Solid state disk intelligent temperature control method and system based on edge computing

Abstract

The invention relates to the technical field of intelligent temperature control, in particular to a solid state disk intelligent temperature control method and system based on edge computing. The method comprises the following steps: acquiring solid state disk array data; performing array topology analysis on the solid state disk array data to generate solid state disk array topology data; performing initial self-organizing heat dissipation network construction on the solid state disk array data based on the solid state disk array topological data to obtain an initial self-organizing heat dissipation network; edge computing node deployment is carried out on the initial self-organizing heat dissipation network, and temperature data of the solid state disk are collected according to edge computing nodes; and network heat conduction path optimization is carried out on the initial self-organizing heat dissipation network through the temperature data, and a self-organizing heat dissipation optimization network is generated. According to the invention, through self-organization of the heat dissipation network, edge calculation optimization, thermal-electric mode conversion and long-term prediction, SSD array-level intelligent temperature control is realized, and the heat dissipation efficiency, the response speed and the long-term stability are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent temperature control technology, and in particular to an intelligent temperature control method and system for a solid-state hard disk based on edge computing. Background Art

[0002] Initially, SSD temperature control mainly relied on passive heat dissipation, such as heat sinks and thermal grease. With the development of PCIe and NVMe protocols, data transmission rates have increased significantly, and the problem of SSD heating has become increasingly serious. Some high-end products have begun to use active heat dissipation, such as fans and metal heat dissipation casings. At the same time, temperature sensors are integrated into SSD controllers to achieve basic overheating protection. With the rise of edge computing, computing tasks tend to be deployed in a distributed manner, and SSDs need to operate stably for a long time in a high-temperature environment. Intelligent temperature control technology gradually adopts AI algorithms combined with edge computing to achieve real-time temperature monitoring, load balancing, and dynamic frequency adjustment. Some SSDs also use phase change material heat dissipation and liquid cooling technology to reduce temperature peaks. In addition, combined with the data prediction capabilities of edge computing nodes, SSD temperature control strategies can be adaptively adjusted according to workload and ambient temperature to achieve a balance between efficient heat dissipation and performance. However, the current traditional SSD temperature control strategy is usually based on single-disk monitoring and regulation. For example, a simple temperature threshold triggers frequency reduction, which lacks a global perspective and makes it difficult to achieve balanced heat dissipation in multi-disk arrays. At the same time, traditional temperature control methods mostly use frequency reduction or air cooling, which affects SSD performance. At the same time, the heat dissipation method is single and does not fully utilize factors such as current and voltage for coordinated optimization, which leads to low heat dissipation efficiency and heat dissipation accuracy of solid-state drive temperature control. Summary of the invention

[0003] Based on this, it is necessary to provide a solid-state hard drive intelligent temperature control method and system based on edge computing to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a solid state drive intelligent temperature control method based on edge computing is provided, the method comprising the following steps: Step S1: acquiring solid state drive array data; performing array topology analysis on the solid state drive array data to generate solid state drive array topology data; constructing an initial self-organizing heat dissipation network for the solid state drive array data based on the solid state drive array topology data to obtain an initial self-organizing heat dissipation network; Step S2: deploy edge computing nodes on the initial self-organizing heat dissipation network, and collect temperature data of the solid-state hard disk according to the edge computing nodes; optimize the network thermal conductivity path of the initial self-organizing heat dissipation network through the temperature data to generate a self-organizing heat dissipation optimization network; Step S3: Calculate the current thermodynamic state of the solid-state hard disk array, and perform thermal-electric mode conversion on the temperature data according to the current thermodynamic state to generate heat flow dynamic temperature control data; pre-adjust the heat path of the self-organizing heat dissipation optimization network through the heat flow dynamic temperature control data to generate hard disk global thermal balance optimization data; Step S4: predicting the long-term change trend of the temperature data, and performing hard disk optimization feedback in advance on the hard disk global thermal balance optimization data according to the long-term change trend, and generating a solid state hard disk intelligent temperature control strategy.

[0005] The present invention establishes an efficient heat conduction path through the construction and optimization of a self-organizing heat dissipation network, so that heat can be quickly and evenly distributed and dissipated, avoiding local overheating and improving the overall heat dissipation efficiency. The edge computing node is used to monitor the hard disk temperature in real time, and the global thermal balance is achieved through the heat flow dynamic temperature control technology to ensure that the temperature distribution of all hard disks is within the optimal range, thereby improving the stability of the system. Excessive temperature can cause the performance of the solid-state hard disk to decline or even damage the data. Through the intelligent temperature control strategy, the impact of high temperature can be effectively reduced, the hard disk read and write speed can be ensured to be stable, and the I / O performance loss caused by temperature fluctuations can be reduced. Being in a high temperature environment for a long time will accelerate the wear of the hard disk, and this method optimizes the feedback mechanism in advance to make the temperature control more accurate, thereby reducing the hard disk aging caused by overheating and improving the service life of the hard disk. The long-term temperature change trend prediction model is used to adjust the heat dissipation strategy in advance to avoid the impact caused by sudden temperature changes, realize active temperature control, and reduce the emergency frequency reduction or shutdown protection triggered by sudden overheating of the system. Due to the optimization of thermal management, the hard disk runs in a reasonable temperature range for a long time, which can greatly reduce the physical damage caused by thermal expansion and contraction, as well as the failure of the controller and NAND chip due to overheating, and improve the reliability of data storage. Overheating can increase the bit error rate (BER) of stored data, affecting data integrity. This method ensures that the data storage environment temperature is suitable and reduces the risk of data damage caused by temperature fluctuations through thermal balance optimization and intelligent temperature control strategy. Therefore, the present invention realizes SSD array-level intelligent temperature control through self-organizing heat dissipation network, edge computing optimization, thermal-electric mode conversion and long-term prediction, thereby improving heat dissipation efficiency, response speed and long-term stability.

[0006] Preferably, step S1 comprises the following steps: Step S11: Acquire solid state drive array data; Step S12: performing array topology analysis on the solid state drive array data to generate solid state drive array topology data; Step S13: performing intelligent material integration on the solid state drive array using the solid state drive array topology data to generate solid state drive array integrated material data, wherein the intelligent material integration includes integrated phase change materials, thermoelectric materials, and intelligent thermal conductive films; Step S14: performing microfluidic heat conduction path identification on the solid state drive array topology data based on the solid state drive array integrated material data, and generating hard drive self-assembly microstructure data; Step S15: local sensing units are implanted into the solid state hard disk array topology data through the hard disk self-assembly microstructure data to generate an initial self-organizing heat dissipation network, wherein the local sensing units include temperature sensors and heat flow sensing devices.

[0007] The present invention realizes efficient heat dissipation management, improves the overall thermal balance ability, and reduces the risk of local overheating through the integration of intelligent materials (phase change materials, thermoelectric materials, intelligent thermal conductive films) and the optimization of microfluidic thermal conduction paths. Intelligent thermal conductive films and thermoelectric materials are used to optimize heat distribution, and combined with microfluidic thermal conduction paths, the heat transfer efficiency is improved, the heat energy is quickly diffused, and the problem of local overheating is reduced. By implanting local sensing units (temperature sensors, heat flow sensing devices), real-time temperature monitoring of the hard disk array is realized, so that the system can accurately sense temperature changes and optimize thermal management strategies. Combined with microfluidic thermal conduction path identification and intelligent material integration, an automatically adjusted self-organized heat dissipation network is realized, the intelligent level of thermal management is improved, and manual intervention is reduced. Overheating can cause material expansion and contraction, affecting the long-term stability of storage chips. Through the thermal buffering effect of phase change materials, the physical stress of the hard disk caused by drastic temperature fluctuations can be effectively reduced, and the system reliability can be improved. Intelligent materials are used for thermal management optimization, so that the hard disk can operate in a low-temperature stable state, reduce the damage of high temperature to the storage unit, extend the life of the hard disk, and reduce the replacement and maintenance costs caused by overheating damage. By combining high-precision temperature monitoring of local sensing units with the rapid thermoelectric conversion capability of thermoelectric materials, the hard disk system can quickly adjust the heat dissipation strategy to avoid damage caused by sudden temperature changes. Traditional heat dissipation solutions usually rely on air cooling or liquid cooling systems, which consume a lot of energy. This method uses smart materials (thermoelectric materials, phase change materials) and microfluidic heat conduction mechanisms to make heat dissipation more efficient, thereby reducing additional energy consumption and improving the system energy efficiency ratio.

[0008] Preferably, step S2 comprises the following steps: Step S21: deploy edge computing nodes on the initial self-organizing heat dissipation network to obtain edge computing nodes of the heat dissipation network; Step S22: Based on a preset time period, the temperature sensor is used to sense the hard disk temperature of the edge computing node of the heat dissipation network to generate hard disk temperature sensing data; the hard disk temperature sensing data is preprocessed to generate standard hard disk temperature sensing data, wherein the data preprocessing includes data cleaning, data denoising and data standardization; Step S23: performing temperature distribution analysis on the standard hard disk temperature sensing data to generate a hard disk thermal distribution matrix; performing thermal conduction path optimization on the hard disk self-assembled microstructure data based on the hard disk thermal distribution matrix to generate thermal conduction path optimization data; Step S24: regional temperature discrimination is performed on the hard disk temperature sensing data according to the thermal conduction path optimization data. When the hard disk temperature sensing data is greater than or equal to the preset temperature threshold, the initial self-organizing heat dissipation network is reorganized locally through the edge computing node of the heat dissipation network to generate a self-organizing heat dissipation optimization network.

