Method and system for testing and evaluating adaptation degree of domestic hardware of power grid system

Abstract

The invention provides a method and system for testing and evaluating the adaptation degree of domestic hardware of a power grid system, and the method specifically comprises the following steps: S1, obtaining the basic information of system hardware, and obtaining a system hardware performance index, a system application performance index and a system application stability performance index through an automatic test; the system application performance indexes comprise performance indexes of eight types of applications and database applications in the power grid system; s2, calculating a system hardware performance score, a system application performance score and a system application stability score according to the system hardware basic information collected in S1 and each performance index, and obtaining a comprehensive performance score; and S3, generating an adaptation degree test report of the power grid system to the current hardware based on each performance index and each score. According to the method, the whole process from hardware basic parameter collection to high-load testing and stability and functional evaluation is covered, and comprehensive testing and evaluation of the adaptation degree of domestic hardware in specific platforms such as a power grid system are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of software and hardware testing, and in particular to a method and system for testing and evaluating localized hardware adaptation of a power grid system. Background Art

[0002] In recent years, the country has been vigorously promoting the process of replacing domestic hardware. As the core business system supporting the operation of large power grids, the power grid system needs to accelerate the pace of replacing domestic hardware. However, the domestic complete equipment currently used still contains a large number of imported components. It is urgent to establish an evaluation method and system for the adaptability of domestic hardware to verify the operating capabilities in the domestic hardware environment and ensure the stable and reliable operation of the power grid system.

[0003] Since the application scope of domestic hardware was relatively limited before, the existing software and hardware testing methods are mostly based on foreign hardware, and only a small number of methods involve domestic hardware testing. There is a relative lack of specialized and comprehensive evaluation methods for power grid systems, and the test content is often focused on general software or only evaluates from a single aspect such as system databases and Web applications. Existing hardware detection tools on the market (such as 360 Hardware Master, Master Lu, etc.) mainly provide general performance scores, fail to meet specific industry needs, and lack evaluation of the adaptability of hardware in specific platforms such as power grid systems.

[0004] Under the background that the country vigorously promotes domestic hardware to replace imported hardware, in order to comprehensively compare the business support capabilities of domestic hardware and imported hardware for power grid systems, the present invention proposes a method and system for testing and evaluating the adaptation of domestic hardware of power grid systems. Summary of the invention

[0005] The purpose of the present invention is to propose a method and system for testing and evaluating the adaptation of domestic hardware in power grid systems, covering the entire process from hardware basic parameter collection to high-load testing, stability and functionality evaluation, and realizing comprehensive testing and evaluation of the adaptability of domestic hardware in specific platforms such as power grid systems.

[0006] To achieve the above purpose, the technical solution of the present invention is: a method for testing and evaluating the adaptability of localized hardware of a power grid system, which specifically includes the following steps:

[0007] S1. Obtain basic information of system hardware, and obtain system hardware performance indicators, system application performance indicators, and system application stability performance indicators through automated testing;

[0008] The system hardware performance indicators include the performance indicators of the CPU, memory, hard disk and graphics card under high load;

[0009] The system application performance indicators include the performance indicators of eight categories of applications and database applications in the power grid system; the eight categories of applications include real-time monitoring, automatic control, analysis and verification, training simulation, spot market, new energy forecasting, operation evaluation and dispatch management;

[0010] S2. Calculate the system hardware performance score, system application performance score and system application stability score based on the system hardware basic information and various performance indicators collected in S1, and add the system hardware performance score, system application performance score and system application stability score to obtain a comprehensive performance score;

[0011] S3. Generate a test report on the adaptability of the power grid system to the current hardware based on various performance indicators and scores.

[0012] Preferably, the acquisition of basic hardware information of the system is specifically to collect basic hardware information by running a Shell script, including the number of cores, number of threads, cache size and main frequency of the CPU, the frequency, capacity and bus width of the memory, and the total capacity of the hard disk and the video memory size and video memory width of the graphics card.

[0013] Preferably, the acquisition of the system hardware performance index is specifically as follows:

[0014] CPU performance indicators under high load: running a highly computationally intensive program and keeping the program running until the CPU usage reaches a preset threshold and stabilizes for a preset time, recording the CPU temperature under high load and the number of threads when the test program is running; the highly computationally intensive program includes a program that calculates the approximate value of π based on the Monte Carlo method;

[0015] Performance indicators of memory under high load: first record the initial physical memory usage; then run a memory stress test program that increases the memory load by continuously applying for large blocks of physical memory space, so that the memory reaches the peak usage rate, and record the physical memory usage at this time, and calculate the difference in physical memory usage before and after running the memory stress test program; run the memory rate test program, through continuous read / write operations, and divide the total amount of data read / written by the running time to measure the memory read speed and write speed respectively;

[0016] Performance indicators of hard disks under high load: Sequential read and write tests are performed by continuously reading and writing large files, and the total data volume is counted and divided by the running time to calculate the sequential read and write speed; random read and write tests are performed by randomly reading and writing small blocks of data at different locations in the file, and the total data volume is counted and divided by the operation time to obtain the random read and write speed;

[0017] Performance indicators of graphics cards under high load conditions: First, use the GPU to run a matrix operation test program, set up two large-scale matrices and load them into the GPU's video memory; then, set up a loop structure in the program to repeatedly perform preset rounds of matrix multiplication operations; after each round of multiplication, the result is written back to the video memory, and the next round of operations uses this result as input to continue calculations; during the calculation process, the graphics card's operating frequency is read in real time through the driver interface, its frequency changes are monitored and frequency data at multiple times are recorded; finally, by averaging the frequency data at multiple times, the representative operating frequency of the graphics card under a high load environment is obtained.

