An agent-based power grid protection setting value matching and checking method

Abstract

The application provides an agent-based power grid protection setting value matching and checking method, and relates to the technical field of power system automation. The steps of the application comprise: implementing OCR identification on a setting value single image to obtain a text sequence, and synchronously extracting protection setting values of corresponding equipment; utilizing a finite state automaton to analyze the text to generate a preliminary field-value pair, and combining state description to perform fault-tolerant reasoning by an agent to output a corrected field-value pair; executing fuzzy comparison to generate a difference list and mark an alarm level; storing difference information into a database and triggering an alarm or recording a log according to a result, so as to realize high-precision and strong-robust full-process automatic checking.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system automation, in particular to a power grid protection setting value matching and checking method based on an agent. BACKGROUND

[0002] The setting value parameters of the power grid relay protection device are the basis for ensuring the rapid and accurate removal of faulty elements under fault conditions, and need to be set comprehensively considering multiple factors such as power grid topology, equipment capacity and operation mode. With the popularization of digital substations and smart distribution networks, the order of magnitude and update frequency of protection setting values have been significantly improved, and traditional manual checking or simple script comparison methods cannot meet the real-time and reliability requirements.

[0003] In the prior art, information extraction is usually performed on the setting value sheet by means of optical character recognition (OCR) and deep learning methods such as graph neural networks, and then compared with the measured protection setting values in the database. However, when the OCR misrecognizes, the image resolution decreases, or the file contains abbreviations and wrong characters, key information may be missing or mismatched, and this error may be amplified along the deep model inference chain, ultimately affecting the checking accuracy, and the model inference process is not interpretable, so the problem root cause cannot be located. SUMMARY

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present application is to provide a power grid protection setting value matching and checking method based on an agent, which realizes high-precision, strong-robustness and full-process automatic checking of power grid protection setting values from structured extraction to difference comparison and risk warning by fusing the deterministic structure of finite state automata and the fault-tolerant reasoning ability of agents.

[0005] To achieve the above purpose, the present application provides the following solutions:

[0006] A power grid protection setting value matching and checking method based on an agent, comprising:

[0007] Performing OCR recognition on the file or image of the target protection setting value sheet, obtaining a text sequence, and synchronously extracting the current protection setting value of the corresponding device to form an initial data set;

[0008] Inputting the text sequence into a finite state automaton in a predetermined order, and outputting a preliminary field-value correspondence relationship according to a state transition function;

[0009] Submitting the preliminary field-value correspondence relationship and the state description of the finite state automaton to an agent as a prompt word to perform fault-tolerant reasoning on character missing, wrong characters and industry abbreviations, and outputting a corrected field-value pair;

[0010] Taking the field-value pair as a first comparison object and the initial data set as a second comparison object, the agent executes a comparison rule containing fuzzy matching and ignoring a specific identification of "field setting", generates a difference list, and marks an alarm level;

[0011] The difference list is stored in a risk database, and if the difference list is not empty, an alarm is pushed to an operation and maintenance end, otherwise a check pass log is recorded.

[0012] Preferably, OCR recognition is performed on a file or image of a target protection setting sheet to obtain a text sequence, and current protection settings of a corresponding device are synchronously extracted to form an initial data set, including:

[0013] An image or electronic file of the target protection setting sheet is obtained and stored as a to-be-recognized object;

[0014] An OCR recognition method is used to perform character extraction on the to-be-recognized object, and a text sequence is obtained by combination in a left-up-to-right-down order;

[0015] Device identification information is parsed from the text sequence and temporarily stored as device identification data; the device identification information includes a nameplate label of an electronic screen cabinet;

[0016] According to the device identification data, the current protection settings of the corresponding device are called and obtained through an interface of a power grid information system to form the initial data set.

[0017] Preferably, the text sequence is input into a finite state automaton in a preset order, and a preliminary field-value correspondence is output according to a state transition function, including:

[0018] The state set, input set, state transition function, initial state and acceptance state set of the finite state automaton are preloaded according to a preset setting sheet template;

[0019] The text sequence is sequentially split into multiple text segments in a left-up-to-right-down order to form an input sequence;

[0020] For each text segment in the input sequence, a state transition function is called to realize deterministic transition of a state;

[0021] After each state transition, a field name and value corresponding to a current state and an input symbol are written into an output set according to a defined output function to gradually generate the preliminary field-value correspondence.

