Heat dissipation control method and device, storage medium and electronic equipment

Abstract

The embodiment of the invention provides a heat dissipation control method and device, a storage medium and electronic equipment, and relates to the technical field of computers, the method is applied to the electronic equipment, the electronic equipment comprises a heat dissipation device, and the method comprises the steps that a physical parameter set of the electronic equipment and noise generated by the heat dissipation device are obtained; determining control parameters for heat dissipation according to the physical parameter set and the noise; and controlling a heat dissipation device of the electronic equipment according to the control parameters. Therefore, the control parameter of the heat dissipation device is dynamically adjusted according to the temperature of each key component of the electronic equipment and the noise generated by the heat dissipation device, the overall heat dissipation requirement of the electronic equipment is considered, the power consumption of the electronic equipment can also be considered, and the resource utilization efficiency is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technologies, and in particular, to a heat dissipation control method, apparatus, storage medium, and electronic device. Background Art

[0002] A Baseboard Management Controller (BMC) is a controller used to monitor and manage a server. Its main functions include: 1) Device information management: recording detailed information of the server, such as model, manufacturer, production date, as well as production and technical information of each component, chassis and motherboard information, etc. At the same time, it also records information about the BMC itself, such as server hostname, IP address, and BMC firmware version, etc. 2) Server status monitoring and management: real-time detecting the health status of temperature and voltage of each key component of the server (such as CPU, memory, hard disk, fan, and chassis, etc.). According to the data of temperature acquisition points, the BMC adjusts the fan speed to ensure that the server does not overheat and controls the overall power consumption within a reasonable range. If any component of the server appears abnormal, the BMC will report relevant information to the upper-layer network administrator in a timely manner through various industry-standard specifications (such as SNMP protocol, SMTP protocol, and Redfish protocol). 3) Remote control and management of the server: allowing the administrator to remotely execute operations such as power on / off, restart, maintenance, firmware update, and system installation of the server. 4) Maintenance management: including log management, user management, BIOS management, and alarm management, etc., to ensure the stable operation of the server. Generally speaking, the BMC plays a key role in the operation and management of the server. The BMC is mainly responsible for monitoring, managing, and remotely controlling the server to ensure its stable operation and efficient management.

[0003] Server heat dissipation is crucial for maintaining the stable operation of hardware. Especially in high-load situations, overheating problems may occur. Heat dissipation includes passive heat dissipation and active heat dissipation. Passive heat dissipation may include heat sinks and heat pipe technologies, while active heat dissipation mainly relies on fans. The fan is controlled by the BMC. The key is how to optimize the fan control strategy to balance heat dissipation effect, noise, and energy consumption. Common techniques may include Pulse Width Modulation (PWM) to control the fan speed, or using temperature sensor feedback to achieve dynamic adjustment.

[0004] In the related art, the server's active cooling mainly relies on the BMC to read the temperature data through the temperature sensors of key components such as the CPU, memory, and network card, and calculates the maximum PWM according to the preset parameters (such as fixed thresholds) through the PID (Proportional Integral Derivative) algorithm, and uses the PWM signal to control the fan rotation speed. This solution depends on component temperature reading, has a single control source, and there is a risk of the fan running at full speed for a long time. Specifically, 1) The temperature sampling dimension is single, and a multi-sensor data fusion mechanism is not established, which easily leads to frequent full-speed operation of the fan due to misjudgment of local hot spots; 2) The static configuration of PID parameters lacks dynamic adaptability, and when facing sudden cooling pressure, it can only passively trigger the limit speed, exacerbating mechanical wear; 3) The control strategy is completely coupled with the hardware temperature, lacking environmental perception ability, and unable to achieve acoustic optimization while ensuring the cooling efficiency.

[0005] This solution will directly cause the fan to be in a high-load working condition for a long time, shortening the bearing life to 60%-70% of the nominal value, and the noise during full-speed operation generally exceeds the 55dB limit specified by the TIA-942 standard. Moreover, it cannot meet the customer's demand for differential noise reduction: it is necessary to balance cooling and silence during the day, and it is necessary to deeply reduce the speed to below 48dB at night.

[0006] In addition, the original solution lacks the design of redundant control channels, and the failure of a single-point sensor will cause system misjudgment, further increasing the risk of abnormal fan loss. This crude temperature control mechanism can no longer meet the composite requirements of modern IDC computer rooms for acoustic comfort, equipment reliability, and energy efficiency ratio. Summary of the Invention

[0007] The present disclosure provides a heat dissipation control method, device, storage medium, and electronic device to at least solve the above technical problems existing in the prior art.

[0008] The technical solution of the embodiment of the present disclosure is realized as follows: In a first aspect, the embodiment of the present disclosure provides a heat dissipation control method, which is applied to an electronic device, and the electronic device includes a heat dissipation device. The method includes: Obtain the physical parameter set of the electronic device and the noise generated by the heat dissipation device; Determine the control parameters for heat dissipation according to the physical parameter set and the noise; Control the heat dissipation device of the electronic device according to the control parameters.

[0009] In the above solution, the obtaining of the physical parameter set of the electronic device includes: Collect at least one physical parameter of the electronic device, and determine the physical parameter set according to the collected at least one physical parameter; Among them, the at least one physical parameter includes at least one of the following: component temperature, rotation speed of the fan; the components include: CPU, memory, and PCIE device.

[0010] In the above solution, determining the control parameter for heat dissipation according to the set of physical parameters and the noise includes: Based on the first rule, determining the evaluation coefficient of at least one component according to the temperature parameter of at least one component and the noise; Based on the second rule, determining the control parameter according to the evaluation coefficient of the at least one component.

[0011] In the above solution, based on the first rule, determining at least one control index according to the temperature parameter of at least one component and the noise includes: Determining the temperature weight and temperature threshold corresponding to each component according to each component; Determining the evaluation coefficient of each component according to the temperature parameter, the temperature threshold, the temperature weight, the noise, and the noise weight of each component.

[0012] In the above solution, determining the evaluation coefficient of each component according to the temperature parameter, the noise, the temperature weight, and the noise weight of each component includes: Determining the temperature control index according to the temperature parameter and the temperature threshold of each component; Determining the noise control index according to the noise and the reference noise; Determining the evaluation coefficient of each component according to the temperature control index, the noise control index, the temperature weight, and the noise weight.

[0013] In the above solution, the determining the control parameter according to the evaluation coefficient of the at least one component based on the second rule includes: Determining an evaluation coefficient vector according to the evaluation coefficient of the at least one component; Inputting the evaluation coefficient vector into the first model to obtain the output result of the first model, and the output result is used to indicate the adjustment range of the PID parameter.

