A motherboard image overheat point detection method and system based on thermal imaging

Abstract

The application discloses a kind of mainboard image overheat point detection method and system based on thermal imaging, applied to the technical field of electronic equipment thermal performance detection, by independent measurement, the first thermodynamic characteristic data of each unit itself is obtained;And by the first thermodynamic characteristic data, generate expected thermodynamic characteristic data;When at least two functional units work simultaneously, the second thermodynamic characteristic data of circuit board is obtained, the difference between the second thermodynamic characteristic data and expected thermodynamic characteristic data is compared, to determine whether there is overheat point caused by synergistic effect.There is the advantage that the method can accurately identify overheat point caused by synergistic effect by comparing the difference between actual heat distribution and expected heat distribution based on independent contribution superposition, to avoid misjudgment.

Description

Technical Field

[0001] This invention belongs to the field of thermal performance testing technology for electronic devices, and specifically relates to a method and system for detecting hot spots in motherboard images based on thermal imaging. Background Technology

[0002] In a company focused on the research and development and production of high-performance, high-density server motherboards, the product testing department bears the crucial responsibility of ensuring product quality. They typically employ a mature testing process: newly manufactured motherboards are installed on a dedicated test rack, powered on, and run testing software that subjects the motherboards to high-load operating conditions to simulate the intense operation in real-world application environments. During this process, a high-precision thermal imaging camera is used to capture the temperature distribution on the motherboard surface, generating a visual image of the heat distribution. Subsequently, a computer analysis program compares the temperature of various key areas on the thermal image with preset design temperature limits. If the temperature at any point exceeds the set threshold, the system issues an alarm, and the motherboard is deemed unqualified, requiring further analysis and repair. This method has demonstrated high effectiveness in identifying obvious hot spots caused by component defects (such as poor chip soldering or inherent component quality issues).

[0003] However, with the market's ever-increasing demand for computing performance, motherboard design trends are towards higher integration and more compact component layouts. To achieve more powerful functions and higher performance within limited space, various electronic components are integrated onto circuit boards at unprecedented densities. Under this high-density layout, existing detection methods based on simple temperature threshold comparisons are beginning to show their limitations. For example, when testing a flagship blade server motherboard with an extremely compact component layout, the test system frequently reported overheating issues at the same location: the network controller chip on the motherboard. Following standard procedure, the network controller chip was replaced, but the problem persisted. Further investigation revealed that a power module (VRM) supplying power to the central processing unit was designed very close to the network controller chip, with minimal spacing between them. Under high-load testing conditions, this power module generates a large amount of heat. Although its own temperature is within design specifications, the heat it generates is transferred to the adjacent network controller chip through conduction, convection, and radiation. The network controller chip itself also generates heat during operation. When these two sources of heat are combined, the local temperature of the network controller chip exceeds its safe limit.

[0004] In this situation, existing detection systems mistakenly identify network controller chips that passively heat up due to the thermal influence of neighboring components as the source of the fault. This misjudgment leads repair personnel to repeatedly replace functional components, wasting valuable parts and repair time without truly solving the problem. The crux of the problem lies in the lack of in-depth analysis capabilities of existing methods regarding heat distribution on the circuit board, heat transfer paths, and thermal interactions between components. It cannot distinguish whether an overheating point is due to "endogenous" overheating caused by defects in the component itself, or "exogenous" overheating caused by the influence of nearby high-heat-generating components. Simply identifying temperature exceedances as faults fails to accurately pinpoint the true "source" of heat, thus failing to provide accurate basis for fault diagnosis and design optimization.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] The purpose of this application is to provide a method and system for detecting hot spots on motherboard images based on thermal imaging, which has the advantage of being able to accurately identify hot spots caused by synergistic effects, thereby avoiding misjudgment.

[0007] In a first aspect, a method for detecting hot spots in motherboard images based on thermal imaging is provided, the method comprising the following steps:

[0008] S1: Obtain the first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they work independently;

[0009] S2: Based on the first thermodynamic feature data, generate expected thermodynamic feature data characterizing the linear superposition of the independent thermal contributions of the at least two functional units;

[0010] S3: Under the condition that at least two functional units are working simultaneously, acquire the second thermodynamic characteristic data of the circuit board;

[0011] S4: Based on the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, determine whether the circuit board has a hot spot caused by a synergistic effect.

[0012] This application proposes a method for detecting hot spots on a motherboard image based on thermal imaging. By measuring independently, it avoids heat interference from other units during operation, thus more accurately acquiring the thermodynamic characteristics of each unit. First thermodynamic characteristic data is acquired to generate expected thermodynamic characteristic data, which represents the simple summation of the thermal contributions of each unit in the absence of synergistic effects. Second thermodynamic characteristic data of the circuit board is acquired when at least two functional units are operating simultaneously, reflecting the overall heat distribution under actual operating conditions. The difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data is compared to determine whether a hot spot caused by a synergistic effect exists. If there is a significant difference between the actual heat distribution and the expected heat distribution, it indicates the presence of a synergistic effect, meaning that the heat interaction between some units exceeds the range of simple summation, leading to additional overheating. In summary, this method identifies hot spots caused by synergistic effects by comparing the difference between the actual heat distribution and the expected heat distribution based on the summation of independent contributions, offering the advantage of accurately identifying hot spots caused by synergistic effects and avoiding misjudgments.

[0013] Furthermore, step S1 includes:

[0014] S11: For any one of the at least two functional units:

[0015] Before stimulating the functional unit, obtain pre-stimulation thermodynamic data;

[0016] After stimulating the functional unit, obtain post-stimulation thermodynamic data;

[0017] S12: Based on the difference between the thermodynamic data after excitation and the thermodynamic data before excitation, determine the first thermodynamic characteristic data of the functional unit.

