A method for predicting dynamic creep life loss of high-temperature heating surface pipeline of a power plant boiler

Abstract

This invention relates to a method for dynamically predicting the creep life loss of high-temperature heating surface pipes in power plant boilers. The method includes: collecting pipe wall temperature data of the boiler heating surface pipes over a certain operating period; calculating the creep life loss data of the pipe heating surface during that operating period; and establishing a radial basis function neural network model using the pipe wall temperature data and the creep life loss data to calculate the predicted creep life of the boiler heating surface pipes. The beneficial effects of this invention are: it can achieve accurate prediction of the creep life of boiler heating surface pipes, which is of great significance for extending the service life of thermal power units and ensuring the long-term safe and stable operation of the units.

Description

Technical Field

[0001] This invention relates to the field of high-temperature mechanical properties research of metallic materials, and more specifically, it relates to a method for dynamically predicting creep life loss of pipes on high-temperature heating surfaces of power plant boilers. Background Technology

[0002] Creep life is one of the important performance characteristics for the safe service of heat-resistant steel. The service temperature of high-temperature heating surfaces in power plant boilers is generally greater than 500℃. Under long-term high-temperature and high-pressure service environments, creep damage will occur. Creep damage continues to accumulate until creep fracture occurs. Currently, the assessment and prediction of creep loss of boiler high-temperature heating surface pipes relies on the Larson-Miller empirical formula. In this formula, creep life loss is a function of operating temperature, working fluid pressure, and time. In typical calculations, the parameters used are the rated temperature and rated pressure from the wall temperature calculation report. In actual operation of thermal power units, the actual operating temperature and working fluid pressure vary with the unit load, and the actual temperature of different pipe groups within the same component also differs. Therefore, the creep life results obtained by the static prediction method of directly substituting design parameters into the Larson-Miller empirical formula are very conservative, sometimes even lower than the design life, and cannot accurately reflect the operating state of high-temperature heating surface pipes. Therefore, it is necessary to propose a method for dynamic prediction of creep life loss of boiler high-temperature heating surface pipes under varying operating conditions. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method for dynamically predicting the creep life loss of pipes on high-temperature heating surfaces in power plant boilers.

[0004] Firstly, a method for dynamically predicting creep life loss of high-temperature heating surface pipes in power plant boilers is provided, including:

[0005] Step 1: Collect the pipe wall temperature data of the boiler heating surface pipes over a period of time by using the pipe wall temperature measuring points inside the furnace.

[0006] Step 2: Using the pipe wall temperature data, calculate the creep life loss data of the pipe heating surface during this operating time.

[0007] Step 3: Using the pipe wall temperature data and the creep life loss data of the pipe heating surface, establish a radial basis function (RBF) neural network model to calculate the prediction result of the creep life of the boiler heating surface pipe.

[0008] Preferably, step 1 includes:

[0009] Step 1.1: Use finite element software to calculate and analyze the temperature distribution of the boiler's heating surface pipes, and determine the installation positions of N different boiler component wall temperature measuring points based on the calculation results;

[0010] Step 1.2: Based on the installation location determined in Step 1.1, use N integrated thermocouples to insert them through holes in the furnace top cover and ceiling, and position the heat collection block to the tube assembly. Then, use double-sided welding to weld the heat collection block to the tube wall so that it can fully transfer heat with the tube wall.

[0011] Step 1.3: Set the data acquisition rate of N measuring points to once every X minutes, run for M hours, and obtain N sets of temperature data for the measuring points.

[0012] Preferably, in step 1.2, the thermocouple is fixed to the unexposed side of the tube during installation, the heat collection block is wrapped with an aluminum silicate needle-punched blanket, and finally protected with abrasion-resistant tiles.

[0013] Preferably, step 2 includes:

[0014] Step 2.1: Complete the high-temperature creep endurance test X group and the uniaxial tensile endurance long-term test Y group for the metal materials of the boiler heating surface pipes. Perform polynomial regression and least squares method on the test results to obtain the material constant C and polynomial coefficients c1, c2, c3, and c4.

