CN115045855B - Cooling fan control method and system based on temperature closed-loop control - Google Patents
Cooling fan control method and system based on temperature closed-loop control Download PDFInfo
- Publication number
- CN115045855B CN115045855B CN202210958164.XA CN202210958164A CN115045855B CN 115045855 B CN115045855 B CN 115045855B CN 202210958164 A CN202210958164 A CN 202210958164A CN 115045855 B CN115045855 B CN 115045855B
- Authority
- CN
- China
- Prior art keywords
- fan control
- control
- information
- fan
- heat dissipation
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04D—NON-POSITIVE-DISPLACEMENT PUMPS
- F04D27/00—Control, e.g. regulation, of pumps, pumping installations or pumping systems specially adapted for elastic fluids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F1/00—Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
- G06F1/16—Constructional details or arrangements
- G06F1/20—Cooling means
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- Feedback Control In General (AREA)
- Control Of Temperature (AREA)
Abstract
The application discloses cooling fan control method and system based on temperature closed-loop control, when the target heat dissipation state information of the target heat dissipation partition uploaded by a service host is obtained, temperature situation data and environment state data according to time domain distribution can be obtained, the target heat dissipation state information is loaded into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, a fan control instruction sequence aiming at each heat dissipation control node of the target heat dissipation partition is generated, and therefore the fan control instruction sequence of each heat dissipation control node is issued to the service host, so that a CPU of the service host controls a cooling fan according to the fan control instruction sequence of each heat dissipation control node, and fan control decision is carried out by combining the temperature situation and the environment state, and the heat dissipation effect is improved.
Description
Technical Field
The application relates to the technical field of intelligent control, in particular to a cooling fan control method and system based on temperature closed-loop control.
Background
In the process of practical application, a high-end service host adopts a configuration with high power consumption and high rotating speed, so that the heat dissipation state of the high-end service host is related to the power consumption of the host and the running performance of the host. How to effectively control heat dissipation of a cooling fan of a service host is a technical problem to be researched urgently at present. For example, in the related art, a closed-loop feedback strategy for monitoring temperature situation data in real time is generally adopted for controlling the cooling fan, but only considering the temperature situation data has a problem of poor heat dissipation effect.
Disclosure of Invention
The application provides a cooling fan control method and system based on temperature closed-loop control.
In a first aspect, the present application provides a cooling fan control method based on temperature closed-loop control, applied to a distributed control system, including:
the method comprises the steps that target heat dissipation state information obtained by a heat dissipation state monitoring device through monitoring a target heat dissipation partition of a host case of a service host at present is obtained, when the target heat dissipation state information reaches a cooling fan control triggering condition, the service host is controlled to send the target heat dissipation state information of the corresponding target heat dissipation partition, the target heat dissipation state information comprises temperature situation data and environment state data distributed according to time domains, the environment state data comprises state data of a plurality of environment state dimensions, and the environment state dimensions comprise environment drying parameter dimensions and environment dust parameter dimensions;
loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition;
and issuing the fan control instruction sequence of each heat dissipation control node to the service host, so that the CPU of the service host controls the cooling fan according to the fan control instruction sequence of each heat dissipation control node.
The step of obtaining target heat dissipation state information obtained by monitoring a target heat dissipation partition of a host case of a service host by a heat dissipation state monitoring device at present comprises the following steps:
determining whether the heat dissipation state monitoring processing is needed, and generating heat dissipation state monitoring notification information when the heat dissipation state monitoring processing is needed;
sending the heat dissipation state monitoring notification information to the heat dissipation state monitoring device, wherein the heat dissipation state monitoring device is used for monitoring the internal space of the host case after receiving the heat dissipation state monitoring notification information to obtain corresponding target heat dissipation state information, and sending the target heat dissipation state information to the heat dissipation state monitoring processing component;
and acquiring target heat dissipation state information obtained by monitoring the internal space of the host case by the heat dissipation state monitoring device based on the heat dissipation state monitoring notification information.
For example, the step of determining whether the heat dissipation state monitoring processing is required, and generating the heat dissipation state monitoring notification information when the heat dissipation state monitoring processing is required includes:
determining whether a heat dissipation state monitoring trigger instruction is received, wherein the heat dissipation state monitoring trigger instruction is generated when a CPU in an internal space of the host case runs at least one target application program and/or runs power consumption which is larger than or equal to a preset power consumption threshold value;
if the heat dissipation state monitoring trigger instruction is received, determining that heat dissipation state monitoring processing is needed, and when the heat dissipation state monitoring processing is needed, analyzing the heat dissipation state monitoring trigger instruction to obtain a first operation parameter carried by the heat dissipation state monitoring trigger instruction, wherein the first operation parameter has a positive correlation with the number of target application programs currently operated by the CPU and/or the operation power consumption of the CPU;
the method comprises the steps of determining corresponding state monitoring frequency information based on the first operating parameters, generating corresponding heat dissipation state monitoring notification information based on the state monitoring frequency information, wherein the state monitoring frequency information has positive correlation with the first operating parameters, and the heat dissipation state monitoring device is used for monitoring the internal space of the host case based on the state monitoring frequency information after receiving the heat dissipation state monitoring notification information.
For example, the step of determining corresponding state monitoring frequency information based on the first operating parameter and generating corresponding heat dissipation state monitoring notification information based on the state monitoring frequency information includes:
determining corresponding state monitoring frequency information based on the first operation parameter, wherein the state monitoring frequency information and the first operation parameter have positive correlation;
determining corresponding state monitoring duration information based on the first operation parameter, wherein the state monitoring duration information and the first operation parameter have a negative correlation;
and generating corresponding heat dissipation state monitoring notification information based on the state monitoring frequency information and the state monitoring duration information, wherein the heat dissipation state monitoring device is used for monitoring the internal space of the host case based on the state monitoring frequency information and the state monitoring duration information after receiving the heat dissipation state monitoring notification information, the state monitoring duration information is used for representing that target heat dissipation state information obtained by monitoring is sent to the heat dissipation state monitoring processing assembly at intervals corresponding to duration, and the duration of each piece of target heat dissipation state information is the state monitoring duration information.