[0018] This application proposes a hot spot detection method for motherboard images based on thermal imaging. By more accurately acquiring the first thermodynamic characteristic data of functional units when they are operating independently, the accuracy of hot spot detection is improved.

[0019] Furthermore, step S2 includes:

[0020] S21: For each of the at least two functional units, based on the first thermodynamic characteristic data of the other functional units, determine the physical quantity characterizing the background thermal effect applied to that functional unit;

[0021] S22: Based on the physical quantity characterizing the background thermal effect, determine a correction value for adjusting the first thermodynamic characteristic data of the functional unit;

[0022] S23: Apply the correction value to the first thermodynamic characteristic data of the functional unit to generate the adjusted thermal contribution data of the functional unit;

[0023] S24: The adjusted thermal contribution data of the at least two functional units are superimposed to generate the expected thermodynamic characteristic data.

[0024] This application proposes a hot spot detection method for motherboard images based on thermal imaging. By considering and correcting the background thermal influence of other functional units on the superimposed functional units before superimposing their thermal contributions, the accuracy of the expected thermodynamic characteristic data is improved.

[0025] Furthermore, step S22 includes:

[0026] S221: Determine the feature points that characterize the location of the heat generation of the functional unit;

[0027] S222: Generate data characterizing the spatial distribution of background thermal effects on the functional unit as the physical quantity characterizing the background thermal effects;

[0028] S223: Obtain the background thermodynamic parameters corresponding to the feature points in the data representing the spatial distribution;

[0029] S224: Calculate the correction value based on the background thermodynamic parameters.

[0030] This application proposes a method for detecting hot spots in motherboard images based on thermal imaging. By accurately determining the correction value, the first thermodynamic characteristic data of the functional unit is adjusted to more accurately reflect the background thermal influence.

[0031] Furthermore, step S221 includes:

[0032] S2211: Calculate the spatial gradient field of thermodynamic parameters based on the first thermodynamic characteristic data of the functional unit;

[0033] S2212: The specific location in the spatial gradient field that represents the heat source is determined as the feature point.

[0034] Furthermore, step S3 includes:

[0035] S31: Monitor non-thermodynamic parameters associated with the operating state of the circuit board;

[0036] S32: Determine whether the non-thermodynamic parameter meets the preset triggering condition;

[0037] S33: When the non-thermodynamic parameters meet the preset triggering conditions, acquire the second thermodynamic characteristic data.

[0038] Furthermore, step S32 includes:

[0039] S321: Compare the value of the non-thermodynamic parameter with a preset parameter threshold;

[0040] S322: When the value of the non-thermodynamic parameter exceeds the preset parameter threshold, compare the duration for which the non-thermodynamic parameter continues to exceed the parameter threshold with the preset time threshold to obtain a comparison result;

[0041] S323: Determine whether the non-thermodynamic parameter meets the preset triggering condition based on the comparison result; if the comparison result is that the duration exceeds the preset time threshold, determine that the non-thermodynamic parameter meets the preset triggering condition; otherwise, determine that the non-thermodynamic parameter does not meet the preset triggering condition.

[0042] Furthermore, step S4 includes:

[0043] S41: Generate spatial distribution data characterizing the difference between the second thermodynamic feature data and the expected thermodynamic feature data;

[0044] S42: Identify a continuous region in the spatially distributed data composed of data points whose difference exceeds a preset discrimination threshold;

[0045] S43: The continuous region is identified as a hot spot caused by the synergistic effect.

[0046] Furthermore, step S42 includes:

[0047] S421: Rasterize the spatial distribution data to generate raster data;

[0048] S422: In the rasterized data, search for a set of data points whose difference exceeds the discrimination threshold and whose spatial locations are adjacent;

[0049] S423: Define the set as the continuous region.

[0050] Secondly, a motherboard image hotspot detection system based on thermal imaging, characterized in that, in the steps of any of the methods described above, the system comprises:

[0051] First acquisition module: acquires the first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they work independently;

[0052] Generation module: Based on the first thermodynamic feature data, generate expected thermodynamic feature data that characterizes the linear superposition of the independent thermal contributions of the at least two functional units;

[0053] Second acquisition module: Under the condition that at least two functional units are working simultaneously, acquire the second thermodynamic characteristic data of the circuit board;

[0054] Detection module: Based on the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, determine whether the circuit board has a hot spot caused by a synergistic effect.

[0055] Beneficial Effects: This application provides a method and system for detecting hot spots on a motherboard image based on thermal imaging. By measuring independently, it avoids heat interference generated by other units during operation, thus more accurately acquiring the thermodynamic characteristics of each unit. It generates expected thermodynamic characteristic data from the acquired first thermodynamic characteristic data, which represents the simple superposition of the thermal contributions of each unit in the absence of synergistic effects. It acquires second thermodynamic characteristic data of the circuit board when at least two functional units are operating simultaneously, reflecting the overall heat distribution under actual operating conditions. By comparing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, it determines whether there are hot spots caused by synergistic effects. If there is a significant difference between the actual heat distribution and the expected heat distribution, it indicates the presence of a synergistic effect, meaning that the heat interaction between some units exceeds the range of simple superposition, leading to additional overheating. In summary, this method identifies hot spots caused by synergistic effects by comparing the difference between the actual heat distribution and the expected heat distribution based on the superposition of independent contributions, and has the advantage of accurately identifying hot spots caused by synergistic effects and avoiding misjudgments. Attached Figure Description

[0056] Figure 1 A flowchart of a motherboard image hotspot detection method based on thermal imaging provided in this application.

[0057] Figure 2 This application provides a structural diagram of a motherboard image hotspot detection system based on thermal imaging.

