Intelligent electromechanical equipment test management method and system, electronic equipment and storage medium

Abstract

The invention relates to the technical field of intelligent electromechanical equipment test management, and discloses an intelligent electromechanical equipment test management method and system, electronic equipment and a storage medium, and the method comprises the steps: collecting the information of intelligent electromechanical equipment; constructing a state space model of the equipment; a deep learning algorithm is applied to train the reinforcement learning model, and an optimal test strategy is generated; generating a test script according to the optimal test strategy, and executing the test script to obtain a test result; optimizing the reinforcement learning model based on the test result; the system comprises an information acquisition module, a state space construction module, a reinforcement learning module, a test execution module and a data storage module. According to the method, the technical scheme of dynamic test script generation is adopted, the technical effect of quickly adapting to the test requirements of different types of equipment is achieved by analyzing the equipment information and the environmental conditions in real time, the test delay is reduced, and the overall test efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent electromechanical equipment test management, and specifically to an intelligent electromechanical equipment test management method, system, electronic device and storage medium. Background Art

[0002] In the modern rail transit industry, intelligent electromechanical equipment is increasingly widely used. These devices include automatic ticket vending machines, security inspection systems, elevator control systems, etc., and undertake the important tasks of ensuring operation safety and improving the passenger experience. With the continuous progress of technology, the complexity and diversity of the devices are also increasing, which puts higher requirements on the test management of the devices.

[0003] Existing rail transit industries generally adopt automated test systems to manage the tests of intelligent electromechanical equipment. These systems usually rely on preset test scripts and can improve the test efficiency to a certain extent. Through automated tools, testers can quickly execute a series of standardized test steps, reducing the time and errors of manual operations.

[0004] However, for the existing intelligent electromechanical equipment test technology, its fixed test scripts are difficult to cope with the rapid updates and diverse requirements of the equipment, resulting in an insufficiently comprehensive test coverage. In addition, the existing systems mainly rely on manual experience in fault diagnosis, with a slow response speed and easy to make mistakes. The lack of data analysis capabilities also limits the effective utilization of test data and is difficult to provide strong support for equipment maintenance and strategy optimization. Therefore, the present invention provides an intelligent electromechanical equipment test management method, system, electronic device and storage medium to solve the deficiencies existing in the prior art. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides an intelligent electromechanical equipment test management method, system, electronic device and storage medium, which solves the problems of insufficient test flexibility, low level of strategy intelligence, low fault handling efficiency and lack of data analysis capabilities in the prior art.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: An intelligent electromechanical equipment test management method includes the following steps: Collect the type, model, configuration and current status information of the intelligent electromechanical equipment to form an equipment information data set; Construct a state space model of the equipment based on the equipment information data set; Utilize the state space model to train a reinforcement learning model using a deep learning algorithm to generate an optimal test strategy; Generate a test script according to the optimal test strategy and execute the test script to obtain a test result; Collect the test results, update the status information of the intelligent electromechanical device, and optimize the reinforcement learning model based on the test results to form an adaptive optimization process.

[0007] Preferably, the device information data set includes the historical test data of the device, the current operating parameters, and the environmental condition information.

[0008] Preferably, the steps of constructing the state space model of the device include: Map each device state in the device information data set to a state vector, where the state vector includes the current operating parameters and historical performance indicators of the device; Use the state vector to construct a state transition matrix, where the state transition matrix is used to describe the state change probability of the device under different operating conditions; Generate a complete state space model through the state transition matrix and the operation constraint conditions of the device to support the subsequent training of the reinforcement learning model.

[0009] Preferably, the steps of training the reinforcement learning model by applying the deep learning algorithm using the state space model include: Approximate the action value function through a deep Q-network ; Use the following update formula to adjust the parameters of the deep Q-network: ; Where is the learning rate; is the immediate reward; is the discount factor; is the next state after executing the action ; is the possible action under the state ; represents the expected cumulative reward after taking the action in the state ; represents the maximum expected return among all possible actions in the state ;

[0010] Preferably, the steps of generating the test script according to the optimal test strategy include: Parse the action sequence in the optimal test strategy and map each action to a specific test instruction; Combine the test instructions to form a complete test script.

