Method and system for monitoring data in real time in chip test

Abstract

The invention relates to the technical field of scheme design for monitoring data in real time in chip testing, in particular to a method and system for monitoring data in real time in chip testing. The method comprises the following steps: collecting a test data stream in real time; performing real-time analysis based on the initial monitoring rule set, and controlling the machine to stop when an alarm condition is met; the core of the method is that in the test process, a monitoring rule set is dynamically updated based on historical test data flow, and a self-adaptive monitoring rule set is generated for subsequent analysis. The system comprises a data acquisition module, a rule management module, a dynamic learning engine module and a real-time analysis and control execution module. According to the invention, the monitoring threshold can adapt to the process drift through a dynamic learning mechanism, so that the false alarm is effectively reduced; through multi-dimensional analysis, trend abnormity can be pre-warned in a prospective manner. The system adopts an external mounting type design, a main test program does not need to be modified, the flexibility is high, and the intelligent level and the production efficiency of chip test monitoring are remarkably improved.

Description

Technical Field

[0001] This invention relates to the technical field of scheme design for real-time data monitoring in chip testing, and specifically to a method and system for real-time data monitoring in chip testing. Background Technology

[0002] With the rapid development of semiconductor technology, the complexity and integration of chips are increasing daily, posing unprecedented challenges to the reliability, efficiency, and intelligence of their testing processes. Real-time monitoring of test data is a crucial step in ensuring product quality during chip mass production testing. Traditional monitoring methods primarily rely on fixed upper and lower thresholds preset within the test program. During testing, if a collected data point exceeds this static threshold range, the system determines the test has failed and triggers an alarm. While simple to implement, this method has significant limitations. First, fixed thresholds cannot accommodate normal process drift caused by equipment aging, environmental temperature fluctuations, and probe card wear during testing, easily generating numerous false alarms and leading to frequent unnecessary shutdowns of testing equipment, severely impacting production efficiency. Second, this static monitoring only focuses on whether a single data point exceeds the limit, lacking the ability to analyze data trends and failing to identify potential anomalies such as data mutations or slow drift. This prevents early warning of impending equipment failures or process problems, missing opportunities for preventative maintenance. Furthermore, any modification to the monitoring logic requires code-level changes, recompilation, and deployment of the test program itself, resulting in a cumbersome process, poor flexibility, and difficulty in responding to rapidly changing testing requirements. Although some existing technologies employ Statistical Process Control (SPC) methods for quality monitoring, they are typically implemented as offline analysis tools or in higher-level data analysis systems, making it difficult to integrate them into the test equipment for truly real-time, closed-loop control. They also cannot make intelligent decisions the instant test data is generated. Therefore, there is an urgent need in this field for a real-time monitoring solution that can adaptively learn, perform multi-dimensional intelligent analysis, and does not affect the flexibility of the main test program.

[0003] Therefore, existing technologies still need further development. Summary of the Invention

[0004] The purpose of this invention is to overcome the above-mentioned technical deficiencies and provide a method and system for real-time data monitoring in chip testing, so as to solve the problems existing in the prior art.

[0005] To achieve the above-mentioned technical objectives, according to a first aspect of the present invention, the present invention provides a method for real-time data monitoring in chip testing, comprising: S1: Real-time acquisition of test data streams generated during chip testing; S2: Perform real-time analysis on the test data stream based on a preset initial monitoring rule set, which is defined by a configuration file; S3: When the real-time analysis results meet the preset alarm conditions, trigger the machine alarm and control the test machine to stop testing; S4: During the testing process, the monitoring rule set is dynamically updated based on the historical test data stream to generate an adaptive monitoring rule set, and the adaptive monitoring rule set is used to perform real-time analysis on subsequent test data streams.

[0006] Specifically, the dynamic updating of the monitoring rule set includes: Based on statistical process control methods, the historical test data stream is learned to update the thresholds in the monitoring rule set.

[0007] Specifically, the statistical process control method is based on a normal distribution model, and the threshold update is calculated using the following formula: Upper Control Limit = μ + k σ; Lower Control Limit = μ-k σ; Wherein, Upper Control Limit represents the updated upper control limit; Lower Control Limit represents the updated lower control limit; μ represents the average value of the historical test data stream; σ represents the standard deviation of the historical test data stream; and k represents the control limit coefficient, which is a real number greater than 0.

[0008] Specifically, the control limit coefficient k is dynamically set according to the criticality of the test item or customer specifications.

