Mainboard image overheat point detection method and system based on thermal imaging

Abstract

The invention discloses a mainboard image overheat point detection method and system based on thermal imaging, and is applied to the technical field of thermal performance detection of electronic equipment. First thermodynamic characteristic data of each unit is obtained through independent measurement; generating expected thermodynamic characteristic data through the first thermodynamic characteristic data; when the at least two functional units work at the same time, second thermodynamic characteristic data of the circuit board are obtained, and the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data is compared, so that whether an overheated point caused by the synergistic effect exists or not is determined. According to the method, the difference between the actual heat distribution and the expected heat distribution based on independent contribution superposition is compared, the overheat points caused by the synergistic effect are recognized, and the method has the advantages that the overheat points caused by the synergistic effect can be accurately recognized, and misjudgment is avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of thermal performance detection of electronic equipment, and in particular relates to a method and system for detecting hot spots in motherboard images based on thermal imaging. Background Art

[0002] At a company specializing in the development and production of high-performance, high-density server motherboards, the product testing department shoulders the crucial responsibility of ensuring product quality. They typically employ a sophisticated testing process: newly produced motherboards are mounted on a dedicated test rack, powered on, and run through a suite of test software that subjects the motherboards to high-load conditions, simulating the intense operating conditions encountered in real-world applications. During this process, a high-precision thermal imaging camera captures the temperature distribution on the motherboard surface, generating intuitive thermal images. A computer analysis program then compares the temperatures of key areas within the thermal image with preset design temperature limits. If the temperature at any point exceeds the set threshold, the system issues an alarm, deeming the motherboard unqualified and requiring further analysis and repair. This method has proven highly effective in identifying obvious hot spots caused by component defects, such as poor solder connections or component quality issues.

[0003] However, as market demand for computing performance continues to rise, motherboard designs are trending towards higher integration and more compact component layouts. To achieve greater functionality and higher performance within limited space, various electronic components are being packed onto circuit boards at unprecedented density. Under this high-density layout, existing detection methods based on simple temperature threshold comparisons are beginning to show their limitations. For example, while testing a flagship blade server motherboard with an extremely compact component layout, the test system frequently reported excessive temperatures at the same location: the network controller chip on the motherboard. Following standard procedures, maintenance personnel replaced the network controller chip, but the problem persisted. A thorough investigation revealed that a power module (VRM) on the motherboard, which powers the central processing unit (CPU), was designed to be located very close to the network controller chip, with minimal spacing between them. Under high-load testing conditions, the VRM generates significant heat. Although its own temperature is within design specifications, this heat is transferred to the adjacent network controller chip through conduction, convection, and radiation. The network controller chip also generates heat during operation. When these two heat sources are combined, the local temperature of the network controller chip exceeds the safety limit.

[0004] In this case, existing detection systems mistakenly identify the network controller chip, which is passively heated due to the influence of heat from adjacent components, as the source of the fault. This misjudgment causes maintenance personnel to repeatedly replace properly functioning components, wasting valuable components and repair time while failing to truly resolve the problem. The crux of the problem lies in the fact that existing methods lack the ability to deeply analyze the heat distribution, transfer paths, and thermal interactions between components on the circuit board. They cannot distinguish whether an overheating spot is "intrinsic" due to a defect in the component itself or "exogenous" due to the influence of neighboring high-heat-producing components. Simply identifying the point where the temperature exceeds the specified limit as a fault fails to accurately identify the true "source" of the heat, thus failing to provide an accurate basis for fault diagnosis and design optimization.

