Temperature self-adapting based micro host dynamic flow guiding temperature control system
Patent Information
- Application Number
- CN202610925498.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-25
- Publication Date
- 2026-09-25
AI Technical Summary
微型主机CPU、GPU的负载状态是影响其温度变化的重要因素，在高负载状态下容易快速产生较多热量，目前微型主机温度控制往往仅通过监测温度状态来调节风扇转速进而实现散热，然而，该方式难以对温度变化做出及时的响应，缺乏负载对温度变化影响的分析，存在微型主机温控效率不足、精度较低的缺陷
本申请通过在功率修正模块，分别对各时刻CPU、GPU的功率进行修正，能够动态评估负载状态下的功率消耗，有助于实时调整温控策略，根据CPU、GPU在执行程序时功率的变化状态，完成对功率的修正，提高了各时刻CPU、GPU的功率的测量准确性，从而有助于获取更为准确的CPU、GPU的功耗状态，提升CPU和GPU在不同工作负载下的功率管理效率，增强了功率修正与温度控制的精准性；进一步，通过主机温控模块分析CPU、GPU功耗与温度之间的相关性，以及历史温度趋势，能够精确评估不同时间窗口内的温升显著系数，便于提前预判温度变化趋势，有助于散热装置在高温负载时能够快速响应，调节风扇转速，以达到最佳散热效果；然后，进一步考虑功耗状态对其温度变化的影响特征以及微型主机内不同部件之间的温度分布差异特征，对微型主机散热的需求程度进行分析，得到微型主机的散热需求系数，反映了微型主机需要进行散热处理的迫切程度，基于此对散热风扇的转速进行调控，提高了散热风扇转速控制的合理性与适宜性，从而提升了微型主机温控效率与精度。
Smart Images
Figure CN122816420A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of host temperature control technology, specifically to a micro-host dynamic airflow temperature control system based on temperature adaptation. Background Technology
[0002] A microcomputer is a small, powerful computer device designed to provide users with a more flexible, portable, and efficient way to work or play. With continuous technological advancements, the performance of microcomputers has been continuously improved while maintaining a small size and low power consumption, making them widely used in various scenarios.
[0003] Mini PCs typically use a combination of heatsinks and small fans for cooling. Alloy or copper heatsinks, with their high thermal conductivity, transfer heat to the air, while cooling fans further accelerate airflow to dissipate heat from within the PC. Since the internal space of a mini PC is usually limited, airflow is restricted, necessitating the use of air ducts to dissipate heat from specific areas quickly. The load status of the CPU and GPU in a mini PC is a significant factor affecting its temperature changes; under high load, a large amount of heat can be generated rapidly. Currently, mini PC temperature control often relies solely on monitoring temperature and adjusting fan speed to achieve cooling. However, this method is insufficient for timely responses to temperature changes and lacks analysis of the impact of load on temperature variations, resulting in inefficient and inaccurate temperature control for mini PCs. Summary of the Invention
[0004] To address the aforementioned technical problems, the purpose of this application is to provide a temperature-adaptive micro-host dynamic airflow temperature control system, the specific technical solution of which is as follows: This application proposes a temperature-adaptive microcomputer dynamic flow-guiding temperature control system, the system comprising: The data acquisition module is used to collect the temperature of each component inside the microcomputer at various times, as well as the power of the CPU and GPU inside the microcomputer at various times. The power correction module corrects the power of the CPU and GPU at each time point based on the rate of change of CPU and GPU power at each time point, as well as the power change state of CPU and GPU when executing programs, in order to determine the power consumption of CPU and GPU within each time window. The host temperature control module is used to analyze the correlation between the power consumption of the CPU and GPU and their average temperature in each time window, as well as the trend of the average temperature of the CPU and GPU in the historical time windows of each time window, to determine the temperature rise significance coefficient of the CPU and GPU in each time window, and to obtain the high temperature significance coefficient of the micro-host in each time window. The fan speed adjustment module is used to obtain the temperature deviation coefficient of the microcomputer in each time window by measuring the temperature difference between different components inside the microcomputer in each time window, and to determine the heat dissipation demand coefficient of the microcomputer in each time window by combining the high temperature significance coefficient, so as to control the speed of the cooling fan of the microcomputer.