[0018] Preferably, the acquisition of the system application performance indicator is specifically as follows;

[0019] Performance indicators of eight categories of applications in the power grid system: Run the power grid system software stress test program to test these eight categories of applications respectively, and monitor and record the system's performance indicators in real time, including response time, throughput, concurrency, function error rate, CPU utilization, and hardware temperature;

[0020] Performance indicators of database applications in power grid systems: First, by writing a multi-threaded program and setting a fixed number of threads, multiple threads send data read and write requests to the database at the same time to simulate concurrent access in actual applications; during the test run, the monitoring program collects CPU and memory usage in real time; at the same time, the response time of each thread to complete the data read and write operation is measured, and the average response time of all threads is taken as the average response time of the database under concurrent load.

[0021] Preferably, the power grid system software stress testing program is used to simulate the high load of eight major types of power grid applications, including increasing the load of front-end acquisition applications by sending large-scale data packets, or simulating the load pressure on the eight major types of applications by increasing the number of concurrent requests.

[0022] Preferably, the system application stability performance index is obtained as follows:

[0023] Robotic process automation technology RPA is introduced to build a simulation operation process to simulate the actual operation scenarios of various applications in the power grid system. During the test, the system monitors and collects performance indicator data in real time, including the average usage of CPU and memory, and compares and analyzes it with the preset stable operation time requirements to verify whether the system can meet the conditions for long-term stable operation. The system monitoring program records the start and end time of the test, calculates the actual operation time, and compares it with the preset target time to determine whether the test time meets the standard and obtain the indicator value of whether the test time meets the standard. At the same time, the monitoring program will continue to track the status of each process, detect whether there is unstable behavior including process crash or abnormal exit, and record detailed log information.

[0024] Preferably, the calculation of the system hardware performance score is specifically as follows:

[0025] The CPU performance score is calculated based on the basic system hardware information obtained by S1 and the CPU performance indicators under high load:

[0026]

[0027] Among them, S1 represents the upper limit of CPU temperature, which is a preset constant, S2 represents the CPU temperature under high load, n represents the number of CPUs, S 3_i Indicates the number of physical cores of the i-th CPU, S' 3_i represents the number of threads supported by the core of the i-th CPU, S4 represents the cache, S5 represents the main frequency parameter, and S6 represents the number of threads when the test program is running; W2-W6 are the weights of the corresponding parameters;

[0028] The memory performance score is calculated based on the basic system hardware information obtained by S1 and the performance indicators of the memory under high load:

[0029] MemScore=W7×(1-S7)+W8×S8+W9×S9+W 10 ×(1-S 10 )+W 11 ×S 11 +W 12 ×S 12 (2)

[0030] Among them, S7 represents the memory frequency, S8 represents the memory capacity, S9 represents the bus width, and S 10 Indicates the difference in physical memory usage before and after running the memory stress test program, S 11 Indicates the read speed measured by running the memory speed test program, S 12 Indicates the write speed measured by running a memory speed test program; W7-W 12 is the weight of the corresponding parameter;

[0031] The hard disk performance score is calculated based on the basic system hardware information obtained by S1 and the hard disk performance indicators under high load:

[0032] HDScore=W 13 ×S 13 +W 14 ×S 14 +W 15 ×S 15 +W 16 ×S 16 +W 17 ×S 17 (3)

[0033] Where S 13 Indicates the hard disk capacity, S 14 Indicates the measured sequential read speed, S 15 Indicates the measured sequential write speed, S 16 Indicates the measured random read speed, S 17 Indicates the measured random write speed; W 13 -W 17 is the weight of the corresponding parameter;

[0034] The graphics card performance score is calculated based on the basic system hardware information obtained by S1 and the performance indicators of the graphics card under high load:

[0035] GPUScore = W 18 ×S 18 +W 19 ×S 19 +W 20 ×S 20 (4)

[0036] Where S 18 Indicates the size of video memory, S 19 Indicates the memory bit width, S 20 Indicates the representative operating frequency under high load environment; W 18 -W 20 is the weight of the corresponding parameter;

[0037] Calculate system hardware performance score:

[0038] Q1=CPUScore+MemScore+HDScore+GPUScore.

[0039] Preferably, the calculation of the system application performance score is specifically as follows:

[0040] The fuzzy analytic hierarchy process is used to calculate the performance scores of eight categories of applications;

[0041] First, a hierarchical model of applications in eight categories in the power grid system is constructed:

[0042] Target layer A is the degree of autonomy and controllability of the power grid dispatching technical support system;

[0043] Criteria layer B includes real-time monitoring B1, automatic control B2, analysis and verification B3, training simulation B4, spot market B5, new energy forecast B6, operation evaluation B7, and dispatch management B8;

[0044] The indicator layer C corresponding to real-time monitoring B1 includes response time C11, concurrency C12, and other indicators C13;

[0045] The indicator layer C corresponding to the automatic control B2 includes the function error rate C21, the response time C22 and other indicators C23;

[0046] The indicator layer C corresponding to the analysis and verification B3 includes CPU utilization C31, hardware temperature C32 and other indicators C33;

[0047] The indicator layer C corresponding to the training simulation B4 includes function error rate C41, hardware temperature C42, CPU utilization C43 and other indicators C44;

[0048] The indicator layer C corresponding to the spot market B5 includes throughput C51, concurrency C52, and other indicators C53;

[0049] The indicator layer C corresponding to the new energy prediction B6 includes CPU utilization C61 and other indicators C62;

[0050] The indicator layer C corresponding to the operation evaluation B7 includes throughput C71, CPU utilization C72, and other indicators C73;

[0051] The indicator layer C corresponding to the scheduling management B8 includes hardware temperature C81, response time C82 and other indicators C83;

[0052] Then, the judgment matrix is ​​constructed. Before constructing the judgment matrix, the triangular fuzzy number scale is introduced and combined with the hierarchical analysis method to quantify the indicators of each target layer and assign corresponding weights. Among them, the triangular fuzzy number scales with equal importance are The slightly more important triangular fuzzy number scale is A very important triangular fuzzy number scale is The extremely important triangular fuzzy number scale is