[0009] The present invention transfers data processing from the cloud to the local through the deployment of edge computing nodes of the heat dissipation network, improves the real-time performance of temperature data collection and calculation, reduces data transmission delay, and realizes efficient distributed thermal management. Data preprocessing (data cleaning, data denoising, and data standardization) is adopted to reduce sensor noise interference, ensure that the acquired hard disk temperature data is accurate and reliable, and provide high-quality data support for subsequent temperature control optimization. The hard disk temperature distribution is analyzed by the hard disk heat distribution matrix, and the heat conduction path optimization is combined to make the heat flow more uniform, improve the overall heat dissipation effect, and reduce local overheating. The regional temperature discrimination mechanism is adopted. When the hard disk temperature exceeds the preset threshold, the heat path reorganization is automatically triggered, so that the heat energy can be quickly diffused to avoid the impact of temperature accumulation on the hard disk performance. Through the local heat path reorganization, the heat channel can be intelligently adjusted according to the real-time temperature change to form a self-organized heat dissipation optimization network, ensure that the hard disk operates within a reasonable temperature range, and improve the long-term stability of the system. Traditional air cooling or liquid cooling heat dissipation methods often rely on high-power consumption equipment. This solution is based on thermal conduction path optimization and edge computing collaborative management to reduce unnecessary heat dissipation resource consumption and improve the overall energy efficiency ratio.

[0010] Preferably, optimizing the heat conduction path of the hard disk self-assembly microstructure data based on the hard disk heat distribution matrix includes: Confirming the state of the phase change material in the hard disk self-assembly microstructure data based on the hard disk heat distribution matrix, and obtaining the phase change material state data, wherein the phase change material state data includes solid state data and liquid state data; The heat flow direction and heat exchange rate of the SSD array data are calculated through the hard disk heat distribution matrix, and a local heat flow model is constructed based on the heat flow direction and heat exchange rate; The local heat flow model is used to locate the hard disk hotspot area of ​​the initial self-organized heat dissipation network and generate the hard disk hotspot area; the regional thermal buffer analysis of the hard disk hotspot area is performed to generate the hard disk regional thermal buffer data; based on the dynamic phase change control formula, the local phase change control of the hard disk hotspot area is performed according to the phase change material state data to generate the phase change material solid-liquid area dynamic adjustment data, where the local phase change control formula is as follows: ; In the formula, For phase change material in position The solid-liquid conversion ratio at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to adjust the influence weight of temperature deviation on the PCM solid-liquid conversion ratio, The temperature of the SSD array Over time The rate of change, In order to adjust the influence weight of temperature change rate on PCM solid-liquid conversion ratio, is the horizontal coordinate of the position, is the vertical coordinate of the position; Based on the thermoelectric material in the hard disk self-assembly microstructure data, the electric field application mode of the hard disk hot spot area is confirmed, and the potential difference of the hard disk hot spot area is calculated according to the power plant application mode to obtain the hot spot area potential difference; the hot spot area potential difference is controlled by the thermoelectric material electric field distribution using the heat flow guidance optimization formula to generate dynamic electric field control data, where the heat flow guidance optimization formula is as follows: ; In the formula, For phase change material in position The local potential difference at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to control the influence weight of temperature deviation on TEG potential difference, The temperature of the SSD array Over time The rate of change, In order to control the weight of the influence of temperature change rate on TEG potential difference, is the horizontal coordinate of the position, is the vertical coordinate of the position; According to the dynamic adjustment data of the solid-liquid region of the phase change material and the dynamic electric field control data, the heat diffusion path of the hard disk self-assembly microstructure data is optimized to generate the thermal conduction path optimization data.

[0011] The present invention realizes the solid-liquid intelligent conversion of phase change materials under different temperature conditions through a dynamic phase change control formula, so that heat can be quickly absorbed and exported in high-temperature areas, reducing local overheating and improving heat diffusion efficiency. Combined with the analysis of heat flow direction and heat exchange rate, a local heat flow model is constructed, which can accurately identify and guide the heat flow path, so that heat can diffuse to the low-temperature area more efficiently and improve the overall heat dissipation capacity. The local heat flow model is used to accurately locate the hot spot area of ​​the hard disk, and combined with the regional thermal buffer analysis, it is helpful to perceive the high-temperature area in advance and take corresponding temperature control measures to avoid system failure caused by excessive temperature. The phase change material state data (solid / liquid) can be used to dynamically control according to the real-time changes in the hard disk temperature, so that the temperature fluctuation is smooth, the material stress caused by excessive or low temperature is reduced, and the long-term stability and reliability of the hard disk are improved. Through the calculation of the potential difference in the hot spot area, the electric field application mode is reasonably adjusted, and the electric field effect of the thermoelectric material is used to further optimize the heat diffusion path, so that the thermal conduction path is more efficient and heat accumulation is reduced. By using the heat flow guidance optimization formula and dynamically regulating the potential difference of TEG (thermoelectric material) based on temperature deviation and temperature change rate, the thermal energy conversion efficiency can be improved, and the excess heat can be partially converted into electrical energy, thereby improving the energy utilization rate of the system. By combining phase change material regulation with dynamic electric field control, the heat diffusion path can be optimized, so that heat can diffuse outward in the shortest path, reducing ineffective heat dissipation power consumption and improving overall energy efficiency.

[0012] Preferably, the step S3 of calculating the current thermodynamic state of the solid state drive array includes: Extracting operation data of the solid state drive array, and calculating the total power consumption of the solid state drive array according to the operation data; Analyze the thermal power distribution of the SSD array through the total power consumption of the SSD array; Based on the finite element method, the temperature field of the thermal power distribution of the solid-state drive array is simulated to generate the array temperature field model; According to the array temperature field model, the heat exchange amount of the solid state drive array is evaluated and the curve is converted to obtain a heat exchange curve; The time point with a slope of 0 in the heat exchange curve was selected and marked as thermal steady-state time data; The heat exchange amount and thermal steady-state time data of the solid-state drive array are used to perform thermal runaway risk analysis on the operation data of the solid-state drive array, thereby generating the current thermodynamic state of the solid-state drive array.

[0013] The present invention can accurately grasp the energy consumption of the hard disk by extracting the operation data of the solid-state hard disk array and calculating its total power consumption, provide data support for subsequent thermal management optimization, help reduce unnecessary power consumption, and improve the overall energy efficiency ratio. The thermal power distribution analysis method can be used to identify the heat distribution in different areas, find out the high heat area and the low heat area, and achieve more refined temperature management to avoid overheating in the hot spot area, resulting in performance degradation or equipment damage. The finite element method (FEM) is used for temperature field simulation to accurately calculate the temperature distribution of the solid-state hard disk array under different power consumption conditions, establish an array temperature field model, improve the accuracy of temperature prediction, and ensure that the hard disk runs within the optimal temperature range. By calculating the heat exchange amount and performing curve conversion, the heat flow of the hard disk array can be quantified, the heat dissipation effect can be accurately evaluated, and data support can be provided for the heat dissipation optimization strategy to improve the refinement of thermal management. The heat exchange curve analysis is used to screen the time point when the heat exchange slope is 0, automatically identify the time when the hard disk reaches thermal steady state, help optimize the heat dissipation strategy, avoid temperature overshoot, and improve the long-term operation stability of the hard disk. Combined with the heat exchange volume and thermal steady-state time data, the thermal runaway risk analysis of the hard disk operation status is carried out to discover potential temperature abnormality trends in advance, avoid hardware damage, data loss or system crash caused by overheating, and improve the safety and reliability of the storage system.

[0014] Preferably, the step S3 of converting the temperature data into a thermal-electrical mode according to the current thermodynamic state includes: Perform thermal aggregation analysis on the temperature data according to the current thermodynamic state to generate thermal aggregation data; Use thermal clustering data to perform cluster analysis on SSD array data to generate hard disk hotspots, hard disk cooling areas, and hard disk neutral areas, and eliminate hard disk neutral areas; Through the thermoelectric materials in the SSD array integrated material data, the hot spot area, cooling area and neutral area of ​​the hard disk are analyzed for thermo-electric mode, and the Peltier effect mode and Seebeck effect mode are generated. The optimal heat flow guiding path for the hard disk hot spot area and the hard disk cooling area is calculated, and the electric field direction of the Peltier effect mode and the Seebeck effect mode is adjusted through the optimal heat flow guiding path to generate heat flow dynamic temperature control data.

[0015] The present invention can accurately identify the heat distribution through heat aggregation analysis, and generate heat aggregation data, provide a scientific basis for subsequent thermal management, optimize the heat dissipation strategy, and improve the overall system energy efficiency. Cluster analysis is adopted to divide the solid-state hard disk array into hard disk hot spot area, hard disk cooling area and hard disk neutral area, and the neutral area is eliminated to ensure that the heat dissipation strategy is only for key areas, reduce unnecessary heat dissipation power consumption, and improve heat dissipation efficiency. Through the thermoelectric materials in the solid-state hard disk array integrated material data, the Peltier effect mode and the Seebeck effect mode analysis are performed on different areas respectively, and the conversion of thermal energy to electrical energy is realized, the intelligent degree of thermal management is improved, and part of the energy can be recovered to reduce system power consumption. The optimal heat flow guiding path is calculated to ensure that heat can be efficiently transferred from the hot spot area to the cooling area, reduce local overheating, optimize the overall temperature distribution, and improve the stability of the long-term operation of the hard disk. Through the heat flow dynamic temperature control data, the electric field direction of the Peltier effect and the Seebeck effect is adjusted to achieve dynamic temperature balance, prevent local overheating or overcooling, ensure that the hard disk runs within the optimal operating temperature range, and extend the service life. The use of advanced thermal-electrical mode conversion strategies can effectively reduce hard disk temperature fluctuations and reduce mechanical stress caused by thermal expansion and contraction, thereby reducing hard disk failures and improving data security and reliability of storage systems.