[0005] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0006] The purpose of this application is to provide a method and system for detecting hot spots in 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 a motherboard image based on thermal imaging is provided, the method comprising the steps of:

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

[0009] S2: generating expected thermodynamic characteristic data representing the linear superposition of independent thermal contributions of the at least two functional units based on the first thermodynamic characteristic data;

[0010] S3: Acquiring second thermodynamic characteristic data of the circuit board under the condition that the at least two functional units operate simultaneously;

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

[0012] This application proposes a method for detecting hot spots in motherboard images based on thermal imaging. By independent measurement, the heat interference generated by other units when working can be avoided, thereby more accurately obtaining the thermodynamic characteristics of each unit itself; by obtaining the first thermodynamic characteristic data, expected thermodynamic characteristic data is generated. This data represents the simple superposition of the thermal contributions of each unit in the absence of a synergistic effect; by obtaining the second thermodynamic characteristic data of the circuit board when at least two functional units are working simultaneously, this data reflects the overall thermal distribution under the actual working state; by comparing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, it is determined whether there are hot spots caused by a synergistic effect. If there is a significant difference between the actual thermal distribution and the expected thermal distribution, it indicates the presence of a synergistic effect, that is, the thermal interaction between certain units exceeds the range of simple superposition, resulting in additional overheating. In summary, this method identifies hot spots caused by a synergistic effect by comparing the difference between the actual thermal distribution and the expected thermal distribution based on the superposition of independent contributions. It has the advantage of being able to accurately identify hot spots caused by a synergistic effect and avoid misjudgment.

[0013] Furthermore, step S1 includes:

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

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

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

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

[0018] This application proposes a method for detecting hot spots in motherboard images based on thermal imaging, which improves the accuracy of hot spot detection by more accurately acquiring the first thermodynamic characteristic data of the functional unit when it works independently.

[0019] Furthermore, step S2 includes:

[0020] S21: for each of the at least two functional units, determining a physical quantity representing a background thermal influence applied to the functional unit based on first thermodynamic characteristic data of other functional units except the functional unit;

[0021] S22: determining a correction value for adjusting first thermodynamic characteristic data of the functional unit based on the physical quantity characterizing the background thermal influence;

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

[0023] S24: Superimposing the adjusted heat contribution data of the at least two functional units to generate the expected thermodynamic characteristic data.

[0024] This application proposes a method for detecting hot spots in motherboard images based on thermal imaging. Before superimposing the thermal contribution of functional units, the background thermal impact generated by other functional units is considered and corrected, thereby improving the accuracy of the expected thermodynamic characteristic data.

[0025] Furthermore, step S22 includes:

[0026] S221: Determine a characteristic point of the functional unit that represents a heat generation location of the functional unit;

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

[0028] S223: Obtaining 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 parameter.

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

[0031] Furthermore, step S221 includes:

[0032] S2211: Calculating a spatial gradient field of a thermodynamic parameter based on the first thermodynamic characteristic data of the functional unit;

[0033] S2212: Determine a specific position representing a heat source in the spatial gradient field as the feature point.

[0034] Furthermore, step S3 includes:

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

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

[0037] S33: When the non-thermodynamic parameter satisfies the preset trigger condition, acquiring the second thermodynamic characteristic data.

[0038] Furthermore, step S32 includes:

[0039] S321: Comparing 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 exceeds the parameter threshold with a preset time threshold to obtain a comparison result;

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

[0042] Furthermore, step S4 includes:

[0043] S41: generating spatial distribution data representing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data;

[0044] S42: Identifying a continuous region in the spatial distribution data consisting of data points whose difference exceeds a preset discrimination threshold;

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

[0046] Furthermore, step S42 includes:

[0047] S421: performing rasterization processing on the spatial distribution data to generate rasterized data;

[0048] S422: searching the rasterized data for a set of data points whose difference exceeds the discrimination threshold and whose spatial positions are adjacent to each other;

[0049] S423: Determine the set as the continuous area.

[0050] In a second aspect, a motherboard image hotspot detection system based on thermal imaging is characterized in that, when applied to the steps of any of the above methods, the system includes:

[0051] A first acquisition module is configured to acquire first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they operate independently;

[0052] A generating module: generating, based on the first thermodynamic characteristic data, expected thermodynamic characteristic data representing the linear superposition of independent thermal contributions of the at least two functional units;

[0053] A second acquisition module: acquiring second thermodynamic characteristic data of the circuit board under the condition that the at least two functional units work simultaneously;

[0054] A detection module is configured to determine whether the circuit board has a hot spot caused by a synergistic effect based on a difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data.

