Thermal resistance measuring probe based on heat flow focusing effect and optimization method thereof

Abstract

The invention belongs to the technical field of thermal resistance measurement, and discloses a thermal resistance measurement probe based on a heat flow focusing effect and an optimization method thereof. The probe is composed of a silicon chip layer, a copper substrate layer and a diamond inverted pyramid frustum layer in sequence, and a heat flow focusing channel is formed by the inclination angle of the side wall of the inverted pyramid frustum layer and used for guiding transverse heat diffusion into longitudinal conduction, so that the heat flux density and distribution uniformity of a contact surface are improved. In order to obtain optimal structural parameters, a multi-objective optimization method based on Taguchi orthogonal design and finite element analysis is adopted to carry out combined optimization on the size and the thickness of the diamond prismatic table and the thickness of the copper substrate. Through simulation verification, the designed probe can realize high-precision measurement of the interface thermal resistance, and the heat flow distribution uniformity and the heat penetration capability are improved compared with those of a traditional structure. The technology can be used for testing the thermal performance of multi-layer structure materials such as a power module solder layer and microelectronic packaging.

Description

Technical Field

[0001] This invention relates to the field of thermal resistance measurement technology, and in particular to a probe for analyzing the thermal properties of materials and its optimization method, which is especially suitable for the accurate measurement of interface thermal resistance of high power density electronic devices. Background Technology

[0002] In high-power electronic systems such as electric vehicles, new energy converters, and industrial drives, power modules such as IGBTs are core components. Within their multi-layered structures, the structural integrity of the solder layer directly determines the heat dissipation performance and long-term reliability of the power module. However, under long-term service conditions involving thermal cycling and mechanical stress, the solder layer is prone to defects such as voids, cracks, and delamination. These defects significantly degrade the heat conduction path, leading to increased chip junction temperature, the formation of localized hot spots, and ultimately, premature module failure. Therefore, accurate detection of solder layer defects is a core requirement for power module manufacturing and quality assessment.

[0003] IGBT failure detection mainly relies on technologies such as infrared thermal imaging, X-rays, and cyclosonography (C-SAM). However, these methods all have significant limitations: infrared thermal imaging requires opening the device, making it a destructive test with insufficient transient response; while X-rays and ultrasound can achieve non-destructive testing, they have poor penetration into high-density or excessively thick materials, which can easily affect the test results.

[0004] To overcome the aforementioned limitations, thermal resistance measurement technology has been introduced as an effective means of evaluating the thermal performance of solder layers. While existing thermal resistance measurement probes possess basic functions, they still have significant shortcomings in practical applications: traditional probe structures fail to effectively control the heat flow path, resulting in uneven heat flow distribution and limited heat flux density, making accurate measurements difficult in high-power-density scenarios. Especially in multi-layer interface thermal resistance analysis, the lateral diffusion of heat flow easily causes measurement deviations, affecting interface identification capabilities.

[0005] While the interface thermal resistance measurement probe designed by Gao Shan's team at Virginia Tech can achieve basic thermal resistance detection and 2D distribution layout, its heat flow focusing relies on passive heat dissipation area reduction. The copper block conduction structure is prone to significant lateral heat diffusion, resulting in uneven heat flow distribution at the contact surface and a larger effective measurement area (the smallest detectable defect is about 2×2mm). 2 Furthermore, parameter design relies heavily on trial and error based on experience, making it difficult to balance heat flow uniformity and heat penetration capability. It also lacks quantitative support and has weak accuracy in separating and identifying interfaces of multi-layer structures. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention aims to provide a thermal resistance measurement probe and its optimization method. Based on thermal simulation and Taguchi orthogonal experimental design, the key structural parameters of the probe are optimized and determined, resulting in a significant improvement in heat flux distribution uniformity and heat flux density, thereby achieving high-precision and high-reliability thermal resistance measurement.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a thermal resistance measurement probe based on the heat flow focusing effect, comprising, arranged sequentially from top to bottom:

[0008] The silicon chip layer serves as a heat source.

[0009] A copper substrate layer, located below the silicon chip layer, is used for support and connection to the measurement equipment;

[0010] The diamond frustum layer is an inverted frustum structure with an upper surface larger than the lower surface. Its upper surface is connected to the copper substrate layer. Its sidewall inclination angle and geometry together form a heat flow focusing channel, which is used to focus lateral heat diffusion into longitudinal heat conduction, thereby improving the heat flux density and distribution uniformity of the contact surface.

