Unmanned intelligent monitoring system and monitoring method for power plant

Abstract

The invention discloses an unmanned intelligent monitoring system and method for a power plant, and relates to the technical field of intelligent monitoring, and the system comprises a power plant pipeline data acquisition unit which is used for collecting the real-time related data of a pipeline in the power plant by taking the unmanned intelligent monitoring system for the power plant pipeline as an example. And the real-time related data of the pipeline in the power plant is collected, the collected real-time related data of the pipeline in the power plant is processed for subsequent use, and the processed real-time related data of the pipeline in the power plant is stored and recorded. The invention provides an unmanned intelligent monitoring system and a monitoring method for a power plant, which take the unmanned intelligent monitoring system for a power plant pipeline as an example, collects and processes real-time data of the power plant pipeline through an arranged power plant pipeline data acquisition unit so as to provide data support for subsequent pipeline early warning detection. A pipeline early-warning rule is constructed based on pipeline early-warning knowledge through the arranged pipeline early-warning knowledge base construction unit, and a basis is provided for subsequent pipeline early-warning detection.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent monitoring panels, and in particular to an unmanned intelligent monitoring panel system and a monitoring panel method for a power plant. Background Art

[0002] With the increase of the single-unit capacity and total installed capacity of power plants, the technical requirements for the operation, maintenance, and fault handling of super-large power plants are becoming higher and higher, with greater difficulty and risk. Operation, maintenance, and management personnel need to monitor a large number of equipment parameters. In the Chinese patent application number 202210826590.8, "an intelligent monitoring and fault warning system for smart power plants is disclosed, including: an intelligent interface module, configured with different intelligent interfaces; a data acquisition, storage, and processing module, which collects real-time and historical data of unit equipment operation from a DCS system or a SIS system in seconds; a configuration data modeling module, Through the existing algorithm configuration, graphical configuration modeling or code interactive modeling; intelligent analysis module, parameter warning and fault analysis; intelligent monitoring module, real-time display of warning and fault information; monitoring assistant module, operation data recording, typical working conditions automatic meter reading, record the historical data of the characteristic values ​​of important parameters when the parameters on the monitoring screen are abnormal. The present invention has the advantages of intelligent interface, convenient deployment, easy maintenance and intelligent modeling, which meets the current situation of uneven progress in the construction of smart power plants, can reduce the workload of operators, improve the technical level of operation monitoring and equipment management, and ensure the safe operation of equipment. "

[0003] This prior art only solves the problem that as the number of access systems increases, the amount of alarm information in the current power plant monitoring system increases, there are a large number of invalid process alarms, the alarms lack correlation analysis, and early warnings cannot be issued, resulting in heavy monitoring tasks for on-duty personnel, inability to grasp the most important events, and a high risk of missed alarms, which directly affects the safety of centralized control equipment. It does not take into account that when monitoring the power plant pipeline, it is necessary to determine the pipeline warning rules based on warning knowledge to provide a basis for pipeline warning detection, and when conducting pipeline warning detection, it is necessary to comprehensively consider the results obtained by various detection methods to judge and identify the pipeline warning level. Summary of the invention

[0004] The object of the present invention is to provide an unmanned intelligent monitoring system and a monitoring method for a power plant to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an unmanned intelligent monitoring system for a power plant, comprising a power plant pipeline data acquisition unit. Taking the unmanned intelligent monitoring system for a power plant pipeline as an example, the power plant pipeline data acquisition unit is used to collect real-time relevant data of the pipelines in the power plant, and process the collected real-time relevant data of the pipelines in the power plant for subsequent use, and store and record the processed real-time relevant data of the pipelines in the power plant;

[0006] A pipeline early warning knowledge base construction unit, the pipeline early warning knowledge base construction unit is used to acquire power plant pipeline early warning knowledge, generate a power plant pipeline early warning knowledge base based on the acquired power plant pipeline early warning knowledge, further construct power plant pipeline early warning rules based on the power plant pipeline early warning knowledge through different representation methods, and self-update the power plant pipeline early warning knowledge in the power plant pipeline early warning knowledge base;

[0007] A pipeline early warning model construction unit, which is used to construct a variety of pipeline early warning models, and perform model training and model accuracy verification on various pipeline early warning models, and determine the reliability of each pipeline early warning model based on the accuracy verification results of each pipeline early warning model, and collect and record the output results of each pipeline early warning model;

[0008] The intelligent early warning unit is used to identify and judge the early warning type of the power plant pipeline from the two aspects of reliability and repeatability based on the output results of each pipeline early warning model, and to notify the early warning information based on the judgment result.