[0005] In one embodiment, the step of correcting the power of the CPU and GPU at each time step includes: For the CPU, the power of its acquisition at all times is nonlinearly fitted to obtain the power fitting curve of the CPU. The slope of the power fitting curve at each time is calculated. If the positive and negative signs of the slopes at multiple consecutive times are consistent, the power of the next time of the time period corresponding to the multiple consecutive times is taken as the power correction value of the multiple consecutive times, and the power correction value of the other times is its acquisition power value. For the GPU, the same power correction method as for the CPU is used to obtain the power correction value of the GPU at each time step.
[0006] In one embodiment, determining the significant coefficients of temperature rise for the CPU and GPU in each time window includes: For the CPU, calculate the average temperature and power consumption at all times within each time window, and calculate the metric distance of the average temperature and power consumption of the CPU in each time window and its historical time windows, which is denoted as the first metric distance. Trend analysis is used to extract the trend characteristics of the average CPU temperature in each time window and its historical time windows. The significant coefficient of CPU temperature rise in each time window is positively correlated with the trend characteristics and negatively correlated with the first metric distance. The same calculation method as that used for the CPU's temperature rise significance coefficient was employed to obtain the GPU's temperature rise significance coefficient for each time window.
[0007] In one embodiment, further determination of the CPU temperature rise significance coefficient in each time window includes: The exponential mapping result of the trend feature is determined, and the first metric distance is forward mapped. The significant coefficient of CPU temperature rise in each time window is the ratio of the exponential mapping result to the forward mapping result.
[0008] In one embodiment, the high temperature significance coefficient of the microcomputer under each time window is the fusion result of the temperature rise significance coefficient of the CPU under each time window and the temperature rise significance coefficient of the GPU under each time window.
[0009] In one embodiment, obtaining the temperature deviation coefficient of the microcontroller under each time window includes: Calculate the average temperature of each component in the microcomputer at all times within each time window, determine the metric distance between any two components of the microcomputer in each time window and its historical time window, and record it as the second metric distance. Merge all the second metric distances of the microcomputer in each time window to obtain the temperature deviation coefficient of the microcomputer in each time window.
[0010] In one embodiment, the temperature deviation coefficient of the microcontroller in each time window is the average of all the second metric distances of the microcontroller in each time window.
[0011] In one embodiment, the heat dissipation demand coefficient of the microcomputer in each time window is the product of the temperature deviation coefficient and the high temperature significance coefficient of the microcomputer in each time window.
[0012] In one embodiment, controlling the speed of the cooling fan of the microcomputer includes: A smoothing algorithm is used to obtain the smoothed result of the heat dissipation demand coefficient of the microcomputer under each time window. Based on the smoothed result, the speed of the cooling fan of the microcomputer is controlled by PWM pulse speed regulation.
[0013] In one embodiment, the smoothing result is used as the duty cycle of PWM control to adjust the speed of the cooling fan in the next time window of each time window.
[0014] This application has the following beneficial effects: This application, through a power correction module, corrects the power consumption of the CPU and GPU at various times, enabling dynamic evaluation of power consumption under load. This facilitates real-time adjustment of temperature control strategies. Based on the power changes of the CPU and GPU during program execution, power correction is performed, improving the accuracy of CPU and GPU power measurement at each moment. This leads to more accurate CPU and GPU power consumption status, improving power management efficiency under different workloads and enhancing the precision of power correction and temperature control. Furthermore, by analyzing the correlation between CPU and GPU power consumption and temperature, as well as historical temperature trends, the host temperature control module can accurately assess the temperature rise significance coefficient within different time windows. This facilitates early prediction of temperature change trends, enabling the cooling device to respond quickly under high-temperature loads and adjust fan speed for optimal heat dissipation. Then, considering the impact of power consumption status on temperature changes and the temperature distribution differences among different components within the micro-host, the application analyzes the micro-host's heat dissipation needs, obtaining a heat dissipation demand coefficient. This coefficient reflects the urgency of heat dissipation for the micro-host, and based on this, the fan speed is adjusted, improving the rationality and suitability of fan speed control, thereby enhancing the micro-host's temperature control efficiency and accuracy. Attached Figure Description
[0015] Figure 1 A block diagram of a temperature-adaptive micro-host dynamic flow temperature control system provided in one embodiment of this application; Figure 2 This is a diagram showing the adjustment of the cooling fan speed for a miniature computer. Detailed Implementation
[0016] The following description, in conjunction with the accompanying drawings, details the specific scheme of the temperature-adaptive micro-host dynamic flow temperature control system provided in this application.
