Portable AVC automatic joint debugging test device and method

Abstract

The invention provides a portable AVC automatic joint debugging test device and method, and belongs to the technical field of power equipment. The device comprises a communication module, a data acquisition module, a user interface module, an application program module, a communication protocol stack module, a security software module and a data processing library module. The method comprises the following steps: generating a task request, sending the task request to a command theme subscribed by each AVC substation, obtaining real-time data collected by each AVC substation, decrypting and verifying to obtain plaintext data, inputting a dynamic and static curve real-time discrimination model constructed based on a deep reinforcement learning network, outputting a dynamic and static curve discrimination result, and outputting the dynamic and static curve discrimination result. And inputting a PID parameter cost function constructed by a model prediction control algorithm, solving optimized PID parameters of each substation, and issuing the optimized PID parameters to each AVC substation to recollect data until the comprehensive performance grade of each AVC substation is qualified. According to the invention, the blank of the AVC substation joint debugging test device is filled, and the judgment accuracy of the multi-AVC substation joint debugging test result is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power equipment, and more particularly to a portable AVC automatic commissioning test device and method. BACKGROUND

[0002] The automatic voltage control system (AVC) is a system for automatically adjusting the voltage and reactive power of a power system, and realizes the safe and economic operation of the power grid through centralized control of reactive resources. The core function of AVC is to automatically adjust the generator excitation, transformer tap changer, capacitor switching and other devices through real-time monitoring of the grid voltage, reactive load and device state, to maintain the voltage qualification rate and reduce the network loss. The goal is to ensure the economic operation of the power grid under the condition of meeting the safety constraints.

[0003] Therefore, the grid dispatching institution and the power generation enterprise believe that the network test of the AVC system necessary for ensuring the safety of the power system, the safety check of the power secondary system, and other network source key works should be carried out in a timely and high-quality manner to ensure that the performance indicators of the AVC system meet the requirements of relevant standards.

[0004] Currently, AVC substation test work mainly relies on manual recording mode, and there is no dedicated portable AVC automatic commissioning test device and test method. AVC substation closed-loop trial operation test requires manual reading of more than 1,000 adjustment results for 7 days, and the test personnel have a huge workload. AVC substation test work relies on manual calculation to determine performance indicators, which is limited by the differences in information reading functions of various AVC substations, and is prone to information omission or calculation errors, making it difficult to ensure the accuracy of test determination. The current technical means can only carry out AVC substation single test work, and cannot carry out multi- AVC substation test work.

[0005] Prior art document 1 (CN114123321B) discloses a new energy grid-connected test system. Its disadvantages are that it focuses on new energy grid-connected comprehensive testing and does not design for AVC substation commissioning; it relies on manual recording and calculation of data, which is low in efficiency and large in error; it only supports single station testing and has no multi- AVC substation collaborative commissioning capability; it has no portability and cannot adapt to field testing; and it needs to manually prepare reports and has no dynamic customization function.

[0006] Prior art document 2 (CN102064610B) discloses a decentralized power plant automatic voltage control substation system based on a device. Its disadvantages are that the test device of the present application is used to test the substation system, which has no AVC substation performance test capability; it is only suitable for specific substations and has poor compatibility; it cannot be used for multi-substation commissioning and has no automatic test mechanism; it has no test data security protection; and it needs manual data processing and report preparation. SUMMARY

[0007] To solve the problems in the prior art, the application provides a portable AVC automatic joint debugging test device and method, proposes a signal transceiver mechanism between the portable AVC automatic joint debugging test device and multiple AVC sub-stations based on the combination of a message queue telemetry transmission protocol (Message Queuing Telemetry Transport, MQTT) and a standard data structure JSON (JavaScript Object Notation, JSON), an information interaction mechanism between the portable AVC automatic joint debugging test device and the multiple AVC sub-stations based on a cyclic redundancy check (Cyclic Redundancy Check, CRC), and realizes automatic joint debugging test of an AVC system; proposes an information security isolation method between the portable AVC automatic joint debugging test device and the multiple AVC sub-stations based on multiple security layers, and maintains the safety and stability of a test environment; proposes an architecture design and module classification of the portable AVC automatic joint debugging test device, fills the technical gap of no AVC system dedicated test instrument in the industry, and provides a reliable technical means for developing AVC system test work; and proposes a dynamic customization test report template generation technology of the portable AVC automatic joint debugging test device. The portable AVC automatic joint debugging test device can be conveniently operated by a user, and interaction and customization options in a report generation process are realized according to the behavior, preference and historical data of the user, so as to meet different test requirements.

[0008] The application adopts the following technical solutions.

[0009] The first aspect of the application provides a portable AVC automatic joint debugging test device, which comprises a communication module, a data acquisition module, a user interface module, an application program module, a communication protocol stack module, a security software module and a data processing library module, and comprises: The application program module is used for generating a task request according to a parameter setting target and a limit condition set by the user interface module; The communication module is used for packaging and adding a VLAN identifier to the task request generated by the application program module in a JSON format, and determining a JSON test task; The communication module sends the JSON test task to a command topic subscribed by each AVC sub-station, and the communication module obtains a new JSON data packet filled with real-time data collected by each AVC sub-station according to the JSON test task from a data feedback topic of each AVC sub-station; The communication protocol stack module is used for performing CRC check on the new JSON data packet obtained by the communication module; The data acquisition module is used for extracting an initial test data set from the new JSON data packet that passes the check of the communication protocol stack module; The security software module is used for decrypting and verifying the initial test data set to obtain plaintext data; The data processing library module is configured to input the plaintext data into a dynamic and static curve real-time discrimination model constructed based on a deep reinforcement learning network to obtain a dynamic and static curve discrimination result, input the discrimination result into a PID parameter cost function constructed based on a model predictive control algorithm, and solve to obtain the optimized PID parameters of each substation, and the optimized PID parameters are sent to a command topic subscribed by each AVC substation by a communication module, and each AVC substation re-collects data until the comprehensive performance level of each AVC substation is qualified.

[0010] Preferably, the communication module initiates a connection request to each AVC substation, exchanges and verifies the digital certificate of each AVC substation to complete two-way authentication, and after verification, negotiates to generate a shared session encryption key and establishes a TLS secure channel that is authenticated and encrypted in both directions; based on the established TLS secure channel, each AVC substation acts as an MQTT subscriber to subscribe to its exclusive command topic and data feedback topic, and the communication module subscribes to the data feedback topic of each substation to receive the feedback data of each substation; after the AVC substation confirms the successful subscription of the TLS secure channel and its exclusive command topic, the AVC substation publishes a ready state message to the communication module through MQTT.

[0011] Preferably, establishing a TLS secure channel that is authenticated and encrypted in both directions comprises: The security software module activates a hardware-level firewall policy to divide an independent communication link and limit the communication direction; Based on the independent communication link and the limited communication direction, the communication module sends a TCP connection request to each AVC substation according to the VLAN identifier in the JSON test task and starts a TLS handshake protocol, and the substation and the communication module establish a TLS secure channel that is authenticated and encrypted in both directions.

[0012] Preferably, decrypting and verifying the initial test data set to obtain plaintext data comprises: The security software module generates a dynamic decryption key by fusing a hardware random number generator and dynamic parameters; The security software module decrypts the data by using the dynamic decryption key through an AES-256 algorithm to obtain decrypted plaintext data; The security software module verifies the CRC check code of the decrypted plaintext data to check the integrity of the data packet structure; After the integrity verification is passed, the security software module checks the digital certificate and access permission of the decrypted plaintext data, and after the checking is passed, the verified plaintext data is obtained, and the security software module transfers the plaintext data to the data processing library module.

[0013] The second aspect of the application provides a portable AVC automatic commissioning test method, which is run in the portable AVC automatic commissioning test device described in the first aspect, and comprises: According to the set parameter setting target and limit condition generation task request, the task request is packaged in JSON format and added with VLAN identification to determine the JSON test task; The JSON test task is sent to the command topic subscribed by each AVC substation, and the new JSON data packet filled with real-time data collected by each AVC substation according to the JSON test task is obtained from the data feedback topic of each AVC substation; The new JSON data packet is subjected to CRC check, and the initial test data set is extracted from the JSON data packet passing the check; The initial test data set is decrypted and verified to obtain plaintext data, which is preprocessed and input into the dynamic and static curve real-time discrimination model constructed based on the deep reinforcement learning network to output the dynamic and static curve discrimination result. The discrimination result is input into the PID parameter cost function constructed by the model predictive control algorithm, and the optimized PID parameters of each substation are solved and issued to the command topic subscribed by each AVC substation. Each AVC substation re-collects data until the comprehensive performance level of each AVC substation is qualified.

[0014] Preferably, determining the JSON test task comprises: When the user logs in, the user's authority is checked based on the preset authority matrix, the operation entrance matched with the authority is opened, and the encrypted authority token is generated. The user inputs the set parameter setting target and limit condition through the operation entrance matched with the authority, and generates the user instruction combined with the encrypted authority token; When the test object is multiple substations, based on the historical running data of each AVC substation, a collaborative scheduling model is constructed according to the model predictive control algorithm, with the minimum bus voltage deviation, reactive power deviation and control amount as the target, and the test control sequence of each AVC substation is solved. When the test object is a single substation, the test control sequence of the substation is set; the voltage qualified range is solved according to the voltage threshold setting algorithm, and the reactive power adjustment threshold is solved according to the reactive power output adjustment algorithm; According to the test control sequence, the voltage qualified range and the reactive power adjustment threshold, the test task parameter set is generated; The substation set in the test task parameter set is parsed, the VLAN identification is added to each substation, and the JSON test task is generated.