In a second aspect, the present application provides a distributed control system comprising:
a processor;
a memory having stored therein a computer program that, when executed, implements the cooling fan control method based on temperature closed-loop control of the first aspect.
In a third aspect, an embodiment of the present application provides a cooling fan control system based on temperature closed-loop control, where the cooling fan control system based on temperature closed-loop control includes a distributed control system and one or more service hosts communicatively connected to the distributed control system, where the service hosts include a cooling status monitoring device, and the distributed control system is configured to:
acquiring target heat dissipation state information obtained by monitoring a target heat dissipation partition of a host case of a service host by a heat dissipation state monitoring device at present;
when the target heat dissipation state information reaches a cooling fan control trigger condition, controlling the service host to send target heat dissipation state information of a corresponding target heat dissipation partition, wherein the target heat dissipation state information comprises temperature situation data distributed according to a time domain and environment state data, the environment state data comprises state data of a plurality of environment state dimensions, and the environment state dimensions comprise an environment drying parameter dimension and an environment dust parameter dimension;
loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition;
and issuing the fan control instruction sequence of each heat dissipation control node to the service host, so that the CPU of the service host controls the cooling fan according to the fan control instruction sequence of each heat dissipation control node.
By adopting the technical scheme of any one aspect, when the target heat dissipation state information of the target heat dissipation partition uploaded by the service host is obtained, the temperature situation data and the environment state data distributed according to the time domain can be obtained, so that the target heat dissipation state information is loaded into the cooling fan control model meeting the optimization and selection deployment requirements of the parameter layer information, a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition is generated, the fan control instruction sequence of each heat dissipation control node is issued to the service host, a CPU of the service host controls a cooling fan according to the fan control instruction sequence of each heat dissipation control node, a fan control decision is made by combining the temperature situation and the environment state, and the heat dissipation effect is improved.
Drawings
FIG. 1 is a flowchart illustrating steps of a cooling fan control method based on temperature closed-loop control according to an embodiment of the present disclosure;
fig. 2 is a schematic block diagram of a structure of a distributed control system according to an embodiment of the present application.
Detailed Description
As shown in fig. 1-2, an embodiment of the present application provides a cooling fan control method based on temperature closed-loop control, which is applied to a distributed control system, and includes:
step S110, obtaining target heat dissipation state information obtained by the current monitoring of the target heat dissipation partition of the host enclosure of the service host by the heat dissipation state monitoring device, and controlling the service host to send the target heat dissipation state information of the corresponding target heat dissipation partition when the target heat dissipation state information reaches the cooling fan control trigger condition.
For example, the target heat dissipation state information includes temperature situation data according to time domain distribution and environment state data, the temperature situation data according to the time domain distribution may refer to variation trend data of a temperature state according to the time domain distribution, and the environment state data includes state data of a plurality of environment state dimensions, for example, the environment state dimensions include an environment drying parameter dimension and an environment dust parameter dimension.
Step S120, loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of the parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition.
For example, the cooling fan control model may be pre-trained, and the trained cooling fan control model may have the capability of outputting a fan control instruction sequence corresponding to the target heat dissipation state information of the target heat dissipation section.
Step S130, issuing the fan control instruction sequence of each heat dissipation control node to the service host, so that the CPU of the service host controls the cooling fan according to the fan control instruction sequence of each heat dissipation control node.
According to the above steps, in this embodiment, when the target heat dissipation state information of the target heat dissipation partition uploaded by the service host is obtained, the temperature situation data and the environmental state data distributed according to the time domain can be obtained, so that the target heat dissipation state information is loaded into the cooling fan control model meeting the tuning and selecting deployment requirements of the parameter layer information, a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition is generated, and thus the fan control instruction sequence of each heat dissipation control node is issued to the service host, so that the CPU of the service host controls the cooling fan according to the fan control instruction sequence of each heat dissipation control node, and a fan control decision is made by combining the temperature situation and the environmental state, thereby improving the heat dissipation effect.
For some exemplary embodiments, step S120 may be implemented by the following technical solutions.
The processor 110 loads the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, analyzes, according to the cooling fan control model, a composite temperature situation feature distribution of each environment state dimension of the environment state data corresponding to the temperature situation data according to time domain distribution in the target heat dissipation state information, and the composite temperature situation feature is configured to represent a basic control feature distribution in which the temperature situation data according to time domain distribution performs weighted aggregation according to the environment state feature distribution of each environment state dimension of the environment state data.
And the Process120 performs fan control instance generation on the composite temperature situation characteristic distribution according to the cooling fan control model, and determines fan control instance information matched with historical fan control strategy data.
The cooling fan control model performs deep learning determination according to composite temperature situation feature distribution of a specified environment state dimension and example fan control instance information of a plurality of target fan control variables, the example fan control instance information and the specified environment state dimension have a specified correlation, and fan control instance information covering the example fan control categories is generated according to the fan control instance information obtained from the example fan control instance information of each fan control category, so that after the composite temperature situation feature distribution is loaded to the cooling fan control model, the fan control instance information of the plurality of target fan control variables can be obtained, and each piece of fan control instance information comprises historical fan control strategy data. The example fan control instance information covering the example fan control category may refer to example fan control instance information having control policy data for a fan control instance with a target environmental state dimension.
The example fan control instance information is parsed from the example fan control instance information for different fan control requirements based on fan control instance information obtained from the example fan control instance information for each fan control category, and then example fan control instance information covering the example fan control categories and matching a plurality of different target fan control variables is generated based on the parsed fan control instance information.
The target fan control variable may be a fan control alternative instruction for generating actual control of the environmental state dimension, and in the process of tuning and selecting parameter layer information, example fan control instance information of a plurality of target fan control variables and composite temperature situation feature distribution corresponding to the specified environmental state dimension may be used to perform deep learning on the cooling fan control model, so as to output the corresponding cooling fan control model.
And the processor 130 processes the fan control instance information according to the cooling fan control model, and outputs application fan control instance information which is associated with each target fan control variable and covers the target fan control category.
