Dry-type rectifier transformer heat dissipation method and system

Abstract

The present application relates to the field of data processing, and relates to a dry-type rectifier transformer heat dissipation method and system, wherein the method comprises: installing a bottom fan at the bottom air duct inlet of the dry-type rectifier transformer, which is used for obliquely blowing the lower coil and the iron core of the dry-type rectifier transformer; installing an outside fan at the middle part outside the coil of the dry-type rectifier transformer, which is used for obliquely blowing the upper coil; the bottom fan adopts a cross-flow type cooling fan, and the outside fan adopts a centrifugal fan; the determination method of the angle between the fan axis of the bottom fan and the horizontal plane and the angle between the fan axis of the outside fan and the horizontal plane comprises: inputting each angle combination into a preset neural network model respectively to obtain the corresponding comprehensive temperature rise of each angle combination; and selecting the angle combination with the minimum comprehensive temperature rise from the multiple angle combinations as the optimal angle combination. The method of the present application can effectively improve the heat dissipation effect of the dry-type rectifier transformer.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing. More particularly, the present application relates to a dry-type rectifier transformer heat dissipation method and system. BACKGROUND

[0002] The dry-type rectifier transformer is a key equipment in the fields of industrial direct current transmission, electrolytic metallurgy, hydrogen production and energy storage, etc. It realizes high-power direct current output through the "primary side phase shift + secondary side split" topology, and has a relatively high rated load efficiency. With the increase of power demand in hydrogen production and other scenarios, the capacity of the dry-type rectifier transformer continues to increase, and the temperature rise problem is increasingly prominent. The temperature rise directly affects the performance, service life and operation reliability of the equipment, and the excessive hot spot temperature can accelerate the insulation aging and even cause the equipment to burn out.

[0003] In the prior art, the dry-type rectifier transformer is cooled by installing a bottom fan at the bottom of the dry-type rectifier transformer coil to blow the lower coil and the core of the dry-type rectifier transformer obliquely upward, thereby achieving heat dissipation of the dry-type rectifier transformer. However, for the dry-type rectifier transformer with an axial double-split structure (upper and lower coils are connected in parallel and the middle is disconnected), the axial double-split structure naturally blocks the heat dissipation air duct, resulting in significant defects in the traditional single-fan cooling scheme. Specifically, the traditional single-fan cooling scheme has the following disadvantages:

[0004] I. Uneven heat dissipation and significant temperature difference: The cooling airflow driven by the single fan (bottom cross-flow type) produces a throttling effect when flowing through the middle disconnected area, and the flow rate of the upper coil main air duct is only 50% or less than that of the lower coil. The accumulation of hot air causes the temperature of the upper coil to be much higher than that of the lower coil. Actual measurement shows that the temperature difference between the upper and lower coils of the low-voltage coil under the single-fan configuration can reach 35-40K, and the hot spot temperature rise of the upper coil can reach 130K.

[0005] II. Low heat exchange efficiency: The heat exchange coefficient of the upper coil area is low due to insufficient flow rate, and the heat dissipation efficiency is reduced by about 41% compared with the lower coil, and the risk of local overheating is prominent.

[0006] III. Poor adaptability: The air duct segmentation characteristics of the axial double-split structure do not match the "one-way blowing" mode of the single fan, and cannot meet the symmetrical heat dissipation demand. The prior art lacks a systematic heat dissipation optimization scheme for this structure. SUMMARY

[0007] To solve the technical problems of uneven heat dissipation, significant temperature difference, low heat exchange efficiency and poor adaptability of the dry-type rectifier transformer heat dissipation method in the prior art, the present application provides solutions in the following aspects.

[0008] In a first aspect, the present application provides a dry-type rectifier transformer heat dissipation method, the dry-type rectifier transformer being an axial double-split structure, the method comprising: installing a bottom fan at a coil bottom air duct inlet of the dry-type rectifier transformer for obliquely upward blowing of a lower coil and a core of the dry-type rectifier transformer; installing an outer fan at a coil outer side of the dry-type rectifier transformer for obliquely upward blowing of an upper coil; the bottom fan being a cross-flow type cooling fan, and the outer fan being a centrifugal fan; and a determination method of an included angle between a fan axis of the bottom fan and a horizontal plane and an included angle between a fan axis of the outer fan comprising:

[0009] taking an angle value every first angle step in a first preset angle range to obtain a plurality of first angles of the bottom fan; and taking an angle value every first angle step in a second preset angle range to obtain a plurality of first angles of the outer fan;

[0010] combining the first angles of the bottom fan and the first angles of the outer fan two by two to obtain a plurality of angle combinations, and inputting each angle combination into a preset neural network model to obtain a comprehensive temperature rise corresponding to each angle combination; the comprehensive temperature rise being a temperature rise obtained by weighting and summing a high-voltage coil average temperature rise, a lower low-voltage coil temperature rise and an upper low-voltage coil temperature rise by using an entropy weight method; the input of the neural network model being the angle combination of the bottom fan and the outer fan, and the output being the comprehensive temperature rise;

[0011] selecting an angle combination with the minimum comprehensive temperature rise from the plurality of angle combinations as an optimal angle combination, and taking an angle of the bottom fan corresponding to the optimal angle combination as the included angle between the fan axis of the bottom fan and the horizontal plane, and taking an angle of the outer fan corresponding to the optimal angle combination as the included angle between the fan axis of the outer fan and the horizontal plane.