[0055] Beneficial effects: The present application provides a method and system for detecting hot spots in motherboard images based on thermal imaging. By independent measurement, the heat interference generated by other units when working can be avoided, thereby more accurately obtaining the thermodynamic characteristics of each unit itself; by obtaining the first thermodynamic characteristic data, expected thermodynamic characteristic data is generated, which represents the simple superposition of the thermal contributions of each unit in the absence of a synergistic effect; by obtaining the second thermodynamic characteristic data of the circuit board when at least two functional units are working simultaneously, the data reflects the overall thermal distribution under the actual working state; by comparing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, it is determined whether there is a hot spot caused by a synergistic effect. If there is a significant difference between the actual thermal distribution and the expected thermal distribution, it indicates the presence of a synergistic effect, that is, the thermal interaction between certain units exceeds the range of simple superposition, resulting in additional overheating. In summary, the present method identifies hot spots caused by a synergistic effect by comparing the difference between the actual thermal distribution and the expected thermal distribution based on the superposition of independent contributions. It has the advantage of being able to accurately identify hot spots caused by a synergistic effect and avoid misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This application provides a flowchart of a method for detecting hot spots in motherboard images based on thermal imaging.

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

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

[0059] Description of reference numerals: 201, first acquisition module; 202, generation module; 203, second acquisition module; 204, detection module. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and marked in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work fall within the scope of protection of the present application.

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

[0062] The following disclosure provides many different implementations or examples to achieve the purpose of the present invention and solve the existing problems. To solve this problem, this application proposes a motherboard image hot spot detection method and system based on thermal imaging, as follows:

[0063] Please refer to Figure 1 In a first aspect, a method for detecting hot spots in a motherboard image based on thermal imaging is provided, the method comprising the steps of:

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

[0065] S2: generating expected thermodynamic characteristic data representing the linear superposition of independent thermal contributions of at least two functional units based on the first thermodynamic characteristic data;

[0066] S3: Acquiring second thermodynamic characteristic data of the circuit board under the condition that at least two functional units operate simultaneously;

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

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

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

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

[0071] Independent operation means that one or a group of functional units on a circuit board operate independently when other functional units are in non-working or low-power states. The purpose is to isolate and measure the heat generated by the functional unit itself and its impact range.

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

[0073] The linear superposition of independent thermal contributions refers to the simple spatial accumulation of the heat distribution diagrams or thermal parameters generated by each functional unit when working independently. It is mainly used to simulate the effect of simple accumulation of heat of each unit under ideal conditions.

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

[0075] The second thermodynamic characteristic data refers to data characterizing the overall thermal state of the circuit board obtained when at least two functional units are working simultaneously, such as an overall temperature distribution diagram. It can be obtained in the same or different way as obtaining the first thermodynamic characteristic data. It is mainly used to reflect the thermal distribution under the actual working state, including the influence of all thermal interactions.

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

[0077] Hot spots caused by synergistic effects refer to areas on the circuit board where the temperature rises abnormally due to the extra heat generated by at least two functional units working simultaneously, rather than due to defects in a single unit itself. They are mainly used to identify unexpectedly high temperature areas caused by thermal coupling between adjacent units.

[0078] The solution of this application analyzes the differences in heat distribution by step-by-step acquisition and comparison of thermodynamic data under different working conditions. Specifically:

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

[0080] The second thermodynamic characteristic data is obtained when all relevant functional units are working simultaneously, which reflects the real thermal distribution including all thermal interactions under the actual working state.

[0081] By comparing the difference between the actually obtained second thermodynamic characteristic data and the theoretically calculated expected thermodynamic characteristic data, those areas of additional heat concentration caused by the thermal synergistic effect, namely, the hot spots caused by the synergistic effect, can be identified.

[0082] In a specific embodiment, using the power supply module and network controller on a motherboard as an example, when the power supply module is operating independently, an infrared thermal imager is used to obtain a temperature distribution map on the motherboard surface as the first thermodynamic characteristic data of the power supply module. Similarly, when the network controller is operating independently, its first thermodynamic characteristic data is obtained.