The Process140 generates a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition according to the application fan control instance information associated with each target fan control variable and covering the target fan control category.
By adopting the above technical solution, in this embodiment, fan control instance information is analyzed from examples of different fan control types, example fan control instance information of a plurality of target fan control variables covering the example fan control types is generated according to the analyzed fan control instance information, deep learning is performed on the cooling fan control model according to the composite temperature situation characteristic distribution of the specified environment state dimension and the example fan control instance information of different target fan control variables covering the example fan control types and having a specified association relationship with the specified environment state dimension, so that different cooling fan control models do not need to be separately configured according to the example fan control instance information of different fan control types of different target fan control variables, and the deep learning effect is improved. And fan regulation and control instance generation is carried out on the composite temperature situation characteristic distribution of the relevant environment state dimension according to the trained cooling fan control model, fan regulation and control instance information matched with historical fan regulation and control strategy data is determined, then the fan regulation and control instance information is processed according to the trained cooling fan control model, and application fan regulation and control instance information of a plurality of target fan regulation and control variables of a target fan regulation and control type is output, so that when one fan regulation and control requirement is changed to another fan regulation and control requirement, the fan regulation and control type of the application fan regulation and control instance information does not need to carry out optimization and selection of parameter layer information again, namely, other cooling fan control models do not need to be dispatched additionally.
For some exemplary embodiments, the Process120 may be implemented by the following technical solutions.
And the Process122 performs a fan control decision on the composite temperature situation characteristic distribution according to the cooling fan control model, and determines a fan control curve of the fan control dimension.
Wherein the fan control curve of the fan control dimension is fan control component information of each fan control dimension in the environmental state dimension.
For some exemplary embodiments, process122, for example, performs the following steps: and carrying out fan regulation decision on each fan regulation dimension composite temperature situation characteristic in the composite temperature situation characteristic distribution according to a fan regulation decision branch in the fan regulation instance output submodel, and determining the fan regulation dimension characteristic of a fan regulation curve covering the fan regulation dimension. For example, assuming that the composite temperature situation characteristic distribution is [ i1, i 2., iU ], loading the composite temperature situation characteristic distribution to a fan control decision branch, and performing a fan control decision on the loaded composite temperature situation characteristic distribution [ i1, i 2., iU ], by the fan control decision branch, and determining a fan control dimension characteristic [ j1, j 2., jU ] of a fan control curve covering the fan control dimension.
The Process124 performs fan control instance generation on the fan control curve of the fan control dimension according to the historical fan control strategy data of the target environment state dimension, and determines fan control instance information including the historical fan control strategy data of the plurality of target fan control variables.
The historical fan control strategy data is the control strategy data of the fan control example of the target environment state dimension learned in the fan control learning process by the cooling fan control model.
In addition, after obtaining the fan control instance information, the method may further include the steps of:
the Process126 processes the fan control instance information according to the cooling fan control model, and outputs application fan control instance information which is associated with each target fan control variable and covers the target fan control category.
The detailed execution steps of the Process126 can refer to the Process130 described above.
The Process128 generates a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition according to the application fan control instance information associated with each target fan control variable and covering the target fan control category.
For some exemplary embodiments, before the foregoing embodiments, the cooling fan control model may be deeply learned, and after tuning and selecting parameter layer information is completed, the cooling fan control model is configured in the distributed control system based on temperature closed-loop control, and the tuning and selecting process of the parameter layer information of the cooling fan control model may be implemented by the following technical solutions:
the Process202 analyzes the example fan control instance information of the target fan control variables one by one, and outputs the example fan control instance information.
The example fan control instance information may refer to fan control instance information configured in different environmental state dimensions, and fan control instance information configured in different environmental state dimensions, where fan control requirements and fan control categories are different. The different environmental state dimensions include a target environmental state dimension and other environmental state dimensions, and the target environmental state dimension may refer to an environmental state dimension of a desired fan regulation category.
For some exemplary embodiments, after obtaining example fan control instance information of a plurality of target fan control variables, analyzing fan control time node and fan control space node information of a fan control instance from the example fan control instance information, and determining a fan control curve of an example fan control dimension of the example fan control instance information according to the fan control time node and the fan control space node information of the fan control instance; and carrying out fan regulation example generation on a fan regulation and control curve of the fan regulation and control dimension of the example, and determining the fan regulation and control time-space domain information.
The Process204 generates first example instance information covering the example fan control category for a plurality of target fan control variables based on the example fan control instance information.
The Process206, when obtaining the example composite temperature situation feature distribution from the specified environmental state dimension corresponding to the example fan control instance information, performs fan control instance generation on the example composite temperature situation feature distribution according to the cooling fan control model to determine second example instance information matching the historical fan control policy data.
The Process208 processes the second example information according to the cooling fan control model, and outputs fan control output example information including a plurality of target fan control variables and covering the example fan control categories.
The Process210 determines corresponding model performance parameter values according to the fan regulation output instance information and the example fan regulation instance information, and performs parameter layer information optimization and selection on the cooling fan control model according to the model performance parameter values.
The processes 206 to 210 are the optimization and selection stages of parameter layer information of the cooling fan control model, and can perform deep learning on the cooling fan control model by using all example fan control instance information as rebate data, and can also perform deep learning on the cooling fan control model by two branches, wherein: the branch 1 is used for performing deep learning by taking the example fan regulation and control instance information corresponding to the target environment state dimension as the example fan regulation and control instance information, and the training of the branch 2 is performed after the training is finished; branch 2, the cooling fan control model continues to be deeply learned using the generated example fan control instance information covering the example fan control categories (i.e., fan control categories for the target environmental state dimension).
The following embodiments will be described in detail for the above two branches:
and branch 1, performing deep learning by taking the example fan control instance information sent according to the target environment state dimension as the example fan control instance information.