[0016] Preferably, the step S3 of pre-adjusting the heat path of the self-organizing heat dissipation optimization network by using the heat flow dynamic temperature control data includes: The heat transfer topology structure of the self-organizing heat dissipation optimization network is mapped through the heat flow dynamic temperature control data to generate a heat transfer path diagram; the heat transfer path impedance characteristics are extracted from the heat transfer path diagram to obtain the heat transfer path impedance characteristic data; The optimization target is set based on the impedance characteristic data of the heat transfer path, and the heat transfer path diagram is dynamically allocated according to the optimization target to generate the heat path control data; The thermal path control data is used to optimize the global network thermal balance of the self-organizing heat dissipation optimization network and generate the global thermal balance optimization data of the hard disk.

[0017] The present invention uses the heat flow dynamic temperature control data to map the heat transfer topology of the self-organizing heat dissipation optimization network, accurately draw the heat flow transmission path in the network, and identify the key heat transfer path. This step optimizes the distribution of heat flow in the hard disk array, so that the heat can be more efficiently guided to the heat dissipation area, avoid hot spot concentration, and improve the overall heat dissipation efficiency. The heat transfer path impedance feature extraction of the heat transfer path diagram can understand the impedance characteristics of each heat path, so as to identify which areas have overheating problems and further strengthen thermal management. By extracting thermal impedance data, more refined heat flow distribution can be achieved to reduce unnecessary energy loss. Based on the heat transfer path impedance feature data, the optimization target is set, and through dynamic path allocation, the heat flow path can be accurately controlled, the heat dissipation capacity of each area can be flexibly adjusted, and the hot spot area can be effectively cooled. At the same time, other areas are avoided from being overcooled, thereby improving the energy efficiency and stability of the system. Using the heat path control data to perform global thermal balance optimization on the self-organizing heat dissipation optimization network can balance the heat flow state of all hard disks in the network, so that the entire hard disk array system maintains the optimal operating temperature during operation, effectively reducing the impact of temperature fluctuations on hard disk performance and extending the life of the hard disk. Through global thermal balance optimization, the burden on the cooling equipment is reduced, while the utilization efficiency of thermal energy is improved. A more energy-efficient cooling solution is achieved, which not only reduces the energy consumption of the cooling system, but also reduces the demand for additional energy, thereby reducing the energy consumption of the entire hard disk array system.

[0018] Preferably, step S4 comprises the following steps: Step S41: Analyze the long-term trend of temperature data to generate long-term trend data of temperature; Step S42: predicting potential hotspot areas of the solid state disk array data according to the long-term temperature trend data to generate potential hotspot area prediction data; Step S43: Pre-optimize the hard disk heat dissipation path and provide feedback to the hard disk global thermal balance optimization data through the potential hot spot area prediction data, and generate an intelligent temperature control strategy for the solid state hard disk.

[0019] The present invention can predict the long-term change of the temperature of the hard disk array by analyzing the long-term change trend of the temperature data, and identify the law and potential problems of temperature fluctuation in advance. The long-term trend data of the temperature provides a reliable basis for the subsequent thermal management strategy, which is helpful to optimize the long-term operation and maintenance of the hard disk array and prevent the hard disk failure caused by abnormal temperature. Based on the long-term trend data of the temperature, the potential hot spot area prediction of the hard disk array can identify the overheating area in advance, which is convenient for implementing targeted thermal control measures. This prediction ability provides foresight for thermal management, and can effectively intervene before the hot spot occurs in the hard disk array, reducing the loss and failure risk caused by overheating. Through the potential hot spot area prediction data, the heat dissipation path of the hard disk global thermal balance optimization data is optimized in advance, and the heat flow path can be adjusted before the actual overheating occurs, ensuring that the heat can be effectively dispersed to the designated heat dissipation area. This step can not only improve the efficiency of heat flow distribution, but also optimize the heat dissipation system of the hard disk array and reduce the system burden. The generated solid-state hard disk intelligent temperature control strategy can dynamically adjust the heat dissipation configuration of the hard disk array according to the long-term temperature change trend and the prediction of the potential hot spot area, so that it can maintain the optimal temperature level under different working conditions. The intelligent temperature control strategy enhances the adaptive capability of the hard disk array, enabling it to optimize and adjust based on real-time and historical temperature data, effectively avoiding abnormal temperature conditions and improving the stability and operating efficiency of the overall system.

[0020] Preferably, step S42 includes the following steps: Step S421: Calculate the temperature change rate of the temperature long-term trend data to obtain the temperature long-term change rate; Step S422: Performing temperature fluctuation analysis on the long-term temperature change rate by Fourier transform to generate temperature fluctuation data; Step S423: performing hotspot confidence calculation on the solid state drive array data based on the temperature fluctuation data, and performing hotspot diffusion trend analysis on the solid state drive array data based on the hotspot confidence to generate hotspot clusters; dividing the hotspot clusters into data sets to generate a model training set and a model test set; Step S424: Perform model training on the model training set by using a convolutional neural network algorithm to generate a hotspot area prediction pre-model; perform model optimization iteration on the hotspot area prediction pre-model by using a model test set to generate a hotspot area prediction model; Step S425: importing the solid state disk array data into the hot spot area prediction model to perform potential hot spot area prediction, thereby generating hot spot area prediction data.

[0021] The present invention helps to accurately capture the temperature change trend through the calculation of the long-term temperature change rate (step S421), reflecting the temperature fluctuation speed of the hard disk array in different time periods. This data provides a key input for the subsequent thermal management strategy, so that the system can perform real-time temperature control on the hard disk array based on the dynamic temperature change rate to avoid sudden temperature anomalies. The temperature fluctuation characteristics of the hard disk array can be analyzed from the frequency domain by analyzing the temperature fluctuation through Fourier transform (step S422). This method can help the system identify the periodic and non-periodic changes of temperature fluctuations, and provide a more accurate fluctuation characteristic analysis for predicting potential problems and adjusting the hard disk array heat dissipation solution. The calculation of the hot spot confidence based on the temperature fluctuation data (step S423) can predict the hot spot area appearing in the hard disk array, and determine the expansion path of the hot spot area through the hot spot diffusion trend analysis. This step can discover and mark the areas in the hard disk array with large temperature fluctuations and overheating in advance, greatly improving the early warning capability of the system and reducing hard disk failures caused by heat concentration. In the hot spot cluster data set division (step S423), the hard disk array data is divided into a training set and a test set, ensuring the scientificity and effectiveness of the model training and optimization process. Through this division, the accurate matching of the training set and the test set can ensure the high-precision prediction ability of the final model in practical applications and improve the system's ability to accurately identify hot spots.

[0022] In this specification, a solid state drive intelligent temperature control system based on edge computing is provided, which is used to execute the above-mentioned solid state drive intelligent temperature control method based on edge computing. The solid state drive intelligent temperature control system based on edge computing includes: A network construction module is used to obtain solid-state drive array data; perform array topology analysis on the solid-state drive array data to generate solid-state drive array topology data; perform initial self-organizing heat dissipation network construction on the solid-state drive array data based on the solid-state drive array topology data to obtain an initial self-organizing heat dissipation network; The heat conduction path optimization module is used to deploy edge computing nodes for the initial self-organizing heat dissipation network and collect temperature data of the solid-state hard disk according to the edge computing nodes; the network heat conduction path of the initial self-organizing heat dissipation network is optimized through the temperature data to generate a self-organizing heat dissipation optimization network; The heat flow path optimization module is used to calculate the current thermodynamic state of the solid-state hard disk array, and convert the temperature data into a thermal-electrical mode according to the current thermodynamic state to generate heat flow dynamic temperature control data; the heat flow dynamic temperature control data is used to pre-adjust the heat path of the self-organizing heat dissipation optimization network to generate hard disk global thermal balance optimization data; The forward feedback module is used to predict the long-term change trend of temperature data, and perform hard disk optimization feedback in advance on the hard disk global thermal balance optimization data based on the long-term change trend to generate an intelligent temperature control strategy for the solid-state hard disk.

[0023] The beneficial effect of the present invention is that the state information of the hard disk array can be fully collected through the solid-state hard disk array data acquisition and array topology analysis (network construction module), and detailed hard disk array structure data can be obtained through array topology analysis, which provides a solid data foundation and accurate network structure view for subsequent thermal management, optimization and intelligent adjustment, so that the system can deeply understand the working environment of the hard disk array. The deployment of edge computing nodes and temperature data collection (thermal path optimization module) enable the system to monitor the temperature data of the hard disk array in real time, respond quickly and adjust the thermal management strategy. The deployment of edge computing nodes not only improves the accuracy of temperature data collection, but also makes the temperature control system have a higher response speed and intelligence level, and adjusts the heat dissipation path of the hard disk array in real time to avoid overheating. The generation of global thermal balance optimization data of the hard disk (heat flow path optimization module) improves the heat dissipation efficiency of the hard disk array by optimizing the heat flow path and balancing the overall heat distribution of the hard disk array, thereby avoiding local overheating. This optimization can effectively improve the working performance and life of the hard disk array, while reducing energy consumption and extending the stable operation time of the system. By predicting the long-term temperature change trend (feedback module), the system can optimize the hard disk array in advance according to the long-term trend of temperature data. This prediction capability enables the system to predict possible thermal problems in advance and adjust the temperature control strategy of the hard disk array according to the predicted data, thereby achieving more intelligent and accurate thermal management, improving the system's ability to respond to abnormal situations, and reducing potential overheating risks. Therefore, the present invention realizes SSD array-level intelligent temperature control through self-organizing heat dissipation network, edge computing optimization, thermal-electric mode conversion and long-term prediction, improving heat dissipation efficiency, response speed and long-term stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the steps of an intelligent temperature control method for a solid-state hard disk based on edge computing; Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart; Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG. The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0025] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0027] It should be understood that although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, and the term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0028] To achieve this, please refer to Figures 1 to 3 , a solid state hard disk intelligent temperature control method based on edge computing, the method comprising the following steps: Step S1: acquiring solid state drive array data; performing array topology analysis on the solid state drive array data to generate solid state drive array topology data; constructing an initial self-organizing heat dissipation network for the solid state drive array data based on the solid state drive array topology data to obtain an initial self-organizing heat dissipation network; Step S2: deploy edge computing nodes on the initial self-organizing heat dissipation network, and collect temperature data of the solid-state hard disk according to the edge computing nodes; optimize the network thermal conductivity path of the initial self-organizing heat dissipation network through the temperature data to generate a self-organizing heat dissipation optimization network; Step S3: Calculate the current thermodynamic state of the solid-state hard disk array, and perform thermal-electric mode conversion on the temperature data according to the current thermodynamic state to generate heat flow dynamic temperature control data; pre-adjust the heat path of the self-organizing heat dissipation optimization network through the heat flow dynamic temperature control data to generate hard disk global thermal balance optimization data; Step S4: predicting the long-term change trend of the temperature data, and performing hard disk optimization feedback in advance on the hard disk global thermal balance optimization data according to the long-term change trend, and generating a solid state hard disk intelligent temperature control strategy.