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

[0084] The second thermodynamic characteristic data is compared with the expected thermodynamic characteristic data, for example, by calculating the temperature difference of corresponding pixels. If the temperature difference in a certain area exceeds a preset threshold, and the area is associated with the location of the network controller, it can be determined that a hot spot exists in the area due to 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 effects of functional units based on thermal imaging data. By comparing the expected linear superposition of the independent thermal contributions of the functional units with the actual thermal distribution when the functional units are operating simultaneously, it is possible to accurately identify abnormally high temperature areas caused by the thermal influence of adjacent components rather than defects within the functional units themselves. This avoids misidentifying heat-affected components as the source of faults, 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 of the above-mentioned embodiments of the present application, it is proposed to obtain first thermodynamic characteristic data corresponding to at least two functional units on a circuit board when they are each working independently. The method of obtaining first thermodynamic characteristic data corresponding to at least two functional units on a circuit board when they are each working independently can be specifically by directly measuring the temperature distribution on the surface of the functional unit when the functional unit is working, so that thermodynamic information of the functional unit under the working state can be obtained. However, in its implementation process, only measuring the temperature distribution under the working state cannot distinguish between the heat generated by the functional unit itself and the heat conducted or radiated from the surrounding environment or adjacent components, which may cause the acquired data to contain background heat effects, making the data inaccurate, and thus affecting the subsequent judgment on whether there are hot spots caused by synergistic effects on the circuit board.

[0087] Therefore, further, step S1 includes:

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

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

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

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

[0092] Here, stimulating the functional unit refers to applying the electric energy or signal required to make the functional unit enter its normal working state.

[0093] Thermodynamic data before excitation 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 in the background thermal environment; thermodynamic data after excitation refers to the thermodynamic characteristic data obtained when the target functional unit is excited and 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, the present application is implemented as follows: assuming that 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, first ensure that functional unit A is in a non-operating state and that other essential components on the circuit board (e.g., a power supply module) are in a stable state. A thermal imaging camera is used to obtain a temperature distribution map of the area on the circuit board where functional unit A is located as pre-stimulation thermodynamic data. Then, only functional unit A is stimulated to enter an operating state (e.g., if functional unit A is a network chip, network data is sent to activate it). After its thermal state stabilizes, a thermal imaging camera is used to obtain a temperature distribution map of the same area as post-stimulation thermodynamic data. Finally, the pixel values ​​of the post-stimulation temperature distribution map are subtracted from the corresponding pixel values ​​of the pre-stimulation 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 the at least two functional units, determining a physical quantity representing a background thermal influence applied to the functional unit based on first thermodynamic characteristic data of other functional units except the functional unit;

[0097] S22: determining a correction value for adjusting first thermodynamic characteristic data of the functional unit based on a physical quantity representing background thermal influence;

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

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

[0100] The physical quantity characterizing the background thermal impact refers to data that quantifies the thermal impact of other functional units on the current functional unit.

[0101] The correction value refers to a 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 diagram or heat flux density distribution diagram. Its purpose is to more accurately characterize the thermal contribution of the functional unit after considering the background thermal influence.

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

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

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

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

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

[0108] Furthermore, step S22 includes:

[0109] S221: Determine a characteristic point of the functional unit that represents a location where heat is generated.

[0110] S222: Generate data representing the spatial distribution of background thermal influence on the functional unit as a physical quantity representing the background thermal influence;

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

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

[0113] The characteristic points of a functional unit that characterize the location of its own heat generation refer to 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 where the functional unit is most sensitive to background heat or where its own heat generation is most concentrated.

[0114] Among them, the data characterizing the spatial distribution of background heat 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. It can be generated by calculation using the heat contribution data measured by thermal imaging, and its purpose is to provide detailed information on the background heat influence at different positions of the functional unit.

[0115] The background thermodynamic parameter corresponding to the characteristic point refers to the thermodynamic parameter value corresponding to the previously determined characteristic point position extracted from the data representing the spatial distribution of the background thermal influence on the functional unit. It can be obtained by searching for the value corresponding to the coordinate position of the characteristic point in the spatial distribution data. Its purpose is to quantify the background thermal influence on the key heat source position of the functional unit.

