Chip temperature and humidity aging test system

Abstract

The invention discloses a chip temperature and humidity aging test system, which relates to the field of chip temperature and humidity aging test and comprises a core control and digital twinborn fusion module, a three-dimensional temperature and humidity field active construction module, a multi-physical field coupling stress loading module, an energy self-adaptive supply module and a full life cycle data traceability and intelligent evaluation module. According to the chip temperature and humidity aging test system, through the core control and digital twinning fusion module, chip cross-scale digital twinning modeling containing process deviation and deep closed loop of digital rehearsal, physical test and feedback regulation can be realized, aging key nodes can be pre-judged in advance, the physical test can be accurately guided, and the chip temperature and humidity aging test efficiency can be improved. The perspectiveness and the accuracy of the test are greatly improved; the three-dimensional temperature and humidity field active building module can build a three-dimensional non-uniform temperature and humidity field which is precisely regulated and controlled in a layered mode and driven by a dynamic flow field, the three-dimensional non-uniform temperature and humidity field is closer to a heat and humidity distribution and airflow disturbance scene in actual application of a chip, and temperature and humidity stress simulation is more real.

Description

Technical Field

[0001] This invention relates to the field of temperature and humidity aging testing of chips, and particularly to a chip temperature and humidity aging testing system. Background Technology

[0002] In the field of chip reliability testing, traditional temperature and humidity aging testing techniques have many limitations. On the one hand, the test environment is often a uniform and static temperature and humidity field, which makes it difficult to reproduce the complex scenarios of dynamic temperature and humidity changes, mechanical stress, electromagnetic interference, and other multi-physical field couplings in actual chip applications. This results in a large deviation between the test results and the actual aging condition of the chip. On the other hand, digital twin technology mostly remains at the level of simple model simulation and has not formed a deep closed loop of digital pre-simulation, physical testing, and data feedback, making it impossible to predict the chip aging trend in advance. In addition, the energy supply mostly adopts a constant voltage or current mode, which does not match the dynamic fluctuation of power consumption, energy recovery, and limited conditions during actual chip operation, and lacks full life cycle data traceability capabilities.

[0003] Therefore, it is necessary to propose a chip temperature and humidity aging test system to solve the above problems. Summary of the Invention

[0004] The main objective of this invention is to provide a chip temperature and humidity aging test system, which can effectively solve the problems in the background art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The chip temperature and humidity aging test system includes a core control and digital twin fusion module, a three-dimensional temperature and humidity field active construction module, a multi-physics field coupled stress loading module, an energy self-adaptive supply module, and a full life cycle data traceability and intelligent evaluation module. The core control and digital twin fusion module is used to construct a cross-scale digital twin model and map process deviations. Through dynamic scene generation and data and material feedback, it realizes closed-loop collaborative testing. The three-dimensional temperature and humidity field active construction module is used to divide three layers of space for precise temperature and humidity control. Combined with vector airflow injection and flow field simulation, it constructs a three-dimensional non-uniform temperature and humidity field. The multi-physics field coupled stress loading module is used to synchronously load micromechanical stress and broadband EMI stress, so as to realize the phase synchronous coupling of mechanical, electromagnetic and temperature and humidity stress. The energy self-adaptive supply module generates a dynamic power supply waveform based on power consumption prediction, recovers test energy, and adaptively adjusts the power supply to match the chip power consumption and environmental stress changes. The full lifecycle data traceability and intelligent assessment module is used to collect and integrate data from the entire chain, use intelligent algorithms to assess aging, predict lifespan, and provide feedback for optimization testing.

[0006] Preferably, the core control and digital twin fusion module includes: The multi-dimensional digital twin modeling submodule includes cross-scale model construction: based on chip material composition and structural design, a micro-macro cross-scale digital twin model is constructed. At the micro level, the temperature and humidity response at the transistor level inside the chip is accurately simulated; at the macro level, the thermal diffusion and moisture penetration characteristics of the entire package are reproduced to capture the multi-scale aging coupling effect from the transistor to the package. Process Deviation Mapping: Integrates process deviation-aging correlation function to transform process data in chip manufacturing into parameter variables of digital twin model; The real-time closed-loop control submodule includes a dynamic scene generation engine: based on the target application scenario of the chip, it generates sudden change, gradient, and cyclic composite temperature and humidity curves. Digital pre-simulation - physical test feedback: When the digital twin model simulates stress concentration inside the chip and material moisture saturation, it automatically adjusts the temperature and humidity parameters of the physical test to ensure that the physical test accurately captures key aging nodes.

