Intelligent prediction method for creep damage of last-stage superheater of power station boiler

Abstract

The invention discloses a power station boiler final-stage superheater creep damage intelligent prediction method, and relates to the field of special equipment safety prevention and control. The problem that the future damage of the material cannot be predicted due to the fact that the traditional creep prediction method adopts temperature stress data to perform damage evaluation on the material is solved. The method comprises the steps of collecting historical operation data of the last-stage superheater of the power station boiler and the like. Abnormal data in the historical operation data are processed, standardized parameters are stored, and a data set is created according to the training requirement of the prediction model; constructing a neural network model according to the selected time sequence prediction algorithm, inputting the data set into the neural network model, storing and compiling the final-stage superheater temperature prediction model, the pressure prediction model and the standardized parameters which meet the prediction precision requirement into codes, and performing packaging and calling to obtain a final-stage superheater temperature prediction model; and the predicted final-stage superheater temperature data and the like are called for creep damage calculation, and creep damage prediction of the final-stage superheater is achieved. The method is suitable for the field of prediction model construction and application methods.

Description

Technical Field

[0001] The present invention relates to the technical field of safety prevention and control of special equipment, and in particular to an intelligent prediction method for creep damage of a final superheater of a power station boiler. Background Art

[0002] Power station boilers are equipment that uses thermal energy such as coal, gas, and biomass to heat water or other media into steam for power generation. They are in service for a long time under high temperature and high pressure, and will inevitably produce local creep, fatigue and other damage, causing equipment failure. The final stage superheater is a typical pressure-bearing component of power station boilers. Once a failure occurs, it will not only cause a large amount of maintenance costs to the enterprise, reduce production efficiency and economic benefits, but also cause safety accidents and casualties.

[0003] A Chinese invention patent document with publication number CN114970241A discloses a method for predicting creep damage of high-temperature pressure-bearing structures. This method determines the creep rate based on the temperature and stress level of the material, thereby conducting creep damage of the material, simplifying the traditional creep assessment method. However, this method is essentially assessing the damage, not predicting the damage, and cannot predict the creep damage state of the material in the future.

[0004] The existing creep prediction method uses temperature stress data to evaluate material damage, but cannot predict future material damage. To predict creep damage, it is necessary to predict the future temperature and pressure data of the material application environment. In actual applications, power plant managers are more concerned about the future status of boiler equipment so that they can make treatment plans in advance to prevent accidents. Summary of the invention

[0005] The present invention aims to solve the problem that the creep prediction method in the prior art uses temperature stress data to perform damage assessment on materials, but cannot predict future damage to the materials.

[0006] To solve the above technical problems, the present invention is achieved through the following technical solutions:

[0007] Solution 1: The present invention proposes an intelligent prediction method for creep damage of a final superheater of a power plant boiler, the prediction method comprising the following steps:

[0008] S1. Collect historical operation data, material data, specification data, and historical test data of the final superheater of the power station boiler to establish a basic database;

[0009] S2. Processing abnormal data in the historical operation data described in S1;

[0010] S3, standardize the abnormal data processed in S2 and save the standardized parameters, and create a data set using the training requirements of the prediction model;

[0011] S4, constructing a neural network model according to the selected time series prediction algorithm, and transferring the data set described in S3 into the neural network model, training and optimizing the neural network model according to the set error constraints, and storing the final superheater temperature prediction model, pressure prediction model and standardized parameters that meet the prediction accuracy requirements;

[0012] S5, compile the final superheater temperature prediction model, final superheater pressure prediction model and standardized parameter file trained in S4 into code, and encapsulate and call them to predict the temperature and pressure data of the final superheater;

[0013] S6. Calling the temperature data, pressure data, material data and specification data of the final superheater predicted in S5 to perform creep damage calculation, so as to realize creep damage prediction of the final superheater.

