Intelligent IV testing and grading system and method for perovskite solar cell

Abstract

The invention discloses an intelligent IV testing and grading system and method for a perovskite solar cell, and belongs to the field of cell testing, and the system comprises a simulation illumination module, an IV testing module, a visual recognition module, an AI analysis module and a grading execution module. The method mainly comprises the steps that millisecond-level pulsed light is emitted through the simulation illumination module, and the IV test module synchronously scans and collects IV data; the AI analysis module analyzes data in real time based on a mixed model of LSTM and XGBoost, automatically judges test convergence according to a peak power change rate, and adaptively selects a stepped or linear scanning mode; after the test is completed, the IV test module is controlled to excite the assembly to generate electroluminescence, the visual identification module collects EL images, and defects are identified by an image model based on YOLOv8; and finally, performing intelligent grading by integrating the electrical performance and the defect result. According to the invention, nondestructive, rapid and adaptive perovskite cell testing and grading are realized, and the problems of thermal damage, ferroelectric hysteresis effect interference, low testing efficiency and the like are effectively solved.

Description

A smart IV testing and grading system and method for perovskite solar cells. Technical Field

[0001] This invention relates to the field of battery testing, and in particular to an intelligent IV testing and grading system and method for perovskite solar cells. Background Technology

[0002] Perovskite solar cells, as a new generation of photovoltaic technology, have attracted widespread attention due to their high light conversion efficiency, low manufacturing cost, and excellent material tunability. In the research and development and production process, accurately and efficiently measuring their current-voltage characteristic curves (IV curves) and calculating core performance parameters such as open-circuit voltage (Voc), short-circuit current (Isc), fill factor (FF), peak power (Pmax), and photoelectric conversion efficiency (PCE) are essential steps for performance evaluation, process optimization, and product classification.

[0003] Currently, the industry's IV testing of perovskite solar cells mainly borrows from the testing methods of traditional crystalline silicon cells. This typically involves continuous steady-state light source illumination (such as a tritium lamp) and voltage scanning using a source measurement unit to obtain data. However, the unique physicochemical properties of perovskite materials present significant challenges to these traditional testing methods in practical applications, primarily due to the following problems:

[0004] First, perovskite materials are highly sensitive to light and heat. Under continuous steady-state illumination, the temperature of the battery or module will continue to rise, generating a significant thermal effect. This temperature rise not only causes the measured parameters such as Pmax to be lower than the material's true "cold-state" performance, resulting in measurement bias, but more seriously, it may induce irreversible photodegradation or ion migration in the material, causing substantial damage to the tested sample and affecting its subsequent performance and reliability.

[0005] Secondly, perovskite solar cells generally exhibit a significant current-voltage hysteresis effect. This means that the IV curve obtained from a single scan is heavily dependent on the scan direction, rate, and voltage bias history, and cannot stably and accurately reflect the steady-state output performance of the device. To address this issue, existing technologies often rely on operators manually performing multiple repeated measurements and subjectively judging data stability by visually observing the curve shape. This method is inefficient, has poor repeatability, and lacks objective and unified quantitative judgment standards, severely restricting the level of automation and standardization in testing.

[0006] Furthermore, existing testing systems lack sufficient intelligence and adaptability. The research and production of perovskite solar cells involve samples of varying specifications, ranging from tiny solar cells to large-area modules, with significant differences in their electrical characteristics (such as Voc and Isc). Traditional testing equipment typically employs fixed scan steps, rates, and illumination strategies, making it difficult to meet the testing needs of different samples: for high-power modules, fixed parameters may lead to insufficient testing accuracy; for low-power samples, it results in unnecessary extensions in testing time, making it impossible to achieve a balance between high throughput and high accuracy on the production line.

[0007] Furthermore, in industrial production lines, performance testing and defect detection are usually separate processes, requiring the use of different equipment for electrical performance (IV) testing and electroluminescence (EL) imaging detection, respectively. This "secondary wiring, step-by-step operation" model not only reduces overall production efficiency but also increases operational complexity. Moreover, the lack of automatic and reliable correlation between electrical performance data and defect image data hinders subsequent data traceability and root cause analysis of processes.