[0009] Preferably, the power plant pipeline data acquisition unit includes a pipeline data acquisition module, which is used to collect real-time relevant data of the pipelines in the power plant. The specific operation is to divide the power plant into areas, and perform a naming operation on each area in the power plant, the content of which is a combination of the area and the digital codes connected in sequence, and perform a naming operation on each pipeline in each area in the power plant, the content of which is a combination of the area naming content, the pipeline and the digital codes connected in sequence, so as to achieve accurate positioning of each section of the pipeline, and obtain real-time data generated during the working process of the pipeline through the set pressure transmitter, temperature transmitter and flow computer.

[0010] Preferably, the power plant pipeline data acquisition unit also includes a pipeline data processing module and a pipeline data recording module. The pipeline data processing module is used to process the collected real-time related data of the pipeline in the power plant. The specific operation is to perform missing value detection and abnormal value detection on the acquired real-time related data of the pipeline in the power plant, and when missing values ​​are detected in the real-time related data of the pipeline in the power plant, the mean of the four node data adjacent to the missing value is used to fill the missing value to retain the time series characteristics of the pipeline data. When abnormal values ​​are detected in the real-time related data of the pipeline in the power plant, the mean of the remaining data is used to replace the abnormal value, and the processed pipeline data is normalized to eliminate the impact of the dimensions and orders of magnitude of different data types. The pipeline data recording module is used to store and record the processed real-time related data of the pipeline in the power plant, and establish an association with the naming content information of the corresponding pipeline.

[0011] Preferably, the pipeline early warning knowledge base construction unit includes a pipeline early warning knowledge acquisition module and a pipeline early warning knowledge base generation module. The pipeline early warning knowledge acquisition module is used to acquire the pipeline early warning knowledge of the power plant, specifically to obtain the data required for early warning from the pipeline database and pipeline early warning materials by manual acquisition, and further to supplement the pipeline early warning knowledge by consulting relevant field literature, pipeline early warning technical manuals, pipeline early warning professional books and pipeline early warning historical cases, and in combination with the semi-automatic acquisition method, to obtain pipeline early warning expert experience by holding talks with pipeline early warning professional technicians and business management personnel. The pipeline early warning knowledge base generation module is used to construct the power plant pipeline early warning knowledge base based on the pipeline early warning knowledge acquired by the pipeline early warning knowledge acquisition module.

[0012] Preferably, the pipeline warning knowledge base construction unit also includes a pipeline warning knowledge representation module and a pipeline warning knowledge self-update module. The pipeline warning knowledge representation module is used to construct power plant pipeline warning rules based on power plant pipeline warning knowledge through different representation methods. The specific operation is to intuitively and clearly represent the pipeline warning knowledge through the production rule representation method and the framework representation method, thereby eliminating the defect that a single representation method cannot accurately represent the pipeline warning knowledge, and construct the power plant pipeline warning rules based on the represented pipeline warning knowledge, thereby providing a basis for subsequent pipeline warning detection. The pipeline warning knowledge self-update module is used to regularly self-update the pipeline warning knowledge in the power plant pipeline warning knowledge base through the extended isolation forest method.

[0013] Preferably, the pipeline early warning model construction unit includes a pipeline early warning model generation module, which is used to construct a variety of pipeline early warning models. The specific operation is to select the LSTM model as the basic model, combine the ARIMA model with the LSTM model to construct an ARIMA-LSTM hybrid model, combine the self-attention mechanism with the LSTM model to construct an SA-LSTM hybrid model, combine the CNN network with the LSTM model to construct a CNN-LSTM hybrid model, optimize the parameters of the LSTM model through the Bayesian method, and construct a BO-LSTM hybrid model.

