Fan strategy adjusting method based on optimization algorithm

Abstract

The invention belongs to the technical field of server fan speed regulation, and particularly relates to a fan strategy regulation method based on an optimization algorithm, a cooling mixed strategy model is created, one or more fans and a plurality of hot spots are arranged in a system, and each fan is responsible for cooling a group of hot spots; constructing a fan speed regulation hybrid strategy matrix, wherein each row in the fan speed regulation strategy matrix corresponds to one strategy distribution; and simulating a scene in which a plurality of fans cover a plurality of hot spots based on an optimal algorithm, solving an optimal solution of the heat dissipation mixed strategy model in a system environment, and realizing heat dissipation strategy deployment based on the optimal solution. According to the method, a model with mutual influence of heat dissipation effects is established, the problem of heat dissipation energy efficiency expectation maximization under limited power consumption is converted into the problem of solving an optimal solution in an optimization model, and the optimal strategy and the maximum energy efficiency of a multi-fan heat dissipation strategy are solved through a mixed integer linear programming formula. And the fan in the equipment can work with the optimal strategy and the maximum energy efficiency.

Description

Technical Field

[0001] The present invention belongs to the technical field of server fan speed regulation, and particularly relates to a fan strategy adjustment method based on an optimization algorithm. Background Art

[0002] The development of electronic technology enables the control of fans to be achieved through electronic circuits and provides more precise speed regulation control. The wide application of microcontrollers provides an intelligent control core for fan speed regulation. Sensing technologies such as temperature sensors and rotational speed sensors can monitor environmental parameters in real time and provide a basis for speed regulation. With the improvement of energy awareness, achieving energy conservation through speed regulation has become an important focus. In electronic devices, good heat dissipation is crucial for performance and lifespan. Therefore, optimizing the fan speed regulation of servers can optimize the heat dissipation effect and extend the service life of servers. Different rotational speeds meet users' requirements for air volume and noise, enhancing the user experience of servers. Automatic speed regulation of fans is achieved through intelligent control and dynamically adjusted according to actual situations. Environmental protection requirements reduce energy consumption through speed regulation and minimize the impact on the environment. The need for system integration makes fan speed regulation need to work in coordination with other system components.

[0003] The purpose of an intelligent fan speed regulation strategy is to minimize fan noise and energy consumption while meeting the heat dissipation requirements. The following are some common intelligent fan speed regulation strategies: Temperature-sensing speed regulation: The environmental or device temperature is monitored through a temperature sensor, and the fan rotational speed is automatically adjusted according to temperature changes. When the temperature rises, the fan rotational speed increases to provide more heat dissipation; when the temperature drops, the fan rotational speed decreases to reduce noise and energy consumption. However, when measuring temperature, the temperature sensor may be affected by interference factors such as line resistance and external electromagnetic fields, resulting in relatively large measurement errors. The response time of the temperature sensor is relatively long and sometimes cannot meet the requirements of actual applications. Sensitive to environmental interference, the temperature sensor may be affected by factors such as environmental temperature changes, humidity, and vibration, resulting in unstable measurement signals.

[0004] Pulse-width modulation (PWM) speed regulation: PWM speed regulation controls the fan rotational speed by changing the duty cycle of the fan power supply. The higher the duty cycle, the faster the fan rotational speed; the lower the duty cycle, the slower the fan rotational speed. PWM speed regulation can achieve fine rotational speed control and is relatively efficient. The design and implementation of a PWM speed regulation system are relatively complex. Precise time control and filter design are required to process the PWM signal to reduce the ripple and electromagnetic interference in the output. This increases the complexity and maintenance difficulty of the system. At the same time, special control circuits and high-speed switches are required, resulting in a relatively high hardware cost. The high-frequency pulse signals generated by PWM technology are prone to electromagnetic interference and noise. Especially in low-frequency PWM signals, audio noise may be generated, affecting the normal operation of electronic devices.

[0005] Fuzzy control speed regulation: Based on the fuzzy logic algorithm, the fan speed is adjusted according to factors such as temperature, humidity, and human flow. Fuzzy control speed regulation can better adapt to complex environmental conditions, but requires more sensors and computational power consumption.