[0009] Specifically, the real-time analysis includes: performing multi-dimensional analysis on the test data stream, wherein the multi-dimensional analysis includes data value dimension, time series dimension, and test batch dimension.

[0010] Specifically, the multi-dimensional analysis includes: calculating the difference between data values ​​between adjacent test periods; when the difference exceeds a preset mutation threshold, even if the data value itself is within the qualified range defined by the monitoring rule set, it is determined to meet the alarm conditions.

[0011] Specifically, the degree of difference is calculated using the following formula: D = |X_{n} - X_{n-1}|; Where D represents the difference; X_{n} represents the data value collected in the nth test period; and X_{n-1} represents the data value collected in the (n-1)th test period.

[0012] Specifically, in step S4, the triggering conditions for dynamically updating the monitoring rule set include at least one of the following: The number of test chips has reached the predetermined value, the test time has reached the predetermined value, and an external update command has been received.

[0013] Specifically, in step S3, when a machine alarm is triggered, the test data that caused the alarm and its context information are recorded, and a traceable alarm log is generated.

[0014] According to a second aspect of the present invention, a system for real-time data monitoring in chip testing is provided, comprising: The data acquisition module is used to acquire test data streams generated during chip testing in real time. The rule management module is used to store and provide monitoring rule sets, which include an initial monitoring rule set and an adaptive monitoring rule set. A dynamic learning engine, connected to the data acquisition module and the rule management module, is used to dynamically update the initial monitoring rule set based on historical test data streams, so as to generate the adaptive monitoring rule set in the rule management module; The real-time analysis module, connected to the data acquisition module and the rule management module, is used to call the currently valid monitoring rule set to perform real-time analysis on the test data stream and output the analysis results; The control execution module is connected to the real-time analysis module and is used to trigger an alarm on the test machine and control the test machine to stop testing when the analysis result meets the preset alarm conditions.

[0015] Beneficial effects: The real-time monitoring method and system for chip test data provided by this invention brings several significant benefits compared to existing technologies, mainly reflected in four aspects: intelligence, precision, efficiency, and flexibility.

[0016] In terms of intelligence, this invention introduces a dynamic learning engine to achieve adaptive updates of monitoring rules. The system can automatically calculate and adjust monitoring thresholds based on historical test data, allowing them to fluctuate according to the actual state of the process. This fundamental improvement endows the monitoring system with "intelligence," enabling it to effectively distinguish between normal fluctuations in the process and genuine anomalies. This significantly reduces false alarms and missed alarms caused by rigid fixed thresholds, thereby improving the intelligence and reliability of the monitoring.

[0017] In terms of precision, thanks to dynamically adjusted thresholds, the monitoring accuracy of this invention has achieved a qualitative leap. The system not only focuses on whether the instantaneous values ​​of data points exceed limits, but also, through multi-dimensional analysis, especially the calculation of data differences between adjacent test units, can keenly capture trend anomalies and sudden mutations hidden in massive amounts of data. This proactive monitoring capability allows engineers to intervene before minor anomalies escalate into serious failures, realizing a shift from "post-event judgment" to "pre-event warning," significantly improving the depth and breadth of product quality control.

[0018] In terms of efficiency, this invention directly improves the overall efficiency of the testing production line by reducing unplanned downtime caused by false alarms. Machines can operate continuously in a more stable state, saving valuable testing time. At the same time, the detailed and traceable alarm logs provided by the system offer strong support for engineers to quickly locate the root cause of problems, shortening the troubleshooting and process debugging cycle, and further improving production and operational efficiency.

[0019] In terms of flexibility, this invention inherits the advantages of an add-on design, separating the core monitoring logic from the main test program. All monitoring rules and learning algorithm parameters can be defined and modified through configuration files without affecting the test program itself. This modular and configurable design gives the solution extremely high flexibility, enabling it to quickly adapt to the diverse needs of different products, different testing stages, and different customer specifications, greatly reducing system maintenance costs and deployment difficulty. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating a method for real-time data monitoring in chip testing provided in a specific embodiment of the present invention. Figure 2 This is a schematic diagram of the system composition for real-time data monitoring in chip testing, provided in a specific embodiment of the present invention. Detailed Implementation

[0021] To enable those skilled in the art to better understand the technical solutions of the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Based on the embodiments in this application, other similar embodiments obtained by those skilled in the art without creative effort should all fall within the scope of protection of this application. Furthermore, directional terms mentioned in the following embodiments, such as "up," "down," "left," and "right," are only for reference to the directions in the accompanying drawings; therefore, the directional terms used are for illustrative purposes and not for limiting the invention.