[0058] Figure 3 This application provides a framework diagram of a motherboard image hotspot detection system based on thermal imaging.

[0059] Labeling explanation: 201, First acquisition module; 202, Generation module; 203, Second acquisition module; 204, Detection module. Detailed Implementation

[0060] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and marked in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0061] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0062] The following disclosure provides many different implementation methods or examples to achieve the objectives of the present invention and solve existing problems. To address this problem, this application proposes a method and system for detecting hot spots in motherboard images based on thermal imaging, as detailed below:

[0063] Please refer to Figure 1 Firstly, a method for detecting hot spots in motherboard images based on thermal imaging, the method comprising the following steps:

[0064] S1: Obtain the first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they work independently;

[0065] S2: Based on the first thermodynamic characteristic data, generate expected thermodynamic characteristic data that characterizes the linear superposition of the independent thermal contributions of at least two functional units;

[0066] S3: Under the condition that at least two functional units are working simultaneously, acquire the second thermodynamic characteristic data of the circuit board;

[0067] S4: Based on the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, determine whether there is a hot spot caused by a synergistic effect on the circuit board.

[0068] Among them, a circuit board refers to a substrate that carries electronic components.

[0069] A functional unit refers to a circuit module or component set on a circuit board that can independently perform a specific function, such as a central processing unit, a graphics processing unit, a power supply module, a network controller, etc. The purpose is to analyze the thermal contribution of different units.

[0070] The first thermodynamic characteristic data refers to data that characterizes the thermal state of a functional unit when it operates independently, such as temperature distribution maps and heat flux density maps.

[0071] Independent operation refers to the operation of one or a group of functional units on a circuit board when other functional units are in a non-working or low-power state. Its purpose is to isolate and measure the heat generated by the functional unit itself and its range of influence.

[0072] Expected thermodynamic characteristic data refers to theoretical heat distribution data predicted by a certain calculation method based on the data of each functional unit working independently, when these functional units work simultaneously and without additional thermal interaction. It is mainly used to establish a benchmark model for comparison with the actual situation.

[0073] Linear superposition of independent thermal contributions refers to the simple spatial accumulation of the heat distribution map or thermal parameters generated when each functional unit works independently. It is mainly used to simulate the effect of simple accumulation of heat from each unit under ideal conditions.

[0074] Simultaneous operation refers to at least two related functional units on the circuit board operating under normal or high load conditions. It is mainly used to simulate the real working state of the circuit board in actual application scenarios and capture all possible thermal interactions.

[0075] The second thermodynamic characteristic data refers to the data that characterizes the overall thermal state of the circuit board when at least two functional units are working simultaneously, such as the overall temperature distribution map. It can be obtained in the same or different ways as the first thermodynamic characteristic data. It is mainly used to reflect the heat distribution under actual working conditions, including the effects of all thermal interactions.

[0076] The difference refers to the difference in numerical or distribution characteristics between the second thermodynamic characteristic data and the expected thermodynamic characteristic data at the same spatial location. It can be calculated by point-by-point subtraction, comparison of regional averages, statistical analysis, etc. It is mainly used to quantify the degree of deviation between the actual heat distribution and the theoretical superimposed heat distribution.

[0077] Synergistic overheating refers to an abnormally high temperature region on a circuit board caused by the interaction of additional heat generated when at least two functional units are working simultaneously, rather than by defects in a single unit. It is mainly used to identify unexpected high-temperature regions caused by thermal coupling between adjacent units.

[0078] This application's solution analyzes the differences in heat distribution by acquiring and comparing thermodynamic data under different operating conditions step by step. Specifically:

[0079] The first thermodynamic characteristic data of each functional unit operating independently is obtained, which provides information on the thermal contribution of each unit without interference from other units. Based on this independent data, expected thermodynamic characteristic data is generated, representing the theoretical heat distribution of the simple superposition of the heat from each unit.

[0080] Acquire the second thermodynamic characteristic data when all relevant functional units are working simultaneously, which reflects the actual heat distribution under actual working conditions, including all thermal interactions.

[0081] By comparing the actual obtained second thermodynamic characteristic data with the theoretically calculated expected thermodynamic characteristic data, the additional heat concentration areas caused by the thermal synergy effect can be identified, namely, the superheated areas caused by the synergy effect.

[0082] In one specific implementation, taking the power supply module and network controller on the motherboard as examples: when the power supply module operates independently, an infrared thermal imager is used to acquire a temperature distribution map of the motherboard surface, which serves as the first thermodynamic characteristic data of the power supply module. Similarly, when the network controller operates independently, its first thermodynamic characteristic data is acquired.

[0083] The two independent temperature distribution maps are linearly superimposed at the pixel level to generate the expected thermodynamic characteristic data. Next, the power supply module and network controller are simultaneously operated under high load, and an infrared thermal imager is used again to acquire the overall temperature distribution map of the motherboard at this time, which serves as the second thermodynamic characteristic data.

[0084] The second thermodynamic feature data is compared with the expected thermodynamic feature data, for example, by calculating the temperature difference of the corresponding pixels. If the temperature difference of a certain region exceeds a preset threshold, and that region is associated with the location of the network controller, it can be determined that there is an overheated hot spot in that region caused by the thermal synergy effect of the power supply module on the network controller.

[0085] The above technical solution effectively distinguishes hot spots on a circuit board caused by the synergistic effect of functional units based on thermal imaging data. By comparing the expected linear superposition of the independent thermal contributions of functional units with the actual heat distribution when functional units are working simultaneously, abnormally high-temperature areas caused by the thermal influence of neighboring components rather than their own defects can be accurately identified. This avoids misjudging heat-affected components as fault sources, thus providing an accurate basis for fault diagnosis and repair of the circuit board, improving fault location accuracy, and reducing unnecessary component replacement and repair costs.