[0011] Preferably, the sidewall inclination angle of the diamond frustum layer is 45°.

[0012] Preferably, the lower surface length L2 and width W2 of the diamond frustum layer are determined by the upper surface length L1, upper surface width W1, and frustum thickness t through the following geometric relationship:

[0013] L2 = L1 - 2t

[0014] W2 = W1 - 2t

[0015] Preferably, the silicon chip layer has dimensions of 3mm in length, 5mm in width, and 0.1mm in thickness.

[0016] Preferably, the copper substrate has dimensions of 4.5 mm in length, 7.5 mm in width, and 0.5 mm in thickness.

[0017] Preferably, the upper surface of the diamond frustum layer has dimensions of 1.5 mm in length, 2.5 mm in width, and 0.5 mm in thickness.

[0018] An optimization method for a thermal resistance measurement probe based on the heat flow focusing effect includes:

[0019] Step 1: Based on heat conduction theory and multi-objective optimization principles, determine the key design variables and their levels for the probe;

[0020] Preliminary finite element analysis identified three key variables that significantly affect the probe's thermal performance: the upper surface dimension A of the diamond frustum, the thickness B of the diamond frustum, and the thickness C of the Cu substrate. Three levels were set for each variable, forming a three-factor, three-level experimental system, specifically:

[0021] Variable A: A1 = 1.5mm × 2.5mm, A2 = 1.6mm × 2.7mm, A3 = 1.8mm × 3.0mm;

[0022] Variable B: B1 = 0.3 mm, B2 = 0.4 mm, B3 = 0.5 mm;

[0023] Variable C: C1 = 0.5 mm, C2 = 0.7 mm, C3 = 0.8 mm;

[0024] Step 2, based on the Taguchi orthogonal experimental design method, construct L9(3 3 Nine sets of simulation experiments were conducted using an orthogonal array.

[0025] Step 3: Using ANSYS finite element analysis software, perform steady-state thermal simulation for each group of experiments; in the simulation model, apply 20W of power to the surface of the silicon chip; after simulation, extract two key evaluation indicators:

[0026] The first indicator is the small size characteristic: the difference in heat flux along the horizontal direction on the contact surface between the lower surface of the diamond frustum and the sample to be tested. The smaller this value is, the more uniform the heat flow distribution.

[0027] The second indicator is the large-scale characteristic: the average heat flux density along the Z-axis on the lower surface of the diamond frustum. The larger this value, the stronger the heat penetration capability.

[0028] Step 4: Calculate the signal-to-noise ratio (SNR) of the two indicators for each experimental group, perform SNR response analysis, and predict the optimal combination of levels for each design variable.

[0029] After calculating the SNR of the two indicators separately, a response analysis is performed on the average SNR value of each factor at different levels; for each factor, the level that brings the highest average SNR value is selected, and the optimal combination can be predicted.

[0030] Step 5: For the predicted optimal combination A1B3C1, achieve simultaneous optimization of heat flux distribution uniformity and flux density.

[0031] Preferably, the formula for calculating the signal-to-noise ratio (SNR) is:

[0032] For the small-scale characteristic index, SNR is calculated using the following formula:

[0033] SNR = -10log 10 (Δ 2 )

[0034] The larger the value, the better the heat flow uniformity and the smaller the fluctuation under this combination;

[0035] For the large-scale characteristic index, the SNR calculation formula is:

[0036]

[0037] The larger the value, the higher the heat flux density under that combination;

[0038] In the formula: Δ is the range, and q is the heat flux.

[0039] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0040] 1) The miniature thermal probe provided by this invention, through a three-layer synergistic structure design of silicon chip layer-copper substrate layer-diamond truncated pyramid layer, especially the optimization of the core contour of the diamond layer with a "larger upper and smaller lower truncated pyramid," achieves a synergistic improvement in heat flow distribution uniformity and heat flux density, significantly improving the accuracy and reliability of interface thermal resistance measurement. The diamond truncated pyramid is not a planar layer structure, but rather adopts a stepped convergence design with an upper surface larger than the lower surface: its sidewalls are machined at a 45° angle to form a heat flow focusing channel. This ensures that when the heat flow generated by the heat source is conducted through the copper substrate layer to the diamond truncated pyramid, the heat flow, which would normally diffuse laterally, is directionally guided and gradually converged by the inclined sidewalls of the truncated pyramid, ultimately being concentrated and output as longitudinal heat conduction from the smaller lower surface of the pyramid. This significantly improves the efficiency of longitudinal heat transfer.

