Power plant boiler real-time monitoring method and system based on multi-parameter fusion

Abstract

The invention discloses a power plant boiler real-time monitoring method and system based on multi-parameter fusion, and relates to the technical field of power plant boiler monitoring, and the method comprises the steps: obtaining historical working condition data and current working condition data of the operation of a power plant boiler; dividing the historical working condition data into a plurality of time periods; expanding the working condition data of each time period into a corresponding historical working condition vector; expanding the current working condition data into a corresponding current working condition vector; respectively calculating the similarity between the current working condition vector and each historical working condition vector to obtain a plurality of similarity results; sorting the plurality of similarity results to generate a candidate set; and performing iterative indexing on the sorting result of the candidate set so as to realize a monitoring service according to a similarity sorting result and obtain a monitoring result. Real-time monitoring and real-time early warning of the power plant boiler can be achieved, the safety accident prevention capacity in the boiler operation process is improved, and risks generated in the boiler operation process are effectively avoided in time.

Description

Technical Field

[0001] The present application relates to the technical field of power plant boiler monitoring, and in particular to a real-time monitoring method and system for power plant boilers based on multi-parameter fusion. Background Art

[0002] Power plant boilers are medium-to-large boilers that provide a specified quantity and quality of steam to the steam turbines in a power plant. They are one of the primary thermal equipment in thermal power plants, and their stable operation is crucial for ensuring a stable power supply. However, power plant boilers often encounter various faults during operation, which can seriously impact their stable operation.

[0003] Traditional boiler fault identification methods usually identify boiler operating faults through external abnormal characteristics of the boiler. Due to the complex structure of power plant boilers and the large number of internal components, this method cannot detect boiler faults in a timely and accurate manner, which has a great impact on the operating stability of power plant boilers.

[0004] Currently, the monitoring method driven by sensor data is mainly used to solve the problem of monitoring boiler fault status, and is rarely used to solve the problem of real-time status monitoring when the boiler operates according to time status. Summary of the Invention

[0005] The purpose of this application is to provide a real-time monitoring method and system for power plant boilers based on multi-parameter fusion, which can realize real-time monitoring and real-time early warning of power plant boilers, improve the ability to prevent safety accidents during boiler operation, and timely and effectively avoid risks arising during boiler operation.

[0006] To achieve the above objectives, this application provides the following solutions:

[0007] In a first aspect, the present application provides a real-time monitoring method for a power plant boiler based on multi-parameter fusion, the real-time monitoring method for a power plant boiler based on multi-parameter fusion comprising:

[0008] Obtain historical operating condition data and current operating condition data of the power plant boiler; the historical operating condition data includes: historical preset and generated parameter data; the current operating condition data includes: current preset and generated parameter data.

[0009] The historical operating condition data is divided into several time periods.

[0010] The operating condition data of each time period is expanded into the corresponding historical operating condition vector.

[0011] The current operating condition data is expanded into a corresponding current operating condition vector.

[0012] The similarity between the current operating condition vector and each of the historical operating condition vectors is calculated respectively to obtain a plurality of similarity results.

[0013] Several similarity results are sorted to generate a candidate set.

[0014] The ranking results of the candidate sets are iteratively indexed to implement monitoring services according to the similarity ranking results to obtain monitoring results.

[0015] Optionally, historical and current operating data of the power plant boiler are obtained, including:

[0016] The collected parameter data are processed to obtain processed parameter data; the data processing includes removing abnormal values ​​and duplicate data.

[0017] Feature extraction is performed on the processed parameter data to obtain target acquisition parameters.

[0018] Add the target acquisition parameters to the information extraction node.

[0019] Key information is acquired from the information extraction node to obtain target key information.

[0020] Target feature data is determined based on the target key information.

[0021] The target feature data is combined with the information extraction node to extract parameter data features of the target analysis object, and the historical operating condition data and current operating condition data of the power plant boiler are obtained.

[0022] Optionally, the operating condition data of each time period is expanded into a corresponding historical operating condition vector, specifically including:

[0023] The factory parameters, preset parameters and historically generated parameters of various components of the power plant boiler are collected and sorted to obtain a first multidimensional time series of the historical operating status of the power plant boiler, and the first multidimensional time series is stored in a multidimensional format.

[0024] The first multidimensional time series is encoded into a high-dimensional vector space to obtain a historical operating condition vector.

