A composite heat dissipation fin for a plate-fin air-cooled radiator and its optimized design method

Abstract

This invention belongs to the field of radiator technology, specifically a composite heat dissipation fin for a plate-fin air-cooled radiator and its optimized design method. It includes heat dissipation fins mounted on a radiator and a fan mounted on the radiator. A vortex copper tube is installed through the interior of the heat dissipation fins in a planar or three-dimensional manner. A turbulence-inducing mechanism that agitates the direct current airflow is also fixedly installed on the surface of the heat dissipation fins. This composite heat dissipation fin for a plate-fin air-cooled radiator and its optimized design method, by setting the interior of the heat dissipation fins to have a vortex copper tube installed through in a planar or three-dimensional manner, ensures that the shape of the copper tube for heat dissipation is consistent with the shape of the fan. The three-dimensional shape is more conducive to heat dissipation and makes full use of space, enabling intelligent heat dissipation. Simultaneously, it utilizes existing data for intelligent neural network processing to optimize various parameters of the radiator, thereby retrieving heat dissipation solutions for different environments.

Description

Technical Field

[0001] This invention relates to the field of radiator technology, and in particular to a composite heat dissipation fin for a plate-fin air-cooled radiator and its optimized design method. Background Technology

[0002] In existing technologies, such as Figure 1 As shown, most of them involve loading plate-fin fins onto the surface of copper pipes, then installing a fan to form a plate-fin air cooler. However, this type of cooler has the following drawbacks:

[0003] 1. The air generated by the fan is in the shape of the fan blades, which are circular. However, the existing copper heat pipes are arranged in a straight line on the heat sink, which makes it difficult for the copper pipes outside the circle to be cooled by the fan.

[0004] 2. In a plate-fin type of heat dissipation, the airflow direction will be as follows: Figure 1 As shown, due to the DC cooling method, the back of the copper pipe is either difficult to cool or heat will accumulate there, ultimately causing uneven heat dissipation of the copper pipe and increasing the difficulty of optimization. Summary of the Invention

[0005] Based on existing technical problems, this invention proposes a composite heat dissipation fin for a plate-fin air-cooled radiator and its optimized design method.

[0006] The present invention proposes a plate-fin type air-cooled radiator with composite heat dissipation fins, including heat dissipation fins mounted on the radiator and a fan mounted on the radiator.

[0007] The heat dissipation fins are internally fitted with vortex copper tubes that penetrate in a planar or three-dimensional manner.

[0008] The surface of the heat dissipation fins is also fixedly equipped with a turbulence mechanism that disturbs the DC heat dissipation airflow to achieve composite heat dissipation.

[0009] Preferably, the center of the vortex copper tube is located on the same axis as the fan, and is mounted on the heat dissipation fins in a planar manner or in a conical shape.

[0010] Preferably, the vortex copper tube is composed of two semi-circular copper tubes joined together, with the outer ends of the two semi-circular copper tubes respectively serving as the inlet and outlet of the cooling body to be cooled.

[0011] The above technical solution increases the heat dissipation path and improves the heat dissipation effect through the two semi-circular copper pipes.

[0012] Preferably, a heat insulation layer is provided at the connection between the two copper tubes. The heat insulation layer is composed of a graphite polystyrene board with a cavity, and the cavity in the heat insulation layer is filled with an inert gas.

[0013] Through the above technical solution, the insulation layer can insulate the semi-circular copper tube and prevent heat radiation from affecting the heat dissipation effect.

[0014] Preferably, the turbulence mechanism includes multiple turbulence grooves formed on the side of the heat dissipation fins. An elastic steel sheet that always pops outward is fixedly installed at one end of the inner wall of the turbulence groove. One end of the elastic steel sheet is movably hinged to the inner wall of the turbulence groove. When one end of the elastic steel sheet slides closer to the other end, the middle surface of the elastic steel sheet pops outward to achieve the action of turbulence on the airflow.

[0015] Through the above technical solution, the elastic steel sheet can achieve the effect of automatic reset, as well as corrosion resistance and high temperature resistance.