[0041] 2) This invention establishes a quantitative evaluation system based on the observability index of separation points, providing a scientific and visual evaluation standard for structural function analysis. The improved probe enhances the ability to identify interface separation points, providing reliable technical support for the accurate analysis of the thermal resistance of multilayer material interfaces. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort.

[0043] Figure 1This is a schematic diagram of the thermal resistance measurement probe structure provided in an embodiment of the present invention. The component marked "1" is the silicon chip layer, which serves as the heat source of the probe and is responsible for conducting heat flow downwards. Its dimensions are 3mm long, 5mm wide, and 0.1mm thick. The component marked "2" is the copper substrate layer, located below the silicon chip layer, which supports the silicon chip layer and conducts heat, while also connecting to the measurement equipment. Its dimensions are 4.5mm long, 7.5mm wide, and 0.5mm thick. The component marked "3" is the diamond frustum layer, a truncated frustum structure with an upper surface larger than the lower surface. Its upper surface connects to the copper substrate layer, and its sidewall inclination angle is 45°. Its core function is to form a heat flow focusing channel, guiding lateral heat diffusion into longitudinal heat conduction. The diagram clearly shows the relative positions, geometric contours, and functional relationships of each layer, reflecting the probe's three-layer functional design logic of "heat source - support and heat conduction - heat flow focusing."

[0044] Figure 2 This is a schematic diagram of the diamond frustum layer structure provided in an embodiment of the present invention. Key geometric parameters of the frustum are highlighted, including the upper surface dimensions (length 1.5mm, width 2.5mm); the frustum thickness, i.e., the vertical distance between the upper and lower surfaces of the frustum, with an optimized value of 0.5mm; and the lower surface dimensions, derived from the geometric relationship between the upper surface dimensions, the frustum thickness, and the 45° sidewall inclination angle. The 45° sidewall inclination angle is also clearly marked in the figure; this angle is a core design feature for achieving directional heat flow focusing and reducing lateral heat diffusion.

[0045] Figure 3 This is a Taguchi orthogonal experimental signal-to-noise ratio (SNR) response table provided in this embodiment of the invention. Two heatmaps visually present the differences in the impact of different factor levels on the SNR of the "small-scale" and "large-scale" characteristics through color gradients and numerical labels. The left heatmap displays the SNR (dB) distribution for different factor level combinations under the "small-scale" characteristic. The vertical axis represents the factors (A, B, C), and the horizontal axis represents the factor levels (1, 2, 3). The value in each cell is the SNR of the corresponding factor-level combination, and the color intensity is positively correlated with the SNR magnitude. For example, factor A at level 3 has an SNR of 3.92 dB (the darkest color), factor B at level 1 has an SNR of 1.97 dB, and factor C at level 2 has an SNR of 3.05 dB. Similarly, the right heatmap displays the SNR (dB) distribution for different factor level combinations under the "large-scale" characteristic, with the same dimensions as the left side; darker colors represent higher SNR.

[0046] Figure 4This is the main effect diagram of signal-to-noise ratio (SNR) provided in this embodiment of the invention. The diagram contains two sub-plots, corresponding to the main effect diagram of SNR for heat flux range and the main effect diagram of SNR for heat flux density, respectively. The horizontal axis of both sub-plots is "factor level," and the vertical axis is "signal-to-noise ratio (SNR) (dB)." The three curves in the diagram represent three key design variables (A: frustum top surface size, B: frustum thickness, C: copper substrate thickness). The curve trends allow for a direct assessment of the impact of different levels of each variable on SNR, thus quickly identifying the optimal level for each variable and providing a more intuitive method for determining the optimal parameter combination (A1B3C1).

[0047] Figure 5 This is a heat flux density curve at the horizontal position of the contact surface between the lower surface of the diamond frustum and the sample under test, provided by the optimal parameter combination (A1: 1.5mm × 2.5mm, B3: 0.5mm, C1: 0.5mm) in this embodiment of the invention. The horizontal axis represents the horizontal position of the contact surface (unit: mm), and the vertical axis represents the heat flux density (unit: W / mm²). 2 The curve shows a generally stable fluctuation trend without obvious peaks or troughs, and the difference (range) between the maximum and minimum heat flux density is only 0.5294 W / mm². 2 The curve visually verifies the uniformity of heat flux distribution and high penetration capability of the optimized probe, directly demonstrating the improvement effect of the optimal parameter combination.