[0014] In the above solution, before determining the control parameter for heat dissipation according to the set of physical parameters and the noise, the method further includes: Determining whether there is a heat dissipation failure according to the set of physical parameters and the noise; If it is determined that there is a heat dissipation failure, generating a failure alarm, and the failure alarm is used to indicate the type of failure that occurs; If it is determined that there is no heat dissipation failure, determining the control parameter for heat dissipation according to the set of physical parameters and the noise.

[0015] In the above solution, determining whether there is a heat dissipation failure according to the physical parameter set and the noise includes: Generating time series data of rotational speed and / or temperature according to the physical parameter set; Extracting a first feature set from the noise, the first feature set including: BPF, HAR; Inputting the first feature set and the time series data of rotational speed and / or temperature into a second model to obtain an output result of the second model, the output result being used to indicate the fault type and the fault anomaly level.

[0016] In the above solution, generating a fault warning includes: If the fault type belongs to the target fault type and the fault anomaly level exceeds the level threshold, generating a fault warning according to the fault type and the fault anomaly level.

[0017] In the above solution, the method further includes: generating the second model; generating the second model includes: Obtaining a training data set; the training data set includes: at least one training data and a label for each training data, each training data including: sample noise, time series data of sample rotational speed and / or temperature; the label being the fault type and the fault anomaly level; Training a hybrid neural network using the training data set to obtain the trained hybrid neural network as the second model; Wherein, the hybrid neural network includes: a front-end network, a feature fusion layer, and an output layer; The front-end network adopts a first branch network and a second branch network, the first branch network being used to extract the features of the sample noise, and the second branch network being used to extract the features of the time series data of the sample rotational speed and / or temperature; The feature fusion layer is used to perform data processing on the features extracted by the front-end network by using an attention mechanism; The output layer is used to obtain an output result according to the data processing result of the feature fusion layer.

[0018] In the above solution, the method further includes: Obtaining a fault database of at least one fan, and fine-tuning the second model according to the fault database to obtain a second model corresponding to the electronic device adopting the fan.

[0019] In the above solution, obtaining the noise generated by the heat dissipation device includes: Respectively obtaining a first noise generated by a fan and / or a second noise of an air duct; The first noise and the second noise are combined to obtain the noise generated by the heat dissipation device.

[0020] In a second aspect, an embodiment of the present disclosure provides a heat dissipation control device, which is applied to an electronic device, and the electronic device includes a heat dissipation device. The device includes: An acquisition module, configured to acquire a set of physical parameters of the electronic device and the noise generated by the heat dissipation device; A first processing module, configured to determine a control parameter for heat dissipation according to the set of physical parameters and the noise; A second processing module, configured to control the heat dissipation device of the electronic device according to the control parameter.

[0021] In a third aspect, an embodiment of the present disclosure provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute any one of the heat dissipation control methods.

[0022] In a fourth aspect, an embodiment of the present disclosure provides a non-transitory computer-readable storage medium storing computer instructions, and the computer instructions are used to cause a computer to execute any one of the heat dissipation control methods.

[0023] The embodiments of the present disclosure have the following beneficial effects: Applying the heat dissipation control method, device, storage medium and electronic device provided by the embodiments of the present disclosure, the method is applied to an electronic device, and the electronic device includes a heat dissipation device. The method includes: acquiring a set of physical parameters of the electronic device and the noise generated by the heat dissipation device; determining a control parameter for heat dissipation according to the set of physical parameters and the noise; controlling the heat dissipation device of the electronic device according to the control parameter. In this way, the control parameters of the heat dissipation device are dynamically adjusted according to the temperatures of the key components of the electronic device and the noise generated by the heat dissipation device, which not only takes into account the overall heat dissipation requirements of the electronic device, but also takes into account the power consumption of the electronic device, and improves the resource utilization efficiency.

[0024] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it used to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a schematic flowchart of a heat dissipation control method provided by an embodiment of the present disclosure; Figure 2 is a schematic diagram of a hybrid neural network architecture provided by an embodiment of the present disclosure; Figure 3 A schematic flowchart of a training method for a second model provided by an embodiment of the present disclosure; Figure 4 A schematic structural diagram of a heat dissipation control device provided by an embodiment of the present disclosure; Figure 5 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure. Detailed implementation manners

[0026] To make the objectives, features, and advantages of the present disclosure more obvious and understandable, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present disclosure.

[0027] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0028] If similar descriptions such as "first / second" appear in the application documents, the following description is added. In the following description, the terms "first\second\third" only distinguish similar objects and do not represent a specific order for the objects. It can be understood that "first\second\third" can be interchanged with a specific order or sequence when permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0030] Figure 1 A schematic flowchart of a heat dissipation control method provided by an embodiment of the present disclosure, as Figure 1 shown. The method is applied to an electronic device, and the electronic device includes a heat dissipation device. The method includes: Step 101, obtaining a set of physical parameters of the electronic device and the noise generated by the heat dissipation device; Step 102, determining a control parameter for heat dissipation according to the set of physical parameters and the noise; Step 103, controlling the heat dissipation device of the electronic device according to the control parameter.

[0031] Among them, the electronic device can be a device with a heat dissipation device such as a server, a computer, a projector, etc.

[0032] In some embodiments, obtaining the set of physical parameters of the electronic device includes: Collecting at least one physical parameter of the electronic device, and determining the set of physical parameters according to the at least one collected physical parameter; Among them, the at least one physical parameter includes at least one of the following: component temperature, rotation speed of the fan (such as the PWM value of the fan); The component can be a component in the electronic device that affects and / or is affected by temperature. For example, the component can include at least one of the following: CPU (Central Processing Unit), memory RAM (Random Access Memory), PCIE (Peripheral Component Interconnect Express) device, etc. Correspondingly, the component temperature includes: CPU temperature, memory temperature, PCIE device temperature.

[0033] Here, PCIe is a high-speed data transmission interface used to connect the computer motherboard to various external hardware devices. PCIE devices can include: graphics cards, network cards, storage devices, etc.

[0034] Based on the designs of different electronic devices, the processing device can also include: GPU (Graphics Processing Unit), TPU (Tensor Processing Unit), FPGA (Field-Programmable Gate Array), ASIC (Application-Specific Integrated Circuit), DSP (Digital Signal Processor), NPU (Neural Processing Unit), VPU (Vision Processing Unit), etc. Correspondingly, the component temperature can also include the temperatures of the above devices.