[0014] Preferably, the pipeline early warning model construction unit further includes a pipeline early warning model training module and a pipeline early warning model output module. The pipeline early warning model training module is used to perform model training and model accuracy verification on four pipeline early warning models. The specific operation is to obtain pipeline early warning data through a network crawler, and construct a pipeline early warning data set based on the obtained pipeline early warning data, and respectively perform model training on the four pipeline early warning models through the constructed pipeline early warning data set, and perform accuracy verification on the four trained pipeline early warning models through a model accuracy algorithm. The specific calculation formula is as follows:

[0015]

[0016] Wherein, PR represents the precision value of the pipeline early warning model, TP represents the number of sample data with correct early warning results, FP represents the number of sample data with wrong early warning results, and the precision verification results of each pipeline early warning model are arranged in descending order, and the reliability of the pipeline early warning model is assigned 0.1, 0.2, 0.3 and 0.4 respectively according to the sorting. The pipeline early warning model output module is used to obtain the output results of each pipeline early warning model based on the processed real-time relevant data of the power plant pipeline, and collect and record the output results of each pipeline early warning model respectively.

[0017] Preferably, the intelligent early warning unit includes an early warning identification and judgment module and an early warning notification module. The early warning identification and judgment module is used to identify and judge the early warning type of the power plant pipeline from the two aspects of reliability and repeatability. The operation of identifying and judging the early warning type of the power plant pipeline from the aspect of reliability is to calculate the early warning terminal value of the pipeline through the pipeline early warning algorithm based on the output results of each pipeline early warning model. The specific calculation formula is as follows:

[0018]

[0019] Where i represents the i-th pipeline early warning model, v i represents the reliability value of the i-th pipeline early warning model, R i represents the output result of the i-th pipeline early warning model, and R i∈[0, 100], and the pipeline warning final value is divided into 5 pipeline warning levels with a tolerance of 20 according to the value range, and when 0≤calculation result<20, it is judged as pipeline warning level one, and when 80≤calculation result≤100, it is judged as pipeline warning level five, wherein the operation of identifying and judging the pipeline warning type of the power plant from the aspect of repetition is that when the output results of the four pipeline warning models are all lower than 60, it is judged as pipeline warning level one, when the output result of one pipeline warning model is greater than 60, it is judged as warning level two, when the output results of two pipeline warning models are greater than 60, it is judged as warning level three, when the output results of three pipeline warning models are greater than 60, it is judged as warning level four, and when the output results of four pipeline warning models are greater than 60, it is judged as warning level five, and the warning notification module is used to notify the warning message based on the judgment result of the warning identification and judgment module, and after the corresponding pipeline is processed based on the warning message, the warning message is ignored to prevent repeated processing.

[0020] A method for unmanned intelligent monitoring of a power plant, comprising the following steps:

[0021] S1: collecting real-time relevant data of pipelines in the power plant through the pipeline data collection module, processing the collected real-time relevant data of pipelines in the power plant through the pipeline data processing module, and storing and recording the processed real-time relevant data of pipelines in the power plant through the pipeline data recording module;

[0022] S2: The pipeline warning knowledge acquisition module is used to acquire the power plant pipeline warning knowledge, the pipeline warning knowledge base generation module is used to build the power plant pipeline warning knowledge base, the pipeline warning knowledge representation module is used to build the power plant pipeline warning rules based on the power plant pipeline warning knowledge through different representation methods, and the pipeline warning knowledge self-update module is used to regularly update the pipeline warning knowledge in the power plant pipeline warning knowledge base;

[0023] S3: Construct multiple pipeline warning models through the pipeline warning model generation module, train four pipeline warning models and verify model accuracy through the pipeline warning model training module, collect and record the output results of each pipeline warning model through the pipeline warning model output module, identify and judge the power plant pipeline warning type from the two aspects of reliability and repeatability through the warning identification and judgment module, and issue warning message notification based on the judgment result of the warning identification and judgment module through the warning notification module.