[0006] The above several fan adjustment strategies are generally "single-temperature single-control" and "multiple-fan single-control"; that is, the fan is only adjusted according to a single hot spot, or multiple fans are regulated by the same strategy. The strategy allocation is regulated according to temperature, and the adjustment factor is single, without considering the mutual influence between adjacent fans.

[0007] Therefore, the present invention considers the influence of the heat dissipation effects of multiple cooling fans in the same device on each other, and accordingly establishes a model that maximizes the heat dissipation energy efficiency under a rated power consumption. Since this model takes into account the mutual influence between the fan heat dissipation strategies, a hybrid strategy will be used to represent and calculate the optimal heat dissipation strategy for each fan. Summary of the Invention

[0008] The purpose of the present invention is to provide a fan strategy adjustment method based on an optimization algorithm, which is used to consider the influence of the heat dissipation effects of multiple cooling fans in the same device on each other, and accordingly establishes a model that maximizes the heat dissipation energy efficiency under a rated power consumption. Since this model takes into account the mutual influence between the fan heat dissipation strategies, a hybrid strategy will be used to represent and calculate the optimal heat dissipation strategy for each fan.

[0009] To solve the above technical problems, the technical solutions adopted by the present invention are as follows: A fan strategy adjustment method based on an optimization algorithm includes the following steps: S1: Create a heat dissipation hybrid strategy model. There is one or more fans and several hot spots in the system, and each fan i is responsible for dissipating heat from a group of hot spots; S2: Construct a fan speed regulation hybrid strategy matrix, and each row in the fan speed regulation strategy matrix corresponds to a strategy distribution; S3: Based on the optimal algorithm, simulate the scenario where multiple fans cover multiple hot spots, solve the optimal solution of the heat dissipation hybrid strategy model in the system environment, and implement the heat dissipation strategy deployment based on the optimal solution.

[0010] Preferably, the specific process in step S1 is as follows: S11: Create a heat dissipation hybrid strategy model. There is one or more fans I in the heat dissipation hybrid strategy model, I = {1, 2, 3, ···, n} and several hot spots; S12: Each fan i is responsible for dissipating heat from a group of hot spots T i , , and , that is, the number of hotspots covered by each fan ≥ 1.

[0011] Preferably, the specific process of constructing the fan speed regulation hybrid strategy matrix in step S2 is as follows: S21: Let the policy distribution set corresponding heat dissipation policies for each hotspot it dissipates heat for, and use the probability to represent the heat dissipation policy of the fan i for the hotspot. S22: The policies of all hotspots covered by the fan i are represented by a hybrid strategy matrix Q , and , and , use v j to represent the temperature reduction value of the hotspot j . Among them, .

[0012] Preferably, the correlation degree between fans is introduced into the fan speed regulation hybrid strategy matrix in step S2, and the specific process is as follows: Let fan A cover a specified number of hotspots. Due to the proximity of hotspots or the coverage of the air duct, other hotspots will be affected by the heat dissipation effect of fan A, and to a certain extent, the heat dissipation effect of fan A is expanded; allocate heat dissipation power consumption to fan A. After improving the heat dissipation energy efficiency of fan A, indirectly increase the heat dissipation energy efficiency of other hotspots associated with this hotspot, and the indirect influence of indirectly increasing the heat dissipation energy efficiency of other hotspots associated with this hotspot is transmitted by the correlation degree , and use to represent the correlation degree between the current hotspot j and the associated hotspot .

[0013] Preferably, the heat dissipation power consumption c is introduced into the fan speed regulation hybrid strategy matrix in step S2, and the specific process is as follows: Calculate the power consumption of the fan through , and the total heat dissipation power consumption is C , where the power consumption required for the fan i to cover the hotspot is C i , , and the total heat dissipation cost C also satisfies: .