[0007] Preferably, the three-dimensional temperature and humidity field active construction module specifically includes: The layered temperature and humidity control submodule includes dividing the test chamber into a 0-5mm near-field layer, a 5-20mm middle layer, and a more than 20mm far-field layer. The near-field layer uses micro-heat flow channels and nanofiber humidity conduction membranes to simulate the near-field high temperature and humidity caused by the chip's own heating. The middle layer uses a gradient heating wire array and a porous ceramic humidity buffer layer to construct a linear temperature and humidity gradient to simulate the transition of the chip with the external environment. The far-field layer uses a high-flow-rate heat and humidity exchanger to maintain the overall energy and humidity balance of the chamber and to simulate the temperature and humidity background field of the external environment. Distributed flexible sensing feedback: Graphene-based flexible sensing nodes are deployed in each layer to collect temperature and humidity data in real time and feed it back to the core control and digital twin fusion module; The dynamic flow field driving submodule includes a vector airflow injection system: through a miniature adjustable nozzle with a diameter of 0.5mm, the airflow temperature, humidity, and injection angle of each nozzle can be independently controlled to simulate complex airflow disturbances; Flow field simulation and optimization unit: Based on computational fluid dynamics algorithms, it simulates the airflow field, temperature field, and humidity field inside the cavity in real time. If the deviation between the flow field and the digital twin pre-simulation exceeds the threshold, it automatically adjusts the nozzle parameters to ensure that the physical flow field is consistent with the actual application scenario.

[0008] Preferably, the multiphysics coupled stress loading module specifically includes: The micromechanical stress loading submodule includes a shape memory alloy microarray: composed of hundreds of independent and controllable SMA units, each unit generating 0-50MPa pressure, and the stress direction and loading frequency are adjustable; Stress feedback and calibration: Mechanical stress is collected in real time by a surface micro piezoresistive sensor that is packaged synchronously with the chip. After feedback, the SMA microarray parameters are adjusted to ensure the coupling accuracy between mechanical stress and temperature and humidity stress. The electromagnetic interference stress loading submodule includes wide-spectrum EMI field construction: generating 10kHz-10GHz wide-band electromagnetic interference with adjustable field strength in the range of 1V / m-100V / m, used to simulate complex electromagnetic environments; Multi-stress phase synchronization: Through the coordination of the core control and digital twin fusion module, phase synchronization coupling of EMI with temperature, humidity and mechanical stress is achieved.

[0009] Preferably, the energy self-adaptive supply module specifically includes: The dynamic power consumption simulation submodule includes real-time power consumption prediction: based on a cross-scale digital twin model, it predicts the power consumption changes of the chip under temperature and humidity, mechanical stress, and EMI coupling in real time. Dynamic power supply waveform generation: Based on the predicted power consumption, a dynamic power supply waveform consistent with the actual power consumption is generated, including voltage fluctuations and current spikes, to simulate the dynamic fluctuations of the power supply in actual applications. The energy recovery and adaptive regulation submodule includes micro energy recovery: through thermoelectric conversion and electromagnetic radiation recovery, the energy generated during chip aging tests is converted into electrical energy and stored in a supercapacitor; Energy self-feedback regulation: Automatically adjusts the voltage, current, and waveform of the external power supply based on the amount of energy recovered and the dynamic power consumption requirements of the chip.

[0010] Preferably, the full lifecycle data traceability and intelligent evaluation module specifically includes: The multi-source data fusion submodule includes full-chain data acquisition: acquiring chip manufacturing process data, aging test multi-physical field data, electrical performance data, and actual application field data; Data association graph construction: Using graph database technology, an association graph of process parameters, stress response, and performance degradation is established; The intelligent aging assessment and prediction submodule includes causal and transfer learning algorithms: first, causal reasoning is used to determine the causal chain of process parameters, stress response, and performance degradation, and then transfer learning is used to transfer laboratory aging data to actual application scenarios to improve prediction accuracy. Aging health report generation: This report outputs the chip's current aging level, remaining lifespan prediction, and key failure risk points. Simultaneously, the report is fed back to the core control and digital twin fusion module, suggesting adjustments to subsequent test parameters.