[0024] Compared with the prior art, the beneficial effects of the present invention at least include: the present invention proposes an unmanned intelligent monitoring system and a monitoring method for a power plant, and takes the unmanned intelligent monitoring system for a power plant pipeline as an example, collects and processes the real-time data of the power plant pipeline through a set power plant pipeline data acquisition unit to provide data support for subsequent pipeline early warning detection, constructs pipeline early warning rules based on pipeline early warning knowledge through a set pipeline early warning knowledge base construction unit to provide a basis for subsequent pipeline early warning detection, constructs multiple pipeline early warning models through a set pipeline early warning model construction unit, and assigns different reliability values ​​to each pipeline early warning model, identifies and judges the pipeline early warning level from two aspects through a set intelligent early warning unit, and notifies the pipeline early warning message of the corresponding level, and after the corresponding pipeline is processed based on the early warning message, the early warning message is ignored to prevent repeated processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a schematic diagram of the overall process of an embodiment of the present invention;

[0026] Figure 2 This is a schematic diagram of the process of a power plant pipeline data acquisition unit according to an embodiment of the present invention;

[0027] Figure 3 This is a schematic diagram of a pipeline early warning knowledge base construction unit process according to an embodiment of the present invention;

[0028] Figure 4 This is a schematic diagram of the pipeline early warning model construction unit process according to an embodiment of the present invention;

[0029] Figure 5 The figure is a flow chart of the intelligent early warning unit according to an embodiment of the present invention.

[0030] In the figure: 100, power plant pipeline data acquisition unit; 101, pipeline data acquisition module; 102, pipeline data processing module; 103, pipeline data recording module; 200, pipeline early warning knowledge base construction unit; 201, pipeline early warning knowledge acquisition module; 202, pipeline early warning knowledge base generation module; 203, pipeline early warning knowledge representation module; 204, pipeline early warning knowledge self-update module; 300, pipeline early warning model construction unit; 301, pipeline early warning model generation module; 302, pipeline early warning model training module; 303, pipeline early warning model output module; 400, intelligent early warning unit; 401, early warning identification and judgment module; 402, early warning notification module. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] See also Figure 1-5 The present invention provides a technical solution: an unmanned intelligent monitoring system for a power plant, comprising a power plant pipeline data acquisition unit 100. Taking the unmanned intelligent monitoring system for a power plant pipeline as an example, the power plant pipeline data acquisition unit 100 is used to collect real-time relevant data of the pipelines in the power plant, and process the collected real-time relevant data of the pipelines in the power plant for subsequent use, and store and record the processed real-time relevant data of the pipelines in the power plant;

[0033] The pipeline early warning knowledge base construction unit 200 is used to acquire power plant pipeline early warning knowledge, generate a power plant pipeline early warning knowledge base based on the acquired power plant pipeline early warning knowledge, further construct power plant pipeline early warning rules based on the power plant pipeline early warning knowledge through different representation methods, and self-update the power plant pipeline early warning knowledge in the power plant pipeline early warning knowledge base;

[0034] The pipeline early warning model construction unit 300 is used to construct a variety of pipeline early warning models, and perform model training and model accuracy verification on various pipeline early warning models, and determine the reliability of each pipeline early warning model based on the accuracy verification results of each pipeline early warning model, and collect and record the output results of each pipeline early warning model;

[0035] The intelligent early warning unit 400 is used to identify and judge the early warning type of the power plant pipeline from the two aspects of reliability and repeatability based on the output results of each pipeline early warning model, and to notify the early warning information based on the judgment result.

[0036] The power plant pipeline data acquisition unit 100 includes a pipeline data acquisition module 101, which is used to collect real-time related data of pipelines in the power plant. The specific operation is to divide the power plant into regions, and perform a naming operation on each region in the power plant, which is a combination of the region and the digital codes connected in sequence, and perform a naming operation on each pipeline in each region in the power plant, which is a combination of the region naming content, the pipeline and the digital codes connected in sequence, so as to achieve accurate positioning of each section of the pipeline, and obtain real-time data generated during the working process of the pipeline through the set pressure transmitter, temperature transmitter and flow computer;

[0037] The power plant pipeline data acquisition unit 100 also includes a pipeline data processing module 102 and a pipeline data recording module 103. The pipeline data processing module 102 is used to process the collected real-time related data of the pipeline in the power plant. The specific operation is to perform missing value detection and abnormal value detection on the acquired real-time related data of the pipeline in the power plant, and when it is detected that there are missing values ​​in the real-time related data of the pipeline in the power plant, the mean of the four node data adjacent to the missing value is used to fill the missing value to retain the time series characteristics of the pipeline data. When it is detected that there are abnormal values ​​in the real-time related data of the pipeline in the power plant, the mean of the remaining data is used to replace the abnormal value, and the processed pipeline data is normalized to eliminate the impact of the dimensions and orders of magnitude of different data types. The pipeline data recording module 103 is used to store and record the processed real-time related data of the pipeline in the power plant, and establish an association with the naming content information of the corresponding pipeline;