For some exemplary embodiments, the target fan control variables include first target fan control variables for which the target environment state dimension issues corresponding example fan control instance information, such that the example composite temperature posture feature distribution includes first example composite temperature posture feature distributions associated with the first target fan control variables. The Process206, for example, performs the steps of: the terminal carries out fan regulation and control instance generation on the first example composite temperature situation characteristic distribution according to the cooling fan control model, and determines first generation instance information matched with historical fan regulation and control strategy data as second example instance information; the Process208, for example, performs the steps of: the terminal processes the first generation instance information according to the cooling fan control model and outputs fan control instance information of a first target fan control variable covering the fan control type of the example; the Process210, for example, performs the steps of: and adjusting and selecting parameter layer information of the cooling fan control model according to a first model performance parameter value between the fan control instance information of the first target fan control variable and the example fan control instance information until the cooling fan control model is converged.
After the first model performance parameter value is obtained, the first model performance parameter value can be reversely propagated in the cooling fan control model, so that the gradient value of each fan control function configuration information in the cooling fan control model is obtained, and each fan control function configuration information in the cooling fan control model is updated according to the gradient value.
Branch 2, the cooling fan control model continues to be deeply learned using the generated first example instance information covering the example fan control category.
For some exemplary embodiments, the target fan control variables include second target fan control variables for other different fan control demand environment state dimensions to issue corresponding example fan control instance information, and the example composite temperature situation feature distribution includes second example composite temperature situation feature distributions associated with the second target fan control variables; the Process206, for example, performs the steps of: when the fan control function configuration information of the cooling fan control model is adjusted according to the first model performance parameter value to meet the model application deployment requirement, fan regulation and control instance generation is carried out on the second example composite temperature situation characteristic distribution according to the cooling fan control model, and second generation instance information matched with historical fan regulation and control strategy data is determined to serve as the second example instance information; the Process208, for example, performs the steps of: processing the second generated instance information according to the cooling fan control model, and outputting fan control instance information of a second target fan control variable covering the example fan control type; the Process210, for example, performs the steps of: and the terminal carries out parameter layer information optimization and selection on the cooling fan control model according to a second model performance parameter value between the fan control instance information of the second target fan control variable and the example fan control instance information until the cooling fan control model is trained, so that the training of a second branch is completed, and the final cooling fan control model is generated.
After the second model performance parameter value is obtained, the second model performance parameter value can be propagated in the cooling fan control model in a reverse direction, so that a gradient value of each fan control function configuration information in the cooling fan control model is obtained, and each fan control function configuration information in the cooling fan control model is updated according to the gradient value.
According to the design, the example fan control instance information covering the example fan control types and different target fan control variables is analyzed from the example fan control instance information, the example fan control instance information covering the example fan control types and different target fan control variables is generated according to the example fan control instance information, the cooling fan control model is deeply learned according to the composite temperature situation characteristic distribution of the specified environment state dimension and the example fan control instance information covering the example fan control types and different target fan control variables and having a specified correlation with the specified environment state dimension, and therefore the cooling fan control model used for deciding and generating the cooling fan control models covering the example fan control types and different target fan control variables is obtained, different cooling fan control models do not need to be independently configured according to the example fan control instance information of different fan control types of different target fan control variables, and accordingly the deep learning effect is improved.
For some exemplary embodiments, for training of the cooling fan control model, in the foregoing embodiments, the fan control instance output sub-model and the fan control instance generation adjustment layer in the cooling fan control model may also be deeply learned separately, for example, the following steps are performed:
the Process302 analyzes the example fan control instance information of the target fan control variables one by one, and outputs the example fan control instance information.
Wherein, the fan control types between the fan control instance information of each example have a relationship.
The Process304 generates first example instance information covering the example fan control category for a plurality of target fan control variables based on the example fan control instance information.
The Process306, when obtaining the example composite temperature situation feature distribution from the specified environment state dimension corresponding to the example fan control instance information, performs fan control instance generation on the example composite temperature situation feature distribution according to the fan control instance output sub-model, and determines second example instance information matching the historical fan control policy data.
The detailed execution steps of the processes 302 to 306 can be referred to as the processes 202 to 206 in the foregoing embodiments.
The Process308 determines corresponding model performance parameter values according to the example fan control instance information and the fan control instance information analyzed from the example fan control instance information, and performs weight adjustment on the fan control instance output submodel according to the model performance parameter values.
For some exemplary embodiments, after the model performance parameter value is obtained, the terminal makes the model performance parameter value propagate in the fan regulation and control instance output submodel in the reverse direction, so as to obtain a gradient value of each fan control function configuration information in the fan regulation and control instance output submodel, and updates each fan control function configuration information in the fan regulation and control instance output submodel according to the gradient value until the fan regulation and control instance output submodel converges.
In the process of performing deep learning determination on the fan control instance output submodel, the distributed control system based on the temperature closed-loop control may further perform deep learning on a fan control instance generation adjustment layer, which may be, for example: analyzing the example fan regulation and control example information, and outputting application fan regulation and control example information; generating an adjusting layer according to the fan regulation and control example, processing the application fan regulation and control example information, and outputting target fan regulation and control output example information of different target fan regulation and control variables, wherein the target fan regulation and control output example information covers the regulation and control type of the example fan; and determining corresponding model performance parameter values according to the target fan regulation and control output instance information and the example fan regulation and control instance information, and performing weight adjustment on a fan regulation and control instance generation adjusting layer according to the model performance parameter values.
According to the design, analyzing the example fan control instance information of different fan control types, generating the example fan control instance information covering the example fan control types and different target fan control variables according to the analyzed example fan control instance information, and performing deep learning on the fan control instance output sub-model according to the composite temperature situation characteristic distribution of the specified environment state dimension and the example fan control instance information covering the example fan control types and different target fan control variables with specified incidence relation with the specified environment state dimension; in addition, the application fan control instance information is obtained by analyzing the application fan control instance information, and the fan control instance generation adjustment layer is deeply learned according to the application fan control instance information and the example fan control instance information, so that a cooling fan control model (comprising a trained fan control instance output sub-model and a fan control instance generation adjustment layer) for deciding and generating different target fan control variables covering the example fan control categories is obtained, and different cooling fan control models do not need to be independently configured according to the example fan control instance information of the different fan control categories of the different target fan control variables, so that the deep learning effect is improved.