[0022] The present invention will be further described below with reference to the accompanying drawings and preferred embodiments.

[0023] Please see Figure 1This invention provides a method for real-time data monitoring during chip testing, comprising: S1: Real-time acquisition of test data streams generated during chip testing.

[0024] It should be further explained that when the testing equipment tests the chip, it periodically generates data containing multiple test items (such as current, voltage, and frequency). The data acquisition module actively captures or passively receives this data through the communication interface and organizes it into a test data stream arranged in time sequence. The data stream contains not only the data values ​​themselves, but also metadata such as test time, chip identifier, and test item name.

[0025] S2: Perform real-time analysis on the test data stream based on a preset initial monitoring rule set, which is defined by a configuration file.

[0026] Specifically, the real-time analysis includes: performing multi-dimensional analysis on the test data stream, wherein the multi-dimensional analysis includes data value dimension, time series dimension, and test batch dimension.

[0027] Specifically, the multi-dimensional analysis includes: calculating the difference between data values ​​between adjacent test periods; when the difference exceeds a preset mutation threshold, even if the data value itself is within the qualified range defined by the monitoring rule set, it is determined to meet the alarm conditions.

[0028] It should be further noted that the initial monitoring rule set is stored in a configuration file, which is predefined by engineers according to the product specifications before the test begins. Its content is as follows: "[Voltage_Test]" Upper_Limit=3.5 Lower_Limit=2.8 Monitoring_Method=ValueCheck".

[0029] Furthermore, the real-time analysis module loads this configuration. For each newly arrived test data point, the module finds the corresponding rule based on its test item name and makes a judgment. For example, if the measured value of the Voltage_Test item is 3.6V, it is judged as unqualified.

[0030] Understandably, the beneficial effects of the above solution include: this step inherits the advantages of the original solution, such as flexibility and no need to modify the main program, and provides basic monitoring capabilities.

[0031] Specifically, the degree of difference is calculated using the following formula: D = |X_{n} - X_{n-1}|; Where D represents the difference; X_{n} represents the data value collected in the nth test period; and X_{n-1} represents the data value collected in the (n-1)th test period.

[0032] S3: When the real-time analysis results meet the preset alarm conditions, trigger the machine alarm and control the test machine to stop testing.

[0033] It should be further explained that the multi-dimensional analysis function of the real-time analysis module also includes time series analysis. Besides checking whether data values ​​exceed limits, it also calculates the difference (D) between data values ​​from adjacent test cycles (i.e., the same test item between the previous chip and the current chip). The difference is calculated using the following formula: D = |X_{n} - X_{n-1}| where, D represents the degree of difference, which is the absolute change in the measured value between adjacent test units.

[0034] X_{n} represents the data value collected in the nth test cycle (current chip).

[0035] X_{n-1} represents the data value collected in the (n-1)th test cycle (the previous chip).

[0036] Preferred values ​​and reasons for selection: The preferred value for the mutation threshold (D_{max}) is 3σ. The rationale is that, based on statistical principles, the normal fluctuation differences between adjacent points should also roughly follow a normal distribution. Setting the mutation threshold to three times the current process standard deviation σ can effectively capture abnormal mutations that occur drastically even when individual data points do not exceed limits (such as sudden poor contact of the probe card). Such mutations are often early signs of impending equipment problems.

[0037] If the calculated D > D_{max}, the real-time analysis module will determine that the alarm condition is met even if X_n itself is within the control limit.

[0038] Understandably, the beneficial effects of the above solution include: enabling more comprehensive and forward-looking monitoring of the testing process, focusing not only on anomalies at "points" but also on anomalies in "lines" (trends), providing early warnings before serious problems occur, and achieving predictive maintenance.

[0039] Specifically, in step S3, when a machine alarm is triggered, the test data that caused the alarm and its context information are recorded, and a traceable alarm log is generated.

[0040] It should be further noted that the trigger conditions for dynamically updating the monitoring rule set can be defined in the configuration file. The preferred combination is: Update_Trigger=(Chips_Tested>=100)OR(Time_Elapsed>=3600)OR(External_Signal==True) Understandably, the benefits of the above solution include: providing a variety of flexible update strategies, which can be triggered by quantity, by time, or manually by engineers, adapting to different production rhythms and engineering analysis needs.

