Self-adaptive testing and fault tracking method and system

Abstract

The invention provides a self-adaptive test and fault tracking method and system. The method comprises the following steps: acquiring original data of a tested system; processing the original data, identifying abnormal data and generating a trigger signal; extracting a fault waveform feature vector according to the trigger signal, and judging a fault type according to the fault waveform vector; generating a corresponding maintenance strategy based on the fault type, and executing online compensation or shutdown operation; and repeatedly executing the steps until the steady-state parameters are output when no abnormal data is detected for continuous N times. Through obtaining the original data of the detected system and carrying out intelligent processing, abnormal data can be identified in time and a trigger signal is generated so that sensitivity and accuracy of fault detection are effectively increased.

Description

Technical Field

[0001] The present invention relates to the field of automated testing technology, and in particular to an adaptive testing and fault tracing method and system. Background Art

[0002] With the increasing complexity of software systems and rising user quality requirements, traditional testing methods and fault tracing technologies face numerous challenges. In modern software development environments, rapid iteration, multi-platform compatibility, and continuous integration / continuous delivery (CI / CD) have become mainstream development models, placing higher demands on testing efficiency and defect management. Existing testing systems generally suffer from issues such as insufficient automation coverage, high test script maintenance costs, and untimely defect feedback, making them unable to meet the growing demand for quality assurance. In particular, during the testing of complex systems, test data collection, processing, and defect location often require extensive manual effort, which is not only inefficient but also prone to missed tests and misjudgments.

[0003] Traditional bug tracking systems and test systems are usually independent of each other, resulting in difficulties in data sharing, serious information island phenomena, and a lack of effective closed-loop management mechanisms. Although existing technologies can achieve the automation of certain test tasks and basic defect recording functions, they still have major deficiencies in the intelligent identification of abnormal data, automatic determination of fault types, and adaptive optimization of maintenance strategies. Especially when faced with complex fault types such as system parameter offsets, hardware component failures, and signal transmission anomalies, existing systems find it difficult to provide accurate diagnosis and effective repair solutions. They also lack the ability to dynamically adjust test strategies and fault determination criteria, and are unable to adapt to ever-changing test environments and needs. Therefore, there is an urgent need for an intelligent adaptive system that integrates test execution, data processing, fault tracking, maintenance execution, and feedback optimization. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems of low efficiency, lack of self-adaptation and the like caused by the separation of the existing test system and the fault tracing mechanism.

[0005] A first aspect of the present invention provides an adaptive testing and fault tracing method, comprising:

[0006] Obtain the original data of the system under test;

[0007] Processing the raw data, identifying abnormal data and generating a trigger signal;

[0008] extracting a fault waveform feature vector according to the trigger signal, and determining the fault type according to the fault waveform vector;

[0009] Generate a corresponding maintenance strategy based on the fault type and perform online compensation or shutdown operations;

[0010] Repeat the above steps until no abnormal data is detected for N consecutive times and the steady-state parameters are output.

[0011] Furthermore, the raw data includes: current, temperature and pressure data of the measured system.

[0012] Furthermore, processing the raw data includes performing gradient filtering and noise suppression, identifying abnormal data exceeding a threshold, and generating a trigger signal.

[0013] Furthermore, determining the fault type according to the fault waveform vector includes: inputting the fault waveform feature vector into a pattern classifier to determine the fault category.

[0014] Furthermore, the fault types include: hardware component failure, parameter deviation exceeding a threshold, abnormal signal transmission, system response delay, and system state instability.

[0015] Furthermore, generating a corresponding maintenance strategy based on the fault type includes: calling a repair instruction in a strategy library based on the fault type.

[0016] Furthermore, it also includes:

[0017] The fault characteristics and the repair process are associated and stored in a historical database, and the abnormal data recognition threshold and fault judgment parameters are dynamically updated.

[0018] Furthermore, the dynamic update includes adjusting the sensitivity of abnormal data identification and the accuracy of fault type determination based on the success rate of fault repair.

[0019] A second aspect of the present invention provides an adaptive testing and fault tracing system, comprising:

[0020] Test execution unit, the step of obtaining the raw data of the system under test;

[0021] The data processing unit processes the raw data, identifies abnormal data and generates a trigger signal;

[0022] a tracking unit, extracting a fault waveform feature vector according to the trigger signal, and determining a fault type according to the fault waveform vector;

[0023] A maintenance unit generates a corresponding maintenance strategy based on the fault type and performs online compensation or shutdown operations;

[0024] Feedback optimization unit optimizes test parameters and fault identification thresholds based on fault handling results and system response data.