[0086] In some embodiments of this application, a method is proposed to obtain first thermodynamic characteristic data corresponding to at least two functional units on a circuit board operating independently. Specifically, this method can be achieved by directly measuring the temperature distribution on the surface of the functional unit when it is operating. This provides thermodynamic information about the functional unit's operating state. However, in its implementation, simply measuring the temperature distribution during operation cannot distinguish between the heat generated by the functional unit itself and the heat conducted or radiated from the surrounding environment or adjacent components. This may result in the acquisition of data containing background thermal effects, making the data inaccurate and affecting subsequent judgments on whether there are hot spots caused by synergistic effects on the circuit board.

[0087] Therefore, further, step S1 includes:

[0088] S11: For any one of at least two functional units:

[0089] Before stimulating the functional unit, obtain pre-stimulation thermodynamic data;

[0090] After stimulating the functional unit, obtain post-stimulation thermodynamic data;

[0091] S12: Based on the difference between the thermodynamic data after excitation and the thermodynamic data before excitation, determine the first thermodynamic characteristic data of the functional unit.

[0092] In this context, "exciting the functional unit" refers to applying the electrical energy or signal required to bring the functional unit into its normal operating state.

[0093] Pre-excitation thermodynamic data refers to the thermodynamic characteristic data obtained when the target functional unit is not excited and is in a non-working state. Its purpose is to record the initial thermal state of the functional unit under the background thermal environment. Post-excitation thermodynamic data refers to the thermodynamic characteristic data obtained when the target functional unit is excited and is in a working state. Its purpose is to record the thermal state of the functional unit under the combined action of its own working heat and the background thermal environment.

[0094] In some preferred embodiments, this application is implemented as follows: Assuming two functional units A and B need to be analyzed on a circuit board, to obtain the first thermodynamic characteristic data of functional unit A, firstly, ensure that functional unit A is in a non-operating state, while other necessary components on the circuit board (e.g., power supply module) are in a stable state. Use a thermal imaging camera to acquire the temperature distribution map of the area where functional unit A is located on the circuit board as the pre-excitation thermodynamic data. Then, only functional unit A is excited to enter the operating state (e.g., if functional unit A is a network chip, send network data to make it work). After its thermal state stabilizes, use a thermal imaging camera again to acquire the temperature distribution map of the same area as the post-excitation thermodynamic data. Finally, subtract the corresponding pixel values ​​of the pre-excitation temperature distribution map from the pixel values ​​of the post-excitation temperature distribution map; the resulting difference temperature distribution map is the first thermodynamic characteristic data of functional unit A. The first thermodynamic characteristic data of functional unit B can be obtained in a similar manner.

[0095] Furthermore, step S2 includes:

[0096] S21: For each of at least two functional units, based on the first thermodynamic characteristic data of the other functional units, determine the physical quantity characterizing the background thermal influence applied to that functional unit;

[0097] S22: Based on the physical quantities characterizing the background thermal effects, determine a correction value for the first thermodynamic characteristic data of the functional unit.

[0098] S23: Apply the correction value to the first thermodynamic characteristic data of the functional unit to generate the adjusted thermal contribution data of the functional unit;

[0099] S24: Superimpose the adjusted thermal contribution data of at least two functional units to generate the expected thermodynamic characteristic data.

[0100] Among them, the physical quantity characterizing the background thermal effect refers to the data that quantifies the thermal effect of other functional units on the current functional unit.

[0101] The correction value refers to the numerical value or function used to adjust the first thermodynamic characteristic data of the functional unit.

[0102] The adjusted thermal contribution data refers to the data obtained after applying the correction value to the first thermodynamic characteristic data of the functional unit. Specifically, it can be a corrected temperature distribution map or heat flux density distribution map. Its purpose is to more accurately characterize the thermal contribution of the functional unit after considering the background thermal effects.

[0103] In some preferred embodiments, this application is implemented as follows: Assume that there are functional units A and functional units B on the circuit board.

[0104] In step S21, for functional unit A, based on the first thermodynamic characteristic data of functional unit B (e.g., the temperature distribution map when functional unit B operates independently), a physical quantity characterizing the background thermal effect applied to functional unit A is determined. For example, the average temperature or maximum temperature of the temperature distribution map of functional unit B in the region near functional unit A can be calculated as the background thermal effect physical quantity applied to functional unit A. Similarly, for functional unit B, based on the first thermodynamic characteristic data of functional unit A, a background thermal effect physical quantity applied to functional unit B is determined.

[0105] In step S22, a correction value for adjusting the first thermodynamic characteristic data of functional unit A is determined based on the background thermal influence physical quantity applied to functional unit A. For example, a model can be preset that outputs a corresponding temperature offset as the correction value according to the background temperature. Similarly, a correction value for functional unit B is determined.

[0106] In step S23, the correction value is applied to the first thermodynamic characteristic data of functional unit A to generate adjusted thermal contribution data for functional unit A. For example, the corresponding correction value is added to or subtracted from each pixel value on the original temperature distribution map of functional unit A. Similarly, adjusted thermal contribution data for functional unit B is generated.

[0107] In step S24, the adjusted thermal contribution data of functional unit A and functional unit B are superimposed, for example, by adding the two adjusted temperature distribution maps at the pixel level to generate the expected thermodynamic characteristic data.

[0108] Furthermore, step S22 includes:

[0109] S221: Determine the feature points that represent the location of heat generation of the functional unit;

[0110] S222: Generate data characterizing the spatial distribution of background thermal effects on functional units as physical quantities characterizing background thermal effects;

[0111] S223: Obtain the background thermodynamic parameters corresponding to the feature points in the data characterizing the spatial distribution;

[0112] S224: Calculate the correction value based on the background thermodynamic parameters.