[0014] Furthermore, a preferred embodiment is provided, wherein the historical operation data in S1 includes the DCS temperature and DCS pressure data of the final superheater of the power plant boiler; the material data includes the material grade and material performance data; the specification data includes the outer diameter and inner wall thickness of the tube of the final superheater; and the historical detection data mainly includes the detection data of the tube wall thickness of the final superheater.

[0015] Furthermore, a preferred implementation is provided, wherein the processing of abnormal data in the historical operation data described in S1 in S2 includes the steps of processing missing data and processing erroneous data.

[0016] Furthermore, a preferred implementation is provided, in which a data set is created in S3 according to the prediction model training requirements, and the created data set includes a training set, a test set, and a validation set.

[0017] Furthermore, a preferred implementation is provided, wherein the training set, test set, and validation set are allocated in a ratio of 7:2:1.

[0018] Furthermore, a preferred implementation is provided, in which the determination in S4 whether the prediction model meets the prediction accuracy requirement is achieved through the root mean square error.

[0019] Furthermore, a preferred implementation manner is provided, in which the final superheater temperature prediction model and the final superheater pressure prediction model in S5 are constructed using an LSTM algorithm.

[0020] Further, a preferred embodiment is provided, in which the method for calculating creep damage by calling the predicted temperature data, pressure data, material data and specification data of the final superheater in S6 is:

[0021]

[0022]

[0023] Wherein, σ is the hoop stress of the final superheater tube, P is the operating pressure, in MPa; D is the outer diameter of the tube, in mm; S is the wall thickness of the tube, in mm; t is the service life of the final superheater at the current temperature and pressure, T is the operating temperature, K 1 , K 2 , C is the material parameter; D creep is the cumulative amount of creep damage, t i is the time interval under the i-th temperature and pressure condition of the final superheater.

[0024] Solution three: A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes any one of the methods described in Solution one.

[0025] Solution 4: A computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the steps of the method described in any one of Solution 1.

[0026] The present invention is beneficial in that:

[0027] The intelligent prediction method for creep damage of the final superheater of a power station boiler described in the present invention realizes the prediction of creep damage of the final superheater of a power station boiler. The power plant manager can grasp the damage condition of the final superheater in advance and deal with it in advance, which can reduce the occurrence of accidents and enhance the safety prevention and control capabilities of the power plant.

[0028] The prediction method of the present invention helps to reduce the number of failure shutdowns, lower the operation and maintenance costs of power plants, and improve economic benefits.

[0029] The present invention is also applicable to the field of construction and application method of a creep damage prediction model for a final superheater of a power station boiler based on machine learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of an intelligent prediction method for creep damage of a final superheater of a power station boiler as described in Implementation Method 1.

[0031] Figure 2 This is a schematic diagram of the neural network structure of the LSTM machine learning algorithm used in Implementation Method Seven.

[0032] Figure 3This is a test effect diagram of the final superheater temperature prediction model constructed using the machine learning algorithm LSTM in implementation mode seven.

[0033] Figure 4 This is a test effect diagram of the final superheater pressure prediction model constructed using the machine learning algorithm LSTM in implementation method seven. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solutions and advantages of the implementation methods of the present application clearer, the technical solutions in the implementation methods of the present application will be clearly and completely described below in conjunction with the drawings in the implementation methods of the present application. Obviously, the described implementation methods are only part of the implementation methods of the present application, not all of the implementation methods.

[0035] Embodiment 1: This embodiment provides an intelligent prediction method for creep damage of a final superheater of a power plant boiler, the prediction method comprising the following steps:

[0036] S1. Collect historical operation data, material data, specification data, and historical test data of the final superheater of the power station boiler to establish a basic database;

[0037] S2. Processing abnormal data in the historical operation data described in S1;

[0038] S3, standardize the abnormal data processed in S2 and save the standardized parameters, and create a data set using the training requirements of the prediction model;

[0039] S4, constructing a neural network model according to the selected time series prediction algorithm, and transferring the data set described in S3 into the neural network model, training and optimizing the neural network model according to the set error constraints, and storing the final superheater temperature prediction model, pressure prediction model and standardized parameters that meet the prediction accuracy requirements;

[0040] S5, compile the final superheater temperature prediction model, final superheater pressure prediction model and standardized parameter file trained in S4 into code, and encapsulate and call them to predict the temperature and pressure data of the final superheater;

[0041] S6. Calling the temperature data, pressure data, material data and specification data of the final superheater predicted in S5 to perform creep damage calculation, so as to realize creep damage prediction of the final superheater.