[0029] The present invention establishes an efficient heat conduction path through the construction and optimization of a self-organizing heat dissipation network, so that heat can be quickly and evenly distributed and dissipated, avoiding local overheating and improving the overall heat dissipation efficiency. The edge computing node is used to monitor the hard disk temperature in real time, and the global thermal balance is achieved through the heat flow dynamic temperature control technology to ensure that the temperature distribution of all hard disks is within the optimal range, thereby improving the stability of the system. Excessive temperature can cause the performance of the solid-state hard disk to decline or even damage the data. Through the intelligent temperature control strategy, the impact of high temperature can be effectively reduced, the hard disk read and write speed can be ensured to be stable, and the I / O performance loss caused by temperature fluctuations can be reduced. Being in a high temperature environment for a long time will accelerate the wear of the hard disk, and this method optimizes the feedback mechanism in advance to make the temperature control more accurate, thereby reducing the hard disk aging caused by overheating and improving the service life of the hard disk. The long-term temperature change trend prediction model is used to adjust the heat dissipation strategy in advance to avoid the impact caused by sudden temperature changes, realize active temperature control, and reduce the emergency frequency reduction or shutdown protection triggered by sudden overheating of the system. Due to the optimization of thermal management, the hard disk runs in a reasonable temperature range for a long time, which can greatly reduce the physical damage caused by thermal expansion and contraction, as well as the failure of the controller and NAND chip due to overheating, and improve the reliability of data storage. Overheating can increase the bit error rate (BER) of stored data, affecting data integrity. This method ensures that the data storage environment temperature is suitable and reduces the risk of data damage caused by temperature fluctuations through thermal balance optimization and intelligent temperature control strategy. Therefore, the present invention realizes SSD array-level intelligent temperature control through self-organizing heat dissipation network, edge computing optimization, thermal-electric mode conversion and long-term prediction, thereby improving heat dissipation efficiency, response speed and long-term stability.

[0030] In the embodiment of the present invention, reference Figure 1 As shown, it is a schematic diagram of the steps of a solid state hard disk intelligent temperature control method based on edge computing of the present invention. In this example, the solid state hard disk intelligent temperature control method based on edge computing includes the following steps: Step S1: acquiring solid state drive array data; performing array topology analysis on the solid state drive array data to generate solid state drive array topology data; constructing an initial self-organizing heat dissipation network for the solid state drive array data based on the solid state drive array topology data to obtain an initial self-organizing heat dissipation network; In an embodiment of the present invention, by connecting to a solid-state drive array (SSD Array), a storage management interface (such as NVMe, SAS or SATA protocol) is called to obtain information such as the hard disk operation status and data storage structure. Key data indicators such as disk I / O rate, storage capacity distribution, temperature sensor data, data block access frequency, etc. are collected. An original solid-state drive array data set is formed, and a connection structure diagram of the solid-state drive array is constructed, including RAID levels, controller distribution, and data channel paths. Data block distribution, index structure, hot data distribution, etc. are analyzed. Based on sensor data, the temperature distribution of each storage unit is calculated, hot spots are identified, and solid-state drive array topology data is generated for subsequent heat dissipation network construction. Based on the solid-state drive array topology data, the SSD unit is temperature partitioned to form a hot data area mapping, and a heat conduction model (such as Fourier heat conduction equation) is used to calculate the heat diffusion relationship between storage units to establish a preliminary heat dissipation network structure. A dynamic adjustment strategy based on data access mode (such as LSTM or reinforcement learning) is used to optimize the data block migration path and reduce hot spot accumulation. Combined with a hot and cold data distribution strategy based on load balancing, the data writing and reading priorities of the storage unit are automatically adjusted to improve the overall heat dissipation efficiency. Calculate the initial balance of the heat dissipation network to ensure that there are no overload points, and generate the initial self-organizing heat dissipation network including: thermal map (displaying temperature distribution), heat migration path (data block migration strategy), and balance index (measuring heat dissipation balance).

[0031] Step S2: deploy edge computing nodes on the initial self-organizing heat dissipation network, and collect temperature data of the solid-state hard disk according to the edge computing nodes; optimize the network thermal conductivity path of the initial self-organizing heat dissipation network through the temperature data to generate a self-organizing heat dissipation optimization network; In an embodiment of the present invention, a reasonable edge computing node deployment location is selected based on the physical distribution of the solid-state drive (SSD). An edge computing device with low latency and high computing power is selected, such as an embedded processor or FPGA. An initial self-organizing heat dissipation network is deployed in the solid-state drive cluster to establish an initial communication link. The data transmission protocol between nodes is set to ensure that the edge nodes can collect data in real time and optimize the network. Temperature sensors (such as NTC thermistors and MEMS temperature sensors) are deployed on the surface of the SSD and key heat dissipation components. The data sampling period is set, such as temperature collection every 10ms, to ensure the real-time nature of the data. The solid-state drive temperature time series change curve is calculated to form a temperature data stream. Through time series analysis, abnormal temperature fluctuations are identified and future temperature trends are predicted. Based on the SSD temperature data, a heat conduction mathematical model is constructed to set the thermal conductivity parameters between each node. Finite element analysis (FEA) or finite difference method (FDM) is used to calculate the heat transfer path. According to the heat flow distribution, the heat conduction path is dynamically adjusted to reduce the heat accumulation in the hot spot area. An adaptive load balancing algorithm is used to evenly distribute the heat and reduce the local temperature peak. Combine machine learning optimization algorithms (such as reinforcement learning or neural networks) to adjust the heat dissipation path between nodes. Based on temperature feedback, optimize the heat dissipation duct and the heat flow channel of the heat dissipation material in real time. Record the optimized thermal conduction path to form a new self-organizing heat dissipation optimization network topology. This network has better thermal conductivity performance and improves the overall thermal management efficiency of the solid-state drive.

[0032] Step S3: Calculate the current thermodynamic state of the solid-state hard disk array, and perform thermal-electric mode conversion on the temperature data according to the current thermodynamic state to generate heat flow dynamic temperature control data; pre-adjust the heat path of the self-organizing heat dissipation optimization network through the heat flow dynamic temperature control data to generate hard disk global thermal balance optimization data; In an embodiment of the present invention, a hard disk array temperature distribution model is constructed by collecting temperature data of various parts of the solid-state hard disk. The steady-state and transient temperature fields of the hard disk array are calculated by finite element analysis (FEA) or finite difference method (FDM). Parameters such as the specific heat capacity, thermal conductivity, and thermal convection coefficient of each SSD unit are calculated. The heat flux between each SSD unit is calculated to evaluate the heat exchange efficiency. The SSD temperature distribution is converted into a thermal-electric equivalent circuit (similar to a thermal resistance network) by a thermal-electric simulation method. Equivalent thermoelectric parameters are set: Thermal resistance (Rth): corresponds to resistance, indicating heat flow resistance. Thermal capacitance (Cth): corresponds to capacitance, indicating heat storage capacity. Heat flow (Ith): corresponds to current, indicating heat transfer rate. A neural network mapping algorithm or least squares regression is used to convert temperature data into equivalent thermal-electric signal data. The distribution of thermal voltage (Vth) and thermal current (Ith) in the SSD array is calculated to generate dynamic temperature control data for thermal flow. The temperature target is set according to the heat flow gradient. Use optimal control algorithms (such as model predictive control MPC) to calculate the optimal thermal control strategy and generate dynamic average temperature control data for heat flow. Dynamically adjust the thermal conduction path of the self-organizing heat dissipation optimization network based on the dynamic average temperature control data for heat flow. Use an adaptive heat dissipation optimization algorithm to ensure that heat is evenly distributed between each SSD unit. Combine reinforcement learning to optimize the thermal path through a feedback mechanism. Calculate the optimal heat dissipation duct structure to ensure an efficient heat conduction path. Calculate the global heat flow balance factor of the SSD array and optimize the heat exchange path. Use multi-objective optimization methods (such as genetic algorithms) to optimize the thermal balance control strategy and generate global thermal balance optimization data for the hard disk as input to the cooling system to ensure temperature balance during long-term operation.

[0033] Step S4: predicting the long-term change trend of the temperature data, and performing hard disk optimization feedback in advance on the hard disk global thermal balance optimization data according to the long-term change trend, and generating a solid state hard disk intelligent temperature control strategy.