[0025] Furthermore, the test execution unit includes a test controller and a test executor, the test controller is used to formulate a test strategy and send test instructions to the test executor, and the test executor is connected to the test instrument and the test tool to directly collect the original data of the system under test;

[0026] The data processing unit includes a data acquisition module, a data processing module and a data storage module. The data acquisition module receives the original data transmitted by the test execution unit, the data processing module performs data filtering and abnormality identification, and the data storage module stores the processing results and exchanges data with other units.

[0027] The tracking unit includes a bug tracking module and a fault location module. The bug tracking module receives a trigger signal from the data processing unit and extracts a feature vector. The fault location module determines the fault type according to the feature vector and sends fault information to the maintenance unit.

[0028] The maintenance unit includes a maintenance module and a repair module. The maintenance module formulates a maintenance strategy according to the fault type, and the repair module performs online compensation or shutdown operations and feeds back the results to the feedback optimization unit.

[0029] The feedback optimization unit includes a parameter optimization module and a threshold adjustment module. The parameter optimization module adjusts the test parameters of the test execution unit, and the threshold adjustment module optimizes the abnormality recognition threshold of the data processing unit and the fault judgment standard of the tracking unit.

[0030] Compared with the prior art, the present invention includes at least the following beneficial effects: by acquiring the raw data of the system under test and performing intelligent processing, abnormal data can be identified in a timely manner and a trigger signal can be generated, effectively improving the sensitivity and accuracy of fault detection. The steps of extracting the fault waveform characteristic vector and determining the fault type based on the trigger signal achieve accurate positioning and classification of the fault, avoiding misjudgments and missed judgments in traditional methods. Automatically generating maintenance strategies and executing corresponding operations based on the fault type greatly improves the efficiency and pertinence of fault handling, and reduces system downtime and maintenance costs. By iterating the test process in a loop until no abnormality is detected for N consecutive times, the stability and reliability of the system are ensured, and the output steady-state parameters provide an important reference for subsequent operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings such as the provided drawings can be obtained without creative work.

[0032] Figure 11 is a flow chart of an adaptive testing and fault tracing method according to an embodiment of the present invention;

[0033] Figure 2 FIG. 1 is a schematic diagram of modules of an adaptive testing and fault tracing system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The present invention will be described in more detail below with reference to schematic diagrams, which illustrate preferred embodiments of the present invention. It should be understood that those skilled in the art may modify the present invention as described herein while still achieving the advantageous effects of the present invention. Therefore, the following description should be understood as being generally known to those skilled in the art and not as limiting the present invention.

[0035] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0036] The following paragraphs describe the present invention in more detail by way of example with reference to the accompanying drawings. The advantages and features of the present invention will become more apparent as described below. It should be noted that the drawings are greatly simplified and not to exact scale, and are provided solely for the purpose of assisting in the description of the embodiments of the present invention.

[0037] Example 1

[0038] The first aspect of the present invention provides an adaptive testing and fault tracing method, please refer to Figure 1 ,include:

[0039] Obtain the original data of the system under test;

[0040] Processing the raw data, identifying abnormal data and generating a trigger signal;

[0041] extracting a fault waveform feature vector according to the trigger signal, and determining the fault type according to the fault waveform vector;

[0042] Generate a corresponding maintenance strategy based on the fault type and perform online compensation or shutdown operations;

[0043] Repeat the above steps until no abnormal data is detected for N consecutive times and the steady-state parameters are output.

[0044] Raw data from the system under test can be directly collected through sensors or test instruments. Raw data includes current, temperature, and pressure. Processing of this raw data involves performing gradient filtering and noise suppression, identifying abnormal data exceeding a threshold, and generating a trigger signal. Gradient filtering effectively removes noise, while noise suppression reduces interference in the data. Identifying abnormal data exceeding a threshold is accomplished by setting a threshold and performing a comparison. Based on the trigger signal, a fault waveform feature vector is extracted. This feature vector is obtained by extracting features from the abnormal data. The fault type is determined based on the feature vector. This feature vector can be input into a pattern classifier to determine the fault type. Based on the fault type, a corresponding maintenance strategy is generated. The maintenance strategy can invoke repair instructions from a strategy library based on the fault type. Online compensation or shutdown operations are then executed. Online compensation can be performed while the system is running, while shutdown operations can be performed when the system cannot continue to operate. Repeat these steps until no abnormal data is detected for N consecutive times. Steady-state parameters are then output, reflecting the system's stable state.