[0017] Please see Figure 1 The diagram illustrates a block diagram of a temperature-adaptive micro-host dynamic airflow temperature control system according to an embodiment of this application. The system includes: a data acquisition module 101, a power correction module 102, a host temperature control module 103, and a fan speed adjustment module 104.
[0018] The data acquisition module 101 collects the temperature of each component inside the microcomputer at various times, as well as the power of the CPU and GPU inside the microcomputer at various times.
[0019] The heat generated during the operation of the microcomputer mainly comes from the CPU, GPU, motherboard, memory, and hard drive—the various components within the microcomputer. This embodiment uses thermal sensors to monitor the temperature of each component in real time. The load status of the CPU and GPU is a significant factor affecting the host temperature; under high load, they tend to generate a large amount of heat rapidly. Therefore, this embodiment collects power data from the sensors built into the CPU and GPU. This embodiment sets the temperature data acquisition interval to 0.5 seconds and the power data acquisition interval to 0.01 seconds. Implementers can set the acquisition interval according to their actual needs; this embodiment does not impose any restrictions on this.
[0020] The power correction module 102 corrects the power of the CPU and GPU at each time point based on the rate of change of power of the CPU and GPU at each time point and the power change state of the CPU and GPU when executing the program, so as to determine the power consumption of the CPU and GPU in each time window.
[0021] Temperature is a crucial factor in the performance of a mini PC, while noise level is an important indicator of user experience. To ensure adequate heat dissipation and safe operation, fan speed should be minimized to reduce noise. Frequent load variations can cause significant temperature fluctuations, and fan speed adjustments can generate substantial non-steady-state noise. Therefore, a stable, efficient, and low-noise temperature control system is essential for mini PCs.
[0022] First, taking the temperature change characteristics of GPU power consumption as an example, the GPU's power consumption increases significantly when running high-resolution games, performing image rendering, or deep learning tasks. However, general detection techniques struggle to directly and accurately measure GPU power data, and power hysteresis occurs when using built-in sensors to collect power data. Specifically, the collected power data gradually approaches the actual power consumption, especially during program execution and termination, and cannot reflect the actual power usage of the device in real time. Consequently, the power consumption data obtained by integrating the power data over the program's runtime is inaccurate. Therefore, this embodiment calibrates the collected power data, including both CPU and GPU power data.
[0023] Specifically, taking a GPU as an example, there will be a discrepancy between the measured power and the true power of the GPU. In idle state, the processor may have a basic power consumption, and the power in idle state is in a non-zero stable state. The power measured by the sensor at this time is the true power data. Taking a single-core processor as an example, from the start to the end of a task, the processor's true power is constant, while the measured power gradually increases from idle power to true power after the task starts. After the program ends, the true power immediately returns to idle power, while the measured power exhibits a certain trailing effect, gradually decreasing from true power to idle power. Therefore, to further improve the accuracy of temperature monitoring, the measured power data is corrected. This embodiment uses the maximum power value during each program's running phase as the processor's true power data, and the idle power data after the program ends as the processor's true power data.
[0024] Since the program's execution causes changes in the rate of change of the measured power, taking the GPU as an example, the least squares method is used to obtain the power fitting curve for all moments of the GPU, denoted as the power fitting curve. Further, the slope value at each moment on the fitting curve is calculated. The time interval where the slope value is greater than 0 for multiple consecutive moments is designated as the first correction interval, and the power data within the first correction interval is corrected to the first collected power value after that interval. The time interval where the slope value is less than 0 for multiple consecutive moments is designated as the second correction interval, and the power data within the second correction interval is corrected to the first collected power value after that interval. Here, "multiple consecutive moments" means two or more moments. For the power data at other moments, the correction value is the collected power data itself.
[0025] Connecting all adjacent corrected power data with a straight line yields the corrected power line of the GPU. By calculating the integral of the corrected power line in each time period, the power consumption data of the GPU in each time period can be obtained. This helps to obtain a more accurate GPU power consumption status, thereby improving the temperature control accuracy of the microcomputer.
[0026] For the CPU, the same power correction method as the GPU is used to obtain the corrected power data of the CPU at each moment, and obtain the corrected power piecewise curve of the CPU.
[0027] The host temperature control module 103 analyzes the correlation between the power consumption of the CPU and GPU and their average temperature in each time window, as well as the trend of the average temperature of the CPU and GPU in the historical time windows of each time window, determines the temperature rise significance coefficient of the CPU and GPU in each time window, and obtains the high temperature significance coefficient of the micro-host in each time window.
