Heat dissipation optimization method and system for power devices based on heat flow path reconstruction

Abstract

The present invention discloses a method and system for optimizing the heat dissipation of power devices based on the reconstruction of the heat flow path, aiming to improve the heat dissipation efficiency of power devices during operation and reduce the adverse effects of thermal stress on the device performance and lifespan. The method includes the following steps: First, obtain the real-time temperature field distribution data of the target power module during the operation of the power device, and determine the initial heat flow path of the module based on this data; Subsequently, conduct a thermal sensitivity analysis on each component in the target power module; Identify the heat flow shielding area according to the thermal sensitivity data, and on this basis, reconstruct the initial heat flow path to generate an optimized heat flow path; Finally, construct an efficient heat dissipation channel according to the optimized heat flow path, thereby realizing the effective heat dissipation optimization of the power device. The present invention can dynamically identify thermal bottlenecks and actively optimize the heat dissipation path, and is applicable to the thermal management design of high-power density electronic devices, with high engineering practical value.

Description

Technical Field

[0001] The present invention relates to the technical field of power device heat dissipation, and particularly relates to a power device heat dissipation optimization method and system based on the reconstruction of the heat flow path. Background Art

[0002] With the continuous development of power electronics technology, power devices have been widely used in fields such as electric vehicles, high-speed trains, new energy power generation, and industrial automation. A large amount of heat is generated during the operation of power devices. If the heat cannot be dissipated in a timely and effective manner, it will not only cause the working temperature of the device to rise and the performance to decline, but may also lead to thermal runaway or even damage. Therefore, improving the heat dissipation efficiency of power devices is one of the key issues to ensure their reliability and extend their service life.

[0003] Currently, the common heat dissipation designs of power devices mainly rely on structural presetting and material optimization, such as using high thermal conductivity materials, heat sinks, air cooling or liquid cooling systems, etc. to conduct overall or local thermal management. However, the traditional methods generally have the following problems: First, the design of the heat dissipation path is mostly based on static thermal simulation, which is difficult to adapt to the dynamic change of the heat source distribution during the actual operation of the device; Second, the sensitivity differences of different components to the thermal environment are ignored, which is likely to cause insufficient thermal management of key areas; Third, there is a lack of a dynamic reconstruction mechanism based on the real-time evolution characteristics of the heat flow, and it is impossible to effectively avoid the heat accumulation area, resulting in prominent local hot spot problems.

[0004] In addition, the existing heat dissipation optimization methods mostly adopt a regular layout method, which cannot accurately guide the heat flow to diffuse along the optimal path, resulting in low heat dissipation efficiency in some high heat density areas, seriously restricting the overall thermal management performance of the power module. Therefore, there is an urgent need for a heat dissipation optimization technology that can combine temperature field distribution, thermal sensitivity analysis, and dynamic regulation of the heat flow path to achieve more efficient and intelligent thermal management of power devices. Summary of the Invention

[0005] In order to solve at least one of the above technical problems, the present invention proposes a power device heat dissipation optimization method and system based on the reconstruction of the heat flow path.

[0006] The first aspect of the present invention provides a power device heat dissipation optimization method based on the reconstruction of the heat flow path, including:

[0007] Obtain the real-time temperature field distribution data of the target power module during the operation of each power device, and determine the initial heat flow path of the target power module according to the real-time temperature field distribution data;

[0008] Conduct a thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component;

[0009] Determine the heat flow shielding area of the target power module according to the thermally sensitive data, and reconstruct the initial heat flow path according to the heat flow shielding area to obtain an optimized heat flow path;

[0010] Construct a heat dissipation channel for the power device by the target power module according to the optimized heat flow path, and optimize the heat dissipation of the power device according to the heat dissipation channel.

[0011] In this solution, the real-time temperature field distribution data of the target power module during the operation of each power device is obtained, and the initial heat flow path of the target power module is determined according to the real-time temperature field distribution data. Specifically:

[0012] Based on the thermal imaging device, obtain the thermal imaging image change data of the target power module during the operation of each power device in a preset time period, extract the thermal imaging image change data at a preset frame rate, and construct a thermal imaging frame image;

[0013] Convert the thermal imaging frame image to grayscale, construct a grayscale image, establish the correspondence between the pixel grayscale value and the actual temperature through blackbody radiation calibration, determine the temperature value of each pixel point of each grayscale image according to the correspondence, and construct a temperature matrix;

[0014] Align and integrate the temperature matrix according to the time series, construct the temperature distribution change matrix of the target power module, obtain the real-time temperature field distribution data, and calculate the temperature gradient vector of each pixel point on the surface of the target power module according to the real-time temperature field distribution data;

[0015] Construct a temperature gradient distribution map according to the amplitude and direction of the temperature gradient vector, extract the area where the temperature change rate is higher than the set threshold as the high heat flux density area according to the temperature gradient distribution map, and fuse adjacent high heat flux density areas based on the region growing algorithm to generate a heat flow aggregation zone;

[0016] Extend the path of the heat flow aggregation zone in the direction of the temperature gradient, calculate the direction consistency coefficient of the temperature gradient vector in the extended path, and smooth and optimize the extended path according to the direction consistency coefficient to form the initial heat flow path of the target power module.

[0017] In this solution, perform a thermally sensitive analysis on each component in the target power module to determine the thermally sensitive data of each component. Specifically:

[0018] Obtain the working parameter change time series data of each component in the target power module in a preset time period, and the working parameter change time series data includes the time change sequences of voltage, current, and power parameters;

[0019] Extract the temperature change time series data of the regions where each component is located according to the real-time temperature field change data, synchronize and align the time of the temperature change time series data with the working parameter change time series data, and construct a temperature-working parameter data matrix;

[0020] Perform correlation analysis on the temperature-working parameter data matrix through the Pearson correlation coefficient algorithm, calculate the correlation coefficient of the temperature change for each working parameter change, perform a significance test on the correlation coefficient through the t-test method, and screen out the effective correlation coefficients with a significance level lower than the preset significance threshold;

[0021] Evaluate the influence degree of the temperature change on the component working parameter change according to the absolute value of the effective correlation coefficient, and determine the thermal sensitivity coefficient of each component according to the influence degree to obtain the thermal sensitivity data of each component.