[0015] Preferably, the new JSON data packet filled with real-time data collected by each AVC substation according to the JSON test task comprises: Each AVC substation obtains the JSON test task from the subscribed command topic, parses the JSON test task, executes the test control sequence in the JSON test task according to the local PID controller, and collects real-time data according to the data collection item in the JSON test task; The collected data is arranged in the original sensor data format and filled into the payload field of a new JSON data packet; A CRC check code is calculated for the new JSON data packet and attached, and the data is published to the data feedback topic of each AVC substation subscribed by the communication module through the MQTT protocol.

[0016] Preferably, the plaintext data is preprocessed and input into a dynamic and static curve real-time discrimination model based on a deep reinforcement learning network to output a dynamic and static curve discrimination result, which includes: A state vector is constructed according to the preprocessed plaintext data, the state vector is used as an initial feature, the output of a deep reinforcement learning network hidden layer is calculated, a dynamic curve and a static wireless are used as an action vector, the output of the deep reinforcement learning network hidden layer is input into a deep reinforcement learning network output layer to calculate a Q value, and the dynamic and static curve is discriminated according to the Q value.

[0017] Preferably, the discrimination result is input into a PID parameter cost function constructed by a model predictive control algorithm, and the optimized PID parameters of each substation are obtained by solving, which includes: The comprehensive performance grade of each AVC substation is evaluated by calculating the voltage qualification rate, the reactive power regulation deviation rate and the regulation response time deviation; If the comprehensive performance grade is unqualified, the PID parameter cost function is constructed according to the discriminated curve type and the model predictive control algorithm; The PID parameter cost function is minimized by the gradient descent method, and the optimized PID parameters of each substation are obtained.

[0018] Preferably, the PID parameter cost function is represented by the following formula:

[0019] In the formula, The PID parameter cost function of the proportional coefficient and the integral time of the mth substation is represented by formula (1), The reactive power deviation weight of the mth substation is represented by formula (2), The response time weight of the mth substation is represented by formula (3), The voltage qualification rate weight of the mth substation is represented by formula (4), The reactive power regulation deviation rate of the mth substation is represented by formula (5), The reactive power regulation response time of the mth substation is represented by formula (6), The voltage qualification rate deviation of the mth substation is represented by formula (7).

[0020] Compared with the prior art, the beneficial effects of the present application at least include: The application program module automatically selects an adaptive algorithm for a single AVC substation or multiple AVC substations according to a test object, a single AVC substation adopts PID independent regulation and control, and multiple AVC substations generate a test control sequence through MPC collaborative scheduling, Overcome the shortcomings of traditional test method can only for single AVC substation test, for single station test and multi-station cooperative test different needs, adopt the optimal algorithm to generate the most suitable test excitation signal, improve the precision and effectiveness of AVC automatic commissioning test, ensure that the test result can truly reflect the performance of the system under test; The application adds VLAN identification to different sub-stations through the communication module to analyze the test task parameter set, realizes communication link isolation, even when testing multiple sub-stations, the communication traffic between the sub-stations is isolated, effectively prevents network problems such as broadcast storm and address conflict from spreading across stations, improves the parallel efficiency and communication reliability of multi-station testing, and enhances the stability and safety of AVC automatic commissioning test; The application adopts the publish / subscribe mode of MQTT, each AVC sub-station subscribes to a dedicated topic, and the test device distributes and collects messages, realizes accurate delivery of test instructions and parallel collection of data, decouples the test device and the sub-station, has strong device expandability, and adding a new test AVC sub-station does not need to change the core architecture, only needs to subscribe to the corresponding topic, meets the needs of large-scale AVC system commissioning test for multiple AVC sub-stations, fills the technical gap in the industry that there is no commissioning test device for AVC sub-stations, provides a reliable technical means for AVC system test work, and improves the efficiency and result determination accuracy of multi- AVC sub-station commissioning test; The application inputs the decrypted data into the dynamic and static curve discrimination model based on deep reinforcement learning through the data processing library module, automatically identifies whether the power grid connected to the current AVC sub-station is in dynamic transient state or static steady state, realizes intelligent performance index determination of each sub-station without manual calculation method, dynamically adjusts the weights of voltage, reactive power and response time indexes in the cost function according to the identified curve type, realizes accurate weighing of the control performance of the AVC sub-station, prioritizes the rapidness of the response of each AVC sub-station in the dynamic process, and prioritizes the accuracy of the control of each AVC sub-station in the static process, compared with the traditional trial-and-error method or empirical formula method, the application improves the comprehensive control performance of the AVC sub-station, and the application delivers the optimized parameters to the sub-station, re-collects data and evaluates the performance, and circulates until qualified, realizes complete automation of the parameter setting process of the AVC sub-station without manual debugging, improves the test and debugging efficiency, and ensures the high-quality and reliable final setting result. The application integrates the communication module, the data acquisition module, the user interface module, the data processing library module and the security software module into the portable device, overcomes the shortcomings of traditional test systems that need to carry multiple instruments and complex wiring, realizes single device and full coverage, and is suitable for on-site mobile and rapid test requirements. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 is a schematic diagram of the overall architecture of the portable AVC automatic commissioning test device provided according to an embodiment of the present application; Figure 2 is a schematic diagram of the signal transceiver mechanism between the portable AVC automatic commissioning test device and the multi- AVC substation provided according to an embodiment of the present application; Figure 3 is a schematic diagram of the information interaction process between the portable AVC automatic commissioning test device and the multi- AVC substation provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. The described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art without creative labor based on the spirit of the present application shall fall within the protection scope of the present application.

[0023] As shown in Figure 1 Embodiment 1 of the present application proposes a portable AVC automatic commissioning test device, which comprises a central processing unit (CPU), a communication module, a storage device module, an embedded operating system module, a data acquisition module, a power management module, a user interface module, an application program module, a communication protocol stack module, a data processing library module and a security software module.

[0024] The application program module is configured to generate a task request according to the target and the limit condition of the parameter setting of the user interface module. Preferably, the target and the limit condition of the parameter setting of the user interface module comprise: The user logs in through the user interface module, and the security software module automatically calls the preset permission matrix of the storage device module to match the user role permission, and substitutes the user role permission into a permission verification function to perform permission verification. After the verification is passed, the security software module generates an encrypted permission token, and the user interface module only opens an operation entry matching the permission. The user inputs the target and the limit condition of the parameter setting through the operation entry matching the permission, the user interface module transmits the target and the limit condition of the parameter setting to the CPU, the security software module generates an encrypted permission token and transmits it to the CPU, and the target and the limit condition of the parameter setting and the encrypted permission token generate a user instruction in the CPU.

[0025] Preferably, generating the task request comprises: The application program module receives user instructions of the central processing unit, when the test object in the user instructions is multiple sub-stations, the application program module selects the MPC cooperative scheduling algorithm to solve the test control sequence of each AVC sub-station; when the test object is a single sub-station, the application program module sets the test control sequence of the sub-station; the application program module solves the voltage qualified range according to the voltage threshold setting algorithm and solves the reactive power regulation threshold according to the reactive power regulation algorithm; the application program module generates a test task parameter set according to the test control sequence, the voltage qualified range and the reactive power regulation threshold, encapsulates the test task parameter set into an initial JSON object according to the JSON standard format, returns the CPU, and the CPU forwards the initial JSON object to the communication module; the communication module parses the sub-station set in the initial JSON object, adds a VLAN identifier for each sub-station, embeds the VLAN identifier in the JSON data packet header, and generates a JSON test task.

[0026] The communication module is configured to encapsulate the task request generated by the application program module in a JSON format and add a VLAN identifier, and determine a JSON test task. The communication module sends the JSON test task to a command topic subscribed by each AVC sub-station, and the communication module obtains real-time data collected by each AVC sub-station according to the JSON test task from a data feedback topic of each AVC sub-station. Preferably, the communication module initiates a connection request to each AVC sub-station, exchanges and verifies the digital certificate of the sub-station with each AVC sub-station to complete two-way authentication, and after verification, negotiates to generate a shared session encryption key and establishes a TLS secure channel that has been authenticated and encrypted in both directions; based on the established TLS secure channel, each AVC sub-station acts as an MQTT subscriber and subscribes to its exclusive command topic and data feedback topic, and the communication module subscribes to the data feedback topic of each sub-station to receive the return data of each sub-station; after each AVC sub-station confirms that the TLS secure channel and the exclusive command topic subscription are successful, the sub-station publishes a ready state message to the communication module through MQTT.

[0027] Further preferably, establishing a TLS secure channel that has been authenticated and encrypted in both directions comprises: The security software module activates a hardware-level firewall policy to divide an independent communication link and limit the communication direction; Based on the independent communication link and the limited communication direction, the communication module sends a TCP connection request to each AVC sub-station according to the VLAN identifier in the JSON test task and starts a TLS handshake protocol, and the sub-station and the communication module establish a TLS secure channel that has been authenticated and encrypted in both directions.