[0045] In this embodiment, N may be an integer greater than or equal to 1.

[0046] By acquiring raw data, processing it, identifying anomalies, extracting features, determining fault types, generating maintenance strategies, and executing actions, this method effectively identifies and processes abnormal data in the system. It then generates appropriate maintenance strategies based on the fault type, executes online compensation or shutdown operations, and ultimately outputs steady-state parameters. This method improves test efficiency and fault tracing accuracy, reduces manual intervention, lowers the risk of missed tests and misjudgments, and adapts to changing test environments and requirements.

[0047] Furthermore, the raw data includes: current, temperature and pressure data of the measured system.

[0048] Specifically, current data reflects the system's electrical state, temperature data monitors the system's thermal stability, and pressure data assesses the system's mechanical performance. This data is received by the data acquisition module, filtered and anomaly identified by the data processing module, and ultimately stored in the data storage module for subsequent analysis.

[0049] Current data can be collected using current sensors, temperature data using temperature sensors, and pressure data using pressure sensors. These sensors convert physical quantities into electrical signals, which are then collected by the test instruments and test tools used to test the actuator. The data processing module further performs gradient filtering and noise suppression to identify abnormal data that exceeds a threshold and generate a trigger signal.

[0050] Abnormal current data may manifest as current fluctuations or current outside the normal range, abnormal temperature data may manifest as excessively high or low temperatures, and abnormal pressure data may manifest as excessively high or low pressure. After filtering and noise suppression in the data processing module, these abnormal data can more accurately identify potential faults in the system.

[0051] Furthermore, processing the raw data includes performing gradient filtering and noise suppression, identifying abnormal data exceeding a threshold, and generating a trigger signal.

[0052] Specifically, gradient filtering is a filtering method based on the rate of change of data, which can effectively remove high-frequency noise in the data while retaining the trend information of the data. Noise suppression further reduces the random noise in the data through specific algorithms (such as wavelet transform or Kalman filtering) and improves the signal-to-noise ratio of the data. After completing filtering and noise suppression, the system will set a dynamic threshold to identify abnormal data that exceeds the normal range. When abnormal data is detected, the system will generate a trigger signal, which is used to start the subsequent fault tracking and maintenance process. As a preferred embodiment, gradient filtering can use Sobel operator or Prewitt operator, noise suppression can use median filtering or Gaussian filtering, and the dynamic threshold can be adaptively adjusted according to historical data or system status.

[0053] Gradient filtering and noise suppression techniques effectively improve the quality of raw data and reduce noise interference with abnormal data identification, thereby enhancing the accuracy and reliability of fault detection. This not only enables more precise identification of abnormal data, but also adapts to different test environments and system states through dynamic threshold adjustment, resolving the issues of misjudgment and missed detection caused by noise interference and fixed thresholds in traditional methods.

[0054] Furthermore, determining the fault type according to the fault waveform vector includes: inputting the fault waveform feature vector into a pattern classifier to determine the fault category.

[0055] Specifically, the pattern classifier can employ a variety of machine learning algorithms, such as support vector machines (SVMs), random forests, or neural networks. As a preferred embodiment, the pattern classifier can be trained using a training dataset containing waveform feature vectors of various known fault types and their corresponding fault category labels. This allows the pattern classifier to output the corresponding fault category based on the input fault waveform feature vector.

[0056] In one possible embodiment, the pattern classifier training process can include the following steps: first, collect a large amount of historical fault data, extract the waveform feature vectors for each fault type, and label them with the corresponding fault category; second, divide the data set into a training set and a test set, use the training set to train the pattern classifier, and verify the classifier's accuracy and generalization ability with the test set; finally, deploy the trained pattern classifier into the system for real-time fault classification. In addition, the pattern classifier can be regularly updated to adapt to new fault types or changes in the system environment.

[0057] Furthermore, the fault types include: hardware component failure, parameter deviation exceeding a threshold, abnormal signal transmission, system response delay, and system state instability.