[0041] It should be further explained that the alarm condition is usually "data value exceeds the upper or lower limit". After receiving the unqualified analysis result, the control execution module immediately sends a command to the test machine through the communication interface, triggering an audible and visual alarm and controlling the machine to stop the current test process. At the same time, the current chip is marked as a special failure Bin (such as Bin99).

[0042] S4: During the testing process, the monitoring rule set is dynamically updated based on the historical test data stream to generate an adaptive monitoring rule set, and the adaptive monitoring rule set is used to perform real-time analysis on subsequent test data streams.

[0043] Specifically, the dynamic updating of the monitoring rule set includes: Based on statistical process control methods, the historical test data stream is learned to update the thresholds in the monitoring rule set.

[0044] Specifically, the statistical process control method is based on a normal distribution model, and the threshold update is calculated using the following formula: Upper Control Limit = μ + k σ; Lower Control Limit = μ-k σ; Wherein, Upper Control Limit represents the updated upper control limit; Lower Control Limit represents the updated lower control limit; μ represents the average value of the historical test data stream; σ represents the standard deviation of the historical test data stream; and k represents the control limit coefficient, which is a real number greater than 0.

[0045] It should be further explained that a preferred approach for dynamic updates is to employ statistical process control (SPC). Assuming the test data follows a normal distribution, the dynamic learning engine periodically (e.g., after testing every 100 chips) calculates the mean (μ) and standard deviation (σ) of the historical data. Subsequently, the new control limits are calculated using the following formula: Upper Control Limit = μ + k σ; Lower Control Limit = μ-k σ; in: Upper Control Limit represents the updated control limit, which is the maximum allowable value for data fluctuations; Lower Control Limit represents the updated lower control limit, which is the minimum allowable value for data fluctuations; μ represents the average value of the historical test data stream, which is the average level of the current process. σ represents the standard deviation of the historical test data stream, which is the degree of fluctuation in the current process.

[0046] Furthermore, the preferred value for k is 3. This can be understood as follows: based on the statistical "3σ principle" or "6σ quality management," for a normal distribution, the probability of data falling within (μ-3σ, μ+3σ) is approximately 99.73%. This means that only about 0.27% of the data will be falsely flagged as abnormal due to normal fluctuations, achieving an optimal balance between controlling false alarm rates and capturing true anomalies. The value of k can also be adjusted according to needs: if the testing items are very critical (such as safety-related), k=2 (approximately 95.45% coverage) can be used to tighten control; if slightly larger fluctuations are permissible, k=4 (approximately 99.99% coverage) can be used.

[0047] The optimal amount of historical data is between 100 and 500 data points. Understandably, too little data (e.g., less than 30) results in insufficient statistical significance for the calculated μ and σ, failing to represent the overall process performance; too much data (e.g., over 1000) leads to slow system adaptation, making it difficult to respond quickly to gradual process drift. A range of 100-500 points strikes a good balance between ensuring statistical reliability and response speed.

[0048] Furthermore, after calculating the new control limits, the dynamic learning engine will update the monitoring rule set in the rule management module, and then the real-time analysis module will use the new adaptive monitoring rule set to make judgments.

[0049] Understandably, the beneficial effects of the above solution include: through dynamic learning, the system no longer relies on fixed and rigid specification limits, but establishes control limits based on the actual performance of the process. It can adaptively follow the drift and fluctuation range of the process center, significantly reducing false alarms and missed alarms caused by equipment aging, changes in ambient temperature, etc., and improving the intelligence and accuracy of monitoring.

[0050] Specifically, the control limit coefficient k is dynamically set according to the criticality of the test item or customer specifications.

[0051] It should be further noted that the control limit coefficient k is not fixed. Multiple alternative k values ​​and trigger conditions can be set for a single test item in the rule configuration file. For example: "[Critical_Voltage_Test]" Upper_Limit=3.5 Lower_Limit=2.8 Monitoring_Method=SPC k_Normal=3 k_Tight=2 Tighten_Condition=Lot_ID=='VIP123'# Enables tightening control when encountering specific important customer batches.

[0052] Furthermore, the dynamic learning engine checks these conditions before updating the rules, thereby dynamically selecting the k value.

[0053] Understandably, the benefits of the above solution include: providing flexibility to meet the differentiated quality requirements of different customers and product grades.