[0014] Preferably, the effective coefficient s and utility are introduced into the fan speed regulation hybrid strategy matrix in step S2, and the specific process is as follows: Use the parameters j to reflect the fan on the hot spot j effectiveness, to reflect the fan on the hot spot j effectiveness, will s j be expressed as the covered hot spot j effective coefficient, when fan A only covers the hot spot j at that time, s j the value is 1; s j the value is defined according to the actual effectiveness of the fan in dissipating heat from the hot spot, s j when the value is 0, it means that fan A is completely ineffective in dissipating heat from this hot spot; for the hot spot j dissipating heat will cause power consumption C j , let fan A only dissipate heat from the hot spot , the resulting utility is denoted as U i,j , using v j to represent the hot spot j temperature reduction value; let the heat dissipation efficiency only depend on the hot spot j temperature reduction value v j and the heat dissipation strategy for this hot spot, the fan i for the hot spot j dissipating heat will consume heat dissipation power c i,j , the hot spot j due to being close in distance to the hot spot or sharing the same air duct, a free-riding effect is generated, and the magnitude of the free-riding effect is determined by the correlation degree ; now the fan dissipates heat from the hot spot , based on the free-riding effect, the total utility value of the fan at this time will be expressed by the following formula: .

[0015] Preferably, the specific process of step S3 is as follows: S31: Express the total available heat dissipation power of the heat dissipation hot spot matrix T as C , and give an initial heat dissipation strategy matrix that satisfies a uniform distribution, and the size of the initial heat dissipation strategy is ; S32: According to the initial heat dissipation strategy matrix , for the hot spot jAllocate the heat dissipation power consumption of the fan. The specific formula is as follows: ; By j allocating the heat dissipation power consumption formula of the fan to the hot spots, the initial heat dissipation power consumption matrix is obtained ; S33: After establishing the initial heat dissipation power consumption matrix of all hot spots , the initial heat dissipation energy efficiency matrix is calculated by the formula ; ; S34: Set the effective coefficient matrix S j , based on the initial heat dissipation power consumption matrix , the initial heat dissipation energy efficiency matrix , and the effective coefficient matrix S j , calculate the corresponding new heat dissipation strategy for each hot spot q j . q j The specific calculation formula is as follows: ; Among them, ; S35: After obtaining the new heat dissipation strategy matrix Q , the new heat dissipation power consumption matrix is obtained through c , and the new heat dissipation energy efficiency matrix is calculated by U , and the results Q , U are output.

[0016] The beneficial effects of the present invention include: The fan strategy adjustment method based on the optimization algorithm provided by the present invention creates a heat dissipation hybrid strategy model. There is one or more fans and several hot spots in the system, and each fan is responsible for dissipating heat for a group of hot spots; a fan speed regulation hybrid strategy matrix is constructed, and each row in the fan speed regulation strategy matrix corresponds to a strategy distribution; based on the optimal algorithm, the scenario of multiple fans covering multiple hot spots is simulated, and the optimal solution of the heat dissipation hybrid strategy model is solved in the system environment, and the heat dissipation strategy deployment is realized based on the optimal solution. By establishing a model in which the heat dissipation effects interact with each other, the problem of maximizing the heat dissipation energy efficiency expectation under limited power consumption is transformed into the problem of finding the optimal solution in the optimization model. Through the mixed integer linear programming formula, the optimal strategy and the maximum energy efficiency of the multi-fan heat dissipation strategy are obtained, so that the fans in the device can work with the optimal strategy and the maximum energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1Schematic diagram of the heat dissipation strategy of the present invention.

[0018] Figure 2 Schematic diagram of the parameter values of the present invention.

[0019] Figure 3 Schematic diagram of the hot spot value, correlation relationship and initial strategy distribution of the present invention.

[0020] Figure 4 Schematic diagram of the growth rate of heat dissipation energy efficiency of the present invention.

[0021] Figure 5 Schematic diagram of the comparison of the total heat dissipation energy efficiency with or without considering the correlation relationship under different heat dissipation powers of the present invention. Detailed implementation mode

[0022] The following combines the attached Figures 1 to 5 The present invention will be further described in detail: Embodiment 1 A method for adjusting the fan strategy based on an optimization algorithm includes the following steps: S1: Create a heat dissipation hybrid strategy model. There is one or more fans and several hot spots in the system, and each fan i is responsible for dissipating heat from a group of hot spots; S2: Construct a fan speed regulation hybrid strategy matrix, and each row in the fan speed regulation strategy matrix corresponds to a strategy distribution; S3: Based on the optimal algorithm, simulate the scenario where multiple fans cover multiple hot spots, solve the optimal solution of the heat dissipation hybrid strategy model in the system environment, and deploy the heat dissipation strategy based on the optimal solution.