[0012] Preferably, the training method of the neural network model comprises: constructing a data set, and training the neural network model by using data in the data set; and the construction method of the data set comprises:

[0013] taking an angle value every second angle step in a first preset angle range to obtain each second angle of the bottom fan; and taking an angle value every second angle step in a second angle range to obtain each second angle of the outer fan; the second angle step being greater than the first angle step;

[0014] combining the second angles of the bottom fan and the second angles of the outer fan two by two to obtain a plurality of test angle combinations;

[0015] The fan shafts of the bottom fan and the outer fan are arranged at angles with the horizontal plane according to the angles in each test angle combination, heat dissipation is performed on the dry-type rectifier transformer, and the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil of the dry-type rectifier transformer are measured; and the entropy weight method is used to perform weighted summation on the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil, so as to obtain the corresponding comprehensive temperature rise of each test angle combination.

[0016] The test angle combination and the corresponding comprehensive temperature rise are taken as sample data in the data set.

[0017] Preferably, the method for constructing the data set further comprises: taking 80% of the sample data in the data set as a training set and 20% of the sample data as a test set.

[0018] Preferably, the weighted summation of the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil by the entropy weight method comprises:

[0019] A sample set is constructed, wherein the samples in the sample set are test angle combinations and corresponding average temperature rises of high-voltage coils, temperature rises of lower low-voltage coils, and temperature rises of upper low-voltage coils;

[0020] The average temperature rises of high-voltage coils, the temperature rises of lower low-voltage coils, and the temperature rises of upper low-voltage coils corresponding to each test angle combination are standardized;

[0021] The entropy values of each temperature rise index are calculated, and the calculation expression is:

[0022] ;

[0023] ;

[0024] The weights of each temperature rise index are calculated, and the calculation expression is:

[0025] ;

[0026] The value T of the comprehensive temperature rise index is calculated, and the calculation expression is:

[0027] ;

[0028] In the above formulas, , and respectively represent the values of the first temperature rise index, the second temperature rise index, and the third temperature rise index; is the standardized value of the jth temperature rise index corresponding to the ith sample, an entropy value of the jth temperature rise index, n represents the number of samples in the sample set, represents a natural logarithm, represents a normalized proportion of the jth temperature rise index of the ith sample; represents a weight of the jth temperature rise index, is a constant.

[0029] Preferably, when the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil and the temperature rise of the upper low-voltage coil corresponding to each test angle combination are normalized, the corresponding calculation expression is:

[0030] ;

[0031] In the formula, is the jth temperature rise index corresponding to the ith sample, represents the maximum value of the jth temperature rise index in the sample set, represents the minimum value of the jth temperature rise index in the sample set.

[0032] Preferably, the first preset angle range and the second preset angle range are both 20 degrees to 70 degrees.

[0033] Preferably, the first angle step is 1 degree.

[0034] Preferably, the second angle step is 10 degrees.

[0035] Preferably, when the dry-type rectifier transformer is cooled, the fan flow rate of the bottom fan and the fan flow rate of the outer fan are both rated flow rates.

[0036] In the second aspect, the present application provides a dry-type rectifier transformer cooling system, comprising:

[0037] a bottom fan installed at the coil bottom air duct inlet of the dry-type rectifier transformer, used for obliquely upward blowing the lower coil and the core of the dry-type rectifier transformer;

[0038] an outer fan installed at the middle part of the coil outer side of the dry-type rectifier transformer, used for obliquely upward blowing the upper coil of the dry-type rectifier transformer;

[0039] a controller connected to the bottom fan and the outer fan to execute the dry-type rectifier transformer cooling method of the present application.

[0040] The present application has the beneficial effect of solving the core pain points of "upper coil heat accumulation, significant temperature difference between upper and lower (up to 35-40K), local hot spot temperature rise (over 130K)" in the traditional single fan cooling scheme of the axial double-split structure dry-type rectifier transformer. On the one hand, the present application designs a collaborative cooling structure of "bottom cross-flow fan + outer centrifugal fan" according to the air duct blocking characteristics of the axial double-split structure - the bottom fan is responsible for the basic cooling of the lower coil and the core, and the outer fan blows upward to form a directional air flow path of "pushing down and supplementing up", which accurately solves the problem of upper hot air retention; on the other hand, by introducing a quantitative optimization model of "entropy weight method + neural network", taking the angle between the fan axis and the horizontal direction as the core parameter, combining with the coil temperature rise multi-index weight calculation, constructing a comprehensive temperature rise index, through sample training and iterative optimization, determining the optimal layout parameters of the double fan, and realizing the precise maximization of cooling efficiency. This scheme not only adapts to the cooling needs of the axial double-split winding from the structural level, but also fully releases the synergistic effect of the double fan through quantitative optimization, which can increase the flow rate of the upper coil area by 81%, increase the heat exchange coefficient by 40%, reduce the hot spot temperature rise of the low-voltage coil from 130K to 83K, and reduce the temperature difference between the upper and lower coils to 7.5K, effectively avoiding the risk of accelerated insulation aging, and having key technical support significance for ensuring the long-term operation reliability of the dry-type rectifier transformer in hydrogen production and energy storage, industrial direct current transmission and other scenes, and prolonging the service life of the equipment.