[0048] Figure 6 This chart contains two sub-charts: the top one compares the integral structure function curves, and the bottom one compares the differential structure function curves. The horizontal axis for both is "cumulative thermal resistance ∑Rth (K / W)", and the vertical axes are "cumulative heat capacity ∑Cth (J / K)" and "rate of change of heat capacity dC / dR (J / (K)"). 2 The figure shows two curves: the darker curve represents the traditional probe before improvement, and the lighter curve represents the optimized probe after improvement. Separation points (corresponding to the interface positions of multilayer materials) are marked with dots in the differential structure function graph. Observation shows that the separation points of the improved curves (such as peaks 1, 2, and 5) are sharper, and the difference between the peak intensity and the average value of adjacent points is larger. The observability index of the corresponding separation point reaches 0.4469 (a 1.34-fold improvement from 0.3330 before improvement). This figure clearly demonstrates the improved ability of the improved probe to identify the thermal resistance of multilayer structure interfaces, with clearer interface inflection points, providing a clear visual basis for quantitative analysis of the thermal resistance and heat capacity of each layer. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, other embodiments obtained by those skilled in the art without creative effort are all within the protection scope of this invention.

[0050] Embodiment 1 of the present invention provides a thermal probe for thermal resistance testing, comprising a silicon chip layer 1, a copper substrate layer 2, and a diamond frustum layer 3 stacked sequentially from top to bottom.

[0051] Silicon chip layer 1 serves as a heat source; the chip model is IRG4CH30K.

[0052] A copper substrate layer 2, located below the silicon chip layer 1, is used for support and connection to the measurement equipment;

[0053] The diamond frustum layer 3 has an inverted frustum structure with an upper surface larger than the lower surface. Its upper surface is connected to the copper substrate layer 2. The inclination angle of its sidewalls and its geometry together form a heat flow focusing channel, which is used to focus lateral heat diffusion into longitudinal heat conduction, thereby improving the heat flux density and distribution uniformity of the contact surface.

[0054] The thermal probe operates as follows: when a constant power is applied to the silicon chip, heat is conducted sequentially through the copper substrate layer 2 and the diamond frustum layer 3 to the sample under test. The unique diamond frustum structure effectively improves the uniformity of heat flow distribution and increases the heat flux at the contact surface.

[0055] Embodiment 2 of the present invention provides an optimized design method for the above-mentioned thermal probe. The present invention employs the Taguchi orthogonal experimental method for system optimization. This method can efficiently handle multi-parameter coupling problems and scientifically predict the global optimal solution with the fewest number of experiments. The method includes the following steps:

[0056] Step 1: Based on the theory of heat conduction and the principle of multi-objective optimization, determine the key design variables of the probe and their levels.

[0057] Preliminary finite element analysis identified three key variables that significantly affect the probe's thermal performance: the upper surface size of the diamond frustum (A), the thickness of the diamond frustum (B), and the thickness of the Cu substrate (C). Three levels were set for each variable, forming a three-factor, three-level experimental system:

[0058] Variable A (the upper surface dimension of the diamond prism): A1 = 1.5mm × 2.5mm, A2 = 1.6mm × 2.7mm, A3 = 1.8mm × 3.0mm.

[0059] Variable B (thickness of diamond frustum): B1 = 0.3 mm, B2 = 0.4 mm, B3 = 0.5 mm.

[0060] Variable C (Cu substrate thickness): C1 = 0.5 mm, C2 = 0.7 mm, C3 = 0.8 mm.

[0061] Step 2, based on the Taguchi orthogonal experimental design method, construct L9(3 3 Nine sets of simulation experiments were conducted using an orthogonal array. This method can effectively analyze the influence of multiple variables on multiple objectives with the fewest number of experiments. The parameter combinations for the nine sets of experiments are shown in Table 1.

[0062] Table 1. Parameters and Results of Taguchi Orthogonal Experimental Design

[0063]

[0064] Step 3: Using ANSYS finite element analysis software, steady-state thermal simulations were performed on each group of experiments. In the simulation model, a power of 20W was applied to the surface of the silicon chip. After the simulation, two key evaluation indicators were extracted:

[0065] Index 1 (small size characteristic): The difference in heat flux along the horizontal direction on the contact surface between the lower surface of the diamond frustum and the sample under test. The smaller the value, the more uniform the heat flow distribution.