[0011] Compared with the prior art, the present invention provides a chip temperature and humidity aging test system, which has the following beneficial effects: This chip temperature and humidity aging test system, through the core control and digital twin fusion module, can realize cross-scale digital twin modeling of chips with process deviations, as well as a deep closed loop of digital pre-simulation, physical testing, and feedback control. It can predict key aging nodes in advance, accurately guide physical testing, and significantly improve the foresight and accuracy of testing. The three-dimensional temperature and humidity field active construction module can create a three-dimensional non-uniform temperature and humidity field with layered precise control and dynamic flow field drive, which is closer to the heat and humidity distribution and airflow disturbance scenarios in actual chip applications, making the temperature and humidity stress simulation more realistic.

[0012] This chip temperature and humidity aging test system features a multi-physics field coupled stress loading module that, for the first time, synchronously and accurately couples mechanical stress, electromagnetic interference stress, and temperature and humidity stress, fully replicating the multi-stress superposition environment of actual chip operation and overcoming the limitations of single-stress testing in existing technologies. The energy self-adaptive supply module achieves dynamic matching between energy supply and chip power consumption and environmental stress, while combining energy recovery and adaptive adjustment to accurately simulate the aging characteristics of the chip under actual energy fluctuations, providing strong support for testing low-power, energy-recovery chips. The full lifecycle data traceability and intelligent evaluation module connects the data chain of production, aging testing, and actual application, using multi-source data fusion and intelligent algorithms to achieve accurate assessment of aging health and prediction of remaining lifespan, providing a complete chain of evidence for improving chip reliability. Attached Figure Description

[0013] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0014] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0015] like Figure 1 As shown, the chip temperature and humidity aging test system includes a core control and digital twin fusion module, a three-dimensional temperature and humidity field active construction module, a multi-physics field coupled stress loading module, an energy self-adaptive supply module, and a full life cycle data traceability and intelligent evaluation module. The core control and digital twin fusion module is used to construct a cross-scale digital twin model and map process deviations. Through dynamic scene generation and data and physical feedback, it achieves closed-loop collaborative testing. The core control and digital twin fusion module includes: The multi-dimensional digital twin modeling submodule includes cross-scale model construction: based on chip material composition and structural design, a micro-macro cross-scale digital twin model is constructed. At the micro level, the temperature and humidity response at the transistor level inside the chip is accurately simulated; at the macro level, the thermal diffusion and moisture penetration characteristics of the entire package are reproduced to capture the multi-scale aging coupling effect from the transistor to the package. Process Deviation Mapping: Integrating process deviation-aging correlation function, it transforms process data in chip manufacturing into parameter variables of digital twin model. For example, if the lithography linewidth deviation of a certain batch of chips is ±10% of the design value, the model will assign correlation parameters of linewidth fluctuation, current density change and temperature rise in the corresponding area, making digital simulation closer to the individual aging differences of actual chips, and solving the problem of existing technology ignoring the impact of process deviation on aging. The real-time closed-loop control submodule includes a dynamic scene generation engine: based on the target application scenario of the chip, it generates sudden change, gradient, and cyclic composite temperature and humidity curves. For example, it can simulate a car accelerating rapidly, where the temperature rises from 25°C to 120°C in 10 seconds and the humidity drops from 60% to 10%; or simulate a gradual change scenario in an industrial production line where the temperature rises from 80°C to 150°C at a rate of 0.5°C / min, while the humidity slowly decreases from 95%RH to 60%RH at a rate of 2%RH / min. Unlike existing fixed temperature and humidity spectra, this engine can reproduce the dynamic stress characteristics of real-world scenarios. Digital pre-simulation - physical test feedback: When the digital twin model predicts stress concentration inside the chip and material moisture saturation, the temperature and humidity parameters of the physical test are automatically adjusted to ensure that the physical test accurately captures the key aging nodes, realizing a closed loop where digital pre-simulation guides physical testing and physical data feeds back into the digital model.