[0028] More preferably, the communication module sends a TCP connection request to each AVC sub-station according to the VLAN identifier in the JSON test task and starts a TLS handshake protocol, which comprises: After each AVC substation receives the connection request, it returns an acknowledgement signal to the communication module and prepares to conduct a TLS handshake. The substation sends its X.509 digital certificate to the communication module, and the communication module sends the device's X.509 digital certificate to the substation. The security software module and the substation respectively perform bidirectional verification on the substation digital certificate and the device digital certificate, obtaining bidirectional certificate verification results, including: The communication module passes the substation certificate to the security software module, and the device verifies the substation certificate, including: the security software module checks the validity of the CA signature of the substation X.509 digital certificate, the certificate validity period, and the test authority range (such as the device certificate allowing the test of 220 kV substations, and only receiving 220 kV substation certificates); the substation verifies the device certificate, including: checking whether the SN code in the device certificate is in the list of trusted test devices stored locally by the substation, to avoid illegal test devices from accessing; the security software module and the substation respectively record the verification results and exchange the verification status through the communication module, to obtain bidirectional certificate verification results.

[0029] After bidirectional certificate verification is passed, the security software module needs to confirm the operator's identity to prevent illegal personnel from operating legal devices, and therefore sends an authentication request to the user interface module; the user interface module pops up an authentication interface, the user inputs a password through the authentication interface, and generates a security token through a USB key or collects biological information through biometric identification (such as fingerprint); the security software module compares whether the user-input password matches the preset value, whether the security token matches the legal code generated in real time, or whether the biological information matches the pre-stored template, to obtain an authentication result.

[0030] After authentication is passed, the communication module sends the authentication pass result and the device X.509 digital certificate to the substation, and the substation completes identity verification based on the received authentication pass result and the device X.509 digital certificate, starts a TLS handshake, negotiates with the communication module to generate a shared session encryption key, and both parties use the session key to establish a TLS secure channel that is bidirectionally authenticated and encrypted.

[0031] The security software module is embedded with a key periodic rotation mechanism, which automatically updates the key every 24 hours.

[0032] After authentication fails, the user interface module pops up a window to alert, and the communication module refuses to initiate a connection.

[0033] The communication protocol stack module is used to perform CRC check on the real-time data obtained by the communication module.

[0034] Notably, the communication module performs bidirectional X.509 digital certificate verification with the substation, and establishes a TLS secure channel after combining the operator identity authentication, thereby building a double security guarantee of device authentication and personnel authentication, completely eliminating the risk of illegal device access and illegal personnel operation. The latest TLS 1.3 protocol is adopted to improve the security while accelerating the connection speed, which is very suitable for the scene of rapid deployment and testing.

[0035] Preferably, after receiving the ready state message, the communication module sends the JSON test task to the command topic subscribed by the AVC substation after adding a CRC check code through the communication protocol stack module, including: The communication module encapsulates the JSON test task according to the MQTT protocol, and delivers the MQTT data packet to the communication protocol stack module. The communication protocol stack module adds a TCP header (source / destination port, sequence number, etc.) to the MQTT data packet at the TCP layer, calculates and adds a CRC check code, and obtains an encapsulated TCP segment; the communication module sends the encapsulated TCP segment to the MQTT broker through the underlying network interface, and the MQTT broker routes the data packet to the AVC substation subscribing to the corresponding command topic.

[0036] Preferably, each AVC substation collects real-time data according to the received JSON test task by subscribing to the command topic, packages the real-time data into a JSON format and adds a CRC check code, and publishes the data packet to the data feedback topic of each AVC substation subscribed by the communication module through MQTT.

[0037] Each AVC substation obtains the JSON test task by subscribing to the command topic, parses the JSON test task, collects real-time data according to the data collection item, arranges the collected data in the original sensor data format, fills it into the payload field of a new JSON data packet, and the packet header contains information such as data type (such as sensor measured data), timestamp, and data source (substation ID); a CRC check code is calculated and added for the new JSON data packet, and the data packet is published to the data feedback topic of each AVC substation subscribed by the communication module through the MQTT protocol.

[0038] The communication module delivers the received data packet to the communication protocol stack module for CRC check; if the CRC check fails, the communication protocol stack module triggers the retransmission mechanism, requests the AVC substation to resend the data packet through the communication module, and retransmits until the check is successful. The data is marked as reliable, and if the CRC check is successful, the data collection module performs preliminary processing on the data that passes the check and delivers it to the central processing unit.

[0039] Preferably, if the check is successful, the data packet is marked as reliable, the communication protocol stack module delivers the data packet to the communication module, and the communication module delivers it to the data collection module.

[0040] If the verification fails, an error detection is triggered, an error code (such as CRC_ERROR) is generated to indicate the error cause (such as data corruption), error information (error code, timestamp, error cause, data source location) is recorded, and the error information is transmitted to the communication module to receive an error signal, and then the error information is pushed to the security software module in synchronization, the security software module writes the information into an audit log, and at the same time, a pop-up window is popped up through the user interface module to alarm, and error recording and prompting are completed; The communication module sends a data retransmission request to each AVC substation, and after the AVC substation retransmits the data, the communication module receives and again submits the data to the communication protocol stack module for verification until the verification passes; After the verification passes, the communication module transmits the data to the data acquisition module, and at the same time, feeds back to the CPU that the retransmission is successful, and terminates the retransmission process.

[0041] The data acquisition module is configured to extract an initial test data set from the JSON data packet that passes the verification of the communication protocol stack module.

[0042] Preferably, the data acquisition module receives the JSON data packet that passes the verification, parses the data packet, and extracts real-time data (such as bus voltage, unit reactive power, etc.) in the payload; according to the AVC system performance evaluation index library preset by the data acquisition module, the data is preliminarily screened to obtain the initial test data set.

[0043] The data acquisition module retains indicators such as voltage fluctuation value, reactive power regulation response time, and remote control instruction execution success rate; eliminates invalid redundant data such as device temperature; and performs data denoising preprocessing (such as applying median filtering and exponential smoothing algorithm to remove noise).

[0044] The security software module is configured to decrypt and verify the initial test data set to obtain plaintext data. Preferably, the security software module decrypts the initial test data set, verifies the integrity of the decryption result, and checks the digital certificate and access permission to determine the plaintext data that passes the verification, including: The security software module generates a dynamic decryption key by fusing a hardware random number generator (HRNG) and dynamic parameters; the security software module decrypts the data by using the dynamic decryption key through an AES-256 algorithm to obtain decrypted plaintext data; verifies the data CRC check code and checks the integrity of the data packet structure (header, payload, and checksum); after the integrity verification passes, the security software module checks the digital certificate and access permission of the decrypted plaintext data to ensure that the data is legal and secure, and obtains the plaintext data that passes the verification after the checking passes. The security software module transmits the plaintext data to the data processing library module.

[0045] The data processing library module is used for inputting the preprocessed plaintext data into a dynamic and static curve real-time discrimination model based on a deep reinforcement learning network to output a dynamic and static curve discrimination result, inputting the discrimination result into a PID parameter cost function constructed by a model predictive control algorithm to solve the optimized PID parameters of each substation, and issuing the optimized PID parameters to each AVC substation to re-collect data until the comprehensive performance level of each AVC substation is qualified, and the data processing library module generates a preliminary judgment result table according to the comprehensive performance level and transmits the table to the CPU.

[0046] The CPU is the brain of the whole system, connected with the user interface module, the data acquisition module, the communication module and the application program module, responsible for controlling and coordinating various hardware and software components, used for receiving user instructions from the user interface module and data from the data acquisition module, the communication module and the application program module and performing data processing, and sending the processed data and instructions to the application program module to coordinate the communication among the modules. The CPU uses a standard data structure JSON (JavaScript Object Notation, JSON) to communicate with each module, and formulates a custom protocol or uses a standard communication protocol as an interface protocol.

[0047] The communication module is used for coordinating communication with different AVC substations, transmitting the parameter setting results and test instructions from the CPU to each AVC substation according to the TCP protocol to implement actual parameter adjustment, and receiving test data from the AVC substation back to the central processing unit.

[0048] The storage device module usually selects a solid-state memory as a storage medium for long-term storage of data, uses a file operation protocol to store the test data, test reports and configuration files processed by the CPU into the storage device module, and the output interface provides data retrieval and backup functions to ensure data accessibility and security, providing fast data read / write speed and reliability.

[0049] The embedded operating system module is an abstraction layer between software and hardware, providing a software and hardware interface to make it easier for developers to access underlying hardware resources using application programs, and also providing some basic system services to simplify application development. The embedded operating system module receives instructions and data from application software for interacting with hardware, sends control instructions to hardware through a hardware driver interface to manage hardware resources, and ensures collaboration between modules.

[0050] The hardware part of the data acquisition module is responsible for the physical acquisition and preprocessing of data, while the software part processes, manages, and transmits data. The original sensor data format is received through the hardware sensor, and the real-time test data collected from the AVC substation, including bus voltage, unit reactive power, remote measurement, remote signaling displacement information, and remote control command pulses, are transmitted through the communication module. These data are used to evaluate the performance of the system during parameter setting. The intelligent research and judgment can be embedded in the data acquisition module to monitor the system performance in real time and transmit the data to the CPU for further analysis, ensuring the collection and transmission of various key information for effective monitoring and control.

[0051] The software part of the power management module involves power management functions in the operating system and application programs, collects data from the hardware to perceive the power state, and manages the hardware power supply to ensure the normal operation of the device. The hardware part includes power supply units, batteries, power switches, etc., which are used to control the power switch and power supply mode, and actually provide and control the power supply.