[0038] The pipeline early warning knowledge base construction unit 200 includes a pipeline early warning knowledge acquisition module 201 and a pipeline early warning knowledge base generation module 202. The pipeline early warning knowledge acquisition module 201 is used to acquire the pipeline early warning knowledge of the power plant, specifically to obtain the data required for early warning from the pipeline database and pipeline early warning materials by manual acquisition, and further supplement the pipeline early warning knowledge by consulting relevant field literature, pipeline early warning technical manuals, pipeline early warning professional books and pipeline early warning historical cases, and in combination with the semi-automatic acquisition method, to obtain pipeline early warning expert experience by holding talks with pipeline early warning professional technicians and business management personnel. The pipeline early warning knowledge base generation module 202 is used to construct the power plant pipeline early warning knowledge base based on the pipeline early warning knowledge acquired by the pipeline early warning knowledge acquisition module 201;

[0039] The pipeline warning knowledge base construction unit 200 also includes a pipeline warning knowledge representation module 203 and a pipeline warning knowledge self-updating module 204. The pipeline warning knowledge representation module 203 is used to construct a power plant pipeline warning rule based on the power plant pipeline warning knowledge through different representation methods. The specific operation is to intuitively and clearly represent the pipeline warning knowledge through the production rule representation method and the framework representation method, thereby eliminating the defect that a single representation method cannot accurately represent the pipeline warning knowledge, and construct the power plant pipeline warning rule based on the represented pipeline warning knowledge, thereby providing a basis for subsequent pipeline warning detection. The pipeline warning knowledge self-updating module 204 is used to regularly self-update the pipeline warning knowledge in the power plant pipeline warning knowledge base through the extended isolation forest method;

[0040] The pipeline early warning model construction unit 300 includes a pipeline early warning model generation module 301, which is used to construct a variety of pipeline early warning models. The specific operation is to select the LSTM model as the basic model, combine the ARIMA model with the LSTM model to construct an ARIMA-LSTM hybrid model, combine the self-attention mechanism with the LSTM model to construct an SA-LSTM hybrid model, combine the CNN network with the LSTM model to construct a CNN-LSTM hybrid model, and optimize the parameters of the LSTM model through the Bayesian method to construct a BO-LSTM hybrid model;

[0041] The pipeline early warning model construction unit 300 also includes a pipeline early warning model training module 302 and a pipeline early warning model output module 303. The pipeline early warning model training module 302 is used to perform model training and model accuracy verification on four pipeline early warning models. The specific operation is to obtain pipeline early warning data through a network crawler, and construct a pipeline early warning data set based on the obtained pipeline early warning data, and respectively perform model training on the four pipeline early warning models through the constructed pipeline early warning data set, and perform accuracy verification on the four trained pipeline early warning models through a model accuracy algorithm. The specific calculation formula is as follows:

[0042]

[0043] In the formula, PR represents the accuracy value of the pipeline early warning model, TP represents the number of sample data with correct early warning results, and FP represents the number of sample data with incorrect early warning results. The accuracy verification results of each pipeline early warning model are arranged in descending order, and the reliability of the pipeline early warning model is assigned 0.1, 0.2, 0.3 and 0.4 respectively according to the order. The pipeline early warning model output module 303 is used to obtain the output results of each pipeline early warning model based on the processed real-time relevant data of the power plant pipeline, and collect and record the output results of each pipeline early warning model respectively;

[0044] The intelligent early warning unit 400 includes an early warning identification and judgment module 401 and an early warning notification module 402. The early warning identification and judgment module 401 is used to identify and judge the early warning type of the power plant pipeline from the two aspects of reliability and repeatability. The operation of identifying and judging the early warning type of the power plant pipeline from the aspect of reliability is to calculate the early warning terminal value of the pipeline through the pipeline early warning algorithm based on the output results of each pipeline early warning model. The specific calculation formula is as follows:

[0045]

[0046] Where i represents the i-th pipeline early warning model, v i represents the reliability value of the i-th pipeline early warning model, R i represents the output result of the i-th pipeline early warning model, and Ri ∈[0, 100], and the pipeline warning final value is divided into 5 pipeline warning levels with a tolerance of 20 according to the value range, and when 0≤calculation result<20, it is judged as pipeline warning level one, and when 80≤calculation result≤100, it is judged as pipeline warning level five, wherein the operation of identifying and judging the pipeline warning type of the power plant from the aspect of repetition is that when the output results of the four pipeline warning models are all lower than 60, it is judged as pipeline warning level one, when the output result of one pipeline warning model is greater than 60, it is judged as warning level two, when the output results of two pipeline warning models are greater than 60, it is judged as warning level three, when the output results of three pipeline warning models are greater than 60, it is judged as warning level four, when the output results of four pipeline warning models are greater than 60, it is judged as warning level five, the warning notification module 402 is used to notify the warning message based on the judgment result of the warning identification and judgment module 401, and after the corresponding pipeline is processed based on the warning message, the warning message is ignored to prevent repeated processing;

[0047] A method for unmanned intelligent monitoring of a power plant, comprising the following steps:

[0048] S1: The pipeline data acquisition module 101 collects the real-time relevant data of the pipelines in the power plant, the pipeline data processing module 102 processes the collected real-time relevant data of the pipelines in the power plant, and the pipeline data recording module 103 stores and records the processed real-time relevant data of the pipelines in the power plant;

[0049] S2: The pipeline early warning knowledge of the power plant is acquired through the pipeline early warning knowledge acquisition module 201, the pipeline early warning knowledge base of the power plant is constructed through the pipeline early warning knowledge base generation module 202, the pipeline early warning rules of the power plant are constructed based on the pipeline early warning knowledge of the power plant through different representation methods through the pipeline early warning knowledge representation module 203, and the pipeline early warning knowledge in the pipeline early warning knowledge base of the power plant is regularly self-updated through the pipeline early warning knowledge self-updating module 204;

[0050] S3: construct multiple pipeline early warning models through the pipeline early warning model generation module 301, perform model training and model accuracy verification on four pipeline early warning models through the pipeline early warning model training module 302, collect and record the output results of each pipeline early warning model through the pipeline early warning model output module 303, identify and judge the power plant pipeline early warning type from the two aspects of reliability and repeatability through the early warning identification and judgment module 401, and issue an early warning message notification based on the judgment result of the early warning identification and judgment module 401 through the early warning notification module 402.

[0051] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0052] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An unmanned intelligent monitoring system for a power plant, characterized by: The invention comprises a power plant pipeline data acquisition unit (100). Taking the unmanned intelligent monitoring system for power plant pipelines as an example, the power plant pipeline data acquisition unit (100) is used to collect real-time relevant data of pipelines in the power plant, process the collected real-time relevant data of pipelines in the power plant for subsequent use, and store and record the processed real-time relevant data of pipelines in the power plant; A pipeline early warning knowledge base construction unit (200), the pipeline early warning knowledge base construction unit (200) is used to acquire power plant pipeline early warning knowledge, generate a power plant pipeline early warning knowledge base based on the acquired power plant pipeline early warning knowledge, further construct power plant pipeline early warning rules based on the power plant pipeline early warning knowledge through different representation methods, and self-update the power plant pipeline early warning knowledge in the power plant pipeline early warning knowledge base; A pipeline early warning model construction unit (300), the pipeline early warning model construction unit (300) is used to construct a plurality of pipeline early warning models, and to perform model training and model accuracy verification on the various pipeline early warning models, and to determine the reliability of each pipeline early warning model based on the accuracy verification results of each pipeline early warning model, and to collect and record the output results of each pipeline early warning model; An intelligent early warning unit (400) is used to identify and judge the power plant pipeline early warning type from the two aspects of reliability and repeatability based on the output results of each pipeline early warning model, and to notify early warning information based on the judgment result.