[0113] Among them, the characteristic points representing the location of the heat generation of the functional unit are one or more points inside or on the surface of the functional unit that represent the location of its main heat source. The purpose of obtaining them is to locate the key areas of the functional unit that are most sensitive to the background heat or where the heat generation of the functional unit is most concentrated.

[0114] Among them, the data characterizing the spatial distribution of background thermal influence on the functional unit refers to the spatial distribution of thermodynamic parameters (such as temperature and heat flux density) generated by other surrounding functional units or environmental factors in the area where the functional unit is located, excluding the heat generated by the functional unit itself during operation. It can be generated by calculating the thermal contribution data measured by thermal imaging, and its purpose is to provide detailed information on the background thermal influence at different locations of the functional unit.

[0115] Among them, the background thermodynamic parameters corresponding to the feature points refer to the thermodynamic parameter values ​​extracted from the data characterizing the spatial distribution of background thermal influence on the functional unit and corresponding to the previously determined feature point positions. These values ​​can be obtained by searching for the values ​​corresponding to the coordinate positions of the feature points in the spatial distribution data. The purpose is to quantify the background thermal influence on the key heat source positions of the functional unit.

[0116] Among them, calculating the correction value based on the background thermodynamic parameters refers to obtaining a value based on the background thermodynamic parameters corresponding to the feature point through a predetermined calculation method. This value is used to adjust the first thermodynamic characteristic data measured when the functional unit works independently. It can be done by directly using the background thermodynamic parameters as the correction value, or by converting the background thermodynamic parameters into correction values ​​through a functional relationship, or by looking up the correction value based on the background thermodynamic parameters. Its purpose is to generate thermal contribution data that can more accurately reflect the functional unit in the actual background thermal environment.

[0117] Specifically, assuming that the heat generated by the functional unit itself is excluded, the temperature distribution caused by other surrounding factors is as follows: Then, the data characterizing the spatial distribution of background thermal influence on the functional unit can be expressed as: Distribution of the functional unit within its area.

[0118] For each point in the feature point set P From the spatial distribution data of background thermal influence Extract the corresponding background thermodynamic parameters ,Right now: .

[0119] Suppose there exists a function g, based on the background thermodynamic parameters Calculate the correction value Then: ΔT_corr_p = g(T_bg_p). For example, it could be a simple proportional relationship. , This is the proportionality coefficient.

[0120] Furthermore, step S221 includes:

[0121] S2211: Calculate the spatial gradient field of thermodynamic parameters based on the first thermodynamic characteristic data of the functional unit;

[0122] S2212: Determine the specific location in the spatial gradient field that represents the heat source as a feature point.

[0123] The solution of this application calculates the spatial gradient field of thermodynamic parameters based on the first thermodynamic characteristic data of the functional unit. This spatial gradient field can intuitively reflect the flow trend of heat inside the functional unit and the degree of change of thermodynamic parameters.

[0124] Based on this, specific locations in the spatial gradient field representing heat sources are identified as feature points. These locations are typically key components or regions within the functional unit that actually generate heat. Feature points identified in this way more accurately reflect the center of the functional unit's thermal behavior compared to simple preset locations or highest temperature points. Applying these accurately identified feature points to subsequent steps—such as obtaining background thermodynamic parameters corresponding to these feature points after generating data characterizing the spatial distribution of background thermal effects on the functional unit, and calculating correction values ​​based on these parameters—ensures that the correction values ​​are calculated based on the background thermal effects experienced by the functional unit's own heat-generating locations. This makes the correction process more precise, more effectively isolates background thermal effects, and thus more accurately assesses the actual thermal state of the functional unit, improving the ability to distinguish between endogenous and exogenous overheating.

[0125] In some preferred embodiments, the spatial gradient field of thermodynamic parameters is calculated: Let the first thermodynamic characteristic data of the functional unit be... Then the spatial gradient field of thermodynamic parameters It can be represented as: in, and These are the partial derivatives of temperature in the x and y directions, respectively, which can be calculated using the image gradient algorithm.

[0126] Points with large gradient magnitudes in the spatial gradient field are identified as feature points. For example, the top k% of points with the largest gradient magnitudes are selected as feature points. Assuming the gradient magnitude threshold is Th, the feature point set P can be represented as: These feature points are typically located at the heat source locations of functional units, such as the surface or critical heat-generating areas inside the chip.

[0127] Furthermore, step S3 includes:

[0128] S31: Monitor non-thermodynamic parameters associated with the circuit board's operating status;

[0129] S32: Determine whether the non-thermodynamic parameters meet the preset triggering conditions;

[0130] S33: When the non-thermodynamic parameters meet the preset triggering conditions, acquire the second thermodynamic characteristic data.

[0131] Among them, non-thermodynamic parameters associated with the circuit board's operating status refer to non-temperature-related physical quantities or logical states that can reflect the circuit board's current operating load, resource usage, or environmental conditions. These can be achieved by monitoring CPU load rate, memory usage, hard disk read / write speed, network traffic, power consumption, fan speed, system error logs, and the operating status of specific software.

[0132] Preset trigger conditions refer to logical rules or combinations of numerical thresholds that are pre-set based on experience, design specifications, or historical data to determine whether a circuit board may enter a high-load or abnormal state. They can be implemented by setting thresholds for a single non-thermodynamic parameter (e.g., CPU load > 80%), combined thresholds for multiple non-thermodynamic parameters (e.g., CPU load > 70% and memory utilization > 90%), parameter change rate thresholds (e.g., power consumption increases sharply in a short period of time), or specific system events (e.g., starting a high-load application).