[0042] Implementation method 2. This implementation method is a further limitation of the intelligent prediction method for creep damage of a power plant boiler final superheater described in implementation method 1. The historical operation data in S1 includes the DCS temperature and DCS pressure data of the power plant boiler final superheater; the material data includes the material grade and material performance data; the specification data includes the outer diameter and inner wall thickness of the final superheater tube; and the historical detection data mainly includes the final superheater tube wall thickness detection data.

[0043] Implementation method three: This implementation method is a further limitation of the intelligent prediction method for creep damage of a power plant boiler final superheater described in implementation method one. In S2, the abnormal data in the historical operation data described in S1 are processed, including the steps of missing data processing and erroneous data processing.

[0044] Implementation method 4: This implementation method is a further limitation of the intelligent prediction method for creep damage of a power plant boiler final superheater described in implementation method 1. In S3, a data set is created according to the training requirements of the prediction model. The created data set includes a training set, a test set, and a validation set.

[0045] Implementation method five: This implementation method is a further limitation of the intelligent prediction method for creep damage of a final superheater of a power plant boiler described in implementation method four. The training set, test set, and validation set are allocated in a ratio of 7:2:1.

[0046] Embodiment 6: This embodiment further limits the intelligent prediction method for creep damage of a final superheater of a power plant boiler described in embodiment 1. In S4, whether the prediction model meets the prediction accuracy requirement is determined through the root mean square error.

[0047] Implementation method seven: This implementation method is a further limitation of the intelligent prediction method for creep damage of a power plant boiler final superheater described in implementation method one. The final superheater temperature prediction model and the final superheater pressure prediction model described in S5 are constructed using the LSTM algorithm.

[0048] Embodiment 8: This embodiment further limits the method for intelligently predicting creep damage of a power plant boiler final superheater described in Embodiment 1. The method for calculating creep damage by calling the predicted final superheater temperature data, pressure data, material data and specification data in S6 is:

[0049]

[0050] Wherein, σ is the hoop stress of the final superheater tube, P is the operating pressure, in MPa; D is the outer diameter of the tube, in mm; S is the wall thickness of the tube, in mm; t is the service life of the final superheater at the current temperature and pressure, T is the operating temperature, K 1 , K 2, C is the material parameter; D creep is the cumulative amount of creep damage, t i is the time interval under the i-th temperature and pressure condition of the final superheater.

[0051] Embodiment 9. This embodiment proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the method described in any one of embodiments 1 to 8.

[0052] Embodiment 10: This embodiment proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the method described in any one of Embodiments 1 to 8 are implemented.

[0053] Implementation eleven: This implementation provides an example, which is used to explain the above implementations one to eight. The specific example is as follows:

[0054] See also Figures 1 to 4 This embodiment is described as follows: S1. Data collection: historical operation data, material data, specification data, and historical test data of the final superheater of the power station boiler are collected to establish a basic database. The historical operation data includes the DCS temperature and DCS pressure data of the final superheater of the power station boiler, the material data includes the material grade and material performance data, the specification data includes the outer diameter and inner wall thickness of the tube of the final superheater, and the historical test data mainly includes the tube wall thickness test data of the final superheater.

[0055] S2. Data processing: Process the abnormal data in the historical operation data to improve the data quality and ensure the prediction accuracy of the model. Data processing includes missing data processing and erroneous data processing.