[0041] In addition, the dry-type rectifier transformer cooling method of the present embodiment can avoid the subjectivity and limitations of the traditional trial-and-error method, and ultimately obtain the optimal fan angle with "cooling efficiency" and "temperature rise uniformity", providing a quantitative basis for the engineering design of the double fan cooling system. BRIEF DESCRIPTION OF DRAWINGS

[0042] The above and other objects, features and advantages of the exemplary embodiments of the present application will be more apparent from the following detailed description read in conjunction with the accompanying drawings, in which several embodiments of the present application are shown by way of example, and wherein the same or corresponding elements are referred to by the same or corresponding reference numerals. In the drawings:

[0043] Figure 1 is a flow chart schematically showing a bottom fan corresponding angle and an outer fan corresponding angle determination method according to an embodiment of the present application;

[0044] Figure 2 is a schematic diagram of a transformer structure model according to an embodiment of the present application;

[0045] Figure 3 is a schematic diagram of a multi-layer structure of a coil according to an embodiment of the present application;

[0046] Figure 4is a schematic diagram illustrating a local mesh partition in a computing domain according to an embodiment of the present application;

[0047] Figure 5 is a temperature rise distribution diagram illustrating a single-fan cooling condition of a dry-type rectifier transformer according to an embodiment of the present application;

[0048] Figure 6 is a flow field distribution diagram illustrating a single-fan cooling condition of a dry-type rectifier transformer according to an embodiment of the present application;

[0049] Figure 7 is a temperature rise distribution diagram illustrating a double-fan cooling condition of a dry-type rectifier transformer according to an embodiment of the present application;

[0050] Figure 8 is a flow field distribution diagram illustrating a double-fan cooling condition of a dry-type rectifier transformer according to an embodiment of the present application;

[0051] Figure 9 is a structural diagram of a dry-type rectifier transformer cooling system according to an embodiment of the present application;

[0052] Figure 10 is a structural diagram of a controller according to an embodiment of the present application. DETAILED DESCRIPTION

[0053] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0054] Embodiment of a dry-type rectifier transformer cooling method:

[0055] The dry-type rectifier transformer cooling method of the present application is applicable to a dry-type rectifier transformer with an axial double-split structure, such as Figure 1 As shown in the figure, the method comprises: installing a bottom fan at the bottom of the coil of the dry-type rectifier transformer for blowing the lower coil and the core of the dry-type rectifier transformer obliquely upwards; installing an outside fan at the middle of the coil of the dry-type rectifier transformer for blowing the upper coil obliquely upwards; the bottom fan adopts a cross-flow cooling fan, and the outside fan adopts a centrifugal fan; and the determination method of the angle between the fan axis of the bottom fan and the horizontal plane and the angle between the fan axis of the outside fan and the horizontal plane comprises:

[0056] S101, obtaining a plurality of first angles of the bottom fan and a plurality of first angles of the outside fan, specifically: taking an angle value every first angle step in a first preset angle range to obtain a plurality of first angles of the bottom fan; and taking an angle value every first angle step in a second preset angle range to obtain a plurality of first angles of the outside fan;

[0057] In the embodiment, the first preset angle range and the second preset angle range are both 20 degrees to 70 degrees. In other embodiments, other suitable ranges can also be taken.

[0058] In the embodiment, the first angle step is 1 degree. In other embodiments, other suitable steps such as 0.5, 0.1, etc. can also be taken.

[0059] S102, obtain a plurality of angle combinations and corresponding comprehensive temperature rises, specifically: combine the first angle of the bottom fan with the first angle of the outer fan two by two to obtain a plurality of angle combinations, and input each angle combination into a preset neural network model to obtain the comprehensive temperature rise corresponding to each angle combination; the comprehensive temperature rise is the temperature rise obtained by weighting and summing the high-voltage coil average temperature rise, the lower low-voltage coil temperature rise, and the upper low-voltage coil temperature rise using the entropy weight method; the input of the neural network model is the angle combination of the bottom fan and the outer fan, and the output is the comprehensive temperature rise;

[0060] S103, obtain the angle between the fan axis of the bottom fan and the horizontal plane, and the angle between the fan axis of the outer fan and the horizontal plane, specifically: select the angle combination with the smallest comprehensive temperature rise from the plurality of angle combinations as the optimal angle combination, and take the angle of the bottom fan corresponding to the optimal angle combination as the angle between the fan axis of the bottom fan and the horizontal plane, and take the angle of the outer fan corresponding to the optimal angle combination as the angle between the fan axis of the outer fan and the horizontal plane.

[0061] Through experiments, when the angle between the fan axis of the bottom fan and the horizontal plane is 20° and the angle between the fan axis of the outer fan and the horizontal plane is 35°, the comprehensive temperature rise T reaches the minimum value. Compared with the non-optimized arrangement scheme in the actual project, it is reduced by 18%, and the coil area temperature rise distribution is more uniform.