[0023] In this embodiment, the optimization algorithm is applied to the research of the fan heat dissipation speed regulation strategy, a heat dissipation model with multiple fans and multiple hot spots is established, and considering factors such as the temperature values of multiple hot spots in the system, the mutual influence of the fan heat dissipation effect, and the fan power consumption, with the goal of maximizing the heat dissipation energy efficiency, a corresponding algorithm is proposed to calculate the optimal heat dissipation strategy to ensure the maximization of the overall heat dissipation energy efficiency of the system. When only one fan runs in the system, it will have different heat dissipation effects on different hot spots; similarly, when multiple fans run simultaneously in the system, they will also affect each other, and this phenomenon is called the "proximity effect".

[0024] The established heat dissipation hybrid strategy model. The basic heat dissipation strategy is also regulated by negative feedback according to the hot spot temperature value. However, on this basis, the influence of the "proximity effect" on fan heat dissipation is additionally considered. Since the heat dissipation effect of the fan will be promoted or offset by each other due to factors such as duct design, this positive superposition of heat dissipation effect is similar to "hitchhiking", and the negative mutual cancellation will reduce the actual energy efficiency of the fan. A reasonable heat dissipation strategy for hot spots will be established according to the energy efficiency of fan heat dissipation under a given total power consumption. Ensure that the overall heat dissipation energy efficiency of the system reaches the optimal or maximum. Based on this, this problem is transformed into the calculation problem of the optimal solution in the optimization model.

[0025] Embodiment 2 On the basis of Embodiment 1, the specific process in step S1 is as follows: S11: Create a heat dissipation hybrid strategy model, in which there are one or more fans I, I = {1, 2, 3, ···, n} and several hot spots in the heat dissipation hybrid strategy model; S12: Each fan i is responsible for dissipating heat from a group of hot spots T i , , and , that is, the number of hot spots covered by each fan ≥ 1.

[0026] In this embodiment, the specific process of constructing the fan speed regulation hybrid strategy matrix in step S2 is as follows: S21: Let the strategy distribution set corresponding heat dissipation strategies for each hot spot it dissipates heat from, and use the probability to represent the heat dissipation strategy of fan i for the hot spot; S22: The strategies of fan i for all the hot spots it covers are represented by a hybrid strategy matrix Q , , and , and use v j to represent the temperature reduction value of hot spot j ; Among them, .

[0027] In step S2, the correlation degree between fans is introduced into the fan speed regulation hybrid strategy matrix, and the specific process is as follows: Let fan A cover a specified number of hotspots. Due to the proximity of the hotspots or the coverage of the air duct, other hotspots will be affected by the heat dissipation effect of fan A, which to a certain extent expands the heat dissipation effect of fan A; allocate the heat dissipation power consumption to fan A. After improving the heat dissipation energy efficiency of fan A, indirectly increase the heat dissipation energy efficiency of other hotspots associated with this hotspot, and the indirect influence of indirectly increasing the heat dissipation energy efficiency of other hotspots associated with this hotspot is transmitted by the correlation degree to be passed, and use to represent the current hotspot j and the associated hotspot The correlation degree between them.

[0028] In step S2, the heat dissipation power consumption is introduced into the fan speed regulation hybrid strategy matrix c , and the specific process is as follows: Calculate the power consumption of the fan through , and the total heat dissipation power consumption is C , where the fan i covering the hotspot needs to consume power of C i , , and the total heat dissipation cost C also satisfies: .

[0029] In step S2, the effective coefficient s and utility are introduced into the fan speed regulation hybrid strategy matrix, and the specific process is as follows: Use the parameter s j to reflect the effectiveness of the fan on the hotspot j , s j ∈[0, 1], and s j is expressed as the effective coefficient of the covered hotspot j . When fan A only covers the hotspot j , s j The value of is 1; s j The value of is defined according to the actual effective degree of the fan's heat dissipation on the hotspot, s j When the value is 0, it means that the heat dissipation of fan A on this hotspot is completely ineffective; The heat dissipation power consumption for the hotspot j will cause power consumption C j . When fan A only dissipates heat from the hotspot j ( j ∈T), the generated utility is denoted as U i,j , and use v jIndicates a hot spot j Temperature reduction value; Make the heat dissipation energy efficiency only depend on the hot spot j Of the temperature reduction value v j And the heat dissipation strategy for this hot spot, the fan i For the hot spot j Performing heat dissipation on the hot spot will consume heat dissipation power c i,j , the hot spot j Due to being close to the hot spot Or sharing the same air duct, a free-riding effect is generated, and the magnitude of the free-riding effect is determined by the correlation degree Determined. Now the fan dissipates heat from the hot spot j Based on the free-riding effect, the total utility value of the fan at this time will be expressed by the following formula: .