For some exemplary embodiments, the present application provides a method for training a cooling fan control model, including the following steps:
the Process402 analyzes the example fan control instance information of the target fan control variables one by one, and outputs the example fan control instance information.
The fan control types of the fan control instance information are related.
For some demonstrative embodiments, the example fan regulation instance information includes fan regulation time-space domain information; process402 may be, for example: analyzing fan regulation and control time node information and fan regulation and control space node information of a fan regulation and control example from example fan regulation and control example information of a plurality of target fan regulation and control variables respectively; determining a fan control curve of an example fan control dimension of the example fan control instance information according to the fan control time node and the fan control space node information of the fan control instance, wherein the fan control curve of the example fan control dimension is used for representing fan control variable distribution of each fan control instance in the example fan control instance information, wherein the fan control instances are associated with the target composite temperature situation characteristic; and carrying out fan regulation example generation on a fan regulation and control curve of the fan regulation and control dimension of the example, and determining the fan regulation and control time-space domain information.
The Process404 generates first example instance information of a plurality of target fan control variables covering the example fan control category according to the example fan control instance information.
The Process406, when obtaining the example composite temperature situation feature distribution from the specified environmental state dimension corresponding to the example fan control instance information, performs fan control instance generation on the example composite temperature situation feature distribution according to the cooling fan control model to determine second example instance information matching the historical fan control policy data.
The Process408 processes the second example information according to the cooling fan control model, and outputs fan control output instance information of a plurality of target fan control variables, where the fan control output instance information covers target fan control types corresponding to the example fan control type information.
And the Process410 determines corresponding model performance parameter values according to the fan regulation output instance information and the example fan regulation instance information, and performs parameter layer information optimization and selection on the cooling fan control model according to the model performance parameter values.
The detailed execution steps of the processes 402 to 410 refer to the processes 202 to 210 and the processes 302 to 308.
According to the design, the example fan control instance information covering the example fan control types and different target fan control variables is analyzed from the example fan control instance information, the example fan control instance information covering the example fan control types and different target fan control variables is generated according to the example fan control instance information, the cooling fan control model is deeply learned according to the composite temperature situation characteristic distribution of the specified environment state dimension and the example fan control instance information covering the example fan control types and different target fan control variables and having a specified correlation with the specified environment state dimension, and therefore the different cooling fan control models do not need to be separately configured according to the example fan control instance information of the different fan control types of the different target fan control variables, and accordingly the deep learning effect is improved.
For some possible embodiments, in step S110, it may be determined whether heat dissipation state monitoring processing is required, and when heat dissipation state monitoring processing is required, heat dissipation state monitoring notification information is generated, and the heat dissipation state monitoring notification information is sent to the heat dissipation state monitoring device, where the heat dissipation state monitoring device is configured to monitor an internal space of the host chassis after receiving the heat dissipation state monitoring notification information, obtain corresponding target heat dissipation state information, and send the target heat dissipation state information to the heat dissipation state monitoring processing component;
and acquiring target heat dissipation state information obtained by monitoring the internal space of the host case by the heat dissipation state monitoring device based on the heat dissipation state monitoring notification information.
For example, the step of determining whether the heat dissipation state monitoring process is required, and generating the heat dissipation state monitoring notification information when the heat dissipation state monitoring process is required includes: determining whether a heat dissipation state monitoring trigger instruction is received, wherein the heat dissipation state monitoring trigger instruction is generated when a CPU in an internal space of the host case runs at least one target application program and/or runs power consumption which is larger than or equal to a preset power consumption threshold value; if the heat dissipation state monitoring trigger instruction is received, determining that heat dissipation state monitoring processing is needed, and when the heat dissipation state monitoring processing is needed, analyzing the heat dissipation state monitoring trigger instruction to obtain a first operation parameter carried by the heat dissipation state monitoring trigger instruction, wherein the first operation parameter has a positive correlation with the number of target application programs currently operated by the CPU and/or the operation power consumption of the CPU; the method comprises the steps of determining corresponding state monitoring frequency information based on the first operating parameters, generating corresponding heat dissipation state monitoring notification information based on the state monitoring frequency information, wherein the state monitoring frequency information has positive correlation with the first operating parameters, and the heat dissipation state monitoring device is used for monitoring the internal space of the host case based on the state monitoring frequency information after receiving the heat dissipation state monitoring notification information.
For example, the step of determining corresponding state monitoring frequency information based on the first operating parameter, and generating corresponding heat dissipation state monitoring notification information based on the state monitoring frequency information includes: determining corresponding state monitoring frequency information based on the first operation parameter, wherein the state monitoring frequency information and the first operation parameter have positive correlation; determining corresponding state monitoring duration information based on the first operation parameter, wherein the state monitoring duration information and the first operation parameter have a negative correlation; and generating corresponding heat dissipation state monitoring notification information based on the state monitoring frequency information and the state monitoring duration information, wherein the heat dissipation state monitoring device is used for monitoring the internal space of the host case based on the state monitoring frequency information and the state monitoring duration information after receiving the heat dissipation state monitoring notification information, the state monitoring duration information is used for representing that target heat dissipation state information obtained by monitoring is sent to the heat dissipation state monitoring processing assembly at intervals corresponding to duration, and the duration of each piece of target heat dissipation state information is the state monitoring duration information.
Further, as shown in fig. 2, the embodiment of the present application also provides a distributed control system, and the distributed control system 100 may generate a relatively large difference due to different configurations or performances, and may include one or more Central Processing Units (CPUs) 112 (e.g., one or more processors) and a memory 111. Wherein the memory 111 may be a transient storage or a persistent storage. The program stored in the memory 111 may include one or more modules, each of which may include a series of instruction operations in the distributed control system 100. Still further, a central processor 112 may be provided in communication with the memory 111 to execute a series of instruction operations in the memory 111 on the distributed control system 100.
The distributed control system 100 may also include one or more power supplies, one or more communication units 113, one or more pass-to-output interfaces, and/or one or more operating systems, such as Windows Server, mac OS XTM, unixTM, linuxTM, freeBSDTM, and the like.