[0028] Taking a GPU as an example, considering the impact of GPU power consumption on its temperature change, if the GPU power consumption increases, its corresponding temperature will also increase. The temperature change of the copper solder wire changes rapidly with power and there is a certain delay. Therefore, this embodiment sets each time window, and the length of each time window in this embodiment is 5 seconds. Implementers can set it according to the actual situation, and this embodiment does not impose any restrictions on it. It should be noted that there is no overlap between adjacent time windows, and there are no idle moments between adjacent time windows. For example, the first time window is from the 1st to the 5th time window, and the second time window is from the 6th to the 10th time window. The first time window and the second time window are adjacent time windows.
[0029] Based on the GPU's corrected power data, the GPU's power consumption data within each time window can be obtained, and the average temperature of the GPU at all times within each time window can be calculated. The average temperatures of the i-th time window and its N preceding consecutive time windows are arranged in chronological order to form a temperature sequence, and the power consumption of the i-th time window and its N preceding consecutive time windows are arranged in chronological order to form a power consumption sequence. In this embodiment, N=10, but implementers can set this value according to their actual situation; this embodiment does not impose any restrictions on it.
[0030] The metric distance between the temperature sequence and the power consumption sequence in the i-th time window is calculated and denoted as the first metric distance. Furthermore, the temperature sequence in the i-th time window is analyzed for trend using the Thales-Sen estimation method, and the slope output by the Thales-Sen estimation method is obtained as the trend feature of the temperature sequence in the i-th time window. The Thales-Sen estimation method is a well-known existing technology, and implementers can choose other feasible existing trend analysis algorithms, such as the STL trend decomposition algorithm, etc. This embodiment does not impose any restrictions on this. In this embodiment, the first metric distance is calculated using the SBD (Shape Based Distance) distance; implementers can choose other feasible existing metric distance calculation methods.
[0031] The smaller the first metric distance, the greater the impact of power consumption on GPU temperature within the local area of the i-th time window. Therefore, a significant coefficient of temperature rise due to power consumption influence is calculated for the GPU in the i-th time window. This significant coefficient is positively correlated with the trend feature and negatively correlated with the first metric distance. Specifically, the exponential mapping result of the trend feature is determined, and the first metric distance is forward mapped. The significant coefficient of CPU temperature rise in each time window is the ratio of the exponential mapping result to the forward mapping result. The purpose of the forward mapping is to map the first metric distance to a positive number to avoid the denominator being 0 during the calculation of the significant coefficient of temperature rise, which would prevent calculation.
[0032] In this embodiment, the expression for the significant coefficient of GPU temperature rise in each time window is as follows: In the formula, This represents the significance coefficient of GPU temperature rise in the i-th time window. Let represent the trend characteristics of the temperature sequence in the i-th time window, and exp() be an exponential function with the natural constant as the base. Let the first metric distance be the distance in the i-th time window. To ensure that the value is greater than 0 and to avoid a denominator of 0, this embodiment... The implementer can set it according to the actual situation. In this embodiment... This is the forward mapping result of the first metric distance; the implementer can choose other forward mapping methods at their own discretion. This reflects the significant degree to which GPUs exhibit temperature rise characteristics due to power consumption.
[0033] For the CPU, the same calculation method as that used for the GPU's temperature rise significance coefficient in the i-th time window is adopted to obtain the CPU's temperature rise significance coefficient in the i-th time window.
[0034] The overheating of microcomputers is usually caused by the high power consumption of the CPU and GPU. Furthermore, some tasks require the two to work together, which can easily exacerbate the risk of high temperature in the computer. Therefore, the fusion result of the temperature rise significance coefficient of the CPU and the temperature rise significance coefficient of the GPU in each time window is used as the high temperature significance coefficient of the microcomputer in each time window.
[0035] It should be noted that fusion means combining multiple variables. Specifically, it can be calculated by adding, multiplying, combining addition and multiplication, or taking the average. In this embodiment, the average of the temperature rise significance coefficient of the CPU and the temperature rise significance coefficient of the GPU in each time window is taken as the high temperature significance coefficient of the microcomputer in each time window. The larger the high temperature significance coefficient, the more obvious the high temperature risk characteristics of the microcomputer.