2. According to claim 1, the unmanned intelligent monitoring system for power plants is characterized by: The power plant pipeline data acquisition unit (100) comprises a pipeline data acquisition module (101), and the pipeline data acquisition module (101) is used to collect real-time relevant data of pipelines in the power plant. The specific operation is to divide the power plant into regions, and perform a naming operation on each region in the power plant, the content of which is a combination of the region and the digital codes connected in sequence, and perform a naming operation on each pipeline in each region in the power plant, the content of which is a combination of the region naming content, the pipeline and the digital codes connected in sequence, so as to achieve accurate positioning of each section of the pipeline, and acquire real-time data generated during the operation of the pipeline through the provided pressure transmitter, temperature transmitter and flow computer.

3. According to claim 1, the unmanned intelligent monitoring system for power plants is characterized by: The power plant pipeline data acquisition unit (100) further comprises a pipeline data processing module (102) and a pipeline data recording module (103). The pipeline data processing module (102) is used to process the collected real-time related data of the pipeline in the power plant. Specifically, the operation is to perform missing value detection and abnormal value detection on the acquired real-time related data of the pipeline in the power plant. When missing values ​​are detected in the real-time related data of the pipeline in the power plant, the mean of the four node data adjacent to the missing value is used to fill the missing value so as to retain the time series characteristics of the pipeline data. When abnormal values ​​are detected in the real-time related data of the pipeline in the power plant, the mean of the remaining data is used to replace the abnormal values, and the processed pipeline data is normalized to eliminate the influence of the dimensions and magnitudes of different data types. The pipeline data recording module (103) is used to store and record the processed real-time related data of the pipeline in the power plant, and establish association with the naming content information of the corresponding pipeline.

4. The unmanned intelligent monitoring system for a power plant according to claim 1 is characterized by: The pipeline early warning knowledge base construction unit (200) comprises a pipeline early warning knowledge acquisition module (201) and a pipeline early warning knowledge base generation module (202). The pipeline early warning knowledge acquisition module (201) is used to acquire the pipeline early warning knowledge of the power plant, specifically to acquire data required for early warning from the pipeline database and pipeline early warning materials by manual acquisition, and further to supplement the pipeline early warning knowledge by consulting relevant field literature, pipeline early warning technical manuals, pipeline early warning professional books and pipeline early warning historical cases, and in combination with a semi-automatic acquisition method, to acquire pipeline early warning expert experience by holding talks with pipeline early warning professional technicians and business management personnel. The pipeline early warning knowledge base generation module (202) is used to construct a power plant pipeline early warning knowledge base based on the pipeline early warning knowledge acquired by the pipeline early warning knowledge acquisition module (201).

5. The unmanned intelligent monitoring system for a power plant according to claim 1 is characterized by: The pipeline early warning knowledge base construction unit (200) further comprises a pipeline early warning knowledge representation module (203) and a pipeline early warning knowledge self-updating module (204). The pipeline early warning knowledge representation module (203) is used to construct power plant pipeline early warning rules based on power plant pipeline early warning knowledge through different representation methods. The specific operation is to intuitively and clearly represent the pipeline early warning knowledge through a production rule representation method and a framework representation method, thereby eliminating the defect that a single representation method cannot accurately represent the pipeline early warning knowledge, and to construct power plant pipeline early warning rules based on the represented pipeline early warning knowledge, thereby providing a basis for subsequent pipeline early warning detection. The pipeline early warning knowledge self-updating module (204) is used to regularly self-update the pipeline early warning knowledge in the power plant pipeline early warning knowledge base through an extended isolation forest method.

6. The unmanned intelligent monitoring system for a power plant according to claim 1 is characterized by: The pipeline early warning model construction unit (300) comprises a pipeline early warning model generation module (301), wherein the pipeline early warning model generation module (301) is used to construct a plurality of pipeline early warning models, and the specific operation comprises selecting an LSTM model as a basic model, combining an ARIMA model with an LSTM model to construct an ARIMA-LSTM hybrid model, combining a self-attention mechanism with an LSTM model to construct an SA-LSTM hybrid model, combining a CNN network with an LSTM model to construct a CNN-LSTM hybrid model, and optimizing parameters of the LSTM model by a Bayesian method to construct a BO-LSTM hybrid model.