[0062] When the dry-type rectifier transformer is cooled, the fan flow rate of the bottom fan and the fan flow rate of the outer fan are both rated flow rates.

[0063] The bottom fan and the outer fan are combined to form a "push down and supplement up" air flow path, which specifically solves the upper cooling bottleneck of the axial double-split structure.

[0064] In order to verify the cooling effect of the combination of the bottom fan and the outer fan on the dry-type rectifier transformer, in the embodiment, a dry-type rectifier transformer for hydrogen production and energy storage with a model number of ZSCZB-7500 / 10 is taken as an example. The finite volume method (FVM) is used to simulate and calculate the dry-type rectifier transformer under two cooling modes of bottom single fan cooling and bottom fan and outer fan cooperative cooling, so as to obtain the temperature rise distribution and the air flow field distribution. The main performance parameters of the transformer prototype of the dry-type rectifier transformer for hydrogen production and energy storage with a model number of ZSCZB-7500 / 10 are shown in Table 1:

[0065] Table 1

[0066]

[0067] The specific process of the simulation calculation is as follows:

[0068] S1, a three-dimensional simulation model of the dry-type rectifier transformer is established.

[0069] The established three-dimensional simulation model includes a transformer structure model and a fluid domain model.

[0070] The transformer structure model includes a core, a winding, a cooling system, a shell, and a surrounding environment.

[0071] The fluid domain model includes an air flow area inside the transformer and an air flow area outside the transformer. It is ensured that the air flow path and the air flow exchange of each part in the transformer can be truly represented.

[0072] The core adopts a conjugate structure, the high-voltage coil is a multi-layer cylindrical structure (parallel connection of upper and lower parts), and the low-voltage coil is an axial double split structure (upper and lower height ratio 1:1).

[0073] Specifically, based on the design parameters of the dry-type transformer, the core and the coil are subjected to three-dimensional equivalent treatment to obtain a numerical calculation model of the dry-type transformer; the three-dimensional model should include structures that have a greater impact on the calculation results of physical fields. The core structure parameters include window height, main column center distance, main column piece width, yoke piece width, yoke offset distance, hole distance, and conjugate piece width; the coil structure parameters include electrical height, inner diameter, outer diameter, and electrical turn number of the coil; the coil arrangement parameters include excitation coil inter-phase distance, excitation coil distance from the upper iron yoke, and distance from the lower iron yoke; and the coil electrical parameters include connection group, rated voltage, turns per pie, and radial parallel winding number per pie.

[0074] The established transformer structure model is shown in Figure 2 .

[0075] The cooling system includes a bottom fan and an outer fan. The bottom fan is a conventional cross-flow type cooling fan, which is installed at the air duct inlet at the bottom of the coil, and its function is consistent with that of the traditional single fan: by forcibly sucking in ambient cold air, it blows upward along the air duct from the bottom to directly remove the heat of the lower part of the coil and the bottom of the core; the other is an outer centrifugal fan, which is installed at the middle position of the outer side of the coil, and by adjusting the included angle between the air outlet and the surface of the coil, the airflow is obliquely blown upward to the upper part of the winding.

[0076] The configuration parameters of the bottom fan and the outer fan are shown in Table 2.

[0077] Table 2

[0078]

[0079] The material property parameters of the established transformer structure model include air thermal characteristic parameters and solid material thermal characteristic parameters, wherein the air thermal characteristic parameters are shown in Table 3, and the solid material thermal characteristic parameters are shown in Table 4.

[0080] Table 3

[0081]

[0082] Table 4

[0083]

[0084] In this embodiment, the insulation layer of the dry-type rectifier transformer coil is simplified as two coupled thin layers, namely an insulation paper layer and an epoxy resin layer.

[0085] The insulation layer of the dry-type transformer coil is equivalently processed: the dry-type transformer coil adopts a three-layer composite insulation structure of “copper wire-insulation paper-epoxy resin”, wherein the thermal conductivity of copper is about 400 W / (m•K), the thermal conductivity of insulation paper is about 0.11 W / (m•K), and the thermal conductivity of epoxy resin is about 0.2 W / (m•K). The thermal conductivities of the latter two insulation materials are only 1 / 2000-1 / 3000 of that of copper, and are the main resistance of heat transfer from the wire to the environment.

[0086] In order to consider the influence of the insulation paper and the epoxy resin on the temperature rise of the coil, the insulation paper and the epoxy resin are equivalently simplified, and two coupled thin layer structures are defined on the surface of the coil. The key influence of the insulation paper and the epoxy resin on heat dissipation is retained, the simulation can accurately reflect the actual temperature rise law of the coil, the calculation complexity caused by the modeling of the multi-layer complex structure is avoided, the model is simplified without losing the core heat dissipation characteristics, and the feasibility and calculation efficiency of the simulation are improved. The schematic diagram of the multi-layer structure of the coil is shown in Figure 3 .

[0087] The dry-type transformer is parameterized. Including:

[0088] Material parameters: magnetic-thermal related physical parameters of the core, the coil, the insulation paper, and the epoxy resin material;

[0089] Loss model: the coil loss is decomposed into direct current resistance loss (increasing with temperature rise) and eddy current loss (decreasing with temperature rise), which is fitted as a function of temperature and current:

[0090] ;

[0091] In the formula, represents the coil loss, represents the direct current resistance loss of the coil under the condition of 75℃, R represents the eddy current loss of the coil at 75°C, and T represents the actual temperature of the coil.