[0016] The three-dimensional temperature and humidity field active construction module is used to divide a three-layer space for precise temperature and humidity control. Combined with vector airflow injection and flow field simulation, it constructs a three-dimensional non-uniform temperature and humidity field. The three-dimensional temperature and humidity field active construction module specifically includes: The layered temperature and humidity control submodule includes dividing the test chamber into a 0-5mm near-field layer, a 5-20mm middle layer, and a far-field layer larger than 20mm. The near-field layer utilizes micro-heat flow channels and nanofiber humidity-conducting membranes to achieve temperature accuracy of ±0.1℃ and humidity accuracy of ±1%RH, simulating the near-field high temperature and humidity caused by the chip's own heating. The middle layer constructs a linear temperature and humidity gradient through a gradient heating wire array and a porous ceramic humidity buffer layer, simulating the thermal and humidity gradient transition between the chip and the external environment. The far-field layer uses a high-flow-rate heat and humidity exchanger to maintain the overall energy and humidity balance of the chamber, simulating the temperature and humidity background field of the external environment. Distributed flexible sensing feedback: Graphene-based flexible sensing nodes are deployed in each layer to collect temperature and humidity data in real time and feed it back to the core control and digital twin fusion module; The dynamic flow field driving submodule includes a vector airflow injection system: through a miniature adjustable nozzle with a diameter of 0.5mm, the airflow temperature, humidity and injection angle of each nozzle can be independently controlled to simulate complex airflow disturbances, such as crosswinds and eddies when consumer electronics are used outdoors, or the impact of directional airflow near industrial fans on the temperature and humidity environment of the chip, which is different from the simple simulation of existing unidirectional uniform airflow. Flow field simulation and optimization unit: Based on computational fluid dynamics algorithms, it simulates the airflow field, temperature field, and humidity field inside the cavity in real time. If the deviation between the flow field and the digital twin pre-simulation exceeds the threshold, it automatically adjusts the nozzle parameters to ensure that the physical flow field is consistent with the actual application scenario.

[0017] The multi-physics field coupled stress loading module is used to synchronously load micromechanical stress and broadband EMI stress, achieving phase synchronization coupling of mechanical, electromagnetic, and temperature and humidity stresses. Specifically, the multi-physics field coupled stress loading module includes: The micromechanical stress loading submodule includes a shape memory alloy microarray: composed of hundreds of independent and controllable SMA units, each unit generates 0-50MPa pressure, and the stress direction and loading frequency are adjustable. For example, it can simulate the periodic extrusion stress of the chip in the package when the car is bumping, or the high-frequency shear stress in the industrial vibration environment. Stress feedback and calibration: Mechanical stress is collected in real time by a surface micro piezoresistive sensor that is packaged synchronously with the chip. After feedback, the SMA microarray parameters are adjusted to ensure the coupling accuracy between mechanical stress and temperature and humidity stress. The electromagnetic interference stress loading submodule includes wide-spectrum EMI field construction: generating 10kHz-10GHz wide-band electromagnetic interference with adjustable field strength in the range of 1V / m-100V / m, used to simulate complex electromagnetic environments; Multi-stress phase synchronization: Through the coordination of the core control and digital twin fusion module, phase synchronization coupling of EMI with temperature, humidity and mechanical stress is achieved.

[0018] The energy self-adaptive supply module generates a dynamic power supply waveform based on power consumption prediction, recovers test energy, and adaptively adjusts the power supply to match changes in chip power consumption and environmental stress. Specifically, the energy self-adaptive supply module includes: The dynamic power consumption simulation submodule includes real-time power consumption prediction: based on a cross-scale digital twin model, it predicts the power consumption changes of the chip under temperature and humidity, mechanical stress, and EMI coupling in real time. For example, when the chip's leakage current increases due to high temperature and its parasitic capacitance changes due to mechanical stress, the submodule calculates the chip's actual power consumption curve at this time. Dynamic power supply waveform generation: Based on the predicted power consumption, a dynamic power supply waveform consistent with the actual power consumption is generated, including voltage fluctuations and current spikes, to simulate the dynamic fluctuations of the power supply in actual applications. The energy recovery and adaptive regulation submodule includes micro energy recovery: through thermoelectric conversion and electromagnetic radiation recovery, the energy generated during chip aging tests is converted into electrical energy and stored in a supercapacitor; Energy self-feedback regulation: Based on the amount of energy recovered and the dynamic power consumption requirements of the chip, the voltage, current and waveform of the external power supply are automatically adjusted to realize the dynamic cycle of sufficient energy, limited energy and energy recovery and reuse, and to reproduce chip aging under actual power supply instability or energy recovery scenarios.