[0016] Preferably, a through hole extending to the top or bottom of the heat dissipation fins is provided at the center line of the length direction of the turbulence groove, and a steel wire rope is movably installed inside the through hole, with one end of the steel wire rope fixed to the movable end of the elastic steel sheet.

[0017] Miniature cylinders are fixedly mounted in an array on the upper or lower surface of the heat dissipation fins. The piston rod of the miniature cylinder is fixedly connected to the other end of the steel wire rope to realize the action of pulling the elastic steel sheet.

[0018] Through the above technical solution, each cylinder controls each elastic steel sheet individually, achieving a dual control effect of independent control and linkage control.

[0019] An optimized design method for composite heat dissipation fins in a plate-fin air-cooled radiator includes step one: data acquisition. A temperature sensor is installed on the back of the vortex copper tube in a vortex array, and the temperature inside the vortex copper tube is acquired in real time through the temperature sensor.

[0020] Step 2: Group the data collected by the temperature sensor, the control signal data of the miniature cylinder, and the fan speed data to form a mapping relationship. .

[0021] Step 3: Based on the mapping relationship For each of the corresponding parameters, set a threshold for these parameters. The upper and lower limits of each parameter are obtained. Then, the dataset composed of these parameters is imported into the neural network model for training, verification and parameter tuning. Finally, the temperature data of the temperature sensor, the control signal data of the micro cylinder and the fan speed data are derived from the trained parameters.

[0022] Step 4: Import the data from the temperature sensor, miniature cylinder, and fan speed in Step 3 into the database and reverse map out the respective models of the temperature sensor, miniature cylinder, and fan.

[0023] Step 5: Input the model or parameter data of the temperature sensor, miniature cylinder and fan after reverse mapping into Step 3 to simulate and verify whether they exceed the upper and lower limits of the model or parameter data.

[0024] If it exceeds the limit, reduce the adjustment threshold by one level. Then proceed with steps three through five.

[0025] If the parameters are not exceeded, then export the models or parameters of the temperature sensor, miniature cylinder, and fan.

[0026] Step Six: End.

[0027] Preferably, the temperature sensors in step one are grouped and then matched one-to-one with the miniature cylinders. The grouping method is based on the temperature sensor's sensing radius. As a standard, if one of the micro-cylinders controls the elastic steel sheet to simultaneously fall into the temperature sensing radius of two adjacent temperature sensors... At this time, the miniature cylinder moves into the temperature sensor group that is closest to the outlet end.

[0028] The above technical solution allows the turbulence effect of the elastic steel sheet to be directly displayed through data.

[0029] Preferably, in step three, the mapping relationship The dataset is normalized and divided into training and testing sets according to a certain ratio. The model is then trained, validated, and its parameters are tuned to build a neural network model.

[0030] During training, the loss function will calculate the sum of the predicted values ​​of the neural network model and... The errors between the labels of the dataset are backpropagated to update the parameters of the neural network model.

[0031] The neural network model is evaluated using mean absolute error (MAE), mean squared error (MSE), and coefficient of determination. As an evaluation indicator, if the indicator does not meet expectations, the parameter size will be adjusted. When the accuracy of the neural network model reaches the required level, the normalized information of the neural network model during training will be saved into the database.

[0032] The above technical solutions enable the addition of input normalization actions during training when using neural network models for data processing.

[0033] Preferably, in step four, an inversion model of the neural network model is established during the reverse mapping process.

[0034] When the inversion model no longer matches the actual temperature sensor, miniature cylinder, and fan model or parameter data from step five with the model or parameter data already stored in the database, the monitored temperature sensor, miniature cylinder, and fan values ​​are stabilized: the temperature fluctuation variance is calculated for each set of parameters. When temperature fluctuation variance Below the preset threshold When the monitored temperature sensor, micro cylinder, and fan values ​​are in an effective heat dissipation stage, the values ​​of the temperature sensor, micro cylinder, and fan are taken as inputs, and the output is obtained through the inversion model to complete the acquisition of the optimized working parameters of the radiator.