[0015] Step 2.2: Substitute the temperature data obtained from the measuring points in Step 1.3 and the material constant C and polynomial coefficients c1, c2, c3, and c4 obtained in Step 2.1 into equation (1):

[0016]

[0017] In the formula, T is the absolute temperature in K; C is the material constant; t is the creep rupture time in h; p(σ) is the thermal strength parameter; σ is the load stress in MPa; c1, c2, c3, and c4 are constants.

[0018] The load stress σ is calculated from the working fluid pressure at the measuring point and the pipe diameter and wall thickness.

[0019]

[0020] In the formula, p is the working fluid pressure in MPa; d is the inner radius of the pipe in mm; and S is the pipe wall thickness in mm.

[0021] Step 2.3: Calculate the creep life loss of the boiler heating surface pipes within the operating time M hours;

[0022] In the formula, Ti represents temperature and Pi represents pressure; according to formula (1), the creep rupture time is expressed as:

[0023]

[0024] The creep life of the pipeline during the interval between data collections is:

[0025]

[0026] Therefore, the creep life of the boiler heating surface pipes over a certain period of time is:

[0027]

[0028] In the formula, n represents the number of temperature data points.

[0029] Preferably, step 3 includes:

[0030] Step 3.1: Based on the operating temperature, working fluid pressure, and creep life loss during furnace wall temperature acquisition, establish a radial basis function neural network model, using operating temperature and working fluid pressure as inputs and creep life loss as output; select group D of temperature, pressure, and creep life data for data normalization, the normalization formula is expressed as:

[0031]

[0032] In the formula, x0 represents the normalized data, x p For the data set, x max x is the maximum value of the vector. min The minimum value of the vector. The vector average is used; and training and test samples are selected.

[0033] Step 3.2: Select the radial basis function (RBF) neural network creation function and the scattering constant SPREAD; use the training samples selected in Step 3.1 to train the established RBF neural network. After training, input the test samples selected in Step 3.1 into the established RBF neural network; if the relative error of the test result is less than 15%, the establishment of the RBF neural network is completed; otherwise, the RBF neural network is retrained.

[0034] Step 3.3: Input the historical operating temperature and working fluid pressure data of the system into the trained radial basis neural network, and accumulate the obtained historical creep life loss data to obtain the creep life loss of the heated surface pipe.

[0035] In a second aspect, a dynamic prediction device for creep life loss of high-temperature heating surface pipes in a power plant boiler is provided, used to execute the dynamic prediction method for creep life loss of high-temperature heating surface pipes in a power plant boiler as described in the first aspect, including:

[0036] The data acquisition module is used to collect the pipe wall temperature data of the boiler heating surface pipes over a period of time through the pipe wall temperature measuring points inside the furnace.

[0037] The first calculation module is used to calculate the creep life loss data of the pipe heating surface during the operating time using the pipe wall temperature data.

[0038] The second calculation module is used to establish a radial basis function neural network model using the pipe wall temperature data and the creep life loss data of the pipe heating surface, and calculate the prediction result of the creep life of the boiler heating surface pipe.

[0039] The beneficial effects of this invention are as follows: This invention uses finite element method (FEM) calculations to accurately determine the temperature distribution of boiler heating surface pipes and to identify the installation locations of furnace wall temperature measuring points. By collecting furnace wall temperature data and conducting mechanical tests, the creep life loss of the boiler heating surface pipes over a certain operating time is calculated. Based on the calculation results, an RBF neural network model is established, and combined with historical operating data, the creep life loss of the unit's boiler heating surface pipes is predicted. Applying this invention enables accurate prediction of the creep life of boiler heating surface pipes, which is of great significance for extending the service life of thermal power units and ensuring their long-term safe and stable operation. Attached Figure Description

[0040] Figure 1 Diagram showing the installation locations of temperature measuring points on the inner wall of the high-temperature reheater furnace;

[0041] Figure 2 This is a training graph for the RBF neural network. Detailed Implementation

[0042] The present invention will be further described below with reference to embodiments. The description of the embodiments below is only for the purpose of helping to understand the present invention. It should be noted that those skilled in the art can make several modifications to the present invention without departing from the principle of the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

[0043] As one embodiment, the present invention provides a method for dynamically predicting the creep life loss of high-temperature heating surface pipes in power plant boilers, comprising the following steps:

[0044] Step 1: Collect the pipe wall temperature data of the boiler heating surface pipes over a period of time by using the pipe wall temperature measuring points inside the furnace.