[0035] The physical parameters can also include: ambient temperature, ambient humidity, air pressure inside the chassis, etc.

[0036] Here, the electronic device has a BMC, and the BMC is designed with a data acquisition function to collect the above physical parameters and record the corresponding time. For example, the BMC monitors the temperature sensors, humidity sensors, etc. corresponding to the above components to determine the above various physical parameters.

[0037] In some embodiments, obtaining the noise generated by the heat dissipation device includes: Obtaining the first noise generated by the fan and / or the second noise of the air duct respectively; Combining the first noise and the second noise to obtain the noise generated by the heat dissipation device.

[0038] Here, the heat dissipation device may include: a fan, an air duct; the first noise generated by the fan is: the noise generated when the fan runs. The second noise of the air duct is: the noise generated when the air flow flows in the air duct.

[0039] For combining the first noise and the second noise to obtain the noise generated by the heat dissipation device, a mathematical model (such as logarithmic weighting, spectrum analysis, etc.) can be used to integrate the contributions of different noise sources together to obtain a result reflecting the comprehensive noise level. The specific combination method should be selected according to the actual situation (such as the frequency characteristics, action intensity, etc. of the noise source).

[0040] In one example, weighted combination is adopted. Here, considering that the noises generated by the fan and the air duct may have different effects on the total noise, they can be combined through a weighting coefficient. For example, after verification, the noise of the fan dominates, so a higher weight can be given to the fan noise and a lower weight to the air duct noise.

[0041] In another example, spectrum analysis is adopted. Here, considering that the frequency characteristics of the noise sources may be different, the spectrum analysis method can be used to superimpose the spectrograms of each noise source to more accurately reflect the contributions of each noise source to different frequency bands. For the noise intensity of each frequency band, the noise level can be graded and different weights can be set, and the spectrum of the entire noise signal can be obtained by combining the weighted combination method.

[0042] Of course, other methods can also be used for combination, which are not limited here.

[0043] Here, the electronic device has a BMC, and the BMC is designed with a data acquisition function to collect the above noises and record the corresponding acquisition time. For example, the first noise generated by the fan and / or the second noise of the air duct are collected in real time through a high-precision microphone array (for example, the signal-to-noise ratio ≥ 60 dB) (for example, the designed sampling rate ≥ 48 kHz and the resolution is 24 bit).

[0044] After obtaining the above physical parameters and noise, the timestamps of each data can be aligned based on the BMC clock to obtain a set of data in the following form: time-noise-physical parameter set, so as to clearly and accurately record the data set.

[0045] In some embodiments, determining the control parameter for heat dissipation according to the physical parameter set and the noise includes: Based on the first rule, determine the evaluation coefficient of at least one component according to the temperature parameter of at least one component and the noise; Based on the second rule, determine the control parameter according to the evaluation coefficient of the at least one component.

[0046] Here, the first rule is used to analyze the temperature parameter and noise of at least one component to determine or reflect the heat dissipation situation. Specifically, it indicates an analysis method. By performing corresponding processing or calculation on the temperature parameter of at least one component and the noise, the evaluation coefficient of each component is obtained. This evaluation coefficient is used to measure the working state of the heat dissipation device (such as a fan), or rather, to reflect the control index for strengthening or weakening the heat dissipation currently required.

[0047] The second rule is used to indicate how to analyze and determine the control parameter according to the evaluation coefficient after obtaining the evaluation coefficient.

[0048] For example, in a heat dissipation system, there is a PID controller, whose function is to maintain the electronic device within the ideal temperature range by adjusting control variables such as the fan speed in real time. The control parameter can be the PID parameter adjustment range, that is, proportional (P), integral (I), and derivative (D). Among them, the proportional (P) parameter affects the response speed of the system. A suitable P parameter can make the temperature approach the target value in a shorter time. The integral (I) parameter can eliminate the steady-state error of the system. An appropriate I parameter helps to eliminate the temperature drift of the system, but too high an integral parameter may cause the system to over-regulate and oscillate. The derivative (D) parameter can reduce the oscillation and overshoot of the system. The role of the D parameter is to help the system react quickly and reduce the excessive temperature fluctuation. These three parameters determine how to adjust the output of the fan to achieve more precise heat dissipation control.

[0049] For the PID parameters, their adjustment ranges can be discretized into multiple actions. For example, for the P parameter, its adjustment range is discretized into 5 actions: -10%, -5%, 0%, +5%, +10%, indicating the adjustment amplitude of the P parameter relative to the current value. Similarly, the I parameter and D parameter are also discretized. In this way, the control parameter is designed as a discrete action space, and each action corresponds to a set of PID parameter adjustment ranges.

[0050] In some embodiments, based on the first rule, at least one control index is determined according to the temperature parameter of at least one component and the noise, including: Determine the temperature weight and temperature threshold corresponding to each component according to each component; Determine the evaluation coefficient of each component according to the temperature parameter, the temperature threshold, the temperature weight, the noise and the noise weight of each component.

[0051] Here, the first rule may include the temperature weight and temperature threshold corresponding to each component.

[0052] The temperature weight represents the degree of influence of the temperature of this component on the overall system control. The temperature of different components may have different effects on the system. Therefore, different weights can be assigned to different components.

[0053] The temperature threshold is a boundary value of the temperature of this component. When the temperature reaches or exceeds this value, certain control measures or alarms can be triggered. For example, the temperature threshold for the CPU can be designed as 85 °C.

[0054] In some embodiments, determining the evaluation coefficient of each component according to the temperature parameter, the noise, the temperature weight and the noise weight of each component includes: Determine the temperature control index according to the temperature parameter and the temperature threshold of each component; Determine the noise control index according to the noise and the reference noise; Determine the evaluation coefficient of each component according to the temperature control index, the noise control index, the temperature weight and the noise weight.

[0055] Here, an evaluation function is provided for calculating the evaluation coefficient, as shown in the following formula:

[0056] Where Tcr is the temperature threshold (as in the above example, assuming the evaluation coefficient of the CPU is calculated, it is 85 °C), Tcu is the current temperature; Nba is the reference noise (such as 45 dB), Ncu is the current noise, a is the temperature weight, b is the noise weight, and a + b = 1.

[0057] Through the calculated evaluation coefficient, the control effect of the fan on the temperature of this component can be measured. Among them, represents the temperature control index, represents the noise control index.

[0058] The larger the value of the temperature control index, the closer the current temperature Tcu is to the temperature threshold Tcr (that is, the higher the temperature). In the evaluation function, the larger the value of this part, the worse the temperature control effect.