[0052] When the user inputs the target and limit conditions of parameter setting through user input devices such as keyboards, mice, and touchscreens, the user interface module provides intelligent suggestions and information display for the user, helping the operator to select appropriate performance indicators and parameter ranges. The user interface module receives user input, transmits the input to the application program module, and displays information and feedback operation results. The data format includes user interface elements, text, graphics, etc.

[0053] The application program module receives user instructions (such as test targets, limit conditions) and related data from the CPU, uses intelligent research and judgment technology, combines with AVC test scene requirements (such as single / multiple substation testing, voltage / reactive power regulation indicators), selects the best control algorithm from the preset algorithm library, and accurately controls the test parameters to match the system optimization target (such as improving test efficiency and ensuring data accuracy). The result provides the core basis for generating test instructions and supports subsequent substation test task execution.

[0054] The parameter setting algorithm is used in step 1 of the test method (the operator sets the parameter setting target and limit conditions). The test parameters configured by the parameter setting algorithm (such as voltage qualification range, reactive power regulation threshold) are cooperatively adapted with the control algorithm results obtained by intelligent research and judgment, and together form the final test instructions that meet the test requirements. After being returned to the CPU, the test is started by coordinating the modules.

[0055] The communication protocol stack module receives data from the communication module through a standard communication protocol stack, processes data packet transmission, analysis, and error handling of the underlying communication protocol, and sends the processed data back to the communication module to ensure data reliability and integrity. Intelligent research and judgment improve the communication protocol to achieve efficient data transmission and remote monitoring.

[0056] In combination with the AVC test scenarios, the communication protocol stack module optimizes the protocol adaptability and transmission efficiency through intelligent research and judgment: first, for multi-substation joint debugging, the MQTT protocol "publish / subscribe" mechanism is optimized, the topics are divided according to the priority of the substation, and the message redundancy is reduced; second, the TCP protocol handshake process is simplified, the data packet header is compressed, and the bandwidth demand of the portable device is adapted. Technically, the MQTT and JSON protocol stacks are integrated, the CRC check algorithm is embedded to handle error retransmission, and efficient transmission and remote monitoring are achieved through dynamic adjustment of protocol parameters.

[0057] Intelligent research and judgment is not limited to the application program module, but a cross-module technical means. The intelligent research and judgment of the application program module focuses on the selection of control / parameter tuning algorithms, while the intelligent research and judgment of the communication protocol stack module revolves around communication optimization: the former outputs communication performance targets to the latter based on test requirements, and the latter improves the protocol based on the targets, both of which work together to ensure efficient transmission of test data and adapt to the AVC test whole process.

[0058] The data processing library module receives the data collected by the data acquisition module and processes and analyzes the data using signal processing algorithms, data filtering, and feature extraction techniques, and performs parameter tuning calculations. The processed data is sent to the CPU, which can be used to generate test reports or passed to the communication protocol stack module.

[0059] The communication protocol stack is a special scenario - when the substation performance does not meet the standard (such as the reactive power regulation deviation exceeding the threshold), the optimized parameters after tuning need to be fed back to the substation through the protocol stack to guide the substation to adjust the control strategy and achieve closed-loop optimization.

[0060] The security software module communicates with other modules through embedded intelligent research and judgment technology and data transmission, responsible for data encryption, user authentication and permission management, to implement an anomaly detection mechanism to identify abnormal situations in system operation, protect the system from unauthorized access, generate security logs and send security status and alarm information to other modules, and take appropriate measures to ensure the security of the system and the confidentiality of the data.

[0061] Data analysis: first, remove the noise and outliers of the telemetry data (bus voltage, unit reactive power, etc.) through median filtering and exponential smoothing algorithm; then use wavelet transform to extract data features (such as voltage fluctuation frequency, reactive power regulation rate), and identify data trends through time series analysis. Parameter tuning calculation: based on the extracted feature data, call the PID parameter tuning algorithm, compare the preset performance threshold (such as voltage eligibility rate ≥ 98%), calculate the optimal regulation parameters (proportion coefficient, integral time, etc.), and output the processed data and tuning results.

[0062] This close collaboration structure enables the system to efficiently complete the automatic test task, and the user inputs instructions through a friendly interface, and the CPU interprets the instructions and controls the cooperative work of each module, so as to realize the functions of data acquisition, processing, communication and storage. The modular design of the overall architecture ensures the flexibility and maintainability of the system, so that the system can be expanded or upgraded according to the needs. The design of the overall architecture can meet the complex test requirements, while ensuring the safety and stability of the system.

[0063] Embodiment 2 of the present application provides a portable AVC automatic test method, which runs in the portable AVC automatic test system described in embodiment 1, and comprises: Step 1, generate a task request according to the set parameter setting target and limit condition, encapsulate the task request in JSON format and add VLAN identification to obtain a JSON test task.

[0064] In the preferred but non-limiting embodiment of the present application, step 1 comprises: Step 1.1, when the user logs in, the user's authority is checked based on the pre-set authority matrix, the operation entrance matched with the authority is opened, and an encrypted authority token is generated.

[0065] Further preferably, step 1.1 comprises: Pre-set authority matrix , R represents a set of user roles, ; S represents a set of substation identifications, , d represents a set of data types, ; o represents a set of operation authorities, .

[0066] For example, but not limited to, the authority matrix can be represented as follows:

[0067] Pre-set authority verification function , which is represented by the following formula:

[0068] In the formula, r represents the user role, s represents the substation identification, d represents the data type, and o represents the operation authority.

[0069] The user logs in, calls the authority matrix to match the user role authority, and substitutes the user role authority into the authority verification function for authority checking; After the check, the portable AVC automatic commissioning test device only opens the operation entrance matching the permission (such as the test operator cannot see the key configuration menu), and generates an encrypted permission token Token=AES-256(user ID, user role, token validity period) for subsequent data access check.

[0070] Step 1.2, the user inputs the target and limit conditions of the parameter setting through the operation entrance matching the permission, and generates a user instruction combined with the encrypted permission token.

[0071] Further preferably, step 1.2 includes: The user inputs the target performance indicators and limit conditions through the user input device such as keyboard, mouse and touch screen, and the target performance indicators include: bus voltage qualified rate > 98%, reactive power regulation response time ≤ 3s; the limit conditions include available communication bandwidth and data acquisition frequency.

[0072] The input target performance indicators and limit conditions and the encrypted permission token are transmitted to the CPU to generate a user instruction, and the user instruction={encrypted permission token, test object (single AVC substation / multi- AVC substation cluster), target performance indicators, limit conditions}.

[0073] It is worth noting that the security software module calls the preset permission matrix for verification, and generates an encrypted permission token after passing and only opens the corresponding operation entrance, realizes strict permission separation and the principle of least privilege, effectively prevents unauthorized operation, and guarantees the safety and compliance of the test operation from the source. System administrator, test operator, auditor, each plays its own role, avoids the risk of misoperation and malicious operation, and improves the safety of the AVC substation automatic commissioning process.

[0074] Step 1.3, when the test object is multiple substation, select MPC cooperative scheduling algorithm, solve the test control sequence of each AVC substation; when the test object is a single substation, select PID independent control algorithm, set the test control sequence of the substation.

[0075] Further preferably, step 1.3 includes: Step 1.3.1, when the test object is multiple substation, based on the historical running data of each AVC substation, taking the minimum bus voltage deviation, reactive power deviation and control amount as the target, according to the model predictive control algorithm, a cooperative scheduling model is constructed.

[0076] More preferably, step 1.3.1 includes: The objective function of the cooperative scheduling model is expressed as follows:

[0077] In the formula, N represents the number of substation, denotes the total sampling time step, denotes the bus voltage deviation weight, denotes the reactive power deviation weight, denotes the control quantity smoothing weight, denotes the actual bus voltage of the mth substation at the nth time step, denotes the bus voltage reference value of the mth substation at the nth time step, denotes the actual unit reactive power output of the mth substation at the nth time step, denotes the reactive power output reference value of the mth substation at the nth time step, denotes the control quantity of the mth substation at the nth time step and the increment of the control quantity at the n-1th time step, .

[0078] The constraint conditions include reactive power constraints, control quantity rate of change constraints, bus voltage deviation constraints, and communication delay constraints.

[0079] The reactive power constraints are expressed by the following formula:

[0080] In the formula, denotes the test reactive power signal value of the mth substation at the nth time step, denotes the minimum reactive power output capacity of the mth substation, denotes the maximum reactive power output capacity of the mth substation.

[0081] The bus voltage deviation constraints are expressed by the following formula:

[0082] In the formula, denotes the test bus voltage signal value of the mth substation at the nth time step, denotes the bus voltage reference value of the mth substation at the nth time step, denotes the maximum allowed bus voltage deviation.

[0083] The control quantity rate of change constraints are expressed by the following formula:

[0084] In the formula, , denotes the maximum increment of the bus voltage control quantity of the mth substation at the nth sampling time step, denotes the maximum rate of change of the control quantity of the mth substation.

[0085] The communication delay constraints are expressed by the following formula:​

[0086] wherein, denotes the nth sampling time step, denotes the sampling period, denotes the mth substation communication delay, which must be ensured from the current time step to that time step when sending the control instruction at the nth sampling time step is greater than the communication delay .

[0087] solving the coordinated dispatching model obtains the test control sequence of the plurality of substations , including the optimal bus voltage and the optimal reactive power output, and is expressed in the following formula:

[0088] wherein, denotes the optimal bus voltage of the mth substation at the kth sampling time step, denotes the optimal reactive power output of the mth substation at the kth sampling time step.