[0019] The full lifecycle data traceability and intelligent assessment module is used to collect and integrate data from the entire chain, use intelligent algorithms to assess aging, predict lifespan, and provide feedback for optimization testing. Specifically, the full lifecycle data traceability and intelligent assessment module includes: The multi-source data fusion submodule includes full-chain data acquisition: acquiring chip manufacturing process data, aging test multi-physical field data, electrical performance data, and actual application field data; Data association graph construction: Using graph database technology, an association graph of process parameters, stress response, and performance degradation is established. For example, it is found that due to the deviation of the photolithography linewidth, the leakage current of a certain batch of chips rises 30% faster than that of other batches under a specific temperature and humidity cycle, thus clarifying the influence of the process on aging. The intelligent aging assessment and prediction submodule includes causal and transfer learning algorithms: first, causal reasoning is used to determine the causal chain of process parameters, stress response, and performance degradation, and then transfer learning is used to transfer laboratory aging data to actual application scenarios to improve prediction accuracy. Aging health report generation: This report outputs the chip's current aging level, remaining lifespan prediction, and key failure risk points. Simultaneously, the report is fed back to the core control and digital twin fusion module, suggesting adjustments to subsequent test parameters.

[0020] It should be noted that this invention is a chip temperature and humidity aging test system. When in use, the digital pre-simulation is started: the core control and digital twin fusion module imports chip design drawings and process data, constructs a cross-scale digital twin model, and generates dynamic test scenarios such as sudden changes in engine compartment temperature and humidity, turbulence mechanical stress, and EMI interference.

[0021] Temperature and humidity field and flow field construction: The three-dimensional temperature and humidity field active construction module regulates the temperature and humidity of the cavity in layers, and the dynamic flow field module simulates the airflow disturbance in the engine compartment to form a three-dimensional non-uniform environment consistent with reality.

[0022] Multi-stress coupling loading: The multi-physics field coupled stress loading module synchronously loads mechanical stress and EMI, and precisely couples it with temperature and humidity stress.

[0023] Dynamic energy supply: The self-adaptive energy supply module generates a dynamic power supply waveform based on the chip power consumption predicted by the digital twin, and at the same time recovers the chip's heat energy for adaptive adjustment.

[0024] Data traceability and assessment: The full life cycle data traceability and intelligent assessment module collects multi-source data, performs correlation analysis to generate an aging health report, and feeds it back to the core control and digital twin fusion module to optimize test parameters.

[0025] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A chip temperature and humidity aging test system, comprising a core control and digital twin fusion module, a three-dimensional temperature and humidity field active construction module, a multi-physical field coupling stress loading module, an energy self-adaptive supply module, a full life cycle data traceability and intelligent evaluation module, characterized in that: The core control and digital twin fusion module is used for constructing a cross-scale digital twin model and mapping process deviation, generating a dynamic scene, and feeding back numbers and objects to realize test closed-loop cooperation. The active three-dimensional temperature and humidity field construction module is used for dividing three layers of space to accurately control temperature and humidity, combining vector air jet and flow field simulation to construct a three-dimensional non-uniform temperature and humidity field. The multi-physical field coupling stress loading module is used for synchronously loading micro-mechanical stress and wide-band EMI stress to realize mechanical, electromagnetic and temperature and humidity stress phase synchronous coupling. The energy self-adaptive supply module generates a dynamic power supply waveform based on power consumption prediction, recovers test energy, and self-adapts the power supply to match the chip power consumption and environmental stress changes. The full life cycle data traceability and intelligent evaluation module is used for collecting full-chain data and fusion, evaluating aging and predicting life span by using intelligent algorithms, and feeding back optimization test.

2. The system of claim 1, wherein: The core control and digital twin fusion module comprises: A multi-dimensional digital twin modeling submodule comprises cross-scale model construction: based on chip material components and structure design, a micro-macro cross-scale digital twin model is constructed, on a micro level, a temperature and humidity response of a transistor level inside a chip is accurately simulated, and on a macro level, thermal diffusion and wet permeation characteristics of an entire package are reproduced to capture multi-scale aging coupling effects from a transistor to a package; Process deviation mapping: integrating process deviation-aging correlation functions, process data of a chip production link is converted into parameter variables of a digital twin model; A real-time closed-loop control submodule comprises a dynamic scene generation engine: according to a chip target application scene, a sudden change, a gradient, and a cycle compound temperature and humidity curve are generated; Digital preview-physical test feedback: when a digital twin model is previewed to stress concentration inside a chip and material hygroscopic saturation, temperature and humidity parameters of a physical test are automatically adjusted to ensure that a physical test accurately captures aging key nodes.