[0058] Among them, hardware component failure refers to system function abnormality caused by damage or aging of hardware components; parameter deviation failure exceeding the threshold refers to the system operating parameters exceeding the preset normal range; signal transmission abnormality failure involves signal distortion or interruption during transmission; system response delay failure is manifested as the system's response time to the input signal exceeds expectations; system state instability failure refers to the system's frequent state fluctuations or inability to maintain a stable state during operation.

[0059] Specifically, hardware component failures can be prevented by regularly checking the status of hardware components; parameter deviation failures exceeding the threshold can be avoided by real-time monitoring of system parameters and timely adjustment; abnormal signal transmission failures can be solved by optimizing the signal transmission path and enhancing the signal's anti-interference capability; system response delay failures can be improved by optimizing the system scheduling algorithm and resource allocation strategy; and system state instability failures can be addressed by strengthening system stability and redundant design.

[0060] Furthermore, generating a corresponding maintenance strategy based on the fault type includes: calling a repair instruction in a strategy library based on the fault type.

[0061] Specifically, the policy library stores repair instructions corresponding to various fault types, which are optimized and updated based on historical fault handling experience. Once the fault type is determined, the system automatically matches the corresponding repair instructions from the policy library and performs the corresponding maintenance operations based on the instructions. For example, for hardware component failure, the repair instructions may include replacing the faulty component or adjusting system parameters to bypass the faulty component; for parameter offset failures exceeding the threshold, the repair instructions may include recalibrating the sensor or adjusting the system control parameters; for signal transmission abnormality failures, the repair instructions may include reconfiguring the communication module or switching to a backup communication channel; for system response delay failures, the repair instructions may include optimizing the system scheduling algorithm or increasing computing resources; for system state instability failures, the repair instructions may include restarting the system or executing a system self-test program. By calling the repair instructions in the policy library, the system can quickly respond to faults and perform effective maintenance operations, thereby reducing downtime and maintenance costs.

[0062] Specifically, the construction and maintenance of the policy library is a dynamic process, in which repair instructions are generated and optimized based on historical fault handling data and system response data. For example, when a certain fault type is successfully repaired, the system records the key parameters and operating steps of the repair process and adds them to the policy library as new repair instructions. Simultaneously, the system optimizes and adjusts existing repair instructions based on the success rate and effectiveness of the fault repair to improve the accuracy and effectiveness of maintenance strategies. Furthermore, the policy library supports multi-dimensional query and matching capabilities, enabling the system to quickly locate and invoke the most appropriate repair instructions based on different fault characteristics and environmental conditions.

[0063] Furthermore, it also includes:

[0064] The fault characteristics and the repair process are associated and stored in a historical database, and the abnormal data recognition threshold and fault judgment parameters are dynamically updated.

[0065] Specifically, fault characteristics include fault waveform feature vectors, fault types, and system status data when the fault occurs. The repair process includes the execution results of the maintenance strategy and feedback on the effects of online compensation or shutdown operations. These data are associated and stored in the historical database for subsequent query and analysis. The process of dynamically updating abnormal data recognition thresholds and fault judgment parameters is based on the fault repair records and system response data in the historical database. For example, when the repair success rate of a certain type of fault is high, the recognition threshold of this type of fault can be appropriately lowered to improve the fault tolerance of the system; conversely, if the repair success rate of a certain type of fault is low, the recognition threshold of this type of fault can be appropriately increased to avoid misjudgment. In addition, fault judgment parameters can also be optimized based on historical data, such as adjusting the weight of the pattern classifier or updating the criteria for fault type judgment.

[0066] Specifically, the associated storage of fault characteristics and repair processes can be achieved through the design of database tables. For example, a fault record table can be created, which contains fields such as fault ID, fault waveform feature vector, fault type, system status data, maintenance strategy, and repair results. Every time a fault occurs, the system automatically generates a record and stores the relevant information in the table. The process of dynamically updating abnormal data recognition thresholds and fault judgment parameters can be achieved through a regularly running optimization algorithm. For example, a machine learning-based optimization model can be designed, which automatically adjusts the abnormal data recognition thresholds and fault judgment parameters based on the fault repair records and system response data in the historical data. The input of the optimization model includes historical fault records, system status data, repair results, etc., and the output is the updated abnormal data recognition thresholds and fault judgment parameters.

[0067] Furthermore, the dynamic update includes adjusting the sensitivity of abnormal data identification and the accuracy of fault type determination based on the success rate of fault repair.