[0008] Therefore, developing an intelligent IV testing and grading system and method that can overcome the limitations of perovskite material properties, achieve non-destructive, rapid, adaptive testing, and integrate performance testing and defect detection is of great significance for promoting the technological advancement and industrialization of perovskite solar cells.

[0009] The information disclosed in this background section is intended only to enhance the understanding of the overall background of the invention and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0010] The purpose of this invention is to provide an intelligent IV testing and grading system and method that overcomes the limitations of perovskite material properties, achieves non-destructive, rapid, adaptive testing, and integrates performance testing and defect detection. This invention aims to overcome the shortcomings of existing perovskite solar cell IV testing technologies, solve the problems of thermal damage and performance testing distortion caused by continuous illumination, eliminate measurement instability caused by hysteresis, improve the adaptability of the testing system to balance the testing accuracy and efficiency of samples of different specifications, and achieve integrated and intelligent electrical performance testing and defect detection.

[0011] To achieve the above objectives, the present invention provides an intelligent IV testing and grading system and method for perovskite solar cells.

[0012] The system includes a simulated illumination module for providing controllable simulated illumination to the perovskite solar cell or module under test. This module includes an adjustable spectral LED array and a closed-loop feedback control unit, capable of outputting millisecond-level pulsed light with an adjustable pulse width in the range of 10ms to 100ms, and ensuring that the light intensity non-uniformity is less than 1% through feedback control.

[0013] The IV test module, electrically connected to the simulated illumination module and the battery or component under test, is used to perform current-voltage (IV) scanning and acquire IV data under illumination conditions. This module includes a high-speed source measurement unit (SMU) based on an FPGA architecture, with a minimum sampling interval of 10 microseconds and a maximum scanning speed of 2000 points / second. It is configured to synchronously trigger IV scanning sampling when the light pulse reaches the stable region.

[0014] A vision recognition module is used to acquire images of the battery or component under test under specific stimuli. This module includes an industrial camera. The system is configured to, after the IV electrical performance test is completed, control the IV test module to switch to injecting a forward bias current into the device under test to excite it to generate electroluminescence (EL), and simultaneously trigger the camera to capture the EL image.

[0015] The AI ​​analysis module is communicatively connected to both the IV testing module and the visual recognition module, and is used to receive and analyze the IV data and image data. This module includes a locally deployed industrial control computer and a hybrid AI model running on it. The hybrid AI model includes a numerical analysis model (built on LSTM and XGBoost) for analyzing IV data time series to predict steady-state performance parameters, and an image analysis model (built on the YOLOv8 architecture) for analyzing EL images to identify defects. The AI ​​analysis module is configured to calculate the peak power (Pmax) change rate (Δ) based on real-time IV data and automatically decide whether to continue testing, stop testing, or handle anomalies based on the Δ value.

[0016] The hierarchical execution module is communicatively connected to the AI ​​analysis module and is used to execute corresponding hierarchical operations according to the hierarchical instructions output by the AI ​​analysis module.

[0017] The triggering of the IV test module, the simulated lighting module, and the visual recognition module is controlled by the AI ​​analysis module to collaboratively complete the test.

[0018] The intelligent IV testing and grading method applied to the above system includes the following steps: S1: System initialization, placing the perovskite solar cell or module to be tested in the test position.

[0019] S2: Control the simulated illumination module to apply millisecond-width pulsed light to the battery or component under test, and synchronously control the IV test module to perform an IV scan and acquire IV data during the stable period of the light pulse. Preferably, the IV test module triggers the scan 5ms after the light pulse is turned on, and immediately turns off the light source after the scan is completed.

[0020] S3: The collected IV data is input into the AI ​​analysis module for analysis. This module determines whether the test has converged based on preset decision rules. The decision rules include: calculating the rate of change Δ between the current test peak power Pmax(N) and the previous test peak power Pmax(N-1); if Δ > 0.1%, it is determined that the test has not converged and the test continues; if |Δ| < 0.1% occurs twice consecutively, it is determined that the test has converged and the test stops; if Δ < -0.2%, it is determined to be abnormal and an emergency stop is triggered.

[0021] S4: If it is determined that the test has not converged, repeat steps S2-S3 for the next test until the AI ​​analysis module determines that the test has converged or an abnormality is triggered and the test stops.