[0054] Specifically, in step S4, the triggering conditions for dynamically updating the monitoring rule set include at least one of the following: The number of test chips has reached the predetermined value, the test time has reached the predetermined value, and an external update command has been received.

[0055] It should be further noted that the trigger conditions for dynamically updating the monitoring rule set can be defined in the configuration file. The preferred combination is: Update_Trigger=(Chips_Tested>=100)OR(Time_Elapsed>=3600)OR(External_Signal==True) Understandably, the benefits of the above solution include: providing a variety of flexible update strategies, which can be triggered by quantity, by time, or manually by engineers, adapting to different production rhythms and engineering analysis needs.

[0056] Please see Figure 2 The present invention provides another embodiment, which provides a system for real-time data monitoring in chip testing. The system for real-time data monitoring in chip testing includes: The data acquisition module 100 is used to acquire the test data stream generated during the chip testing process in real time.

[0057] It should be further explained that this module is a resident communication service program that is responsible for establishing a connection with the test machine interface, configuring data parsing rules, parsing the received raw data packets into structured test data points, adding metadata such as timestamps, and then sending them to the real-time analysis module and the dynamic learning engine.

[0058] The rule management module 200 is used to store and provide monitoring rule sets, which include an initial monitoring rule set and an adaptive monitoring rule set.

[0059] It should be further explained that this module manages an in-memory rule database and a configuration file on disk. Upon system startup, it loads the initial monitoring rule set from the configuration file into the in-memory database. After receiving update instructions from the dynamic learning engine, it updates the rules in the in-memory database and can optionally persist the adaptive rule set to disk.

[0060] The dynamic learning engine 300 is connected to the data acquisition module and the rule management module, and is used to dynamically update the initial monitoring rule set based on historical test data streams, so as to generate the adaptive monitoring rule set in the rule management module.

[0061] It should be further explained that this module is a background computing process. It continuously subscribes to the data stream from the data acquisition module and maintains a historical data cache. When preset trigger conditions are met (such as the number of chips reaching 100), the engine retrieves historical data from the cache, executes the SPC algorithm (calculating μ and σ), and calculates the result according to the formula UCL=μ+k. σ,LCL=μ-k σ calculates the new control limits. Finally, it sends the new control limits to the rule management module for updating.

[0062] The real-time analysis module 400 is connected to the data acquisition module and the rule management module, and is used to call the currently effective monitoring rule set to perform real-time analysis on the test data stream and output the analysis results.

[0063] It should be further explained that this module contains the core decision-making logic. It receives each real-time data point from the data acquisition module and immediately queries the rule management module for the currently active rules for that project. It executes all decision-making logic, including checking upper and lower limits, calculating and judging differences, etc. The analysis results are then sent to the control execution module.

[0064] The control execution module 500 is connected to the real-time analysis module and is used to trigger an alarm on the test machine and control the test machine to stop testing when the analysis result meets the preset alarm conditions.

[0065] It should be further noted that this module performs operations based on the output of the real-time analysis module. If the result is satisfactory, no operation is performed. If the result is unsatisfactory or an abnormal alarm is triggered, the module sends standard stop and alarm commands to the testing machine via the communication interface and calls the log service to record detailed alarm logs.

[0066] Furthermore, the output of the data acquisition module is simultaneously connected to both the real-time analysis module and the dynamic learning engine. The rule management module is bidirectionally connected to both the dynamic learning engine (which receives updates) and the real-time analysis module (which provides rule queries). The output of the real-time analysis module is connected to the control execution module.

[0067] In a preferred embodiment, this application also provides an electronic device, the electronic device comprising: The computer device includes a memory and a processor, wherein the memory stores computer-readable instructions that, when executed by the processor, implement the method for real-time data monitoring in chip testing. The computer device can be broadly categorized as a server, terminal, or any other electronic device with the necessary computing and / or processing capabilities. In one embodiment, the computer device may include a processor, memory, network interface, communication interface, etc., connected via a system bus. The processor of the computer device can be used to provide the necessary computing, processing, and / or control capabilities. The memory of the computer device may include a non-volatile storage medium and internal memory. The non-volatile storage medium may store an operating system, computer programs, etc. The internal memory can provide an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface and communication interface of the computer device can be used to connect and communicate with external devices via a network. When the computer program is executed by the processor, it performs the steps of the method of the present invention.

