Self-adaptive temperature control adjusting system

Abstract

The invention discloses a self-adaptive temperature control adjusting system. The invention discloses a thermal limit adaptive performance management method which is suitable for intelligent temperature control of a mobile terminal in a high-load scene. According to the method, a heat generation index (HIndex) and a heat dissipation efficiency index (CIndex) are constructed based on a nonlinear heat generation relation, the temperature trend of equipment is predicted through a double-index model, core frequency and load distribution are dynamically regulated and controlled accordingly, and self-adaptive balance of performance and temperature is achieved. The system adopts a three-stage closed-loop control architecture, comprises a prediction layer, a calibration layer and a regulation and control layer, and adjusts a control strategy in combination with real-time load data, a temperature change rate and environmental conditions. Compared with a traditional passive response temperature control method, the method has higher response sensitivity, environment adaptability and resource scheduling efficiency, and can be widely applied to high-performance application scenes such as games and video rendering.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of mobile terminal temperature control management, and specifically relates to an intelligent temperature control method for realizing adaptive balance of device performance and temperature through heat generation characterization based on a nonlinear relationship, which is suitable for real-time temperature control optimization under high-load scenarios (such as games and video rendering) and solves the core defects of traditional solutions in response delay, environmental adaptability and resource utilization. BACKGROUND

[0002] The existing mobile terminal temperature control technology has the following defects: Passive response mechanism: traditional solutions (such as frequency reduction and core shutdown) only trigger regulation after the temperature exceeds the standard, resulting in sudden performance degradation and lag; Insufficient environmental adaptability: without dynamically sensing changes in heat dissipation conditions (such as environmental temperature and user holding state), fixed thresholds are used, resulting in performance waste when heat dissipation is good and frequent frequency locking when heat dissipation is poor; Coarse-grained regulation: relying on white-listed preset frequencies or load monitoring, it cannot match the heat dissipation capacity in real time, causing performance and temperature imbalance; The present application solves the above problems through dynamic temperature prediction and predictive closed-loop regulation. SUMMARY

[0003] Core innovation 1. Nonlinear temperature abstraction model: Heat generation index (H_Index): dynamically characterizing the real-time heat generation capacity of the device core, calculated based on the nonlinear relationship between load (CPU / GPU occupancy, thread number) and frequency; Heat dissipation efficiency index (C_Index): combining device surface temperature change rate and environmental temperature data to quantify environmental heat dissipation efficiency; 2. Three-level closed-loop control architecture: Prediction layer: predicting future temperature trends based on a double exponential model; Calibration layer: dynamically correcting model parameters according to real-time temperature deviation; Regulation layer: adjusting core frequency and task allocation according to priority to achieve performance and temperature balance. System composition

[0004] The system is composed of the following parts 1. Data acquisition unit: Temperature sensor: collecting temperature of key areas inside the device (such as SoC and battery); Load monitoring module: real-time acquisition of CPU / GPU thread state and power consumption data; 2. Double exponential calculation unit: Heat generation index module: H_Index is calculated based on non-linear relationship; Heat dissipation index module: C_Index is calculated based on temperature rise rate and ambient temperature; ‌3. Dynamic regulation unit: ‌ Thermal balance decision maker: Predict temperature value (T_pred) = Current temperature value (T_current) + (H_Index - C_Index) × Time coefficient (t) by formula; Hierarchical regulator: According to the difference between T_pred and threshold, allocate computing resources according to core load weight, dynamically adjust core frequency (such as preferentially limit low-load cores to ensure high-load core performance); ‌4. Adaptive calibration unit: ‌ Parameter feedback loop: Through sliding window statistics of prediction error, dynamically adjust heat generation and heat dissipation weight; Fault-tolerant controller: When temperature fluctuation is out of limit, trigger emergency frequency reduction and reset model. DETAILED DESCRIPTION