[0022] Preferably, the preliminary field-value correspondence and a state description of the finite state automaton are submitted as a prompt word to an agent to perform fault-tolerant reasoning on missing characters, wrong characters and industry abbreviations, and output corrected field-value pairs, including:

[0023] The state name, state transition condition and input sequence of the finite state automaton are described in natural language, and the preliminary field-value correspondence relationship is embedded to form a prompt word text for calling by an intelligent agent;

[0024] The prompt word text is submitted to the intelligent agent with reasoning capability to trigger a fault-tolerant reasoning process;

[0025] Based on the intelligent agent, the fault-tolerant analysis of missing characters, misspelled words and industry abbreviations is performed according to the prompt word, and the corrected field-value pair is output;

[0026] The structured format of the field-value pair is checked; if the check fails, the intelligent agent is called again within a preset retry upper limit until it passes or reaches the upper limit.

[0027] Preferably, the field-value pair is the first comparison object, the initial data set is the second comparison object, the intelligent agent performs comparison rules containing fuzzy matching and ignoring specific identification of "field setting", generates a difference list, and marks the alarm level, including:

[0028] The field-value pair is determined as the first comparison object, and the initial data set is determined as the second comparison object;

[0029] The comparison rules are provided to the intelligent agent; the comparison rules include: fuzzy matching of letter-number similar characters, half-angle-full-angle symbols and commonly used unit abbreviations, and when the field value contains "field setting" identification, skip the numerical consistency comparison of the field value;

[0030] The intelligent agent compares the first comparison object and the second comparison object one by one based on the comparison rules, and generates a difference list containing field name, first value, second value and difference type;

[0031] According to the priority order of "missing> numerical inconsistency> format exception", the alarm level of each difference in the difference list is marked, and the graded difference list is output.

[0032] Preferably, the difference list is stored in a risk database, and if the difference list is not empty, an alarm is pushed to the operation and maintenance end, otherwise a review pass log is recorded, including:

[0033] The difference list is written into a preset risk database;

[0034] Determine whether the difference list is empty;

[0035] If yes, send information containing difference details and alarm level to the operation and maintenance end through an operation and maintenance alarm interface;

[0036] If not, record the check pass result in the check log.

[0037] The present application discloses the following technical effects:

[0038] The present application ensures that the field extraction has a clear path and logical rules by using a finite state automaton to drive the state analysis of the text sequence after OCR recognition, avoiding the fuzzy matching and misrecognition problems in traditional information extraction methods, and improving the reliability of the analysis link. The present application combines the agent's prompt of the preliminary field-value pair correspondence and the automatic machine state description, realizes the automatic correction of common abnormalities such as character missing, wrong characters and abbreviations in actual scenarios, thereby avoiding manual intervention and reducing false positives. The present application automatically identifies the on-site setting and other exempt fields by executing comparison rules with fuzzy matching capability through the agent, and skips invalid comparison, effectively dealing with the flexible on-site setting of the power grid and the problem of non-uniform expression. The comparison result difference of the present application is classified and labeled and persisted to the risk database, and if an abnormality is found, an active alarm is given, and if the check passes, a log is automatically recorded, realizing the full-process automatic management from recognition, comparison to alarm. BRIEF DESCRIPTION OF DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0040] Figure 1 The method flowchart provided for the embodiments of the present application;

[0041] Figure 2 The technical route schematic diagram provided for the embodiments of the present application;

[0042] Figure 3 The state transition diagram of the setting single analysis provided for the embodiments of the present application;

[0043] Figure 4 The flow line schematic diagram provided for the embodiments of the present application;

[0044] Figure 5 The system architecture diagram provided for the embodiments of the present application. DETAILED DESCRIPTION

[0045] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of the present application.

[0046] The present application aims to provide an agent-based power grid protection setting matching and checking method, which combines the deterministic structure of finite state automata and the fault-tolerant reasoning ability of agents, and realizes high-precision, strong-robustness and full-process automatic checking of power grid protection setting from structured extraction to difference comparison and risk warning.