[0053] Next, we use the performance indicators of eight categories of applications to construct a judgment matrix for the criterion layer and the indicator layer. The judgment matrix for the criterion layer is as follows:

[0054]

[0055] The indicator layer judgment matrix is ​​constructed as follows:

[0056]

[0057] b ij represents the relative importance of the i-th indicator in the criterion layer to the j-th indicator, expressed in terms of triangular fuzzy number scale, then b ji =1 / b ij , C ij Indicates the relative importance of the i-th indicator in the indicator layer to the j-th indicator, expressed by a triangular fuzzy number scale, then C ji =1 / C ij ;

[0058] Solve the eigenvalue and eigenvector, normalize the judgment matrix, and add them row by row to get the weight vector W b :

[0059]

[0060] Vector W b After normalization, solve the maximum eigenvalue of B:

[0061]

[0062] When comparing the importance of evaluation indicators at each level and assigning weights, the defuzzified judgment matrix is ​​subjected to consistency check;

[0063]

[0064] Where: max represents the maximum eigenvalue of the judgment matrix, and n represents the order of the judgment matrix;

[0065] The consistency of the matrix is ​​judged by the index, and its mathematical expression is:

[0066]

[0067] RI represents the random consistency index, CR represents the consistency ratio, when CR≤0.1, the judgment matrix passes the consistency test, otherwise the weight needs to be reset;

[0068] Conduct fuzzy comprehensive evaluation:

[0069] (1) Determine the factor set and the evaluation set: The factor set U is the index layer element: U = {u1,u2,...,u n Evaluation grade score vector: V = {extremely poor (20), poor (40), good (60), excellent (80), extremely good (100)}

[0070] (2) Constructing the membership matrix: Through data analysis, determine the membership of the indicators to the evaluation level and construct the membership matrix R;

[0071]

[0072] where r ij Indicator u i Evaluation level v j The degree of membership;

[0073] (3) Defuzzification, calculation of performance score: using the weight vector W b , membership matrix R, evaluation grade score vector V, the final performance score formula is:

[0074] APPScore=W b ·R·V T ;

[0075] Calculate the system database application performance score based on the performance indicators of the database application in the power grid system:

[0076]

[0077] Where S 23 Indicates the average CPU usage during database testing, S 24 Indicates the average memory usage during database testing, S 25_i represents the i-th response time of the database during the test; W 23 -W 24 is the weight of the corresponding parameter;

[0078] Calculate the system application performance score:

[0079] Q2=APPScore+DBScore.

[0080] Preferably, the calculation of the system application stability score is as follows:

[0081] Calculate the system application stability performance score based on the obtained system application stability performance indicators:

[0082] Q3=(W 26 ×(1-S 26 )+W 27 ×(1-S 27 ))×S 28 ×S 29 (7)

[0083] Among them, S 26 Indicates the average CPU usage during system stability test, S 27 Indicates the average memory usage during system stability test, S 28 Indicates whether the test duration meets the standard. If the actual running time ≥ the preset stable running time, S 28 The value is 1, indicating that the duration is up to the standard. If the actual running time is less than the preset stable running time, S 28 The value is calculated based on the ratio of the actual running time to the preset time, specifically:

[0084]

[0085] S 29 Indicates process stability; W 26 -W 27 is the weight of the corresponding parameter.

[0086] A localized hardware adaptability testing and evaluation system for a power grid system comprises a processor, a memory and a computer program stored in the memory. When the processor executes the computer program, the steps in the above-mentioned adaptability testing and evaluation method are specifically performed.

[0087] Compared with the prior art, the present invention has the following beneficial effects:

[0088] The present invention is applicable to a variety of domestic hardware platforms, supports consistency testing of chips of different architectures and models, and compares them with imported hardware, helping to clarify the business support capabilities of domestic hardware in the power grid system, and providing strong data support for subsequent hardware selection and system optimization.

[0089] The present invention provides a comprehensive performance evaluation method specifically for power grid systems, covering the entire process from hardware basic parameter collection to high-load testing, stability and functionality evaluation, filling the technical gap in the evaluation of the adaptability of domestically produced hardware in power grid systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 Schematic diagram of the power grid system hardware adaptability evaluation system of the present invention. DETAILED DESCRIPTION

[0091] The following is combined with Figure 1 , the technical solution of the present invention is specifically described.

[0092] In view of the fact that current evaluation methods pay little attention to the specific needs of the power grid system, the present invention proposes a power grid system adaptability evaluation method based on the existing evaluation technology to achieve the following functions:

[0093] 1. The system automatically collects various hardware parameters of the domestic system in the current network environment by running Shell scripts in the background, generates basic hardware information and displays it.

[0094] 2. Perform system hardware stress testing. Develop hardware stress testing programs based on Linux system instructions or C language programs. By running these stress testing programs, increase the load on each hardware component, simulate a high-intensity operating environment, and obtain performance indicator data of the hardware under peak load.

[0095] 3. Conduct system application stress testing. For power grid system applications, develop a software stress testing program based on C language program. For front-end acquisition applications, increase the order of magnitude of data packets sent, for eight major types of applications, increase the number of concurrent requests, and for database applications, perform high-frequency read and write operations to conduct software stress testing and obtain software performance indicator data.

[0096] 4. Conduct system application stability tests, build simulation operation processes based on RPA (robotic process automation) technology, simulate the actual operation scenarios of various types of applications in the power grid system, monitor the operation status of the entire system in real time, and collect key performance indicator data of system applications under stability tests.

[0097] 5. Based on the parameter indicator data collected under hardware stress testing, application stress testing, and application stability testing, a scientific fitness scoring model is constructed to generate an evaluation score.

[0098] 6. Based on the performance indicators collected above and the final evaluation scores, generate a test report on the adaptability of the power grid system to the current hardware and provide detailed test conclusions.

[0099] The present invention conducts comprehensive testing around multiple dimensions of system hardware performance, system application performance, and system application stability performance; conducts in-depth research on the business support capabilities of the power grid system in a domestic chip environment, and conducts comparative analysis with the business support capabilities in an imported chip environment, thereby verifying the feasibility and operational reliability of the power grid system in a highly autonomous and controllable environment. The present invention proposes a method for testing and evaluating the adaptability of localized hardware of a power grid system, which specifically includes the following steps:

[0100] S1. Obtain basic information of system hardware, and obtain system hardware performance indicators, system application performance indicators, and system application stability performance indicators through automated testing;

[0101] The system hardware performance indicators include the performance indicators of the CPU, memory, hard disk and graphics card under high load;

[0102] The system application performance indicators include the performance indicators of eight major types of applications and database applications in the power grid system;

[0103] S2. Calculate the system hardware performance score, system application performance score and system application stability score based on the system hardware basic information and various performance indicators collected in S1, and add the system hardware performance score, system application performance score and system application stability score to obtain a comprehensive performance score;

[0104] S3. Generate a test report on the adaptability of the power grid system to the current hardware based on various performance indicators and scores.

[0105] In this embodiment, the basic information of the system hardware is obtained by running a Shell script to collect the basic information of the hardware, including the number of cores, number of threads, cache size and main frequency of the CPU, the frequency, capacity and bus width of the memory, and the total capacity of the hard disk and the memory size and memory width of the graphics card.

[0106] In this embodiment, the system hardware performance indicators are obtained as follows:

[0107] As the core computing unit of the system, the performance of the CPU is directly related to the efficiency of task scheduling and data processing. The performance indicators of the CPU under high load are as follows:

[0108] Run a highly computationally intensive program to simulate a high-load state, and keep the program running until the CPU usage reaches a preset threshold and stabilizes for a period of time, and record the CPU temperature under high load and the number of threads when the test program is running; the highly computationally intensive program includes a program for calculating the approximate value of π based on the Monte Carlo method;

[0109] As a temporary data storage unit of the system, the performance of memory has an important impact on data processing speed and system response efficiency. The performance indicators of memory under high load are as follows:

[0110] In order to exclude the influence of other programs in the system environment during the evaluation, the initial physical memory usage is first recorded. Then, a memory stress test program is run to increase the memory load by continuously applying for large blocks of physical memory space, so that the memory reaches the peak usage rate. After the program reaches a stable state, the physical memory usage at this time is recorded, and the difference in physical memory usage before and after running the memory stress test program is calculated. The memory rate test program is run to measure the memory reading speed and writing speed respectively through continuous read / write operations and the total amount of data read / written is divided by the running time.

[0111] The hard disk is a long-term storage unit of the system, and its read and write speed directly affects the data loading and storage efficiency. The performance indicators of the hard disk under high load are as follows:

[0112] Write 500MB of random character data to an empty file through a preset program, calculate the write speed by dividing the total data size by the write time, then read all the data in the file, and get the read speed by dividing the total data size by the read time; perform random read and write tests by randomly reading and writing small blocks of data (such as 4KB) at different locations in the file, calculate the total amount of data and divide it by the operation time to get the random read and write speed;

[0113] The graphics card is responsible for processing graphics rendering and complex calculations in the power grid system. Its performance directly affects the smoothness and response speed of tasks such as data visualization and graphic analysis. In order to accurately obtain the peak frequency of the graphics card, a multi-step measurement method is used to ensure the accuracy and representativeness of the data. The performance indicators of the graphics card under high load are as follows:

[0114] First, the GPU is used to run a matrix operation test program. Two large-scale matrices are set up and loaded into the GPU's video memory. Then, a loop structure is set in the program to repeatedly perform a preset number of rounds (thousands of times) of matrix multiplication operations. After each round of multiplication, the result is written back to the video memory, and the next round of operations uses this result as input to continue the calculation. This design enables the graphics card to continuously process large amounts of data and complex calculations, so that its core components (such as CUDA cores or stream processors) are fully engaged and reach a high-load state. During the calculation process, the graphics card's operating frequency is read in real time through the driver interface, its frequency changes are monitored, and frequency data at multiple times are recorded. Finally, by averaging the frequency data at multiple times, the representative operating frequency of the graphics card under a high-load environment is obtained.

[0115] In this embodiment, the system application performance indicator is obtained as follows:

[0116] The execution efficiency and stability of the eight categories of applications directly affect the system's ability to support daily tasks. Table 1 shows which eight categories of applications are which and the indicators that each category of application focuses on:

[0117] Table 1 Eight categories of applications and their focus indicators

[0118] Application Name Focus on indicators Real-time monitoring Response time, concurrency Automatic Control Functional error rate, response time Analysis and verification CPU utilization, hardware temperature Training Simulation Function error rate, hardware temperature, CPU utilization Spot Market Throughput and concurrency New Energy Forecast CPU Utilization Run the assessment Throughput, CPU utilization Scheduling Management Hardware temperature, response time

[0119] To evaluate application performance, we run software stress testing programs designed specifically for power grid systems. For example, we increase the load on front-end acquisition applications by sending large-scale data packets, or we increase the number of concurrent requests to simulate the load pressure of eight types of power grid applications. The high-load scenarios created in this way can effectively test the system's carrying capacity and response performance under extreme conditions. During the test, we monitor and record the system's performance indicators in real time, including response time, throughput, concurrency, function error rate, CPU utilization, and hardware temperature.

[0120] The database is responsible for high-frequency data storage and query tasks, and its performance determines the system data processing efficiency. The performance indicators of database applications in power grid systems are as follows:

[0121] In order to comprehensively evaluate the application performance of the database, a multi-threaded test method is used to simulate high-load scenarios and collect performance data. First, by writing a multi-threaded program and setting a fixed number of threads, multiple threads send data read and write requests to the database at the same time to simulate concurrent access in actual applications. During the test run, the monitoring program collects CPU and memory usage in real time and records the consumption of system resources. At the same time, the response time of each thread to complete the data read and write operation is measured, and the average response time of all threads is taken as the average response time of the database under concurrent load.

[0122] In this embodiment, the stability test is intended to verify whether each component of the system can operate continuously and stably under different loads, temperatures and environmental conditions during long-term operation; the system application stability performance index is obtained as follows:

[0123] The robotic process automation technology RPA is introduced to build a simulation operation process to simulate the actual operation scenarios of various applications in the power grid system. During the test, the system monitors and collects performance indicator data in real time, including the average usage of CPU and memory, and compares and analyzes it with the preset stable operation time requirements to verify whether the system can meet the conditions for long-term stable operation. The system monitoring program records the start and end time of the test, calculates the actual operation time, and compares it with the preset target time to determine whether the test time meets the standard and obtain the indicator value of whether the test time meets the standard. At the same time, the monitoring program will continue to track the status of each process, especially paying attention to whether there are unstable behaviors such as process crashes or abnormal exits, and record detailed log information.

[0124] In this embodiment, the calculation of the system hardware performance score is specifically as follows:

[0125] The CPU performance score is calculated based on the basic system hardware information obtained by S1 and the CPU performance indicators under high load:

[0126]

[0127] Among them, S1 represents the upper limit of CPU temperature, which is a preset constant, S2 represents the CPU temperature under high load, n represents the number of CPUs, S 3_i Indicates the number of physical cores of the i-th CPU, S' 3_i represents the number of threads supported by the core of the i-th CPU, S4 represents the cache, S5 represents the main frequency parameter, and S6 represents the number of threads when the test program is running; W2-W6 are the weights of the corresponding parameters;

[0128] The memory performance score is calculated based on the basic system hardware information obtained by S1 and the performance indicators of the memory under high load:

[0129] MemScore=W7×(1-S7)+W8×S8+W9×S9+W 10 ×(1-S 10 )+W 11 ×S 11 +W 12 ×S 12 (2)

[0130] Among them, S7 represents the memory frequency, S8 represents the memory capacity, S9 represents the bus width, and S 10Indicates the difference in physical memory usage before and after running the memory stress test program, S 11 Indicates the read speed measured by running the memory speed test program, S 12 Indicates the write speed measured by running a memory speed test program; W7-W 12 is the weight of the corresponding parameter;

[0131] The hard disk performance score is calculated based on the basic system hardware information obtained by S1 and the hard disk performance indicators under high load:

[0132] HDScore=W 13 ×S 13 +W 14 ×S 14 +W 15 ×S 15 +W 16 ×S 16 +W 17 ×S 17 (3)

[0133] Where S 13 Indicates the hard disk capacity, S 14 Indicates the measured sequential read speed, S 15 Indicates the measured sequential write speed, S 16 Indicates the measured random read speed, S 17 Indicates the measured random write speed; W 13 -W 17 is the weight of the corresponding parameter;

[0134] The graphics card performance score is calculated based on the basic system hardware information obtained by S1 and the performance indicators of the graphics card under high load:

[0135] GPUScore = W 18 ×S 18 +W 19 ×S 19 +W 20 ×S 20 (4)

[0136] Where S 18 Indicates the size of video memory, S 19 Indicates the memory bit width, S 20 Indicates the representative operating frequency under high load environment; W 18 -W 20 is the weight of the corresponding parameter;

[0137] Calculate system hardware performance score:

[0138] Q1=CPUScore+MemScore+HDScore+GPUScore.

[0139] Preferably, the calculation of the system application performance score is specifically as follows:

[0140] The performance scores of the eight categories of applications in the power grid system are calculated based on the performance indicators of the eight categories of applications in the power grid system.

[0141] The fuzzy analytic hierarchy process is used to calculate the performance scores of the eight categories of applications. This method combines the advantages of the analytic hierarchy process and the fuzzy comprehensive evaluation method. First, a hierarchical model is constructed and a fuzzy judgment matrix is ​​established. The weights of each level are determined after the consistency test by calculating the eigenvector and performing defuzzification. Subsequently, the fuzzy comprehensive evaluation method is used to determine the factor set and construct the judgment set, and finally the comprehensive evaluation result is obtained.

[0142] According to the above calculation steps, we first construct the hierarchical structure types of each application in the eight categories in the power grid system, as shown in Table 2:

[0143] Table 2 Hierarchical structure model of eight types of applications at the power grid system level

[0144]

[0145]

[0146] Constructing a judgment matrix

[0147] Before constructing the judgment matrix, it is also necessary to introduce the triangular fuzzy number scale (Table 3) and combine it with the hierarchical analysis method to quantify the indicators of each target layer and assign corresponding weights.

[0148] Table 3 Triangular fuzzy number scaling criteria

[0149]

[0150]

[0151] Next, we use the performance indicators of eight categories of applications to construct a judgment matrix for the criterion layer and the indicator layer. The judgment matrix for the criterion layer is as follows:

[0152]

[0153] The indicator layer judgment matrix is ​​constructed as follows:

[0154]

[0155] b ij represents the relative importance of the i-th indicator in the criterion layer to the j-th indicator, expressed in terms of triangular fuzzy number scale, then b ji =1 / b ij , C of the index layer ijSame reason.

[0156] Solving for eigenvalues ​​and eigenvectors

[0157] Solve the eigenvalue and eigenvector, normalize the judgment matrix, and add them row by row to get the weight vector W b :

[0158]

[0159] Vector W b After normalization, solve the maximum eigenvalue of B:

[0160]

[0161] Consistency Check

[0162] When comparing the importance of evaluation indicators at each level and assigning weights, the consistency of thinking should be ensured. Therefore, it is necessary to conduct a consistency test on the defuzzified judgment matrix.

[0163]

[0164] Where: max represents the maximum eigenvalue of the judgment matrix, and n represents the order of the judgment matrix.

[0165] In actual use, the consistency of the matrix is ​​generally judged by an indicator, and its mathematical expression is:

[0166]

[0167] Table 4 Random consistency index values

[0168] n 1 2 3 4 5 6 7 8 9 10 11 RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49 1.51

[0169] Generally, when CR≤0.1, the judgment matrix passes the consistency test, otherwise the weights need to be reset.

[0170] Fuzzy comprehensive evaluation

[0171] (1) Determine the factor set and judgment set:

[0172] Factor set U, i.e., index layer elements, such as U = {u1,u2,...,u n}

[0173] Based on the evaluation set level correspondence table (Table 5), the evaluation level score vector is obtained:

[0174] V={Very bad (20), Poor (40), Good (60), Excellent (80), Very good (100)}

[0175] Table 5 Correspondence table of evaluation set levels

[0176] Performance level Very bad Difference good excellent Excellent Corresponding score 20 40 60 80 100

[0177] (2) Construct the membership matrix:

[0178] Through data analysis, the degree of membership of indicators to the evaluation level is determined, and the membership matrix R is constructed.

[0179]

[0180] where r ij Indicator u i Evaluation level v j The degree of membership.

[0181] (3) Defuzzification, calculation of performance score:

[0182] Using the weight vector W b , membership matrix R, evaluation grade score vector V, the final performance score formula is:

[0183] APPScore=W b ·R·V T

[0184] Calculate the system database application performance score based on the performance indicators of the database application in the power grid system:

[0185]

[0186] Where S 23 Indicates the average CPU usage during database testing, S 24 Indicates the average memory usage during database testing, S 25_i represents the i-th response time of the database during the test; W 23 -W 24 is the weight of the corresponding parameter;

[0187] Calculate the system application performance score:

[0188] Q2=APPScore+DBScore.

[0189] Preferably, the calculation of the system application stability score is as follows:

[0190] Calculate the system application stability performance score based on the obtained system application stability performance indicators:

[0191] Q3=(W 26 ×(1-S 26 )+W 27 ×(1-S 27 ))×S 28 ×S 29(7)

[0192] Among them, S 26 Indicates the average CPU usage during system stability test, S 27 Indicates the average memory usage during system stability test, S 28 Indicates whether the test duration meets the standard. If the actual running time ≥ the preset stable running time, S 28 The value is 1, indicating that the duration is up to the standard. If the actual running time is less than the preset stable running time, S 28 The value is calculated based on the ratio of the actual running time to the preset time, specifically:

[0193]

[0194] S 29 Indicates process stability; W 26 -W 27 is the weight of the corresponding parameter.

[0195] By scoring the above sub-indicators, the comprehensive performance score can be obtained as follows:

[0196] Q=Q1+Q2+Q3

[0197] The comprehensive performance score is a comprehensive reflection of the system hardware performance score (Q1), system application performance score (Q2), and system stability performance score (Q3).

[0198] After completing the comprehensive testing of system hardware performance, application performance, and stability performance, and obtaining a comprehensive performance score, this phase will integrate the previous scores into a comprehensive test report to provide a scientific and systematic evaluation basis for the hardware adaptability of the power grid system. The test report provides support for decision-making on domestic hardware adaptation through visual display, in-depth analysis, and adaptability scoring. The following are the detailed steps:

[0199] 1. Visualization of results

[0200] First, the test results are presented intuitively using visualization methods such as charts. The report will cover multi-dimensional data on system hardware performance, application performance, and stability performance, and display the performance of different systems in various key indicators during the test process through various forms such as line charts, bar charts, and radar charts. This display method helps to clearly reveal the adaptability and performance differences between domestic and imported systems under different architectures (such as X86 and ARM), making the test results clear at a glance and facilitating subsequent analysis and decision-making.

[0201] 2. Comparative analysis of various test indicators

[0202] In the comparative analysis section, the report will show in detail the specific performance of domestic and imported hardware systems under the same test scenarios and environmental conditions, as well as the differences in adaptability under different architectures. This section will deeply analyze the pros and cons of hardware performance, application performance, and stability performance, and reveal the impact of different configurations on the overall adaptability of the system, so as to accurately evaluate the degree of match between the actual needs of the system and the adaptability of different hardware.

[0203] 3. Compatibility score and optimization suggestions

[0204] Based on the performance scores of each item and the final comprehensive performance score, a fitness score is generated to quantify the overall fitness of the system. This score not only reflects the degree of match between the hardware and the application in the power grid system, but also serves as an important reference for hardware optimization and software adaptation strategies. For components with low fitness scores, the report will make specific improvement suggestions based on the test data, including feasible hardware adjustment solutions, performance optimization measures, and more suitable hardware selection to improve the overall fitness of the system.

[0205] A localized hardware adaptability testing and evaluation system for a power grid system comprises a processor, a memory and a computer program stored in the memory. When the processor executes the computer program, the steps in the above-mentioned adaptability testing and evaluation method are specifically performed.

[0206] The above are preferred embodiments of the present invention. Any changes made according to the technical solution of the present invention, as long as the resulting functions do not exceed the scope of the technical solution of the present invention, belong to the protection scope of the present invention.