[0092] S2, according to the geometric shape of the transformer, the calculation domain is meshed. Both the inside and outside of the transformer need to be meshed, but the mesh of the airflow area is more fine than that of the solid area, because the airflow and temperature distribution are very sensitive to the mesh resolution. When meshing, the key areas such as the coil and the air duct are meshed densely, and the expansion layer is set on the fluid-structure coupling boundary to adapt to the boundary layer effect.

[0093] Under the condition of thermal convection, there is a boundary layer between the air and the solid flowing through the surface, and the heat exchange between the solid heat source surface and the airflow mainly occurs in the boundary layer. The flow rate of the air directly contacting the solid surface is zero, and the flow rate increases rapidly along the normal direction of the contact surface after leaving the solid surface. In the transformer body, when the air flows through the surface of the coil as a cooling medium, the flow rate of the air at the surface of the coil is zero, and the flow rate increases after leaving the surface of the coil. Due to the air boundary layer effect, the air flow rate gradient and temperature gradient in the boundary layer of the solid material surface are very large, so an appropriate expansion layer network needs to be divided at the fluid-structure coupling boundary to reduce the calculation error caused by insufficient heat exchange at the fluid-structure coupling interface. The local mesh division in the calculation domain is shown in Figure 4 As shown in the figure, fine mesh division is adopted for key positions such as the coil and the air duct, and appropriate mesh sparsification is performed for other structural components according to their influence on the temperature field.

[0094] In this way, the accuracy of the calculation results is ensured, and the excessive consumption of computing resources and the significant increase of calculation time caused by excessive number of meshes are avoided.

[0095] S3, the mass equation, energy equation and momentum equation of the transformer are established.

[0096] The momentum equation is used to describe the velocity field of the airflow, which is usually the Navier-Stokes equation, involving the velocity, pressure and density of the airflow. The energy equation is used to describe the temperature field, involving the heat sources of each part of the transformer (such as the heat generated by the current passing through the coil) and the processes of heat conduction, convection and radiation. The mass equation (also known as the continuity equation) is the basic equation for describing the conservation of mass in fluid flow, ensuring the balance of mass input and output in the entire calculation domain. It expresses the relationship between the mass density of the fluid and the flow rate.

[0097] S4, set the boundary conditions and initial conditions.

[0098] The initial conditions are: the initial temperature and the ambient temperature are both 20°C, when using single fan cooling, the flow rate of the lower fan is the rated flow rate, and the flow rate of the outer fan is 0, when using double fan cooling, both fans are at the rated flow rate.

[0099] The inlet boundary condition is that the air flow inlet velocity is the fan flow rate. The rest of the computational domain boundaries are open boundaries.

[0100] S5, solving equations.

[0101] According to the boundary conditions and initial conditions, the finite volume method (FVM) is used to solve the mass equation, momentum equation and energy equation in the control volume. By discretizing these equations, the air flow velocity, temperature and pressure in each control volume are obtained. In the solving process, the temperature field, air flow field and other variables need to be iteratively updated until the results converge.

[0102] In the working condition of single fan cooling, the bottom fan is at rated flow rate, and the side fan flow rate is 0 m 3 / h. The ambient temperature is 20℃, and after the calculation converges, the temperature rise distribution is as shown in Figure 5 . The global hot spot temperature rise of the body is located on the upper part of the low-voltage coil, reaching 123K; the hot spot temperature rise of the high-voltage coil is 91K, and the overall temperature gradient presents high in the upper part and low in the lower part. The average temperature rise of the upper low-voltage coil is 91K, and the average temperature rise of the lower low-voltage coil is 55K, with a temperature difference of 36K; the temperature difference between the upper and lower parts of the high-voltage coil is smaller than that of the low-voltage coil, but there is still a significant temperature stratification. The obtained air flow field distribution is as shown in Figure 6 .

[0103] In the working condition of double fan cooling, it is assumed that both fans operate at rated flow rate, and the ambient temperature is still 20℃. The results show that the device temperature rise distribution is significantly improved, and the temperature rise distribution is as shown in Figure 7 . The hot spot position of the low-voltage coil shifts from the upper part to the lower part, and the hot spot temperature rise decreases from 120K to 83K, with a decrease of 30.8%; the average temperature rise of the upper low-voltage coil decreases to 57.5K, and the hot spot temperature difference between the upper and lower parts is reduced to 7.5K, which is reduced by 80% compared with the single fan configuration. The hot spot temperature rise of the upper low-voltage coil decreases to 71K, and the hot spot temperature difference between the upper and lower parts is reduced to 12K. The hot spot temperature rise of the high-voltage coil decreases to 66K, and the temperature difference between the upper and lower parts is reduced from 19K to 5.5K, and the temperature distribution uniformity is significantly improved. The air flow field distribution is as shown in Figure 8 . Compared with the air flow field distribution obtained in the working condition of single fan cooling, the side fan increases the flow rate in the upper coil area to more than 3.48m / s, which is increased by about 81% compared with the single fan configuration. The low-temperature air (about 20℃) introduced by the side fan forms strong convection with the hot air in the upper part, breaking the original heat accumulation layer, and increasing the heat transfer coefficient of the upper coil by about 40%. It is worth noting that when the double fan operates, the average temperature rise of the lower low-voltage coil is 65K, and the average temperature rise of the upper low-voltage coil is 57.5K. This "low-high-low" temperature distribution is opposite to that of the single fan, which confirms the strengthening effect of the side flow on the upper part cooling.