[0025] Optionally, expanding the current operating condition data into a corresponding current operating condition vector specifically includes:

[0026] The factory parameters, preset parameters and currently generated parameters of each component of the power plant boiler are collected and sorted to obtain a second multidimensional time series of the current operating state of the power plant boiler, and the second multidimensional time series is stored in a multidimensional format.

[0027] The second multidimensional time series is encoded into a high-dimensional vector space to obtain a current operating condition vector.

[0028] Optionally, sorting the similarity results to generate a candidate set specifically includes:

[0029] The plurality of similarity results are screened according to a preset similarity threshold to obtain screened similarity results.

[0030] According to the filtered similarity results, all corresponding filtered historical operating condition vectors are determined.

[0031] All the filtered historical operating condition vectors are sorted into a list using Euclidean distance according to a clustering algorithm to generate a candidate set.

[0032] Optionally, the sorting results of the candidate sets are iteratively indexed to implement monitoring services based on the similarity sorting results, specifically including:

[0033] Each of the filtered historical operating condition vectors is assigned a number of time period parameter data; the states of the time period parameter data include: continuous / intermittent.

[0034] The power plant boiler operating condition data of the future time period after the timestamp corresponding to each time period parameter data is traversed and loaded to obtain the operating condition status of the power plant boiler under each time parameter at the historical moment.

[0035] According to the list order of the candidate set, the current operating condition vector is compared with the historical operating condition vector to obtain a comparison result.

[0036] Based on the comparison results, the operating status of the power plant boiler in the current and future time periods is judged to achieve real-time monitoring of the power plant boiler.

[0037] Optionally, the method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes: filling in the acquired historical operating condition data of the power plant boilers, specifically including:

[0038] Obtain historical data of boilers in power plants of the same level or all boilers in the power plant, query all normally operating boilers, record the normally operating boilers, and collect data from various sensors.

[0039] The historical operating time of the normally operating boiler is divided into a plurality of time periods, and the nodes between each two adjacent time periods are used as boiler steady-state nodes.

[0040] The sensor data at the two time sides of the steady-state node are recorded and queried, and recorded as data A.

[0041] The data on both time sides of the lost operating condition data are collected and recorded as operating condition data B.

[0042] The data A and the operating condition data B are compared one by one for similarity, and a comparison threshold is set. If the threshold screening requirements are met, the operating condition data of the steady-state node is filled in the location node of the lost operating condition data. If not, the index is traversed until a matching steady-state node is found.

[0043] Optionally, the method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes: establishing a similarity comparison mathematical model, specifically including:

[0044] Load parameter data for a certain time period or the entire time period near the historical operating condition vector that is closest to the current operating condition vector.

[0045] According to the loaded parameter data, the parameter data after the future time period is read with any time as the starting point, and the data after the future time period is used as the current working condition vector to input the similarity comparison mathematical model to obtain the historical prediction results.

[0046] The current operating condition vector is input into the similarity comparison mathematical model to obtain the current prediction result.

[0047] It is determined whether the vector distance between the historical prediction result and the current prediction result is less than a threshold value to obtain a determination result.

[0048] If the judgment result is yes, the final similarity comparison mathematical model is obtained.

[0049] If the judgment result is no, the current prediction result is eliminated, and the eliminated current prediction result is used as an error example to be included in the negative sample training set to correct the parameters of the similarity comparison mathematical model to obtain the final similarity comparison mathematical model.

[0050] Optionally, the method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes: incorporating the scoring results into real-time monitoring judgment, specifically including:

[0051] Through the clustering algorithm, various risk data related to the boiler operating status are screened out, and irrelevant risk data are deleted to form a risk identification report.

[0052] Normalization processing is used to eliminate the amplitude differences between the data in the risk identification report to obtain a risk identification report after elimination.

[0053] A risk assessment model is used to score the importance of each risk data relative to the index in the risk identification report after elimination to obtain an assessment result; the risk assessment model is a model established based on the weight index obtained by expert scoring and the degree of correlation between the boiler.

[0054] The scoring results are incorporated into real-time monitoring and judgment.

[0055] In a second aspect, the present application provides a real-time monitoring system for power plant boilers based on multi-parameter fusion, the real-time monitoring system for power plant boilers based on multi-parameter fusion comprising:

[0056] The data acquisition module is used to obtain historical operating data and current operating data of the power plant boiler; the historical operating data includes: historical preset and generated parameter data; the current operating data includes: current preset and generated parameter data.

[0057] The data division module is used to divide the historical operating condition data into several time periods.

[0058] The first expansion module is used to expand the operating condition data of each time period into a corresponding historical operating condition vector.