[0133] Acquiring the second thermodynamic characteristic data refers to collecting the temperature distribution information on the surface of the circuit board through a thermal imaging device under the condition that at least two functional units of the circuit board are working simultaneously, and converting it into a data format for subsequent analysis. This can be achieved by using an infrared thermal imager to capture thermal images of the circuit board, collecting temperature data from multiple points through a temperature sensor array, or combining visible light images for registration and calibration. The purpose is to obtain the actual thermal distribution of the circuit board under a specific working state.

[0134] In some preferred embodiments, a background monitoring program can be configured to periodically read the system's CPU utilization and temperature sensor data from key areas on the circuit board (such as the CPU and VRM). Preset trigger conditions can be set as follows: when the CPU utilization exceeds a parameter threshold for a continuous period, or when the temperature sensor reading of a key area exceeds a temperature threshold. When the monitoring program detects that either condition is met, it determines that the non-thermodynamic parameter meets the preset trigger condition. At this point, the system immediately triggers a thermal imaging camera to perform one or several consecutive thermal image acquisitions, obtaining secondary thermodynamic characteristic data of the circuit board under the current high-load state. This data is then passed to a subsequent analysis module and compared with pre-calculated expected thermodynamic characteristic data to detect whether there are hot spots caused by synergistic effects.

[0135] Furthermore, step S32 includes:

[0136] S321: Compare the values ​​of non-thermodynamic parameters with preset parameter thresholds;

[0137] S322: When the value of a non-thermodynamic parameter exceeds a preset parameter threshold, compare the duration for which the non-thermodynamic parameter exceeds the parameter threshold with a preset time threshold to obtain a comparison result;

[0138] S323: Determine whether the non-thermodynamic parameter meets the preset triggering condition based on the comparison result; if the comparison result is that the duration exceeds the preset time threshold, determine that the non-thermodynamic parameter meets the preset triggering condition; otherwise, determine that the non-thermodynamic parameter does not meet the preset triggering condition.

[0139] Non-thermodynamic parameters refer to physical quantities that are related to the working state of the circuit board and are not thermodynamic properties such as temperature or heat flow. They can be represented by electrical or performance parameters such as current, voltage, power consumption, CPU load rate, memory utilization, and network throughput. Their purpose is to characterize the real-time workload or abnormal state of the circuit board.

[0140] The preset parameter threshold is a preset limit value used to initially determine whether the value of non-thermodynamic parameters has reached the level that needs attention. It can be set according to the circuit board design specifications, historical operating data or experience, and its purpose is to filter out parameter fluctuations within the normal range.

[0141] The preset time threshold refers to the predetermined time length used to determine whether the non-thermodynamic parameter exceeding the threshold state has lasted for a sufficiently long time. It can be set according to the circuit board's response characteristics, thermal inertia, or experience, with the aim of avoiding false triggering caused by instantaneous interference or brief fluctuations. The duration refers to the length of time during which the value of the non-thermodynamic parameter continuously or cumulatively exceeds the preset parameter threshold. It can be measured using timers, counters, or timestamps, and its purpose is to reflect the stability or severity of the abnormal state of the non-thermodynamic parameter.

[0142] In some preferred embodiments, the non-thermodynamic parameter can specifically be the load rate of the central processing unit (CPU) on the circuit board. A preset parameter threshold can be set to 80%, and a preset time threshold can be set to 5 seconds. The specific judgment process is as follows: The system continuously monitors the real-time load rate of the CPU. First, the real-time load rate is compared with 80%. If the real-time load rate is lower than or equal to 80%, the trigger condition is not met. If the real-time load rate exceeds 80%, the system starts timing and records the duration for which the load rate remains above 80%. Then, the recorded duration is compared with the preset 5-second time threshold. If the duration is less than or equal to 5 seconds, the trigger condition is not met. Only when the real-time load rate exceeds 80% and the duration exceeding 80% also exceeds 5 seconds is the non-thermodynamic parameter (CPU load rate) finally determined to meet the preset trigger condition, at which point the system will trigger the acquisition of the second thermodynamic characteristic data of the circuit board.

[0143] Furthermore, step S4 includes:

[0144] S41: Generate spatial distribution data characterizing the difference between the second thermodynamic feature data and the expected thermodynamic feature data;

[0145] S42: Identify a continuous region in spatially distributed data consisting of data points whose difference exceeds a preset discrimination threshold;

[0146] S43: Identify continuous regions as hot spots caused by synergistic effects.

[0147] In some preferred embodiments, this application is implemented as follows:

[0148] First, acquire the second thermodynamic characteristic data of the circuit board measured under the condition that multiple functional units are operating simultaneously. This data can be a two-dimensional temperature matrix, where each element represents the temperature at the corresponding location on the circuit board. Simultaneously, acquire the expected thermodynamic characteristic data generated by the linear superposition of the independent thermal contributions of each functional unit. This data can also be a two-dimensional temperature matrix, representing the expected temperature distribution without synergistic effects.

[0149] Then, by pointwise subtracting the expected thermodynamic feature data matrix from the second thermodynamic feature data matrix, spatial distribution data characterizing the difference between the two is generated. This data is also a two-dimensional matrix, where each element represents the temperature difference value at the corresponding location. Next, a preset discrimination threshold is set, for example, 5°C. The spatial distribution data matrix of differences is traversed to identify all data points with a difference greater than 5°C.

[0150] Furthermore, connectivity analysis is performed on these data points whose differences are greater than the threshold. For example, a pixel-based 8-connectivity algorithm is used to group data points that are spatially adjacent (horizontally, vertically, or diagonally) and whose differences are all greater than the threshold into a continuous region.