3. The system of claim 1, wherein: The active three-dimensional temperature and humidity field construction module specifically comprises: A hierarchical temperature and humidity regulation submodule comprises dividing a test cavity into a chip near-field layer of 0-5 mm, a middle layer of 5-20 mm, and a far-field layer of more than 20 mm, wherein the near-field layer: a micro hot runner and a nanofiber humidity conduction film are used to simulate near-field high temperature and high humidity caused by chip self-heating; the middle layer: a gradient heating wire array and a porous ceramic humidity buffer layer are used to construct a linear temperature and humidity gradient to simulate a thermal and humidity gradient transition between a chip and an external environment; the far-field layer: a large-flow heat and humidity exchanger is used to maintain overall energy and humidity balance of the cavity to simulate a temperature and humidity background field of an external environment; Distributed flexible sensing feedback: graphene-based flexible sensing nodes are deployed at each layer to collect temperature and humidity data in real time and feed back to the core control and digital twin fusion module; A dynamic flow field driving submodule comprises a vector air jet system: through a micro adjustable nozzle with a diameter of 0.5 mm, air flow temperature, humidity and jet angle of each nozzle are independently controlled to simulate complex air flow disturbance. Flow field simulation and optimization unit: based on computational fluid dynamics algorithm, real-time simulation of the cavity airflow field, temperature field, humidity field, if the flow field and digital twin deviation exceeds the threshold, automatically adjust the nozzle parameters, to ensure that the physical flow field and the actual application scene consistent.

4. The system of claim 1, wherein: The multi-physical field coupling stress loading module specifically comprises: Micro-mechanical stress loading sub-module, including shape memory alloy microarray: composed of hundreds of independently controllable SMA units, each unit generates 0-50MPa pressure, and the stress direction, loading frequency are adjustable; Stress feedback and calibration: real-time collection of mechanical stress through the surface micro piezoresistive sensor packaged synchronously with the chip, adjustment of SMA microarray parameters after feedback, to ensure the coupling accuracy of mechanical stress and temperature and humidity stress; EMI stress loading sub-module, including wide spectrum EMI field construction: generating 10kHz-10GHz wide band electromagnetic interference, field strength adjustable in 1V / m-100V / m range, used to simulate complex electromagnetic environment; Multi-stress phase synchronization: through the coordination of the core control and digital twin fusion module, realize the phase synchronization coupling of EMI, temperature and humidity, mechanical stress.

5. The system of claim 1, wherein: The energy self-adaptive supply module specifically comprises: Dynamic power consumption simulation sub-module, including real-time power consumption prediction: based on cross-scale digital twin model, real-time prediction of chip power consumption change under temperature and humidity, mechanical stress, EMI coupling; Dynamic power supply waveform generation: according to the predicted power consumption, generate dynamic power supply waveform consistent with the actual power consumption, including voltage fluctuation, current peak, used to simulate the dynamic fluctuation of power supply in actual application; Energy recovery and adaptive adjustment sub-module, including micro energy recovery: through thermoelectric conversion, electromagnetic radiation recovery, the energy generated in chip burn-in test is converted into electrical energy and stored in super capacitor; Energy self-feedback adjustment: according to the energy recovery amount and chip dynamic power consumption demand, automatically adjust the voltage, current, waveform of external power supply.

6. The system of claim 1, wherein: The full life cycle data traceability and intelligent evaluation module specifically comprises: Multi-source data fusion sub-module, including full-chain data collection: collecting chip production process data, aging test multi-physical field data, electrical performance data, and actual application site data; Data correlation graph construction: using graph database technology, establish the correlation graph of process parameters, stress response, performance degradation; Intelligent aging evaluation and prediction sub-module, including causal, transfer learning algorithm: first determine the causal chain of process parameters, stress response, performance degradation through causal reasoning, then use transfer learning to migrate laboratory aging data to actual application scenario, improve the prediction accuracy; Aging health report generation: used to output the current aging degree of chip, remaining service life prediction, key failure risk point, at the same time, the report is fed back to the core control and digital twin fusion module, to suggest adjusting the subsequent test parameters.

Images