[0034] In an embodiment of the present invention, temperature data, heat flow data, and heat dissipation efficiency data are collected from multiple edge computing nodes. De-noising and interpolation completion are performed to ensure data integrity. The ARIMA (autoregressive integrated moving average) model or LSTM (long short-term memory network) is used to predict the temperature change trend. Identify temperature anomalies to avoid hot spot accumulation. Combined with the thermal-electric conversion model, the temperature data is mapped into thermal voltage (Vth) and thermal current (Ith) data streams. Construct an equivalent heat flow network topology to form a global heat flow model. The finite element method (FEM) is used to calculate the thermal conductivity matrix between each SSD unit. Combined with the Thermal Resistance Network, the heat flow channel is analyzed to find high blockage areas and optimize the heat exchange path. Set the optimization goal so that the temperature of all SSD units tends to be balanced: ,in For the SSD unit temperature, is the global SSD temperature average, is the total number of SSD units. Particle swarm optimization (PSO) or simulated annealing (SA) is used for thermal balance optimization calculation. The optimal heat flow distribution matrix is ​​calculated to ensure that the global temperature tends to be stable. Based on the optimization calculation results, an intelligent temperature control strategy is generated to dynamically adjust the heat dissipation mode of the SSD. Combined with the PID control algorithm, the fan speed, heat pipe power, and heat sink thermal conductivity are adjusted in real time. Combined with workload scheduling, the task allocation of the SSD is intelligently adjusted to reduce the load of the high-temperature SSD unit and increase the load of the low-temperature SSD unit appropriately. Dynamic load migration strategy is calculated to avoid local overheating. Deep reinforcement learning (DQN) is used to optimize the heat dissipation path based on historical temperature data and current working status. The heat transfer direction is optimized by intelligently adjusting the heat dissipation components (such as heat pipes, heat sinks, and liquid cooling modules). For high-temperature areas, phase change materials or thermoelectric cooling devices are enabled for active heat dissipation. The air duct allocation is adjusted for low-temperature areas to improve the global temperature averaging efficiency.

[0035] Preferably, step S1 comprises the following steps: Step S11: Acquire solid state drive array data; Step S12: performing array topology analysis on the solid state drive array data to generate solid state drive array topology data; Step S13: performing intelligent material integration on the solid state drive array using the solid state drive array topology data to generate solid state drive array integrated material data, wherein the intelligent material integration includes integrated phase change materials, thermoelectric materials, and intelligent thermal conductive films; Step S14: performing microfluidic heat conduction path identification on the solid state drive array topology data based on the solid state drive array integrated material data, and generating hard drive self-assembly microstructure data; Step S15: local sensing units are implanted into the solid state hard disk array topology data through the hard disk self-assembly microstructure data to generate an initial self-organizing heat dissipation network, wherein the local sensing units include temperature sensors and heat flow sensing devices.

[0036] In an embodiment of the present invention, an original data set is extracted from a solid-state hard disk storage system, including but not limited to hard disk model, storage capacity, physical connection mode (SATA, NVMe, PCIe), array configuration (RAID 0, RAID 1, RAID5, etc.) and hard disk operation status parameters (temperature, power consumption, read and write speed, etc.). Data acquisition can be automatically collected by an intelligent monitoring system, and the acquired data is formatted to form a solid-state hard disk array data set to ensure the accuracy of subsequent analysis. A topological modeling method is used to analyze the array structure, specifically, based on the hard disk interface type and physical connection mode, a topological node diagram of the solid-state hard disk array is constructed; data transmission paths, power supply paths and heat dissipation channels are analyzed to establish a complete topological structure; based on thermodynamic analysis and data flow calculation, the solid-state hard disk array topology is optimized and adjusted to improve data throughput efficiency and thermal management capabilities; the above analysis results are integrated to generate solid-state hard disk array topology data to provide input data for subsequent steps. Phase change materials (PCM) are introduced in key areas of the hard disk array (such as high-power chip areas and storage controller modules) to improve heat absorption capacity and reduce local temperature peaks; thermoelectric materials (such as Bi2Te3-based thermoelectric modules) are used to achieve temperature difference power generation, and energy efficiency is optimized through energy recovery systems; smart thermal conductive films are introduced between storage units to enhance heat conduction capabilities, and combined with adaptive heat dissipation technology to improve overall heat dissipation performance. Based on the topological data of the solid-state hard disk array, high heat flux areas are identified, and microfluidic channel structures are designed at corresponding locations; combined with the distribution of smart materials, the microfluidic flow path is optimized to achieve uniform heat dissipation; micro-nano structures are introduced on the hard disk casing and PCB board surface to enhance heat exchange efficiency, and hard disk self-assembly microstructure data is generated to provide a basis for the subsequent construction of a heat dissipation network. Temperature sensors are placed around the hard disk control chip, storage unit and radiator to monitor temperature changes in real time. MEMS heat flow sensors are placed in microfluidic channels and key heat dissipation areas to sense heat flow distribution characteristics. Based on the data collected by the sensor unit, an adaptive thermal management algorithm is used to optimize the heat dissipation path, generate an initial self-organizing heat dissipation network, and improve the heat dissipation efficiency and stability of the solid-state hard disk array.

[0037] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: Step S21: deploy edge computing nodes on the initial self-organizing heat dissipation network to obtain edge computing nodes of the heat dissipation network; Step S22: Based on a preset time period, the temperature sensor is used to sense the hard disk temperature of the edge computing node of the heat dissipation network to generate hard disk temperature sensing data; the hard disk temperature sensing data is preprocessed to generate standard hard disk temperature sensing data, wherein the data preprocessing includes data cleaning, data denoising and data standardization; Step S23: performing temperature distribution analysis on the standard hard disk temperature sensing data to generate a hard disk thermal distribution matrix; performing thermal conduction path optimization on the hard disk self-assembled microstructure data based on the hard disk thermal distribution matrix to generate thermal conduction path optimization data; Step S24: regional temperature discrimination is performed on the hard disk temperature sensing data according to the thermal conduction path optimization data. When the hard disk temperature sensing data is greater than or equal to the preset temperature threshold, the initial self-organizing heat dissipation network is reorganized locally through the edge computing node of the heat dissipation network to generate a self-organizing heat dissipation optimization network.

[0038] In the embodiment of the present invention, according to the topological structure of the initial self-organizing heat dissipation network, nodes with strong computing power, large data flow and drastic temperature changes are selected as edge computing nodes; microprocessing units (MCUs) or low-power edge computing modules are deployed on the selected edge computing nodes to realize local thermal management calculations; the data interaction strategy between the edge computing nodes and the central processing unit (CPU) is optimized to ensure that the calculation results can be fed back to the heat dissipation network in real time to realize dynamic thermal management; the configuration and optimization of the edge computing nodes are completed to form a heat dissipation network edge computing node cluster with intelligent computing capabilities. In each preset time period, the temperature sensor of the edge computing node collects the temperature data of the hard disk; the collected hard disk temperature perception data is cleaned and optimized, including: removing abnormal values, such as data deviations caused by sampling errors or hardware failures; using sliding average filtering or wavelet transform methods to remove high-frequency noise and improve data accuracy; normalizing the data or Z-score standardization to ensure the consistency of different batches of data; after data preprocessing, high-quality, analyzable standard hard disk temperature perception data is obtained to provide input for subsequent analysis. Use interpolation algorithms or thermal field simulation algorithms (such as finite element analysis FEM) to analyze the temperature perception data of standard hard disks; generate temperature gradient distribution on the surface and inside of the hard disk, and construct a hard disk thermal distribution matrix, which describes the temperature differences at different locations; based on the hard disk thermal distribution matrix, determine the relative positions of high-temperature areas and low-temperature areas, and identify the main heat conduction pathways; combine the hard disk self-assembly microstructure data to optimize the microfluidic channel, intelligent thermal film layout and thermoelectric material distribution to enhance local heat dissipation capabilities; generate thermal path optimization data as input for subsequent thermal management decisions. Set a temperature threshold (such as 70°C) and monitor the temperature data of each hard disk area in real time; if the temperature of a certain area reaches or exceeds the threshold, it is determined that the area needs to optimize the heat dissipation strategy; trigger the edge computing node of the heat dissipation network to execute a dynamic thermal management algorithm and recalculate the optimal thermal conduction path; optimize the local heat path by controlling the microfluidic channel, adjusting the thermal conductivity of the intelligent thermal film, adjusting the power of the thermoelectric material, etc.; form a self-organized heat dissipation optimization network, which can be adaptively adjusted according to the operating status of the hard disk to improve heat dissipation efficiency and extend the life of the equipment.

[0039] Preferably, optimizing the heat conduction path of the hard disk self-assembly microstructure data based on the hard disk heat distribution matrix includes: Confirming the state of the phase change material in the hard disk self-assembly microstructure data based on the hard disk heat distribution matrix, and obtaining the phase change material state data, wherein the phase change material state data includes solid state data and liquid state data; The heat flow direction and heat exchange rate of the SSD array data are calculated through the hard disk heat distribution matrix, and a local heat flow model is constructed based on the heat flow direction and heat exchange rate; The local heat flow model is used to locate the hard disk hotspot area of ​​the initial self-organized heat dissipation network and generate the hard disk hotspot area; the regional thermal buffer analysis of the hard disk hotspot area is performed to generate the hard disk regional thermal buffer data; based on the dynamic phase change control formula, the local phase change control of the hard disk hotspot area is performed according to the phase change material state data to generate the phase change material solid-liquid area dynamic adjustment data, where the local phase change control formula is as follows: ; In the formula, For phase change material in position The solid-liquid conversion ratio at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to adjust the influence weight of temperature deviation on the PCM solid-liquid conversion ratio, The temperature of the SSD array Over time The rate of change, In order to adjust the influence weight of temperature change rate on PCM solid-liquid conversion ratio, is the horizontal coordinate of the position, is the vertical coordinate of the position; Based on the thermoelectric material in the hard disk self-assembly microstructure data, the electric field application mode of the hard disk hot spot area is confirmed, and the potential difference of the hard disk hot spot area is calculated according to the power plant application mode to obtain the hot spot area potential difference; the hot spot area potential difference is controlled by the thermoelectric material electric field distribution using the heat flow guidance optimization formula to generate dynamic electric field control data, where the heat flow guidance optimization formula is as follows: ; In the formula, For phase change material in position The local potential difference at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to control the influence weight of temperature deviation on TEG potential difference, The temperature of the SSD array Over time The rate of change, In order to control the weight of the influence of temperature change rate on TEG potential difference, is the horizontal coordinate of the position, is the vertical coordinate of the position; According to the dynamic adjustment data of the solid-liquid region of the phase change material and the dynamic electric field control data, the heat diffusion path of the hard disk self-assembly microstructure data is optimized to generate the thermal conduction path optimization data.