[0047] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] Figure 1 As shown in the method flowchart provided by the embodiments of the present application, Figure 1 The present application provides an agent-based power grid protection setting matching and checking method, which includes:

[0049] Step 100: Perform OCR recognition on the file or image of the target protection setting sheet, obtain the text sequence, and simultaneously extract the current protection setting of the corresponding device to form an initial data set;

[0050] Step 200: Input the text sequence into the finite state automata according to the preset order, and output the preliminary field-value correspondence according to the state transition function;

[0051] Step 300: Submit the preliminary field-value correspondence and the state description of the finite state automata to the agent as a prompt word, to perform fault-tolerant reasoning on character missing, wrong characters and industry abbreviations, and output the corrected field-value pair;

[0052] Step 400: Take the field-value pair as the first comparison object, and the initial data set as the second comparison object, execute the comparison rules containing fuzzy matching and ignoring the specific identifier of "on-site setting" by the agent, generate a difference list, and mark the alarm level;

[0053] Step 500: Store the difference list in the risk database, and if the difference list is not empty, push the alarm to the operation and maintenance end, otherwise record the checking pass log.

[0054] Specifically, the deployment environment required by the present embodiment at least has 800GB solid state disk storage, a computing card and 32GB or more memory; the agent is required to have at least 64K context length and reasoning ability.

[0055] Specifically, the embodiment splits the text structure in the setting value chart / table by using the finite state automaton method to determine the corresponding relationship between the fields and values. Then, the state transition method is described in a natural language as input 1, and the text box recognized by the OCR is arranged in a left-top-right-bottom order as input 2, which are submitted to the intelligent agent for analysis to obtain the corresponding relationship of each table. As shown in the following formula, the technical route of the embodiment is as follows: Figure 2

[0056] Step 1, OCR text acquisition: the text data of the setting sheet picture / document is obtained by OCR or other reasonable methods. Since the setting sheet is a standard, strictly reviewed and formatted text, the structure of the setting sheet is stable and clear, and the field description is clear, so the finite state automaton method can be used for analysis.

[0057] Step 2, definition of finite state machine ; wherein Q represents a finite state set, which represents the state of each step, such as “protection setting value analysis state”, “control word analysis state”, “table header analysis state” and the like; represents a set of input symbols, which is a set of texts analyzed by the OCR in this application scenario; represents a state transition function, and the function represents that the next state can be transferred according to the current state and the input symbol; represents the initial state, which is the state of the beginning of analysis; F represents the set of accepted states, which represents the state of the completion or successful extraction of data. The OCR analysis result contains words and word boxes, and the word boxes are arranged in a left-top-right-bottom order, and the text sequence of the input can be obtained , wherein each s is a text segment.

[0058] Step 3, output function setting: define the output function ; wherein O represents the set of outputs.

[0059] Step 4, state-driven analysis: input the text sequence obtained in step 1 into the finite state automaton, and perform post-processing according to the output function defined in step 3, then for each state point and each input text , there is:

[0060] .

[0061] The entire analysis process can be represented as the following transition sequence:

[0062] .

[0063] ​Step 5, state transition design: according to the setting template, design state transition diagram and equation set, design program, sub-step specific as follows:

[0064] Step 5.1: match keywords and state transition. This scheme drives state transition by regular keywords, assuming that the keyword set is , define the state transition function As follows:

[0065] .

[0066] Where, q represents the current state, Indicates the transition to the next state, s represents the text sequence, Represents the keyword set.

[0067] Step 5.2: define the pattern recognition function To determine the state:

[0068] .

[0069] Where, s represents the text sequence, pattern(s) represents the regular matching mode of the text sequence, title, definition, value, multi-value represent title, table header definition, table value, multi-table value and other states, as the transition point of state, , , Represents the keyword set of title, table header definition, value state, etc. Indicates the length of the cell table header definition. If it is greater than 1, that is, a table header corresponds to multiple values, it is considered that the cell is a multi-value state.

[0070] According to the situation, define the title state, field definition state, value segment state, multi-value segment state, tail column state, page column state, etc.