[0059] The larger the value of the noise control index, the closer the current noise Ncu is to the reference noise Nba (i.e., the greater the noise). In the evaluation function, the larger the value of this part, the worse the noise control effect.

[0060] The weight coefficients a and b are used to balance the importance of temperature control and noise control. By adjusting the values of a and b, temperature control or noise control can be prioritized according to actual needs. For example, if more attention is paid to temperature control, a can be set greater than b, such as a = 0.7, b = 0.3; if more attention is paid to noise control, a can be set less than b, such as a = 0.3, b = 0.7.

[0061] Considering the importance of different components in the system and their different sensitivities to temperature and noise, there can be differences in the a and b values adopted for each component.

[0062] For example, taking the CPU, memory, and PCIE devices as examples, since the CPU is the core component of the computer and high temperature may cause performance degradation or even damage, the CPU pays more attention to temperature control. In the evaluation function of the CPU, a can be set relatively large, such as a = 0.8, b = 0.2. While the memory may have relatively higher requirements for noise, so in the evaluation function of the memory, b can be set relatively large, such as a = 0.4, b = 0.6. For PCIE devices, the degree of attention to temperature and noise may be the same, so in the evaluation function of PCIE devices, a and b are set the same, that is, a = 0.5, b = 0.5.

[0063] In some embodiments, based on the second rule, determining the control parameter according to the evaluation coefficient of the at least one component includes: Determining an evaluation coefficient vector according to the evaluation coefficient of the at least one component; Inputting the evaluation coefficient vector into a first model to obtain an output result of the first model, where the output result is used to indicate the control parameter.

[0064] Here, the second rule may include a first model for obtaining the control parameter according to the evaluation coefficient vector.

[0065] In an example, the first model adopts the DDQN model. DDQN (Double Deep Q - Network) is an algorithm in deep reinforcement learning for solving decision - making problems in a discrete action space. It approximates the Q - function through a neural network to guide the agent (here refers to the fan speed control system) to select the optimal action (here refers to the control parameter, that is, the PID parameter adjustment range).

[0066] In actual application, the evaluation coefficient of each component is used as the input of the DDQN model. The DDQN model processes according to the input and outputs control parameters, that is, the adjustment range of PID parameters. Suppose the components include: CPU, memory, and PCIE device. The evaluation coefficients corresponding to the three components of CPU, memory, and PCIE device are obtained, and the three evaluation coefficients are combined into an evaluation coefficient vector as the input of the DDQN model.

[0067] Correspondingly, the method further includes: training to obtain a DDQN model as the first model. A training method is provided, which specifically includes the following steps: 1. Initialize the network: Initialize a target network and an evaluation network. The structures of these two networks are the same, but the parameters are different. The evaluation network is used to calculate the Q value in real time (determine the behavior of the agent), and the target network is used to calculate the target Q value to help stabilize the training.

[0068] 2. Environment interaction: At each time step, the agent selects an action (here is the adjustment range of PID parameters) according to the current state (here is the evaluation coefficient of each component). Apply this action to the environment (for example, adjust the PID parameters of the fan), and then the environment will feedback the new state and the corresponding reward.

[0069] 3. Calculate the reward: The reward function can be designed according to the needs of the system. If the temperature decreases and the noise does not increase significantly after the fan speed is adjusted, a positive reward can be given; on the contrary, if the temperature increases or the noise is too high, a negative reward can be given.

[0070] 4. Store experience: Store the current state, action, reward, and next state in the experience replay buffer.

[0071] 5. Sampling training: Randomly sample a batch of samples from the experience replay buffer, use the evaluation network to calculate the current Q value, and use the target network to calculate the target Q value. Update the parameters of the evaluation network by minimizing the error between the current Q value and the target Q value.

[0072] 6. Regularly update the target network: Every certain number of time steps, copy the parameters of the evaluation network to the target network to stabilize the training process.

[0073] Through the above steps, the DDQN model can learn the optimal PID parameters according to the environmental feedback, so as to achieve optimal control.

[0074] Thus, for the trained DDQN model, the action space is the PID parameter adjustment range (e.g., P ± 20%, I ± 15%, D ± 10%); the state space includes the noise feature vector, temperature gradient, and rotational speed deviation, and the reward function focuses on temperature rise suppression (reward +5 for each 1°C drop) and noise penalty (penalty -3 for each 1 dB exceedance). The optimal PID parameters are output using this DDQN model.

[0075] For the method provided in the embodiments of the present disclosure, considering that during the actual operation of the fan, its operating conditions will constantly change. For example, factors such as temperature and load size will affect the optimal rotational speed of the fan. To achieve dynamic adjustment of the fan rotational speed, based on the real-time collected physical data (such as vibration (reflected by noise), temperature, rotational speed, etc.), the optimal PID parameters are calculated in real time using the DDQN model, and the PID parameters are dynamically adjusted, thereby precisely controlling the fan rotational speed and improving the heat dissipation effect.

[0076] In some embodiments, before determining the control parameters for heat dissipation according to the set of physical parameters and the noise, the method further includes: Determining whether there is a heat dissipation failure according to the set of physical parameters and the noise; If it is determined that there is a heat dissipation failure, generating a failure alarm, and the failure alarm is used to indicate the type of failure that occurs; If it is determined that there is no heat dissipation failure, determining the control parameters for heat dissipation according to the set of physical parameters and the noise.

[0077] Here, a method for determining whether there is a heat dissipation failure according to physical parameters and noise is provided. If it is determined that there is a heat dissipation failure, a failure alarm signal is sent to inform the user of the failure and indicate the type of failure, avoiding device operation damage caused by temperature. If there is no failure, the control parameters for heat dissipation are continued to be determined, that is, the control parameters for heat dissipation are adjusted according to the physical parameters and noise data (such as adjusting the fan rotational speed, etc.) to ensure good heat dissipation effect.

[0078] In some embodiments, the generating of the failure alarm includes: If the type of failure belongs to the target type of failure and the failure abnormality level exceeds the level threshold, generating a failure alarm according to the type of failure and the failure abnormality level.

[0079] Here, the type of failure and the failure abnormality level can be pre-divided to generate a failure alarm by judging the type of failure and its severity.

[0080] In an example, the type of failure may include at least one of the following: Impeller imbalance: The impeller vibrates or is damaged due to imbalance during operation; Bearing wear: Bearing wear can cause the machine to run smoothly or be damaged; Foreign object interference: Foreign objects inside the machine affect normal operation.