In addition, a storage medium is provided in an embodiment of the present application, and the storage medium is used for storing a computer program, and the computer program is used for executing the method provided in the embodiment.
The embodiment of the present application also provides a computer program product including instructions, which when run on a computer, causes the computer to execute the method provided by the above embodiment.
Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium may be at least one of the following media: various media that can store program codes, such as a Read-only Memory (ROM), a RAM, a magnetic disk, or an optical disk.
It should be noted that, in the present specification, each embodiment is described in a progressive manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment is intended to be different from other embodiments. In particular, for the apparatus and system embodiments, since they are substantially similar to the method embodiments, they are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described embodiments of the apparatus and system are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
The above description is only one specific embodiment of the present application, but the scope of the present application is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present application should be included in the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (9)
1. A cooling fan control method based on temperature closed-loop control is applied to a distributed control system and is characterized by comprising the following steps:
the method comprises the steps that target heat dissipation state information obtained by a heat dissipation state monitoring device through monitoring a target heat dissipation partition of a host case of a service host at present is obtained, when the target heat dissipation state information reaches a cooling fan control triggering condition, the service host is controlled to send the target heat dissipation state information of the corresponding target heat dissipation partition, the target heat dissipation state information comprises temperature situation data and environment state data distributed according to time domains, the environment state data comprises state data of a plurality of environment state dimensions, and the environment state dimensions comprise environment drying parameter dimensions and environment dust parameter dimensions;
loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition;
sending the fan control instruction sequence of each heat dissipation control node to the service host, so that the CPU of the service host controls a cooling fan according to the fan control instruction sequence of each heat dissipation control node;
the loading the target heat dissipation state information into a cooling fan control model meeting the tuning and selecting deployment requirements of parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition specifically includes:
loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, analyzing the composite temperature situation characteristic distribution of each environment state dimension of the environment state data corresponding to the temperature situation data distributed according to the time domain in the target heat dissipation state information according to the cooling fan control model, wherein the composite temperature situation characteristic is configured to represent the basic control characteristic distribution of weighted aggregation of the temperature situation data distributed according to the time domain according to the environment state characteristic distribution of each environment state dimension of the environment state data;
carrying out fan regulation decision on the composite temperature situation characteristic distribution according to the cooling fan control model, and determining a fan regulation curve of each fan regulation dimension;
performing fan control instance generation on a fan control curve of a fan control dimension according to historical fan control strategy data of a target environment state dimension, determining a plurality of target fan control variables and inheriting fan control instance information of the historical fan control strategy data, wherein the historical fan control strategy data is the control strategy data of the fan control instance of the target environment state dimension, which is learned by the cooling fan control model in a fan control learning process; the cooling fan control model is determined by deep learning according to composite temperature situation feature distribution of a specified environment state dimension and example fan control instance information of a plurality of target fan control variables, the example fan control instance information has a specified correlation with the specified environment state dimension, and fan control instance information covering the example fan control categories is generated according to the fan control instance information obtained from the example fan control instance information of each fan control category;
processing the fan regulation and control example information according to the cooling fan control model, and outputting application fan regulation and control example information which is related to each target fan regulation and control variable and covers the target fan regulation and control category;
and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition according to the application fan control instance information which is associated with each target fan control variable and covers the target fan control category.
2. The method according to claim 1, wherein before the loading the target heat dissipation state information into the cooling fan control model satisfying requirements for tuning and selecting deployment of parameter layer information, and analyzing the composite temperature situation feature distribution of each environment state dimension of the environment state data corresponding to the temperature situation data distributed according to the time domain in the target heat dissipation state information according to the cooling fan control model, the method further comprises:
analyzing the example fan control instance information of each target fan control variable one by one, and outputting the example fan control instance information;
generating first example instance information for each of the target fan control variables and covering an example fan control category based on the example fan control instance information;
when obtaining an example composite temperature situation feature distribution from a specified environment state dimension corresponding to the first example instance information, performing fan regulation instance generation on the example composite temperature situation feature distribution according to the cooling fan control model, and determining second example instance information inherited to the historical fan regulation policy data;
processing the second example instance information according to the cooling fan control model, outputting fan control output instance information for each of the target fan control variables and associated with the target fan control category;
and determining corresponding model performance parameter values according to the fan regulation and control output instance information and the example fan regulation and control instance information, and carrying out parameter layer information regulation and selection on the cooling fan control model according to the model performance parameter values.
3. The method according to claim 2, wherein the target fan control variables comprise first target fan control variables scheduled by a target environment state dimension for deep learning the example fan control instance information, and the example composite temperature situation feature distribution comprises first example composite temperature situation feature distributions associated with the first target fan control variables;
the generating of the fan control instance for the example composite temperature situation feature distribution according to the cooling fan control model, and determining second example instance information that inherits the historical fan control policy data, specifically include:
performing fan regulation instance generation on the first example composite temperature situation characteristic distribution according to the cooling fan control model, and determining first generation instance information inherited to the historical fan regulation strategy data as the second example instance information;
processing the second example instance information according to the cooling fan control model, outputting fan control output instance information for each of the target fan control variables and associated with the target fan control category, specifically including:
processing the first generation instance information according to the cooling fan control model, and outputting fan control instance information of a first target fan control variable associated with the target fan control category;
the determining, according to the fan control output instance information and the example fan control instance information, a corresponding model performance parameter value, and performing tuning and selection of parameter layer information on the cooling fan control model according to the model performance parameter value specifically includes:
and adjusting and selecting parameter layer information of the cooling fan control model according to a first model performance parameter value between the fan control instance information of the first target fan control variable and the example fan control instance information.