[0151] Finally, each identified continuous region is determined as a hot spot caused by a synergistic effect, and the spatial location information of these hot spots can be output, such as the boundary coordinates or center point coordinates of the region.

[0152] Furthermore, step S42 includes:

[0153] S421: Rasterize the spatially distributed data to generate rasterized data;

[0154] S422: In rasterized data, search for a set of data points whose differences exceed a threshold and whose spatial locations are adjacent;

[0155] S423: Define the set as a continuous region.

[0156] Rasterization refers to the process of converting raw spatial data, which may have arbitrary resolution or distribution, into a discrete dataset composed of regular grid (raster) units. This can be achieved by dividing the spatial region into uniform square or rectangular grids and aggregating the raw data within each grid unit (e.g., calculating the average, maximum, or sampling center point value). The purpose is to structure and discretize the spatial data to facilitate subsequent calculations and analysis.

[0157] Rasterized data refers to datasets formed after rasterization, which are organized in grid cells. Each grid cell stores the attribute values ​​(such as difference data) of the corresponding spatial region. It can be stored in the form of a two-dimensional array or matrix. Its purpose is to provide a standardized data structure to simplify the representation of spatial location and the determination of adjacency relationships.

[0158] A set of data points whose differences exceed a threshold and whose spatial locations are adjacent refers to one or more groups of data points in raster data that meet the condition that their stored differences are greater than a preset threshold and that these data points are adjacent to each other in the raster grid (e.g., adjacent in the vertical, horizontal, or diagonal directions). It can be identified using connected component analysis algorithms or region growing algorithms. Its purpose is to extract spatially clustered anomalous regions from discrete raster data.

[0159] A continuous region refers to a region with a certain range and shape, consisting of spatially connected data points. These data points collectively exhibit a certain attribute (e.g., the difference exceeds a threshold). It can be represented by labeling or boundary extraction. The purpose is to abstract a discrete set of data points into a physical region with practical meaning and identifiability.

[0160] In some preferred embodiments, specifically, the rasterization process S421 of the spatial distribution data may be to divide an image or matrix representing the difference between the second thermodynamic feature data and the expected thermodynamic feature data into a two-dimensional raster grid according to a preset grid size (e.g., dividing each square millimeter into several grid units). For each grid unit, the average or maximum difference of all original data points in the unit is calculated, and the value is used as the data of the raster unit, thereby generating rasterized data.

[0161] In rasterized data, a set S422 consisting of data points whose differences exceed a discrimination threshold and are spatially adjacent is searched. This can be achieved by traversing the generated rasterized data. For each raster cell, it is first checked whether its stored difference exceeds a preset discrimination threshold. If it does, it is marked as a candidate point. Then, a connected component analysis algorithm, such as breadth-first search or depth-first search, is applied to all candidate points to check for adjacent (e.g., using 8-connectivity, i.e., considering adjacent cells in the horizontal, vertical, and diagonal directions) candidate points, and all interconnected candidate points are grouped into a set.

[0162] By defining the set as a continuous region S423, each connected candidate point set identified in step S422 can be output or labeled as an independent continuous region, for example, by drawing the boundaries of these regions on the original thermal imaging image or highlighting them with color.

[0163] Please refer to Figure 2 , Figure 3 A motherboard image hotspot detection system based on thermal imaging, characterized in that, in the steps applied to any of the above methods, the system includes:

[0164] First acquisition module 201: Acquires first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they work independently;

[0165] Generation module 202: Based on the first thermodynamic feature data, generate expected thermodynamic feature data that characterizes the linear superposition of the independent thermal contributions of at least two functional units;

[0166] Second acquisition module 203: Acquires the second thermodynamic characteristic data of the circuit board under the condition that at least two functional units are working simultaneously;

[0167] Detection module 204: Based on the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, determine whether there is a hot spot caused by a synergistic effect on the circuit board.

[0168] The acquisition module 201 is a unit used to collect thermodynamic information of the circuit board. It can be implemented by using a thermal imaging camera, infrared sensor array or temperature probe, etc., and its purpose is to acquire the thermal performance of the functional unit when it works independently.

[0169] Among them, the generation module 202 refers to the unit used to perform calculations and modeling based on the input data. It can be implemented using processors, computing servers or application-specific integrated circuits, etc., and its purpose is to simulate the linear superposition of the independent thermal contributions of functional units.

[0170] Among them, the second acquisition module 203 refers to the unit used to collect thermodynamic information of the circuit board under a specific working state. It can be implemented using the same hardware as the acquisition module or an independent acquisition device. Its purpose is to acquire the actual heat performance when the functional units work at the same time.

[0171] Among them, the detection module 204 refers to the unit used to analyze and compare thermodynamic data and make a judgment. It can be implemented by a processor, algorithm execution unit or comparison circuit, etc., and its purpose is to identify the difference between the actual heat distribution and the predicted heat distribution.

[0172] Among them, the hot spot caused by synergistic effect refers to the area on the circuit board where the actual measured second thermodynamic characteristic data is significantly higher than the expected thermodynamic characteristic data predicted by linear superposition of independent working data. This difference indicates that the additional heat generated or accumulated in this area is due to the mutual influence (such as thermal radiation, thermal conduction, or airflow obstruction) generated when at least two functional units work at the same time, rather than the defects of a single functional unit itself. Specifically, it can be presented in the form of hot spot area, abnormal temperature rise area, or high value area in difference heat map. Its purpose is to distinguish the overheating caused by different reasons, so as to accurately locate the fault.

[0173] Based on the above modules and data, the solution of this application specifically implements each step of the hot spot detection method for motherboard images based on thermal imaging as an acquisition module 201, a generation module 202, a second acquisition module 203, and a detection module 204, thereby constructing a system capable of automatically executing the detection process.