[0081] For each type of fault, the fault abnormality level is specifically divided to reflect its severity, usually represented by a level from 0 to 5. Level 0 indicates no fault, and level 5 indicates the most serious fault.

[0082] For example, if the detected fault type conforms to the predefined target type of the system (such as impeller imbalance, bearing wear, etc.), then continue to determine whether the fault abnormality level exceeds the level threshold. If the fault abnormality level exceeds a certain set threshold (such as the level threshold is 3 or 4), then a fault alarm is generated. The generated fault alarm will include the fault type and the fault abnormality level, which are used as alarm information to help the operation and maintenance personnel or the system make corresponding processing.

[0083] It should be noted that the above is only an example of the division of a fault type and a fault abnormality level. In actual applications, for different electronic devices and their designs, the fault types and fault abnormality levels can adopt other designs. The above example does not limit it.

[0084] In some embodiments, determining whether there is a heat dissipation fault according to the physical parameter set and the noise includes: Generating time-series data of rotational speed and / or temperature according to the physical parameter set; Extracting a first feature set from the noise, where the first feature set includes at least one of the following: blade passing frequency (BPF), harmonic amplitude ratio (HAR); Inputting the first feature set, the time-series data of the rotational speed and / or temperature into a second model to obtain the output result of the second model, where the output result is used to indicate the fault type and the fault abnormality level.

[0085] Here, the time-series data of the rotational speed and / or temperature is the time-series data including the rotational speed and each temperature established according to the acquisition time; each temperature includes: the temperature of each component, the ambient temperature, etc.; if parameters such as humidity and air pressure are collected, it can also include: ambient humidity, air pressure inside the chassis.

[0086] Specifically, BPF refers to the frequency at which each blade passes a certain fixed point when the blades are rotating. It is generally used for rotating equipment such as the fan blower. Since the blades interact with the surrounding air or other objects during rotation, periodic noise or vibration will be generated, and this frequency is usually proportional to the rotational speed of the blades. BPF can be used to analyze the working state of the fan. If abnormal BPF noise is detected, it may indicate that the blades are unbalanced, worn, or damaged. Here, considering the possible noise frequencies generated when the blades are rotating, the BPF is extracted to analyze the state of the blades (such as problems like abnormal rotational speed, missing blades, or damage). For example, if the BPF suddenly changes, it may mean that the rotational speed of the blades is abnormal or the number of blades has decreased.

[0087] Specifically, HAR is an index that describes the amplitude ratio of each frequency component in the noise signal. The noise signal usually consists of multiple frequency components, including the fundamental frequency (original frequency) and its harmonic components. HAR evaluates the characteristics of the noise by calculating the amplitude ratio of each harmonic in the signal (such as the second harmonic, third harmonic, etc.) relative to the fundamental frequency. The change of HAR can also reflect the performance state of the device. For example, as the usage time of the device increases, due to reasons such as wear and aging, HAR may change. By monitoring the change trend of HAR, the performance degradation of the device can be evaluated, and maintenance and overhaul can be arranged in advance. For example, if the amplitude of the harmonic of a certain frequency increases abnormally, it may indicate that a certain component is faulty. Therefore, by extracting HAR, it is analyzed whether there are abnormal vibrations or other fault signs.

[0088] In actual application, the extracted noise feature set (BPF and / or HAR) and the time-series data of other physical parameters (such as rotational speed, temperature, etc.) are input into the second model together. Among them, the time-series data characterizes that the physical parameters can change over time, that is, it can reflect the changes during operation; the second model generates an output result according to the input data.

[0089] The following provides an application example. BMC has a data preprocessing function. After obtaining the noise generated by the heat dissipation device, an adaptive filter (such as the LMS algorithm) is used to eliminate the environmental background noise (such as the sound of the computer room air conditioner) to obtain the noise signal to be analyzed, and then the noise signal is subjected to short-time Fourier transform (STFT) to generate a time-frequency spectrum diagram (for example, the time window length is 20 ms and the overlap rate is 50%); then the characteristic BPF and / or HAR are extracted using a specific physical model. Through BPF and / or HAR, BMC can help detect the change trend of the server fan and issue an alarm.

[0090] In some embodiments, the method further includes: generating the second model; the generating of the second model includes: Obtain a training data set; the training data set includes: at least one training data and the label of each training data, and each training data includes: time series data of sample noise, sample rotational speed, and / or temperature; the label is the fault type and the fault abnormality level; Use the training data set to train a hybrid neural network to obtain the trained hybrid neural network as the second model; Among them, the hybrid neural network includes: a front-end network, a feature fusion layer, and an output layer; The front-end network adopts a first branch network and a second branch network. The first branch network is used to extract the features of the noise, and the second branch network is used to extract the features of the time series data of the rotational speed and / or temperature; The feature fusion layer is used to perform data processing on the features extracted by the front-end network by using an attention mechanism; The output layer is used to obtain an output result according to the data processing result of the feature fusion layer.

[0091] Specifically, the second model can adopt a hybrid neural network architecture, such as Figure 2 As shown, a schematic diagram of a hybrid neural network architecture is provided. The hybrid neural network includes: 1) A front-end network, which can adopt a multi-branch CNN (such as a parallel first branch network (1D-CNN) and a second branch network (2D-CNN)); 2) A feature fusion layer, which dynamically weights multi-modal features (referring to the time series data of noise, rotational speed, and / or temperature) through an attention mechanism. Here, the feature fusion layer calculates the similarity between each modal feature and the query vector in the attention mechanism according to the features of the time series data of noise, rotational speed, and / or temperature to obtain the attention weights, and weights and sums each modal feature with the weights to fuse the multi-modal features, highlighting important information and assisting in the identification of abnormal vibrations.

[0092] 3) An output layer, which outputs the fault type and the fault abnormality level.

[0093] Specifically, when obtaining the training data set, abnormal conditions can be simulated by applying random frequency domain perturbations (±10% harmonic offset) to the normal noise. The specific steps for obtaining the training data set may include: 1) Data preprocessing, including: collecting normal noise data from a device that is actually running and in a normal state through BMC. These data are a series of sampling points in the time domain, and their frequency spectrum distribution can be obtained through Fourier transform in the frequency domain; analyzing the frequency spectrum of the normal noise data to find the main harmonic frequency components. For example, the time domain signal is converted into a frequency domain signal through fast Fourier transform (FFT), and then the peak frequencies in the frequency spectrum are identified. These peak frequencies are the main harmonic frequencies.