4. The method of claim 3, wherein the target fan control variables comprise second target fan control variables scheduled for deep learning of the example fan control instance information in the environmental state dimension for different other fan control requirements, the example composite temperature situation feature distribution comprising second example composite temperature situation feature distributions associated with the second target fan control variables;
the generating of the fan control instance of the example composite temperature situation feature distribution according to the cooling fan control model and the determining of the second example instance information inherited to the historical fan control policy data specifically include:
when the fan control function configuration information of the cooling fan control model is adjusted according to the first model performance parameter value to meet the model application deployment requirement, fan regulation instance generation is carried out on the second example composite temperature situation characteristic distribution according to the cooling fan control model, and second generation instance information inherited to the historical fan regulation strategy data is determined to serve as the second example instance information;
processing the second example instance information according to the cooling fan control model, outputting fan control output instance information for each of the target fan control variables and associated with the target fan control category, specifically including:
processing the second generated instance information according to the cooling fan control model, and outputting fan control instance information of a second target fan control variable associated with the target fan control category;
the determining, according to the fan control output instance information and the example fan control instance information, a corresponding model performance parameter value, and performing tuning and selection of parameter layer information on the cooling fan control model according to the model performance parameter value specifically includes:
and adjusting and selecting parameter layer information of the cooling fan control model according to a second model performance parameter value between the fan control instance information of the second target fan control variable and the example fan control instance information.
5. The cooling fan control method based on temperature closed-loop control according to claim 1, wherein the cooling fan control model comprises a fan control instance output sub-model;
the determining a fan control decision for the composite temperature situation characteristic distribution according to the cooling fan control model to determine a fan control curve of each fan control dimension specifically includes:
and carrying out fan regulation and control decision on the composite temperature situation characteristic of each fan regulation and control dimension in the composite temperature situation characteristic distribution according to the fan regulation and control decision branch in the fan regulation and control instance output submodel, and determining the fan regulation and control dimension characteristic of a fan regulation and control curve covering the fan regulation and control dimension, wherein the fan regulation and control curve of the fan regulation and control dimension is fan regulation and control component information of each fan regulation and control dimension in the environment state dimension.
6. The method of claim 1, further comprising:
analyzing the example fan control instance information of the target fan control variables one by one, and outputting the example fan control instance information; the fan control types between the information of the example fan control instances are associated;
generating first example instance information for each of the target fan control variables and covering an example fan control category based on the example fan control instance information;
when example composite temperature situation feature distribution is obtained from the designated environment state dimension corresponding to the first example instance information, fan control instance generation is carried out on the example composite temperature situation feature distribution according to a cooling fan control model, and second example instance information matching historical fan control strategy data is determined;
processing the second example instance information according to the cooling fan control model, and outputting fan control output instance information of each target fan control variable; the fan regulation output instance information covers a target fan regulation category associated with the historical fan regulation strategy data;
determining corresponding model performance parameter values according to the fan regulation and control output instance information and the example fan regulation and control instance information, and carrying out parameter layer information regulation and selection on the cooling fan control model according to the model performance parameter values;
wherein the example fan regulation instance information comprises fan regulation time-space domain information;
analyzing the example fan control instance information of the target fan control variables one by one, and outputting the example fan control instance information, specifically comprising:
analyzing fan regulation and control time node information and fan regulation and control space node information of a fan regulation and control example from example fan regulation and control example information of a plurality of target fan regulation and control variables respectively;
determining a fan control curve of an example fan control dimension of the example fan control instance information according to the fan control time node and the fan control space node information of the fan control instance, the fan control curve of the example fan control dimension being used to represent fan control variable distribution of each fan control instance in the example fan control instance information associated with a target composite temperature situation characteristic;
and carrying out fan regulation example generation on the fan regulation and control curve of the example fan regulation and control dimension, and determining the fan regulation and control time-space domain information.
7. The method according to any one of claims 1 to 6, wherein the step of obtaining target heat dissipation state information obtained by the heat dissipation state monitoring device monitoring a target heat dissipation partition of a host chassis of the service host at present comprises:
determining whether the heat dissipation state monitoring processing is needed, and generating heat dissipation state monitoring notification information when the heat dissipation state monitoring processing is needed;
sending the heat dissipation state monitoring notification information to the heat dissipation state monitoring device, wherein the heat dissipation state monitoring device is used for monitoring the internal space of the host case after receiving the heat dissipation state monitoring notification information to obtain corresponding target heat dissipation state information, and sending the target heat dissipation state information to a heat dissipation state monitoring processing component;
and acquiring target heat dissipation state information obtained by monitoring the internal space of the host case by the heat dissipation state monitoring device based on the heat dissipation state monitoring notification information.
8. A distributed control system, comprising:
a processor;
a memory having stored therein a computer program that, when executed, implements the cooling fan control method based on temperature closed-loop control of any one of claims 1-7.