[0056] S3, data set construction: First, standardize the data and save the standardization parameters. Then, create a data set according to the prediction model training requirements, and use the ratio of 7:2:1 for training set, test set, and validation set, which are independent of each other.

[0057] S4, model training and optimization: First, build the neural network model according to the selected time series prediction algorithm, then pass the data into the neural network model, train and optimize the model according to the set error constraints, and finally store the prediction model and standardized parameters that meet the prediction accuracy requirements.

[0058] S5. Model deployment: The trained final superheater temperature prediction model, pressure prediction model and standardized parameter file are compiled into code and packaged and called to realize the temperature and pressure data prediction of the final superheater.

[0059] S6. Damage prediction: Call the predicted temperature data, pressure data, material data and specification data of the final superheater to perform creep damage calculation to achieve creep damage prediction of the final superheater. The creep damage calculation method adopts the creep damage calculation method of pressure-bearing parts in the national standard "Damage Pattern Identification of Pressure Equipment".

[0060] Figure 1 In step 101, the historical temperature and pressure data of the last stage superheater in the past six months are extracted from the DCS control system of the power plant, in units of °C and MPa, respectively, and the time interval is controlled at 1 minute;

[0061] Figure 1 In step 102, there may be data errors in the DCS data of the power plant, such as the appearance of error characters such as bad, Bad, nan, Nan, etc., so these abnormal data need to be processed. In this embodiment, these error data are processed by numerical interpolation;

[0062] Figure 1 In step 103, the data set is standardized and the standardized parameters are saved to ensure that the same standardized parameters can be used when the model is deployed in step 105; when constructing the model training data, the temperature data and pressure data of the final superheater are constructed into a new data set in the manner of taking the first three numbers as input and the last number as output, and then divided into a training set, a test set, and a validation set in a ratio of 7:2:1;

[0063] Figure 1 In step 104 of the present invention, the LSTM algorithm is used to construct a final superheater temperature and pressure prediction model. The neural network structure of the LSTM is set to 3 inputs, 1 output, and 2 hidden layers, wherein each hidden layer has 78 hidden nodes, and its structural framework is as follows: Figure 2 As shown; the training set data is passed into the model for training, and the prediction model file with the smallest error is saved, wherein the prediction model uses the root mean square error as the evaluation index, as shown in formula (1-1), and the model is saved in a manner that the entire model is completely saved; the present invention constructs two prediction models, namely, the final superheater temperature prediction model and the final superheater pressure prediction model;

[0064]

[0065] Figure 1In step 105, the standardized parameter file saved in step 103 and the prediction model file saved in step 104 are first loaded, and then the temperature and pressure data of the final superheater monitored in real time are respectively standardized, and then the corresponding prediction models are input to perform temperature and pressure prediction, and the above step 105 is encapsulated into a function module for easy calling, the module input is the temperature and pressure data monitored in real time, and the output is the predicted temperature and pressure data;

[0066] Figure 1 In step 106, the creep damage calculation method of pressure-bearing components in the national standard "Damage Pattern Identification of Pressure Equipment" is first written into a program, and then the historical monitoring data and prediction data of the final superheater are passed into the creep damage calculation program to calculate the creep damage amount in each period of time, and then calculate the cumulative amount of damage, and finally realize the creep damage prediction of the final superheater. The creep calculation formula is as follows:

[0067]

[0068] Wherein, σ is the hoop stress of the final superheater tube, P is the operating pressure, unit: MPa; D is the outer diameter of the tube, unit: mm; S is the tube wall thickness, unit: mm; t is the service life of the final superheater at the current temperature and pressure, T is the operating temperature, K 1 , K 2 , C is the material parameter; D creep is the cumulative amount of creep damage, t i is the time interval under the i-th temperature and pressure condition of the final superheater.