[0045] Step 2: Using the pipe wall temperature data, calculate the creep life loss data of the pipe heating surface during this operating time.

[0046] Step 3: Using the pipe wall temperature data and the creep life loss data of the pipe heating surface, establish a radial basis neural network model to calculate the prediction result of the creep life of the boiler heating surface pipe.

[0047] Step 1 includes:

[0048] Step 1.1: Using finite element analysis software, the temperature distribution of the high-temperature reheater tube bundle of a 1000MW unit was calculated and analyzed. Based on the calculation results, 29 installation locations for the wall temperature measuring points of the high-temperature reheater tube bundle of the 1000MW unit were determined; for example... Figure 1 As shown, measuring points 57, 65, and 73 are located in the 20th tube, while the remaining measuring points are located in the 1st tube. Among them, measuring points 78 and 79 are located on the unexposed side, and measuring points 80 and 81 are located on the side of the tube.

[0049] Step 1.2, the installation location determined in step 1.1, such as... Figure 1 As shown, 29 integrated thermocouples are used, inserted through openings in the furnace top hood and ceiling, and the heat collector blocks are positioned on the high-temperature reheater tube bundle. Then, the heat collector blocks are welded to the tube wall using double-sided welding to ensure sufficient heat transfer between the heat collector blocks and the tube wall. In order to avoid corrosion from high-temperature flue gas and wear from fly ash particles and to maintain stable operation of the measuring points, the thermocouples are fixed to the backfire side of the tubes during installation, the heat collector blocks are wrapped with aluminum silicate needle-punched blankets, and finally protected with wear-resistant tiles.

[0050] Step 1.3: Set the data acquisition rate of the 29 measuring points to once every 10 minutes, run for 1464 hours, and obtain 29 sets of temperature data from 8784 measuring points.

[0051] Step 2 includes:

[0052] Step 2.1: Complete 10 sets of high-temperature creep endurance tests and 16 sets of uniaxial tensile endurance long-term tests on S30432 high-temperature reheater material of a 1000MW unit. Perform polynomial regression and least squares method processing on the test results to obtain the material constant C = 21.716 and polynomial coefficients c1, c2, c3, c4.

[0053] Step 2.2: Substitute the temperature data from the 8784 measuring points obtained in Step 1 and the material constant C = 21.716 and polynomial coefficients c1, c2, c3, c4 obtained in Step 2.1 into equation (1).

[0054]

[0055] In the formula, T(K) is the absolute temperature, C is the material constant, t(h) is the creep rupture time, p(σ) is the thermal strength parameter, σ(MPa) is the load stress, and c1, c2, c3, and c4 are constants.

[0056] The load stress σ can be calculated from the working fluid pressure at the measuring point and the pipe diameter and wall thickness.

[0057]

[0058] In the formula, P (MPa) is the working fluid pressure, d (mm) is the inner radius of the pipe, and S (mm) is the pipe wall thickness.

[0059] Step 2.3: Calculate the creep life loss of the boiler heating surface pipes within 1464 hours of operation.