[0011] Preferably, the execution of the test script includes the following steps: Load the test script in the test execution module and execute each test instruction step by step; Monitor the test execution process in real time, record the actual results of each test step, and compare them with the expected results; When a deviation is detected, trigger the exception handling process.

[0012] Preferably, the reinforcement learning model is optimized by updating the policy parameters of the reinforcement learning model through the policy gradient method, and the policy gradient is calculated using the following formula: ; Wherein, is the policy gradient; is the expected return of the policy; is the policy parameter; is the policy function; is the action value under the policy; is the logarithmic gradient of the policy; represents the calculation of the expected value for all possible states and actions.

[0013] There is also provided an intelligent electromechanical device test management system, including: An information collection module for collecting the type, model, configuration, and current status information of intelligent electromechanical devices; A state space construction module for constructing a state space model of the device based on the device information dataset; A reinforcement learning module for training a reinforcement learning model using the state space model and deep learning algorithms to generate an optimal test strategy; A test execution module for generating a test script according to the optimal test strategy and executing the test script to obtain test results; A data storage module for storing test results and optimized policy data.

[0014] There is also provided an electronic device, including a memory and a processor, the memory stores a computer program executable by the processor, and when the processor executes the computer program, it can implement an intelligent electromechanical device test management method.

[0015] There is also provided a storage medium storing a computer program, and when the computer program is executed by a processor, it can implement an intelligent electromechanical device test management method.

[0016] The present invention provides an intelligent electromechanical device test management method, system, electronic device, and storage medium. It has the following beneficial effects: 1. The present invention adopts a technical solution for generating dynamic test scripts. By analyzing device information and environmental conditions in real time, it achieves the technical effect of quickly adapting to the test requirements of different types of devices. Compared with the automated test system with fixed scripts in the prior art, it solves the problem of insufficient flexibility during device updates and iterations. This dynamic generation ability ensures that the system can respond promptly to device changes, reduces test delays, and improves the overall test efficiency.

[0017] 2. The present invention utilizes a technical solution that combines deep learning and reinforcement learning to generate optimal test strategies, achieving the technical effects of improving test accuracy and efficiency. Compared with traditional test methods based on preset rules, it solves the deficiency of difficultly optimizing the test process in complex device environments. Through deep Q-network and policy gradient methods, the system can adaptively adjust test strategies to ensure test coverage and precision.

[0018] 3. The present invention realizes the technical effect of quickly locating and handling device failures by introducing an AI fault diagnosis and remote repair technical solution. Compared with the method relying on manual diagnosis in the prior art, it solves the problems of long fault response time and low processing efficiency. The application of AI technology enables the system to monitor the device status in real time, automatically identify faults, and provide repair guidelines, greatly shortening the fault handling time.

[0019] 4. The present invention adopts a technical solution of a data storage and analysis module, which supports big data analysis and fault mode mining, achieving the technical effect of providing data support for device maintenance and strategy optimization. Compared with the prior art solutions lacking systematic data analysis, it solves the deficiency of low data utilization rate in long-term device management. Through in-depth analysis of historical data, the system can identify potential problems, optimize maintenance plans, and improve the reliability and lifespan of devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flowchart of the method steps of the present invention; Figure 2 is a system architecture diagram of the present invention; Figure 3 is a schematic diagram of the device of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] Please refer to the attached Figure 1, an embodiment of the present invention provides an intelligent electromechanical device test management method, including the following steps: S1. Collect the type, model, configuration, and current status information of the intelligent electromechanical device to form a device information data set; S2. Based on the device information data set, construct a state space model of the device; S3. Use the state space model to train a reinforcement learning model by applying a deep learning algorithm to generate an optimal test strategy; S4. Generate a test script according to the optimal test strategy and execute the test script to obtain test results; S5. Collect the test results, update the state information of the intelligent electromechanical device, and optimize the reinforcement learning model based on the test results to form an adaptive optimization process.