[0035] The above technical solution utilizes temperature fluctuation variance. This is used to stabilize and limit the various models or parameters of temperature sensors, miniature cylinders, and fans, avoiding excessive fluctuations that could cause inversion errors.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. By installing vortex copper pipes through the interior of the heat dissipation fins in a planar or three-dimensional manner, the shape of the copper pipes can be made consistent with the shape of the fan's heat dissipation, and the three-dimensional shape is more conducive to heat dissipation and makes full use of space.

[0038] 2. By setting up a turbulence mechanism, the straight heat dissipation airflow can be turbulent, increasing the heat dissipation area, reducing the direct area of ​​the airflow at the horizontal diameter of the copper pipe, reducing heat dissipation dead angles, especially reducing heat dissipation dead angles on the back of the copper pipe.

[0039] 3. By setting up an optimized design method, intelligent heat dissipation can be achieved. At the same time, existing data can be used for intelligent neural network processing to optimize various parameters of the heat sink and generate heat dissipation solutions for different environments. Attached Figure Description

[0040] Figure 1 This is a diagram of existing technology;

[0041] Figure 2 This is a perspective view of a composite heat dissipation fin of a plate-fin air-cooled radiator proposed in this invention.

[0042] Figure 3 This is a perspective view of the conical installation of the vortex copper tube and the heat dissipation fins of the composite heat dissipation fins of the plate-fin air-cooled radiator proposed in this invention.

[0043] Figure 4This is a three-dimensional view of the planar distribution of the vortex copper tubes in a plate-fin type air-cooled radiator according to the present invention.

[0044] Figure 5 This is a three-dimensional schematic diagram of the vortex copper tube distribution of the composite heat dissipation fins of a plate-fin air-cooled radiator proposed in this invention.

[0045] Figure 6 This is a three-dimensional view of the heat insulation layer of the composite heat dissipation fins of a plate-fin air-cooled radiator proposed in this invention.

[0046] Figure 7 This is a perspective view of the turbulence mechanism of the composite heat dissipation fins of a plate-fin air-cooled radiator proposed in this invention.

[0047] Figure 8 This is a rear view of the temperature sensor and vortex copper tube installed, representing an optimized design method for composite heat dissipation fins in a plate-fin air-cooled radiator proposed in this invention.

[0048] In the diagram: 1. Heat dissipation fins; 2. Fan; 3. Scroll copper tube; 31. Insulation layer; 4. Fluctuation channel; 41. Elastic steel sheet; 42. Steel wire rope; 43. Miniature cylinder; 5. Temperature sensor. Detailed Implementation

[0049] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0050] Example 1

[0051] Reference Figure 2-7 A plate-fin air-cooled radiator with composite heat dissipation fins includes heat dissipation fins 1 mounted on the radiator and a fan 2 mounted on the radiator. The heat dissipation fins 1 can be made of any one of aluminum alloy, copper, or copper-aluminum composite.

[0052] The interior of the heat dissipation fins 1 is equipped with a vortex copper tube 3 that runs through it in a planar or three-dimensional manner.

[0053] To maintain concentricity, the center of the vortex copper tube 3 and the axis of the fan 2 are positioned on the same axis line, and it is mounted on the heat dissipation fins 1 in a planar manner or in a conical shape.

[0054] To increase the heat dissipation path, the vortex copper tube 3 is composed of two semi-circular copper tubes joined together. The outer ends of the two semi-circular copper tubes are respectively set as the inlet and outlet of the cooling body to be cooled. The two semi-circular copper tubes can increase the heat dissipation path and improve the heat dissipation effect.

[0055] To prevent the heat dissipation of the two semicircular copper tubes from affecting each other, a heat insulation layer 31 is provided at the connection between the two copper tubes. The heat insulation layer 31 is made of a graphite polystyrene board with cavities, and the cavities in the heat insulation layer 31 are filled with inert gas. The heat insulation layer 31 can insulate the semicircular copper tubes and prevent heat radiation from affecting the heat dissipation effect.

[0056] By installing vortex copper pipes 3 through the interior of the heat dissipation fins 1 in a planar or three-dimensional manner, the shape of the copper pipes can be made consistent with the heat dissipation shape of the fan 2, and the three-dimensional shape is more conducive to heat dissipation and makes full use of space.