[0068] Specifically, dynamic updates analyze historical fault repair records, calculate the repair success rate, and adjust the sensitivity threshold for abnormal data identification and the accuracy parameters for fault type determination based on this success rate. For example, when the fault repair success rate is high, the sensitivity of abnormal data identification can be appropriately lowered to reduce the false alarm rate; when the fault repair success rate is low, the sensitivity of abnormal data identification can be increased to ensure that faults are caught promptly. As a preferred embodiment, dynamic updates can be achieved through machine learning algorithms, using historical data to train models and automatically adjust sensitivity and accuracy parameters.

[0069] Example 2

[0070] The second aspect of the present invention provides an adaptive testing and fault tracing system, please refer to Figure 2 ,include:

[0071] Test execution unit, the step of obtaining the raw data of the system under test;

[0072] The data processing unit processes the raw data, identifies abnormal data and generates a trigger signal;

[0073] a tracking unit, extracting a fault waveform feature vector according to the trigger signal, and determining a fault type according to the fault waveform vector;

[0074] A maintenance unit generates a corresponding maintenance strategy based on the fault type and performs online compensation or shutdown operations;

[0075] Feedback optimization unit optimizes test parameters and fault identification thresholds based on fault handling results and system response data.

[0076] Furthermore, the test execution unit includes a test controller and a test executor. The test controller is used to formulate test strategies and send test instructions to the test executor. The test executor connects to test instruments and test tools to directly collect raw data from the system under test. The test controller can be configured with different test strategies to meet different testing requirements, such as periodic testing and event-triggered testing. The test executor, based on the test controller's instructions, collects raw data such as current, temperature, and pressure from the system under test through the test instruments and test tools.

[0077] In one possible embodiment, the test controller includes a control panel and a control circuit, the control panel is fixed to the front end of the test controller, and the control circuit is embedded in the interior of the test controller. The control panel is provided with a plurality of operation buttons and a display screen, the operation buttons are connected to the control circuit, and the display screen is connected to the control circuit. The test executor includes an execution panel and an execution circuit, the execution panel is fixed to the front end of the test executor, and the execution circuit is embedded in the interior of the test executor. The execution panel is provided with a plurality of indicator lights and interfaces, the indicator lights are connected to the execution circuit, and the interfaces are connected to the data cable of the test equipment. The test executor includes a variety of sensors and detection instruments, and the test tools include simulators and emulators, all of which are fixed by a test tool bracket.

[0078] The data processing unit comprises a data acquisition module, a data processing module, and a data storage module. The data acquisition module receives raw data from the test execution unit, the data processing module performs data filtering and anomaly identification, and the data storage module stores the processing results and interacts with other units. After receiving the raw data, the data processing module first performs gradient filtering and noise suppression to remove noise interference from the data. It then identifies anomalies by setting thresholds and generates trigger signals. The data storage module not only stores the processed data but also interacts with other units to ensure real-time and consistent data.

[0079] In one possible embodiment, the data acquisition module includes multiple data acquisition cards and data acquisition circuits, with the data acquisition cards embedded within the data acquisition modules and the data acquisition circuits embedded within the data acquisition cards. The data processing module includes a processing chip and a data processing circuit, with the processing chip embedded within the data processing module. The processing chip includes a data cleaning unit, a data conversion unit, and a data verification unit. The data processing circuit is embedded within the processing chip. The data storage module includes multiple storage media and storage circuits, with the storage media embedded within the data storage modules and the storage circuits embedded within the storage media.

[0080] The tracking unit includes a bug tracking module and a fault location module. The bug tracking module receives a trigger signal from the data processing unit and extracts a feature vector. The fault location module determines the fault type based on the feature vector and sends the fault information to the maintenance unit. After receiving the trigger signal, the bug tracking module extracts the fault waveform feature vector and inputs it into a pattern classifier to determine the fault type. Based on the determination result, the fault location module sends detailed fault information to the maintenance module.

[0081] In one possible embodiment, the bug tracking module includes a tracking chip and a tracking circuit, wherein the tracking chip is embedded in the bug tracking module, and the tracking circuit is embedded in the tracking chip. The fault location module includes a positioning chip and a positioning circuit, wherein the positioning chip is embedded in the fault location module, and the positioning circuit is embedded in the positioning chip. The maintenance module includes a maintenance chip and a maintenance circuit, wherein the maintenance chip is embedded in the maintenance module, and the maintenance circuit is embedded in the maintenance chip.