[0059] The second expansion module is used to expand the current operating condition data into a corresponding current operating condition vector.

[0060] The similarity calculation module is used to respectively calculate the similarity between the current operating condition vector and each of the historical operating condition vectors to obtain a plurality of similarity results.

[0061] The sorting module is used to sort the similarity results and generate a candidate set.

[0062] The iterative indexing module is used to iteratively index the sorting results of the candidate sets to implement monitoring services according to the similarity sorting results and obtain monitoring results.

[0063] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0064] The present application provides a method and system for real-time monitoring of power plant boilers based on multi-parameter fusion. First, the historical operating condition data and current operating condition data of the power plant boiler are obtained; the historical operating condition data include: historical preset and generated parameter data; the current operating condition data include: current preset and generated parameter data; secondly, the historical operating condition data are divided into several time periods; the operating condition data of each time period is expanded into a corresponding historical operating condition vector; the current operating condition data is expanded into a corresponding current operating condition vector; then, the similarity between the current operating condition vector and each of the historical operating condition vectors is calculated to obtain several similarity results; the several similarity results are sorted to generate a candidate set; finally, the sorted results of the candidate set are iteratively indexed to implement monitoring services based on the similarity sorting results to obtain monitoring results. The present application predicts and evolves historical parameter data through a model to achieve real-time assessment of the current boiler operating status and whether there are risks. Real-time monitoring and real-time early warning of power plant boilers are achieved, improving the ability to prevent safety accidents during boiler operation and timely and effectively avoiding risks generated during boiler operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0066] Figure 1 This is an application environment diagram of a real-time monitoring method for power plant boilers based on multi-parameter fusion in one embodiment of the present application.

[0067] Figure 2 A flowchart of a method for real-time monitoring of power plant boilers based on multi-parameter fusion provided in one embodiment of the present application.

[0068] Figure 3 A schematic diagram of the functional modules of a real-time monitoring system for power plant boilers based on multi-parameter fusion provided in one embodiment of the present application. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0070] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0071] The real-time monitoring method for power plant boilers based on multi-parameter fusion provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the terminal 102 communicates with the server 104 via a network. The data storage system can store data that the server 104 needs to process. The data storage system can be set up separately, integrated on the server 104, or placed on the cloud or other servers. Terminal 102 may transmit the acquired historical and current operating condition data of the power plant boiler to server 104. The historical operating condition data includes historical preset and generated parameter data; the current operating condition data includes current preset and generated parameter data. After receiving the historical and current operating condition data of the power plant boiler, server 104 may divide the historical operating condition data into a number of time periods; expand the operating condition data for each time period into a corresponding historical operating condition vector; expand the current operating condition data into a corresponding current operating condition vector; calculate the similarity between the current operating condition vector and each of the historical operating condition vectors to obtain a number of similarity results; sort the similarity results to generate a candidate set; and iteratively index the sorted results of the candidate set to implement monitoring services based on the similarity sorting results to obtain monitoring results. Server 104 may provide feedback of the obtained monitoring results to terminal 102. In addition, in some embodiments, the real-time monitoring method of a power plant boiler based on multi-parameter fusion can also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 can directly monitor the historical operating condition data and current operating condition data of the power plant boiler, or the server 104 can obtain the historical operating condition data and current operating condition data of the power plant boiler from the data storage system and monitor the historical operating condition data and current operating condition data of the power plant boiler.

[0072] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, and tablet computers. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers, or a cloud server.

[0073] In an exemplary embodiment, Figure 2 As shown, a method for real-time monitoring of power plant boilers based on multi-parameter fusion is provided. The method is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to Figure 1The server 104 in the example is used for explanation, and the steps include the following steps S1 to S7.

[0074] in:

[0075] S1: Acquire historical operating condition data and current operating condition data of a power plant boiler; the historical operating condition data includes historical preset and generated parameter data; the current operating condition data includes current preset and generated parameter data.

[0076] S2: Divide the historical operating condition data into several time periods.

[0077] S3: Expand the operating condition data of each time period into the corresponding historical operating condition vector.

[0078] S4: Expand the current operating condition data into a corresponding current operating condition vector.

[0079] S5: Calculate the similarity between the current operating condition vector and each of the historical operating condition vectors respectively to obtain a plurality of similarity results.

[0080] S6: Sort the similarity results to generate a candidate set.

[0081] S7: Iteratively index the ranking results of the candidate set to implement monitoring services according to the similarity ranking results to obtain monitoring results.