[0094] 2) Apply random frequency-domain perturbations, including: For each major harmonic frequency, generate a random offset within the range of ±10%. For example, if a certain harmonic frequency is f0, the range of the random offset Δf is -0.1f0 ≤ Δf ≤ 0.1f0. A random number generation function can be used to generate this offset. Add the generated random offset to the original harmonic frequency to obtain the adjusted harmonic frequency f′ = f0 + Δf. Reconstruct the spectrum according to the adjusted harmonic frequency. The perturbation can be simulated by increasing or decreasing the corresponding amplitude at the adjusted harmonic frequency. For example, a Gaussian function or other appropriate function can be used to generate the amplitude at the new harmonic frequency position while keeping the amplitudes of other frequency components unchanged or making appropriate adjustments.

[0095] 3) Frequency-domain to time-domain conversion, including: Convert the reconstructed spectrum back to a time-domain signal through the inverse Fourier transform (IFFT) to obtain the vibration signal data simulating anomalies.

[0096] Specifically, during the training process of the second model, when the model confidence is lower than the threshold, trigger the robotic arm to simulate real anomalies in the laboratory environment to generate labeled data. As Figure 3 shown, the training steps include: Step 301, Model confidence evaluation, including: According to the task requirements and model performance, preset a confidence threshold (denoted as T) in advance; in the fault identification task, if the maximum classification probability of the model for a certain sample is lower than the threshold T, it is considered that the prediction result of the model for this sample is not reliable enough, and trigger the robotic arm to simulate real anomalies in the laboratory environment and enter Step 302.

[0097] Input the collected physical parameters into the trained model, and the model will output the probability distribution of each possible fault type. For example, for three anomaly types of impeller imbalance, bearing wear, and foreign object interference, the model will output three probability values P1, P2, and P3, corresponding to the probabilities of these three anomaly types respectively. For the calculation of confidence, usually take the maximum value in the probability distribution as the confidence C of the model, that is, C = max(P1, P2, P3).

[0098] Step 302, Trigger the robotic arm to simulate anomalies and generate a dataset to be labeled, including: Identify low-confidence samples: When C < T, determine that this sample is a low-confidence sample; Determine the anomaly type: Although the prediction result of the model for low-confidence samples is not reliable, the possible direction of the anomaly type can still be roughly judged according to the probability distribution; for example, if P1 is relatively large, it can be preliminarily considered that there may be an anomaly related to impeller imbalance.

[0099] Robotic arm simulation of anomalies: Operate the equipment in a laboratory environment by controlling the robotic arm to simulate corresponding abnormal situations. For example, if it is preliminarily determined that the impeller is unbalanced, the robotic arm can simulate the working condition of impeller imbalance by adjusting the position of the counterweight block of the impeller or changing the installation angle of the impeller. During the process of the robotic arm simulating anomalies, use BMC and relevant sensors to collect multi-modal data such as the vibration, speed, and temperature of the equipment as the generated labeled data.

[0100] Step 303: Perform data annotation on the dataset to be annotated to obtain an annotated dataset, including: Annotate the data generated by the robotic arm simulating anomalies. For example, for the data simulating impeller imbalance, the annotator can determine it as an impeller imbalance anomaly and annotate the severity level (level 3) according to the degree of imbalance. Integrate the annotation results with the collected multi-modal data to form a complete annotated dataset.

[0101] Step 304: Use the annotated dataset to update and optimize the model, including: Add the newly generated labeled data to the original training dataset to expand the scale and diversity of the training data. Retrain the neural network model using the expanded training dataset. An incremental learning method can be adopted, that is, continue to train with new data based on the original model parameters to retain the model's learning ability for the original data and improve the model's recognition ability for new abnormal situations.

[0102] Evaluate the updated model on an independent test set to check whether the performance of the model has been improved. If indicators such as the accuracy rate and recall rate of the model have increased, it indicates that the active learning strategy is effective; otherwise, the reason needs to be further analyzed, and the strategy or model structure needs to be adjusted.

[0103] In some embodiments, the method further includes: Obtain a failure database of at least one fan, and fine-tune the second model according to the failure database to obtain a second model corresponding to the electronic device using the fan.

[0104] Here, after obtaining the above second model, the second model can be fine-tuned based on the failure databases of each fan manufacturer (such as the EBMPapst NBR dataset) to adapt to each fan model, that is, obtain a second model for each fan model. In this way, the second model adopted by BMC can be obtained by matching according to the fan model of the electronic device, and using the matched second model for recognition can improve the recognition accuracy.

[0105] The method provided by the embodiments of the present disclosure dynamically adjusts the control parameters of the cooling fan according to the temperatures of the key components of the electronic device and the real-time noise, taking into account the overall heat dissipation requirements of the electronic device, and better managing and controlling the power consumption of the electronic device, improving the resource utilization efficiency, and reducing the operating cost. It realizes the leap from "passive response" to "intelligent prediction" and from "single temperature control" to "multi-objective optimization", and is especially suitable for harsh scenarios sensitive to noise such as high-density computer rooms and medical equipment computer rooms, with both economic and technological foresight.

[0106] During the control process, the second model can be used to accurately identify the operating state of the fan (including normal and various abnormal states), and then the first model is used to dynamically control the fan speed to ensure that the fan can operate efficiently and stably under different working conditions.

[0107] Among them, by identifying abnormal states, potential faults (such as fan imbalance, hard disk mechanical faults, or heat sink looseness) can be detected, providing accurate diagnostic basis for maintenance personnel. Predictive maintenance reduces the need for emergency repairs, optimizes spare parts inventory and manpower scheduling, and extends the equipment life cycle (such as avoiding premature replacement of expensive components).

[0108] By calculating the optimal control parameters, the BMC dynamically controls the fan speed, while optimizing multi-dimensional objectives such as heat dissipation efficiency, noise control, energy consumption, and hardware life, breaking the limitations of traditional single-objective optimization (such as only reducing temperature). Through reinforcement learning, continuously explore the optimal control strategy to adapt to long-term environmental changes (such as seasonal temperature fluctuations) and hardware aging (such as fan efficiency decline), realizing precise heat dissipation control and meeting the customer's noise requirements.

[0109] Figure 4 It is a schematic structural diagram of a heat dissipation control device provided by the embodiments of the present disclosure; as Figure 4 shown, the device is applied to an electronic device, the electronic device includes a heat dissipation device, and the device includes: An acquisition module, configured to acquire the physical parameter set of the electronic device and the noise generated by the heat dissipation device; A first processing module, configured to determine control parameters for heat dissipation according to the physical parameter set and the noise; A second processing module, configured to control the heat dissipation device of the electronic device according to the control parameters.