9. A cooling fan control system based on temperature closed-loop control, the cooling fan control system based on temperature closed-loop control comprising a distributed control system and one or more service hosts communicatively connected to the distributed control system, the service hosts comprising a thermal state monitoring device, the distributed control system being configured to:
acquiring target heat dissipation state information obtained by monitoring a target heat dissipation partition of a host case of a service host by a heat dissipation state monitoring device at present;
when the target heat dissipation state information reaches a cooling fan control triggering condition, controlling the service host to send target heat dissipation state information of a corresponding target heat dissipation partition, wherein the target heat dissipation state information comprises temperature situation data and environment state data distributed according to a time domain, the environment state data comprises state data of a plurality of environment state dimensions, and the environment state dimensions comprise an environment drying parameter dimension and an environment dust parameter dimension;
loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition;
sending the fan control instruction sequence of each heat dissipation control node to the service host, so that the CPU of the service host controls a cooling fan according to the fan control instruction sequence of each heat dissipation control node;
the loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition specifically includes:
loading the target heat dissipation state information into a cooling fan control model meeting the optimization and selection deployment requirements of parameter layer information, analyzing the composite temperature situation characteristic distribution of each environment state dimension of the environment state data corresponding to the temperature situation data distributed according to the time domain in the target heat dissipation state information according to the cooling fan control model, wherein the composite temperature situation characteristic is configured to represent the basic control characteristic distribution of weighted aggregation of the temperature situation data distributed according to the time domain according to the environment state characteristic distribution of each environment state dimension of the environment state data;
carrying out fan regulation decision on the composite temperature situation characteristic distribution according to the cooling fan control model, and determining a fan regulation curve of each fan regulation dimension;
performing fan control instance generation on a fan control curve of a fan control dimension according to historical fan control strategy data of a target environment state dimension, determining a plurality of target fan control variables and inheriting fan control instance information of the historical fan control strategy data, wherein the historical fan control strategy data is the control strategy data of the fan control instance of the target environment state dimension, which is learned by the cooling fan control model in a fan control learning process; the cooling fan control model is determined by deep learning according to composite temperature situation feature distribution of a specified environment state dimension and example fan control instance information of a plurality of target fan control variables, the example fan control instance information has a specified correlation with the specified environment state dimension, and fan control instance information covering the example fan control categories is generated according to the fan control instance information obtained from the example fan control instance information of each fan control category;
processing the fan regulation and control example information according to the cooling fan control model, and outputting application fan regulation and control example information which is related to each target fan regulation and control variable and covers the target fan regulation and control category;
and generating a fan control instruction sequence for each heat dissipation control node of the target heat dissipation partition according to the application fan control instance information which is associated with each target fan control variable and covers the target fan control category.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210958164.XA CN115045855B (en) | 2022-08-11 | 2022-08-11 | Cooling fan control method and system based on temperature closed-loop control |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210958164.XA CN115045855B (en) | 2022-08-11 | 2022-08-11 | Cooling fan control method and system based on temperature closed-loop control |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115045855A CN115045855A (en) | 2022-09-13 |
| CN115045855B true CN115045855B (en) | 2022-12-20 |
Family
ID=83167822
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210958164.XA Active CN115045855B (en) | 2022-08-11 | 2022-08-11 | Cooling fan control method and system based on temperature closed-loop control |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115045855B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004185254A (en) * | 2002-12-03 | 2004-07-02 | Hitachi Ltd | Cooling system |
| CN106321218A (en) * | 2015-06-15 | 2017-01-11 | 徐工集团工程机械股份有限公司 | Cooling control system and method and excavator |
| CN110062561A (en) * | 2019-03-25 | 2019-07-26 | 深圳市商宇电子科技有限公司 | A kind of UPS heat dissipation regulating system |
| CN111949101A (en) * | 2020-09-11 | 2020-11-17 | 苏州浪潮智能科技有限公司 | Server fan rotating speed regulating method, system, terminal and storage medium |
| CN113126403A (en) * | 2021-04-19 | 2021-07-16 | 深圳市和天创科技有限公司 | Heat dissipation control method and system for projector |
| CN114326989A (en) * | 2021-11-29 | 2022-04-12 | 苏州浪潮智能科技有限公司 | Server cooling fan control method and device, server and storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10653044B2 (en) * | 2013-01-10 | 2020-05-12 | International Business Machines Corporation | Energy efficiency based control for a cooling system |
-
2022
- 2022-08-11 CN CN202210958164.XA patent/CN115045855B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004185254A (en) * | 2002-12-03 | 2004-07-02 | Hitachi Ltd | Cooling system |
| CN106321218A (en) * | 2015-06-15 | 2017-01-11 | 徐工集团工程机械股份有限公司 | Cooling control system and method and excavator |
| CN110062561A (en) * | 2019-03-25 | 2019-07-26 | 深圳市商宇电子科技有限公司 | A kind of UPS heat dissipation regulating system |
| CN111949101A (en) * | 2020-09-11 | 2020-11-17 | 苏州浪潮智能科技有限公司 | Server fan rotating speed regulating method, system, terminal and storage medium |
| CN113126403A (en) * | 2021-04-19 | 2021-07-16 | 深圳市和天创科技有限公司 | Heat dissipation control method and system for projector |
| CN114326989A (en) * | 2021-11-29 | 2022-04-12 | 苏州浪潮智能科技有限公司 | Server cooling fan control method and device, server and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115045855A (en) | 2022-09-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113534936B (en) | Server fan rotating speed control method, device, equipment and medium | |
| CN111096094B (en) | Cooling unit energy optimization via intelligent supply air temperature setpoint control | |
| US8185893B2 (en) | Starting up at least one virtual machine in a physical machine by a load balancer | |
| US11334671B2 (en) | Adding adversarial robustness to trained machine learning models | |
| EP2793229B1 (en) | Noise cancellation for voice activation | |
| EP3462705B1 (en) | Provisioning of software applications on edge devices in an internet-of-things environment | |
| CN110673951A (en) | Mimicry scheduling method, system and medium for general operation environment | |
| CN105371420A (en) | Refrigeration control method, device and system | |
| CN114065994A (en) | Energy consumption optimization method, device and equipment for air conditioning system and computer storage medium | |
| US11275429B2 (en) | Reducing power consumption of a data center utilizing reinforcement learning framework | |
| CN114330099A (en) | Network card power consumption adjusting method, device, equipment and readable storage medium | |
| CN115045855B (en) | Cooling fan control method and system based on temperature closed-loop control | |
| Song et al. | A deep reinforcement learning-based task scheduling algorithm for energy efficiency in data centers | |
| Suliman et al. | Shark smell optimization (SSO) algorithm for cloud jobs scheduling | |
| Lima et al. | Fast learning set theoretic estimation | |
| CN111488208A (en) | Edge cloud cooperative computing node scheduling optimization method based on variable step length bat algorithm | |
| US20220308529A1 (en) | Extremum Seeking Control with Stochastic Gradient Estimation | |
| CN114254545A (en) | Parallel connection refrigerator system load control optimization method, system, equipment and medium | |
| CN114117778A (en) | Control parameter determination method and device, electronic equipment and storage medium | |
| CN113806077A (en) | Data center server regulation and control method and device based on artificial intelligence | |
| Li et al. | Intelligent trainer for model-based reinforcement learning | |
| Belhaj et al. | Self-adaptive decision making for the management of component-based applications | |
| CN115355603B (en) | Optimal control method and device for refrigeration station, electronic equipment and storage medium | |
| CN116345451B (en) | Operation control method and device for variable frequency temperature control load and terminal equipment | |
| US20230350722A1 (en) | Apparatuses and methods for determining an interdependency between resources of a computing system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |