A method and system for automatically controlling and supervising chip functions

Abstract

The application discloses a method and system for automatically controlling and supervising chip functions, relates to the field of machine vision detection, and is especially suitable for functional detection and active protection of a chip module in a complex application environment. The method for automatically controlling and supervising chip functions comprises the following steps: placing the chip module in a temperature box and connecting a test circuit; guiding the temperature box to locally heat based on position information of each element in the chip module, so as to obtain a temperature-position influence infrared image set of all elements; starting detection and acquiring an original infrared time sequence image sequence of the chip module; correcting the original infrared time sequence image sequence based on the temperature-position influence infrared image set of each element position and an environmental temperature in the original infrared time sequence image sequence, so as to obtain a self-heat infrared time sequence image sequence of the chip module; and supervising the elements in the self-heat infrared time sequence image sequence based on a time sequence model. Through the above steps, the application accurately reflects functional changes of chip elements, effectively eliminates interference of the environment and other elements, and greatly improves the accuracy of chip health monitoring.

Description

Technical Field

[0001] This invention relates to the field of machine vision inspection, specifically to a method and system for automatically controlling and monitoring chip functions, which is particularly suitable for the functional inspection and active protection of chip modules in complex application environments. Background Technology

[0002] With the rapid development of artificial intelligence technology, intelligent chips have been widely used in high-performance and high-security scenarios such as cloud computing, autonomous driving, and smart terminals. In modern industrial production processes, functional testing and temperature monitoring of chips have become crucial links in ensuring the safe and stable operation of systems. After prolonged operation, the internal semiconductor structure, nodes, and other critical components of a chip undergo irreversible aging effects, leading to significant deviations between actual operating physical parameters (such as energy consumption, heat generation, and operating current) and factory calibration values. If the chip is mechanically managed and scheduled according to the factory calibration parameters during subsequent use, it can easily lead to major safety risks such as overheating, performance degradation, or even failure. Therefore, how to dynamically and accurately detect the actual operating status of a chip throughout its entire lifecycle and adjust management measures accordingly has become a pressing technical challenge for the industry.

[0003] Existing chip testing technologies mainly fall into two categories. The first involves separate testing of the functionality and performance indicators of individual chip components. This type of testing is often conducted in an idealized laboratory environment, effectively eliminating interference from the external environment and other components, and obtaining the performance limits of the chip component itself. However, in practical applications, chips often operate in the form of modules or integrated circuit boards, where the surrounding environment (such as temperature, humidity, and electromagnetic interference) is complex and variable, and there are interactions with other chips or components in terms of energy, heat, and noise. In this case, test data obtained under ideal conditions often cannot accurately reflect the functional performance of the chip component under actual operating conditions. The heat generation characteristics, aging rate, and heat dissipation efficiency of semiconductor materials will change due to environmental differences, resulting in limited reference value of laboratory test conclusions in practical applications. The second approach involves functional testing of the entire chip module, that is, testing a complete module integrating various types of components. Due to the coupling and interaction between the components within the module, the temperature rise of a single chip component is not only affected by its own characteristics but also fluctuates due to the heat generation of nearby components and changes in the external environment. Existing technologies struggle to effectively isolate the influence of these complex factors on target chip components, leading to significant discrepancies in test results and an inability to accurately reflect the true functionality and health status of the target component. Therefore, both discrete component testing and overall module testing suffer from prominent issues such as inconsistencies between the testing environment and the actual operating environment, complex and difficult-to-decouple influencing factors, and insufficient reliability and specificity of test results. These problems directly impact the accuracy of chip functional assessment and the scientific rigor of subsequent management measures, ultimately posing potential risks to the safe operation of chips throughout their entire lifecycle.

[0004] To address the shortcomings of existing technologies, this invention proposes a method and system for automatically controlling and monitoring chip functions. This system acquires thermal distribution information of the chip module under real-world operating conditions based on infrared image sequences, and combines this with standardized testing conditions to obtain the actual heat generation of the semiconductor materials inside the chip. By introducing a time-series network model, dynamic thermal features in the infrared image sequences are extracted and analyzed, enabling real-time prediction of changes in the chip's operating conditions. This system can proactively predict risks such as abnormal chip overheating, dynamically adjust chip operating parameters, or issue protection commands, thereby achieving intelligent management and proactive protection throughout the chip's entire lifecycle. Compared to existing detection methods, this invention can accurately reflect functional changes in chip components without affecting normal chip operation, effectively eliminating interference from the environment and other components, and greatly improving the accuracy of chip health monitoring and the safety of engineering applications. Summary of the Invention

[0005] In view of this, the present invention provides a method for automatically controlling and supervising chip functions, the method specifically including the following steps: S1: Place the chip module in the temperature chamber and connect the test circuit; S2: Based on the position information of each component in the chip module, the temperature chamber is guided to perform local heating, and an infrared image set of the temperature-position influence of all components is obtained; S3: Start detection and acquire the original infrared timing image sequence of the chip module; S4: Based on the temperature-position influence infrared atlas of each component location and the ambient temperature in the original infrared timing image sequence, correct the original infrared timing image sequence to obtain the self-heating infrared timing image sequence of the chip module; S5: Based on the time series model, perform heat generation supervision on the components in the self-generating infrared time series image sequence.

[0006] The present invention also provides a system for automatically controlling and monitoring chip functions, the system comprising: Temperature chamber: Place the chip module in the temperature chamber and connect the test circuitry; Local heating control module: Based on the position information of each component in the chip module, it guides the temperature chamber to perform local heating, and obtains an infrared image set of the temperature-position influence of all components; Infrared temporal image sequence acquisition module: Start detection and use an infrared thermal imager to acquire the original infrared temporal image sequence of the chip module; Image correction module: Based on the temperature-position influence infrared atlas of each component location and the ambient temperature in the original infrared time-series image sequence, correct the original infrared time-series image sequence to obtain the self-heating infrared time-series image sequence of the chip module; Supervision module: Based on the time series model, it performs heat generation supervision on the components in the infrared time series image sequence of its own heat generation.

[0007] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for automatically controlling and monitoring chip functions.

[0008] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for automatically controlling and monitoring chip functions.

[0009] Compared with existing technologies, this invention discloses a method and system for automatically controlling and monitoring chip functions. First, it accurately decouples and quantifies the actual heat generation contribution of each independent component within the chip module. Then, based on this self-heating data, it utilizes time-series and correlation analysis models to achieve high-precision, intelligent monitoring and early warning of anomalies for the operating status of each component. This method overcomes the fundamental problem in traditional chip thermal management and fault diagnosis of the difficulty in distinguishing between the component's own heat generation and the thermal influence of the environment and neighboring components. Specifically, this invention, through dual-cycle temperature-position influence infrared atlas calibration and triple-cycle correction algorithms, can extract the net heat generated by each component due to its own workload from complex, mixed raw thermal maps. This allows the control system to clearly and accurately understand which components are truly operating under high load and which components may have abnormal heat generation, rather than simply being passively heated. Based on the infrared time-series image sequence and time-series model of self-generated heat, it is possible not only to detect anomalies in the heat generation time of individual components, but also to analyze the complex correlation between the heat generation modes of different components. The analysis of these self-time-series characteristics and mutual mode characteristics at the self level can reveal chip self-degradation, hardware defects and system-level collaborative anomalies that are difficult to detect by traditional methods. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating the automatic control and monitoring of chip functions in this application. Detailed Implementation

[0012] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0013] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] It should be noted that various aspects of embodiments within the scope of the appended claims are described below. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0015] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.

[0016] The technical solutions provided by the various embodiments of this application are described below with reference to the accompanying drawings.

[0017] This specification provides a method for automatically controlling and monitoring chip functions, which specifically includes the following steps: S1: Place the chip module in the temperature chamber and connect the test circuit; S2: Based on the position information of each component in the chip module, the temperature chamber is guided to perform local heating, and an infrared image set of the temperature-position influence of all components is obtained; S3: Start detection and acquire the original infrared timing image sequence of the chip module; S4: Based on the temperature-position influence infrared atlas of each component location and the ambient temperature in the original infrared timing image sequence, correct the original infrared timing image sequence to obtain the self-heating infrared timing image sequence of the chip module; S5: Based on the time series model, perform heat generation supervision on the components in the self-generating infrared time series image sequence.

[0018] In this embodiment, the temperature chamber is an integrated high-precision environmental control device specifically designed for local and overall temperature regulation and infrared imaging detection of the chip module. The main body of the chamber employs a double-layer thermal insulation structure. The outer shell can be made of high-strength engineering plastics or metal materials, while the inner liner is constructed from materials with low thermal conductivity and high infrared reflectivity, such as polished stainless steel or coated alloys. This aims to minimize internal and external heat exchange and ensure the uniformity and stability of the internal temperature environment. Inside the temperature chamber is a chip module support platform, preferably constructed from insulating materials with a low coefficient of thermal expansion, such as ceramics or special composite materials. The platform integrates fixtures for mounting the chip module under test and provides ample wiring space and multiple electrical signal feedthrough interfaces. All wiring is introduced through a sealed cable tray at the rear of the temperature chamber. The main control system of the temperature chamber has multiple signal channels, enabling communication with an external host computer system to achieve unified management and automated control of the temperature chamber heating process, camera acquisition, and chip module testing status. The system supports remote setting of overall and local temperature parameters of the temperature chamber, and can record infrared and visible light image sequences of the chip module under different temperature control conditions in real time.

[0019] To achieve precise simulation and control of the operating environment temperature of the chip module, a global temperature control system is installed inside the chamber. This system includes an air circulation fan, evenly distributed resistance heating elements, and a Peltier semiconductor cooling module, and monitors the air temperature inside the chamber in real time via at least one temperature sensor. This temperature sensor is connected to the main control system of the chamber, and through a PID control algorithm, precisely adjusts the power output of the heating elements and the operating state of the cooling module. This enables rapid setting, stable maintenance, and programmed heating and cooling of the overall ambient temperature inside the chamber over a wide temperature range. Preferably, the internal temperature adjustment range of the chamber is set to -40°C to +60°C.

[0020] Directly beneath the chip module's mounting platform, an array of independently addressable and controllable micro-local heating units is integrated. These local heating units, preferably micro-ceramic heating elements, thin-film resistance heaters, or small thermoelectric module arrays, are connected to a multi-channel precision power control system. This system can independently and precisely control the power output of one or more selected local heating units based on component location information and testing requirements. This enables precise, rapid, and independent heating of specific locations on the chip module. Component location information can be obtained from the chip design layout or through visible light visual inspection.

[0021] To enable real-time, non-contact monitoring of the temperature distribution of the chip module during testing, one or more optical observation windows are provided on the top of the chamber. At least one window is made of a special material with high transmittance in the mid- and far-infrared bands. This infrared window faces the chip module's mounting platform, allowing an infrared thermal imager to capture images of the chip module's complete surface temperature distribution under various operating conditions. For further analysis, preferably, the chamber also includes another visible light observation window. Combined with an adjustable LED lighting source inside the chamber, a visible light camera can capture images of the chip module's physical morphology for positioning, alignment, or status assessment. The infrared thermal imager and visible light camera are fixed to a precision adjustment bracket outside the chamber via a standard interface. All data output interfaces of the imaging devices are led out to the outside of the chamber and connected to the image acquisition and processing system.

[0022] In addition, the chamber is equipped with a dry nitrogen purging port to prevent internal frost buildup, an over-temperature protection device, and a door locking mechanism to ensure operational safety and airtightness. All controllable components of the entire chamber system, including global temperature control, local heating unit array control, and internal lighting control, can be connected to the main control system through a unified control interface to achieve automated control of the testing process and synchronous data acquisition.

[0023] To further illustrate the structure and working principle of the local heating unit in this invention, a specific embodiment based on a thermocouple module array is provided. In this embodiment, the local heating unit array does not employ traditional resistance wire heating, but rather utilizes the precise, high spatial resolution temperature control characteristics of the semiconductor cooling chip in reverse working (i.e., heating) mode to achieve targeted heating of any critical component location on the chip module.

[0024] In this embodiment, the localized heating unit array is integrated directly beneath the chip module carrier platform, with a high-density matrix of miniature TEC modules at its core. The array's substrate is made of a ceramic material with high thermal conductivity, such as aluminum nitride or beryllium oxide. This high thermal conductivity ensures that waste heat generated by the TEC modules during operation is rapidly and uniformly conducted to the shared heat dissipation system beneath the carrier platform, preventing heat accumulation within the substrate and thus maintaining the temperature baseline stability and operational lifespan of the entire heating array. The ceramic substrate is processed such that its upper surface, facing the chip module carrier platform, is polished to maximize thermal contact efficiency.

[0025] On the upper surface of the ceramic substrate, micro-TEC modules are mounted in an array using thermally conductive silver paste. These TEC modules are small in size, preferably between 3mm x 3mm and 5mm x 5mm in size, forming a heating matrix with high spatial resolution. For example, a 50mm x 50mm area can integrate an array of 10x10, or 100, 5mm x 5mm TEC modules.

[0026] Each TEC module consists of multiple P-type and N-type semiconductor thermocouples connected in series, sandwiched between two ceramic substrates. When a direct current is applied, one ceramic substrate absorbs heat, while the other releases heat, depending on the direction of the current. In this invention, we utilize its heating function to independently control the temperature of its heated surface by precisely controlling the direction and magnitude of the current applied to each TEC module.

[0027] The hot side of the TEC module array faces upwards, directly towards the bottom of the chip module carrier platform. To eliminate the tiny air gaps between the TEC modules and the carrier platform, a thin, uniform layer of high-performance thermal interface material is used. Preferably, a phase change thermal conductive material is employed, which changes from a solid to a semi-fluid state at a specific temperature. This material can fill all microscopic gaps, achieving near-zero interfacial thermal resistance. This ensures that the heat generated by the TEC modules can be efficiently and losslessly transferred to the carrier platform and ultimately applied precisely to the corresponding location on the chip module under test.

[0028] This embodiment employs a matrix addressing architecture to control the TEC array. Multiple layers of wiring are etched onto the ceramic substrate, forming row and column drive lines. Each TEC module's two electrodes are connected to a specific row and column line, respectively. When the central control system needs to activate a TEC module at a specific location, it simply applies a drive signal to the corresponding row and column line. For each selected TEC module, the system does not simply apply a constant DC current, but instead uses high-frequency pulse-width modulation (PWM) for power regulation. The central control system outputs a PWM signal to the drive unit of the corresponding TEC module. By adjusting the duty cycle of the PWM signal, the average power delivered to the TEC module per unit time can be controlled. A higher duty cycle results in greater heating power and a higher target temperature.

[0029] To achieve precise and stable target temperature, this system employs a closed-loop feedback control strategy. A miniaturized temperature sensor is integrated at regular intervals within the TEC module array. Preferably, miniature NTC thermistors or thin-film platinum resistance thermometers are placed at intervals between every four or nine TEC modules. The sensors and TEC modules coexist on the same substrate, enabling real-time and accurate measurement of the actual temperature of localized areas.

[0030] When performing a heating task at a component location, the control system reads the values ​​from one or more miniature temperature sensors closest to that location. This real-time temperature value is input as a process variable to a dedicated PID controller. The setpoint of the PID controller is the target temperature value required for the current inner loop. The PID controller calculates the error between PV and SP in real time and dynamically adjusts the duty cycle of the output to the TEC module at that location based on the error until the error is eliminated. This closed-loop control method overcomes the influence of individual TEC module differences and minor environmental fluctuations, ensuring the accuracy and stability of the heating temperature and meeting thermal steady-state conditions.

[0031] Based on the location information of each component in the chip module, the temperature chamber is guided to perform localized temperature increases, including: Number the N key components in the chip module and obtain their corresponding location information. ; Set M discrete local target heating temperature values. and overall temperature, wherein the overall temperature is the operating ambient temperature of the chip module; A double loop structure is used to traverse the locations of N key components and M discrete local target heating temperature values ​​to obtain M temperature-location effect infrared images for each component, forming a set of temperature-location effect infrared images for all components. Specifically, the infrared image set that employs a dual-loop structure to compose all components and its temperature-position influence includes: Outer loop: Select the position of the component to be calibrated. , The central control system is based on Activate the local heating unit at its corresponding location.

[0032] Inner loop: For the currently activated local heating unit, select the target heating temperature value sequentially. Adjust the power applied to the local heating unit so that the temperature of the corresponding area gradually increases and stabilizes at the target heating temperature. In the inner loop, under thermal steady-state conditions, the temperature distribution of the entire chip module is collected when each component is heated to the target temperature value. The inner loop also includes a steady-state determination process: During the heating process, an infrared thermal imager continuously captures infrared thermal images of the entire chip module at preset time intervals; the system monitors the temperature distribution changes of the entire module in real time, and analyzes the position of the component to be calibrated in the continuous infrared images. Temperature changes determine whether a thermal steady state has been reached.

[0033] The thermal steady state is set such that, within a preset time window, the average temperature change rate of the pixels in the current component location area to be calibrated is less than a preset temperature threshold.

[0034] Once thermal steady state is confirmed, that is, in the local heating unit... Under stable heating conditions, the temperature distribution in the area of ​​the component to be calibrated no longer changes significantly. The central control system then triggers the infrared thermal imager to acquire and store an infrared image. The infrared image Records when the k-th element is heated to At that time, the temperature distribution of the entire chip module, ; The inner circulation also includes a cooling and recirculation initiation process: Complete an infrared image After data acquisition, the central control system shuts down the current local heating unit. Once the entire chip module cools to the ambient operating temperature, the inner layer circulation continues, applying the same method to the same component location. Apply the next target temperature rise value Repeat the heating, steady-state judgment, infrared image acquisition and storage until all M target heating temperature values ​​have been completed.

[0035] The outer loop also includes: when the component position After all M target heating temperature values' infrared images have been acquired, the outer loop enters the next iteration, selecting a new component location. And repeat all the operations of the inner loop.

[0036] This invention quantifies and calibrates the thermal impact of chip modules. It actively and controllably heats each key component independently while simultaneously recording the infrared thermal image of the entire module, thus constructing a comprehensive thermal impact database—a temperature-location impact infrared atlas. This atlas records the extent and degree to which the heat from any specific component affects all other components and areas within the module, providing crucial calibration data for accurately isolating the thermal effects between components. This atlas forms the basis of the correction algorithm. Without this atlas, it is impossible to accurately calculate and deduct the contribution from neighboring heating components to the observed temperature of a component during actual operation. This solves the problem that traditional infrared thermal imaging can only see the surface mixed temperature. Simultaneously, this invention enhances the understanding of the thermal characteristics of specific chip modules through independent experiments. Chip modules with different packages and layouts exhibit vastly different internal heat transfer characteristics. This invention establishes unique thermal impact model parameters for each specific module through actual measurements, avoiding reliance on general, inaccurate theoretical models or empirical estimations.

[0037] The detection process begins, and the original infrared time-series image sequence of the chip module is acquired, including: The overall temperature control system of the temperature chamber is used to adjust the overall ambient temperature inside the temperature chamber to the operating ambient temperature, which is the room temperature of the chip module's working environment. During the overall ambient temperature regulation process, the heating device and air circulation fan inside the chamber work together, and the temperature sensor provides feedback to ensure that the air temperature inside the chamber reaches the operating ambient temperature and remains stable. Once the temperature in the chamber stabilizes at the ambient operating temperature, the original infrared timing image sequence of the chip module is acquired. The acquisition of the original infrared time-series image sequence of the chip module includes: Set the start time, apply power to the chip module through the test system, and execute the predetermined workload to trigger the infrared thermal imager to start continuously capturing infrared thermal images of the chip module surface.

[0038] The workload can be an application program that is actually working on the chip module, a standard benchmark suite such as SPEC CPU or Linpack, or a series of preset instruction sequences that simulate user operations.

[0039] As the workload is executed, the various key components inside the chip module begin to perform calculations and data processing, generating heat as a result. This heat is conducted to the module surface through the chip's heat dissipation structure, causing dynamic changes in the surface temperature distribution.

[0040] The infrared thermal imager generates raw infrared image frames at a set frame rate. , where i is the image frame number from the start time, and each frame All of these accurately reflect the complete temperature distribution on the surface of the chip module at a precise point in time.

[0041] This invention records the complete surface temperature field of a chip over time as it executes an application or specific test program. The infrared time-series image sequence not only shows the spatial distribution of heat on the chip surface but also reveals patterns of heat change over time. This raw two-dimensional data is far richer in information than that from traditional single-point temperature sensors.

[0042] After acquisition, a preliminary integrity check is performed on the original infrared time-series image sequence. This integrity check includes checking for missing frames and the continuity of timestamps. Once confirmed to be correct, the original infrared time-series image sequence, along with all relevant metadata, is stored. The metadata should include at least: the set overall operating temperature of the chamber, a detailed description or identifier of the executed workload, the model and acquisition parameters of the infrared thermal imager, the start and end times of the test, and the model of the chip module. Based on the temperature-position effect infrared atlas of each component location and the ambient temperature in the original infrared timing image sequence, the original infrared timing image sequence is corrected to obtain the self-heating infrared timing image sequence of the chip module. A triple loop structure is used to obtain the self-heating infrared time-series image sequence of the chip module, wherein the first loop iteration includes extracting any one frame of the original infrared time-series image sequence. Generate a corresponding corrected image for one frame. The correction process targets N key components. Each critical component needs to have its thermal effects from all other components calculated and removed.

[0043] For each raw infrared image frame The correction process sets up a second loop structure, which iterates through all the critical components to be corrected. For the currently selected key components To obtain this key component Corresponding position Average observed temperature ; For each critical component to be corrected The correction process establishes a third loop structure, which iterates through all other critical components. , , ; Each key component to be corrected is modified based on the third loop structure. include: For other key components To obtain this key component exist Corresponding position Average observed temperature Accessing the temperature-location influence infrared dataset corresponds to key components. The M infrared images are used to search for... Position corresponding to M infrared images The average temperature in the current frame Average observed temperature The closest one Calculations yielded At temperature Time The resulting temperature rise is: ; in, This represents the effect of critical component j on the temperature rise of critical component k in the i-th frame. This indicates the target temperature rise value for critical component j. The acquired temperature-location effect influences the average temperature at position k, a key element in the infrared image. Operating ambient temperature; The calculated key components For key components Thermal influence Added to the total external heat impact: ; in, Key element of frame i The total temperature rise effect; The third cycle continues, focusing on the same key component. Apply the next critical component, and continue until all other critical components have been traversed, to obtain the total temperature rise effect of all other components on critical component k. .

[0044] Utilizing key components Location Average observed temperature Obtain the corrected temperature at this location: ;in, Position in the i-th frame The corrected temperature; Each original infrared image frame is corrected based on the second loop structure. Also includes: Set the original infrared image frame Each critical element in the process is a critical element to be corrected, and a third loop structure is executed until the original infrared image frame is reached. Each critical component in the process completed the thermal impact elimination of other critical components, resulting in the corresponding corrected image. ; Repeated execution for a single frame image The correction process continues until every frame in the original infrared time-series image sequence has been processed and its corresponding corrected image has been generated. All the corrected image frames are combined in chronological order to form the self-heating infrared time-series image sequence of the chip module, denoted as... , This represents the number of frames in the original infrared time-series image sequence. Each frame in the self-heating infrared time-series image sequence of the chip module represents the net heat contribution of each key component due to its own operation at a specific moment.

[0045] Through the above corrections, this invention achieves precise separation and extraction of the heat generated by each component itself. Utilizing the temperature-position influence infrared image set obtained in S2, the original infrared time-series image sequence acquired in S3 is corrected, effectively eliminating contamination of the observed temperature of the current target component by other heat-generating components through their heat dissipation. The resulting self-heating infrared time-series image sequence more accurately reflects the net heat generated by each independent component due to its own operating power consumption. In traditional methods, a component with a high surface temperature might have a high self-heating rate, or it might be due to its proximity to a large heat-generating component. This correction step allows for a clear distinction between these two situations, thereby accurately identifying the component whose abnormal heating is truly due to its own problem and avoiding misjudgment.

[0046] The heat generation supervision of components in the self-generating infrared time-series image sequence based on the time-series model includes: Extraction of two-dimensional image block sequences of components; Detection of abnormal heat generation in individual component regions based on ConvLSTM autoencoder; Correlation analysis of changes in the self-heating modes of key component regions; Generation and output of heat generation monitoring information.

[0047] The extraction of two-dimensional image block sequences of components includes extracting each key component. ,from Each frame of the image Based on its location information Extract the corresponding two-dimensional image patch Therefore, for each key component Construct its independent two-dimensional image patch time series ; After the extraction of the two-dimensional image block sequence of the element, sliding window segmentation is also included; a fixed-length time window is defined. and sliding step size At any time ,element The input data is a sequence of image patches. ; To enable the analysis and anomaly detection of the heat generation behavior of each critical component region, this embodiment provides a method for each component. The image patch sequence is configured with an independent convolutional long short-term memory network autoencoder, denoted as . . Corresponding components The data was obtained by training on data generated under normal, fault-free operating conditions.

[0048] encoder in Composed of ConvLSTM layers, the encoder take over A sequence of image patches As input, spatiotemporal features are extracted through stacking multiple ConvLSTM layers, and the image patch sequence is compressed and flattened into a one-dimensional feature vector. , , decoder in Its structure is symmetrical to that of the encoder; it receives a one-dimensional feature vector. As input, it is decoded and reconstructed to produce the original image patch sequence. . .

[0049] For image patch sequences Input it into the pre-trained In the process, the reconstructed image patch sequence window is obtained. Calculate the reconstruction error of the current window as its spatiotemporal anomaly score. The abnormal score will be compared with the component. Preset Compare. If Then determine the component Region in the current time window The internal heat generation behavior exhibits spatiotemporal anomalies, generating an anomaly marker. Otherwise .

[0050] For the current analysis window Calculate any two different components and , , one-dimensional eigenvector and similarity between ,when , If the similarity threshold is used, then the component is considered... and The spatiotemporal patterns of regional heat generation within the current window show significant correlations.

[0051] This invention utilizes a timing model to process two-dimensional self-heating infrared timing image sequences. It can learn and identify the spatiotemporal dynamic patterns of self-heating in each component region under normal operating conditions. Compared to simple threshold judgment, the timing model can capture subtle temporal changes in the heating patterns, such as abnormal heating rise rates and irregular heating fluctuations. These anomaly detections based on self-timing characteristics can identify potential functional degradation or early signs of failure in components earlier than traditional methods. By analyzing the correlation and synergy between the spatiotemporal patterns of self-heating in different component regions, it is possible to rule out factors beyond mutual influence of component heat and discover more complex system-level problems. Abnormal synchronous or asynchronous changes in the self-heating patterns of multiple components indicate problems with shared resources or chain reactions triggered by specific task flows within the chip. This analysis of mutual patterns provides system-level detection beyond the perspective of a single component.

[0052] The present invention also provides a system for automatically controlling and monitoring chip functions, the system comprising: Temperature chamber: Place the chip module in the temperature chamber and connect the test circuitry; Local heating control module: Based on the position information of each component in the chip module, it guides the temperature chamber to perform local heating, and obtains an infrared image set of the temperature-position influence of all components; Infrared temporal image sequence acquisition module: Start detection and use an infrared thermal imager to acquire the original infrared temporal image sequence of the chip module; Image correction module: Based on the temperature-position influence infrared atlas of each component location and the ambient temperature in the original infrared time-series image sequence, correct the original infrared time-series image sequence to obtain the self-heating infrared time-series image sequence of the chip module; Supervision module: Based on the time series model, it performs heat generation supervision on the components in the infrared time series image sequence of its own heat generation.

[0053] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for automatically controlling and monitoring chip functions.

[0054] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for automatically controlling and monitoring chip functions.

[0055] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0056] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for automatically controlling and monitoring chip functions, characterized in that, The method includes the following steps: S1: Place the chip module in the temperature chamber and connect the test circuit; S2: Based on the location information of each key component in the chip module, the temperature chamber is guided to perform local heating, and an infrared image set of the temperature-position influence of all key components is obtained; S3: Start detection and acquire the original infrared timing image sequence of the chip module; S4: Based on the temperature-position influence infrared atlas of each key component location and the ambient temperature in the original infrared time-series image sequence, correct the original infrared time-series image sequence to obtain the self-heating infrared time-series image sequence of the chip module. S5: Heat generation supervision of key components in the self-generating infrared time-series image sequence based on the time-series model; Step S2 includes: numbering the N key components in the chip module and obtaining their corresponding location information. Set M discrete local target heating temperature values. The overall temperature is the operating environment temperature of the chip module. A dual-loop structure is used to traverse the locations of N key components and M discrete local target heating temperature values ​​to obtain M temperature-location effect infrared images of each key component, forming a set of temperature-location effect infrared images of all key components.

2. The method for automatically controlling and monitoring chip functions according to claim 1, characterized in that, The infrared image set, which employs a dual-loop structure to compose all key components, specifically includes the following temperature-position influences: Outer loop: Select the location of the critical component to be calibrated. , The central control system is based on Activate the local heating unit at its corresponding location; Inner loop: For the currently activated local heating unit, select the target heating temperature value sequentially. Adjust the power applied to the local heating unit so that the temperature of the corresponding area gradually increases and stabilizes at the target heating temperature. .

3. The method for automatically controlling and monitoring chip functions according to claim 2, characterized in that, In the inner loop, under thermal steady-state conditions, the temperature distribution of the entire chip module is collected when each key component is heated to the target temperature value; Once thermal steady state is confirmed, the central control system triggers the infrared thermal imager to acquire and store an infrared image. ; Records when the k-th critical component is heated to At that time, the temperature distribution of the entire chip module.

4. The method for automatically controlling and monitoring chip functions according to claim 3, characterized in that, The inner loop also includes a cooling and restart process: completing an infrared image. After data acquisition, the central control system shuts down the current local heating unit. Once the entire chip module cools to the operating ambient temperature, the inner layer circulation continues, targeting the same critical component location. Apply the next target temperature rise value Repeat the heating, steady-state judgment, infrared image acquisition and storage until all M target heating temperature values ​​are completed; The outer loop also includes: when the critical component location... After all M target heating temperature values' infrared images have been acquired, the outer loop enters the next iteration, selecting new key component locations. And repeat all the operations of the inner loop.

5. The method for automatically controlling and monitoring chip functions according to claim 4, characterized in that: The detection process begins, and the raw infrared time-series image sequence of the chip module is obtained, including: The global temperature control system of the incubator adjusts the overall ambient temperature inside the incubator to the operating ambient temperature. Power is applied to the chip module via the testing system, and a predetermined workload is executed, triggering the infrared thermal imager to continuously capture infrared thermal images of the chip module's surface. The infrared thermal imager generates raw infrared image frames at a set frame rate. , where i is the image frame number from the start time.

6. The method for automatically controlling and monitoring chip functions according to claim 5, characterized in that: A triple loop structure is used to obtain the self-heating infrared time-series image sequence of the chip module. The first loop iteration includes extracting any one frame of the original infrared time-series image sequence. Generate a corresponding corrected image for one frame. .

7. The method for automatically controlling and monitoring chip functions according to claim 6, characterized in that: For each raw infrared image frame The correction process sets up a second loop structure, which iterates through all critical components to be corrected; for the currently selected critical component, it obtains the information about that critical component in... Corresponding position Average observed temperature For each critical component to be corrected The correction process establishes a third loop structure, which iterates through all other critical components. , , .

8. The method for automatically controlling and monitoring chip functions according to claim 7, characterized in that: Each key component to be corrected is modified based on the third loop structure. include: For other key components Comp j Obtain the other key component Comp j in I raw i The corresponding position pos j Average observed temperature T obs i,j Access temperature-location influences infrared image concentration corresponding to this other key component Comp. j M infrared images were used to locate the other critical component Comp. j The corresponding position pos in M ​​infrared images j The average temperature of the other key components in the current frame is similar to that of the Comp component. j Average observed temperature T obs i,j The closest T set_t The calculation yielded the other key component Comp. j At temperature T obs i,j For key components Comp k The resulting temperature rise is: ΔT j→k (i)=T target (I j t ,pos k )-T baseline ; Where, ΔT j→k (i) indicates other key components Comp j For key components Comp k The effect of temperature rise in the i-th frame; T target (I j t ,pos k ) indicates other key components Comp j Target heating temperature T set_t The acquired temperature-position influences the key component Comp in infrared images k Average temperature at location; T baseline Operating ambient temperature; Other key components Comp obtained from the calculation j For key components Comp k The effect of temperature rise on ΔT j→k (i) Accumulated to the total external heat effect ΔT k (i): ΔT k (i)=ΔT k (i)+ΔT j→k (i); where ΔT k (i) is the key element Comp in the i-th frame. k Total external heat impact; The third cycle continues, focusing on the same critical component, Comp. k Apply the next other key element, until all other key elements have been traversed, to obtain all other key elements in relation to key element Comp. k Total external heat influence ΔT k (i); Utilizing key component Comp k position pos k Average observed temperature T obs i,k The corrected temperature at this location is obtained: T self (pos k ,i)=T obs i,k -ΔT k (i); where T self (pos k ,i) represents the position pos in the i-th frame. k The corrected temperature; Set the original infrared image frame I raw i Each critical element in the process is a critical element to be corrected, and a third loop structure is executed until the original infrared image frame I is reached. raw i Each critical component in the process completed the thermal impact elimination of other critical components, resulting in the corresponding corrected image I. self i ; Repeated execution for a single frame image I raw i The correction process continues until every frame in the original infrared time-series image sequence has been processed and its corresponding corrected image has been generated; all the corrected image frames are combined in chronological order to obtain the self-heating infrared time-series image sequence of the chip module.

9. A system for automatically controlling and monitoring chip functions, used to perform the method for automatically controlling and monitoring chip functions as described in claim 1, characterized in that, The system includes: Temperature chamber: Place the chip module in the temperature chamber and connect the test circuitry; Local heating control module: Based on the location information of each key component in the chip module, the temperature chamber is guided to perform local heating, and an infrared image set of the temperature-position influence of all key components is obtained; Infrared temporal image sequence acquisition module: Start detection and use an infrared thermal imager to acquire the original infrared temporal image sequence of the chip module; Image correction module: Based on the temperature-position influence infrared atlas of each key component location and the ambient temperature in the original infrared time-series image sequence, the original infrared time-series image sequence is corrected to obtain the self-heating infrared time-series image sequence of the chip module; Supervision module: Based on the time series model, it performs heat generation supervision on key components in the infrared time series image sequence of its own heat generation.

Images