[0089] Step 1.3.2, when the test object is a single substation, a PID independent control algorithm is selected, and the test control sequence of the single substation is set , and is expressed in the following formula:

[0090] wherein, denotes the bus voltage at the kth sampling time step, denotes the reactive power output at the kth sampling time step, denotes the reactive power output at the kth sampling time step, denotes the kth sampling time step.

[0091] Step 1.4, according to the voltage threshold setting algorithm, the voltage qualified range is solved, and according to the reactive power output adjustment algorithm, the reactive power adjustment threshold is solved.

[0092] It is worth noting that the present application automatically solves the voltage qualified range and the reactive power adjustment threshold according to the voltage threshold setting algorithm and the reactive power output adjustment algorithm, rather than relying on artificial experience to set fixed values, so that the test standard (threshold) can adapt to the actual operation characteristics and power grid state of different substations, and the performance evaluation is more scientific, accurate and fair.

[0093] ​​​Step 1.5. Generating a test task parameter set according to the test control sequence, the voltage qualified range and the reactive power regulation threshold: {substation communication protocol: MQTT; data acquisition items: bus voltage, unit reactive power, remote signaling variable position information, remote control instruction pulse; parameter threshold: bus voltage qualified range, reactive power regulation threshold; test control sequence}, and encapsulating the test task parameter set into an initial JSON object in the JSON standard format: {task identifier; test task parameter set; substation set}.

[0094] The data acquisition items are not completely manually inputted by the user, and specifically include: In combination with the historical test database and the AVC test standard, the data acquisition items, such as bus voltage, unit reactive power, remote signaling variable position information and remote control instruction pulse, are automatically supplemented.

[0095] Step 1.6. Analyzing the substation set in the test task parameter set, adding a VLAN identifier to different substations, such as binding VLAN 10 to the 110kV substation and binding VLAN 20 to the 220kV substation, embedding the identifier information into the JSON data packet header, and generating a JSON test task.

[0096] Step 2. Establishing a TLS secure channel with each AVC substation, and subscribing to the data feedback topics of each AVC substation based on the TLS secure channel.

[0097] In the preferred but non-limiting embodiments of the application, step 2 includes: Step 2.1. Initiating a connection request to each AVC substation, exchanging and verifying the substation digital certificate with each AVC substation to complete the two-way identity authentication, and after the verification is passed, negotiating to generate a shared session encryption key, and establishing a TLS secure channel that has been authenticated and encrypted in both directions.

[0098] Further preferably, step 2.1 includes: Step 2.1.1. Activating a hardware-level firewall policy to divide and limit the communication direction of the independent communication link with each AVC substation.

[0099] More preferably, step 2.1.1 includes: Dividing the independent communication link according to the VLAN identifier, and prohibiting data interaction between different VLANs (such as the VLAN 10 where the 1# substation is located cannot directly access the VLAN 20 where the 3# substation is located); Only open test dedicated port (TCP1883, MQTT protocol port), close HTTP (80 port), SSH (22 port) and other redundant ports, and limit the communication direction to portable AVC automatic test device to substation for command issuance and substation to portable AVC automatic test device for data return, prohibit the substation from initiating non-test connection to the portable AVC automatic test device.

[0100] Step 2.1.2, based on independent communication link and limited communication direction, sends a TCP connection request to each AVC substation according to the VLAN identifier in the JSON test task and starts the TLS handshake protocol, and establishes a TLS secure channel with each AVC substation through bidirectional authentication and encryption.

[0101] Further preferably, step 2.1.2 includes: Based on independent communication link and limited communication direction, send a TCP connection request to each AVC substation according to the VLAN identifier in the JSON test task and start the TLS handshake protocol, and after the AVC substation receives the connection request, return an acknowledgement signal and prepare for TLS handshake, and the substation sends its X.509 digital certificate to the portable AVC automatic test device, and the portable AVC automatic test device sends the X.509 digital certificate of the portable AVC automatic test device to the substation.

[0102] Each AVC substation performs substation digital certificate verification, and each AVC substation performs bidirectional certificate verification on the device digital certificate sent by the portable AVC automatic test device, including: The substation certificate verification includes: verifying the validity of the CA signature of the substation X.509 digital certificate, the certificate validity period, and the test authority range (such as the device certificate allowing the test of 220kV substation, and only receiving 220kV substation certificate); The substation verifies the device certificate, including: checking whether the SN code in the device certificate is in the trusted test device list stored locally by the substation, to avoid illegal test device access; The portable AVC automatic test device and the substation record and exchange the verification results respectively to obtain the bidirectional certificate verification results.

[0103] After the bidirectional certificate verification is passed, the operator identity needs to be confirmed to prevent illegal personnel from operating the legal portable AVC automatic test device, so the authentication request is triggered; the authentication interface is popped up, the user inputs the password through the authentication interface, and generates a security token through the USB key or collects biological information through biological recognition (such as fingerprint); compare whether the user input password matches the preset value, whether the security token matches the legal code generated in real time, or whether the biological information matches the pre-stored template to obtain the authentication result.

[0104] After the authentication is passed, the authentication passing result and the portable AVC automatic commissioning test device X.509 digital certificate are sent to the substation, the substation completes identity verification according to the received authentication passing result and the portable AVC automatic commissioning test device X.509 digital certificate, starts a TLS handshake, negotiates a shared session encryption key with the portable AVC automatic commissioning test device, and establishes a TLS secure channel that is authenticated and encrypted in both directions by using the session key.

[0105] The embedded key is periodically rotated, and the key is automatically updated every 24 hours.

[0106] After the authentication fails, a pop-up window is used to alarm and refuse the connection initiated by each substation.

[0107] Compared with the commonly used TLS1.2 and below versions of the existing AVC system, the latest TLS1.3 transmission layer security protocol is adopted in the application, the handshake step is reduced (from 2-RTT to 1-RTT) to adapt to the fast connection requirement of the field test, and the weak encryption suite such as SHA-1 and RC4 is disabled, only the strong encryption suite such as AES-GCM and ChaCha20-Poly1305 is reserved, and the anti-hijacking ability is improved.

[0108] Step 2.2, based on the TLS secure channel established in step 2.1, each AVC substation is an MQTT subscriber, and subscribes to its exclusive command topic (such as AVC / Command / Substation_01) and data feedback topic, and the portable AVC automatic commissioning test device subscribes to the data feedback topics of all substations to receive the returned data.

[0109] Step 2.3, after the AVC substation confirms that the TLS secure channel and the topic subscription are successful, the ready state message is published to the portable AVC automatic commissioning test device through MQTT.

[0110] As Figure 2As shown, the core part of the signal transceiver mechanism is request and response, and MQTT is used as the communication protocol. MQTT uses the publish / subscribe model, which allows the portable AVC automatic debugging test device to act as a publisher and publish data to a specific topic, while the AVC substation acts as a subscriber to subscribe to the topic of interest. This model supports real-time information interaction and has a topic identifier to ensure that data packets are delivered to the correct AVC substation. Standard data structure JSON is used as the data format to create well-structured data packets. The portable AVC automatic debugging test device generates collected sensor data and packages it into JSON format data packets. These data packets are published to the relevant topic through MQTT and establish a real-time connection with the subscriber AVC substation. The AVC substation receives the data packet, parses the JSON data to obtain information, and then performs appropriate operations to feedback data, which can also be sent back to the portable AVC automatic debugging test device using the same protocol and format for subsequent decision-making and operation.

[0111] The signal transceiver mechanism provides an efficient and reliable way of information interaction, high scalability and flexibility, while allowing adaptation to different data types and needs. Through the combination of MQTT and JSON, the integrity and real-time nature of the data are ensured, providing a solid foundation for system debugging and testing.

[0112] Step 3, send the JSON test task to the AVC substation; the AVC substation collects real-time data according to the received JSON test task, packages the real-time data into data packets, and publishes them to the data feedback topic of each AVC substation subscribed to. The portable AVC automatic debugging test device performs CRC check on the received data packets, and passes the data packets that pass the check to the central processing unit, such as Figure 3 as shown.

[0113] In a preferred but non-limiting embodiment of the present application, step 3 includes: Step 3.1, after the portable AVC automatic debugging test device receives the ready state message, the JSON test task is attached with a CRC check code, and then sent to the command topic subscribed by the AVC substation.

[0114] Further preferably, step 3.1 includes: Step 3.1.1, the portable AVC automatic debugging test device encapsulates the JSON test task according to the MQTT protocol, adds a TCP header (source / destination port, sequence number, etc.) to the MQTT data packet at the TCP layer, calculates and attaches a CRC check code, and obtains the encapsulated TCP segment.

[0115] Step 3.1.2, the portable AVC automatic commissioning test device sends the encapsulated TCP packet segment to the MQTT broker through the underlying network interface, and the MQTT broker routes the data packet to the AVC substation subscribing to the corresponding topic.

[0116] Step 3.2, each AVC substation obtains a JSON test task from the subscribed topic, executes the test control sequence in the JSON test task according to the local PID controller, collects real-time data according to the data collection items in the JSON test task, packages the data into JSON format and adds a CRC check code, and publishes the data to the data feedback topic subscribed by the portable AVC automatic commissioning test device.

[0117] Further preferably, step 3.2 includes: Step 3.2.1, each AVC substation obtains a JSON test task from the subscribed command topic, parses the JSON test task, executes the corresponding bus voltage setting and reactive power output adjustment operation in the test control sequence in the JSON test task according to the local PID controller, and collects real-time data according to the data collection items in the JSON test task.