7. The unmanned intelligent monitoring system for a power plant according to claim 1 is characterized by: The pipeline early warning model construction unit (300) further comprises a pipeline early warning model training module (302) and a pipeline early warning model output module (303). The pipeline early warning model training module (302) is used to perform model training and model accuracy verification on four pipeline early warning models. The specific operation is to obtain pipeline early warning data through a network crawler, and construct a pipeline early warning data set based on the obtained pipeline early warning data. The four pipeline early warning models are respectively trained through the constructed pipeline early warning data set. The accuracy of the trained four pipeline early warning models is verified through a model accuracy algorithm. The accuracy verification results of each pipeline early warning model are arranged in descending order, and the reliability of the pipeline early warning model is assigned 0.1, 0.2, 0.3 and 0.4 respectively according to the order. The pipeline early warning model output module (303) is used to obtain the output results of each pipeline early warning model based on the processed real-time relevant data of the power plant pipeline, and the output results of each pipeline early warning model are collected and recorded respectively.

8. The unmanned intelligent monitoring system for a power plant according to claim 1 is characterized by: The intelligent early warning unit (400) comprises an early warning identification and judgment module (401) and an early warning notification module (402), wherein the early warning identification and judgment module (401) is used to identify and judge the early warning type of the power plant pipeline from the two aspects of reliability and repeatability, wherein the operation of identifying and judging the early warning type of the power plant pipeline from the aspect of reliability is to calculate the early warning final value of the pipeline through the pipeline early warning algorithm based on the output results of each pipeline early warning model, and to divide the pipeline early warning final value into 5 pipeline early warning levels with a tolerance of 20 according to the value range of the pipeline early warning final value, and when 0≤calculation result<20, it is judged as pipeline early warning level one, and when 80≤calculation result≤100, it is judged as pipeline early warning level five, wherein the operation of identifying and judging the early warning type of the power plant pipeline from the aspect of repeatability is to calculate the early warning final value of the pipeline through the pipeline early warning algorithm based on the output results of each pipeline early warning model, and to divide the pipeline early warning final value into 5 pipeline early warning levels with a tolerance of 20, and to judge the pipeline early warning level one when 0≤calculation result<20, and to judge the pipeline early warning level five when 80≤calculation result≤100, The operation of identifying and judging the model is that when the output results of the four pipeline early warning models are all lower than 60, it is judged as the pipeline early warning level one; when the output result of one pipeline early warning model is greater than 60, it is judged as the early warning level two; when the output results of two pipeline early warning models are greater than 60, it is judged as the early warning level three; when the output results of three pipeline early warning models are greater than 60, it is judged as the early warning level four; when the output results of the four pipeline early warning models are all greater than 60, it is judged as the early warning level five; the early warning notification module (402) is used to notify the early warning message based on the judgment result of the early warning identification and judgment module (401), and after the corresponding pipeline is processed based on the early warning message, the early warning message is ignored to prevent repeated processing.

9. An unmanned intelligent monitoring method for a power plant, characterized in that The following steps are involved: S1: collecting real-time relevant data of pipelines in a power plant through a pipeline data collection module (101), processing the collected real-time relevant data of pipelines in the power plant through a pipeline data processing module (102), and storing and recording the processed real-time relevant data of pipelines in the power plant through a pipeline data recording module (103); S2: acquiring power plant pipeline warning knowledge through a pipeline warning knowledge acquisition module (201), constructing a power plant pipeline warning knowledge base through a pipeline warning knowledge base generation module (202), constructing power plant pipeline warning rules based on the power plant pipeline warning knowledge through different representation methods through a pipeline warning knowledge representation module (203), and regularly self-updating the pipeline warning knowledge in the power plant pipeline warning knowledge base through a pipeline warning knowledge self-updating module (204); S3: constructing a plurality of pipeline early warning models through the pipeline early warning model generation module (301), performing model training and model accuracy verification on four pipeline early warning models through the pipeline early warning model training module (302), collecting and recording the output results of each pipeline early warning model through the pipeline early warning model output module (303), identifying and judging the power plant pipeline early warning type from the two aspects of reliability and repeatability through the early warning identification and judgment module (401), and issuing an early warning message notification based on the judgment result of the early warning identification and judgment module (401) through the early warning notification module (402).

Images