[0023] For step S1, in this embodiment, the device information module communicates with the intelligent electromechanical device through a variety of sensors and interface protocols (such as Modbus, CAN) to collect the type, model, configuration, and current status information of the device in real time. Specifically, the device information module can identify the hardware configuration, software version, operating status, and historical performance data of the device. As an option, the device information module can also integrate environmental sensors to obtain information such as the temperature and humidity of the environment where the device is located, and these data are of great significance for the performance analysis and fault prediction of the device.

[0024] In a possible implementation manner, the device information module adopts a distributed architecture and can collect information of multiple devices simultaneously. The advantage of such a design is to improve the efficiency of information collection and the scalability of the system. In some embodiments, the device information module also has a self-diagnosis function, can detect its own working status, and send an alarm signal when an abnormality occurs.

[0025] To ensure the accuracy of the data, the device information module adopts a multi-layer data verification mechanism. Generally, the data will be subjected to preliminary formatting processing after collection and then verified. Specifically, the verification process includes data integrity check, outlier detection, and data consistency verification. As an option, the device information module can be synchronized with the cloud database to ensure the real-time update and backup of the data.

[0026] During the data collection process, the device information module can also perform preliminary data analysis using machine learning algorithms. By learning historical data, the module can identify the normal operation mode of the device and send a warning signal when the data deviates from the normal mode. This design improves the intelligent level of the system and can prevent problems before they occur.

[0027] In some embodiments, the device information module also supports remote access and control. Through a secure network connection, users can remotely view the status information of the device and make necessary configuration adjustments. This function is particularly suitable for the management of devices distributed in different geographical locations.

[0028] As an extension, the device information module can also be integrated with other systems. For example, the integration with an enterprise resource planning (ERP) system can enable the sharing of device information and enterprise management data, optimizing resource allocation and production scheduling.

[0029] In terms of data transmission, the device information module adopts encryption technology to ensure the security of data during transmission. Specifically, the module uses a combination of symmetric encryption and asymmetric encryption to protect the confidentiality and integrity of the data.

[0030] In a possible implementation, the device information module also has the ability to self-learn. By continuously interacting with the device, the module can optimize its own acquisition strategy, improving the efficiency and accuracy of data acquisition.

[0031] For step S2, in this embodiment, the multi-dimensional information of the device is transformed into a state space that can be used for subsequent training of the reinforcement learning model. Generally, the accuracy and integrity of the state space model directly affect the generation of the test strategy and the accuracy of fault diagnosis. Therefore, it is crucial to construct a comprehensive and accurate state space model.

[0032] The state space model of the device is constructed by mapping each device state in the device information dataset to a state vector. Specifically, the state vector includes the current operating parameters of the device, historical performance indicators, and environmental condition information. As an option, the state vector can also include the device's fault history and maintenance records to more comprehensively describe the state of the device.

[0033] In a possible implementation, the state space model describes the probability of state change of the device under different operating conditions by constructing a state transition matrix. The state transition matrix represents the probability of transitioning from state to state after taking action a. Here and represent the current state and the next state of the device respectively represents the operating action of the device. Generally, the construction of the state transition matrix needs to be combined with the operating constraint conditions and historical data of the device.

[0034] To improve the accuracy of the state space model, a multi-level state description method is adopted in this embodiment. Specifically, the state space model not only considers the physical state of the device, but also includes the logical state and environmental state of the device. The physical state describes the hardware configuration and operating parameters of the device, the logical state describes the control logic and software state of the device, and the environmental state describes the external environmental conditions where the device is located.

[0035] In some embodiments, the state space model can also be optimized through machine learning algorithms. By learning from historical data, the model can identify the normal operating mode of the device and make adjustments when the state deviates from the normal mode. As an extension, the state space model can also be integrated with other systems. For example, the integration with a predictive maintenance system can achieve real-time monitoring and prediction of the device state.