[0030] See Figure 1 , the corresponding mixed strategy matrix q Is: .

[0031] In the strategy matrix, the m Row represents the fan m , the n Column represents the n th hot spot covered by the fan. Therefore, Figure 1 The corresponding strategy matrix represents that the first one is only responsible for dissipating heat from hot spots 3 and 7, and the heat dissipation strategies are 0.5 and 0.5 respectively; similarly, the second row represents that the second fan covers hot spots 1, 4, and 8, and the heat dissipation strategies are 0.333333, 0.333333, and 0.333333 respectively; the third row represents that the third fan covers hot spots 2, 5, and 6, and the heat dissipation strategies are 0.333333, 0.333333, and 0.333333 respectively.

[0032] See Figure 2 , there are 2 hot spots. First, assume that there are two fans each dissipating heat from one hot spot separately. The temperature value of each hot spot is 1 (normalized), the correlation degree between them is 0.5, and the heat dissipation power is 0.5; assume that the temperature of hot spot 1 drops from 1 to 0, and its effective coefficient is 0.5; considering the "free-riding" effect, from It can be obtained that the heat dissipation energy efficiencies of hot spot 1 and hot spot 2 are 1.75 and 2.25 respectively.

[0033] Embodiment 3 On the basis of Embodiment 1 or Embodiment 2, the specific process of step S3 is as follows: S31: Represent the total available heat dissipation power of the heat dissipation hot spot matrix T as C , and give the initial heat dissipation strategy matrix that satisfies the uniform distribution , where the size of the initial heat dissipation strategy is ; S32: According to the initial heat dissipation strategy matrix , allocate the heat dissipation power of the fan to the hot spot j . The specific formula is as follows: ; By using the formula for allocating the heat dissipation power of the fan to the hot spot j , obtain the initial heat dissipation power matrix ; S33: After establishing the initial heat dissipation power matrix C of all hot spots * , calculate the initial heat dissipation energy efficiency matrix from the formula ; S34: Set the effective coefficient matrix S j . Based on the initial heat dissipation power matrix , the initial heat dissipation energy efficiency matrix , and the effective coefficient matrix S j , calculate the corresponding new heat dissipation strategy q j for each hot spot. q j The specific calculation formula of is as follows: Among them, ; S35: After obtaining the new heat dissipation strategy matrix Q , obtain the new heat dissipation power matrix through c , and calculate the new heat dissipation energy efficiency matrix from U , and output the results Q , U .

[0034] To intuitively describe how the above algorithm implements heat dissipation strategy deployment in the system environment, the following example is given: Assume that there are 8 hot spots in a certain server. Let the temperature reduction value of each hot spot be v , and the correlation degree is represented by r . The initial heat dissipation strategy is represented by q ; Simplify the following topological structure Figure 3 according to the system relationship between computers, and represent the influence between any two fans by the correlation degree r . The total heat dissipation power is C, the currently defined total heat dissipation power consumption C Under this condition, through the optimization algorithm, a reasonable allocation strategy is given to maximize the utility of heat dissipation.

[0035] For the convenience of calculation, we set the temperature reduction value of each hot spot to 1, and the correlation degree r between hot spots is unified to 0.5, and the given total heat dissipation cost C = 4.