[0110] In some embodiments, the acquisition module is configured to collect at least one physical parameter of the electronic device and determine the physical parameter set according to the collected at least one physical parameter; Wherein, the at least one physical parameter includes at least one of the following: component temperature, fan speed; the components include: CPU, memory, and PCIE device.

[0111] In some embodiments, the first processing module is configured to determine an evaluation coefficient of at least one component based on a first rule according to the temperature parameter of the at least one component and the noise. Based on a second rule, determine the control parameter according to the evaluation coefficient of the at least one component.

[0112] In some embodiments, the first processing module is configured to determine the temperature weight and temperature threshold corresponding to each component according to each component. Determine the evaluation coefficient of each component according to the temperature parameter, the temperature threshold, the temperature weight, the noise, and the noise weight of each component.

[0113] In some embodiments, the first processing module is configured to determine a temperature control index according to the temperature parameter and the temperature threshold of each component. Determine a noise control index according to the noise and a reference noise. Determine the evaluation coefficient of each component according to the temperature control index, the noise control index, the temperature weight, and the noise weight.

[0114] In some embodiments, the first processing module is configured to determine an evaluation coefficient vector according to the evaluation coefficient of the at least one component. Input the evaluation coefficient vector into a first model to obtain an output result of the first model, where the output result is used to indicate an adjustment range of PID parameters.

[0115] In some embodiments, before determining a control parameter for heat dissipation according to the physical parameter set and the noise, the first processing module is configured to determine whether there is a heat dissipation failure according to the physical parameter set and the noise. If it is determined that there is a heat dissipation failure, generate a fault alarm, where the fault alarm is used to indicate the type of the fault that occurs. If it is determined that there is no heat dissipation failure, determine a control parameter for heat dissipation according to the physical parameter set and the noise.

[0116] In some embodiments, the first processing module is configured to generate time series data of rotation speed and / or temperature according to the physical parameter set. Extract a first feature set according to the noise, where the first feature set includes: BPF, HAR. Input the first feature set and the time series data of the rotation speed and / or temperature into a second model to obtain an output result of the second model, where the output result is used to indicate the type of the fault and the fault abnormality level.

[0117] In some embodiments, the first processing module is configured to generate a fault alarm according to the fault type and the fault anomaly level if the fault type belongs to a target fault type and the fault anomaly level exceeds a level threshold.

[0118] In some embodiments, the first processing module is further configured to obtain a training data set; the training data set includes: at least one training data and the label of each training data, and each training data includes: time series data of sample noise, sample rotation speed, and / or temperature; the label is a fault type and a fault anomaly level; Train a hybrid neural network using the training data set to obtain the trained hybrid neural network as the second model; Wherein, the hybrid neural network includes: a front-end network, a feature fusion layer, and an output layer; The front-end network adopts a first branch network and a second branch network. The first branch network is used to extract the features of the sample noise, and the second branch network is used to extract the features of the time series data of the sample rotation speed and / or temperature; The feature fusion layer is configured to perform data processing using an attention mechanism according to the features extracted by the front-end network; The output layer is configured to obtain an output result according to the data processing result of the feature fusion layer.

[0119] In some embodiments, the first processing module is further configured to obtain a fault database of at least one fan, and fine-tune the second model according to the fault database to obtain a second model corresponding to the electronic device using the fan.

[0120] In some embodiments, the acquisition module is configured to respectively acquire a first noise generated by a fan and / or a second noise of an air duct; Combine the first noise and the second noise to obtain the noise generated by the heat dissipation device.

[0121] It can be understood that when the heat dissipation control device provided in the above embodiments implements the corresponding heat dissipation control method, the above processing can be allocated to different program modules as needed to complete all or part of the processing described above. In addition, the device provided in the above embodiments and the corresponding method embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be elaborated here.

[0122] The embodiments of the present application provide a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the heat dissipation control method.

[0123] An embodiment of the present application provides a computer-readable storage medium storing executable instructions, where the executable instructions, when executed by a processor, cause the processor to execute the heat dissipation control method provided by the embodiment of the present application.

[0124] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disc, or CD-ROM; or may be various devices including one or any combination of the above memories.

[0125] In some embodiments, the executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including being deployed as an independent program or being deployed as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0126] As an example, the executable instructions may or may not correspond to a file in the file system, may be stored as part of a file that stores other programs or data, for example, stored in one or more scripts in a Hyper Text Markup Language (HTML) document, stored in a single file dedicated to the program under discussion, or stored in multiple cooperating files (for example, files that store one or more modules, subroutines, or code portions).

[0127] As an example, the executable instructions may be deployed to execute on one computing device, or on multiple computing devices located at one location, or, on multiple computing devices distributed at multiple locations and interconnected by a communication network.

[0128] Figure 5 A schematic structural diagram of an electronic device provided for an embodiment of the present disclosure; as Figure 5 shown, the electronic device 50 includes: a processor 501, and a memory 502 communicatively connected to the processor 501; the memory 502 stores instructions executable by the processor 501. The instructions are executed by the processor 501 to enable the processor 501 to execute: Obtain a set of physical parameters of the electronic device and the noise generated by the heat dissipation device; Determine control parameters for heat dissipation according to the set of physical parameters and the noise; Control the heat dissipation device of the electronic device according to the control parameters.

[0129] The embodiments of the electronic device and the corresponding control method provided in the above embodiments belong to the same concept. For the specific implementation process, please refer to the method embodiments, which will not be elaborated here.

[0130] In practical applications, the electronic device 50 may further include: at least one network interface 503. Each component in the electronic device 50 is coupled together through a bus system 504. It can be understood that the bus system 504 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 504 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clear description, in Figure 5 the various buses are all labeled as the bus system 504. Among them, the number of the processors 501 can be at least one, and the number of the memories 502 can be at least one. The network interface 503 is used for the communication between the electronic device 50 and other devices in a wired or wireless manner.

[0131] The memory 502 in the embodiments of the present disclosure is used to store various types of data to support the operation of the electronic device 50.

[0132] The methods disclosed in the above embodiments of the present disclosure can be applied to the processor 501 or implemented by the processor 501. The processor 501 may be an integrated circuit chip with signal processing capabilities. In the implementation process, the steps of the above methods can be completed by the integrated logic circuit in the hardware of the processor 501 or the instructions in the form of software. The above-mentioned processor 501 may be a general-purpose processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 501 can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present disclosure. The general-purpose processor may be a microprocessor or any conventional processor, etc. Combining the steps of the methods disclosed in the embodiments of the present disclosure, it can be directly embodied as being executed and completed by the hardware decoding processor, or by a combination of the hardware and software modules in the decoding processor. The software module may be located in the storage medium, and this storage medium is located in the memory 502. The processor 501 reads the information in the memory 502 and combines its hardware to complete the steps of the foregoing control method.