[0069] Figure 1Any process or method description in the flowchart described in or otherwise described herein can be understood as representing a module, fragment or portion of a code including one or more executable instructions for implementing the steps of a custom logic function or process, and the scope of the preferred embodiment of the present invention includes other implementations, in which the functions may be performed in a substantially simultaneous manner or in a reverse order according to the functions involved, which should be understood by a person skilled in the art of the present invention. The logic and / or steps represented in the flowchart or otherwise described herein illustrate the possible implementation architecture, functions and operations of the apparatus and methods according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logic function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It is also noted that each block in the block diagram or flowchart, and combinations of blocks in the block diagram or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions. For example, what may be considered as an ordered list of executable instructions for implementing a logical function may be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other system that can fetch instructions from and execute instructions in an instruction execution system, apparatus, or device).

[0070] Those skilled in the art will appreciate that the above are only preferred embodiments of the present invention, and the various embodiments of the present disclosure and / or the features described in the claims may be combined or combined in various ways, even if such combinations or combinations are not explicitly described in the present disclosure. It is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art may still modify the technical solutions described in the aforementioned embodiments, or perform equivalent substitutions on some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

[0071] Although preferred embodiments of the present invention have been described, additional changes and modifications may be made to these embodiments by those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent prediction method for creep damage of a final superheater of a power plant boiler, characterized in that: The prediction method comprises the following steps: S1. Collect historical operation data, material data, specification data, and historical test data of the final superheater of the power station boiler to establish a basic database; S2. Processing abnormal data in the historical operation data described in S1; S3, standardize the abnormal data processed in S2 and save the standardized parameters, and create a data set using the training requirements of the prediction model; S4, constructing a neural network model according to the selected time series prediction algorithm, and transferring the data set described in S3 into the neural network model, training and optimizing the neural network model according to the set error constraints, and storing the final superheater temperature prediction model, pressure prediction model and standardized parameters that meet the prediction accuracy requirements; S5, compile the final superheater temperature prediction model, final superheater pressure prediction model and standardized parameter file trained in S4 into code, and encapsulate and call them to predict the temperature and pressure data of the final superheater; S6. Calling the temperature data, pressure data, material data and specification data of the final superheater predicted in S5 to perform creep damage calculation, so as to realize creep damage prediction of the final superheater.

2. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 1 is characterized in that: The historical operation data in S1 include the DCS temperature and DCS pressure data of the final superheater of the power plant boiler; the material data include the material brand and material performance data; the specification data include the outer diameter and inner wall thickness of the final superheater tube; the historical inspection data mainly include the final superheater tube wall thickness inspection data.

3. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 1 is characterized in that: The processing of abnormal data in the historical operation data described in S1 in S2 includes the steps of processing missing data and processing erroneous data.

4. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 1 is characterized in that: In S3, a data set is created based on the prediction model training requirements. The created data set includes a training set, a test set, and a validation set.

5. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 4 is characterized in that: The training set, test set, and validation set are allocated in a ratio of 7:2:

1.

6. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 1 is characterized in that: In S4, whether the prediction model meets the prediction accuracy requirement is determined by the root mean square error.

7. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 1 is characterized in that: The final superheater temperature prediction model and the final superheater pressure prediction model described in S5 are constructed using the LSTM algorithm.

8. The intelligent prediction method for creep damage of the final superheater of a power plant boiler according to claim 1 is characterized in that: The creep damage calculation method for calling the predicted temperature data, pressure data, material data and specification data of the final superheater in S6 is: Wherein, σ is the hoop stress of the final superheater tube, P is the operating pressure, in MPa; D is the outer diameter of the tube, in mm; S is the wall thickness of the tube, in mm; t is the service life of the final superheater at the current temperature and pressure, T is the operating temperature, K1, K2, C are material parameters; D creep is the cumulative amount of creep damage, t i is the time interval under the i-th temperature and pressure condition of the final superheater.

9. A computer device comprising a memory and a processor, characterized in that A computer program is stored in the memory. When the processor runs the computer program stored in the memory, the processor executes the method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images