[0118] Each AVC substation executes the corresponding bus voltage setting and reactive power output adjustment operation at the specified time step according to the test control sequence, and synchronously collects actual response data.

[0119] Step 3.2.2, the collected data is arranged in the original sensor data format and filled into the payload field of the new JSON data packet, and the packet header includes data type (such as sensor measured data), timestamp, data source (substation ID) and other information; Step 3.2.3, calculate and attach a CRC check code for the new JSON data packet, and publish the data packet to the data feedback topic of each AVC substation subscribed by the communication module through the MQTT protocol.

[0120] Step 3.3, after the portable AVC automatic commissioning test device receives the data packet, it performs CRC check.

[0121] Further preferably, step 3.3 includes: Step 3.3.1, receive the JSON data packet published by each AVC substation from the data feedback topic; Step 3.3.2, extract the CRC check code of the data packet for check calculation to determine whether the data transmission is complete and correct, if the check is successful, if the data is damaged or incomplete, the check fails; Step 3.3.3, if the check fails, trigger error detection and request data retransmission from the substation until the check is passed.

[0122] More preferably, step 3.3.3 includes: If the verification is successful, the data packet is marked as reliable; If the verification fails, error detection is triggered, an error code (such as CRC_ERROR) is generated to indicate the error cause (such as data corruption), error information (error code, timestamp, error cause, data source location) is recorded, the error information is written into the audit log, a pop-up window is alarmed, and error recording and prompting are completed; A data retransmission request is sent to each AVC substation. After each AVC substation retransmits the data, the portable AVC automatic commissioning test device receives and re-CRC verifies the data until the verification is passed; After the verification is passed, the retransmission process is terminated.

[0123] Step 3.3.4, implement a timing mechanism (such as a heartbeat signal and a timeout monitoring), and if a response is not received within a maximum response time threshold, it is considered as a timeout.

[0124] Step 3.4, the portable AVC automatic commissioning test device receives the JSON data packet that passes the verification, parses the data packet, extracts real-time data (such as bus voltage, unit reactive power, etc.) in the payload, and preliminarily filters the data according to a preset AVC system performance evaluation index library to obtain an initial test data set.

[0125] The preliminary filtering includes retaining voltage fluctuation values, reactive power regulation response times, remote control instruction execution success rates, and the like; eliminating invalid redundant data such as device temperature; and performing data denoising preprocessing (such as applying median filtering and exponential smoothing algorithm to remove noise).

[0126] Step 4, the portable AVC automatic commissioning test device decrypts the initial test data set and inputs the dynamic and static curve real-time discrimination model constructed based on a deep reinforcement learning network to output a dynamic and static curve discrimination result. The dynamic and static curve discrimination result is input into a PID parameter cost function constructed by a model predictive control algorithm to solve the optimized PID parameters of each substation, which are then issued to each AVC substation to reacquire data until the comprehensive performance level of each AVC substation is qualified.

[0127] In a preferred but non-limiting embodiment of the present application, step 4 includes: Step 4.1, decrypt the initial test data set, verify the integrity of the decryption result, and check the digital certificate and access permission to determine the plaintext data that passes the verification.

[0128] Further preferably, step 4.1 includes: Step 4.1.1, a dynamic decryption key is generated by fusing a hardware random number generator HRNG and dynamic parameters.

[0129] More preferably, step 4.1.1 includes: Call the hardware random number generator HRNG to generate a basic random number; Fusion real-time dynamic parameters (such as current timestamp, device hardware serial number, substation IP check value), generate the final decryption key through SHA-256 hash algorithm; If the preset period (such as 7 days) is reached, the portable AVC automatic debugging test device generates a new key, sends a key rotation request and a new key request to the substation through the existing TLS channel, and the substation receives and replaces the old key and feeds back the confirmation. The portable AVC automatic debugging test device updates the key library according to the confirmation information, without the need to re-negotiate the connection, only to replace the key value.

[0130] Step 4.1.2, use the dynamic decryption key to decrypt the data through the AES-256 algorithm to obtain the decrypted plaintext data; verify the data CRC check code and check the integrity of the data packet structure (header, payload, checksum).

[0131] Step 4.1.3, after the integrity verification, check the digital certificate and access permission of the decrypted plaintext data to ensure that the data is legal and safe, and obtain the verified plaintext data.

[0132] Checking the digital certificate of the decrypted plaintext data includes: verifying the certificate signature (CA validity), validity period, and permission range (such as substation test permission).

[0133] Checking the access permission of the decrypted plaintext data includes comparing the substation ID with the preset authorized list and checking the user permission token to prevent unauthorized access. If the certificate is invalid or the permission does not match, a security alarm is triggered and the data is discarded.

[0134] Step 4.2, input the verified plaintext data into the dynamic and static curve real-time discrimination model based on deep reinforcement learning network to output dynamic and static curve discrimination results, and automatically optimize PID parameters through gradient descent method according to the dynamic and static curve discrimination results until the comprehensive performance level of the AVC substation is qualified, realizing the portable AVC automatic debugging test.

[0135] Further preferably, step 4.2 includes: Step 4.2.1, pre-process the plaintext data: remove outliers and noise (use median filtering, exponential smoothing algorithm), normalize, and standardize.

[0136] More preferably, step 4.2.1 includes: The plaintext data is removed from outliers, noise and other interference. Due to measurement error, equipment failure or other factors, outliers are detected based on statistical methods, clustering methods or machine learning algorithms; Smooth the noise in the data by using moving average, exponential smoothing and median filtering and other smoothing algorithms; Estimate missing values according to the trend of existing data by interpolation algorithm.

[0137] The preprocessed data is normalized and standardized to eliminate the difference between different orders of magnitude data and ensure the comparability of the data.

[0138] Step 4.2.2. Input the preprocessed data into the dynamic and static curve real-time discrimination model constructed based on the deep reinforcement learning network, and output the dynamic and static curve discrimination results, including: According to the preprocessed plaintext data, a state vector is constructed, which is expressed by the following formula:

[0139] In the formula, represents the state vector, represents the preprocessed voltage value of the i-th sampling point, represents the preprocessed reactive power value of the i-th sampling point, L represents the length of the time window, t represents the current time point index, and L≤t≤T.

[0140] Taking the state vector as the initial feature, the output of the DQN hidden layer is calculated, which is expressed by the following formula:

[0141] In the formula, represents the output of the i-th hidden layer, represents the output of the i-th hidden layer, represents the weight matrix of the i-th hidden layer, represents the bias vector of the i-th hidden layer, represents the ReLU activation function. The Q value of the DQN output layer is calculated , which is expressed by the following formula:

[0142] In the formula, represents the Q value vector,

[0143] represents the action, , represents the static curve, represents the dynamic curve, represents the output layer weight matrix, represents the output layer bias vector.

[0144] ​​The static and dynamic curve discrimination is expressed by the following formula:

[0145] In the formula, represents the type of the discriminated curve, is discriminated as static, is discriminated as dynamic.

[0146] DQN discriminant result corresponding to the power system state Table 1

[0147] Step 4.2.3, the comprehensive performance level of each AVC substation is evaluated by calculating the voltage qualification rate, reactive power regulation deviation rate and regulation response time deviation, and the PID parameter cost function is constructed based on the discriminated curve type through the model predictive control algorithm to obtain the optimized PID parameters of each substation until the comprehensive performance level of the AVC substation is qualified.

[0148] Voltage qualification rate = (voltage qualification time / total test time) x 100% (voltage qualification threshold 98%); Reactive power regulation deviation rate = | (actual value - target value) / target value | x 100% (qualified threshold ± 5%); Regulation response time deviation = difference between instruction issuance and regulation completion timestamp (qualified threshold ≤ 3 seconds).

[0149] According to the voltage qualification rate, reactive power regulation deviation rate and regulation response time deviation, the comprehensive performance level Perf is generated, which is expressed by the following formula: (15) In the formula, represents the voltage qualification rate margin, , represents the voltage qualification rate, represents the voltage qualification threshold, represents the reactive power regulation overrun margin, , represents the reactive power deviation rate threshold, represents the response time overrun margin, , represents the response time threshold, A represents the excellent level threshold, B represents the good level threshold, B > A, which can be adjusted through the user interface module.

[0151] If the comprehensive performance level of the substation is excellent or good, it is marked as qualified in the user interface module; If the comprehensive performance level of the substation is unqualified, the user interface module marks the unqualified, and constructs a PID parameter cost function according to the identified curve type and model predictive control algorithm , which is expressed by the following formula:

[0152] In the formula, δm represents the reactive power regulation deviation rate of the mth substation, which is expressed by the following formula:

[0153] In the formula, A represents the maximum deviation base, represents the proportional gain inhibition coefficient, represents the proportional coefficient of the mth substation, represents the integral time sensitive coefficient, represents the integral time of the mth substation; represents the reactive power regulation response time of the mth substation, D represents the system response constant, represents the voltage qualification rate deviation of the mth substation, which is expressed by the following formula:

[0154] In the formula, represents the steady-state error base, represents the integral time improvement coefficient, represents the proportional gain improvement coefficient, represents the reactive power deviation weight of the mth substation, represents the response time weight of the mth substation, represents the voltage qualification rate weight of the mth substation, when the curve type is a dynamic curve, which is expressed by the following formula:

[0155] represents the response time importance weight when the load has no mutation, represents the sensitivity coefficient of the load change rate, represents the load change rate, The greater the response time is, The greater the response time is, and the response time is preferentially guaranteed; represents the voltage qualification rate importance weight when the load has no mutation, represents the attenuation coefficient of the load change rate, and the secondary target is voltage stability, avoiding being negative.

