching.

7. The method according to claim 6, characterized in that, The method further includes: when the filtration temperature exceeds the preset threshold range of each temperature range and the temperature range switching condition is met, temperature range switching is performed; after the temperature range switching is completed, repeated switching of temperature range is prohibited within the locking time.

8. The method according to claim 6, characterized in that, The method further includes: when the filtering temperature exceeds the preset threshold range of each temperature range and meets the condition of switching the temperature range back to the normal temperature range, restoring the read / write priority policy, read / write bandwidth, retry resources, reread / rewrite mechanism, and read / write window to default parameters in multiple steps according to a preset recovery order. The preset recovery order is as follows: read / write bandwidth quota and read / write priority policy, retry count of reread resources and read / write window duration, flash memory write rate, write reference voltage, read reference voltage, and ECC error correction level. Each recovery step is executed according to a preset step size. After each step is executed, an observation period is set. The read / write error rate is collected and verified within the observation period. After the verification is passed, the next recovery step is executed.

9. An industrial wide-temperature-range adaptive read / write storage system, characterized in that, An industrial control terminal for intelligent manufacturing training in industry-education integration is deployed in training classrooms and workshops under wide-temperature conditions. It is used to store industrial defect image samples, model weights, training operation logs, and evaluation index data. The system includes: The acquisition and matching module is used to acquire the operating temperature of the storage medium in real time, divide it into three temperature ranges according to a preset temperature threshold, and match the flash memory performance degradation characteristic parameters under the three temperature ranges. The control module is used to maintain default read / write parameters and not initiate adjustment when the current temperature range is within the normal temperature range; when the current temperature range is within the extreme temperature range, it determines the severity of performance degradation based on the performance degradation characteristic parameters corresponding to the extreme temperature range, and adaptively and synchronously adjusts the storage read / write deviation, read / write priority strategy, read / write bandwidth, and retry resources for the same performance degradation severity in the extreme temperature range. Specifically, it includes: adaptively adjusting flash memory hardware parameters to compensate for storage read / write deviation caused by the wide temperature environment based on the differences in flash memory degradation characteristics in different temperature ranges, wherein the hardware parameters include flash memory read / write reference voltage and ECC error correction bits; synchronously and dynamically generating a training data read / write priority strategy based on the severity of performance degradation, classifying high-value training business data, setting industrial defect image samples and evaluation index data as high priority, model weights as medium priority, and temporary cache data as low priority, and dynamically allocating read / write bandwidth and retry resources according to priority; The verification module is used to verify the integrity of each training data read and write result. If the verification is abnormal, the reread and rewrite mechanism and read and write window are adaptively adjusted based on the current temperature range parameters and the severity of performance degradation.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program for electronic data interchange, wherein the computer program causes a computer to perform some or all of the steps of the method described in any one of claims 1-8.