5G power network consistency protocol test method

Abstract

The invention provides a 5G power network consistency protocol testing method, and belongs to the technical field of power network testing. The method comprises the following steps: S1, analyzing a 5G power network protocol specification, identifying a static consistency demand and a dynamic consistency demand, and generating an abstract test set; s2, calling the abstract test set through a test executor, and performing multi-dimensional test on the tested system through a distributed test architecture; s3, collecting input and output sequences and network performance data in a multi-dimensional test process in a mode of combining active detection and passive monitoring, and forming a test result; and S4, analyzing a test result based on the formalized model, performing consistency judgment by adopting a hierarchical judgment mechanism, forming consistency grade evaluation, and generating a test report containing the test result and the consistency grade evaluation. According to the invention, the coverage comprehensiveness and the test efficiency of the 5G power network consistency protocol can be improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power network testing, and particularly relates to a 5G power network consistency protocol test method. BACKGROUND

[0002] With the rapid development of smart grids and energy internet, 5G communication technology has become a key infrastructure supporting the digital transformation and upgrading of power systems. 5G networks, with their low latency, high reliability and massive connectivity characteristics, provide communication support for core power businesses such as distribution automation, precise load control and distributed energy regulation.

[0003] In the power 5G converged networking environment, protocol consistency testing is the basis for ensuring efficient collaboration between power businesses and communication networks, and is directly related to the safety and stability of the power grid. Traditional protocol testing methods are mainly designed for general network environments and lack specific adaptation to power business characteristics, resulting in incomplete test coverage and low efficiency, making it difficult to verify the compliance and reliability of 5G power protocols in real industrial environments. SUMMARY

[0004] The technical problem to be solved by the application is how to improve the comprehensive coverage and test efficiency of 5G power network consistency protocols. In view of the shortcomings of the prior art, a 5G power network consistency protocol test method is provided.

[0005] To solve the above technical problems, the technical solution adopted by the application is: The application provides a 5G power network consistency protocol test method, comprising: S1, parsing the 5G power network protocol specification, identifying static consistency requirements and dynamic consistency requirements, and generating an abstract test set; S2, calling the abstract test set through a test executor, and performing multi-dimensional testing on the tested system through a distributed testing architecture; S3, collecting input-output sequences and network performance data during the multi-dimensional testing through a combination of active probing and passive listening, and forming a test result; S4, analyzing the test result based on a formal model, using a hierarchical decision mechanism for consistency determination, forming a consistency level evaluation, and generating a test report containing the test result and the consistency level evaluation.

[0006] Compared to existing technologies, the beneficial effects of this invention include: First, by parsing the 5G power network protocol specification, static consistency requirements can be identified to obtain the minimum capability constraining network interconnection, and dynamic consistency requirements can also be identified to obtain observable behaviors through external communication. Furthermore, an abstract test set containing test cases, test sequences, and expected results can be generated. This setup allows for the parsing of protocol text using natural language processing technology to extract key protocol elements and constraints, thereby improving the coverage and efficiency issues caused by traditional test case design relying on manual design. This achieves the effect of automatically generating a high-coverage test set, improving the comprehensiveness of 5G power network consistency protocol coverage and testing efficiency. Next, by calling the abstract test set through a test executor, multi-dimensional testing is performed using a distributed testing architecture. Protocol entity behavior is simulated through test proxies, and connections are maintained through a heartbeat mechanism. The timeout retransmission mechanism ensures reliability, thereby resolving the single-point bottleneck problem of centralized testing architectures and achieving high-performance concurrent testing and flexible scalability. Then, through a combination of active probing and passive listening, input / output sequences and network performance data are collected. This setup uses controlled traffic injection through active probing to measure path performance and passive listening to capture real business traffic. Timestamp synchronization technology is used to align multi-source data, thus addressing the limitations of single data sources and achieving a comprehensive understanding of network status. Finally, test results are analyzed based on formal models such as extended finite state machines, and a layered decision mechanism involving the physical layer, protocol layer, and business layer is used for consistency determination. This generates a test report containing detailed results and consistency level assessments, effectively addressing the lack of in-depth analysis and comprehensive evaluation in traditional testing, achieving a multi-dimensional and accurate assessment of protocol compliance.

[0007] Optionally, S1 includes: S11. Parse the text of the 5G power network protocol specification using natural language processing technology, extract key protocol elements and constraints, and identify the static consistency requirements and the dynamic consistency requirements. S12. Based on the key protocol elements, the constraints, the static consistency requirements, and the dynamic consistency requirements, design test cases using protocol state machines and message sequences. The test cases cover normal and abnormal scenarios. S13. Organize the test cases using XML format to form the abstract test set, which includes the test cases, test sequences, and expected results.

[0008] Optionally, the multi-dimensional testing of the system under test using a distributed testing architecture in step S2 includes: S21. Deploy the test execution engine on the test control node to coordinate multiple test agents to be executed, forming the distributed test architecture. S22. Simulate the behavior of the protocol entity through the test agent to send and receive protocol messages; S23. Based on the heartbeat mechanism and timeout retransmission mechanism, the multi-dimensional test is performed on the system under test through the distributed test architecture.

[0009] Optionally, S3 includes: S31. Measure network performance indicators by sending specially designed test data packets through active probing methods; S32. Capture network traffic and record protocol interaction processes using passive monitoring methods; S33. Align the data from multiple collection points of the distributed test architecture using timestamp synchronization technology; S34. Based on the network performance indicators and the protocol interaction process, the test results are generated.

[0010] Optionally, the active detection method in S31 includes: S311. Send a sequence of probe data packets with specific characteristics, measure the transmission delay and loss of each data packet, and form a delay change trend; S312. Based on the aforementioned latency change trend, the statistical measure of the latency increase trend and the normalized measure of the latency change degree are measured using the following formula, and the network performance index is formed: , , Wherein, the S PCT S is a statistical measure of the increasing trend of time delay. PDT As a normalized measure of the degree of time delay variation, the B k Here, H is the latency measurement value of the kth data packet, H is the probe window value of the Pathload algorithm, and I(·) is an indicator function, which takes the value of 1 when the condition of the indicator function is met and takes the value of 0 when the condition is not met.

[0011] Optionally, the active detection method in S31 further includes: S313. Adjust the sending rate of the data packets according to the network performance indicators, and converge the sending rate of the probe data packet sequence using a binary search algorithm; S314. When the difference in the transmission rate of three adjacent data packets in the probe data packet sequence is less than a preset threshold, the transmission of the probe data packet sequence is stopped.

[0012] Optionally, the measurement of transmission delay and loss for each data packet in S311 includes: S3111. Simulate real business traffic patterns and generate test traffic through protocol simulation and the probe data packet sequence; S3112. Calculate the performance indicators of the data packet based on the measurement data from the receiving end and the following formula, and obtain the transmission delay and the data loss situation: , , Among them, the The number of data packets successfully received by the receiving end. The maximum sequence number value in the received data packet, The starting sequence number of the receiving window is the reference value, and J is the dynamic estimate of the delay jitter. This is the measured delay difference between adjacent data packets.

[0013] Optionally, the formal model in S4 analyzes the test results including: S41. Establish a protocol model based on an extended finite state machine; S42. Use model checking technology to verify protocol consistency; S43. Verify the properties of complex protocols using theorem proof methods.

[0014] Optionally, the formal model in S4 further analyzes the test results by including: S43. Train the anomaly detection model using historical test data; S44. Extract the feature vector of the protocol message sequence; S45. Based on the feature vector, identify abnormal patterns using a deep learning algorithm.

[0015] Optionally, the adoption of a hierarchical decision mechanism in step S4 to determine consistency and form a consistency level assessment includes: S46. Verify signal quality and connection stability at the physical layer; S47. Verify message format and timing compliance at the protocol layer; S48. Verify service quality and performance metrics at the business layer; S49. A weighted scoring mechanism is used to comprehensively evaluate the consistency level. Attached Figure Description

[0016] The present invention will now be described in further detail with reference to the accompanying drawings.

[0017] Figure 1 : A flowchart illustrating the 5G power network consistency protocol testing method in this embodiment of the invention. Detailed Implementation

[0018] To better understand the present invention, the following embodiments further illustrate the content of the invention, but the scope of protection of the present invention is not limited to the following embodiments. Numerous specific details are set forth in the following description to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the present invention can be practiced without one or more of these details.

[0019] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0020] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0021] An embodiment of the present invention provides a 5G power network conformance protocol testing method, comprising: S1, parsing the 5G power network protocol specification, identifying static and dynamic conformance requirements, and generating an abstract test set; S2, calling the abstract test set through a test executor, and performing multi-dimensional testing on the system under test through a distributed test architecture; S3, collecting input-output sequences and network performance data during the multi-dimensional testing process through a combination of active probing and passive listening, and forming test results; S4, analyzing the test results based on a formal model, using a hierarchical decision mechanism to determine conformance, forming a conformance level assessment, and generating a test report containing the test results and the conformance level assessment.

[0022] In this embodiment, firstly, as... Figure 1As shown in S1, by parsing the 5G power network protocol specification, static consistency requirements can be identified to obtain the minimum capability constraining network interconnection. Dynamic consistency requirements can also be identified, thereby obtaining observable behaviors through external communication. Furthermore, an abstract test set containing test cases, test sequences, and expected results can be generated. This setup allows for the parsing of protocol text using natural language processing technology to extract key protocol elements and constraints, thus improving the coverage and efficiency issues caused by traditional test case design relying on manual design. This achieves the effect of automatically generating high-coverage test sets, improving the comprehensiveness of 5G power network consistency protocol coverage and testing efficiency. Next, as... Figure 1 As shown in S2, the abstract test set is invoked through the test executor, and multi-dimensional tests are executed using a distributed test architecture. The behavior of protocol entities is simulated through a test proxy, the connection is maintained through a heartbeat mechanism, and reliability is ensured through a timeout retransmission mechanism. This solves the single-point-of-failure bottleneck problem of centralized test architectures, achieving high-performance concurrent testing and flexible scalability. Then, as... Figure 1 As shown in S3, by combining active probing and passive listening, input / output sequences and network performance data are collected. This setup allows for controlled traffic injection through active probing to measure path performance, while passive listening captures real business traffic. Timestamp synchronization technology is used to align multi-source data, thus addressing the limitations of single data sources and achieving a comprehensive understanding of network status. Finally, as... Figure 1 As shown in S4, the test results are analyzed based on formal models such as extended finite state machines. A layered decision mechanism of physical layer, protocol layer and business layer is used to determine consistency, thereby generating a test report containing detailed results and consistency level assessment. This effectively solves the problem of traditional testing lacking in-depth analysis and comprehensive evaluation, and achieves the effect of multi-dimensional and accurate evaluation of protocol compliance.

[0023] Optionally, S1 includes: S11, parsing the text of the 5G power network protocol specification using natural language processing technology, extracting key protocol elements and constraints, and identifying static consistency requirements and dynamic consistency requirements; S12, designing test cases based on key protocol elements, constraints, static consistency requirements, and dynamic consistency requirements using protocol state machines and message sequences, with test cases covering normal and abnormal scenarios; S13, organizing test cases using XML format to form an abstract test set, which includes test cases, test sequences, and expected results.

[0024] In this optional embodiment, the process of generating the abstract test set first involves parsing the protocol text using NLP techniques such as dependency parsing and entity recognition to extract key elements such as message type and parameter range, as well as constraints such as latency ≤10ms. This effectively solves the problems of low efficiency and easy omissions in manual parsing, achieving the effect of automated and accurate extraction of requirements. Next, based on the extracted content, test cases can be designed using protocol state machines and message sequences. The protocol state machine describes the state transition logic, and the message sequence defines the interaction process, covering normal scenarios such as successful connection and abnormal scenarios such as packet loss and retransmission. This solves the problem of single test scenarios and achieves the effect of comprehensive protocol verification. Finally, test cases are organized in XML format to describe test steps and expected outputs in a tagged way, thereby solving the problems of test set readability and interoperability, achieving the effect of machine parsing and easy maintenance.

[0025] Optionally, S2 performs multi-dimensional testing on the system under test using a distributed testing architecture, including: S21, deploying a test execution engine on the test control node to coordinate multiple test agents to be executed, forming a distributed testing architecture; S22, simulating protocol entity behavior through test agents to send and receive protocol messages; and S23, performing multi-dimensional testing on the system under test using a distributed testing architecture based on a heartbeat mechanism and a timeout retransmission mechanism.

[0026] In this optional embodiment, for the multi-dimensional testing process, firstly, an execution engine such as a Kubernetes-based scheduler is deployed on the test control node to coordinate test agents such as containerized instances, thereby solving the complex resource management problem and achieving elastic scaling. Next, the test agent simulates protocol entities, such as the gNodeB simulator sending RRC messages, which can solve the problem of high deployment costs of real devices and achieve low-cost, high-fidelity simulation effects. Finally, through heartbeat mechanisms such as periodically sending liveness signals and timeout retransmissions such as retransmitting test commands when packets are lost, the problem of test interruption caused by network fluctuations can be solved, achieving the effect of reliable execution of long-term tests.

[0027] Optionally, S3 includes: S31, sending specially designed test data packets through an active probing method to measure network performance indicators; S32, capturing network traffic and recording the protocol interaction process through a passive listening method; S33, aligning the data of multiple collection points in the distributed test architecture through timestamp synchronization technology; and S34, forming test results based on network performance indicators and protocol interaction processes.

[0028] In this optional embodiment, the test result generation process first involves actively sending specialized test packets, such as UDP packets with precise timestamps, to measure metrics like latency and packet loss rate. This addresses the issue of passive measurement failing to actively stimulate the network, achieving accurate acquisition of performance parameters. Next, passively monitoring network traffic, such as capturing data through port mirroring and recording protocol interactions like HTTP request-response sequences, solves the problem of actively detecting interference with services, achieving non-intrusive monitoring. Then, using the PTP protocol to synchronize clocks at multiple sampling points and align data timestamps resolves the difficulty of correlating asynchronous data, achieving multi-source data fusion analysis. Finally, combining network metrics and protocol interaction data to generate test results enables data interaction and verification, effectively solving the data silo problem and achieving comprehensive evaluation.

[0029] Optionally, the active probing method in S31 includes: S311, sending a sequence of probe data packets with specific characteristics, measuring the transmission delay and loss of each data packet, and forming a delay change trend; S312, based on the delay change trend, measuring the statistics of the delay increase trend and the normalized measure of the delay change degree using the following formula, and forming a network performance index: (1.1), (1.2), Among them, S PCT S is a statistical measure of the increasing trend of time delay. PDT B is a normalized measure of the degree of time delay variation. k H is the latency measurement value of the k-th data packet, H is the probe window value of the Pathload algorithm, and I(·) is the indicator function, which takes the value of 1 when the indicator function condition is true and 0 when the condition is false.

[0030] In this optional embodiment, during the process of actively testing and measuring network performance indicators, firstly, by sending a sequence of probe packets, such as 50 packets each with a length of 1500 bytes, the transmission delay and loss of each packet are measured. This effectively solves the problem of slow convergence in traditional bandwidth measurement and improves testing efficiency. Next, the statistics of the delay increase trend are calculated using equation (1.1), where H is obtained by taking the square root of k. Thus, the statistics of the delay increase trend can quantify the delay increase ratio. When S... PCT A value greater than 0.66 indicates congestion, while S is calculated using equation (1.2). PDT Quantifiable cumulative delay variation, S PDT A value >0.45 indicates a significant trend, thus through S PCT and S PDT The measurement of network performance indicators can solve the problem of misjudgment caused by network jitter and improve the accuracy of bandwidth measurement.

[0031] Optionally, the active probing method in S31 further includes: S313, adjusting the data packet sending rate according to network performance indicators, and converging the sending rate of the probe data packet sequence through a binary search algorithm; S314, stopping the sending of the probe data packet sequence when the difference in the sending rates of three adjacent data packets in the probe data packet sequence is less than a preset threshold.

[0032] In this optional embodiment, for the process of measuring network performance indicators through active testing, firstly, it can also be based on S PCT and S PDT As a result, the sending rate is dynamically adjusted through a binary search algorithm, such as gradually converging from 1Gbps, effectively solving the problem that a fixed rate is not adapted to network changes and achieving the effect of quickly converging to the actual bandwidth. Furthermore, the probe can be stopped when the difference between three adjacent rates is less than a threshold, which can be 0.1Mbps, thereby solving the problem of excessive probes wasting resources and achieving the effect of efficient measurement.

[0033] Optionally, S311 measures the transmission delay and loss of each data packet, including: S3111, simulating real service traffic patterns by generating test traffic through protocol simulation and probing data packet sequences; S3112, calculating the data packet performance metrics based on the receiver's measurement data and the following formula to obtain the transmission delay and loss: (2.1), (2.2), in, The number of data packets successfully received by the receiving end. The maximum sequence number value in the received data packet. Here, J is the baseline value for the starting sequence number of the receiving window, and J is the dynamic estimate of the delay jitter. This is the measured delay difference between adjacent data packets.

[0034] In this optional embodiment, for the specific process of measuring the transmission delay and loss of each data packet, firstly, simulate real business traffic, such as the burst traffic mode of power distribution automation. Test traffic is generated by simulating protocols such as Modbus / TCP over 5G, which can solve the problem of abstract traffic not being representative, thereby achieving an effect close to actual business. Next, performance indicators can be calculated based on the data received by the receiving end through equations (2.1) and (2.2). For equation (2.1), by setting the actual number of received packets, the maximum sequence number, and the initial sequence number, sequence number wrapping is prevented, and transmission reliability can be accurately quantified. For equation (2.2), D(i-1, i) = (R i -R {i-1} )-(S i -S{i-1} R is the receive timestamp and S is the send timestamp. The instantaneous jitter can be smoothed by exponential weighted averaging, which solves the problem of jitter sensitivity and achieves the effect of stable evaluation of network quality.

[0035] Optionally, the formal model in S4 analyzes the test results, including: S41, establishing a protocol model based on an extended finite state machine; S42, verifying protocol consistency using model checking techniques; and S43, verifying complex protocol properties using theorem proof methods.

[0036] In this optional embodiment, during the analysis of test results using a formal model, firstly, an Extended Finite State Machine (EFSM) model is established, defined as a six-tuple (I, O, S, S0, X, T), where X is a set of variables (e.g., a counter), and T is a transition with a conditional predicate P and an assignment operation A. For example, if the count > 10, a reset is performed, effectively solving the problem that simple state machines cannot handle data dependencies and achieving the effect of accurately modeling protocol logic. Next, model checking tools, such as SPIN, are used to verify protocol consistency, such as the absence of deadlock and satisfaction of latency constraints, thereby solving the problem of low efficiency in manual verification. Finally, theorem proofs, such as those using Coq, are used to verify complex properties such as the correctness of encryption algorithms, which can solve the problem of insufficient rigor in empirical verification and achieve the effect of mathematical rigor guarantee.

[0037] Optionally, the formal model in S4 for analyzing the test results also includes: S43, training an anomaly detection model using historical test data; S44, extracting feature vectors from the protocol message sequence; and S45, identifying anomaly patterns based on the feature vectors using a deep learning algorithm.

[0038] In this optional embodiment, for the process of analyzing test results using a formal model, firstly, a deep learning model such as an LSTM network model is trained using historical test data such as normal / abnormal protocol message sequences to solve the problem that the rule system is difficult to adapt to new anomalies; then, message sequence feature vectors, such as message type distribution and time interval variance, are extracted to solve the problem of high dimensionality of the original data; finally, anomaly patterns such as DDoS attack characteristics can be identified through deep learning algorithms such as convolutional neural networks to achieve a high-accuracy anomaly detection effect.

[0039] Optionally, S4 employs a layered decision-making mechanism to determine consistency, forming a consistency level assessment that includes: S46, verifying signal quality and connection stability at the physical layer; S47, verifying message format and timing compliance at the protocol layer; S48, verifying service quality and performance indicators at the service layer; and S49, comprehensively assessing the consistency level using a weighted scoring mechanism.

[0040] In this optional embodiment, during the process of using a hierarchical decision-making mechanism to determine consistency and form a consistency level assessment, firstly, the physical layer verifies signal quality, such as RSRP > -100dBm, and also verifies connection stability, such as handover success rate > 99%, ensuring upper-layer stability and resolving the issue of lower-layer defects affecting upper layers. Next, the protocol layer verifies message format, such as the correctness of ASN.1 encoding, and timing compliance. For example, the T351 timer timeout behavior can resolve protocol logic errors. Then, the service layer verifies service quality, such as video transmission bitrate ≥ 2Mbps, and performance indicators, such as latency ≤ 20ms, resolving poor service experience issues. Finally, a weighted scoring mechanism is used, such as a physical layer weight of 0.2 and a service layer weight of 0.5, to comprehensively evaluate the consistency level, resolving the issue of the one-sidedness of a single indicator, achieving a scientific and quantitative evaluation effect, and improving the comprehensiveness and stability of the detection.

[0041] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A test method for 5G power network consensus protocols, characterized in that, include: S1. Analyze the 5G power network protocol specification, identify static consistency requirements and dynamic consistency requirements, and generate an abstract test set; S2. The abstract test set is invoked through the test executor, and the system under test is tested in multiple dimensions through a distributed test architecture. S3. By combining active detection and passive listening, the input-output sequences and network performance data during the multi-dimensional test are collected to form test results; S4. Analyze the test results based on the formal model, use a hierarchical decision mechanism to determine consistency, form a consistency level assessment, and generate a test report containing the test results and the consistency level assessment.

2. The 5G power network conformance protocol testing method as described in claim 1, characterized in that, S1 includes: S11. Parse the text of the 5G power network protocol specification using natural language processing technology, extract key protocol elements and constraints, and identify the static consistency requirements and the dynamic consistency requirements. S12. Based on the key protocol elements, the constraints, the static consistency requirements, and the dynamic consistency requirements, design test cases using protocol state machines and message sequences. The test cases cover normal and abnormal scenarios. S13. Organize the test cases using XML format to form the abstract test set, which includes the test cases, test sequences, and expected results.

3. The 5G power network conformance protocol testing method as described in claim 1, characterized in that, The multi-dimensional testing of the system under test using a distributed testing architecture, as described in S2, includes: S21. Deploy the test execution engine on the test control node to coordinate multiple test agents to be executed, forming the distributed test architecture. S22. Simulate the behavior of the protocol entity through the test agent to send and receive protocol messages; S23. Based on the heartbeat mechanism and timeout retransmission mechanism, the multi-dimensional test is performed on the system under test through the distributed test architecture.

4. The 5G power network conformance protocol testing method as described in claim 1, characterized in that, S3 includes: S31. Measure network performance indicators by sending specially designed test data packets through active probing methods; S32. Capture network traffic and record protocol interaction processes using passive monitoring methods; S33. Align the data from multiple collection points of the distributed test architecture using timestamp synchronization technology; S34. Based on the network performance indicators and the protocol interaction process, the test results are generated.

5. The 5G power network conformance protocol testing method as described in claim 4, characterized in that, The active detection method in S31 includes: S311. Send a sequence of probe data packets with specific characteristics, measure the transmission delay and loss of each data packet, and form a delay change trend; S312. Based on the aforementioned latency change trend, the statistical measure of the latency increase trend and the normalized measure of the latency change degree are measured using the following formula, and the network performance index is formed: ， ， Wherein, the S PCT S is a statistical measure of the increasing trend of time delay. PDT As a normalized measure of the degree of time delay variation, the B k Here, H is the latency measurement value of the kth data packet, H is the probe window value of the Pathload algorithm, and I(·) is an indicator function, which takes the value of 1 when the condition of the indicator function is met and takes the value of 0 when the condition is not met.

6. The 5G power network conformance protocol testing method as described in claim 5, characterized in that, The active detection method in S31 further includes: S313. Adjust the sending rate of the data packets according to the network performance indicators, and converge the sending rate of the probe data packet sequence using a binary search algorithm; S314. When the difference in the transmission rate of three adjacent data packets in the probe data packet sequence is less than a preset threshold, the transmission of the probe data packet sequence is stopped.

7. The 5G power network conformance protocol testing method as described in claim 5, characterized in that, The measurement of transmission delay and loss of each data packet in S311 includes: S3111. Simulate real business traffic patterns and generate test traffic through protocol simulation and the probe data packet sequence; S3112. Calculate the performance indicators of the data packet based on the measurement data from the receiving end and the following formula, and obtain the transmission delay and the data loss situation: ， ， Among them, the The number of data packets successfully received by the receiving end. The maximum sequence number value in the received data packet, The starting sequence number of the receiving window is the reference value, and J is the dynamic estimate of the delay jitter. This is the measured delay difference between adjacent data packets.

8. The 5G power network conformance protocol test method as described in any one of claims 1 to 7, characterized in that, The formal model described in S4 analyzes the test results, including: S41. Establish a protocol model based on an extended finite state machine; S42. Use model checking technology to verify protocol consistency; S43. Verify the properties of complex protocols using theorem proof methods.

9. The 5G power network conformance protocol testing method as described in claim 8, characterized in that, The formal model described in S4 further includes the following for analyzing the test results: S43. Train the anomaly detection model using historical test data; S44. Extract the feature vector of the protocol message sequence; S45. Based on the feature vector, identify abnormal patterns using a deep learning algorithm.

10. The 5G power network conformance protocol test method as described in any one of claims 1 to 7, characterized in that, The adoption of a hierarchical decision-making mechanism in S4 to determine consistency and form a consistency level assessment includes: S46. Verify signal quality and connection stability at the physical layer; S47. Verify message format and timing compliance at the protocol layer; S48. Verify service quality and performance metrics at the business layer; S49. A weighted scoring mechanism is used to comprehensively evaluate the consistency level.

Images