[0156] when the curve type is a static curve, which is expressed by the following formula: ​

[0157] represents the importance weight of response time when voltage is qualified, represents the decay coefficient of voltage qualification rate; represents the importance weight of voltage when voltage is qualified, represents the compensation coefficient of voltage deviation.

[0158] Minimizing the PID parameter cost function by gradient descent method , which is expressed as follows:

[0159] In the formula, represents the learning rate, represents the optimized proportional coefficient; represents the optimized integral time; represents the proportional coefficient adjustment factor, represents the integral time adjustment factor.

[0160] The data processing library module solves the optimization PID parameters of each comprehensive performance level unqualified substation, combines the optimization PID parameters of each comprehensive performance level unqualified substation and the identifiers of each comprehensive performance level unqualified substation with the test task parameter set, and encapsulates the initial JSON object in the communication module according to the JSON standard format, and then issues it to the command topic of the corresponding substation subscribed by the communication module. According to the received instruction set, each AVC substation updates the parameters of the local PID controller, adjusts the reactive power through the PID controller, reacquires data, and sends it to the data feedback topic of the corresponding AVC substation subscribed by the communication module. Repeat the solving process of the comprehensive performance level until the comprehensive performance level of all AVC substations is qualified.

[0161] The data processing library module generates a preliminary judgment result table according to the comprehensive performance level of each AVC substation and transmits it to the CPU.

[0162] Step 5, according to the preprocessed data set of step 4, dynamically generate a customized test report and encrypt and archive it to realize automatic integration and test.

[0163] Step 5.1, the operator selects or customizes the style, chart type and content structure of the report template through the user interface module, and the application module matches the preprocessed data set of step 4 with the placeholders in the template to obtain the matching data features.

[0164] Further preferably, step 5.1 includes: Step 5.1.1, the operator selects or customizes the style, chart type and content structure of the test report template through the user interface module, and obtains the test report template.

[0165] Step 5.1.2, the pre-processed data in step 4 is matched with the test report template in step 5.1.1 by the template engine. According to the placeholders and rules in the test report template, the pre-processed data is filled into the corresponding positions of the test report template, and the style and format are applied. Data charts are generated according to user needs.

[0166] The template engine is included in the application module. The template engine matches the "preliminary judgment result table" with the "performance indicator module" of the report template: Automatically fill in the judgment results, and mark the abnormal reasons (such as "substation adjustment response delay") for unqualified indicators (such as voltage qualification rate 95%); When multi-substation joint debugging, generate "substation judgment result comparison table", highlight the performance differences of each substation.

[0167] Step 5.2, the application module drives the report engine to automatically convert data results into descriptive text using natural language generation (NLG) technology, and generates charts, finally rendering a complete draft of the report.

[0168] According to user needs, based on the characteristics and analysis results of the data, using NLG technology, automatically selecting expression methods according to user's custom needs, automatically generating report content, including report title, abstract, data table, chart and explanatory text, according to the information interaction process and signal transceiver mechanism between different AVC substations, providing detailed performance and interoperability description to ensure that users can clearly understand and analyze the content of the report. Provide a deeper understanding for users. Adjust the style of the report through style rules, including font, color, layout and other aspects, generate the final test report.

[0169] The application module uses natural language generation (NLG) technology: Convert the judgment results into descriptive text (such as "2# substation reactive power regulation accuracy 6.2%, exceeding the ±5% threshold, suggest checking the excitation regulation module"); Automatically generate "regulation response time trend chart" and "multi-substation voltage qualification rate comparison chart" to intuitively present the judgment results.

[0170] Step 5.3, after the user confirms the report content, the application module issues instructions, and the storage device module exports the draft of the report to standard formats such as PDF and stores it in the solid state hard disk for long-term archival management.

[0171] In order to facilitate user operation, a user-friendly interface is designed according to the behavior, preference and historical data of the user, interaction and customization options in the report generation process are realized, suitable report content or style is recommended, and the user can select the style, layout and content of the system recommended or redefined report and configuration parameters according to the needs, so as to meet different test requirements.

[0172] 1) Comprehensive audit log construction Log collection: the security software module collects key information in the whole process in real time to form three types of core logs (including time stamp, device / user identification): Access log: user login / logout, data access (such as reading remote control data of a substation), permission operation record; Communication log: TLS handshake result, certificate verification status, data transmission CRC check condition; Security event log: permission interception, certificate expiration / cancellation, abnormal connection attempt.

[0173] Storage protection: when the report is archived in step 5.3, the log is automatically packaged by day, encrypted by AES-256, and stored in the "tamper-proof partition" of the storage device module (based on hash value tamper-proofing), and only the administrator can export and view.

[0174] (2) Security assessment and vulnerability repair Regular automatic assessment: the security software module automatically starts the assessment on the 1st day of each month, checks three core indicators based on the security baseline of the power industry (such as "Provisions for the Security Protection of Electric Power Secondary Systems"), and generates a concise assessment report and archives it with the test report.

[0175] Vulnerability handling: if the indicators are not up to standard (such as certificate expiration), the user interface module immediately pops up a window alarm and pushes a repair suggestion (such as "update XX substation CA certificate"); after repair, the system automatically records the results and synchronously optimizes the monitoring strategy (such as advancing the certificate expiration reminder time).

[0176] Combine the dynamic customized test report template generation technology with the overall process of the automatic integration and test device to realize the automatic test and report generation process.

[0177] Export the generated test report into common file formats such as PDF, Word, etc., for easy user saving and sharing.

[0178] Embodiment 3 of the present application provides a portable AVC automatic integration and test process, I, user setting content (test engineer operation) The test engineer sets the test task through the "user interface module" of the device: Test object: 3# AVC substation (10kV bus) Core test indicators: ① Bus voltage qualified rate ≥ 98%; ② Unit reactive power regulation response time ≤ 3 seconds; ③ Remote control command execution success rate = 100% Test duration: 10 minutes (sampling frequency 1 time / second) Report requirements: Include "Substation Performance Data Table" and "Voltage Time Sequence Line Chart", highlight unqualified indicators and cause analysis II. Device internal processing and JSON data packet generation Application module processing: After receiving user settings, select "single station independent control algorithm" and "voltage threshold setting algorithm" through intelligent research and judgment technology, and convert test parameters into structured task instructions; Communication protocol stack module encapsulates JSON: generates data packets according to MQTT protocol requirements Transmission trigger: The communication module publishes the JSON data packet to the topic "AVC / Test / 3#Substation" through the MQTT protocol, and the 3# AVC substation has subscribed to this topic.

[0179] III. Substation data acquisition and returned JSON data packet Substation task execution: After receiving the JSON instruction, start the self-sensor and data acquisition module, collect 10 minutes of data at a frequency of 1 time / second, and select key data (example: the first 5 groups are intercepted); Substation generates returned JSON: Encapsulate the measured data according to the device requirements format and return to the topic IV. Device generates test report (user perspective) Data preprocessing: After receiving the returned JSON, the device communication module eliminates 1 group of abnormal data (voltage 9.4kV at 09:30:02 is a transient fluctuation) by the data acquisition module, and recalculates the voltage qualified rate as 97.2%; NLG technology generates report: Application module calls template engine and NLG algorithm to generate report according to user requirements: Title: "3# AVC Substation 10kV Bus Joint Test Report (2024-06-15)" Abstract: "This test covers 3# substation 10-minute running data, bus voltage qualified rate 97.2% (unqualified, 0.8% away from threshold), reactive power regulation response time 2.8 seconds (qualified), remote control command execution success rate 100% (qualified); voltage unqualified is caused by 1 time transient fluctuation, substation overall regulation performance is good." Data table: "3# Substation Core Indicator Measurement Table" (columns: indicator name, threshold, measured value, determination result) Chart: Bus voltage time sequence line chart (marking fluctuation point 09:30:02) User operation: the user views the report in the interface module, can click "export PDF" to save, or select "reanalysis" (such as eliminating more abnormal data), and the system automatically updates the report content.