[0040] In an embodiment of the present invention, edge computing nodes are deployed at key heat dissipation locations in the initial self-organizing heat dissipation network to form a set of edge computing nodes in the heat dissipation network. A distributed control strategy is adopted to enable edge computing nodes to independently process local temperature perception, heat flow optimization, and heat path reorganization tasks. High-precision temperature sensors are used to regularly collect hard disk temperature data of edge computing nodes in the heat dissipation network to generate initial hard disk temperature perception data. Abnormal data is eliminated to ensure the accuracy of temperature data, and Kalman filtering or wavelet transform is used to remove noise in temperature signals to improve data quality. The data is normalized to unify analysis standards and generate standard hard disk temperature perception data. Based on the standard hard disk temperature perception data, a hard disk heat distribution matrix is ​​constructed to describe the temperature changes in different regions. The state of the phase change material in the self-assembled microstructure of the hard disk is confirmed by the heat distribution matrix, and the state data of the phase change material (including solid data and liquid data) is generated. The heat flow direction and heat exchange rate of the solid-state hard disk array data are calculated, and a local heat flow model is established. The local heat flow model is used to identify the area with higher temperature in combination with the heat distribution matrix to generate hard disk hot spot area data. Calculate the heat transfer delay and heat capacity of the hot spot area to generate hard disk area thermal buffer data. Use the dynamic phase change control formula: ; In the formula, For phase change material in position The solid-liquid conversion ratio at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to adjust the influence weight of temperature deviation on the PCM solid-liquid conversion ratio, The temperature of the SSD array Over time The rate of change, In order to adjust the influence weight of temperature change rate on PCM solid-liquid conversion ratio, is the horizontal coordinate of the position, is the position ordinate; calculate the electric field application mode of the hard disk hot spot area and calculate the potential difference in the hot spot area. Use the heat flow guidance optimization formula to control the electric field: ; In the formula, For phase change material in position The local potential difference at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to control the influence weight of temperature deviation on TEG potential difference, The temperature of the SSD array Over time The rate of change, In order to control the weight of the influence of temperature change rate on TEG potential difference, is the horizontal coordinate of the position, The vertical coordinate of the position; according to the dynamic adjustment data of the solid-liquid region of the phase change material and the dynamic electric field control data, the thermal conductivity path of the hard disk self-assembled microstructure is optimized to generate the final thermal conductivity path optimization data. According to the thermal conductivity path optimization data, the hard disk temperature perception data is analyzed. If the temperature of a certain area is greater than or equal to the set threshold, the local heat path reorganization is triggered. The heat diffusion path is dynamically adjusted through the edge computing node of the heat dissipation network, so that the heat in the overheated area can be more effectively exported. Combined with the phase adjustment of the phase change material and the electric field optimization of the thermoelectric material, more efficient heat dissipation management is achieved, and finally a self-organized heat dissipation optimization network is generated.

[0041] Preferably, the step S3 of calculating the current thermodynamic state of the solid state drive array includes: Extracting operation data of the solid state drive array, and calculating the total power consumption of the solid state drive array according to the operation data; Analyze the thermal power distribution of the SSD array through the total power consumption of the SSD array; Based on the finite element method, the temperature field of the thermal power distribution of the solid-state drive array is simulated to generate the array temperature field model; According to the array temperature field model, the heat exchange amount of the solid state drive array is evaluated and the curve is converted to obtain a heat exchange curve; The time point with a slope of 0 in the heat exchange curve was selected and marked as thermal steady-state time data; The heat exchange amount and thermal steady-state time data of the solid-state drive array are used to perform thermal runaway risk analysis on the operation data of the solid-state drive array, thereby generating the current thermodynamic state of the solid-state drive array.

[0042] In the embodiment of the present invention, the key operating parameters of the solid-state hard disk array are collected, including: workload (read and write IOPS, throughput), power consumption (instantaneous power, current, voltage), temperature data (sensor temperature, ambient temperature), fan status (speed, heat dissipation efficiency), to generate a solid-state hard disk array operation data set. The power is calculated based on the voltage V and current I of the hard disk: P=V×I; the total power consumption of the array is obtained by summing all hard disk nodes: ;in is the number of hard disks, For the The power consumption of each hard disk is recorded, and the total power consumption data is recorded as the basis for subsequent thermal analysis. According to the physical structure and power consumption distribution of the solid-state drive array, the thermal power of different areas is calculated: ;in The ratio of power converted to heat energy (generally 0.8~0.9) is used to generate the hard disk array thermal power distribution data. Based on the physical arrangement structure of the hard disk, a two-dimensional thermal power matrix is ​​constructed. The finite element analysis (FEA) method is used to establish a mathematical model of the temperature field. The finite element difference method is used for discretization and solution to obtain the temperature field distribution under different time steps. The array temperature field model is generated, which includes the temperature change data of each time step. The heat transfer of each hard disk node is calculated through the temperature field model, and the change of heat exchange amount at different time points is recorded to obtain the heat exchange curve. The slope of the heat exchange curve is calculated, and the time point with a slope of 0 is identified. This time point is recorded as the thermal steady-state time data. Combined with the heat exchange amount and thermal steady-state time data, it is determined whether the hard disk temperature continues to rise or fall: if the temperature continues to rise and there is no stable state, there is a risk of thermal runaway. If the temperature fluctuates near the steady-state time point, the system is in normal working condition. Combined with the thermal runaway evaluation results, the current thermodynamic state data of the solid-state hard disk array is output for subsequent thermal management optimization.

[0043] Preferably, the step S3 of converting the temperature data into a thermal-electrical mode according to the current thermodynamic state includes: Perform thermal aggregation analysis on the temperature data according to the current thermodynamic state to generate thermal aggregation data; Use thermal clustering data to perform cluster analysis on SSD array data to generate hard disk hotspots, hard disk cooling areas, and hard disk neutral areas, and eliminate hard disk neutral areas; Through the thermoelectric materials in the SSD array integrated material data, the hot spot area, cooling area and neutral area of ​​the hard disk are analyzed for thermo-electric mode, and the Peltier effect mode and Seebeck effect mode are generated. The optimal heat flow guiding path for the hard disk hot spot area and the hard disk cooling area is calculated, and the electric field direction of the Peltier effect mode and the Seebeck effect mode is adjusted through the optimal heat flow guiding path to generate heat flow dynamic temperature control data.

[0044] In the embodiment of the present invention, the current thermodynamic state data is calculated by the preceding steps.

[0045] Temperature sensor data of the SSD array (temperature distribution of each drive).

[0046] The operation load information of the device (such as IOPS, throughput, etc.). Use spatial interpolation methods (such as Kriging interpolation and inverse distance weighted method) to continuously model the temperature data of different hard disks, obtain the heat aggregation function, calculate the heat aggregation degree of different regions, form a heat aggregation data matrix, set K=3, and cluster the temperatures of all hard disks to obtain: hard disk hot spot area (high temperature area), hard disk cooling area (low temperature area), hard disk neutral area (moderate temperature, no significant heat dissipation demand). Since the hard disk neutral area has no obvious effect on heat flow control, it does not participate in the subsequent thermal-electric mode conversion, and only the hot spot area and cooling area are retained for thermal-electric optimization. In the hard disk hot spot area, the Peltier effect is used for active heat dissipation: Qc=πI; where: Qc is the heat dissipation, π is the Peltier coefficient, and I is the current. By controlling the direction of the current, the heat in the hot spot area is actively extracted to reduce the temperature. In the hard disk cooling area, the Seebeck effect is used to recover heat energy: V=S·ΔT; where: V is the thermovoltage, S is the Seebeck coefficient, and ΔT is the temperature difference. This mode is used to convert the temperature difference into electrical energy and optimize energy utilization. The heat conduction model is established using the Fourier heat conduction equation: q=-k▽T; where: q is the heat flux density, k is the thermal conductivity of the material, and ▽T is the temperature gradient. The shortest path algorithm (such as the Dijkstra algorithm) is used to calculate the optimal heat flow path from the hot spot area to the cooling area: ;in: is the thermal resistance of each point on the path. The electric field direction of the Peltier and Seebeck modes is optimized according to the heat flow path to enable efficient heat transfer. By optimizing the current direction, the electric field direction matches the heat flow path: E=-▽V; where: E is the electric field, V is the electric potential, and the heat flow dynamic temperature control data is generated to adjust the working mode of the Peltier effect and Seebeck effect in real time to improve the heat dissipation efficiency.

[0047] Preferably, the step S3 of pre-adjusting the heat path of the self-organizing heat dissipation optimization network by using the heat flow dynamic temperature control data includes: The heat transfer topology structure of the self-organizing heat dissipation optimization network is mapped through the heat flow dynamic temperature control data to generate a heat transfer path diagram; the heat transfer path impedance characteristics are extracted from the heat transfer path diagram to obtain the heat transfer path impedance characteristic data; The optimization target is set based on the impedance characteristic data of the heat transfer path, and the heat transfer path diagram is dynamically allocated according to the optimization target to generate the heat path control data; The thermal path control data is used to optimize the global network thermal balance of the self-organizing heat dissipation optimization network and generate the global thermal balance optimization data of the hard disk.