[0082] By implementing steps S1 to S7 above, the model predicts the evolution of historical parameter data to achieve real-time assessment of the current boiler operating status and whether there are risks. This enables real-time monitoring and early warning of power plant boilers, improves the ability to prevent safety accidents during boiler operation, and effectively and promptly mitigates risks arising from boiler operation.

[0083] In another exemplary embodiment of the present application, step S1 specifically includes:

[0084] S11: Processing the collected parameter data to obtain processed parameter data; the data processing includes removing abnormal values ​​and duplicate data.

[0085] S12: Perform feature extraction on the processed parameter data to obtain target acquisition parameters; the feature extraction obtains initial information collected by at least one data acquisition terminal, and obtains target acquisition information for the target analysis object from the initial information, and converts the target acquisition information into target acquisition parameters.

[0086] S13: Add the target acquisition parameters to the information extraction node.

[0087] S14: Acquire key information from the information extraction node to obtain target key information.

[0088] S15: Determine target feature data based on the target key information.

[0089] S16: Combining the target feature data with the information extraction node to perform parameter data feature extraction on the target analysis object, thereby obtaining historical and current operating condition data of the power plant boiler. Specifically, the target feature data is combined with at least one data information extraction node of the data acquisition terminal to perform parameter data feature extraction on the target analysis object.

[0090] It should be noted that if the historical working condition data are retained, they can be directly exported for acquisition, or acquired based on the historical project list and details. These parameter data only need to include the pipeline factory parameter data, environmental parameter data during construction, and damaged parameter data. These parameters can be preset in advance and generated during the process.

[0091] As an optional implementation manner, the aforementioned several time periods can be understood as several historical time periods of the boiler.

[0092] The factory parameters, preset parameters, and real-time parameters of each boiler component are collected and organized to obtain a multidimensional time series that can represent the operating status of the boiler. These data are stored in a multidimensional format and encoded into a high-dimensional vector space. This allows the historical parameter data to be converted into a historical operating condition vector.

[0093] As an optional implementation, in step S3, it specifically includes:

[0094] S31: Collect and organize factory parameters, preset parameters, and historically generated parameters of various components of the power plant boiler to obtain a first multidimensional time series of historical operating states of the power plant boiler, and store the first multidimensional time series in a multidimensional format.

[0095] S32: Encode the first multidimensional time series into a high-dimensional vector space to obtain a historical operating condition vector.

[0096] Similarly, in step S4, it specifically includes:

[0097] S41: Collect and organize factory parameters, preset parameters, and currently generated parameters of various components of the power plant boiler to obtain a second multidimensional time series of the current operating state of the power plant boiler, and store the second multidimensional time series in a multidimensional format.

[0098] S42: Encode the second multi-dimensional time series into a high-dimensional vector space to obtain a current operating condition vector.

[0099] Specifically, the current operating condition data is read, and data acquisition coding is used to generate the current operating condition vector. This is then compared to the historical operating condition vector, and the distance between the historical operating condition vector and the current operating condition vector is calculated. Then, the construction parameter data for the entire time period, or the historical operating condition vector with the smallest distance that meets a threshold (the distance threshold should be set in advance), is loaded. As can be understood, since this is historical data, the results are known and can therefore be used as a reference for the changing trend of the current operating condition data after continued operation, that is, as a reference for real-time monitoring and risk assessment and early warning.

[0100] This application also has the ability to predict the operating status of the current boiler in the future. Specifically, it includes the following steps:

[0101] A1: Set a similarity threshold to obtain all corresponding historical operating condition vectors.

[0102] A2: Using the Euclidean distance according to the clustering algorithm, the historical operating condition vectors are sorted into a list to generate a candidate set.

[0103] A3: Assign time period parameters t1, t2, ..., tn to the historical operating condition vector, where the states of the time period parameters include continuous / discontinuous.

[0104] A4: Traverse and load the boiler operating condition data for the t0 time period after the timestamp corresponding to each time period parameter to obtain the operating condition status of the boiler under each time parameter at the historical moment.

[0105] A5: According to the order of the similarity candidate set list, the current operating condition vector is compared with the historical operating condition vector, and the operating condition status of the boiler in the current and future time period t0 is judged based on the comparison result to achieve real-time monitoring of the boiler.

[0106] As an optional implementation, in step S5, it specifically includes:

[0107] S51: screening a plurality of similarity results according to a preset similarity threshold to obtain screened similarity results.

[0108] S52: Determine all corresponding filtered historical operating condition vectors according to the filtered similarity results.