[0060] According to equation (1), the creep rupture time can be expressed as:

[0061] t i =φ(Ti,Pi) (3)

[0062] The creep life of the pipeline during the interval between data collections is:

[0063]

[0064] Therefore, the creep life of the boiler heating surface pipes over a certain period of time is:

[0065]

[0066] Step 3 includes:

[0067] Step 3.1: Based on the obtained operating temperature, working fluid pressure, and creep life loss during the high-temperature reheater wall temperature acquisition, establish an RBF neural network model, using operating temperature and working fluid pressure as inputs and creep life loss as output. Select 1464 sets of operating temperature, pressure, and creep life data for data normalization. The normalization formula can be expressed as:

[0068]

[0069] In the formula, x0 represents the normalized data, x p For the data set, x max x is the maximum value of the vector. min The minimum value of the vector. The average value is the vector mean. Data normalization can effectively reduce errors caused by insufficient convergence and improve the training accuracy of the model. 1300 sets were then used as the training data set, and 164 sets as the test data set.

[0070] Step 3.2: Selection of the radial basis function (RBF) creation function and the scattering constant SPREAD. The RBF creation function is selected, with an error set to 0.0001 and a scattering constant of 0.8362. Figure 2As shown, the RBF neural network is trained and tested using the training samples selected in step 3.1. After training, the test samples selected in step 3.1 are fed into the established RBF neural network. If the relative error of the test result is less than 15%, the establishment of the RBF neural network is complete; otherwise, the RBF neural network is retrained.

[0071] Step 3.3 involves inputting the historical operating temperature and working fluid pressure data from the system into the trained RBF neural network system, accumulating the obtained creep life loss data, and obtaining that the creep life loss of the high-temperature reheater pipe is approximately 31.65%.

[0072] This invention uses finite element method (FEM) calculations to accurately determine the temperature distribution of boiler heating surface pipes and to identify the installation locations of temperature measuring points on the furnace wall. By collecting furnace wall temperature data and conducting mechanical tests, the creep life loss of the boiler heating surface pipes over a certain operating time is calculated. Based on the calculation results, an RBF neural network model is established, and combined with historical operating data, the creep life loss of the boiler heating surface pipes is predicted. Applying this invention enables accurate prediction of the creep life of boiler heating surface pipes, which is of great significance for extending the service life of thermal power units and ensuring their long-term safe and stable operation.

[0073] In this invention, the heating surface structure, pipe material, pipe specifications, operating temperature, working fluid pressure, and furnace wall temperature are selected as variable parameters, and the applicable scope is as follows:

[0074] Heating surface structure: high-temperature reheater, high-temperature superheater, screen-type superheater

[0075] Pipe materials: 9% Cr steel, 12Cr1MoV steel, S30432 steel, HR3C steel

[0076] Pipe wall temperature: 0~650℃;

[0077] Operating temperature: 0~650℃;

[0078] Working fluid pressure: 0~30MPa.

[0079] Furthermore, the dynamic prediction method for creep life loss of boiler high-temperature heating surface pipes involved in this invention is compared with the conventional creep life loss prediction method (Larson-Miller method). The rated temperature of the outlet BRL of a 1000MW ultra-supercritical unit is 603℃. According to the wall temperature calculation, the wall temperature of the high-temperature reheater under rated conditions is 637℃, and the rated pressure is 4.664MPa. The creep life loss results calculated by the Larson-Miller method with the rated wall temperature and rated pressure and the method of this invention are listed in Table 1.

[0080] Table 1

[0081]

[0082] The results show that the creep life of the high-temperature reheater calculated using the Larson-Miller formula with the design parameters is only 20 years. However, using this method to calculate the creep life of the high-temperature reheater under varying operating conditions over a certain period, the creep life is approximately 36.06 years, which is close to the unit's design life of 30 years.