[0082] The maintenance unit includes a maintenance module and a repair module. The maintenance module formulates a maintenance strategy according to the fault type, and the repair module performs online compensation or shutdown operations and feeds back the results to the feedback optimization unit.

[0083] In one possible embodiment, the maintenance module includes a maintenance chip and a maintenance circuit, wherein the maintenance chip is embedded in the maintenance module and the maintenance circuit is embedded in the maintenance chip. The repair module includes a repair chip and a repair circuit, wherein the repair chip is embedded in the repair module and the repair circuit is embedded in the repair chip.

[0084] The feedback optimization unit includes a parameter optimization module and a threshold adjustment module. The parameter optimization module adjusts the test parameters of the test execution unit, and the threshold adjustment module optimizes the abnormality recognition threshold of the data processing unit and the fault judgment standard of the tracking unit.

[0085] The above examples are used to illustrate the present invention, which are only used to help understand the present invention and are not intended to limit the present invention. Those skilled in the art can make several simple deductions, modifications or substitutions based on the concept of the present invention.

Claims

1. An adaptive testing and fault tracing method, characterized in that: include: Obtain the original data of the system under test; Processing the raw data, identifying abnormal data and generating a trigger signal; extracting a fault waveform feature vector according to the trigger signal, and determining the fault type according to the fault waveform vector; Generate a corresponding maintenance strategy based on the fault type and perform online compensation or shutdown operations; Repeat the above steps until no abnormal data is detected for N consecutive times and the steady-state parameters are output.

2. The adaptive testing and fault tracing method according to claim 1, wherein: The raw data includes: current, temperature and pressure data of the measured system.

3. The adaptive testing and fault tracing method according to claim 1, wherein: Processing the raw data includes performing gradient filtering and noise suppression, identifying abnormal data exceeding a threshold, and generating a trigger signal.

4. The adaptive testing and fault tracing method according to claim 1, wherein: Determining the fault type according to the fault waveform vector includes: inputting the fault waveform feature vector into a pattern classifier to determine the fault category.

5. The adaptive testing and fault tracing method according to claim 4, wherein: The fault types include: hardware component failure, parameter deviation exceeding the threshold, signal transmission abnormality, system response delay and system status instability.

6. The adaptive testing and fault tracing method according to claim 1, wherein: Generating a corresponding maintenance strategy based on the fault type includes: calling a repair instruction in a strategy library based on the fault type.

7. The adaptive testing and fault tracing method according to claim 1, wherein: Also includes: The fault characteristics and the repair process are associated and stored in a historical database, and the abnormal data recognition threshold and fault judgment parameters are dynamically updated.

8. The adaptive testing and fault tracing method according to claim 7, wherein: The dynamic update includes: adjusting the sensitivity of abnormal data identification and the accuracy of fault type determination based on the success rate of fault repair.

9. An adaptive testing and fault tracing system, characterized in that: include: Test execution unit, the step of obtaining the raw data of the system under test; The data processing unit processes the raw data, identifies abnormal data and generates a trigger signal; a tracking unit, extracting a fault waveform feature vector according to the trigger signal, and determining a fault type according to the fault waveform vector; A maintenance unit generates a corresponding maintenance strategy based on the fault type and performs online compensation or shutdown operations; Feedback optimization unit optimizes test parameters and fault identification thresholds based on fault handling results and system response data.

10. The system according to claim 9, characterized in that The test execution unit includes a test controller and a test executor, the test controller is used to formulate test strategies and send test instructions to the test executor, and the test executor is connected to the test instrument and the test tool to directly collect the original data of the system under test; The data processing unit includes a data acquisition module, a data processing module and a data storage module. The data acquisition module receives the original data transmitted by the test execution unit, the data processing module performs data filtering and anomaly identification, and the data storage module stores the processing results and exchanges data with other units. The tracking unit includes a bug tracking module and a fault location module. The bug tracking module receives a trigger signal from the data processing unit and extracts a feature vector. The fault location module determines the fault type according to the feature vector and sends fault information to the maintenance unit. The maintenance unit includes a maintenance module and a repair module. The maintenance module formulates a maintenance strategy according to the fault type, and the repair module performs online compensation or shutdown operations and feeds back the results to the feedback optimization unit. The feedback optimization unit includes a parameter optimization module and a threshold adjustment module. The parameter optimization module adjusts the test parameters of the test execution unit, and the threshold adjustment module optimizes the abnormality recognition threshold of the data processing unit and the fault judgment standard of the tracking unit.

Images