[0022] In this solution, the determination of the heat flow shielding area of the target power module according to the thermal sensitivity data is specifically as follows:

[0023] Perform position mapping on the thermal sensitivity coefficients of the components in the target power module according to the thermal sensitivity data to generate a thermal sensitivity coefficient distribution map;

[0024] Introduce a density clustering algorithm to perform spatial clustering analysis on the thermal sensitivity coefficient distribution map, set the neighborhood radius and minimum number of samples of the density clustering algorithm, and initialize the data points in the thermal sensitivity coefficient distribution map as an unlabeled set;

[0025] Select any data point from the unlabeled set, calculate the number of samples included within its neighborhood radius. If the number of samples is greater than or equal to the minimum number of samples, mark this point as a core point, and traverse all unvisited points within its neighborhood, and recursively merge the adjacent points that meet the core point conditions into the same clustering cluster;

[0026] If the number of samples in the neighborhood is less than the minimum number of samples, mark this point as a noise point and temporarily store it in an independent set. After traversing all unlabeled data points, output multiple thermal sensitivity data clusters composed of core points and density-reachable points;

[0027] Calculate the average value of the thermal sensitivity coefficients within each thermal sensitivity data cluster, and mark the thermal sensitivity data clusters with an average value higher than the preset sensitivity threshold as high thermal sensitivity data clusters;

[0028] According to the geometric center coordinates and boundary range of the high thermal sensitivity data cluster, extract the path segments in the initial heat flow path that intersect with the boundary of the high thermal sensitivity data cluster or have a distance less than the preset safety distance as the path segments to be shielded;

[0029] Calculate the product of the temperature gradient vector corresponding to the path segment to be masked and the average thermal sensitivity coefficient of the high thermal sensitivity data cluster to generate a path masking factor, and mark the path segment to be masked with a path masking factor greater than the preset masking threshold as a high masking priority path segment;

[0030] Based on the position coordinates of the high masking priority path segments, use the region fusion algorithm to merge adjacent high masking priority path segments to generate the boundary of a continuous heat flow masking region;

[0031] Calculate the spatial distance between the temperature gradient vector in the initial heat flow path and the high thermal sensitivity data cluster. If the distance is less than the preset protection distance, include the corresponding area in the heat flow masking region, and finally generate a heat flow masking region covering all high masking priority path segments.

[0032] In this solution, reconstructing the initial heat flow path according to the heat flow masking region to obtain an optimized heat flow path is specifically as follows:

[0033] Extract the initial heat flow path segments that intersect with the heat flow masking region as path adjustment objects according to the boundary coordinates of the heat flow masking region and the node distribution of the initial heat flow path;

[0034] Calculate the equivalent heat dissipation efficiency of each region according to the thermal conductivity distribution data of the circuit board substrate of the target power module and the heat diffusion characteristics of the copper layer coverage area;

[0035] Based on the equivalent heat dissipation efficiency and the distribution of low-temperature regions in the real-time temperature field distribution data, screen out the non-heat flow masking regions with a heat dissipation efficiency higher than the preset efficiency threshold and a temperature gradient amplitude lower than the preset gradient threshold as candidate guiding regions;

[0036] Determine the optimal heat flow guiding region of the path adjustment object according to the spatial distance between the candidate guiding region and the path adjustment object, combined with the consistency coefficient of the temperature gradient direction;

[0037] Based on the geometric center coordinates of the optimal heat flow guiding region, offset the nodes of the initial heat flow path intersecting with the heat flow masking region in the heat flow diffusion direction to generate an optimized heat flow path that bypasses the heat flow masking region and extends along the copper layer coverage area.

[0038] In this solution, constructing a heat dissipation channel for the power device of the target power module according to the optimized heat flow path and optimizing the heat dissipation of the power device according to the heat dissipation channel is specifically as follows:

[0039] Compare the initial heat flow path with the optimized heat flow path, judge the difference in the heat flow path of the target power module, and determine the heat flow change direction according to the heat flow path difference;

[0040] Determine the heat dissipation channel of the target power module for the power device according to the changed direction of the heat flow, determine the installation position of the heat dissipation device of the target power module according to the heat dissipation channel, and the heat dissipation device includes a heat dissipation micro copper tube, a heat sink, and a heat dissipation coating material;

[0041] Optimize the heat dissipation of the power device according to the heat dissipation device.

[0042] The second aspect of the present invention also provides a power device heat dissipation optimization system based on heat flow path reconstruction, which includes: a memory and a processor. The memory includes a power device heat dissipation optimization method program based on heat flow path reconstruction. When the power device heat dissipation optimization method program based on heat flow path reconstruction is executed by the processor, the following steps are realized:

[0043] Obtain the real-time temperature field distribution data of the target power module during the operation of each power device, and determine the initial heat flow path of the target power module according to the real-time temperature field distribution data;

[0044] Conduct a thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component;

[0045] Determine the heat flow shielding area of the target power module according to the thermal sensitivity data, and reconstruct the initial heat flow path according to the heat flow shielding area to obtain an optimized heat flow path;

[0046] Construct a heat dissipation channel of the target power module for the power device according to the optimized heat flow path, and optimize the heat dissipation of the power device according to the heat dissipation channel.