[0036] The fan speed adjustment module 104 obtains the temperature deviation coefficient of the microcomputer in each time window by measuring the temperature difference between different components inside the microcomputer in each time window, and determines the heat dissipation demand coefficient of the microcomputer in each time window by combining the high temperature significance coefficient, so as to control the speed of the cooling fan of the microcomputer.
[0037] Furthermore, most of the heat generated by the microcontroller comes from the high-power operation of the GPU and CPU. In contrast, the heat generated by other components is relatively low. With the help of heat dissipation components and airflow devices, the temperature within the microcontroller is transferred from high to low. Under good heat dissipation conditions, the temperature transfer efficiency from high-power components to low-power components is faster, and the temperature difference between different components is smaller. Therefore, for the i-th time window, the same method used for obtaining the temperature sequence of the GPU in the i-th time window is adopted to obtain the temperature sequence of each component within the microcontroller in the i-th time window. This embodiment calculates the metric distance between any two components of the microcontroller in the i-th time window's temperature sequence, denoted as the second metric distance. Then, the mean of all second metric distances of the microcontroller in the i-th time window is calculated as the temperature deviation coefficient of the microcontroller in the i-th time window, denoted as... The resulting temperature deviation coefficient reflects the temperature deviation characteristics between different components inside the microcontroller. The greater the temperature difference between components, the more timely the microcontroller needs to dissipate heat. Therefore, the heat dissipation demand coefficient of the microcontroller in the i-th time window is calculated based on the internal temperature changes of the microcontroller. Its formula is: In the formula, Let be the high temperature significance coefficient of the microcontroller in the i-th time window. The temperature deviation coefficient of the microcontroller under the i-th time window is obtained. This reflects the differences in temperature distribution among the components within the microcontroller and its significant high-temperature characteristics within this time window.
[0038] Furthermore, the limited space in a microcomputer means that its internal temperature may fluctuate frequently due to load changes and thermal adjustments. Conventional cooling methods typically only adjust fan speeds based on temperature variations to meet cooling needs. However, this approach often struggles to adapt to temperature changes in a timely manner, impacting normal operation. Therefore, this application combines power variation and temperature-heat dissipation efficiency characteristics to control the cooling fan speed, helping to ensure good cooling efficiency while minimizing noise interference from fan speed adjustments.
[0039] Based on the above analysis, this embodiment predicts and analyzes the heat dissipation demand coefficient of the microcomputer in each time window to obtain the heat dissipation trend within a local time period. A first-order exponential smoothing algorithm is used to calculate the predicted value of the heat dissipation demand coefficient for the i-th time window and the preceding N consecutive time windows. The predicted value of the heat dissipation demand coefficient for the i-th time window is denoted as... The result The larger the value, the more timely the microcomputer needs to be cooled down during that time window.
[0040] This embodiment analyzes the deviation between the measured power and actual power state of the GPU and CPU during the operation of the micro-host, corrects the collected power data, and further considers the impact of power consumption on temperature changes and the temperature distribution differences among different components within the micro-host to analyze the host's heat dissipation requirements. Based on this, the fan speed is adjusted for airflow temperature control. Specifically, a larger predicted value of the heat dissipation demand coefficient in the i-th time window indicates a significant temperature anomaly within the micro-host, requiring a corresponding increase in the cooling fan speed; conversely, a smaller predicted value indicates a good temperature condition within the micro-host, allowing for a lower cooling fan speed. This embodiment uses the Z-Score normalization method to normalize the predicted value of the heat dissipation demand coefficient in the i-th time window and utilizes PWM pulse speed control for fan control. The result of the normalization process is used as the duty cycle of the PWM control to adjust the fan speed in the next time window after the i-th time window. To prevent excessively low fan speeds from causing overheating within the mini-PC and affecting its normal operation, a minimum duty cycle of 20% is set to maintain adequate and continuous cooling, helping to compensate for the mini-PC's insufficient temperature control efficiency. PWM pulse speed control is a well-known technology, and its specific process will not be elaborated upon. The block diagram for adjusting the cooling fan speed of the mini-PC is shown below. Figure 2 As shown.
[0041] The dynamic airflow temperature control system for the microcomputer also includes heat dissipation components and airflow guiding devices. The heat dissipation components consist of heat sinks, ducts, and a cooling fan, utilizing the high thermal conductivity of the materials to quickly transfer heat to the air. The cooling fan automatically adjusts its speed according to the temperature, accelerating airflow and carrying away heat. The airflow guiding device is a surround-type convection duct. This duct optimizes airflow within the computer, ensuring that cool air flows smoothly over key heat-generating components, significantly improving heat dissipation efficiency and coverage, and preventing heat buildup.
