A high-temperature tower device temperature field reconstruction method

Abstract

The application relates to a high-temperature tower-type equipment temperature field reconstruction method, and relates to the field of temperature field reconstruction. The reconstruction method comprises the following steps: S1, arranging a temperature monitoring device on the tower-type equipment, and determining the outer surface temperature of the tower-type equipment based on the temperature monitoring device; S2, determining the base function of each stage of the high-temperature tower-type equipment, that is, a characteristic matrix; S3, determining the current process stage of the tower-type equipment, and calling the base function corresponding to the current process stage from S2; S4, inversely deducing the base coefficient based on the outer surface temperature of the tower-type equipment and the base function corresponding to the current process stage; S5, determining a preliminary reconstructed temperature field based on the base coefficient; and S6, correcting the preliminary reconstructed temperature field to obtain a final reconstructed temperature field. The above method in the application can realize fast and high-precision reconstruction of a temperature field.

Description

Technical Field

[0001] This invention relates to the field of temperature field reconstruction, and in particular to a method for reconstructing the temperature field of a high-temperature tower device. Background Technology

[0002] The overall trend in global petroleum product demand structure is a continuous decline in heavy oil demand, an increase in demand for light liquid fuels for engines such as gasoline, kerosene, and diesel, and an increase in the extraction of heavy and extra-heavy oils. Therefore, in the 21st century, the lightening of heavy oils and residues has become a crucial issue for the global refining industry. Coking is an important thermal processing method for residues. It uses residues as feedstock and involves deep thermal cracking reactions at high temperatures (500–505°C). Delayed coking is a process where residues are rapidly cracked at high temperatures within the furnace tubes, "delaying" the coking reaction to the coke tower, hence its name. Currently, the main technologies for increasing the depth of heavy oil and the yield of light oil products remain coking, residue catalytic cracking, and residue hydrotreating. Among these, coking has become a mature method for heavy oil thermal processing, offering advantages in terms of investment, operating costs, technical reliability, feedstock applicability, and conversion. Domestic and international refining companies regard coking technology as an important means of heavy oil processing. The coke tower is a key piece of equipment in delayed coking units. Once it fails, it will cause serious economic losses and even casualties. Therefore, research on coke towers is receiving increasing attention.

[0003] The coke tower is a crucial piece of equipment in a delayed coking unit. Its main characteristics include its height, high length-to-slenderness ratio, and heavy load. The process involves regular intermittent continuous operation with complex load conditions. Prolonged exposure to these complex alternating and external loads ultimately leads to the failure of the coke tower. Common failure modes of coke towers include crack propagation and initiation at the bottom skirt, tower bulging deformation, tower bending and tilting, tower swaying and jumping, weld cracking at the connection points, and material degradation. Among these, weld cracking at the connection between the tower body and the skirt, and tower bulging deformation are particularly serious problems. A 1996 survey of 54 coke towers by the American Petroleum Institute (API) found that: section expansion accounted for 61%, tower body circumferential weld cracking accounted for 97% (mainly occurring on the 3rd, 4th, and 5th circumferential welds above the lower end cap), and skirt cracking accounted for 78%.

[0004] Petrochemical high-temperature tower equipment involves complex processes, typically undergoing alternating heating and cooling during operation. Drastic temperature changes subject the equipment to significant thermal stress, in addition to loads from the tower's own weight, operating internal pressure, wind loads, and processes such as coke removal. Numerical simulation analysis of the coke tower reveals that thermal stress accounts for the largest proportion of the combined stresses, meaning that the thermal stress generated by the temperature gradient has a much greater impact on the safety of the coke tower than the stress generated by the weight of the medium and its own weight. Therefore, the primary factor determining the process cycle of petrochemical high-temperature tower equipment is the thermal stress generated during temperature changes. Thermal stress can lead to tower bulging, tilting, and weld cracking, particularly around the skirt welds and coke-blocking valves, ultimately causing equipment failure. These failure modes eventually result in inadequate sealing, leaks, and even fires. To ensure the safe operation of petrochemical high-temperature tower equipment and extend its service life, the rate of temperature change is usually limited, but this also means extending the production cycle, creating a trade-off between equipment lifespan and production efficiency. Resolving this trade-off requires accurate and real-time monitoring of the temperature field at each process stage.

[0005] Currently, temperature monitoring in petrochemical high-temperature tower equipment mainly relies on a limited number of temperature measuring points arranged between the outer wall and the insulation layer. Whether it's fatigue analysis or stability analysis under cyclic loading, it's essential to start with the temperature field, as temperature difference is one of the conditions for thermal stress generation in coke towers, inextricably linked to daily life and industrial applications. However, the equipment's temperature field is difficult to measure directly. The internal measurement environment of high-temperature tower equipment is harsh, potentially involving phase transition processes such as vaporization and condensation, making it impossible to arrange temperature or heat flow measuring devices inside the tower or on its inner wall. Isolated, limited temperature measuring points cannot directly represent the temperature field, only characterizing local temperatures. The finite element simulation method used in conventional reconstruction processes is computationally time-consuming and cannot meet real-time monitoring requirements. Furthermore, the forward problems of traditional reconstruction methods are typically calculated using finite element or finite volume methods, which are computationally expensive and slow, exhibiting a certain lag in addressing practical engineering problems. Summary of the Invention

[0006] The purpose of this invention is to provide a method for reconstructing the temperature field of high-temperature tower equipment, which improves computational efficiency, enables real-time monitoring of the equipment's temperature field, and allows for structural health detection and diagnostic assessment of the equipment, thereby ensuring the safety of workers and the property of the enterprise in a timely manner.