[0047] The present invention discloses a power device heat dissipation optimization method and system based on heat flow path reconstruction, aiming to improve the heat dissipation efficiency of power devices during operation and reduce the adverse effects of thermal stress on device performance and lifespan. The method includes the following steps: First, obtain the real-time temperature field distribution data of the target power module during the operation of the power device, and determine the initial heat flow path of the module based on this data; subsequently, conduct a thermal sensitivity analysis on each component in the target power module; identify the heat flow shielding area according to the thermal sensitivity data, and on this basis, reconstruct the initial heat flow path to generate an optimized heat flow path; finally, construct an efficient heat dissipation channel according to the optimized heat flow path, thereby realizing effective heat dissipation optimization of the power device. The present invention can dynamically identify thermal bottlenecks and actively optimize the heat dissipation path, is applicable to the thermal management design of high-power density electronic devices, and has high engineering practical value. Description of the Drawings

[0048] Figure 1 Shows a flowchart of a power device heat dissipation optimization method based on heat flow path reconstruction according to the present invention;

[0049] Figure 2 The flowchart of the present invention for determining the thermal sensitivity data of each component is shown;

[0050] Figure 3 The flowchart of the present invention for optimizing the heat dissipation of power devices is shown;

[0051] Figure 4 The block diagram of a power device heat dissipation optimization system based on thermal flow path reconstruction according to the present invention is shown. Detailed implementation manners

[0052] In order to more clearly understand the above objects, features and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific implementation manners. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other.

[0053] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited by the specific embodiments disclosed below.

[0054] Figure 1 The flowchart of a power device heat dissipation optimization method based on thermal flow path reconstruction according to the present invention is shown.

[0055] As Figure 1 shown, in the first aspect of the present invention, a power device heat dissipation optimization method based on thermal flow path reconstruction is provided, including:

[0056] S102, obtaining the real-time temperature field distribution data of the target power module during the operation of each power device, and determining the initial thermal flow path of the target power module according to the real-time temperature field distribution data;

[0057] S104, performing thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component;

[0058] S106, determining the thermal flow shielding area of the target power module according to the thermal sensitivity data, and reconstructing the initial thermal flow path according to the thermal flow shielding area to obtain an optimized thermal flow path;

[0059] S108, constructing a heat dissipation channel for the power device by the target power module according to the optimized thermal flow path, and optimizing the heat dissipation of the power device according to the heat dissipation channel.

[0060] It should be noted that by obtaining the real-time temperature field distribution data of the target power module during the operation of the power device and determining the initial heat flow path of the target power module according to the real-time temperature field distribution data, the actual conduction and aggregation trends of heat in the power module can be comprehensively revealed. Secondly, by implementing thermal sensitivity analysis, the response intensity of each component to temperature changes can be mined. Through the correlation mining with working electrical parameters, the coupling judgment from "thermal" to "electrical" behavior can be realized, and targeted thermal sensitivity data can be formed, so as to identify the key components that are most sensitive to temperature rise and need key protection. Subsequently, based on the thermal sensitivity data, the heat flow shielding area is further identified, and the heat risk concentration area is accurately delimited by means of clustering and geometric analysis, which can effectively prevent the high-temperature path from continuing to pass through these sensitive areas to reduce the damage of local thermal shock to the reliability of the device. Then, by combining the module structure and the difference in thermal conductivity, the initial heat flow path is reconstructed, which can not only achieve intelligent avoidance of high-thermal-sensitivity areas, but also guide the heat flow to preferentially spread along the high-thermal-conductivity efficiency area, improve the overall heat diffusion rate, and then enhance the heat transfer ability. Finally, according to the optimized heat flow path, the heat dissipation channels and heat dissipation devices (such as heat sinks, microchannel cooling structures, thermal conductive coatings, etc.) are reasonably arranged, which can realize the optimal deployment of the thermal management hardware configuration, so that the heat dissipation structure not only conforms to the heat flow direction in space, but also significantly improves the heat dissipation efficiency of the power device in function, and finally realizes the reduction of the device operating temperature and the mitigation of thermal stress.

[0061] According to an embodiment of the present invention, the obtaining of the real-time temperature field distribution data of the target power module during the operation of each power device and the determination of the initial heat flow path of the target power module according to the real-time temperature field distribution data are specifically as follows:

[0062] Based on the thermal imaging device, the thermal imaging image change data of the target power module during the operation of each power device in a preset time period is obtained in real time, and the thermal imaging image change data is extracted at a preset frame rate to construct a thermal imaging frame image;

[0063] The thermal imaging frame image is subjected to gray conversion to construct a gray image, the corresponding relationship between the pixel gray value and the actual temperature is established through blackbody radiation calibration, and the temperature value of each pixel point of each gray image is determined according to the corresponding relationship to construct a temperature matrix;

[0064] The temperature matrix is aligned and integrated according to the time series to construct a temperature distribution change matrix of the target power module, and the real-time temperature field distribution data is obtained. According to the real-time temperature field distribution data, the temperature gradient vector of each pixel point on the surface of the target power module is calculated;

[0065] Construct a temperature gradient distribution map according to the amplitude and direction of the temperature gradient vector, extract the area where the temperature change rate is higher than the set threshold as the high heat flux density area according to the temperature gradient distribution map, and fuse adjacent high heat flux density areas based on the region growing algorithm to generate a heat flux aggregation zone;

[0066] Extend the path of the heat flux aggregation zone in the direction of the temperature gradient, calculate the direction consistency coefficient of the temperature gradient vectors in the extended path, and smooth and optimize the extended path according to the direction consistency coefficient to form the initial heat flux path of the target power module.