[0109] S53: sorting all the filtered historical operating condition vectors into a list using the Euclidean distance according to the clustering algorithm to generate a candidate set.

[0110] As an optional implementation, in step S6, it specifically includes:

[0111] S61: assigning a number of time period parameter data to each of the filtered historical operating condition vectors; the status of the time period parameter data includes: continuous / intermittent.

[0112] S62: Traverse the power plant boiler operating condition data of the future time period after the timestamp corresponding to the loaded parameter data of each time period to obtain the operating condition status of the power plant boiler under each time parameter at the historical moment.

[0113] S63: Compare the current operating condition vector with the historical operating condition vector according to the list order of the candidate set to obtain a comparison result.

[0114] S64: Based on the comparison result, the operating status of the power plant boiler in the current and future time periods is determined to achieve real-time monitoring of the power plant boiler.

[0115] Due to sensor failure and maintenance, data loss in the database should be a common problem, but this leads to incomplete data and affects the robustness of the condition monitoring solution of this application.

[0116] Acquiring historical operating data of power plant boilers also includes:

[0117] S8: Fill in the acquired historical operating data of the power plant boiler, including:

[0118] S81: Obtain historical data of boilers in power plants of the same level or all boilers in the power plant, query all normally operating boilers, record the normally operating boilers, and collect data from various sensors;

[0119] S82: Divide the historical operating time of the normally operating boiler into multiple time periods, and use the nodes between each two adjacent time periods as boiler steady-state nodes;

[0120] S83: Record and query the sensor data at two time sides of the steady-state node, recorded as data A;

[0121] S84: Collect data on both time sides of the lost operating condition data and record them as operating condition data B;

[0122] S85: Compare the data A and the operating condition data B one by one for similarity, set a comparison threshold, and if they meet the threshold screening requirements, fill the operating condition data of the steady-state node at the location node where the lost operating condition data is located; if they do not meet the requirements, traverse the index until a matching steady-state node is found.

[0123] In another exemplary embodiment of the present application, the method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes:

[0124] S9: Establishing a similarity comparison mathematical model, specifically including:

[0125] S91: Load parameter data of a certain time period or a full time period near the historical operating condition vector closest to the current operating condition vector.

[0126] S92: Based on the loaded parameter data, parameter data after a future time period is read starting at any time, and the data after the future time period is used as the current operating condition vector to input a similarity comparison mathematical model to obtain a historical prediction result.

[0127] S93: Input the current operating condition vector into the similarity comparison mathematical model to obtain the current prediction result.

[0128] S94: Determine whether the vector distance between the historical prediction result and the current prediction result is less than a threshold, and obtain a determination result.

[0129] S95: If the judgment result is yes, then obtain the final similarity comparison mathematical model.

[0130] S96: If the judgment result is no, the current prediction result is eliminated, and the eliminated current prediction result is used as an error example to be included in the negative sample training set to correct the parameters of the similarity comparison mathematical model to obtain a final similarity comparison mathematical model.

[0131] Specifically, it includes the following:

[0132] B1: Obtain historical boiler parameters x, operating parameters y, and damaged parameters z to form a historical operating condition vector [x, y, z].

[0133] B2: Obtain the current boiler parameter x', operating parameter y', and damaged parameter z to form the current operating condition vector [x', y', z'].

[0134] B3: Load parameter data for a certain time period or the entire time period near the historical operating condition vector that is closest to the current operating condition vector.

[0135] B4: Based on the loaded parameter data, the parameter data after the time period t0 is read with any time as the starting point, and the data after the time period t0 is used as the current operating condition vector input to obtain the historical prediction result.

[0136] B5: Input the current operating condition vector and compare it with the historical prediction results. The vector distance between the two should be less than the threshold. Otherwise, the prediction results with a distance greater than the threshold are eliminated and the model parameters are corrected.

[0137] B6: Include the eliminated prediction results as incorrect examples in the negative sample training set to retrain the model, and update and save the prediction model package and sample data in real time.

[0138] In another exemplary embodiment of the present application, the method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes:

[0139] S10: Incorporate scoring results into real-time monitoring and judgment, including:

[0140] S101: Filter out various risk data related to the boiler operating status through a clustering algorithm, delete irrelevant risk data, and form a risk identification report.

[0141] S102: using normalization processing to eliminate the amplitude differences between the data in the risk identification report to obtain a risk identification report after elimination.