[0068] This invention can be implemented as a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the steps of the methods of embodiments of the invention to be performed. In one embodiment, the computer program is distributed across multiple network-coupled computer devices or processors, such that the computer program is stored, accessed, and executed in a distributed manner by one or more computer devices or processors. A single method step / operation, or two or more method steps / operations, may be executed by a single computer device or processor or by two or more computer devices or processors. One or more method steps / operations may be executed by one or more computer devices or processors, and one or more other method steps / operations may be executed by one or more other computer devices or processors. One or more computer devices or processors may execute a single method step / operation, or execute two or more method steps / operations.

[0069] Those skilled in the art will understand that the method steps of this invention can be performed by a computer program instructing related hardware, such as a computer device or processor, to perform the steps of this invention when executed. Depending on the context, any references herein to memory, storage, databases, or other media may include non-volatile and / or volatile memory. Examples of non-volatile memory include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, magnetic tape, floppy disk, magneto-optical data storage device, optical data storage device, hard disk, solid-state drive, etc. Examples of volatile memory include random access memory (RAM), external cache memory, etc.

[0070] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0071] The specific embodiments of the present invention described above do not constitute a limitation on the scope of protection of the present invention. Any other corresponding changes and modifications made in accordance with the technical concept of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for real-time data monitoring in chip testing, characterized in that, Includes the following steps: S1: Real-time acquisition of test data streams generated during chip testing; S2: Perform real-time analysis on the test data stream based on a preset initial monitoring rule set, which is defined by a configuration file; S3: When the real-time analysis results meet the preset alarm conditions, trigger the machine alarm and control the test machine to stop testing; S4: During the testing process, the monitoring rule set is dynamically updated based on the historical test data stream to generate an adaptive monitoring rule set, and the adaptive monitoring rule set is used to perform real-time analysis on the subsequent test data stream. The real-time analysis includes: performing multi-dimensional analysis on the test data stream, the multi-dimensional analysis including data value dimension, time series dimension and test batch dimension; The multidimensional analysis includes: The difference between data values ​​between adjacent test periods is calculated. When the difference exceeds a preset mutation threshold, even if the data value itself is within the qualified range defined by the monitoring rule set, it is determined to meet the alarm conditions.

2. The method according to claim 1, characterized in that, The dynamic updating of the monitoring rule set includes: Based on statistical process control methods, the historical test data stream is learned to update the thresholds in the monitoring rule set.

3. The method according to claim 2, characterized in that, The statistical process control method is based on a normal distribution model, and the threshold update is calculated using the following formula: Upper Control Limit=μ+k s; Lower Control Limit=μ-k s; Wherein, Upper Control Limit represents the updated upper control limit; Lower Control Limit represents the updated lower control limit; μ represents the average value of the historical test data stream; σ represents the standard deviation of the historical test data stream; and k represents the control limit coefficient, which is a real number greater than 0.

4. The method according to claim 3, characterized in that, The control limit coefficient k is dynamically set according to the criticality of the test item or customer specifications.

5. The method according to claim 1, characterized in that, The degree of difference is calculated using the following formula: D=|X_{n}-X_{n-1}|; Where D represents the difference; X_{n} represents the data value collected in the nth test period; and X_{n-1} represents the data value collected in the (n-1)th test period.

6. The method according to claim 1, characterized in that, In step S4, the triggering conditions for dynamically updating the monitoring rule set include at least one of the following: The number of test chips has reached the predetermined value, the test time has reached the predetermined value, and an external update command has been received.

7. The method according to claim 1, characterized in that, In step S3, when a machine alarm is triggered, the test data that caused the alarm and its context information are recorded, and a traceable alarm log is generated.

8. A system for real-time data monitoring in chip testing, used to implement the method as described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire test data streams generated during chip testing in real time. The rule management module is used to store and provide monitoring rule sets, which include an initial monitoring rule set and an adaptive monitoring rule set. A dynamic learning engine, connected to the data acquisition module and the rule management module, is used to dynamically update the initial monitoring rule set based on historical test data streams, so as to generate the adaptive monitoring rule set in the rule management module; The real-time analysis module, connected to the data acquisition module and the rule management module, is used to call the currently valid monitoring rule set to perform real-time analysis on the test data stream and output the analysis results; The control execution module is connected to the real-time analysis module and is used to trigger an alarm on the test machine and control the test machine to stop testing when the analysis result meets the preset alarm conditions.

Images