[0067] It should be noted that by obtaining the thermal imaging image change data of the target power module during the operation of each power device, the thermal imaging image is first calibrated through gray conversion and blackbody radiation to accurately obtain the temperature value of each pixel point, and a temperature matrix with high spatio-temporal resolution is constructed; subsequently, a temperature distribution change matrix in the time series is constructed based on the temperature matrix, and the temperature gradient vector of each pixel point in the temperature field is obtained through gradient calculation, so as to reveal the main direction and intensity of temperature change. On this basis, the area where the temperature change rate exceeds the set threshold is extracted as the high heat flux density area through the temperature gradient distribution map, and the region growing algorithm is used to fuse adjacent heat flux regions to form a heat flux aggregation zone, accurately capturing the "trunk" structure of heat concentrated conduction. Further, the path of the heat flux aggregation zone is extended in the direction of the temperature gradient, and the direction consistency coefficient of the gradient vectors in the path is calculated to ensure the continuity and rationality of the heat flux direction, and finally an initial heat flux path with clear physical meaning and representativeness is formed, effectively revealing the main direction and bottleneck position of heat conduction in the module; the target power module is jointly constructed by a number of power devices.

[0068] Figure 2 The flowchart of determining the thermal sensitivity data of each component according to the present invention is shown.

[0069] According to an embodiment of the present invention, the thermal sensitivity analysis of each component in the target power module to determine the thermal sensitivity data of each component is specifically as follows:

[0070] S202, obtain the working parameter change time series data of each component in the target power module during a preset time period, and the working parameter change time series data includes the time change sequences of voltage, current and power parameters;

[0071] S204, extract the temperature change time series data of the area where each component is located according to the real-time temperature field change data, synchronize and time-align the temperature change time series data with the working parameter change time series data, and construct a temperature - working parameter data matrix;

[0072] S206. Perform a correlation analysis on the temperature - operating parameter data matrix using the Pearson correlation coefficient algorithm, calculate the correlation coefficient of the temperature change with respect to the change in each operating parameter, and perform a significance test on the correlation coefficient using the t - test method to screen out the effective correlation coefficients with a significance level lower than the preset significance threshold.

[0073] S208. Evaluate the influence degree of the temperature change on the change in the operating parameters of the components according to the absolute value of the effective correlation coefficient, and determine the thermal sensitivity coefficient of each component based on the influence degree to obtain the thermal sensitivity data of each component.

[0074] It should be noted that by synchronously collecting the time - series data of the temperature change and the time - series data of the operating parameter change of each component in the target power module, constructing an association matrix of temperature, voltage, current, and power parameters, and performing a significance analysis based on the Pearson correlation coefficient algorithm and statistical test methods, the influence degree of temperature fluctuations on the electrical performance of components can be accurately quantified. By establishing a dynamic association model between the thermal field change and the electrical parameter fluctuation, the key components that respond strongly to temperature changes can be effectively identified. Based on the effective correlation coefficients screened by statistical significance, the performance stability differences of different components under temperature rise conditions can be objectively evaluated, providing a data basis for the hierarchical protection of thermal - sensitive components. By calculating the thermal sensitivity coefficient and establishing a quantitative evaluation system, it is possible to prioritize the implementation of key thermal protection for high - sensitive components, significantly improving the pertinence and effectiveness of the optimization of the heat dissipation system. The significance test (t - test) determines whether the correlation coefficient is significant by calculating the p - value (significance level). If the p - value is higher than the set threshold (such as 0.05), it indicates that the correlation may not be significant, and such unreliable results need to be excluded. The higher the influence degree, the higher the thermal sensitivity coefficient, indicating that the increase in temperature has a greater impact on the change in the operating parameters of the component, that is, the change in temperature has a greater impact on the operating stability and efficiency of the component.

[0075] According to an embodiment of the present invention, the determination of the heat - flow shielding area of the target power module according to the thermal sensitivity data is specifically as follows:

[0076] Perform a position mapping on the thermal sensitivity coefficients of each component in the target power module according to the thermal sensitivity data to generate a thermal sensitivity coefficient distribution map.

[0077] Introduce a density - based spatial clustering of applications with noise (DBSCAN) algorithm to perform spatial clustering analysis on the thermal sensitivity coefficient distribution map, set the neighborhood radius and minimum sample number of the DBSCAN algorithm, and initialize the data points in the thermal sensitivity coefficient distribution map as an unlabeled set.

[0078] Select any data point from the unlabeled set, calculate the number of samples contained within its neighborhood radius. If the number of samples is greater than or equal to the minimum number of samples, mark this point as a core point, and traverse all unvisited points within its neighborhood, recursively merging adjacent points that meet the core point criteria into the same clustering cluster;

[0079] If the number of samples within the neighborhood is less than the minimum number of samples, mark this point as a noise point and temporarily store it in a separate set. After traversing all unlabeled data points, output multiple heat-sensitive data clusters composed of core points and density-reachable points;

[0080] Calculate the average value of the heat-sensitive coefficients within each heat-sensitive data cluster, and mark the heat-sensitive data clusters with an average value higher than the preset sensitivity threshold as high-heat-sensitivity data clusters;

[0081] According to the geometric center coordinates and boundary range of the high-heat-sensitivity data cluster, extract the path segments in the initial heat flow path that intersect with the boundary of the high-heat-sensitivity data cluster or have a distance less than the preset safety distance as the path segments to be shielded;

[0082] Calculate the product of the temperature gradient vector corresponding to the path segment to be shielded and the average value of the heat-sensitive coefficients of the high-heat-sensitivity data cluster to generate a path shielding factor, and mark the path segments to be shielded with a path shielding factor greater than the preset shielding threshold as high-shielding-priority path segments;

[0083] Based on the position coordinates of the high-shielding-priority path segments, use the region fusion algorithm to merge adjacent high-shielding-priority path segments to generate the boundary of a continuous heat flow shielding region;

[0084] Calculate the spatial distance between the temperature gradient vector in the initial heat flow path and the high-heat-sensitivity data cluster. If the distance is less than the preset protection distance, include the corresponding area in the heat flow shielding region, and finally generate a heat flow shielding region covering all high-shielding-priority path segments.