[0036] In terms of data processing, the state space model adopts efficient data structures and algorithms to ensure the computational efficiency and storage efficiency of the model. Specifically, the model uses a sparse matrix to represent the state transition matrix to reduce storage space and computational complexity. As an option, the model can also adopt a distributed computing architecture to support the state modeling of large-scale devices.

[0037] In a possible implementation, the state space model also has an adaptive ability. By continuously interacting with the device, the model can dynamically adjust its own parameters to adapt to the changes and updates of the device. This design improves the flexibility and adaptability of the model and can maintain the effectiveness of the model when the device is updated iteratively.

[0038] For step S3, in this embodiment, a reinforcement learning model is trained through a deep learning algorithm to generate an optimal test strategy. The successful implementation of this step depends on the accurate acquisition of previous device information and the precise construction of the state space model. Generally, the training quality of the reinforcement learning model directly affects the effectiveness of the test strategy and the overall performance of the system.

[0039] In this embodiment, the training of the reinforcement learning model uses a deep Q-network (DQN) to approximate the action value function . Specifically, the state represents the current state of the device, and the action represents the test operations that can be taken in this state. The DQN updates its parameters by continuously interacting with the environment to approximate the optimal action value function. As an option, the update process of the DQN uses the following formula: ; where, is the learning rate; is the immediate reward; is the discount factor; Is the next state after performing the action ; Is a possible action in state ; Represents the expected cumulative reward after taking action a in state s; Represents in state All possible actions The maximum expected return among them.

[0040] In a possible implementation, the training process of the reinforcement learning model also incorporates the experience replay technique. By storing and reusing past experiences, the model can update its parameters more stably and avoid training instability caused by data correlation. In some embodiments, the size and update frequency of the experience replay buffer are adjustable to adapt to different device and test requirements.

[0041] To improve the training efficiency of the model, a target network is introduced in this embodiment. The parameters of the target network are copied from the main network every certain number of steps to provide a more stable target value. Specifically, the update formula for the target network is: ; Where Are the parameters of the main network; Are the parameters of the target network; Is the soft update coefficient.

[0042] In some embodiments, the design of the reward function of the reinforcement learning model takes into account multiple factors, including test time, resource consumption, coverage, and accuracy. Generally, the form of the reward function is: Where Represents the comprehensive reward obtained by taking action a in state s; Represents the scope or degree of test coverage after taking action a in state s; Represents in state Taking action The required time; Represents in state The amount of resources consumed by taking action a; Represents the accuracy of the test result after taking action a in state s; , , , Are the weight coefficients indicating the importance of each factor.

[0043] The reinforcement learning model can also combine the policy gradient method to optimize the expected return of the policy. Through the policy gradient, the model can directly optimize the policy parameters , improve the flexibility and adaptability of the strategy.

[0044] In practical applications, the training process of the reinforcement learning model is an iterative process. Through continuous interaction with the environment, the model can gradually approach the optimal strategy, and the generated test strategy can effectively guide the test execution module to perform tests.

[0045] For step S4, in this embodiment, it is responsible for generating test scripts according to the optimal test strategy and executing these scripts to obtain test results. The successful implementation of this step depends on the optimal strategy generated by the previous reinforcement learning model. Generally, the quality of test script generation and execution directly affects the effectiveness of testing and the overall performance of the system.

[0046] In this embodiment, the test script generation module parses the action sequence in the optimal test strategy and maps each action to a specific test instruction. Specifically, the optimal test strategy consists of a series of actions, and each action corresponds to a specific operation or parameter setting of the device. As an option, the test script generation module converts these action sequences into an executable set of test instructions.

[0047] In a possible implementation, the test script generation module combines the specific requirements of the device (such as new function testing, special configuration item verification) to refine and adjust the test script. In this way, the generated test script can comprehensively cover the device characteristics while maintaining high test efficiency.

[0048] In some embodiments, the test execution module is responsible for loading the generated test scripts and executing each test instruction step by step. Generally, the test execution module monitors the test process in real time, records the actual results of each step, and compares them with the expected results. If a deviation is detected, an exception handling process is triggered.

[0049] To improve the accuracy and efficiency of testing, a policy gradient method is introduced in this embodiment to optimize the execution of the test strategy. Specifically, the policy gradient method calculates the policy gradient through the following formula: ; where is the policy gradient; is the expected return of the policy; are the policy parameters; is the policy function; is the action value under the policy; is the logarithmic gradient of the policy; represents calculating the latest state and configuration of the device for the expected value of all possible states and actions, ensuring that the execution of the test script always conforms to the actual situation of the device.

[0050] In practical applications, the generation and execution of test scripts is an iterative process. Through continuous interaction with the device, the system can gradually optimize the test strategy, and the generated test scripts can effectively guide the test execution module to conduct tests.

[0051] For step S5, in this embodiment, it is responsible for collecting test results, updating the status information of intelligent electromechanical devices, and optimizing the reinforcement learning model based on the test results to form an adaptive optimization process. The successful implementation of this step depends on the execution of the previous test scripts and the accurate collection of results. Generally, the quality of the analysis and feedback of test results directly affects the adaptive ability and overall performance of the system.

[0052] In this embodiment, the collection and analysis of test results is a dynamic process. Specifically, during the execution of test scripts by the test execution module, the results of each test step are recorded in real time. As an option, the test results include data such as the performance metrics of the device, fault information, and environmental conditions. These data are transmitted to the data storage and analysis module for further analysis and processing.

[0053] In a possible implementation, the analysis of test results combines machine learning and data mining techniques. By comparing historical data with current test results, the system can identify the performance trends and potential problems of the device. In some embodiments, the data analysis module uses clustering algorithms to classify test results for a better understanding of the state changes of the device.

[0054] To optimize the reinforcement learning model, a feedback mechanism based on test results is introduced in this embodiment. Specifically, the system adjusts the parameters of the reinforcement learning model according to the test results to improve the prediction accuracy and adaptability of the model. Generally, the optimization process of the model uses the following formula: ; where, are the parameters of the model; is the learning rate, which controls the step size of parameter update; is the policy gradient, which represents the optimization direction of the model.

[0055] In some embodiments, the system also combines an adaptive learning rate adjustment mechanism. By monitoring the convergence speed and performance changes of the model, the system can dynamically adjust the learning rate to accelerate the optimization process of the model. As an extension, the system also combines genetic algorithms or particle optimization evolution algorithms to further improve the optimization effect of the model.

[0056] In practical applications, the collection and analysis of monitoring results is an iterative process. Through continuous interaction with the device, the system can gradually optimize the reinforcement learning model, and the generated test strategies can effectively guide the test execution module to conduct tests.

[0057] In this embodiment, it is responsible for storing the test results and the optimized policy data to provide a basis for long-term optimization and policy adjustment. The successful implementation of this step depends on the accurate collection of the previous test results and the effectiveness of model optimization. Generally, the quality of data storage and management directly affects the long-term performance and adaptability of the system.

[0058] In this embodiment, the data storage and analysis module is responsible for systematically storing the test results and the optimized policy data. Specifically, the test results include data such as the performance indicators, fault information, and environmental conditions of the device. These data are stored in a structured database for subsequent query and analysis. As an option, the database can use a relational database or a NoSQL database to adapt to different data types and query requirements.

[0059] In a possible implementation, the data storage module also incorporates data compression and deduplication techniques. By compressing and deduplicating redundant data, the system can effectively reduce the occupied storage space. In some embodiments, the data compression algorithm uses lossless compression technology to ensure the integrity and accuracy of the data.