[0104] S6, test verification.

[0105] The comparison between the test data and the simulation results is shown in Table 5 and Table 6. Table 5 shows the calculated and measured average temperature rise and hot spot temperature rise of each coil of the transformer under the condition of single fan cooling. Table 6 shows the calculated and measured average temperature rise and hot spot temperature rise of each coil of the transformer under the condition of double fan cooling.

[0106] As shown in Table 5, under the condition of single fan cooling, the calculated average temperature rise of the high-voltage coil is 71.1K, and the measured value is 68.1K, with an error of 4.40%; the calculated average temperature rise of the upper low-voltage coil is 91K, and the measured value is 94.7K, with an error of 3.9%, both of which are relatively small. Under the condition of double fan cooling, the calculated average temperature rise of the high-voltage coil is 52.0K, and the measured value is 49.5K, with an error of 5.05%; the calculated average temperature rise of the upper low-voltage coil is 57.5K, and the measured value is 59K, with an error of 2.54%, which further verifies the reliability of the simulation model.

[0107] Table 5

[0108]

[0109] Table 6

[0110]

[0111] In summary, the double fan configuration effectively solves the upper cooling bottleneck of the traditional single fan scheme by optimizing air flow organization, significantly reduces the temperature rise level of the equipment, and makes the temperature rise more uniform, which provides a quantitative basis for improving the operation reliability of the rectifier transformer. The test verification shows that this scheme can reduce the hot spot temperature rise of the low-voltage coil by 37K, which meets the cooling demand of high-power density equipment.

[0112] To solve the core pain points of "upper coil heat accumulation, significant temperature difference between upper and lower (up to 35-40K), local hot spot temperature rise (over 130K)" in the traditional single fan cooling scheme of the axial double-split structure dry-type rectifier transformer. On the one hand, the application designs a collaborative cooling structure of "bottom cross-flow fan + outer centrifugal fan" according to the wind channel blocking characteristics of the axial double-split structure - the bottom fan is responsible for the basic cooling of the lower coil and the core, and the outer fan blows upward to form a directional air flow path of "pushing down and supplementing up", which accurately solves the problem of upper hot air retention; on the other hand, by introducing a quantitative optimization model of "entropy weight method + neural network", taking the angle between the fan axis and the horizontal direction as the core parameter, combining the coil temperature rise multi-index weight calculation, constructing a comprehensive temperature rise index, and through sample training and iterative optimization, the optimal layout parameters of the double fan are determined, and the cooling efficiency is accurately maximized. This scheme not only adapts to the cooling needs of the axial double-split winding from the structural level, but also releases the synergistic effect of the double fan through quantitative optimization, which can increase the flow rate of the upper coil area by 81%, increase the heat exchange coefficient by 40%, reduce the hot spot temperature of the low-voltage coil from 130K to 83K, and reduce the temperature difference between the upper and lower coils to 7.5K, effectively avoiding the risk of accelerated insulation aging. It has key technical support significance for ensuring the long-term operation reliability of dry-type rectifier transformers in hydrogen production and energy storage, industrial direct current transmission and other scenarios, and prolonging the service life of the equipment.

[0113] In addition, the dry-type rectifier transformer cooling method of the embodiment can avoid the subjectivity and limitations of the traditional trial-and-error method, and ultimately obtain an optimal fan angle that combines "cooling efficiency" and "temperature rise uniformity", providing a quantitative basis for the engineering design of the double fan cooling system.

[0114] In one embodiment, the training method of the neural network model includes: constructing a data set, using 80% of the sample data in the data set as a training set and 20% of the sample data as a test set. And use the training set and the test set to train the neural network model, and use the cross-entropy loss function to calculate the loss during training.

[0115] In this embodiment, the topology of the neural network model is: the neural network model includes an input layer, a hidden layer and an output layer, wherein the number of neurons in the input layer is 2, which are used to receive the angle of the bottom fan and the angle of the outer fan. The neurons of the hidden layer are 2 layers, each layer has 16 neurons, which are used to fit the nonlinear relationship between the angle and the temperature rise; the number of neurons in the output layer is 1, which is used to output the comprehensive temperature rise.

[0116] In this embodiment, the data set construction method includes:

[0117] S201, take an angle value every second angle step in the first preset angle range, so as to obtain each second angle of the bottom fan; take an angle value every second angle step in the second angle range, so as to obtain each second angle of the outer fan; the second angle step is greater than the first angle step;

[0118] In the embodiment, the second angle step is 10°.

[0119] The first preset angle range is 20°-70°, so an angle is taken every 10°, and the second angles of the bottom fans obtained are 20°, 30°, 40°, 50°, 60° and 70°, a total of six second angles of the bottom fans.