[0085] It should be noted that, due to the different sensitivities of the components in the target power module to temperature rise during operation, in order to avoid the heat generated during the operation of the power device from affecting the components sensitive to temperature rise, a density clustering algorithm is introduced to perform spatial clustering analysis on the thermal sensitivity data of the target power module. By mapping the spatial positions of the thermal sensitivity coefficients of the components, a thermal sensitivity coefficient distribution map is constructed, and density clustering is performed on the map using the set neighborhood radius and minimum sample number, which can effectively identify the regions where the thermal sensitivity is significantly concentrated, namely the high thermal sensitivity data clusters. Furthermore, by calculating the average thermal sensitivity coefficients of these clusters and combining their geometric boundary information, the path segments that intersect or are adjacent to the initial heat flow path are identified as the objects to be shielded, and a thermal risk quantification assessment is performed through the path shielding factor to screen out the path segments with high shielding priority. Subsequently, a region fusion algorithm is used to merge these path segments to construct a continuous and closed heat flow shielding boundary region, and a protection distance standard is introduced to further expand the shielding range to ensure that the heat flow path is fully isolated from the sensitive region, significantly reducing the thermal shock risk in the high thermal sensitivity region and improving the operating stability of the target power module. The path shielding factor refers to an index used to quantify the thermal interference intensity between the heat flow path and the high thermal sensitivity region, and its value is calculated by the product of the temperature gradient vector corresponding to the path segment and the average thermal sensitivity coefficient of the intersecting high thermal sensitivity data cluster, reflecting the strength of the potential thermal shock of the path segment to the sensitive region.

[0086] According to an embodiment of the present invention, reconstructing the initial heat flow path according to the heat flow shielding region to obtain an optimized heat flow path specifically includes:

[0087] Extracting the initial heat flow path segments that intersect with the heat flow shielding region as the path adjustment objects according to the boundary coordinates of the heat flow shielding region and the node distribution of the initial heat flow path;

[0088] Calculating the equivalent heat dissipation efficiency of each region according to the heat conductivity distribution data of the circuit board substrate of the target power module and the heat diffusion characteristics of the copper layer coverage area;

[0089] Based on the equivalent heat dissipation efficiency and the low temperature region distribution in the real-time temperature field distribution data, screening out the non-heat flow shielding regions with heat dissipation efficiency higher than the preset efficiency threshold and temperature gradient amplitude lower than the preset gradient threshold as the candidate guiding regions;

[0090] Determining the optimal heat flow guiding region of the path adjustment object according to the spatial distance between the candidate guiding region and the path adjustment object and the consistency coefficient of the temperature gradient direction;

[0091] Based on the geometric center coordinates of the optimal heat flow guiding region, the initial heat flow path intersecting with the heat flow shielding region is subjected to path node offset in the heat flow diffusion direction to generate an optimized heat flow path that bypasses the heat flow shielding region and extends along the copper layer coverage region.

[0092] It should be noted that by reconstructing the path segments of the intersection part between the initial heat flow path and the heat flow shielding region, the local overheating risk brought by the high-temperature sensitive region can be effectively avoided, and the thermal safety of the overall heat dissipation channel can be improved. Specifically, by utilizing the thermal conductivity distribution of different regions of the circuit board and the heat diffusion ability of the copper layer, combined with the distribution of low-temperature regions and high heat dissipation efficiency regions in the real-time temperature field, a region with good thermal conductivity and low thermal load is scientifically selected as the heat flow guiding direction. At the same time, through the evaluation of the temperature gradient direction consistency coefficient, the path adjustment process not only considers the path with the minimum thermal resistance, but also ensures the physical continuity and stability of the heat flow guiding. On this basis, the heat flow path is smoothly adjusted by the node offset method, so that the optimized heat flow path can efficiently bypass the sensitive region and preferentially extend along the copper layer region with strong heat diffusion ability, and finally achieve the collaborative optimization effect of uniform heat flow guiding, local hot spot mitigation and overall heat dissipation performance improvement. Path node offset refers to the position adjustment of the path nodes close to or passing through the heat flow shielding region in the original path in the heat flow diffusion direction during the heat flow path reconstruction process.

[0093] Figure 3 The flowchart of optimizing the heat dissipation of the power device according to the present invention is shown.

[0094] According to an embodiment of the present invention, the heat dissipation channel of the power device is constructed according to the optimized heat flow path, and the heat dissipation of the power device is optimized according to the heat dissipation channel, specifically:

[0095] S302, comparing the initial heat flow path with the optimized heat flow path, judging the heat flow path difference of the target power module, and determining the heat flow change direction according to the heat flow path difference;

[0096] S304, determining the heat dissipation channel of the target power module for the power device according to the heat flow change direction, and determining the installation position of the heat dissipation device of the target power module according to the heat dissipation channel. The heat dissipation device includes a heat dissipation micro copper tube, a heat sink, and a heat dissipation coating material;

[0097] S306, optimizing the heat dissipation of the power device according to the heat dissipation device.

[0098] It should be noted that by precisely analyzing the changing direction of the heat flow path, the optimal layout area of the heat dissipation device can be dynamically identified to ensure a high degree of matching between the heat dissipation channel and the optimized heat flow direction. Secondly, based on the heat flow path differences, the selection and layout of the heat dissipation device are guided, enabling devices such as heat dissipation micro copper tubes, heat sinks, and heat dissipation coating materials to be accurately deployed on the key heat flow paths to achieve the optimal allocation of heat dissipation resources.