[0180] Compared with the prior art, the present application has at least the following beneficial effects: The application adopts an application module to automatically select an adaptive algorithm according to a test object being a single AVC substation or multiple AVC substations, the single AVC substation adopts PID independent regulation and control, and the multiple AVC substations generate a test control sequence through MPC collaborative scheduling, The application overcomes the shortcoming of the conventional test method that can only test a single AVC substation, adopts an optimal algorithm to generate a most suitable test excitation signal according to different requirements of single-station testing and multi-station collaborative testing, improves the accuracy and effectiveness of AVC automatic commissioning testing, and ensures that the test result can truly reflect the performance of the tested system; The application adds VLAN identification for different substations through a communication module to analyze test task parameter sets, realizes communication link isolation, even when multiple substations are tested, the communication traffic between the substations is isolated, effectively prevents network problems such as broadcast storm and address conflict from being transmitted across stations, improves the parallel efficiency and communication reliability of multi-station testing, and enhances the stability and safety of AVC automatic commissioning testing; The application adopts the publish / subscribe mode of MQTT, each AVC substation subscribes to a dedicated topic, and a test device performs message distribution and collection, realizes accurate delivery of test instructions and parallel collection of data, decouples the test device and the substation, has strong device expansion, newly added test AVC substations do not need to change the core architecture, only need to subscribe to the corresponding topic, meets the demand of large-scale AVC system commissioning testing for multiple AVC substations, fills the technical gap in the industry that there is no commissioning testing device for AVC substations, provides a reliable technical means for AVC system testing work, and improves the efficiency and result determination accuracy of multi- AVC substation commissioning testing; The application inputs the decrypted data into a dynamic and static curve discrimination model based on deep reinforcement learning through a data processing library module, automatically identifies whether the power grid connected with the current AVC substation is in a dynamic transient state or a static steady state, realizes the intelligentization of the performance index judgment of each substation, and does not need to determine the performance index by an artificial calculation method; according to the discriminated curve type, the weights of the voltage, the reactive power, the response time and other indexes in the cost function are dynamically adjusted, the accurate trade-off of the control performance of the AVC substation is realized, the rapidity of the response of each AVC substation is preferentially ensured in the dynamic process, the accuracy of the control of each AVC substation is preferentially ensured in the static process, compared with the traditional artificial trial and error method or the empirical formula method, the comprehensive control performance of the AVC substation is improved; the optimized parameters are issued to the substation, the data is re-collected and the performance is evaluated, and the cycle is continued until the qualification is achieved, the complete automation of the parameter setting process of the AVC substation is realized, the artificial debugging is not needed, the efficiency of the test and debugging is improved, and the high-quality and reliable final setting result is ensured. The application integrates the communication module, the data acquisition module, the user interface module, the data processing library module and the security software module and other modules in the portable device, overcomes the shortcomings of the traditional test system that needs to carry multiple instruments and the wiring is complex, realizes single device and full coverage, and is suitable for the on-site mobile and rapid test demand.

[0181] The present disclosure can be a system, a method, and / or a computer program product. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present disclosure.

[0182] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application rather than limit it, although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced equivalently without departing from the spirit and scope of the present application, and any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. A portable AVC automatic commissioning test device comprising a communication module, a data acquisition module, a user interface module, an application module, a communication protocol stack module, a security software module and a data processing library module, characterized in that: the application module is configured to generate a task request according to the target and limit conditions set by the user interface module; the communication module is configured to encapsulate the task request generated by the application module in JSON format and add a VLAN identifier to determine a JSON test task; the communication module sends the JSON test task to the command topic subscribed by each AVC substation, and obtains a new JSON data packet filled with real-time data collected by each AVC substation according to the JSON test task from the data feedback topic of each AVC substation; the communication protocol stack module is configured to perform CRC check on the new JSON data packet obtained by the communication module; the data acquisition module is configured to extract an initial test data set from the new JSON data packet that passes the check of the communication protocol stack module; the security software module is configured to decrypt and verify the initial test data set to obtain plaintext data; the data processing library module is configured to input the plaintext data into a dynamic and static curve real-time discrimination model constructed based on a deep reinforcement learning network to output a dynamic and static curve discrimination result, input the discrimination result into a PID parameter cost function constructed by a model predictive control algorithm to solve and obtain optimized PID parameters of each substation, and send the optimized PID parameters to the command topic subscribed by each AVC substation through the communication module, so that each AVC substation re-collects data until the comprehensive performance level of each AVC substation is qualified. 2.The portable AVC automatic commissioning test device according to claim 1, characterized in that: the communication module initiates a connection request to each AVC substation, exchanges and verifies a digital certificate of each AVC substation to complete a two-way identity authentication, negotiates to generate a shared session encryption key after the two-way identity authentication is passed, and establishes a TLS secure channel that is authenticated and encrypted in both directions. Based on the established TLS secure channel, each AVC substation acts as an MQTT subscriber to subscribe to its exclusive command topic and data feedback topic, the communication module subscribes to the data feedback topic of each substation to receive the backhaul data of each substation, and each AVC substation confirms the successful subscription of the TLS secure channel and its exclusive command topic, and publishes a ready state message to the communication module through MQTT. 3.The portable AVC automatic commissioning test device according to claim 2, characterized in that: the establishment of the TLS secure channel that is authenticated and encrypted in both directions comprises: the security software module activates a hardware-level firewall policy to divide an independent communication link and limit a communication direction; based on the independent communication link and the limited communication direction, the communication module sends a TCP connection request to each AVC substation according to the VLAN identifier in the JSON test task and starts a TLS handshake protocol, and the substation and the communication module establish the TLS secure channel that is authenticated and encrypted in both directions. 4.The portable AVC automatic commissioning test device according to claim 1, characterized in that: the decryption and verification of the initial test data set to obtain plaintext data comprises: ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ The security software module generates a dynamic decryption key by fusing a hardware random number generator and dynamic parameters; The security software module decrypts the data by using the dynamic decryption key through the AES-256 algorithm to obtain decrypted plaintext data; The security software module verifies the CRC check code of the decrypted plaintext data to check the integrity of the data packet structure; After the integrity verification passes, the security software module checks the digital certificate and access permission of the decrypted plaintext data, and obtains the verified plaintext data after the verification passes. The security software module transmits the plaintext data to the data processing library module.

5. A portable AVC automatic commissioning test method, running on a portable AVC automatic commissioning test device according to any one of claims 1-4, characterized in that: According to the set parameter setting target and the limit condition, a task request is generated, the task request is packaged in JSON format and added with a VLAN identifier to determine the JSON test task; The JSON test task is sent to the command topic subscribed by each AVC substation, and a new JSON data packet filled with real-time data collected by each AVC substation according to the JSON test task is obtained from the data feedback topic of each AVC substation; The new JSON data packet is subjected to CRC check, and the initial test data set is extracted from the JSON data packet that passes the check; The initial test data set is decrypted and verified to obtain plaintext data, which is preprocessed and input into a dynamic and static curve real-time discrimination model based on a deep reinforcement learning network to output a dynamic and static curve discrimination result. The discrimination result is input into a PID parameter cost function constructed by a model predictive control algorithm to solve the optimized PID parameters of each substation, which are issued to the command topic subscribed by each AVC substation, and each AVC substation re-collects data until the comprehensive performance level of each AVC substation is qualified.

6. The portable AVC automatic commissioning test method according to claim 5, characterized in that: Determining the JSON test task includes: When a user logs in, the user's authority is checked based on a preset authority matrix, an operation entrance matching the authority is opened, and an encrypted authority token is generated; the user inputs the set parameter setting target and the limit condition through the operation entrance matching the authority, and generates a user instruction combined with the encrypted authority token; When the test object is multiple substations, a collaborative scheduling model is constructed based on the historical running data of each AVC substation according to the model predictive control algorithm, and the test control sequence of each AVC substation is solved with the minimum bus voltage deviation, reactive power deviation and control amount as the target; when the test object is a single substation, the test control sequence of the substation is set; the voltage qualified range is solved according to the voltage threshold setting algorithm, and the reactive power adjustment threshold is solved according to the reactive power output adjustment algorithm; Test task parameter set is generated according to test control sequence, voltage qualified range and reactive power adjustment threshold; The substation set in the test task parameter set is parsed, a VLAN identifier is added to each substation, and a JSON test task is generated.

7. The portable AVC automatic commissioning test method according to claim 5, characterized in that: The new JSON data packet filled with real-time data collected by each AVC substation according to the JSON test task includes: Each AVC substation obtains the JSON test task from the subscribed command topic, parses the JSON test task, executes the test control sequence in the JSON test task according to the local PID controller, and collects real-time data according to the data collection item in the JSON test task; The collected data is arranged in the original sensor data format and filled into the payload field of the new JSON data packet; The CRC check code is calculated and attached to the new JSON data packet, and is published to the data feedback topic of each AVC substation subscribed by the communication module through the MQTT protocol.

8. The portable AVC automatic commissioning test method of claim 5, wherein: The preprocessed plaintext data is input into the dynamic and static curve real-time discrimination model based on the deep reinforcement learning network to output the dynamic and static curve discrimination result, including: According to the preprocessed plaintext data, a state vector is constructed, the state vector is used as the initial feature, the output of the deep reinforcement learning network hidden layer is calculated, the dynamic curve and the static wireless are used as the action vector, the output of the deep reinforcement learning network hidden layer is input into the deep reinforcement learning network output layer to calculate the Q value, and the dynamic and static curve is discriminated according to the Q value.

9. The portable AVC automatic commissioning test method of claim 5, wherein: The discrimination result is input into the PID parameter cost function constructed by the model predictive control algorithm, and the optimized PID parameters of each substation are obtained by solving, including: The comprehensive performance grade of each AVC substation is evaluated by calculating the voltage qualification rate, the reactive power regulation deviation rate and the regulation response time deviation; If the comprehensive performance grade is unqualified, the curve type discriminated and the PID parameter cost function constructed by the model predictive control algorithm are used; The PID parameter cost function is minimized by the gradient descent method to obtain the optimized PID parameters of each substation.

10. The portable AVC automatic commissioning test method of claim 5, wherein: The PID parameter cost function is represented by the following formula: wherein, a PID parameter cost function representing a proportional coefficient and an integral time of the mth substation, a reactive power deviation weight of the mth substation, a response time weight of the mth substation, a voltage qualification weight of the mth substation, a reactive power regulation deviation rate of the mth substation, a reactive power regulation response time of the mth substation, a voltage qualification deviation of the mth substation.

Images