[0120] Similarly, the second preset angle range is also 20°-70°, so an angle is taken every 10°, and the second angles of the outer fans obtained are 20°, 30°, 40°, 50°, 60° and 70°, a total of six second angles of the outer fans.

[0121] S202, combine the second angles of the bottom fans with the second angles of the outer fans two by two, so as to obtain a plurality of test angle combinations;

[0122] The second angles of the six bottom fans are combined with the second angles of the six outer fans two by two, and 36 test angle combinations can be obtained, covering the main included angle working conditions, and all samples are calculated at the rated flow rate of the fan.

[0123] S203, respectively, according to the angles in each test angle combination, the fan axis of the bottom fan and the included angle with the horizontal plane, and the fan axis of the outer fan and the included angle with the horizontal plane are arranged, the dry-type rectifier transformer is cooled, and the average temperature rise of the high-voltage coil, the lower low-voltage coil temperature rise and the upper low-voltage coil temperature rise of the dry-type rectifier transformer are measured; and the entropy weight method is used to weight and sum the average temperature rise of the high-voltage coil, the lower low-voltage coil temperature rise and the upper low-voltage coil temperature rise, so as to obtain the comprehensive temperature rise corresponding to each test angle combination;

[0124] S204, the test angle combination and the corresponding comprehensive temperature rise are taken as sample data in the data set.

[0125] In one embodiment, the entropy weight method is used to weight and sum the average temperature rise of the high-voltage coil, the lower low-voltage coil temperature rise and the upper low-voltage coil temperature rise, including:

[0126] S301, construct a sample set, wherein the samples in the sample set are test angle combinations and corresponding average temperature rises of high-voltage coils, lower low-voltage coil temperature rises and upper low-voltage coil temperature rises;

[0127] S302, standardize the high-voltage coil average temperature rise, the lower low-voltage coil temperature rise and the upper low-voltage coil temperature rise corresponding to each test angle combination;

[0128] Since the three temperature rise indicators have the same dimension (all in K), the maximum and minimum value method, the range method or the standard deviation method can be used to normalize the temperature rise indicators in the sample set.

[0129] S303, calculate the entropy value of each temperature rise indicator, the calculation expression is:

[0130] (1)

[0131] (2)

[0132] S304, calculate the weight of each temperature rise indicator, the calculation expression is:

[0133] (3)

[0134] S305, calculate the value T of the comprehensive temperature rise indicator, the calculation expression is:

[0135] (4)

[0136] In formula (1) to formula (4), , and respectively represent the value of the first temperature rise indicator, the second temperature rise indicator and the third temperature rise indicator; is the normalized value of the jth temperature rise indicator corresponding to the ith sample, represents the entropy value of the jth temperature rise indicator, and n represents the number of samples in the sample set, represents the natural logarithm, represents the standardization proportion of the jth temperature rise indicator of the ith sample; represents the weight of the jth temperature rise indicator, is a constant.

[0137] In one embodiment, when the high-voltage coil average temperature rise, the lower low-voltage coil temperature rise and the upper low-voltage coil temperature rise corresponding to each test angle combination are standardized, the corresponding calculation expression is:

[0138] (5)

[0139] In the formula, is the jth temperature rise indicator corresponding to the ith sample, denotes the maximum value of the jth temperature rise indicator in the sample set, denotes the minimum value of the jth temperature rise indicator in the sample set, denotes the normalized value of the jth temperature rise indicator corresponding to the ith sample.

[0140] Dry-type rectifier transformer heat dissipation system embodiment:

[0141] The present application also provides a dry-type rectifier transformer heat dissipation system. As shown in Figure 9 The dry-type rectifier transformer heat dissipation system includes a bottom fan installed at the coil bottom air duct inlet of the dry-type rectifier transformer, for obliquely upward blowing the lower coil and core of the dry-type rectifier transformer.

[0142] An outer fan is installed at the middle part of the coil outer side of the dry-type rectifier transformer, for obliquely upward blowing the upper coil of the dry-type rectifier transformer.

[0143] A controller is connected to the bottom fan and the outer fan, to execute the dry-type rectifier transformer heat dissipation method described in the above embodiment.

[0144] As shown in Figure 10 The controller includes a processor and a memory, and the memory stores computer program instructions, which, when executed by the processor, implement the dry-type rectifier transformer heat dissipation method described in the above embodiment.

[0145] The controller also includes a communication bus and a communication interface, and other components well known to those skilled in the art, which are set and function as known in the art, and thus will not be described here.

[0146] In the description of the present specification, the meaning of "a plurality of", "several" is at least two, for example, two, three or more, etc., unless otherwise explicitly specified.

[0147] Although the present specification has shown and described several embodiments of the present application, it will be apparent to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, modifications and alternatives without departing from the idea and spirit of the present application. It should be understood that various alternatives to the embodiments of the present application described herein can be employed in practicing the present application.