[0099] According to an embodiment of the present invention, it further includes:

[0100] Real-time monitor the current, voltage, and switching frequency parameters of each load loop in the target power module, establish a mapping relationship between the load parameter time series data and the temperature field change rate through a long short-term memory neural network, and generate a load change rate prediction result;

[0101] Analyze the thermal conductivity difference distribution between the copper layer and the insulating substrate in the multi-layer PCB board according to the load change rate prediction result, extract the continuous area of the copper layer as the priority heat conduction channel, and mark the heat flow impedance coefficient at the position of the via;

[0102] Based on the thermal conductivity difference distribution and the heat flow impedance coefficient, construct a heat conduction topology network of the multi-layer PCB board, input the load change rate prediction result into the heat conduction topology network for transient thermal simulation, and pre-generate a multi-level heat conduction compensation path set including the planar heat conduction path of the copper layer and the vertical heat conduction path of the via;

[0103] Verify the effectiveness of the multi-level heat conduction compensation path set according to the temperature field distribution data collected in real time. When it is detected that the deviation between the actual heat flow direction and the predicted path exceeds the tolerance threshold, switch to the compensation path with the highest thermal conductivity matching degree in the adjacent layer copper layer, and optimize the impedance matching of the cross-layer heat flow channel by adjusting the via current density.

[0104] The switching to the compensation path with the highest thermal conductivity matching degree in the adjacent layer copper layer is specifically:

[0105] Calculate the equivalent thermal conductivity attenuation coefficient of each layer of copper foil according to the temperature rise gradient distribution of the copper layer in the current heat flow path, and establish a comprehensive efficiency evaluation model for the cross-layer heat flow path in combination with the vertical thermal resistance parameters of the vias in the heat conduction topology network;

[0106] Based on the comprehensive efficiency evaluation model, iteratively screen the multi-level heat conduction compensation path set, eliminate the via connection paths with abnormal spacing caused by the thermal expansion of the insulating substrate, and retain the candidate paths that meet the copper layer heat flow continuity constraint;

[0107] According to the high-frequency load fluctuation component in the load change rate prediction result, preferentially select the redundant heat conduction path with a parallel via structure as the heat flow switching buffer zone, and verify the local hot spot generation probability during the switching process through electromagnetic-thermal coupling field simulation;

[0108] When the thermal sensitivity coefficient of the target copper layer region is detected to exceed the preset safety threshold, a distributed current injection mechanism of the via array is triggered to reduce the contact thermal resistance. At the same time, the heat in the original heat flow path is shunted to the low-temperature region of the adjacent copper layer through the parallel via structure, realizing the collaborative reconstruction of the multi-layer heat flow path at the millisecond level.

[0109] It should be noted that in dynamic load scenarios such as frequency converters and electric vehicle controllers, the rapid load switching of power modules causes the heat flow path to shift frequently. However, due to the significant difference in thermal conductivity between the internal copper layers and insulating substrates of multi-layer PCBs, traditional thermal management solutions are difficult to track the dynamic changes of heat flow in real time, resulting in problems such as delays in the reconstruction of cross-layer heat conduction paths and local hot spot aggregation, seriously affecting the device life and system reliability. The present invention uses a long short-term memory neural network (LSTM) to predict the load change rate in real time, constructs a multi-layer heat conduction topology network by combining the thermal conductivity distribution of the continuous copper layer region and the heat flow impedance coefficient of vias, pre-generates a set of multi-level heat conduction compensation paths, and dynamically verifies the path effectiveness based on real-time temperature field data. When the deviation of the actual heat flow direction is detected, by calculating the equivalent thermal conductivity attenuation coefficient of the copper layer and the vertical thermal resistance parameter, redundant parallel via paths are selected as the switching buffer area, and a distributed current injection mechanism is triggered to reduce the contact thermal resistance, realizing the collaborative reconstruction of the cross-layer heat flow path at the millisecond level. This technology can significantly improve the heat dissipation response speed of multi-layer PCBs under dynamic loads, accurately shunt heat to the low-temperature regions of adjacent copper layers, suppress the via deformation caused by the thermal expansion of the insulating substrate, reduce the probability of local hot spot generation by more than 30%, and at the same time avoid eddy current losses during path switching through electromagnetic-thermal coupling optimization, ultimately extending the life of power modules and ensuring the operation stability in high-frequency load scenarios.

[0110] Figure 4 The block diagram of a power device heat dissipation optimization system based on heat flow path reconstruction according to the present invention is shown.

[0111] In a second aspect of the present invention, a power device heat dissipation optimization system 4 based on heat flow path reconstruction is further provided. The system includes: a memory 41 and a processor 42. The memory includes a program for the power device heat dissipation optimization method based on heat flow path reconstruction. When the program for the power device heat dissipation optimization method based on heat flow path reconstruction is executed by the processor, the following steps are realized:

[0112] Obtain the real-time temperature field distribution data of the target power module during the operation of each power device, and determine the initial heat flow path of the target power module according to the real-time temperature field distribution data;

[0113] Conduct a thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component;

[0114] Determine the heat flow shielding area of the target power module according to the thermally sensitive data, and reconstruct the initial heat flow path according to the heat flow shielding area to obtain an optimized heat flow path;

[0115] Construct a heat dissipation channel for the power device by the target power module according to the optimized heat flow path, and optimize the heat dissipation of the power device according to the heat dissipation channel.

[0116] The present invention discloses a method and system for optimizing the heat dissipation of a power device based on the reconstruction of a heat flow path, aiming to improve the heat dissipation efficiency of the power device during operation and reduce the adverse effects of thermal stress on the device performance and lifespan. The method includes the following steps: First, obtain the real-time temperature field distribution data of the target power module during the operation of the power device, and determine the initial heat flow path of the module based on this data; Subsequently, perform a thermally sensitive analysis on each component in the target power module; Identify the heat flow shielding area according to the thermally sensitive data, and on this basis, reconstruct the initial heat flow path to generate an optimized heat flow path; Finally, construct an efficient heat dissipation channel according to the optimized heat flow path, thereby realizing the effective heat dissipation optimization of the power device. The present invention can dynamically identify heat bottlenecks and actively optimize the heat dissipation path, is applicable to the thermal management design of high-power density electronic devices, and has high engineering practical value.