[0057] To reduce heat dissipation dead zones, a turbulence mechanism is also fixedly installed on the surface of the heat dissipation fins 1 to turbulent the DC heat dissipation airflow, so as to achieve composite heat dissipation.

[0058] The airflow disturbance is achieved as follows: the disturbance mechanism includes multiple disturbance grooves 4 formed on the side of the heat dissipation fins 1. An elastic steel sheet 41, which always extends outwards, is fixedly installed at one end of the inner wall of each disturbance groove 4. One end of the elastic steel sheet 41 is movably hinged to the inner wall of the disturbance groove 4. When one end of the elastic steel sheet 41 slides closer to the other end, the middle surface of the elastic steel sheet 41 springs outwards, thus disturbing the airflow. The elastic steel sheet 41 has an automatic reset effect and is corrosion-resistant and high-temperature resistant.

[0059] To facilitate linkage control of the turbulence, a through hole extending to the top or bottom of the heat dissipation fin 1 is provided at the center line of the length direction of the turbulence groove 4. A steel wire rope 42 is movably installed inside the through hole, and one end of the steel wire rope 42 is fixed to the movable end of the elastic steel sheet 41.

[0060] Miniature cylinders 43 are fixedly mounted in an array on the upper or lower surface of the heat dissipation fins 1. The piston rod of each miniature cylinder 43 is fixedly connected to the other end of the steel wire rope 42 to pull the elastic steel sheet 41. Each cylinder controls each elastic steel sheet 41 individually, providing a dual control effect of independent control and linkage control.

[0061] By setting up a turbulence mechanism, the straight heat dissipation airflow can be turbulent, increasing the heat dissipation area, reducing the direct area of ​​the airflow at the horizontal diameter of the copper pipe, and reducing heat dissipation dead angles, especially reducing heat dissipation dead angles on the back of the copper pipe.

[0062] Example 2

[0063] Reference Figure 2-8An optimized design method for composite heat dissipation fins in a plate-fin air-cooled radiator includes step one: data acquisition. A temperature sensor 5 is installed on the vortex array on the back of the vortex copper tube 3 to collect the temperature inside the vortex copper tube 3 in real time. The probe of the temperature sensor 5 can be installed on the back of the vortex copper tube 3, with the front end of the probe extending into two semi-circular copper tubes to collect the temperature data of the cooled object. This method avoids affecting airflow heat dissipation and can detect dead zone temperatures. The wiring harness can be threaded through both ends of the insulation layer 31.

[0064] To achieve precise digital control, the temperature sensors 5 in step one are grouped and then matched one-to-one with the miniature cylinders 43. The grouping method is based on the temperature sensing radius of the temperature sensor 5. As a standard, if one of the micro cylinders 43 controls the elastic steel sheet 41 to simultaneously fall into the temperature sensing radius of two adjacent temperature sensors 5 At this time, the miniature cylinder 43 enters the temperature sensor group 5 that is closest to the outlet. The turbulence effect of the elastic steel sheet 41 is directly displayed through data.

[0065] Step 2: Group the data collected by temperature sensor 5, the control signal data of micro cylinder 43, and the speed data of fan 2 to form a mapping relationship. .

[0066] Step 3: Based on the mapping relationship For each of the corresponding parameters, set a threshold for these parameters. The upper and lower limits of each parameter are obtained. Then, the dataset composed of these upper and lower limits is imported into the neural network model for training, verification and parameter tuning. Finally, the temperature data of temperature sensor 5, the control signal data of micro cylinder 43 and the speed data of fan 2 are derived from the trained parameters.

[0067] The neural network model consists of a four-layer ANN model. The input layer consists of data collected by temperature sensor 5, control signal data of micro cylinder 43, and fan speed data. There are two hidden layers, each with 64 neurons. The output data parameters are provided, and the activation function used is ReLU.

[0068] Preferably, in step three, the mapping relationship The dataset is normalized and divided into training and testing sets according to a certain ratio. The model is then trained, validated, and its parameters are tuned to build a neural network model.