[0116] Among them, calculating the correction value based on the background thermodynamic parameters means obtaining a numerical value through a predetermined calculation method based on the background thermodynamic parameters corresponding to the characteristic points obtained. The numerical value is used to adjust the first thermodynamic characteristic data measured when the functional unit works independently. It can directly use the background thermodynamic parameters as the correction value, or convert the background thermodynamic parameters into the correction value through a functional relationship, or determine the correction value based on the background thermodynamic parameter table lookup. The 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 , then the data characterizing the spatial distribution of background thermal influence on the functional unit can be expressed as The distribution of the functional unit in the area where it is located.

[0118] For each point in the feature point set P , from the background heat impact spatial distribution data Extract the corresponding background thermodynamic parameters ,Right now: .

[0119] Assume that there exists a function g, according to the background thermodynamic parameters Calculate correction value , then: ΔT_corr_p=g(T_bg_p) For example, it can be a simple proportional relationship , is the proportional coefficient.

[0120] Furthermore, step S221 includes:

[0121] S2211: Calculating a spatial gradient field of a thermodynamic parameter based on the first thermodynamic characteristic data of the functional unit;

[0122] S2212: Determine a specific position representing a heat source in the spatial gradient field as a feature point.

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

[0124] On this basis, specific locations in the spatial gradient field that characterize the heat source are determined as feature points. These locations are usually the key elements or areas that actually generate heat inside the functional unit. The feature points determined in this way can more accurately reflect the center of thermal behavior of the functional unit itself compared to simple preset positions or highest temperature points. The feature points accurately determined in this way are applied to subsequent steps. For example, after generating data that characterizes the spatial distribution of background thermal effects on the functional unit, the background thermodynamic parameters corresponding to these feature points are obtained, and correction values ​​are calculated based on these parameters. This ensures that the correction values ​​are calculated based on the background thermal effects on the heat generation positions of the functional unit itself. This makes the correction process more accurate, can more effectively remove background thermal effects, and thus more accurately assess the actual thermal state of the functional unit itself, improving the ability to distinguish between endogenous overheating and exogenous overheating.

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

[0126] The points with larger gradient modulus values ​​in the spatial gradient field are determined as feature points. For example, the points in the top k% of the gradient modulus values ​​are selected as feature points. Assuming that the gradient modulus threshold is Th, the feature point set P can be expressed as: These feature points are usually located at the heat source of the functional unit, such as the surface of the chip or the key heat-generating area inside.

[0127] Furthermore, step S3 includes:

[0128] S31: monitoring non-thermodynamic parameters associated with the working state of the circuit board;

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

[0130] S33: When the non-thermodynamic parameter meets a preset trigger condition, obtaining second thermodynamic characteristic data.

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

[0132] A preset trigger condition refers to a combination of logical rules or 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. This can be achieved by setting a threshold for a single non-thermodynamic parameter (e.g., CPU load > 80%), a combined threshold for multiple non-thermodynamic parameters (e.g., CPU load > 70% and memory usage > 90%), a parameter change rate threshold (e.g., a sharp increase in power consumption within a short period of time), or a specific system event (e.g., starting a high-load application).

[0133] Acquiring the second thermodynamic characteristic data means 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 a thermal image of the circuit board, collecting temperature data of multiple points through a temperature sensor array, or combining visible light images for alignment and calibration. Its 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 set up to periodically read the system's CPU usage and temperature sensor data from key areas on the circuit board (such as the CPU and VRM). Preset trigger conditions can be set: when CPU usage exceeds a parameter threshold for a continuous period of time, or when the temperature sensor reading in a key area exceeds a temperature threshold. When the monitoring program detects that any of these conditions are met, it is determined that the non-thermodynamic parameter meets the preset trigger condition. At this point, the system immediately triggers the thermal imaging camera to capture one or more consecutive thermal images to obtain secondary thermodynamic characteristic data of the circuit board under the current high load state. This data is then transmitted to a subsequent analysis module for comparison with pre-calculated expected thermodynamic characteristic data to detect any hot spots caused by synergistic effects.