[0022] S5: After the test converges, control the IV test module to inject current into the battery or component under test to excite electroluminescence (EL), and control the vision recognition module to acquire EL images.

[0023] S6: The AI ​​analysis module comprehensively analyzes the IV data results and the defect recognition results of the EL image, and outputs a grading instruction.

[0024] S7: The hierarchical execution module executes the corresponding hierarchical operation according to the hierarchical instruction.

[0025] Preferably, before step S2, an adaptive scanning mode selection step is also included: performing a pre-scan to obtain the open-circuit voltage (Voc) and short-circuit current (Isc) of the battery or component under test; if Voc ≥ 1.2V and Isc ≥ 2A, then a stepped scanning mode (e.g., step size 0.01V) is selected for subsequent IV scans; if Voc < 0.8V or Isc < 0.5A, then a linear scanning mode (e.g., scanning speed 5V / s) is selected for subsequent IV scans.

[0026] Compared with the prior art, the present invention has the following beneficial effects: non-destructive and accurate measurement: using millisecond-level pulsed light irradiation technology, the test is completed before the battery generates a significant thermal effect, and the temperature rise during the test process is controlled within 0.1℃, which effectively avoids the performance test distortion and irreversible photothermal damage caused by the temperature rise of perovskite materials, and truly restores the "cold state" electrical performance of the material.

[0027] Intelligent and efficient decision-making: Edge computing through locally deployed hybrid AI models enables "testing and computing on the go." By utilizing adaptive convergence rules based on peak power change rate, it replaces traditional fixed-number testing or manual subjective judgment, automatically determining the test endpoint, reducing the average test cycle time by more than 40%, and greatly improving testing efficiency.

[0028] Fully automatic adaptive: The system can automatically switch between high-precision stepped scanning mode and fast linear scanning mode based on the electrical parameters (such as Voc and Isc) obtained from the first pre-scan, thereby intelligently adapting to the testing needs of different specifications, from R&D-grade high-power components to production-line-grade low-power samples, achieving a balance between high throughput and high precision.

[0029] Integrated testing and traceability: In-vitro electrical (IV) performance testing and electronic defect (EL) detection are completed sequentially at the same workstation without requiring changes to wiring or equipment, achieving "one-time wiring, dual testing." Simultaneously, by establishing a link between IV data and EL image data, a complete test data chain is formed, facilitating production traceability and process analysis, and resolving the problem of the separation between electrical performance data and visual defect data in traditional production lines. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 is a schematic diagram of the system structure provided in an embodiment of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] The purpose of this invention is to provide an intelligent IV testing and grading system and method that can overcome the limitations of perovskite material properties, achieve non-destructive, rapid, adaptive testing, and integrate performance testing and defect testing.

[0034] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] Example 1: This example provides an intelligent IV testing and grading system and method for perovskite solar cells, aiming to achieve non-destructive, rapid, and adaptive integrated performance testing and defect detection.

[0036] 1. System Hardware Composition The system mainly includes the following core modules: Simulated Illumination Module: This module uses an AAA+ grade tunable spectral LED array as the light source. Its spectrum conforms to the IEC 60904-9 standard and has microsecond-level switching response capability. This module has a built-in closed-loop feedback system that monitors irradiance in real time at a frequency of 1kHz via a reference battery, dynamically adjusting the LED drive current to ensure light intensity non-uniformity <1% and long-term instability <0.5%. It is also equipped with a TEC (thermal refrigeration) temperature-controlled chuck to maintain the test environment temperature at 25℃±0.5℃. The core function of this module is to generate millisecond-level pulsed light with an adjustable pulse width within the range of 10ms to 100ms to address the thermal sensitivity issue of perovskite materials.

[0037] IV Test Module: The core device is a high-speed source measurement unit (SMU) based on an FPGA architecture. Its performance specifications include: voltage range 0-200V, current range 10nA-20A, sampling accuracy up to 24-bit, minimum sampling interval of 10 microseconds (i.e., maximum sampling rate of 100kHz), and a maximum scan speed of 2000 points / second. This module is configured to work in conjunction with the analog illumination module, synchronously triggering IV scan sampling when the light pulse reaches the flat-top stable region (e.g., 5ms after pulse activation), and immediately turning off the light source after sampling.