[0048] In an embodiment of the present invention, the heat flow dynamic temperature control data is input, including the temperature distribution of the hard disk array, the heat flow path and the thermal balance control strategy. The spatial distribution data of the hard disk array (the location of the hard disk, the location of the heat source, etc.). The heat transfer network of the hard disk array is modeled according to the heat flow dynamic temperature control data: the thermal connection relationship between each hard disk in the array and the heat dissipation device (such as a heat sink, a fan, etc.) is graphically represented. Each node (hard disk or heat dissipation device) represents a heat source or heat dissipation point, each edge represents a heat flow path, and the weight represents the thermal resistance (the impedance characteristic of the thermal path). A heat transfer path diagram is generated in the form of a directed graph. For each path in the heat transfer path diagram, its thermal impedance characteristics are calculated. Thermal impedance is calculated according to the heat conduction equation: ;in is the temperature difference, is the heat flow, is the length of the transmission path, is the thermal conductivity of the material, is the cross-sectional area of ​​the heat transfer path. The thermal impedance characteristic data of each path is extracted to form an impedance matrix. Based on the impedance characteristic data of the heat transfer path, the optimization goals are set, which mainly include: by reducing the thermal resistance on the heat transfer path, ensuring that heat can be efficiently transferred from the hot spot area to the cooling area. Optimize the path allocation to achieve thermal balance of the hard disk array to avoid local overheating or insufficient cooling. According to the set optimization goals, the heat flow path allocation is optimized through a dynamic path allocation algorithm (such as a greedy algorithm or a simulated annealing algorithm): dynamically adjust the allocation of the heat flow path to make the temperature distribution as balanced as possible and reduce the temperature difference between the hot spot area and the cooling area. Determine the heat flow distribution of each heat flow path so that the heat flow passes through the path with the least thermal resistance, generate heat path control data, and record the heat flow and adjustment strategy of each heat flow path. In the self-organizing heat dissipation optimization network, perform global network thermal balance optimization. This process adjusts the heat flow path and heat source distribution through a global optimization algorithm (such as a genetic algorithm or a particle swarm optimization) to optimize the heat transfer network. Consider the thermal behavior and heat transfer relationship of each hard disk in the hard disk array to ensure uniform overall heat distribution. Optimize the synergy between heat source and heat sink so that heat flow can be dispersed to the greatest extent in the hard disk array. Output hard disk global thermal balance optimization data, including optimized heat flow path, heat control strategy of heat sink, and temperature distribution data under thermal balance state. Ensure that the temperature difference of all hard disks in the array is minimized and effectively avoid hot spot problems.

[0049] As an example of the present invention, refer to Figure 3 As shown, in this example, step S4 includes: Step S41: Analyze the long-term trend of temperature data to generate long-term trend data of temperature; Step S42: predicting potential hotspot areas of the solid state disk array data according to the long-term temperature trend data to generate potential hotspot area prediction data; Step S43: Pre-optimize the hard disk heat dissipation path and provide feedback to the hard disk global thermal balance optimization data through the potential hot spot area prediction data, and generate an intelligent temperature control strategy for the solid state hard disk.

[0050] In an embodiment of the present invention, historical temperature data is input, including temperature change data of the solid state drive array in different time periods. Timestamp data: time stamp of temperature data, used to analyze the temperature change trend. Based on historical temperature data, the long-term change trend of temperature data is analyzed by time series analysis method (such as sliding average method, exponential smoothing method, or trend analysis algorithm). The average change trend of temperature in different time dimensions such as year, month, day, hour, etc. is calculated to identify seasonal fluctuations, periodic changes and trend changes of temperature. Regression analysis (such as linear regression, polynomial regression) or machine learning method (such as support vector regression) is used to predict the temperature change trend in the future. According to the above analysis, long-term temperature trend data is generated, which includes the long-term change trend of temperature, periodic fluctuation law, and future predicted temperature curve. The long-term temperature trend data is in the following form: T(t)=T0+ΔT(t), where T(t) represents the predicted temperature at time t, T0 is the initial temperature, and ΔT(t) is the temperature change value obtained based on time series analysis. Based on the long-term temperature trend data and the working data of the hard disk array, the potential hotspot areas are predicted using thermodynamic modeling and machine learning methods (such as clustering algorithms, decision trees, random forests, etc.). Considering the usage load, power consumption data and temperature change trend of the hard disk, the areas where the temperature will rise due to increased load and power consumption in the future period are identified. Different areas of the hard disk array are partitioned to analyze the thermal sensitivity of each area under different load and temperature conditions. Using the heat conduction model, the temperature changes of each hard disk and its surrounding areas are predicted, and then the potential hotspot areas are identified. The temperature prediction results of each area are combined with the actual load conditions to identify the areas that will become hotspots in the future, and the potential hotspot area prediction data is generated. The data includes the predicted temperature of each hard disk and hard disk array area and the potential hotspot areas. Example of hotspot area prediction data: predicted temperature and hotspot warning value of each hard disk area. According to the potential hotspot area prediction data, the current heat dissipation path is optimized in advance. The optimization goal is to avoid overheating near the hotspot area as much as possible, and to adjust the heat flow path so that the heat flow can be smoothly transferred and dissipated to the cooling area. Redesign the heat flow path using heat conduction simulation and optimization algorithms, such as particle swarm optimization (PSO) and simulated annealing. Enhance the heat flow transmission path in the hot spot area by adjusting the heat sink layout and fan speed regulation of the hard disk array to avoid heat accumulation. Generate an intelligent temperature control strategy for the solid-state drive based on the optimized heat dissipation path and predicted data. The strategy includes: adjusting the fan speed and the working status of the cooling device according to the actual temperature data, and performing temperature control in a timely manner. Before the potential hot spot area appears, the system automatically warns and makes necessary heat flow adjustments. Based on the temperature change trend and the predicted data of the hot spot area, dynamically adjust the heat flow path to achieve global thermal balance optimization.

[0051] Preferably, step S42 includes the following steps: Step S421: Calculate the temperature change rate of the temperature long-term trend data to obtain the temperature long-term change rate; Step S422: Performing temperature fluctuation analysis on the long-term temperature change rate by Fourier transform to generate temperature fluctuation data; Step S423: performing hotspot confidence calculation on the solid state drive array data based on the temperature fluctuation data, and performing hotspot diffusion trend analysis on the solid state drive array data based on the hotspot confidence to generate hotspot clusters; dividing the hotspot clusters into data sets to generate a model training set and a model test set; Step S424: Perform model training on the model training set by using a convolutional neural network algorithm to generate a hotspot area prediction pre-model; perform model optimization iteration on the hotspot area prediction pre-model by using a model test set to generate a hotspot area prediction model; Step S425: importing the solid state disk array data into the hot spot area prediction model to perform potential hot spot area prediction, thereby generating hot spot area prediction data.

[0052] In an embodiment of the present invention, the temperature change rate is calculated by performing time difference on the temperature long-term trend data. The temperature change rate can be calculated on multiple time scales (such as hours, days, and months) to generate corresponding temperature long-term change rate data. The temperature long-term change rate data is subjected to frequency domain analysis by Fourier transform. Fourier transform can help extract the periodic fluctuation part in the temperature change. The data after Fourier transform can show the intensity of temperature fluctuation at different frequencies, revealing the periodicity and fluctuation pattern of temperature change. Based on the temperature fluctuation data, the hotspot confidence of each hard disk area is calculated by analyzing the temperature change rate of each hard disk area. The hotspot confidence can indicate the possibility of a certain hard disk area becoming a hotspot. According to the hotspot confidence data, the influence of temperature fluctuation on the hard disk array is analyzed, and the diffusion trend of the hotspot is predicted. A spatial diffusion model can be used to generate the diffusion trend of the hotspot area by considering the spatial propagation characteristics of the temperature change. Based on the hotspot confidence and the diffusion trend, the areas in the hard disk array are clustered and analyzed to form multiple hotspot clusters. Each hotspot cluster contains hard disk areas with similar temperature change characteristics. The clustering method can use clustering algorithms such as K-means or DBSCAN to divide the hotspot clusters into data sets, generate model training sets and model test sets; and use convolutional neural networks (CNNs) to train the model training sets. CNN can effectively process data with spatial structural characteristics, so it is suitable for hotspot area prediction in this task. The spatial features of the data are extracted through multiple convolutional layers to generate a hotspot area prediction pre-model. Through cross-validation and hyperparameter adjustment, the trained pre-model is optimized and iterated using the model test set, and the parameters such as the network structure and learning rate of CNN are adjusted to improve the prediction accuracy of the model. The back propagation algorithm is used to optimize the model so that it can better capture the laws of temperature changes and hotspot diffusion. The hotspot area prediction model is output, that is, the optimized and trained convolutional neural network model, which can be used to predict hotspot areas for future temperature data. The real-time data of the solid-state drive array is input into the trained hotspot area prediction model, and the model predicts possible hotspot areas in the future based on the input data. The prediction results include the temperature prediction of each hard disk area, the hotspot confidence, and whether the area is likely to become a hotspot.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A solid state hard disk intelligent temperature control method based on edge computing, characterized in that: The following steps are involved: Step S1: Acquire solid state drive array data; Performing array topology analysis on the solid-state hard disk array data to generate solid-state hard disk array topology data; constructing an initial self-organizing heat dissipation network for the solid-state hard disk array data based on the solid-state hard disk array topology data to obtain an initial self-organizing heat dissipation network; Step S2: deploy edge computing nodes on the initial self-organizing heat dissipation network, and collect temperature data of the solid-state hard disk according to the edge computing nodes; optimize the network thermal conductivity path of the initial self-organizing heat dissipation network through the temperature data to generate a self-organizing heat dissipation optimization network; Step S3: Calculate the current thermodynamic state of the solid-state hard disk array, and perform thermal-electric mode conversion on the temperature data according to the current thermodynamic state to generate heat flow dynamic temperature control data; pre-adjust the heat path of the self-organizing heat dissipation optimization network through the heat flow dynamic temperature control data to generate hard disk global thermal balance optimization data; Step S4: predicting the long-term change trend of the temperature data, and performing hard disk optimization feedback in advance on the hard disk global thermal balance optimization data according to the long-term change trend, and generating a solid state hard disk intelligent temperature control strategy.