[0133] In some embodiments, the electronic device 50 may be implemented by one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontroller units (MCUs), microprocessors, or other electronic components for performing the foregoing methods.

[0134] It should be understood that the various forms of the processes shown above can be used, and steps can be reordered, added, or deleted. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. There is no limitation herein.

[0135] In the above description, reference is made to "some embodiments", which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0136] Unless otherwise defined, all technical and scientific terms used in this disclosure have the same meaning as commonly understood by those of ordinary skill in the technical field to which this disclosure belongs. The terms used in this disclosure are only for the purpose of describing the embodiments of this disclosure and are not intended to limit this disclosure.

[0137] It should be understood that in the various embodiments of this disclosure, the magnitudes of the sequence numbers of the various implementation processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation processes of the embodiments of this disclosure.

[0138] In addition, the terms "first" and "second" are used only for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this disclosure, "a plurality" means two or more, unless otherwise specifically defined.

[0139] As described above, it is only the specific implementation manner of the present disclosure. However, the protection scope of the present disclosure is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present disclosure can easily think of changes or substitutions, which should all be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.

Claims

1. A heat dissipation control method, characterized in that, The method is applied to an electronic device, and the electronic device includes a heat dissipation device. The method includes: Obtaining a set of physical parameters of the electronic device and the noise generated by the heat dissipation device; Determining control parameters for heat dissipation according to the set of physical parameters and the noise; Controlling the heat dissipation device of the electronic device according to the control parameters.

2. The method according to claim 1, wherein The obtaining of the set of physical parameters of the electronic device includes: Collecting at least one physical parameter of the electronic device, and determining the set of physical parameters according to the collected at least one physical parameter; Wherein, the at least one physical parameter includes at least one of the following: component temperature, rotation speed of the fan; the components include: central processing unit CPU, memory, and peripheral component interconnect express PCIE device.

3. The method according to claim 1, wherein The determining of the control parameters for heat dissipation according to the set of physical parameters and the noise includes: Based on a first rule, determining an evaluation coefficient of at least one component according to the temperature parameter of at least one component and the noise; Based on a second rule, determining the control parameters according to the evaluation coefficient of the at least one component.

4. The method according to claim 3, characterized in that, Based on the first rule, determining at least one control index according to the temperature parameter of at least one component and the noise, including: Determining a temperature weight and a temperature threshold corresponding to each component according to each component; Determining an evaluation coefficient of each component according to the temperature parameter, the temperature threshold, the temperature weight, the noise, and the noise weight of each component.

5. The method according to claim 4, wherein Determining an evaluation coefficient of each component according to the temperature parameter, the noise, the temperature weight, and the noise weight of each component includes: Determining a temperature control index according to the temperature parameter and the temperature threshold of each component; Determining a noise control index according to the noise and a reference noise; Determining an evaluation coefficient of each component according to the temperature control index, the noise control index, the temperature weight, and the noise weight.

6. The method according to claim 3, characterized in that, The determining of the control parameters based on the second rule according to the evaluation coefficient of the at least one component includes: Determining an evaluation coefficient vector according to the evaluation coefficient of the at least one component; Inputting the evaluation coefficient vector into a first model to obtain an output result of the first model, and the output result is used to indicate the adjustment range of the PID parameters.

7. The method according to claim 1, wherein Before the determining of the control parameters for heat dissipation according to the set of physical parameters and the noise, the method further includes: Determining whether there is a heat dissipation fault according to the set of physical parameters and the noise; If it is determined that there is a heat dissipation fault, generating a fault alarm, and the fault alarm is used to indicate the type of the fault that occurs; If it is determined that there is no heat dissipation fault, determining the control parameters for heat dissipation according to the set of physical parameters and the noise.

8. The method according to claim 7, characterized in that, Determining whether there is a heat dissipation fault according to the set of physical parameters and the noise includes: Generating time series data of rotation speed and / or temperature according to the set of physical parameters; Extracting a first feature set according to the noise, and the first feature set includes: blade passing frequency BPF, harmonic component amplitude ratio HAR; Input the first feature set and the time series data of the rotational speed and / or temperature into a second model to obtain an output result of the second model, where the output result is used to indicate a fault type and a fault anomaly level.

9. The method according to claim 7, wherein The generating of the fault warning includes: If the fault type belongs to a target fault type and the fault anomaly level exceeds a level threshold, generate a fault warning according to the fault type and the fault anomaly level.

10. The method according to claim 8, wherein The method further includes: generating the second model; the generating of the second model includes: Obtain a training data set; the training data set includes: at least one training data and a label for each training data, and each training data includes: sample noise, time series data of sample rotational speed and / or temperature; the label is a fault type and a fault anomaly level; Use the training data set to train a hybrid neural network to obtain the trained hybrid neural network as the second model; Wherein, the hybrid neural network includes: a front-end network, a feature fusion layer, and an output layer; The front-end network adopts a first branch network and a second branch network. The first branch network is used to extract features of the sample noise, and the second branch network is used to extract features of the time series data of the sample rotational speed and / or temperature; The feature fusion layer is used to perform data processing on the features extracted by the front-end network by adopting an attention mechanism; The output layer is used to obtain an output result according to the data processing result of the feature fusion layer.

11. The method according to claim 10, wherein The method further includes: Obtain a fault database of at least one fan, and fine-tune the second model according to the fault database to obtain a second model corresponding to the electronic device adopting the fan.

12. The method according to claim 1, characterized in that, Obtaining the noise generated by the heat dissipation device includes: Obtain the first noise generated by the fan and / or the second noise of the air duct respectively; Combine the first noise and the second noise to obtain the noise generated by the heat dissipation device.

13. A heat dissipation control device, characterized in that, The device is applied to an electronic device, the electronic device includes a heat dissipation device, and the device includes: An obtaining module, configured to obtain a physical parameter set of the electronic device and the noise generated by the heat dissipation device; A first processing module, configured to determine a control parameter for heat dissipation according to the physical parameter set and the noise; A second processing module, configured to control the heat dissipation device of the electronic device according to the control parameter.

14. An electronic device, characterized in that, Includes: At least one processor; And a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1 to 12.

15. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 12.

Images