[0174] The acquisition module 201 is responsible for collecting thermodynamic data when the functional units operate independently, providing a basis for subsequent analysis. The generation module 202 calculates the theoretical heat distribution under conditions of no mutual influence based on this independent data. The second acquisition module 203 captures the actual overall heat distribution when the functional units operate simultaneously. Finally, the detection module 204 accurately identifies the extra heat accumulation areas caused by the mutual influence of the functional units—i.e., hot spots caused by synergistic effects—by comparing the actual heat distribution with the theoretical heat distribution.

[0175] This systematic approach enables the efficient and accurate execution of previously complex and manual methods, transforming abstract detection logic into operable system functions. This overcomes the challenge of implementing methods in practical applications. It is precisely this method of transforming methods into system modules that makes large-scale, high-precision hotspot detection of circuit boards possible, significantly improving detection efficiency and reliability.

[0176] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0177] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A thermal imaging-based motherboard image hotspot detection method, characterized in that, The method comprises steps of: S1: obtaining, by an infrared thermal imager, first thermal imaging images corresponding to each of at least two functional units on a mainboard when the functional units work independently; S2: based on the first thermal imaging images, performing pixel-level linear superposition on the first thermal imaging images corresponding to the at least two functional units to generate an expected thermal imaging image representing independent thermal contributions of the at least two functional units; Step S2 comprises: S21: for each of the at least two functional units, determining a physical quantity representing background thermal influence applied to the functional unit based on the first thermal imaging images of other functional units except the functional unit; S22: based on the physical quantity representing background thermal influence, determining a correction value for adjusting the first thermal imaging image of the functional unit; S23: applying the correction value to the first thermal imaging image of the functional unit to generate adjusted thermal contribution data of the functional unit; S24: superimposing the adjusted thermal contribution data of the at least two functional units to generate the expected thermal imaging image; S3: under the condition that the at least two functional units work simultaneously, obtaining, by the infrared thermal imager, a second thermal imaging image of the mainboard, the second thermal imaging image representing an actual temperature distribution when working simultaneously; S4: based on the difference between the second thermal imaging image and the expected thermal imaging image, determining whether there is an overheating point caused by a synergistic effect on the mainboard.

2. The method according to claim 1, wherein, Step S1 comprises: S11: for any functional unit of the at least two functional units: obtaining pre-excitation thermal data before exciting the functional unit; obtaining post-excitation thermal data after exciting the functional unit; S12: based on the difference between the post-excitation thermal data and the pre-excitation thermal data, determining the first thermal imaging image of the functional unit.

3. The method of claim 1, wherein the method further comprises: Step S22 comprises: S221: determining a feature point representing a heat generation position of the functional unit; S222: generating data representing spatial distribution of background thermal influence on the functional unit as the physical quantity representing background thermal influence; S223: obtaining a background thermal parameter corresponding to the feature point in the data representing spatial distribution; S224: calculating the correction value according to the background thermal parameter.

4. The method of claim 3, wherein the method further comprises: Step S221 comprises: S2211: based on the first thermal imaging image of the functional unit, calculating a spatial gradient field of a thermal parameter; S2212: determining a specific position representing a heat source in the spatial gradient field as the feature point.

5. The method of claim 1, wherein the method is characterized by: Step S3 comprises: S31: monitoring a non-thermal parameter associated with a working state of the mainboard; S32: determining whether the non-thermal parameter meets a preset triggering condition; S33: when the non-thermal parameter meets the preset triggering condition, obtaining the second thermal imaging image.

6. The method of claim 5, wherein the method further comprises: Step S32 comprises: S321: comparing a value of the non-thermal parameter with a preset parameter threshold value; S322: When the value of the non-thermodynamic parameter exceeds the preset parameter threshold, compare the duration of the non-thermodynamic parameter continuously exceeding the parameter threshold with a preset time threshold to obtain a comparison result; S323: Determine whether the non-thermodynamic parameter meets the preset triggering condition according to the comparison result; when the duration exceeds the preset time threshold, it is determined that the non-thermodynamic parameter meets the preset triggering condition, otherwise, it is determined that the non-thermodynamic parameter does not meet the preset triggering condition.

7. The method of claim 1, wherein the method further comprises: Step S4 includes: S41: Generate spatial distribution data representing the difference between the second thermal image and the expected thermal image; S42: Identify a continuous region in the spatial distribution data composed of data points whose difference exceeds a preset discrimination threshold; S43: Determine the continuous region as the overheating point caused by the synergistic effect.

8. The method of claim 7, wherein the method further comprises: Step S42 includes: S421: Perform rasterization processing on the spatial distribution data to generate rasterized data; S422: In the rasterized data, search for a set composed of data points whose difference exceeds the discrimination threshold and whose spatial positions are adjacent; S423: Determine the set as the continuous region.

9. A thermal imaging based motherboard image hotspot detection system, characterized in that, The system is applied to the steps of the method of any one of claims 1-8, and includes: A first acquisition module: acquiring, by an infrared thermal imager, first thermal images corresponding to at least two functional units of a mainboard when each of the functional units works independently; A generation module: based on the first thermal images, performing pixel-level linear superposition on the first thermal images corresponding to the at least two functional units to generate an expected thermal image representing the independent thermal contribution of the at least two functional units; A second acquisition module: under the condition that the at least two functional units work simultaneously, acquiring, by the infrared thermal imager, a second thermal image of the mainboard, the second thermal image representing the actual temperature distribution when working simultaneously; A detection module: based on the difference between the second thermal image and the expected thermal image, determining whether the mainboard has an overheating point caused by the synergistic effect.

Images