[0036] Initial heat dissipation strategy matrix: Q * = [0.125, 0.125, 0.125, 0.125, 0.125, 0.125, 0.125, 0.125]; Available total heat dissipation power consumption C , through the formula, the initial heat dissipation power consumption matrix is obtained: C * = [0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5]; According to the formula, the heat dissipation energy efficiency matrix of all hot spots is obtained: U ini = [3.75, 3.0, 3.0, 4.5, 3.75, 3.0, 2.25, 2.25]; According to the hot spot reduction value, the effective coefficient matrix is set: ; Substitute the above known conditions into , , and the heat dissipation strategy matrix after iteration is obtained as: ; After obtaining the heat dissipation strategy matrix, we update the heat dissipation power consumption matrix based on it: ; Finally, according to the formula, the heat dissipation energy efficiency matrix is updated: .

[0037] To verify the effectiveness of the above optimal algorithm, the same total heat dissipation power consumption C is given, and the calculation results of the optimal algorithm are compared with the results obtained without considering the correlation relationship. When making a decision without considering the correlation relationship between hot spots, the corresponding heat dissipation power is often arranged according to the temperature reduction value of the hot spot; the basic relationship is satisfied: ; Furthermore, the heat dissipation power consumption without considering the correlation relationship can be obtained. c The matrix of ; When making decisions, the correlation relationship is not considered, but when calculating the heat dissipation energy efficiency, it needs to be calculated under actual conditions (considering the correlation relationship). Therefore, the heat dissipation power consumption C is substituted into to obtain the heat dissipation energy efficiency matrix .

[0038] Then, the results with and without considering the correlation relationship under different power consumptions are compared. The results are as shown in Figure 5 . It can be seen from Figure 4 and Figure 5 that the decisions considering the mutual influence of heat dissipation often produce higher utility than those without considering the mutual influence relationship. And as the energy consumption increases, provided that the total energy consumption does not exceed the upper limit of the fan energy efficiency, the growth rate of the heat dissipation utility also becomes larger and larger. This shows that when the given power consumption is larger, the mutual influence between the fans should be considered more. This algorithm can help fully consider the mutual influence between the fan heat dissipations when regulating the fan strategy, arrange a reasonable distribution method for it, and ensure the optimal heat dissipation energy efficiency under the rated power consumption.

[0039] In summary, the fan strategy adjustment method based on the optimization algorithm provided by the present invention creates a heat dissipation hybrid strategy model. There is one or more fans and several hot spots in the system, and each fan is responsible for dissipating heat for a group of hot spots; constructs a fan speed regulation hybrid strategy matrix, and each row in the fan speed regulation strategy matrix corresponds to a strategy distribution; simulates the scenario of multiple fans covering multiple hot spots based on the optimal algorithm, solves the optimal solution of the heat dissipation hybrid strategy model in the system environment, and realizes the heat dissipation strategy deployment based on the optimal solution. By establishing a model of the mutual influence of heat dissipation effects, the problem of maximizing the expected heat dissipation energy efficiency under limited power consumption is transformed into the problem of finding the optimal solution in the optimization model. Through the mixed integer linear programming formula, the optimal strategy and the maximum energy efficiency of the multi-fan heat dissipation strategy are obtained, so that the fans in the device can work with the optimal strategy and the maximum energy efficiency.

Claims

1. A method for adjusting a fan strategy based on an optimization algorithm, characterized in that Including the following steps: S1: Create a heat dissipation hybrid strategy model. There is one or more fans and several hotspots in the system, and each fan is responsible for dissipating heat from a group of hotspots; S2: Construct a fan speed regulation hybrid strategy matrix. Each row in the fan speed regulation strategy matrix corresponds to a strategy distribution; S3: Based on an optimal algorithm, simulate the scenario of multiple fans covering multiple hotspots, solve the optimal solution of the heat dissipation hybrid strategy model in the system environment, and implement the heat dissipation strategy deployment based on the optimal solution.

2. The method for adjusting the fan strategy based on the optimization algorithm according to claim 1, wherein The specific process in step S1 is as follows: S11: Create a heat dissipation hybrid strategy model. There is one or more fan Is in the heat dissipation hybrid strategy model, I = {1, 2, 3, ···, n} and several hotspots; S12: Each fan i is responsible for cooling a group of hot spots T i , , and , that is, the number of hot spots covered by each fan ≥ 1.