[0038] The visual recognition module includes a 20-megapixel GigE interface industrial camera with a resolution of 5472×3648, equipped with an infrared filter lens. This camera is vertically mounted approximately 1.5 meters above the test bench to cover the test field of view for components up to 1600mm×1200mm in size.

[0039] The AI ​​analysis module (computing center) is deployed locally (edge ​​computing), using an industrial control computer (IPC) configured with an Intel Core i7 processor and an NVIDIA RTX 4060 or higher graphics card to leverage CUDA for accelerated model inference. The software environment is based on Python 3.9 and PyTorch. This module runs a "hybrid model architecture," including: a numerical analysis model employing a combination of LSTM (Long Short-Term Memory) and XGBoost. LSTM is used to capture the perovskite-specific hysteresis and photoresponse dynamics in the IV data time series, while XGBoost is used for fast regression prediction. Model inputs include real-time acquired voltage (V), current (I) sequences, component temperature (T), current test count (N), and historical Pmax rate of change. Outputs include control commands (0 to continue testing / 1 to stop testing), as well as the predicted final Pmax value, photoelectric conversion efficiency (Eff), and grading recommendations (A / B / C).

[0040] Image analysis model: A convolutional neural network (CNN) based on the YOLOv8 architecture is used to process EL images and output the defect category (such as hidden cracks, black spots) and location coordinates.

[0041] The grading execution module connects to the AI ​​analysis module and drives the sorting mechanism to perform corresponding physical sorting operations based on the grading instructions (such as A, B, and C products) issued by the AI ​​analysis module.

[0042] Module Connection and Collaboration: Modules are connected via a high-speed bus. Industrial Ethernet (GigE) is used to transmit image data and IV waveform data; the PCIe bus is used for high-speed communication of the acquisition card; Digital I / O (TTL level) is used to achieve microsecond-level hardware synchronization triggering between the light source, source meter, and camera. The entire system follows a collaborative core of "event triggering + data driving." For example, barcode recognition triggers parallel initialization of each module (total time ≤ 2 seconds), IV data is pushed to the AI ​​module in real time for analysis and feedback decision-making, and any module failure can trigger a system pause and redirect the component to the fault channel.

[0043] 2. The intelligent testing and grading method applied to the above system includes the following steps: S1: System initialization and pre-scan. Place the perovskite solar cell or module under test at the test location. The system first performs a quick "pre-scan" (taking approximately 50ms) to obtain the initial values ​​of the module's open-circuit voltage (Voc) and short-circuit current (Isc).

[0044] Adaptive scan mode selection: The AI ​​analysis module automatically selects the subsequent IV scan mode based on the pre-scan results. The judgment thresholds are set as follows: high power standard (Voc≥1.2V and Isc≥2A), low power standard (Voc<0.8V or Isc<0.5A).

[0045] If the high power standard is met, select the Step Scan mode: set the voltage step size to 10mV (0.01V), and delay each point by 1ms to eliminate the capacitance effect and improve accuracy.

[0046] If the low power standard is met, select the Linear Sweep mode: set the scan rate to 5V / s and the sampling interval to 200μs, sacrificing a little accuracy for faster testing speed.

[0047] S2: Millisecond-level pulse IV test. The simulated illumination module applies pulsed light with an adjustable pulse width (e.g., 10-100ms) to the device under test. When the light pulse reaches its stable region (e.g., 5ms after activation), the IV test module is synchronously controlled to perform IV scanning sampling according to the mode selected in S1. The light source is immediately turned off after sampling, and the duty cycle is controlled (<5%) to ensure no heat accumulation. This design keeps the component temperature rise during the test within 0.1℃.

[0048] S3: AI Intelligent Analysis and Convergence Judgment. The IV data collected in S2 is input into the numerical analysis model of the AI ​​analysis module in real time. The model calculates the peak power Pmax(N) of the current Nth test and compares it with Pmax(N-1) of the previous test, calculating the rate of change Δ: Δ = (Pmax(N) - Pmax(N-1)) / Pmax(N-1). The system makes a judgment based on preset decision rules: Case A (Temperature / Non-convergence): If Δ > 0.1%, the data is determined to be non-converged, the instruction "Continue Testing" (N=N+1) is output, and the system returns to step S2 for the next scan.

[0049] Case B (Stable / Converging): If the case of |Δ|<0.1% occurs twice consecutively, it is determined that a steady state has been reached, and the command "Stop testing" is output.

[0050] Scenario C (Abnormal): If Δ < -0.2%, it is determined that light decay or damage may occur, an "emergency stop" command is output, and the component is marked as needing to be re-inspected.

[0051] This closed-loop logic replaces manual judgment and enables adaptive determination of the test endpoint.

[0052] S4: Iterative test. If it is determined to be non-convergent in S3 (case A), then repeat S2 and S3 until the convergence (case B) or abnormal (case C) conditions are met.

[0053] S5: Synchronous Acquisition of EL Defect Images. Once the test is deemed convergent, the AI ​​analysis module controls the IV test module to switch from "measurement mode" to "power mode," injecting a forward bias current approximately 1.5 times its short-circuit current (Isc) into the component, exciting it to generate infrared electroluminescence (EL). Simultaneously, the industrial camera of the vision recognition module is synchronously triggered via hardware to capture EL images.

[0054] S6: Comprehensive Analysis and Grading Decision. The image analysis model (YOLOv8) of the AI ​​analysis module processes the EL image captured by S5 to identify the defect category and location. Subsequently, the AI ​​module integrates the final electrical performance parameters (such as Pmax, Eff) output by the numerical analysis model and the defect results output by the image analysis model, and generates the final grading instruction based on preset grading standards (such as Grade A: high efficiency and no defects; Grade B: qualified efficiency or minor defects; Grade C: low efficiency or serious defects).

[0055] S7: Execute grading. The grading execution module receives grading instructions from the AI ​​analysis module and drives the sorting mechanism to move components to the corresponding A, B, or C product areas or fault handling channels.

[0056] 3. Supplementary Implementation Details: Automatic Size Recognition: If the component barcode does not contain size information, the system can use a visual recognition module (using an auxiliary camera with ≥5 million pixels) to capture the component outline, automatically calculate the size based on the image edge detection algorithm (error ≤±2mm), and match the corresponding test parameters.

[0057] Data Traceability and Association: The system establishes a unique data file for each component. Using the component ID, the IV waveform data, EL image data, and final grading results obtained from the test are automatically associated with the production process data (such as coating temperature and annealing time) in the Manufacturing Execution System (MES), forming a complete traceability chain of "production process - test parameters - grading results". Simultaneously, an MD5 fingerprint association is established between the IV data and EL image data to ensure data consistency and facilitate integrated analysis and process optimization.

[0058] Data Collaborative Management: Based on the native adaptation interface of the self-developed MYS-MES system, test data achieves real-time backup and high-speed push. Synchronization to the MES system occurs within one second after test completion, ensuring the timeliness of data flow. The system employs a dual mechanism of "real-time push + scheduled verification," performing MD5 value comparison verification on the stored data in the test system and MES system every hour to prevent data loss, tampering, or inconsistency, ensuring the security and accuracy of cross-system data links. Furthermore, the test system can receive production work orders, component specifications, test priorities, and other instructions from the MES system, achieving bidirectional linkage of "work order-driven testing and test feedback to process," deeply integrating the testing process into the entire production line process control.

[0059] This embodiment combines the aforementioned hardware system with intelligent methods to achieve non-destructive, accurate, and rapid testing and grading of perovskite solar cells, effectively overcoming the problems of thermal damage and hysteresis interference, and improving the adaptive capability and integrated detection level of the testing system.

[0060] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0061] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A smart IV testing and grading system for perovskite solar cells, characterized in that, include: The simulated illumination module is used to provide controllable simulated illumination to the perovskite solar cell or module under test; The IV test module is electrically connected to the simulated illumination module and the battery or component under test, and is used to perform current-voltage scanning and collect IV data under illumination conditions; The system includes a visual recognition module for acquiring images of the battery or component under test under specific stimuli; an AI analysis module, communicatively connected to both the IV testing module and the visual recognition module, for receiving and analyzing the IV data and image processing data; and a hierarchical execution module, communicatively connected to the AI ​​analysis module, for performing hierarchical operations based on the output of the AI ​​analysis module. The simulated illumination module is configured to output pulsed light with millimeter-level pulse widths. The penalties imposed by the IV testing module, the simulated illumination module, and the visual recognition module are controlled by the AI ​​analysis module to collaboratively complete the test.