[0066] Index 2 (Magnification Characteristics): The average heat flux density along the Z-axis on the lower surface of the diamond frustum. The larger this value, the stronger the heat penetration capability.

[0067] Step 4: Calculate the signal-to-noise ratio (SNR) of the two indicators for each experimental group, and perform signal-to-noise ratio response analysis to predict the optimal combination of levels for each design variable.

[0068] For the small characteristic index (range), the SNR calculation formula is:

[0069] SNR = -10log 10 (Δ 2 )

[0070] The larger the value, the better the heat flow uniformity and the smaller the fluctuation under this combination.

[0071] For the large-scale characteristic index (heat flux), the SNR calculation formula is:

[0072]

[0073] The higher the value, the higher the heat flux density under that combination.

[0074] In the formula:

[0075] Δ is the range, and q is the heat flux.

[0076] After calculating the SNR of the two indicators separately, response analysis was performed on the average SNR value of each factor at different levels. For each factor, the level that brings the highest average SNR value was selected to predict the optimal combination. Analysis showed that to simultaneously achieve optimal uniformity (smaller target) and highest heat flux density (larger target), the predicted optimal parameter combination is A1 (1.5mm × 2.5mm), B3 (0.5mm), and C1 (0.5mm).

[0077] Step 5: Verify the predicted optimal combination (A1B3C1) through experiments. Under the same simulation conditions, modeling and analysis were performed on this combination, and the average heat flux density at the contact surface was measured to be 92.736 W / mm². 2 The heat flux distribution range is 0.5294 W / mm². 2 The results significantly outperformed any set of experiments in the orthogonal array, confirming the superiority of this optimized combination, and therefore it was determined as the final structural parameters of the probe.

[0078] Embodiment 3 of the present invention provides an application example of the above-mentioned miniature thermal probe in measuring the thermal structure function of multilayer materials.

[0079] In an exemplary but non-limiting embodiment of the present invention, the sample under test is a copper-silver solder-copper three-layer structure. The optimal probe described in Example 2 is placed on the sample surface, power is applied to the silicon chip, and the temperature change curve over time is recorded using Ansys. The transient temperature response data is processed using Matlab programming, and the thermal resistance-heat capacity structure function curve of the sample can be obtained through calculation. The inflection points on this curve correspond to the interfaces of different material layers, thus allowing for quantitative analysis of the thermal resistance and heat capacity values ​​of each layer, providing crucial data for evaluating weld quality and interface characteristics. To quantitatively evaluate the identifiability of interface peaks in the structure function graph, the present invention introduces a separation point observability index, the calculation formula of which is:

[0080] Separation point observability index = (peak value - average of adjacent points) / peak value

[0081] The average value of adjacent points is taken as the average value of data points ±2 points away from the peak point to avoid the influence of adjacent fluctuations. The ability to identify separation points is graded according to the observability index value: ≥0.5 is excellent, 0.3-0.5 is good, 0.1-0.3 is average, and <0.1 is poor.

[0082] Structure function analysis of the probes before and after the improvement showed that the average observability index of the improved probe was 0.4469, a 1.34-fold improvement compared to 0.3330 before the improvement. Specifically, the improved probe achieved excellent results in 3 out of the 5 main peaks, completely eliminating the poor results; while the probe before the improvement only achieved 1 excellent result and had 1 poor result. The minimum observability index improved from 0.0665 to 0.1278, an increase of 92%, demonstrating the enhanced interface recognition capability of the improved probe. Detailed analysis results of the separation point observability index are shown in Table 2.

[0083] Table 2. Detailed Analysis Results of the Separation Point Observability Index

[0084]

[0085] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0086] 1) The miniature thermal probe provided by this invention achieves a synergistic improvement in heat flow uniformity and heat flux through optimized three-layer structural parameters, which significantly improves the accuracy and reliability of interface thermal resistance measurement.

[0087] 2) The optimization design method provided by this invention systematically solves the problem of multi-parameter coupling optimization by using the Taguchi orthogonal method. The design process is efficient and scientific, and is superior to traditional methods.