[0117] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined, or can be integrated into another system, or some features can be ignored, or not executed. In addition, the couplings, direct couplings, or communication connections between the various components shown or discussed can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be electrical, mechanical, or other forms.

[0118] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units; they can be located in one place or distributed to multiple network units; Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0119] In addition, in each embodiment of the present invention, the various functional units can all be integrated in one processing unit, or each unit can be separately used as one unit, or two or more units can be integrated in one unit; The above integrated units can be implemented in the form of hardware, or in the form of hardware plus software functional units.

[0120] Those of ordinary skill in the art can understand that all or part of the steps to implement the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including those of the above method embodiments. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0121] Alternatively, if the above integrated units of the present invention are implemented in the form of software function modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the embodiments of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the various embodiments of the present invention. The foregoing storage medium includes various media that can store program codes, such as removable storage devices, ROM, RAM, magnetic disks, or optical discs.

[0122] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A heat dissipation optimization method for power devices based on heat flow path reconstruction, characterized in that Including the following steps: Obtain the real-time temperature field distribution data of the target power module during the operation of each power device, and determine the initial heat flow path of the target power module according to the real-time temperature field distribution data; Conduct a thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component; Determine the heat flow shielding area of the target power module according to the thermal sensitivity data, and reconstruct the initial heat flow path according to the heat flow shielding area to obtain an optimized heat flow path; Construct a heat dissipation channel for the power device by the target power module according to the optimized heat flow path, and optimize the heat dissipation of the power device according to the heat dissipation channel; The determining the heat flow shielding area of the target power module according to the thermal sensitivity data is specifically: Perform position mapping on the thermal sensitivity coefficients of each component in the target power module according to the thermal sensitivity data to generate a thermal sensitivity coefficient distribution map; Introduce a density clustering algorithm to perform spatial clustering analysis on the thermal sensitivity coefficient distribution map, set the neighborhood radius and minimum sample number of the density clustering algorithm, and initialize the data points in the thermal sensitivity coefficient distribution map as an unlabeled set; Select any data point from the unlabeled set, calculate the number of samples included within its neighborhood radius. If the number of samples is greater than or equal to the minimum sample number, mark this point as a core point, and traverse all unvisited points within its neighborhood, and recursively merge the adjacent points that meet the core point conditions into the same clustering cluster; If the number of samples within the neighborhood is less than the minimum sample number, mark this point as a noise point and temporarily store it in an independent set. After traversing all unlabeled data points, output multiple thermal sensitivity data clusters composed of core points and density-reachable points; Calculate the average value of the thermal sensitivity coefficients within each thermal sensitivity data cluster, and mark the thermal sensitivity data cluster with an average value higher than the preset sensitivity threshold as a high thermal sensitivity data cluster; According to the geometric center coordinates and boundary range of the high thermal sensitivity data cluster, extract the path segments in the initial heat flow path that intersect with the boundary of the high thermal sensitivity data cluster or have a distance less than the preset safety distance as the path segments to be shielded; Calculate the product of the temperature gradient vector corresponding to the path segment to be shielded and the average value of the thermal sensitivity coefficients of the high thermal sensitivity data cluster to generate a path shielding factor, and mark the path segment to be shielded with a path shielding factor greater than the preset shielding threshold as a high shielding priority path segment; Based on the position coordinates of the high shielding priority path segments, use a region fusion algorithm to merge adjacent high shielding priority path segments to generate a continuous boundary of the heat flow shielding area; Calculate the spatial distance between the temperature gradient vector and the thermal sensitivity data cluster in the initial heat flow path. If the distance is less than the preset protection distance, include the corresponding area in the heat flow shielding area, and finally generate a heat flow shielding area covering all high shielding priority path segments.

2. The heat dissipation optimization method for a power device based on heat flow path reconstruction according to claim 1, wherein The obtaining the real-time temperature field distribution data of the target power module during the operation of each power device, and determining the initial heat flow path of the target power module according to the real-time temperature field distribution data is specifically: Based on the thermal imaging device, real-time acquisition of the thermal imaging image change data of the target power module during the operation of each power device in a preset time period, extracting the thermal imaging image change data at a preset frame rate, and constructing a thermal imaging frame image; Performing gray-scale conversion on the thermal imaging frame image, constructing a gray-scale image, establishing the correspondence between the pixel gray-scale value and the actual temperature through blackbody radiation calibration, determining the temperature value of each pixel point of each gray-scale image according to the correspondence, and constructing a temperature matrix; Performing alignment and integration operations on the temperature matrix according to the time series, constructing the temperature distribution change matrix of the target power module, obtaining the real-time temperature field distribution data, and calculating the temperature gradient vector of each pixel point on the surface of the target power module according to the real-time temperature field distribution data; Constructing a temperature gradient distribution map according to the amplitude and direction of the temperature gradient vector, extracting the region where the temperature change rate is higher than the set threshold as the high heat flux density region based on the temperature gradient distribution map, and fusing adjacent high heat flux density regions based on the region growing algorithm to generate a heat flux aggregation band; Extending the heat flux aggregation band along the temperature gradient direction, calculating the direction consistency coefficient of the temperature gradient vector in the extension path, and smoothing and optimizing the extension path according to the direction consistency coefficient to form the initial heat flux path of the target power module.