[0135] Furthermore, step S32 includes:

[0136] S321: comparing the value of the non-thermodynamic parameter with a preset parameter threshold;

[0137] S322: When the value of the 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 parameters meet the preset trigger conditions based on the comparison results; if the comparison result shows that the duration exceeds the preset time threshold, determine that the non-thermodynamic parameters meet the preset trigger conditions; otherwise, determine that the non-thermodynamic parameters do not meet the preset trigger conditions.

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

[0140] The preset parameter threshold refers to the preset limit value used to preliminarily determine whether the value of the non-thermodynamic parameter has reached a level that requires attention. It can be set based on the design specifications of the circuit board, historical operating data or experience. Its purpose is to filter out parameter fluctuations within the normal range.

[0141] The preset time threshold is used to determine whether a non-thermodynamic parameter exceeds the threshold for a sufficient period of time. This threshold can be set based on the circuit board's response characteristics, thermal inertia, or empirical experience to avoid false triggering due to transient interference or short-term fluctuations. The duration of the non-thermodynamic parameter's continuous or cumulative value exceeding the preset threshold is the length of time. This duration can be measured using a timer, counter, or timestamp recording method to reflect the stability or severity of the abnormal non-thermodynamic parameter state.

[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. The preset parameter threshold can be set to 80%, and the 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 value is compared with 80%. If the real-time load rate is less than or equal to 80%, it is determined that the trigger condition is not met. If the real-time load rate exceeds 80%, the system starts counting and records the length of time the load rate continues to exceed 80%. Then, the recorded duration is compared with the preset time threshold of 5 seconds. If the duration is less than or equal to 5 seconds, it is determined that the trigger condition is not met. Only when the real-time load rate exceeds 80% and the duration of the duration exceeding 80% also exceeds 5 seconds, is it finally determined that the non-thermodynamic parameter (CPU load rate) meets the preset trigger condition, at which point the system triggers the acquisition of the second thermodynamic characteristic data of the circuit board.

[0143] Furthermore, step S4 includes:

[0144] S41: generating spatial distribution data representing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data;

[0145] S42: Identifying, in the spatial distribution data, a continuous region consisting of data points whose difference exceeds a preset discrimination threshold;

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

[0147] In some preferred embodiments, the present application is specifically implemented as follows:

[0148] First, obtain second thermodynamic characteristic data of the circuit board measured under the simultaneous operation of multiple functional units. This data can be a two-dimensional temperature matrix, where each element represents the temperature at a corresponding location on the circuit board. Simultaneously, obtain 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 in the absence of synergistic effects.

[0149] Next, by performing a point-by-point subtraction between the second thermodynamic characteristic data matrix and the expected thermodynamic characteristic data matrix, we generate spatial distribution data representing the difference between the two. This data is also a two-dimensional matrix, with each element representing the temperature difference at the corresponding location. Next, we set a preset threshold, such as 5°C. We traverse the spatial distribution data matrix of differences and identify all data points with a difference greater than 5°C.

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

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

[0152] Furthermore, step S42 includes:

[0153] S421: rasterizing the spatial distribution data to generate rasterized data;

[0154] S422: Searching for a set of data points whose difference exceeds a discrimination threshold and whose spatial positions are adjacent in the rasterized data;

[0155] S423: Determine the set as a continuous area.

[0156] Among them, rasterization processing refers to the process of converting original spatial data that may have arbitrary resolution or distribution into a discrete data set composed of regular grid (raster) units. It can be achieved by dividing the spatial area into uniform square or rectangular grids and aggregating the original data in each grid unit (for example, calculating the average, maximum or sampling center point value). Its purpose is to structure and discretize spatial data to facilitate subsequent calculations and analysis.

[0157] Rasterized data refers to a data set that is formed after rasterization processing and is organized in grid units. Each grid unit stores the attribute value of the corresponding spatial area (such as difference data). 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 positions and the judgment of adjacent relationships.