[0088] 3) This invention establishes a quantitative evaluation system based on the observability index of separation points, providing a scientific and visual evaluation standard for structural function analysis. The improved probe enhances the ability to identify interface separation points, providing reliable technical support for the accurate analysis of the thermal resistance of multilayer material interfaces.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A thermal resistance measurement probe based on the heat flow focusing effect, characterized in that, Including settings from top to bottom: The silicon chip layer (1) serves as a heat source for downward heat conduction. A copper substrate layer (2) is located below the silicon chip layer (1) and is used for support and heat conduction, and is connected to the measuring equipment; The diamond frustum layer (3) is an inverted frustum structure with an upper surface larger than the lower surface. Its upper surface is connected to the copper substrate layer (2). Its sidewall inclination angle and geometry together form a heat flow focusing channel, which is used to focus the lateral heat diffusion into longitudinal heat conduction, thereby improving the heat flux density and distribution uniformity of the contact surface.

2. The thermal resistance measuring probe according to claim 1, characterized in that, The sidewall inclination angle of the diamond frustum layer (3) is 45°.

3. The thermal resistance measuring probe according to claim 2, characterized in that, The length L2 and width W2 of the lower surface of the diamond frustum layer (3) are determined by the upper surface length L1, the upper surface width W1, and the frustum thickness t through the following geometric relationship: L2 = L1 - 2t W2 = W1 - 2t.

4. The thermal resistance measuring probe according to any one of claims 1 to 3, characterized in that, The silicon chip layer (1) has dimensions of 3 mm in length, 5 mm in width, and 0.1 mm in thickness.

5. The thermal resistance measuring probe according to any one of claims 1 to 3, characterized in that, The copper substrate layer (2) has dimensions of 4.5 mm in length, 7.5 mm in width, and 0.5 mm in thickness.

6. The thermal resistance measuring probe according to any one of claims 1 to 3, characterized in that, The upper surface of the diamond frustum layer (3) has dimensions of 1.5 mm in length, 2.5 mm in width, and 0.5 mm in thickness.

7. An optimization method for a thermal resistance measurement probe based on the heat flow focusing effect, characterized in that, include: Step 1: Based on heat conduction theory and multi-objective optimization principles, determine the key design variables and their levels for the probe; Preliminary finite element analysis identified three key variables that significantly affect the probe's thermal performance: the upper surface dimension A of the diamond frustum, the thickness B of the diamond frustum, and the thickness C of the Cu substrate. Three levels were set for each variable, forming a three-factor, three-level experimental system, specifically: Variable A: A1 = 1.5mm × 2.5mm, A2 = 1.6mm × 2.7mm, A3 = 1.8mm × 3.0mm; Variable B: B1 = 0.3 mm, B2 = 0.4 mm, B3 = 0.5 mm; Variable C: C1 = 0.5 mm, C2 = 0.7 mm, C3 = 0.8 mm; Step 2, based on the Taguchi orthogonal experimental design method, construct L9(3 3 Nine sets of simulation experiments were conducted using an orthogonal array. Step 3: Using ANSYS finite element analysis software, perform steady-state thermal simulation for each group of experiments; in the simulation model, apply 20W of power to the surface of the silicon chip; after simulation, extract two key evaluation indicators: The first indicator is the small size characteristic: the difference in heat flux along the horizontal direction on the contact surface between the lower surface of the diamond frustum and the sample to be tested. The smaller this value is, the more uniform the heat flow distribution. The second indicator is the large-scale characteristic: the average heat flux density along the Z-axis on the lower surface of the diamond frustum. The larger this value, the stronger the heat penetration capability. Step 4: Calculate the signal-to-noise ratio (SNR) of the two indicators for each experimental group, perform SNR response analysis, and predict the optimal combination of levels for each design variable. After calculating the SNR of the two indicators separately, a response analysis is performed on the average SNR value of each factor at different levels; for each factor, the level that brings the highest average SNR value is selected, and the optimal combination can be predicted. Step 5: For the predicted optimal combination A1B3C1, achieve simultaneous optimization of heat flux distribution uniformity and flux density.

8. The method according to claim 7, characterized in that, The formula for calculating the signal-to-noise ratio (SNR) is: For the small-scale characteristic index, SNR is calculated using the following formula: SNR=-10log 10 (D 2 ) The larger the value, the better the heat flow uniformity and the smaller the fluctuation under this combination; For the large-scale characteristic index, the SNR calculation formula is: The larger the value, the higher the heat flux density under that combination; In the formula: Δ is the range, and q is the heat flux.

Images