Claims
1. A temperature-adaptive micro-host dynamic flow-guiding temperature control system, characterized in that, The system includes: The data acquisition module is used to collect the temperature of each component inside the microcomputer at various times, as well as the power of the CPU and GPU inside the microcomputer at various times. The power correction module corrects the power of the CPU and GPU at each time point based on the rate of change of CPU and GPU power at each time point, as well as the power change state of CPU and GPU when executing programs, in order to determine the power consumption of CPU and GPU within each time window. The host temperature control module is used to analyze the correlation between the power consumption of the CPU and GPU and their average temperature in each time window, as well as the trend of the average temperature of the CPU and GPU in the historical time windows of each time window, to determine the temperature rise significance coefficient of the CPU and GPU in each time window, and to obtain the high temperature significance coefficient of the micro-host in each time window. The fan speed adjustment module is used to obtain the temperature deviation coefficient of the microcomputer in each time window by measuring the temperature difference between different components inside the microcomputer in each time window, and to determine the heat dissipation demand coefficient of the microcomputer in each time window by combining the high temperature significance coefficient, so as to control the speed of the cooling fan of the microcomputer.
2. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 1, characterized in that, The correction of CPU and GPU power at each time point includes: For the CPU, the power of its acquisition at all times is nonlinearly fitted to obtain the power fitting curve of the CPU. The slope of the power fitting curve at each time is calculated. If the positive and negative signs of the slopes at multiple consecutive times are consistent, the power of the next time of the time period corresponding to the multiple consecutive times is taken as the power correction value of the multiple consecutive times, and the power correction value of the other times is its acquisition power value. For the GPU, the same power correction method as for the CPU is used to obtain the power correction value of the GPU at each time step.
3. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 1, characterized in that, The determination of the significant coefficients of temperature rise of CPU and GPU in each time window includes: For the CPU, calculate the average temperature and power consumption at all times within each time window, and calculate the metric distance of the average temperature and power consumption of the CPU in each time window and its historical time windows, which is denoted as the first metric distance. Trend analysis is used to extract the trend characteristics of the average CPU temperature in each time window and its historical time windows. The significant coefficient of CPU temperature rise in each time window is positively correlated with the trend characteristics and negatively correlated with the first metric distance. The same calculation method as that used for the CPU's temperature rise significance coefficient was employed to obtain the GPU's temperature rise significance coefficient for each time window.
4. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 3, characterized in that, Further determination of the significance coefficient of CPU temperature rise in each time window includes: The exponential mapping result of the trend feature is determined, and the first metric distance is forward mapped. The significant coefficient of CPU temperature rise in each time window is the ratio of the exponential mapping result to the forward mapping result.
5. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 3, characterized in that, The high temperature significance coefficient of the microcomputer under each time window is the fusion result of the temperature rise significance coefficient of the CPU and the temperature rise significance coefficient of the GPU under each time window.
6. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 1, characterized in that, The process of obtaining the temperature deviation coefficient of the microcontroller under each time window includes: Calculate the average temperature of each component in the microcomputer at all times within each time window, determine the metric distance between any two components of the microcomputer in each time window and its historical time window, and record it as the second metric distance. Merge all the second metric distances of the microcomputer in each time window to obtain the temperature deviation coefficient of the microcomputer in each time window.
7. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 6, characterized in that, The temperature deviation coefficient of the microcomputer in each time window is the average of all the second metric distances of the microcomputer in each time window.
8. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 1, characterized in that, The heat dissipation requirement coefficient of the microcomputer in each time window is the product of the temperature deviation coefficient and the high temperature significance coefficient of the microcomputer in each time window.
9. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 1, characterized in that, The control of the speed of the cooling fan of the microcomputer includes: A smoothing algorithm is used to obtain the smoothed result of the heat dissipation demand coefficient of the microcomputer under each time window. Based on the smoothed result, the speed of the cooling fan of the microcomputer is controlled by PWM pulse speed regulation.
10. The temperature-adaptive micro-host dynamic flow-guiding temperature control system according to claim 9, characterized in that, The smoothing result is used as the duty cycle of PWM control to adjust the speed of the cooling fan in the next time window of each time window.