[0069] During training, the loss function will calculate the sum of the predicted values ​​of the neural network model and... The errors between the labels of the dataset are backpropagated to update the parameters of the neural network model.

[0070] The neural network model is evaluated using mean absolute error (MAE), mean squared error (MSE), and coefficient of determination. As an evaluation metric, if the metric does not meet expectations, the parameter values ​​are adjusted. When the accuracy of the neural network model reaches the required level, the normalized information from the neural network model's training is saved to the database. This allows for the restoration of the input normalization process during training when using the neural network model for data processing.

[0071] Step 4: Import the speed data of temperature sensor 5, miniature cylinder 43 and fan 2 from step 3 into the database and reverse map out the respective models of temperature sensor 5, miniature cylinder 43 and fan 2.

[0072] Step 5: Input the model or parameter data of the reverse-mapped temperature sensor 5, miniature cylinder 43 and fan 2 into Step 3 to perform simulation verification to see if they exceed the upper and lower limits of the model or parameter data.

[0073] If it exceeds the limit, reduce the adjustment threshold by one level. Then proceed with steps three through five.

[0074] If the parameters are not exceeded, then export the models or parameters of temperature sensor 5, miniature cylinder 43, and fan 2.

[0075] Preferably, in step four, an inversion model of the neural network model is established during the reverse mapping process.

[0076] When the inversion model no longer matches the actual temperature sensor 5, miniature cylinder 43, and fan 2 data from step five with the data already stored in the database, the monitored values ​​for temperature sensor 5, miniature cylinder 43, and fan 2 are stabilized: the temperature fluctuation variance is calculated for each set of parameters. When temperature fluctuation variance Below the preset threshold When the monitored values ​​of temperature sensor 5, miniature cylinder 43, and fan 2 are determined to be in an effective heat dissipation stage, the values ​​of temperature sensor 5, miniature cylinder 43, and fan 2 are taken as inputs, and the output is obtained through the inversion model, thus completing the acquisition of the optimized operating parameters of the radiator. Temperature fluctuation variance is used. To stabilize and limit the various models or parameters of temperature sensor 5, miniature cylinder 43, and fan 2, so as to avoid excessive fluctuations that could cause inversion errors.

[0077] Step Six: End.

[0078] By setting up an optimized design method, intelligent heat dissipation can be achieved. At the same time, existing data can be used for intelligent neural network processing to optimize various parameters of the heat sink and generate heat dissipation solutions for different environments.

[0079] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A plate-fin air-cooled radiator composite heat dissipation fin, comprising a heat dissipation fin (1) mounted on a radiator and a fan (2) mounted on the radiator; characterized in that The inside of the heat dissipation fin (1) is installed with a vortex copper pipe (3) in a planar or three-dimensional manner; The surface of the heat dissipation fin (1) is also fixedly installed with a spoiler mechanism for disturbing the direct current heat dissipation airflow to realize the action of composite heat dissipation; The spoiler mechanism comprises a plurality of spoiler grooves (4) opened in the side surface of the heat dissipation fin (1), one end of the inner wall of the spoiler groove (4) is fixedly installed with an elastic steel sheet (41) which always pops out to the outside, one end of the elastic steel sheet (41) is hingedly connected to the inner wall of the spoiler groove (4), when one end of the elastic steel sheet (41) slides close to the other end, the middle surface of the elastic steel sheet (41) pops out to the outside to realize the action of airflow disturbance; A through hole extending to the top or bottom of the heat dissipation fin (1) is opened in the middle line of the length direction of the spoiler groove (4), a steel wire rope (42) is movably arranged in the through hole, one end of the steel wire rope (42) is fixed to the movable end of the elastic steel sheet (41); The upper surface and the lower surface of the heat dissipation fin (1) are both arrayed and fixedly installed with micro air cylinders (43), the piston rod of the micro air cylinder (43) is fixedly connected to the other end of the steel wire rope (42) to realize the action of pulling the elastic steel sheet (41).

2. The plate-fin air-cooled heat sink composite fin of claim 1, wherein: The center of the vortex copper pipe (3) and the axis of the fan (2) are located on the same axis line, and the vortex copper pipe (3) is installed on the heat dissipation fin (1) in a planar manner or in a conical shape.