[0158] A set of data points whose difference exceeds the discrimination threshold and whose spatial positions are adjacent refers to one or more data point groups in rasterized data, which satisfy the requirement that their stored difference is greater than the preset discrimination threshold and these data points are adjacent to each other in the raster grid (for example, adjacent in the up, down, left, right, or diagonal directions). They can be identified using the connected component analysis algorithm or the region growing algorithm, with the aim of extracting abnormal areas with spatial aggregation from discrete raster data.

[0159] A continuous region refers to an area with a certain range and shape composed of spatially connected data points. These data points jointly exhibit certain specific properties (such as the difference exceeds a threshold). It can be represented by marking or boundary extraction. Its purpose is to abstract a discrete set of data points into a meaningful and identifiable physical area.

[0160] In some preferred embodiments, specifically, the spatial distribution data is rasterized S421, which can be dividing the image or matrix representing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data into a two-dimensional grid according to a preset grid size (for example, each square millimeter is divided into a number of grid cells), and for each grid cell, calculating the average value or maximum value of the difference of all original data points in the cell, and using the value as the data of the grid cell, thereby generating rasterized data.

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

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

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

[0164] The first acquisition module 201 is used to acquire first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they work independently.

[0165] Generating module 202: generating expected thermodynamic characteristic data representing the linear superposition of independent thermal contributions of at least two functional units based on the first thermodynamic characteristic data;

[0166] The second acquisition module 203 is configured to acquire second thermodynamic characteristic data of the circuit board under the condition that at least two functional units work simultaneously.

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

[0168] The acquisition module 201 is a unit for collecting thermodynamic information of the circuit board, which can be implemented using a thermal imaging camera, an infrared sensor array, or a temperature probe, etc. Its purpose is to obtain the thermal performance of the functional unit when it works independently.

[0169] The generation module 202 refers to a unit for performing calculations and modeling based on input data, which can be implemented using a processor, a computing server, or a dedicated integrated circuit, etc., and its purpose is to simulate the linear superposition of independent thermal contributions of functional units;

[0170] The second acquisition module 203 is a unit for acquiring 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 thermal performance of the functional units when they are working simultaneously.

[0171] The detection module 204 refers to a unit for analyzing and comparing thermodynamic data and making judgments, which can be implemented by a processor, an algorithm execution unit, or a comparison circuit, etc. Its purpose is to identify the difference between the actual thermal distribution and the predicted thermal distribution;

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

[0173] Based on the above modules and data, the solution of the present application implements each step of the motherboard image hotspot detection method based on thermal imaging into an acquisition module 201, a generation module 202, a second acquisition module 203 and a detection module 204, thereby constructing a system that can automatically execute 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. Based on this independent data, the generation module 202 calculates the theoretical thermal distribution in the absence of mutual influence. The second acquisition module 203 captures the actual overall thermal distribution when the functional units operate simultaneously. Finally, the detection module 204 compares the difference between the actual thermal distribution and the theoretical thermal distribution to accurately identify those areas of additional heat accumulation caused by the mutual influence of functional units, that is, hot spots caused by the synergistic effect.

[0175] This systematic approach enables efficient and accurate execution of previously complex and manually-reliant method steps, transforming abstract detection logic into actionable system functions. This overcomes the challenges of relying solely on methodologies for practical implementation. This transformation of method steps into system modules makes large-scale, high-precision hotspot detection for circuit boards possible, significantly improving detection efficiency and reliability.

[0176] In this document, relational terms such as first and second, etc. are used merely to distinguish one entity or operation from another entity or operation, but do not necessarily require or imply any actual relationship or order between these entities or operations.

[0177] The foregoing is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. Persons skilled in the art will readily appreciate that the present application may be modified and altered in various ways. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A method for detecting hot spots in motherboard images based on thermal imaging, characterized in that: The method comprises the steps of: S1: Acquire first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they operate independently; S2: generating expected thermodynamic characteristic data representing the linear superposition of independent thermal contributions of the at least two functional units based on the first thermodynamic characteristic data; S3: Acquiring second thermodynamic characteristic data of the circuit board under the condition that the at least two functional units operate simultaneously; S4: Based on the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data, determining whether the circuit board has a hot spot caused by a synergistic effect.

2. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 1, characterized in that: Step S1 includes: S11: For any functional unit of the at least two functional units: Before stimulating the functional unit, obtaining pre-stimulation thermodynamic data; After stimulating the functional unit, obtaining post-stimulation thermodynamic data; S12: Determine first thermodynamic characteristic data of the functional unit based on the difference between the thermodynamic data after excitation and the thermodynamic data before excitation.

3. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 1, characterized in that: Step S2 includes: S21: for each of the at least two functional units, determining a physical quantity representing a background thermal influence applied to the functional unit based on first thermodynamic characteristic data of other functional units except the functional unit; S22: determining a correction value for adjusting first thermodynamic characteristic data of the functional unit based on the physical quantity characterizing the background thermal influence; S23: applying the correction value to the first thermodynamic characteristic data of the functional unit to generate adjusted heat contribution data of the functional unit; S24: Superimposing the adjusted heat contribution data of the at least two functional units to generate the expected thermodynamic characteristic data.

4. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 3, characterized in that: Step S22 includes: S221: Determine a characteristic point of the functional unit that represents a heat generation location of the functional unit; S222: Generate data representing the spatial distribution of background thermal influence on the functional unit as the physical quantity representing the background thermal influence; S223: Obtaining background thermodynamic parameters corresponding to the feature points in the data representing the spatial distribution; S224: Calculate the correction value based on the background thermodynamic parameter.

5. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 4, characterized in that: Step S221 includes: S2211: Calculating a spatial gradient field of a thermodynamic parameter based on the first thermodynamic characteristic data of the functional unit; S2212: Determine a specific position representing a heat source in the spatial gradient field as the feature point.

6. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 1, characterized in that: Step S3 includes: S31: monitoring non-thermodynamic parameters associated with the working state of the circuit board; S32: Determine whether the non-thermodynamic parameter meets a preset trigger condition; S33: When the non-thermodynamic parameter satisfies the preset trigger condition, acquiring the second thermodynamic characteristic data.

7. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 6, characterized in that: Step S32 includes: S321: Comparing the value of the non-thermodynamic parameter with a preset parameter threshold; S322: When the value of the non-thermodynamic parameter exceeds the 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; S323: Determine whether the non-thermodynamic parameter meets the preset trigger condition based on the comparison result; when the comparison result is that the duration exceeds the preset time threshold, determine that the non-thermodynamic parameter meets the preset trigger condition; otherwise, determine that the non-thermodynamic parameter does not meet the preset trigger condition.

8. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 1, characterized in that: Step S4 includes: S41: generating spatial distribution data representing the difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data; S42: Identifying a continuous region in the spatial distribution data consisting of data points whose difference exceeds a preset discrimination threshold; S43: Determine the continuous region as a hot spot caused by the synergistic effect.

9. The method for detecting hot spots in motherboard images based on thermal imaging according to claim 8, characterized in that: Step S42 includes: S421: performing rasterization processing on the spatial distribution data to generate rasterized data; S422: searching the rasterized data for a set of data points whose difference exceeds the discrimination threshold and whose spatial positions are adjacent to each other; S423: Determine the set as the continuous area.

10. A motherboard image hotspot detection system based on thermal imaging, characterized in that: In the steps of the method according to any one of claims 1 to 9, the system comprises: A first acquisition module is configured to acquire first thermodynamic characteristic data corresponding to at least two functional units on the circuit board when they operate independently; A generating module: generating, based on the first thermodynamic characteristic data, expected thermodynamic characteristic data representing the linear superposition of independent thermal contributions of the at least two functional units; A second acquisition module: acquiring second thermodynamic characteristic data of the circuit board under the condition that the at least two functional units work simultaneously; A detection module is configured to determine whether the circuit board has a hot spot caused by a synergistic effect based on a difference between the second thermodynamic characteristic data and the expected thermodynamic characteristic data.

Images