[0071] Step 5.3: execute the state transition function to get the single-step output, and place the effective output in the output set .

[0072] Step 5.4: as shown in Figure 3 Repeat step 5.3 until the input ends or the result points to the termination. Since the setting sheet must be of finite length, it can be guaranteed that the automaton must exist termination.

[0073] Step 6, Natural Language Description: Based on the state transition function from the previous step, use natural language to describe the state transition diagram and store it as a prompt word for submission to the agent in the next section for analysis. For example: "Consider a state transition mode: in ${condition1}, from ${state1} -> ${state2}; in ${condition2}, from ${state2} -> ${state3} …", and so on, until the entire state transition diagram is described.

[0074] Further, the embodiment combines the setting book matching and setting value checking with the state transition sequence (i.e., state transition diagram) summarized by natural language in the previous step, inputs it into the agent, and at the same time, inputs the sequence of OCR text boxes into the agent as the parsed object, so that the agent analyzes the corresponding relationship between the setting book field and value according to the description. Finite state automaton is rigorous and can uniquely determine the structure of the setting book input, avoiding the probabilistic recognition path that may exist when using methods such as graph neural network in text parsing, and as a black box, it cannot locate the problem. The agent has certain flexibility and can automatically recognize some misspelled words and abbreviations, solving the problem that traditional methods need to use semantic distance to judge the vectorized text, while still not guaranteeing correct recognition of slang and abbreviations. This embodiment uses a finite state automaton scheme, which is rigorous in parsing, has only one corresponding state transition mode for a single input, and is a white box that can be immediately located after an error. It is flexible in matching, ensuring that human errors, abbreviations, and slang are considered in the matching process.

[0075] As another alternative embodiment, in this embodiment, step 1 obtains the information of the nameplate label of the electronic screen cabinet through OCR, temporarily stores it, and uses it as the screen cabinet name in step 3. Step 2 reads the setting book list and stores it in a temporary variable, which is used as the file name list in step 3. Step 3 calls the agent and uses the following prompt words or prompt words with the same semantics to filter the file names, where ${} represents a variable part:

[0076] For a given file name, find the files in the file list that may be related to it, there may be multiple. The file name in the file list may be abbreviated for some unnecessary information, so even if it is not completely matched, it can be considered as matched if the semantics are the same. Since it comes from OCR, the file name may be misrecognized, so the following fuzzy matching rules are followed when matching:

[0077] Only when there is no correct match, confusing letters and numbers can be considered the same.

[0078] Only when there is no correct match, abbreviations and abbreviations of some information are allowed.

[0079] Input screen cabinet name: ${name};

[0080] Input file name list: ${filelist};

[0081] Output content contains only one python list form result.

[0082] Further, step 4 of the embodiment is to extract the content information of the selected setting book. Among them, step 4.1 is to extract the text of the selected setting book using OCR technology or other reasonable technology, denoted as s. Step 4.2 is to collect the file content in the result returned by the agent in step 3, and the result is generated based on the finite state automaton method. Or use natural language description state transition method as prompt word to submit to the agent. The prompt word must include: input sequence, state name, state transition condition, post-processing method. The following is an example, where ${} form represents a variable part:

[0083] "According to the following state transition method, extract ${target} information from the input sequence;

[0084] State contains: ${status_list};

[0085] When ${condition1} is met, state ${x1}->state ${x2};

[0086] When ${condition2} is met, state ${x2}->state ${x3}; ...

[0087] Input sequence: ${s}.

[0088] Optionally, the output content of the embodiment contains only one python dictionary / list syntax form result.

[0089] Further, step 5 of the embodiment is to use the agent to compare the returned result of step 4 with the real data. It is assumed that the real data has been collected by a reasonable method and has been structured stored. The specific sub-steps are as follows:

[0090] Step 5.1: Extract the current real data from the persistent storage / network interface as input .

[0091] Step 5.2: Collect the returned result of the finite state automaton program / agent from step 4 as input .