[0142] S103: Using a risk assessment model to score the importance of each risk data relative to the index in the risk identification report after elimination, to obtain an assessment result; the risk assessment model is a model established based on the weight index obtained by expert scoring and the degree of correlation between the boiler.

[0143] S104: Incorporating the scoring results into real-time monitoring and judgment.

[0144] In another exemplary embodiment of the present application, the real-time monitoring method for power plant boilers based on multi-parameter fusion further includes: safety risk mining, mining the correlation between various risk data and risk events based on the parameter data generated during boiler operation and the basic information of the boiler, and establishing a preliminary risk list through a clustering algorithm, wherein the clustering algorithm selects one object from all data samples and defines it as d x As the first cluster center, calculate the Euclidean distance between all data points and the cluster center. The specific calculation method is as follows:

[0145]

[0146] Where, X a X b All represent sample data, d(X a X b ) represents the Euclidean distance between two samples in the m-dimensional space. The straight-line distance between two points in the Euclidean space is the Euclidean distance. The similarity between samples is evaluated by the Euclidean distance. The risk data related to the boiler operation status are screened out through the clustering algorithm, and irrelevant risk factors are deleted to form a risk identification report.

[0147] The risk assessment model is based on the correlation between the weight index and the boiler obtained by different experts, and the weight of the monitoring index is obtained through the expert scoring. The n experts score the weight of the monitoring index and score the importance of each risk data relative to the monitoring index to obtain the original data of each monitoring index score D = (D1, D2, ... D n ), the original data are numbered starting from 0 in descending order, and reordered to obtain the vector H = (H0, H1, ...H n-1 ), the weight of the data in the vector H is determined by the number of combinations Determine and record the weight vector as J=(J0, J1, ...J k-1 ),in l=0,1,...,k-1, use weighted vector J to perform weighted calculation on vector H to obtain the absolute weight W of risk assessment index o ,as follows:

[0148]

[0149] o=1,2...,z;

[0150] In the formula, the subscript o represents the serial number of the risk indicator among the indicators at the same level, and z is the total number of evaluation indicators at the same level.

[0151] The relative weight of the risk assessment index is R o , the specific calculation formula is as follows:

[0152]

[0153] o=1,2,...z;

[0154] The risk level is defined as a binary function that includes the probability of an adverse event occurring and the consequences of the event. The calculation formula is as follows:

[0155] R = LI;

[0156] Where R∈[0,1] represents the risk level, L∈[0,1] represents the probability of risk occurrence, and I∈[0,1] represents the consequences of risk occurrence.

[0157] The weighted indicators obtained by the experts' scores are divided into four levels according to a, b, c, and d. The membership function of the monitoring indicator safety warning is established. The specific operations are as follows:

[0158]

[0159] Among them, 0, a is the no-alarm interval, a, b is the light-alarm interval, b, c is the medium-alarm interval, and c, d is the heavy-alarm interval. The indicators with alarm forecast are sorted from small to large to determine the alarm level of the safety accident and issue the corresponding alarm;

[0160] The risk assessment model also includes measuring the size of the risk probability and the severity of the consequences after the occurrence using the risk degree. The risk degree r depends on the risk probability g f and the severity of the consequences of risk occurrence C f , the relationship between them is expressed as:

[0161] r=g f +C f -g f C f ;

[0162]

[0163] Where g fi represents the probability of the i-th risk accident occurring, w i represents the weight of the i-th risk accident, C fj Indicates the severity of the j-th risk consequence and the probability of the risk accident occurring g f and risk consequence severity C f The classification of the security status according to this indicator.

[0164] The formula in this application is a formula that is closest to the actual situation obtained by removing the dimension and taking its numerical calculation by collecting a large amount of data and performing software simulation. The preset proportional coefficient in the formula is set by technical personnel in this field according to actual conditions or obtained through large-scale data simulation.

[0165] This application uses a model to predict the evolution of historical parameter data to achieve real-time assessment of the current boiler operating status and whether there are risks. This enables real-time monitoring and early warning of power plant boilers, improves the ability to prevent safety accidents during boiler operation, and effectively and timely avoids risks arising from boiler operation. In addition, this application fills in missing data in the database to improve the robustness of this application method.