Claims

1. A heat dissipation method for a dry-type rectifier transformer, wherein the dry-type rectifier transformer has an axially double-split structure, characterized in that, The method includes: installing a bottom fan at the bottom air duct inlet of the dry-type rectifier transformer coil to blow upwards at an angle onto the lower coil and core of the dry-type rectifier transformer; installing an outer fan at the middle of the outer side of the dry-type rectifier transformer coil to blow upwards at an angle onto the upper coil; the bottom fan is a cross-flow cooling fan, and the outer fan is a centrifugal fan; the method for determining the angle between the fan axis of the bottom fan and the horizontal plane, and the angle between the fan axis of the outer fan and the horizontal plane includes: Within a first preset angle range, an angle value is taken at every first angle step to obtain multiple first angles for the bottom fan; within a second preset angle range, an angle value is taken at every first angle step to obtain multiple first angles for the outer fan. The first angle of the bottom fan and the first angle of the outer fan are combined in pairs to obtain multiple angle combinations. Each angle combination is then input into a preset neural network model to obtain the comprehensive temperature rise corresponding to each angle combination. The comprehensive temperature rise is obtained by weighting and summing the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil using the entropy weight method. The input of the neural network model is the angle combination of the bottom fan and the outer fan, and the output is the comprehensive temperature rise. The angle combination with the smallest overall temperature rise is selected from multiple angle combinations as the optimal angle combination. The angle of the bottom fan corresponding to the optimal angle combination is taken as the angle between the fan axis of the bottom fan and the horizontal plane. The angle of the outer fan corresponding to the optimal angle combination is taken as the angle between the fan axis of the outer fan and the horizontal plane.

2. The heat dissipation method for a dry-type rectifier transformer as described in claim 1, characterized in that, The training method for the neural network model includes: constructing a dataset and training the neural network model using the data in the dataset. The method for constructing the dataset includes: Within the first preset angle range, an angle value is taken every second angle step to obtain the various second angles of the bottom fan; within the second angle range, an angle value is taken every second angle step to obtain the various second angles of the outer fan; the second angle step is greater than the first angle step. The second angle of the bottom fan is combined with the second angle of the outer fan in pairs to obtain multiple test angle combinations; The angles between the bottom fan axis and the horizontal plane, and the angles between the outer fan axis and the horizontal plane, were set according to the angles in each test angle combination to dissipate heat from the dry-type rectifier transformer. The average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil of the dry-type rectifier transformer were measured. The entropy weight method was then used to perform a weighted summation of the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil to obtain the comprehensive temperature rise corresponding to each test angle combination. The test angle combinations and their corresponding overall temperature rise are used as sample data in the dataset.

3. The heat dissipation method for a dry-type rectifier transformer as described in claim 2, characterized in that, The dataset construction method also includes using 80% of the sample data in the dataset as the training set and 20% of the sample data as the test set.

4. The heat dissipation method for a dry-type rectifier transformer as described in claim 2, characterized in that, The entropy weight method is used to perform a weighted summation of the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil, including: Construct a sample set, in which the samples in the sample set are the test angle combinations and the corresponding average temperature rise of the high voltage coil, the temperature rise of the lower low voltage coil, and the temperature rise of the upper low voltage coil; The average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil corresponding to each test angle combination are standardized. Calculate the entropy value of each temperature rise index using the following expression: ； ； The weights of each temperature rise index are calculated using the following expression: ； The numerical value T of the comprehensive temperature rise index is calculated using the following expression: ； In the above formulas, , and These represent the values ​​of the first, second, and third temperature rise indicators, respectively. Let j be the standardized value of the temperature rise index corresponding to the i-th sample. Let represent the entropy value of the j-th temperature rise index, and n represent the number of samples in the sample set. Represents the natural logarithm. This represents the standardized percentage of the j-th temperature rise index for the i-th sample; This represents the weight of the j-th temperature rise index. It is a constant.

5. The heat dissipation method for a dry-type rectifier transformer as described in claim 4, characterized in that, When standardizing the average temperature rise of the high-voltage coil, the temperature rise of the lower low-voltage coil, and the temperature rise of the upper low-voltage coil for each test angle combination, the corresponding calculation expressions are as follows: ； In the formula, For the i-th sample, the j-th temperature rise index is... This represents the maximum value of the j-th temperature rise index in the sample set. This represents the minimum value of the j-th temperature rise index in the sample set.

6. The heat dissipation method for a dry-type rectifier transformer as described in claim 1, characterized in that, Both the first preset angle range and the second preset angle range are 20 degrees to 70 degrees.

7. The heat dissipation method for a dry-type rectifier transformer as described in claim 2, characterized in that, The first angle step size is 1 degree.

8. The heat dissipation method for a dry-type rectifier transformer as described in claim 2, characterized in that, The second angle step size is 10 degrees.

9. The heat dissipation method for a dry-type rectifier transformer as described in any one of claims 1 to 8, characterized in that, When cooling the dry-type rectifier transformer, the fan speed of the bottom fan and the fan speed of the outer fan are both the rated speed.

10. A heat dissipation system for a dry-type rectifier transformer, characterized in that, include: Bottom fan, installed at the bottom air duct inlet of the dry-type rectifier transformer, is used to blow the lower coil and core of the dry-type rectifier transformer at an angle upward. An external fan is installed on the outer middle of the coil of a dry-type rectifier transformer to blow the upper coil of the dry-type rectifier transformer obliquely upward. A controller is connected to the bottom fan and the outer fan to perform the heat dissipation method for the dry-type rectifier transformer as described in any one of claims 1 to 9.

Images