[0092] Step 5.3: Compare with As input, fill in the following prompt word template, submit to the agent, where the content in the form of ${} represents the variable part, and the "key" refers to the unique identifier in the structured field, corresponding to the set value name, soft clamp name, etc. in the setting book. The "value" refers to the numerical value in the structured field, corresponding to the setting value, control word, etc. in the setting book. The prompt word template is:

[0093] Content comparison is performed on the given input p and input q. The following conditions must be followed:

[0094] If there is a field in q that does not exist in p, an alarm is given. Otherwise, it is ignored;

[0095] The content of q is derived from OCR, and if the key in q cannot be matched with the one in p, fuzzy matching is performed. Allowable ambiguous alphanumeric characters are matched, and partial character abbreviations or changes are acceptable without changing the semantics. In this case, try to find the matching item of q in p again.

[0096] When the value segment in q contains synonymous content such as "field setting", it is no longer checked whether the value segment in p is equal.

[0097] When the value segment of q is not equal to the value segment of p, only fuzzy comparison using ambiguous alphanumeric characters is considered. If it is still not equal, it is considered that q and p do not match on this key, and an alarm should be raised.

[0098] Input p: ${p};

[0099] Input q: ${q};

[0100] Optionally, the output of the embodiment only contains the keys that need to be alarmed in the form of a python list.

[0101] Specifically, step 6 of the embodiment is to collect the returned results of step 5, and after verifying the correct format, store them in persistent storage. And through the matching system, the alarm is issued.

[0102] As an optional implementation, the agent-based setting value checking workflow of the embodiment is: using workflow combined with agent, all steps such as collection, identification, analysis, checking, and alarm can be automatically performed, and at the same time, the process output that does not meet the specifications is automatically discovered, and retry or prompt to the user layer is performed.

[0103] Optionally, the setting value checking overall process will run on an automation software pipeline, which supports timed start and manual start, and once configuration long-term operation. The prompt words of the agent will be stored in the server in a template, the pipeline calls the agent through the local area network, and the filled template is transmitted to the agent through the local area network interface to perform reasoning. Through the script, it can automatically check whether the agent returns the content conforming to the output format, and whether the file name and field in the returned content exist. When the check fails, the agent is automatically re-executed. To a certain extent, it can correct the occasional illusion problem of the agent and further improve the availability. The alarm result can be stored in the server for timed release / real-time release, further reducing the labor and time cost in operation and maintenance.

[0104] Figure 4 An automatic pipeline building method is shown, and the automatic pipeline runs the entire analysis and matching process. After the user uploads the picture of the identification screen to the system, the pipeline starts running. The agent is called to obtain the output of each step, which is used as the input of the next step. When the output result is empty or the output format does not meet the specification, retry is triggered. To prevent infinite retry, the maximum number of retries is set to 3, and more than 3 times is prompted as a failure, considering whether the photo or setting book file exists. After the execution is completed, the final result is obtained, the result is published and stored in the business database of the setting value system for subsequent review.

[0105] As shown in Figure 5 The present application is divided into three parts on the system architecture, the terminal equipment collects the picture of the pressure screen cabinet in the substation, uploads it to the background server, and the background server periodically collects setting data and setting book files from the data platform. The background server calls the agent through the automatic pipeline to store the final execution result.

[0106] The beneficial effects of the present application are as follows:

[0107] (1) The present application solves the matching problem caused by the fact that the result of OCR is not always completely accurate, and the conventional method such as fuzzy matching or semantic distance method usually lacks flexibility to deal with such problems.

[0108] (2) The present application combines the advantages of finite state automaton model that can be predicted and determined and the advantages of agent that can be flexible and reasoning, and solves the problem of difficult structured analysis of complex icon files in the OCR process.

[0109] (3) The present application uses the agent workflow to improve the reading, comparison and notification efficiency in the setting single checking process.

[0110] The various embodiments described in this specification are presented for the purpose of illustrating the principles of the present application and its best mode of operation. Each of the embodiments described in this specification has been provided for the purpose of illustration and is not intended to limit the application.