3. The plate-fin type air-cooled heat sink composite heat dissipation fin according to claim 2, characterized in that: The vortex copper pipe (3) is composed of two semicircular copper pipes, and the outer ends of the two semicircular copper pipes are respectively provided as the inlet and outlet of the cooling body to be cooled.

4. The plate-fin type air-cooled heat sink composite heat dissipation fin according to claim 3, characterized in that: The connection part of the two copper pipes is also provided with a heat insulation layer (31), the heat insulation layer (31) is composed of a graphite polystyrene board with a cavity, and the cavity in the heat insulation layer (31) is filled with inert gas.

5. The method of optimizing the design of a plate-fin air-cooled heat sink composite fin according to claim 4, characterized in that: Including; Step one, data collection, a temperature sensor (5) is installed on the back of the vortex copper pipe (3) in a vortex array, the temperature in the vortex copper pipe (3) is collected in real time by the temperature sensor (5); Step two, the temperature sensor (5) of the acquisition data, micro cylinder (43) of control signal data and fan (2) rotation speed data grouping form mapping relationship ; Step three, according to the mapping relationship corresponding parameters, threshold values are set for these parameters , the upper and lower limit ranges of each parameter are obtained, then the data set composed of these upper and lower limit range parameters is imported into the neural network model for training, verification and parameter adjustment, and finally the temperature data of the temperature sensor (5), the control signal data of the micro cylinder (43) and the data of the fan (2) rotating speed are inversely derived according to the trained parameters; Step four, the temperature sensor (5), micro air cylinder (43) and fan (2) speed data in step three are imported into the database to reversely map the temperature sensor (5), micro air cylinder (43) and fan (2) respectively; Step five, the type or parameter data of the temperature sensor (5), micro air cylinder (43) and fan (2) after reverse mapping is brought into step three to verify whether it exceeds the upper and lower limit range of the type or parameter data; If exceeded, then reduce one adjustment threshold The range repeats steps three to five. If not, export the temperature sensor (5), micro air cylinder (43) and fan (2) each type or parameter; Step six, end.

6. The method according to claim 5, wherein: The temperature sensor (5) in the step one is grouped and then corresponds to the micro cylinder (43) one by one. The grouping mode is that the temperature sensor (5) temperature sensing radius is standard. If the elastic steel sheet (41) controlled by one of the micro cylinders (43) falls into the temperature sensing radius of two adjacent temperature sensors (5) at the same time , the micro cylinder (43) draws into the group of the temperature sensor (5) closest to the outlet at this time.

7. The method according to claim 5, wherein: In the third step, the mapping relationship The constituted data set is normalized, and the training set and the test set are divided in a certain proportion, the model is trained, verified and parameterized, so as to realize the construction of the neural network model. During the training process, the loss function will calculate the error between the label of the dataset composed of the predicted value of the neural network model and the propagates in reverse to update the parameters of the neural network model; The neural network model adopts mean absolute error (MAE), mean square error (MSE), and coefficient of determination (R2) during evaluation As an evaluation index, if the index does not reach the expectation, the parameter size is adjusted, and when the accuracy of the neural network model reaches the requirement, the normalization information during the training of the neural network model is saved in the database.

8. The method according to claim 7, wherein: In step four, the inverse model of the neural network model in the reverse mapping process is established; When the inversion model no longer matches the actual temperature sensor (5), micro-cylinder (43) and fan (2) model or parameter data in step five with the model or parameter data already stored in the database, the monitored temperature sensor (5), micro-cylinder (43) and fan (2) values are maintained: calculate the temperature fluctuation variance for each set of parameters When the temperature fluctuation variance is lower than the preset threshold , it is determined that the monitored temperature sensor (5), micro-cylinder (43) and fan (2) values are in the effective heat dissipation stage, at this time, the temperature sensor (5), micro-cylinder (43) and fan (2) values are taken as input, and the output is obtained through the inversion model, completing the acquisition of the working parameters of the optimized radiator.

Images