2. The intelligent IV testing and grading system according to claim 1, characterized in that, The simulated illumination module includes an adjustable spectrum LED array and a closed-loop feedback control unit. The closed-loop feedback control unit is used to dynamically adjust the LED driving current according to the irradiance monitored in real time by the reference battery, so as to control the light intensity non-uniformity to be less than 1%, and the pulse width of the LED array is adjustable in the range of 10ms to 100ms.

3. The intelligent IV testing and grading system according to claim 1, characterized in that, The IV test module includes a high-speed source measurement unit based on an FPGA architecture, with a minimum sampling interval of 10 microseconds and a maximum scanning speed of 2000 points / second. The IV test module is configured to synchronously trigger IV scanning sampling when the light pulse emitted by the simulated illumination module reaches the stable region.

4. The intelligent IV testing and grading system according to claim 3, characterized in that, The AI ​​analysis module includes a locally deployed industrial control computer and a hybrid AI model running on it; the hybrid AI model includes a numerical analysis model for analyzing IV data time series to predict steady-state performance parameters, and an image analysis model for analyzing images acquired by the visual recognition module to identify defects.

5. The intelligent IV testing and grading system according to claim 4, characterized in that, The numerical analysis model is built on Long Short-Term Memory Network and XGBoost, and the image analysis model is built on YOLOv8 architecture. The AI ​​analysis module is configured to calculate the peak power change rate between the current test and the previous test based on real-time IV data, and automatically decide whether to continue testing, stop testing, or handle anomalies based on the change rate Δ value.

6. The intelligent IV testing and grading system according to claim 1, characterized in that, The visual recognition module includes an industrial camera. The system is configured to: after the IV test module completes the electrical performance test determination, control the IV test module to switch to injecting a positive bias current into the battery or component under test to excite it to generate electroluminescence, and simultaneously trigger the industrial camera to capture EL images.

7. A smart IV testing and grading method for perovskite solar cells, applied to the system as described in any one of claims 1 to 6, characterized in that, Includes the following steps: S1: System initialization, placing the perovskite solar cell or module under test in the test position; S2: Controlling the simulated illumination module to apply millisecond-level pulsed light to the cell or module under test, and synchronously controlling the IV test module to perform IV scanning during the stable light pulse period, acquiring the first set of IV data; S3: Inputting the acquired IV data into the AI ​​analysis module for analysis, the AI ​​analysis module determines whether the test has converged based on preset decision rules; S4: If it is determined that the test has not converged, repeating steps S2-S3 for the next test, until the AI ​​analysis module determines that the test has converged or triggers an abnormal stop; S5: After the test converges, controlling the IV test module to inject current into the cell or module under test to excite electroluminescence, and controlling the visual recognition module to acquire EL images; S6: The AI ​​analysis module comprehensively analyzes the IV data results and the defect identification results of the EL images, and outputs a grading instruction; S7: The hierarchical execution module executes the corresponding hierarchical operation according to the hierarchical instruction.

8. The intelligent IV testing and grading method according to claim 7, characterized in that, In step S2, the IV test module is controlled to trigger IV scan sampling 5ms after the light pulse is turned on, and the light source is turned off immediately after the scan is completed.

9. The intelligent IV testing and grading method according to claim 7, characterized in that, In step S3, the preset decision rule includes: calculating the rate of change Δ between the current test peak power Pmax(N) and the previous test peak power Pmax(N-1); if Δ > 0.1%, it is determined to be non-converged, and the test continues; if If the condition occurs twice consecutively, it is considered convergence and the test is stopped; if Δ < -0.2%, it is considered abnormal and an emergency stop is triggered.

10. The intelligent IV testing and grading method according to claim 7, characterized in that, Before step S2, an adaptive scan mode selection step is also included: perform a pre-scan to obtain the open-circuit voltage Voc and short-circuit current Isc of the battery or component under test; if Voc ≥ 1.2V and Isc ≥ 2A, then select the stepped scan mode for subsequent IV scans; if Voc < 0.8V or Isc < 0.5A, then select the linear scan mode for subsequent IV scans.

Images