3. A method for adjusting a fan strategy based on an optimization algorithm according to claim 1, characterized in that, The specific process of constructing the fan speed regulation hybrid strategy matrix in step S2 is as follows: S21: Set corresponding heat dissipation strategies for each hot spot that dissipates heat with the policy distribution, and use probability to represent the heat dissipation strategy of the fan i for the hot spot; S22: Fan i The policies for all hotspots it covers are represented by a hybrid policy matrix Q to represent, and , using v j to represent the hotspot j temperature reduction value; Among them, 。 4. A method for adjusting a fan strategy based on an optimization algorithm according to claim 3, characterized in that, In step S2, the fan speed regulation hybrid strategy matrix introduces the correlation degree between fans , and the specific process is as follows: Let fan A cover a specified number of hotspots. Due to the proximity of the hotspots or the coverage of the air duct, other hotspots will be affected by the heat dissipation effect of fan A, to a certain extent expanding the heat dissipation effect of fan A; allocate the heat dissipation power consumption to fan A. After improving the heat dissipation energy efficiency of fan A, indirectly increase the heat dissipation energy efficiency of other hotspots associated with this hotspot, and use the indirect influence of indirectly increasing the heat dissipation energy efficiency of other hotspots associated with this hotspot to be transmitted by the correlation degree to transfer, and use to represent the current hotspot j and the associated hotspot the correlation degree between them.

5. The method for adjusting a fan strategy based on an optimization algorithm according to claim 3, wherein In step S2, the heat dissipation power consumption is introduced into the fan speed regulation hybrid strategy matrix c , and the specific process is as follows: The power consumption of the fan is calculated by The total dissipated heat power consumption is C Among them, the fan i needs to consume power to cover the hot spot, which is C i , And the total heat dissipation cost C also satisfies: .

6. The method for adjusting a fan strategy based on an optimization algorithm according to claim 3, wherein In step S2, an effective coefficient and utility are introduced into the fan speed regulation hybrid strategy matrix, and the specific process is as follows: s ​ Using parameters s j to reflect the effectiveness of the fan on the hot spot j effectiveness, to reflect the effectiveness of the fan on the hot spot j effectiveness, will s j be expressed as the effective coefficient of the covered hot spot j When fan A only covers the hot spot j at that time, s j the value is 1; s j The value is defined according to the actual effectiveness of the fan in dissipating heat from the hot spot, s j When the value is 0, it means that the heat dissipation of fan A for this hot spot is completely ineffective; For the hot spot j Performing heat dissipation on it will cause power consumption C j Let the fan A only dissipate heat from the hot spot The resulting effect is denoted as U i,j Use v j To represent the temperature reduction value of the hot spot j ; Let the heat dissipation efficiency only depend on the temperature reduction value of the hot spot j and the heat dissipation strategy for this hot spot. The fan v j dissipates heat from the hot spot i and consumes heat dissipation power j in the process. Due to the proximity to the hot spot c i,j or sharing the same air duct, a free-riding effect is generated. The magnitude of the free-riding effect is determined by the correlation j degree; now the fan dissipates heat from the hot spot . Based on the free-riding effect, the total utility value of the fan at this time will be expressed by the following formula:​​ 。 7. A method for adjusting a fan strategy based on an optimization algorithm according to claim 6, characterized in that, The specific process of step S3 is as follows: S31: Represent the total available cooling power consumption of the cooling hot spot matrix T as C , and give an initial cooling strategy matrix that satisfies the uniform distribution , where the size of the initial cooling strategy is ; S32: According to the initial heat dissipation strategy matrix , allocate the heat dissipation power consumption of the fan for the hot spot j , and the specific formula is as follows: ； By allocating the heat dissipation power consumption formula of the fan j to obtain the initial heat dissipation power consumption matrix ; S33: Establish the initial heat dissipation power consumption matrix for all hotspots After that, from the formula, the initial heat dissipation energy efficiency matrix is calculated as ; S34: Set the effective coefficient matrix S j , based on the initial heat dissipation power consumption matrix , and the initial heat dissipation energy efficiency matrix , the effective coefficient matrix S j , calculate the corresponding new heat dissipation strategy for each hot spot q j , q j The specific calculation formula of is as follows: ； Among them, ; S35: Obtain a new heat dissipation strategy matrix Q After that, through obtain a new heat dissipation power consumption matrix c , and from calculate to obtain a new heat dissipation energy efficiency matrix U , and output the results Q , U .

Images