[0166] The present application also provides an application scenario, which applies the above-mentioned real-time monitoring method for power plant boilers based on multi-parameter fusion. Specifically: the real-time monitoring method for power plant boilers based on multi-parameter fusion provided in this embodiment can be applied in the power plant boiler monitoring scenario. The power plant boiler monitoring scenario includes: data acquisition link, data division link, expansion link, similarity calculation link, sorting link and iterative indexing link; first, the historical operating condition data and current operating condition data of the power plant boiler are obtained; the historical operating condition data include: historical preset and generated parameter data; the current operating condition data include: current preset and generated parameter data; secondly, the historical operating condition data are divided into several time periods; the operating condition data of each time period is expanded into the corresponding historical operating condition vector; the current operating condition data is expanded into the corresponding current operating condition vector; then, the similarity between the current operating condition vector and each of the historical operating condition vectors is calculated respectively to obtain several similarity results; the several similarity results are sorted to generate a candidate set; finally, the sorted results of the candidate set are iteratively indexed to implement the monitoring business according to the similarity sorting results, and the monitoring results can be obtained.

[0167] Based on the same inventive concept, embodiments of the present application also provide a monitoring system for implementing the aforementioned method for real-time monitoring of power plant boilers based on multi-parameter fusion. The implementation solution provided by this system is similar to the implementation solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of the real-time monitoring system for power plant boilers based on multi-parameter fusion provided below can be found in the above-mentioned limitations of the method for real-time monitoring of power plant boilers based on multi-parameter fusion, and will not be repeated here.

[0168] In an exemplary embodiment, Figure 3 As shown, a real-time monitoring system for power plant boilers based on multi-parameter fusion is provided, and the real-time monitoring system for power plant boilers based on multi-parameter fusion includes:

[0169] The data acquisition module M1 is used to acquire historical operating condition data and current operating condition data of the power plant boiler; the historical operating condition data includes: historical preset and generated parameter data; the current operating condition data includes: current preset and generated parameter data.

[0170] The data division module M2 is used to divide the historical operating condition data into several time periods.

[0171] The first expansion module M3 is used to expand the operating condition data of each time period into a corresponding historical operating condition vector.

[0172] The second expansion module M4 is configured to expand the current operating condition data into a corresponding current operating condition vector.

[0173] The similarity calculation module M5 is used to respectively calculate the similarity between the current operating condition vector and each of the historical operating condition vectors to obtain a plurality of similarity results.

[0174] The sorting module M6 is used to sort the similarity results to generate a candidate set.

[0175] The iterative indexing module M7 is used to iteratively index the sorting results of the candidate sets to implement monitoring services according to the similarity sorting results and obtain monitoring results.

[0176] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0177] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A real-time monitoring method for power plant boilers based on multi-parameter fusion, characterized in that: The real-time monitoring method for power plant boilers based on multi-parameter fusion includes: Obtain historical operating condition data and current operating condition data of the power plant boiler; the historical operating condition data includes: historical preset and generated parameter data; the current operating condition data includes: current preset and generated parameter data; Dividing the historical operating condition data into a plurality of time periods; Expand the operating condition data of each time period into the corresponding historical operating condition vector; Expanding the current operating condition data into a corresponding current operating condition vector; respectively calculating the similarity between the current operating condition vector and each of the historical operating condition vectors to obtain a plurality of similarity results; Sorting the plurality of similarity results to generate a candidate set; The ranking results of the candidate sets are iteratively indexed to implement monitoring services according to the similarity ranking results to obtain monitoring results.

2. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1 is characterized in that: Obtain historical and current operating data of power plant boilers, including: Processing the collected parameter data to obtain processed parameter data; the data processing includes: removing abnormal values ​​and duplicate data; Performing feature extraction on the processed parameter data to obtain target acquisition parameters; Add the target acquisition parameters to the information extraction node; Acquiring key information from the information extraction node to obtain target key information; Determining target feature data based on the target key information; The target feature data is combined with the information extraction node to extract parameter data features of the target analysis object, and the historical operating condition data and current operating condition data of the power plant boiler are obtained.

3. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1, characterized in that: The operating condition data of each time period is expanded into the corresponding historical operating condition vector, including: Collecting and arranging factory parameters, preset parameters, and historically generated parameters of various components of the power plant boiler to obtain a first multidimensional time series of historical operating states of the power plant boiler, and storing the first multidimensional time series in a multidimensional format; The first multidimensional time series is encoded into a high-dimensional vector space to obtain a historical operating condition vector.

4. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1, characterized in that: Expanding the current operating condition data into a corresponding current operating condition vector specifically includes: Collecting and arranging factory parameters, preset parameters, and currently generated parameters of various components of the power plant boiler to obtain a second multidimensional time series of the current operating state of the power plant boiler, and storing the second multidimensional time series in a multidimensional format; The second multidimensional time series is encoded into a high-dimensional vector space to obtain a current operating condition vector.

5. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1, characterized in that: Sorting the similarity results to generate a candidate set includes: Filtering the plurality of similarity results according to a preset similarity threshold to obtain filtered similarity results; Determining all corresponding filtered historical operating condition vectors according to the filtered similarity results; All the filtered historical operating condition vectors are sorted into a list using Euclidean distance according to a clustering algorithm to generate a candidate set.

6. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 5 is characterized in that: Iteratively indexing the sorting results of the candidate set to implement monitoring services based on the similarity sorting results, specifically including: Assigning a number of time period parameter data to each of the filtered historical operating condition vectors; the states of the time period parameter data include: continuous / intermittent; Traverse and load the power plant boiler operating condition data for the future time period after the timestamp corresponding to the parameter data of each time period to obtain the operating condition status of the power plant boiler under each time parameter at the historical moment; Comparing the current operating condition vector with the historical operating condition vector according to the list order of the candidate set to obtain a comparison result; Based on the comparison results, the operating status of the power plant boiler in the current and future time periods is judged to achieve real-time monitoring of the power plant boiler.

7. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1, characterized in that: The method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes: filling in the acquired historical operating condition data of the power plant boilers, specifically including: Obtain historical data of boilers in power plants of the same level or all boilers in the power plant, query all normally operating boilers, record the normally operating boilers, and collect data from various sensors; Divide the historical operating time of the normally operating boiler into multiple time periods, and use the nodes between each two adjacent time periods as boiler steady-state nodes; Record and query the sensor data of the two time sides of the steady-state node, which is recorded as data A; Collect the data on both time sides of the lost working condition data and record them as working condition data B; The data A and the operating condition data B are compared one by one for similarity, and a comparison threshold is set. If the threshold screening requirements are met, the operating condition data of the steady-state node is filled in the location node of the lost operating condition data. If not, the index is traversed until a matching steady-state node is found.

8. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1, characterized in that: The method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes: establishing a similarity comparison mathematical model, specifically including: Load parameter data of a certain time period or the entire time period near the historical operating condition vector closest to the current operating condition vector; Based on the loaded parameter data, the parameter data after the future time period is read from any time as the starting point, and the data after the future time period is used as the current working condition vector to input the similarity comparison mathematical model to obtain the historical prediction results; Input the current operating condition vector into the similarity comparison mathematical model to obtain the current prediction result; Determine whether the vector distance between the historical prediction result and the current prediction result is less than a threshold value, and obtain a determination result; If the judgment result is yes, then obtaining the final similarity comparison mathematical model; If the judgment result is no, the current prediction result is eliminated, and the eliminated current prediction result is used as an error example to be included in the negative sample training set to correct the parameters of the similarity comparison mathematical model to obtain the final similarity comparison mathematical model.

9. The method for real-time monitoring of power plant boilers based on multi-parameter fusion according to claim 1, characterized in that: The method for real-time monitoring of power plant boilers based on multi-parameter fusion further includes: incorporating the scoring results into real-time monitoring judgment, specifically including: The clustering algorithm is used to filter out various risk data related to the boiler operating status, and irrelevant risk data is deleted to form a risk identification report; Eliminating the amplitude differences between the data in the risk identification report by normalization processing to obtain a risk identification report after elimination; A risk assessment model is used to score the importance of each risk data relative to the index in the risk identification report after elimination to obtain an assessment result; the risk assessment model is a model established based on the weight index obtained by the expert score and the degree of correlation between the boiler; The scoring results are incorporated into real-time monitoring and judgment.

10. A real-time monitoring system for power plant boilers based on multi-parameter fusion, characterized in that: The power plant boiler real-time monitoring system based on multi-parameter fusion includes: The data acquisition module is used to obtain historical operating data and current operating data of the power plant boiler; the historical operating data includes: historical preset and generated parameter data; the current operating data includes: current preset and generated parameter data; A data division module, used to divide the historical operating condition data into several time periods; A first expansion module is used to expand the operating condition data of each time period into a corresponding historical operating condition vector; A second expansion module, configured to expand the current operating condition data into a corresponding current operating condition vector; A similarity calculation module is used to calculate the similarity between the current operating condition vector and each of the historical operating condition vectors to obtain a plurality of similarity results; A sorting module, used to sort the similarity results to generate a candidate set; The iterative indexing module is used to iteratively index the sorting results of the candidate sets to implement monitoring services according to the similarity sorting results and obtain monitoring results.

Images