Claims

1. A method for testing and evaluating the adaptability of localized hardware of a power grid system, characterized in that: The specific steps include: S1. Obtain basic information of system hardware, and obtain system hardware performance indicators, system application performance indicators, and system application stability performance indicators through automated testing; The system hardware performance indicators include the performance indicators of the CPU, memory, hard disk and graphics card under high load; The system application performance indicators include the performance indicators of eight categories of applications and database applications in the power grid system; the eight categories of applications include real-time monitoring, automatic control, analysis and verification, training simulation, spot market, new energy forecasting, operation evaluation and dispatch management; S2. Calculate the system hardware performance score, system application performance score and system application stability score based on the system hardware basic information and various performance indicators collected in S1, and add the system hardware performance score, system application performance score and system application stability score to obtain a comprehensive performance score; S3. Generate a test report on the adaptability of the power grid system to the current hardware based on various performance indicators and scores.

2. According to claim 1, a method for testing and evaluating the adaptability of localized hardware of a power grid system is characterized in that: The acquisition of basic system hardware information specifically includes collecting basic hardware information by running a Shell script, including the number of cores, number of threads, cache size and main frequency of the CPU, the frequency, capacity and bus width of the memory, as well as the total capacity of the hard disk and the memory size and memory width of the graphics card.

3. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 2, characterized in that: The acquisition of the system hardware performance indicators is specifically as follows; CPU performance indicators under high load: running a highly computationally intensive program and keeping the program running until the CPU usage reaches a preset threshold and stabilizes for a preset time, recording the CPU temperature under high load and the number of threads when the test program is running; the highly computationally intensive program includes a program that calculates the approximate value of π based on the Monte Carlo method; Performance indicators of memory under high load: First record the initial physical memory usage; Then run a memory stress test program that increases the memory load by continuously applying for large blocks of physical memory space, so that the memory reaches the peak usage rate, and record the physical memory usage at this time, and calculate the difference in physical memory usage before and after running the memory stress test program; Run the memory speed test program to measure the memory read speed and write speed through continuous read / write operations and divide the total amount of data read / written by the running time; Performance indicators of hard disks under high load: Sequential read and write tests are performed by continuously reading and writing large files, and the total data volume is counted and divided by the running time to calculate the sequential read and write speed; random read and write tests are performed by randomly reading and writing small blocks of data at different locations in the file, and the total data volume is counted and divided by the operation time to obtain the random read and write speed; The performance indicators of the graphics card under high load: First, use the GPU to run a matrix operation test program, set up two large matrices and load them into the GPU's video memory; then set up a loop structure in the program to repeatedly perform the matrix multiplication operation for a preset number of rounds; after each round of multiplication is completed, the result is written back to the video memory, and the next round of operation continues the calculation with this result as input; During the calculation process, the operating frequency of the graphics card is read in real time through the driver interface, its frequency changes are monitored and frequency data at multiple times are recorded; finally, by averaging the frequency data at multiple times, the representative operating frequency of the graphics card under a high-load environment is obtained.

4. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 1, characterized in that: The acquisition of the system application performance indicators is specifically as follows; Performance indicators of eight categories of applications in the power grid system: Run the power grid system software stress test program to test these eight categories of applications respectively, and monitor and record the system's performance indicators in real time, including response time, throughput, concurrency, function error rate, CPU utilization, and hardware temperature; Performance indicators of database applications in power grid systems: First, by writing a multi-threaded program and setting a fixed number of threads, multiple threads send data read and write requests to the database at the same time to simulate concurrent access in actual applications; during the test run, the monitoring program collects CPU and memory usage in real time; at the same time, the response time of each thread to complete the data read and write operation is measured, and the average response time of all threads is taken as the average response time of the database under concurrent load.

5. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 4 is characterized in that: The power grid system software stress test program is used to simulate the high load of eight major types of applications in the power grid, including increasing the load of the front-end collection application by sending large-scale data packets, or simulating the load pressure on the eight major types of applications by increasing the number of concurrent requests.

6. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 1, characterized in that: The system application stability performance index is obtained as follows: Robotic process automation technology (RPA) is introduced to build a simulation operation process to simulate the actual operation scenarios of various applications in the power grid system. During the test, the system monitors and collects performance indicator data in real time, including the average usage of CPU and memory, and compares and analyzes it with the preset stable operation time requirements to verify whether the system can meet the conditions for long-term stable operation. The system monitoring program records the start and end time of the test, calculates the actual operation time, and compares it with the preset target time to determine whether the test time meets the standard and obtain the indicator value of whether the test time meets the standard. At the same time, the monitoring program will continue to track the status of each process, detect whether there is unstable behavior including process crash or abnormal exit, and record detailed log information.

7. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 3 is characterized in that: The calculation of the system hardware performance score is as follows: The CPU performance score is calculated based on the basic system hardware information obtained by S1 and the CPU performance indicators under high load: Among them, S1 represents the upper limit of CPU temperature, which is a preset constant, S2 represents the CPU temperature under high load, n represents the number of CPUs, S 3_i Indicates the number of physical cores of the i-th CPU, S' 3_i represents the number of threads supported by the core of the i-th CPU, S4 represents the cache, S5 represents the main frequency parameter, and S6 represents the number of threads when the test program is running; W2-W6 are the weights of the corresponding parameters; The memory performance score is calculated based on the basic system hardware information obtained by S1 and the performance indicators of the memory under high load: MemScore=W7×(1-W7)+W8×W8+W9×W9+W 10 ×(1-S 10 )+W 11 ×S 11 +W 12 ×S 12 (2) Among them, S7 represents the memory frequency, S8 represents the memory capacity, S9 represents the bus width, and S 10 Indicates the difference in physical memory usage before and after running the memory stress test program, S 11 Indicates the read speed measured by running the memory speed test program, S 12 Indicates the write speed measured by running a memory speed test program; W7-W 12 is the weight of the corresponding parameter; The hard disk performance score is calculated based on the basic system hardware information obtained by S1 and the hard disk performance indicators under high load: HDScore=W 13 ×S 13 +W 14 ×S 14 +W 15 ×S 15 +W 16 ×S 16 +W 17 ×S 17 (3) Where S 13 Indicates the hard disk capacity, S 14 Indicates the measured sequential read speed, S 15 Indicates the measured sequential write speed, S 16 Indicates the measured random read speed, S 17 Indicates the measured random write speed; W 13 -W 17 is the weight of the corresponding parameter; The graphics card performance score is calculated based on the basic system hardware information obtained by S1 and the performance indicators of the graphics card under high load: GPUScore=W 18 ×S 18 +W 19 ×S 19 +W 20 ×S 20 (4) Where S 18 Indicates the size of video memory, S 19 Indicates the memory bit width, S 20 Indicates the representative operating frequency under high load environment; W 18 -W 20 is the weight of the corresponding parameter; Calculate system hardware performance score: Q1=CPUScore+MemScore+HDScore+GPUScore.

8. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 4, characterized in that: The calculation of the system application performance score is as follows: The fuzzy analytic hierarchy process is used to calculate the performance scores of eight categories of applications; First, a hierarchical model of applications in eight categories in the power grid system is constructed: Target layer A is the degree of autonomy and controllability of the power grid dispatching technical support system; Criteria layer B includes real-time monitoring B1, automatic control B2, analysis and verification B3, training simulation B4, spot market B5, new energy forecast B6, operation evaluation B7, and dispatch management B8; The indicator layer C corresponding to real-time monitoring B1 includes response time C11, concurrency C12, and other indicators C13; The indicator layer C corresponding to the automatic control B2 includes the function error rate C21, the response time C22 and other indicators C23; The indicator layer C corresponding to the analysis and verification B3 includes CPU utilization C31, hardware temperature C32 and other indicators C33; The indicator layer C corresponding to the training simulation B4 includes function error rate C41, hardware temperature C42, CPU utilization C43 and other indicators C44; The indicator layer C corresponding to the spot market B5 includes throughput C51, concurrency C52, and other indicators C53; The indicator layer C corresponding to the new energy prediction B6 includes CPU utilization C61 and other indicators C62; The indicator layer C corresponding to the operation evaluation B7 includes throughput C71, CPU utilization C72, and other indicators C73; The indicator layer C corresponding to the scheduling management B8 includes hardware temperature C81, response time C82 and other indicators C83; Then, the judgment matrix is ​​constructed. Before constructing the judgment matrix, the triangular fuzzy number scale is introduced and combined with the hierarchical analysis method to quantify the indicators of each target layer and assign corresponding weights. Among them, the triangular fuzzy number scales with equal importance are The slightly more important triangular fuzzy number scale is A very important triangular fuzzy number scale is The extremely important triangular fuzzy number scale is Next, we use the performance indicators of eight categories of applications to construct a judgment matrix for the criterion layer and the indicator layer. The judgment matrix for the criterion layer is as follows: The indicator layer judgment matrix is ​​constructed as follows: b ij represents the relative importance of the i-th indicator in the criterion layer to the j-th indicator, expressed in terms of triangular fuzzy number scale, then b ji =1 / b ij , C ij Indicates the relative importance of the i-th indicator in the indicator layer to the j-th indicator, expressed by a triangular fuzzy number scale, then C ji =1 / C ij ; Solve the eigenvalue and eigenvector, normalize the judgment matrix, and add them row by row to get the weight vector W b : Vector W b After normalization, solve the maximum eigenvalue of B: When comparing the importance of evaluation indicators at each level and assigning weights, the defuzzified judgment matrix is ​​subjected to consistency check; Where: max represents the maximum eigenvalue of the judgment matrix, and n represents the order of the judgment matrix; The consistency of the matrix is ​​judged by the index, and its mathematical expression is: RI represents the random consistency index, CR represents the consistency ratio, when CR≤0.1, the judgment matrix passes the consistency test, otherwise the weight needs to be reset; Conduct fuzzy comprehensive evaluation: (1) Determine the factor set and the evaluation set: The factor set U is the index layer element: U = {u1,u2,...,u n Evaluation grade score vector: V = {extremely poor (20), poor (40), good (60), excellent (80), extremely good (100)} (2) Constructing the membership matrix: Through data analysis, determine the membership of the indicators to the evaluation level and construct the membership matrix R; where r ij Indicator u i Evaluation level v j The degree of membership; (3) Defuzzification, calculation of performance score: using the weight vector W b , membership matrix R, evaluation grade score vector V, the final performance score formula is: APPScore=W b ·R·V T ; Calculate the system database application performance score based on the performance indicators of the database application in the power grid system: Where S 23 Indicates the average CPU usage during database testing, S 24 Indicates the average memory usage during database testing, S 25_i represents the i-th response time of the database during the test; W 23 -W 24 is the weight of the corresponding parameter; Calculate the system application performance score: Q2=APPScore+DBScore.

9. The method for testing and evaluating the adaptability of localized hardware of a power grid system according to claim 6, characterized in that: The calculation of the system application stability score is as follows: Calculate the system application stability performance score based on the obtained system application stability performance indicators: Q3=(In 26 ×(1-S 26 )+W 27 ×(1-S 27 ))×S 28 ×S 29 (7) Among them, S 26 Indicates the average CPU usage during the system stability test, S 27 Indicates the average memory usage during system stability test, S 28 Indicates whether the test duration meets the standard. If the actual running time ≥ the preset stable running time, S 28 The value is 1, indicating that the duration is up to the standard. If the actual running time is less than the preset stable running time, S 28 The value is calculated based on the ratio of the actual running time to the preset time, specifically: S 29 Indicates process stability; W 26 -W 27 is the weight of the corresponding parameter.

10. A localized hardware adaptability testing and evaluation system for a power grid system, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory. When the processor executes the computer program, the method specifically executes the steps in the method for testing and evaluating the degree of fit as described in any one of claims 1 to 9.

Images