3. A heat dissipation optimization method for a power device based on heat flow path reconstruction according to claim 1, characterized in that Performing thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component, specifically: Obtaining the working parameter change time series data of each component in the target power module during a preset time period, where the working parameter change time series data includes the time change sequences of voltage, current, and power parameters; Extracting the temperature change time series data of the regions where each component is located according to the real-time temperature field change data, synchronizing and aligning the temperature change time series data with the working parameter change time series data, and constructing a temperature - working parameter data matrix; Performing correlation analysis on the temperature - working parameter data matrix through the Pearson correlation coefficient algorithm, calculating the correlation coefficient of the temperature change with the change of each working parameter, performing a significance test on the correlation coefficient through the t - test method, and screening out the effective correlation coefficients with a significance level lower than the preset significance threshold; Evaluating the influence degree of the temperature change on the change of the component working parameters according to the absolute value of the effective correlation coefficient, determining the thermal sensitivity coefficient of each component according to the influence degree, and obtaining the thermal sensitivity data of each component.

4. The heat dissipation optimization method of a power device based on heat flow path reconstruction according to claim 1, wherein Reconstructing the initial heat flux path according to the heat flux shielding region to obtain an optimized heat flux path, specifically: Extracting the initial heat flux path segments that intersect with the heat flux shielding region as the path adjustment objects according to the boundary coordinates of the heat flux shielding region and the node distribution of the initial heat flux path; Calculating the equivalent heat dissipation efficiency of each region according to the thermal conductivity distribution data of the circuit board substrate of the target power module and the heat diffusion characteristics of the copper layer coverage region; Based on the low-temperature region distribution in the equivalent heat dissipation efficiency and the real-time temperature field distribution data, select non-heat flow shielding regions with a heat dissipation efficiency higher than a preset efficiency threshold and a temperature gradient amplitude lower than a preset gradient threshold as candidate guiding regions; According to the spatial distance between the candidate guiding region and the path adjustment object, and in combination with the consistency coefficient of the temperature gradient direction, determine the optimal heat flow guiding region of the path adjustment object; Based on the geometric center coordinates of the optimal heat flow guiding region, offset the path nodes of the initial heat flow path intersecting with the heat flow shielding region in the heat flow diffusion direction to generate an optimized heat flow path that bypasses the heat flow shielding region and extends along the copper layer coverage region.

5. A heat dissipation optimization method for a power device based on heat flow path reconstruction according to claim 1, characterized in that The heat dissipation channel of the power device by the target power module is constructed according to the optimized heat flow path, and the heat dissipation of the power device is optimized according to the heat dissipation channel. Specifically: Compare the initial heat flow path with the optimized heat flow path, judge the heat flow path difference of the target power module, and determine the heat flow change direction according to the heat flow path difference; Determine the heat dissipation channel of the power device by the target power module according to the heat flow change direction, and determine the installation position of the heat dissipation device of the target power module according to the heat dissipation channel. The heat dissipation device includes heat dissipation micro copper tubes, heat sinks, and heat dissipation coating materials; Optimize the heat dissipation of the power device according to the heat dissipation device.

6. A power device heat dissipation optimization system based on heat flow path reconstruction, characterized in that, The power device heat dissipation optimization system based on heat flow path reconstruction includes a memory and a processor. The memory includes a program for the power device heat dissipation optimization method based on heat flow path reconstruction. When the program for the power device heat dissipation optimization method based on heat flow path reconstruction is executed by the processor, the following steps are implemented: Obtain the real-time temperature field distribution data of the target power module during the operation of each power device, and determine the initial heat flow path of the target power module according to the real-time temperature field distribution data; Conduct a thermal sensitivity analysis on each component in the target power module to determine the thermal sensitivity data of each component; Determine the heat flow shielding region of the target power module according to the thermal sensitivity data, and reconstruct the initial heat flow path according to the heat flow shielding region to obtain an optimized heat flow path; Construct a heat dissipation channel for the power device by the target power module according to the optimized heat flow path, and optimize the heat dissipation of the power device according to the heat dissipation channel; The determination of the heat flow shielding region of the target power module according to the thermal sensitivity data is specifically: Perform position mapping on the thermal sensitivity coefficients of each component in the target power module according to the thermal sensitivity data to generate a thermal sensitivity coefficient distribution map; Introduce a density clustering algorithm to perform spatial clustering analysis on the thermal sensitivity coefficient distribution map, set the neighborhood radius and minimum number of samples of the density clustering algorithm, and initialize the data points in the thermal sensitivity coefficient distribution map as an unlabeled set; Select any data point from the unlabeled set, calculate the number of samples included within its neighborhood radius. If the number of samples is greater than or equal to the minimum number of samples, mark this point as a core point, and traverse all unvisited points within its neighborhood, and recursively merge the adjacent points that meet the core point conditions into the same clustering cluster; If the number of samples in the neighborhood is less than the minimum number of samples, mark this point as a noise point and temporarily store it in an independent set. After traversing all unmarked data points, output multiple heat-sensitive data clusters composed of core points and density-reachable points; Calculate the average value of the heat-sensitive coefficients within each heat-sensitive data cluster, and mark the heat-sensitive data clusters with an average value higher than the preset sensitivity threshold as high heat-sensitive data clusters; According to the geometric center coordinates and boundary range of the high heat-sensitive data cluster, extract the path segments in the initial heat flow path that intersect with the boundary of the high heat-sensitive data cluster or have a distance less than the preset safety distance as the path segments to be blocked; Calculate the product of the temperature gradient vector corresponding to the path segment to be blocked and the average value of the heat-sensitive coefficients of the high heat-sensitive data cluster to generate a path blocking factor, and mark the path segments to be blocked with a path blocking factor greater than the preset blocking threshold as high blocking priority path segments; Based on the position coordinates of the high blocking priority path segments, use the region fusion algorithm to merge adjacent high blocking priority path segments to generate the boundary of a continuous heat flow blocking region; Calculate the spatial distance between the temperature gradient vector in the initial heat flow path and the heat-sensitive data cluster. If the distance is less than the preset protection distance, include the corresponding area in the heat flow blocking region, and finally generate a heat flow blocking region covering all high blocking priority path segments.

Images