2. The solid state hard disk intelligent temperature control method based on edge computing according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Acquire solid state drive array data; Step S12: performing array topology analysis on the solid state drive array data to generate solid state drive array topology data; Step S13: performing intelligent material integration on the solid state drive array using the solid state drive array topology data to generate solid state drive array integrated material data, wherein the intelligent material integration includes integrated phase change materials, thermoelectric materials, and intelligent thermal conductive films; Step S14: performing microfluidic heat conduction path identification on the solid state drive array topology data based on the solid state drive array integrated material data, and generating hard drive self-assembly microstructure data; Step S15: local sensing units are implanted into the solid state hard disk array topology data through the hard disk self-assembly microstructure data to generate an initial self-organizing heat dissipation network, wherein the local sensing units include temperature sensors and heat flow sensing devices.

3. The solid state hard disk intelligent temperature control method based on edge computing according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: deploy edge computing nodes on the initial self-organizing heat dissipation network to obtain edge computing nodes of the heat dissipation network; Step S22: Based on a preset time period, the temperature sensor is used to sense the hard disk temperature of the edge computing node of the heat dissipation network to generate hard disk temperature sensing data; the hard disk temperature sensing data is preprocessed to generate standard hard disk temperature sensing data, wherein the data preprocessing includes data cleaning, data denoising and data standardization; Step S23: performing temperature distribution analysis on the standard hard disk temperature sensing data to generate a hard disk thermal distribution matrix; performing thermal conduction path optimization on the hard disk self-assembled microstructure data based on the hard disk thermal distribution matrix to generate thermal conduction path optimization data; Step S24: regional temperature discrimination is performed on the hard disk temperature sensing data according to the thermal conduction path optimization data. When the hard disk temperature sensing data is greater than or equal to the preset temperature threshold, the initial self-organizing heat dissipation network is reorganized locally through the edge computing node of the heat dissipation network to generate a self-organizing heat dissipation optimization network.

4. The solid state hard disk intelligent temperature control method based on edge computing according to claim 3 is characterized in that: The optimization of the thermal path of the hard disk self-assembly microstructure data based on the hard disk thermal distribution matrix includes: Confirming the state of the phase change material in the hard disk self-assembly microstructure data based on the hard disk heat distribution matrix, and obtaining the phase change material state data, wherein the phase change material state data includes solid state data and liquid state data; The heat flow direction and heat exchange rate of the SSD array data are calculated through the hard disk heat distribution matrix, and a local heat flow model is constructed based on the heat flow direction and heat exchange rate; The local heat flow model is used to locate the hard disk hotspot area of ​​the initial self-organized heat dissipation network and generate the hard disk hotspot area; the regional thermal buffer analysis of the hard disk hotspot area is performed to generate the hard disk regional thermal buffer data; based on the dynamic phase change control formula, the local phase change control of the hard disk hotspot area is performed according to the phase change material state data to generate the phase change material solid-liquid area dynamic adjustment data, where the local phase change control formula is as follows: ； In the formula, For phase change material in position The solid-liquid conversion ratio at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to adjust the influence weight of temperature deviation on the PCM solid-liquid conversion ratio, The temperature of the SSD array Over time The rate of change, In order to adjust the influence weight of temperature change rate on PCM solid-liquid conversion ratio, is the horizontal coordinate of the position, is the position ordinate; Based on the thermoelectric material in the hard disk self-assembly microstructure data, the electric field application mode of the hard disk hot spot area is confirmed, and the potential difference of the hard disk hot spot area is calculated according to the power plant application mode to obtain the hot spot area potential difference; the hot spot area potential difference is controlled by the thermoelectric material electric field distribution using the heat flow guidance optimization formula to generate dynamic electric field control data, where the heat flow guidance optimization formula is as follows: ； In the formula, For phase change material in position The local potential difference at For SSD arrays in location The current temperature at Set the optimal operating temperature for the system. In order to control the influence weight of temperature deviation on TEG potential difference, The temperature of the SSD array Over time The rate of change, In order to control the weight of the influence of temperature change rate on TEG potential difference, is the horizontal coordinate of the position, is the position ordinate; According to the dynamic adjustment data of the solid-liquid region of the phase change material and the dynamic electric field control data, the heat diffusion path of the hard disk self-assembly microstructure data is optimized to generate the thermal conduction path optimization data.

5. The solid state hard disk intelligent temperature control method based on edge computing according to claim 1 is characterized in that: The step S3 of calculating the current thermodynamic state of the solid state drive array includes: Extracting operation data of the solid state drive array, and calculating the total power consumption of the solid state drive array according to the operation data; Analyze the thermal power distribution of the SSD array through the total power consumption of the SSD array; Based on the finite element method, the temperature field of the thermal power distribution of the solid-state drive array is simulated to generate the array temperature field model; According to the array temperature field model, the heat exchange amount of the solid state drive array is evaluated and the curve is converted to obtain a heat exchange curve; The time point with a slope of 0 in the heat exchange curve was selected and marked as thermal steady-state time data; The heat exchange amount and thermal steady-state time data of the solid-state drive array are used to perform thermal runaway risk analysis on the operation data of the solid-state drive array, thereby generating the current thermodynamic state of the solid-state drive array.

6. The solid state hard disk intelligent temperature control method based on edge computing according to claim 1 is characterized in that: The step S3 of converting the temperature data into a thermal-electrical mode according to the current thermodynamic state includes: Perform thermal aggregation analysis on the temperature data according to the current thermodynamic state to generate thermal aggregation data; Use thermal clustering data to perform cluster analysis on SSD array data to generate hard disk hotspots, hard disk cooling areas, and hard disk neutral areas, and eliminate hard disk neutral areas; Through the thermoelectric materials in the SSD array integrated material data, the hot spot area, cooling area and neutral area of ​​the hard disk are analyzed for thermo-electric mode, and the Peltier effect mode and Seebeck effect mode are generated. The optimal heat flow guiding path for the hard disk hot spot area and the hard disk cooling area is calculated, and the electric field direction of the Peltier effect mode and the Seebeck effect mode is adjusted through the optimal heat flow guiding path to generate heat flow dynamic temperature control data.

7. The solid state hard disk intelligent temperature control method based on edge computing according to claim 1, characterized in that: The step S3 of pre-adjusting the heat path of the self-organized heat dissipation optimization network by using the heat flow dynamic temperature control data includes: The heat transfer topology structure of the self-organizing heat dissipation optimization network is mapped through the heat flow dynamic temperature control data to generate a heat transfer path diagram; the heat transfer path impedance characteristics are extracted from the heat transfer path diagram to obtain the heat transfer path impedance characteristic data; The optimization target is set based on the impedance characteristic data of the heat transfer path, and the heat transfer path diagram is dynamically allocated according to the optimization target to generate the heat path control data; The thermal path control data is used to optimize the global network thermal balance of the self-organizing heat dissipation optimization network and generate the global thermal balance optimization data of the hard disk.

8. The solid state hard disk intelligent temperature control method based on edge computing according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Analyze the long-term trend of temperature data to generate long-term trend data of temperature; Step S42: predicting potential hotspot areas of the solid state disk array data according to the long-term temperature trend data to generate potential hotspot area prediction data; Step S43: Pre-optimize the hard disk heat dissipation path and provide feedback to the hard disk global thermal balance optimization data through the potential hot spot area prediction data, and generate an intelligent temperature control strategy for the solid state hard disk.

9. The solid state hard disk intelligent temperature control method based on edge computing according to claim 8, characterized in that: Step S42 includes the following steps: Step S421: Calculate the temperature change rate of the temperature long-term trend data to obtain the temperature long-term change rate; Step S422: Performing temperature fluctuation analysis on the long-term temperature change rate by Fourier transform to generate temperature fluctuation data; Step S423: performing hotspot confidence calculation on the solid state drive array data based on the temperature fluctuation data, and performing hotspot diffusion trend analysis on the solid state drive array data based on the hotspot confidence to generate hotspot clusters; dividing the hotspot clusters into data sets to generate a model training set and a model test set; Step S424: Perform model training on the model training set by using a convolutional neural network algorithm to generate a hotspot area prediction pre-model; perform model optimization iteration on the hotspot area prediction pre-model by using a model test set to generate a hotspot area prediction model; Step S425: importing the solid state disk array data into the hot spot area prediction model to perform potential hot spot area prediction, thereby generating hot spot area prediction data.

10. A solid state hard disk intelligent temperature control system based on edge computing, characterized in that: Used to execute the solid-state hard disk intelligent temperature control method based on edge computing as claimed in claim 1, the solid-state hard disk intelligent temperature control system based on edge computing includes: A network construction module is used to obtain solid-state drive array data; perform array topology analysis on the solid-state drive array data to generate solid-state drive array topology data; perform initial self-organizing heat dissipation network construction on the solid-state drive array data based on the solid-state drive array topology data to obtain an initial self-organizing heat dissipation network; The heat conduction path optimization module is used to deploy edge computing nodes for the initial self-organizing heat dissipation network and collect temperature data of the solid-state hard disk according to the edge computing nodes; the network heat conduction path of the initial self-organizing heat dissipation network is optimized through the temperature data to generate a self-organizing heat dissipation optimization network; The heat flow path optimization module is used to calculate the current thermodynamic state of the solid-state hard disk array, and convert the temperature data into a thermal-electrical mode according to the current thermodynamic state to generate heat flow dynamic temperature control data; the heat flow dynamic temperature control data is used to pre-adjust the heat path of the self-organizing heat dissipation optimization network to generate hard disk global thermal balance optimization data; The forward feedback module is used to predict the long-term change trend of temperature data, and perform hard disk optimization feedback in advance on the hard disk global thermal balance optimization data based on the long-term change trend to generate an intelligent temperature control strategy for the solid-state hard disk.

Images