[0060] To support long-term optimization and policy adjustment, a policy optimization mechanism based on data analysis is introduced in this embodiment. Specifically, the system identifies the performance trends and potential problems of the device by analyzing historical data. Generally, the data analysis module uses the following formula for trend analysis: ; where is the trend value at time ; is the number of data points; is the th data point at time

[0061] In some embodiments, the system also incorporates a prediction model to predict the future performance and fault risk of the device. By performing regression analysis on historical data, the system can generate a performance prediction model for the device. As an extension, the prediction model can incorporate machine learning algorithms to improve the accuracy and reliability of the prediction.

[0062] In practical applications, data storage and analysis is an iterative process. By continuously analyzing the data, the system can gradually optimize the test strategy, and the generated strategy can effectively guide the test execution module to conduct tests.

[0063] The intelligent electromechanical device test management system described below can be correspondingly referred to the intelligent electromechanical device test management method described above.

[0064] Please refer to the appendix Figure 2 , the present invention also provides an intelligent electromechanical device test management system, including: An information acquisition module, used to collect the type, model, configuration, and current status information of intelligent electromechanical devices; A state space construction module, used to construct a state space model of the device based on the device information dataset; A reinforcement learning module, used to train a reinforcement learning model using the state space model and deep learning algorithms to generate an optimal test strategy; A test execution module, used to generate a test script according to the optimal test strategy and execute the test script to obtain test results; A data storage module, used to store test results and optimized policy data.

[0065] For the information acquisition module, this module communicates with the device through a variety of sensors and interface protocols (such as Modbus, CAN, Ethernet). The collected information includes the type, model, configuration parameters of the device, and the current operating status. These data are updated in real time and stored in the device information dataset, providing a basis for subsequent state space construction and policy generation.

[0066] In some embodiments, the information acquisition module can also obtain the historical operation data and environmental condition information of the device. These data help to more comprehensively understand the operating conditions of the device and external influencing factors.

[0067] For the state space construction module, this module maps each state of the device to a state vector, describing the behavior of the device under different conditions. The state space model is the basis of reinforcement learning, providing a structured representation of all possible states of the device.

[0068] As an option, the state space construction module can combine the operation constraint conditions and historical performance indicators of the device to generate a more accurate state space model. This model can better support the subsequent training of the reinforcement learning model.

[0069] For the reinforcement learning module, through interaction with the environment, the module generates an optimal test strategy. Generally, the module uses a deep Q-network (DQN) or policy gradient method for training to improve the accuracy and adaptability of the strategy.

[0070] In some embodiments, the reinforcement learning module combines experience replay and target network techniques to improve the stability and efficiency of training. The module can also dynamically adjust the learning rate and other hyperparameters to adapt to different devices and test requirements.

[0071] For the test execution module, it is responsible for loading and parsing test scripts, and executing each test instruction step by step. It monitors the test process in real time, records the results of each step, and compares them with the expected results.

[0072] The test execution module can combine remote monitoring and control technologies to ensure that the execution of test scripts always conforms to the actual situation of the device. The module can also dynamically adjust the execution order and parameter settings of test scripts to adapt to changes in the device and fluctuations in the environment.

[0073] For the data storage module, it supports big data analysis and data mining technologies to discover fault patterns and performance trends. The stored data provides a basis for subsequent test strategy optimization and device maintenance.

[0074] In some embodiments, the data storage module also supports data encryption and access control to ensure data security. The module can be integrated with cloud storage services to achieve remote backup and sharing of data.

[0075] The system of this embodiment can be used to execute the above method embodiments, and their principles and technical effects are similar, so they will not be elaborated here.

[0076] Please refer to the attached Figure 3 , a computer device described below can be correspondingly referred to a smart electromechanical device test management method described above.

[0077] The present invention also provides an electronic device, including: a processor and a memory, the memory stores a computer program executable by the processor, and when the computer program is executed by the processor, it executes the above method.

[0078] The present invention also provides a storage medium, on which a computer program is stored, and when the computer program is run by a processor, it executes the above method.

[0079] Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (abbreviated as SRAM), electrically erasable programmable read-only memory (abbreviated as EEPROM), erasable programmable read-only memory (abbreviated as EPROM), programmable read-only memory (abbreviated as PROM), read-only memory (abbreviated as ROM), magnetic memory, flash memory, magnetic disk or optical disc.

[0080] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent electromechanical equipment test management method, characterized in that, It includes the following steps: Collect the type, model, configuration, and current status information of intelligent electromechanical devices to form a device information dataset; Based on the device information dataset, construct a state space model of the device; Utilize the state space model, apply a deep learning algorithm to train a reinforcement learning model, and generate an optimal test strategy; Generate a test script according to the optimal test strategy, and execute the test script to obtain test results; Collect the test results, update the status information of intelligent electromechanical devices, and optimize the reinforcement learning model based on the test results to form an adaptive optimization process.

2. The intelligent electromechanical equipment test management method according to claim 1, characterized in that, The device information dataset includes the historical test data of the device, current operating parameters, and environmental condition information.

3. The intelligent electromechanical device test management method according to claim 1, characterized in that The step of constructing the state space model of the device includes: Map each device state in the device information dataset to a state vector, where the state vector includes the current operating parameters and historical performance indicators of the device; Utilize the state vector to construct a state transition matrix, and the state transition matrix is used to describe the state change probability of the device under different operating conditions; Generate a complete state space model through the state transition matrix and the operating constraint conditions of the device to support subsequent training of the reinforcement learning model.

4. The intelligent electromechanical device test management method according to claim 1, wherein The step of utilizing the state space model and applying a deep learning algorithm to train the reinforcement learning model includes: Approximating the action-value function with a deep Q-network ; Adjust the parameters of the deep Q-network using the following update formula: ； Among them, is the learning rate; is the immediate reward; is the discount factor; is the next state after executing the action ; is the possible action in the state ; represents the expected cumulative reward after taking the action in the state ; represents the maximum expected return among all possible actions in the state .

5. A method for testing and managing an intelligent electromechanical device according to claim 1, characterized in that, The step of generating a test script according to the optimal test strategy includes: Parse the action sequence in the optimal test strategy, and map each action to a specific test instruction; Combine the test instructions to form a complete test script.

6. The intelligent electromechanical equipment test management method according to claim 5, characterized in that The execution of the test script includes the following steps: Load the test script in the test execution module, and execute each test instruction step by step; Monitor the test execution process in real time, record the actual results of each test step, and compare them with the expected results; When a deviation is detected, trigger an exception handling process.

7. A method for testing and managing an intelligent electromechanical device according to claim 1, characterized in that, The optimization of the reinforcement learning model updates the policy parameters of the reinforcement learning model through a policy gradient method, and calculates the policy gradient using the following formula: ； Among them, is the policy gradient; is the expected return of the policy; is the policy parameter; is the policy function; is the action value under the policy; is the logarithmic gradient of the policy; represents the calculation of the expected value for all possible states and actions.

8. An intelligent electromechanical device test management system, which is applied to an intelligent electromechanical device test management method according to any one of claims 1-7, characterized in that, It includes: An information collection module for collecting the type, model, configuration, and current status information of intelligent electromechanical devices; A state space construction module for constructing a state space model of the device based on the device information dataset; A reinforcement learning module for training a reinforcement learning model using the state space model and a deep learning algorithm to generate an optimal test strategy; A test execution module for generating a test script according to the optimal test strategy and executing the test script to obtain test results; A data storage module for storing test results and optimized policy data.

9. An electronic device, characterized in that, It includes a memory and a processor. The memory stores a computer program executable by the processor. When the processor executes the computer program, it implements an intelligent electromechanical device test management method as described in any one of claims 1-7.

10. A storage medium, characterized in that, A computer program is stored. When the computer program is executed by the processor, it implements an intelligent electromechanical device test management method as described in any one of claims 1-7.

Images