Claims

1. A method for dynamically predicting creep life loss of high-temperature heating surface pipes in power plant boilers, characterized in that, include: Step 1: Collect the pipe wall temperature data of the boiler heating surface pipes over a period of time by using the pipe wall temperature measuring points inside the furnace. Step 1.1: Use finite element software to calculate and analyze the temperature distribution of the boiler's heating surface pipes, and determine the installation positions of N different boiler component wall temperature measuring points based on the calculation results; Step 1.2: Based on the installation location determined in Step 1.1, use N integrated thermocouples, insert them through holes in the furnace top cover and ceiling, and position the heat collector block to the tube assembly. Then, use double-sided welding to weld the heat collector block to the tube wall to ensure sufficient heat transfer between the heat collector block and the tube wall. During installation, fix the thermocouples to the unexposed side of the tube, wrap the heat collector block with aluminum silicate needle-punched blanket, and finally protect it with wear-resistant tiles. Step 1.3: Set the data acquisition rate of N measuring points to once every X minutes, run for M hours, and obtain N sets of temperature data for the measuring points; Step 2: Using the pipe wall temperature data, calculate the creep life loss data of the pipe heating surface during this operating time. Step 2.1: Complete the high-temperature creep endurance test X group and the uniaxial tensile endurance long-term test Y group for the metal materials of the boiler heating surface pipes. Perform polynomial regression and least squares method on the test results to obtain the material constant C and polynomial coefficients c1, c2, c3, c4. Step 2.2: Substitute the temperature data obtained from the measuring points in Step 1.3 and the material constant C and polynomial coefficients c1, c2, c3, and c4 obtained in Step 2.1 into equation (1): In the formula, T is the absolute temperature in K; C is the material constant; t is the creep rupture time in h; σ is the load stress in MPa; and c1, c2, c3, and c4 are constants. The load stress σ is calculated from the working fluid pressure at the measuring point and the pipe diameter and wall thickness. In the formula, p is the working fluid pressure, in MPa; d is the inner radius of the pipe, in mm; and S is the pipe wall thickness, in mm. Step 2.3: Calculate the creep life loss of the boiler heating surface pipes within the operating time M hours; According to equation (1), the creep rupture time can be expressed as: In the formula, Ti is temperature and Pi is pressure; The creep life of the pipeline during the interval between data collections is: Therefore, the creep life of the boiler heating surface pipes over a certain period of time is: In the formula, n is the number of temperature data points; Step 3: Using the pipe wall temperature data and the creep life loss data of the pipe heating surface, establish a radial basis neural network model to calculate the prediction result of the creep life of the boiler heating surface pipe.

2. The method for dynamically predicting creep life loss of high-temperature heating surface pipes in power plant boilers according to claim 1, characterized in that, Step 3 includes: Step 3.1: Based on the operating temperature, working fluid pressure, and creep life loss during furnace wall temperature acquisition, establish a radial basis function neural network model, using operating temperature and working fluid pressure as inputs and creep life loss as output; select group D of temperature, pressure, and creep life data for data normalization, the normalization formula is expressed as: In the formula, x0 represents the normalized data, x p For the data set, x max x is the maximum value of the vector. min The minimum value of the vector. The vector average is used; and training and test samples are selected. Step 3.2: Select the radial basis function (RBF) neural network creation function and the scattering constant SPREAD; use the training samples selected in Step 3.1 to train the established RBF neural network. After training, input the test samples selected in Step 3.1 into the established RBF neural network; if the relative error of the test result is less than 15%, the establishment of the RBF neural network is completed; otherwise, the RBF neural network is retrained. Step 3.3: Input the historical operating temperature and working fluid pressure data of the system into the trained radial basis neural network, and accumulate the obtained historical creep life loss data to obtain the creep life loss of the heated surface pipe.

3. A device for dynamically predicting the creep life loss of high-temperature heating surface pipes in a power plant boiler, characterized in that, The method for dynamically predicting creep life loss of high-temperature heating surface pipes in power plant boilers as described in claim 1 includes: The data acquisition module is used to collect the pipe wall temperature data of the boiler heating surface pipes over a period of time through the pipe wall temperature measuring points inside the furnace. The first calculation module is used to calculate the creep life loss data of the pipe heating surface during the operating time using the pipe wall temperature data. The second calculation module is used to establish a radial basis function neural network model using the pipe wall temperature data and the creep life loss data of the pipe heating surface, and calculate the prediction result of the creep life of the boiler heating surface pipe.

Images