[0005] Some embodiments are listed below: ‌Embodiment 1: Game scene temperature control optimization‌: 1‌. Prediction stage: ‌ Monitor GPU rendering load surge, calculate H_Index non-linear rise; Detect device surface temperature rise rate, reduce C_Index weight; ‌2. Regulation stage: ‌ Because the predicted temperature value (T_pred) is close to the threshold value, migrate background tasks to energy-efficient cores (E-cores), and limit the peak frequency of GPU large cores (P-cores) to a safe range; ‌Embodiment 2: Environmental mutation response‌: 1‌. Heat dissipation deterioration scenario (such as user holding leading to local temperature rise): Update C_Index: Surface temperature rise rate increases → Heat dissipation index decreases; Recalculate the heat balance equation, dynamically adjust the regulation strategy according to the change of heat dissipation capacity, including reducing frequency or adjusting load distribution; 2. Heat dissipation enhancement scenario (such as connecting a heat sink): Improve C_Index value, allow the device to run at a higher frequency to release performance potential.

Claims

1. A device temperature control method based on adaptive parameter calibration, characterized in that, Includes the following steps: Dynamic temperature prediction: Based on multi-source equipment operation data, a temperature prediction model is used to calculate the future temperature trend of the equipment in real time; Employing multiple temperature prediction schemes to adapt to different equipment and environments: Option 1: Temperature prediction based on advance testing. By testing the temperature rise index of different cores at different frequencies in advance and combining it with the current temperature data of the equipment, the heat dissipation index is calculated, thereby deriving the heat dissipation capacity of the equipment in the current environment. Combined with the temperature threshold, the maximum temperature rise index that the equipment can withstand is calculated, and then the equipment frequency is dynamically adjusted to ensure that the equipment operates stably within the temperature threshold range. Option 2: Temperature prediction based on exponential fitting. During equipment operation, temperature data is continuously collected, and the temperature trend is predicted using a nonlinear exponential fitting algorithm. If the predicted temperature exceeds the temperature control threshold, the equipment operating frequency is dynamically adjusted until the temperature converges to the steady-state range. Adaptive calibration of heating rate: Based on the deviation between the real-time temperature measurement and the predicted temperature, the heat accumulation parameter and heat dissipation parameter in the temperature prediction model are dynamically adjusted. Dynamic performance optimization and control: When the heating rate approaches the calculated heat dissipation capacity threshold, a frequency adjustment command is triggered to dynamically adjust the device's computing core frequency, thread scheduling status, and load distribution to optimize device performance. Real-time closed-loop correction: The heating rate model is continuously updated based on changes in ambient temperature and equipment operating status to ensure a balance between equipment temperature and performance; Dynamic tolerance range determination and fault tolerance control: When the equipment temperature fluctuation exceeds the preset dynamic tolerance range, the closed-loop adjustment mechanism is triggered to adjust the equipment operating frequency to the temperature steady-state range.

2. The equipment temperature control method according to claim 1, characterized in that, The temperature prediction model calculates future temperature trends in real time based on multi-source operating data of the equipment, and dynamically selects the appropriate prediction method through various temperature prediction schemes (including temperature prediction based on advance testing and exponential fitting prediction).

3. The equipment temperature control method according to claim 1, characterized in that, The adaptive calibration of the heating rate dynamically optimizes the temperature deviation between real-time data and predicted temperature, calibrating heat accumulation and heat dissipation parameters to ensure that the equipment temperature control is consistent with load changes.

4. The equipment temperature control method according to claim 1, characterized in that, The dynamic performance optimization and control optimizes device performance by adjusting the computing core frequency, thread scheduling, and load distribution, ensuring maximum performance output within the target temperature control range.

5. The equipment temperature control method according to claim 1, characterized in that, The real-time closed-loop correction continuously updates the heating rate model to adapt to environmental changes and equipment operating status, ensuring a long-term balance between equipment temperature and performance.

6. The equipment temperature control method according to claim 1, characterized in that, The dynamic tolerance range determination conditions include a temperature change rate exceeding a threshold or an abnormal statistical distribution of residuals within multiple sampling periods, and a fault-tolerant mechanism that automatically triggers frequency adjustment commands.