[0111] The principles and implementations of the present application have been described in the above embodiments, which are only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, the specific implementation and application range of the present application can be changed according to the idea of the present application. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. An agent-based method for power grid protection setting matching and checking, characterized in that, The method comprises the following steps: applying OCR recognition to a file or an image of a target protection setting sheet to obtain a text sequence and synchronously extract current protection settings of a corresponding device to form an initial data set; inputting the text sequence into a finite state automaton in a preset order and outputting a preliminary field-value correspondence according to a state transition function; submitting the preliminary field-value correspondence and a state description of the finite state automaton to an intelligent agent as a prompt word to perform fault-tolerant reasoning on missing characters, wrong characters and industry abbreviations and output a corrected field-value pair; taking the corrected field-value pair as a first comparison object and the initial data set as a second comparison object, executing comparison rules containing fuzzy matching and ignoring a specific identifier of "field setting" by the intelligent agent, generating a difference list and marking an alarm level; storing the difference list in a risk database and pushing an alarm to an operation and maintenance end if the difference list is not empty, otherwise recording a check-through log; taking the corrected field-value pair as a first comparison object and the initial data set as a second comparison object, executing comparison rules containing fuzzy matching and ignoring a specific identifier of "field setting" by the intelligent agent, generating a difference list and marking an alarm level, comprising: determining the corrected field-value pair as the first comparison object and the initial data set as the second comparison object; providing the comparison rules to the intelligent agent; the comparison rules include fuzzy matching of similar characters, half-angle-full-angle symbols and common unit abbreviations and skipping the value consistency comparison of a field when the field value contains the "field setting" identifier; the intelligent agent compares the first comparison object with the second comparison object one by one based on the comparison rules to generate a difference list containing a field name, a first value, a second value and a difference type; marking an alarm level for each difference in the difference list in the order of "missing> value inconsistency> format exception" and outputting the graded difference list.

2. The agent-based power grid protection setting matching and checking method according to claim 1, characterized in that, applying OCR recognition to a file or an image of a target protection setting sheet to obtain a text sequence and synchronously extract current protection settings of a corresponding device to form an initial data set, comprising: obtaining an image or an electronic file of a target protection setting sheet and storing it as a to-be-recognized object; performing character extraction on the to-be-recognized object by using an OCR recognition method and combining the text sequence in the order of top-left to bottom-right; parsing device identification information from the text sequence and temporarily storing it as device identification data; the device identification information includes a nameplate label of an electronic screen cabinet; obtaining the current protection settings of the corresponding device through an interface of a power grid information system according to the device identification data to form the initial data set.

3. The agent-based power grid protection setting matching and checking method according to claim 1, characterized in that, inputting the text sequence into a finite state automaton in a preset order and outputting a preliminary field-value correspondence according to a state transition function, comprising: preloading a state set, an input set, a state transition function, an initial state and an accepted state set of the finite state automaton according to a preset setting sheet template; splitting the text sequence into multiple text segments in the order of top-left to bottom-right to form an input sequence; For each text segment in the input sequence, a state transition function is called to achieve deterministic transition of states; After each state transition, the current state and the field name and value corresponding to the input symbol are written into the output set according to the defined output function, to gradually generate the preliminary field-value correspondence.

4. The agent-based power grid protection setting matching and checking method according to claim 1, characterized in that, The preliminary field-value correspondence and the state description of the finite state automaton are submitted to an agent as a prompt word, to perform fault-tolerant reasoning on character omissions, typos and industry abbreviations, and output corrected field-value pairs, including: The state name, state transition condition and input sequence of the finite state automaton are described in natural language, and the preliminary field-value correspondence is embedded to form a prompt word text for calling by the agent; The prompt word text is submitted to the agent with reasoning capability to trigger a fault-tolerant reasoning process; Based on the agent, fault-tolerant analysis is performed on character omissions, typos and industry abbreviations according to the prompt word, and corrected field-value pairs are output; The field-value pairs are subjected to structured format verification; if the verification fails, the agent is called again within a preset retry upper limit until it passes or the upper limit is reached.

5. The agent-based power system protection setting matching and checking method of claim 1, wherein, The difference list is stored in a risk database, and if the difference list is not empty, an alarm is pushed to the operation and maintenance end, otherwise a check pass log is recorded, including: The difference list is written into a preset risk database; Determine whether the difference list is empty; If yes, send information containing difference details and alarm levels to the operation and maintenance end through an operation and maintenance alarm interface; If not, record the check pass result in the check log.

Images