[0007] To achieve the above objectives, the present invention provides the following solution:

[0008] In a first aspect, the present invention provides a method for reconstructing the temperature field of a high-temperature tower device, the method comprising:

[0009] S1: Install temperature monitoring equipment on the tower equipment, and determine the outer surface temperature of the tower equipment based on the temperature monitoring equipment;

[0010] S2: Determine the basis functions, i.e., the characteristic matrix, for each stage of the high-temperature tower equipment;

[0011] S3: Determine the current process stage of the tower equipment, and call the basis function corresponding to the current process stage from S2;

[0012] S4: Based on the outer surface temperature of the tower equipment and the basis function corresponding to the current process stage, the basis coefficients are calculated.

[0013] S5: Determine the preliminary reconstructed temperature field based on the basic coefficients;

[0014] S6: Correct the preliminary reconstructed temperature field to obtain the final reconstructed temperature field.

[0015] Optionally, installing temperature monitoring equipment on the tower equipment and determining the outer surface temperature of the tower equipment based on the temperature monitoring equipment specifically includes the following steps:

[0016] Determine the structure and damage mode of the tower equipment;

[0017] A first temperature sensor is arranged at the weld joint between the head and the shell of the tower equipment.

[0018] The second, third, fourth, and fifth temperature sensors are sequentially arranged at the head portion below the weld joint, and at the weld positions of the second, third, and fourth layers of the cylinder.

[0019] Remove the insulation layer from all temperature sensor installation locations and sand off the surface paint, oxide layer, and dirt.

[0020] Wait for the tower equipment process to enter the hydraulic decoking stage, and for the tower body temperature to drop to the preset temperature.

[0021] Mix the high-temperature adhesive thoroughly according to the required ratio;

[0022] Then apply the thoroughly mixed high-temperature adhesive to all temperature sensors;

[0023] Place all temperature sensors coated with high-temperature adhesive in the installation position and secure them with high-temperature resistant Teflon tape, then wait for the high-temperature adhesive to cure.

[0024] Connect all temperature sensors to the transmission cables and initially fix them to the tower body with Teflon tape. After the tower insulation layer is restored, fix the sensor cables to the on-site support and the original cables.

[0025] Connect all temperature sensors to the temperature data acquisition system via transmission cables;

[0026] The output of the temperature data acquisition system is connected to the temperature data processing system via RS232 or Ethernet.

[0027] The temperature data processing system connects to the cloud platform via a 4G or 5G wireless communication module.

[0028] The cloud platform assesses and displays the monitoring results and the safety status of the high-temperature tower equipment.

[0029] Optionally, the preset temperature is 50°C.

[0030] Optionally, the polishing area for removing paint, oxide layer and dirt from the polished surface is 30*30cm.

[0031] Optionally, determining the basis function for each stage of the high-temperature tower equipment specifically includes the following steps:

[0032] Construct a mathematical model for the positive problem of heat conduction;

[0033] The grid independence and heat transfer model were verified using the mathematical model of the heat conduction positive problem.

[0034] The basis functions for each process stage of the tower equipment are determined based on the validated mathematical model of the thermal conductivity positive problem.

[0035] Optionally, determining the current process stage of the tower equipment and calling the basis function corresponding to the current process stage from S2 specifically includes the following steps:

[0036] A piecewise linear representation algorithm is used to extract important local features of the outer surface temperature data of the tower equipment.

[0037] The important local features are classified using a decision tree classifier;

[0038] The basis function corresponding to the current process stage is determined based on the important local features after classification.

[0039] Optionally, the expression for the initial reconstructed temperature field is as follows:

[0040]

[0041] in, This indicates the initial reconstruction of the temperature field. Represents a k-dimensional characteristic matrix. T represents the transpose of the characteristic matrix. mea This represents the predicted temperature of l sensors.

[0042] Optionally, the expression for the final reconstructed temperature field is as follows:

[0043]

[0044] in, This indicates the initial reconstruction of the temperature field. Represents a k-dimensional characteristic matrix. T represents the transpose of the characteristic matrix. mea Let λ represent the predicted temperature of l sensors, λ represent the regularization parameter, and I represent the identity matrix.

[0045] In a second aspect, the present invention provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described high-temperature tower equipment temperature field reconstruction method.

[0046] Thirdly, the present invention provides a computer-readable storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements the temperature field reconstruction method for high-temperature tower equipment as described above.

[0047] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0048] This invention involves deploying temperature monitoring equipment on the tower-type equipment and determining the outer surface temperature of the tower-type equipment based on the temperature monitoring equipment; determining the basis functions, i.e., characteristic matrices, for each stage of the high-temperature tower-type equipment; determining the current process stage of the tower-type equipment and calling the basis function corresponding to the current process stage from step S2; inversely calculating the basis coefficients based on the outer surface temperature of the tower-type equipment and the basis function corresponding to the current process stage; determining the preliminary reconstructed temperature field based on the basis coefficients; and correcting the preliminary reconstructed temperature field to obtain the final reconstructed temperature field. In summary, this invention, based on the automatic identification of the process stages of a high-temperature coke tower and the calling and inverse calculation of the basis coefficients obtained from the corresponding basis functions, forms a rapid reconstruction method for the temperature field of a high-temperature tower-type equipment based on limited temperature information, greatly improving the reconstruction accuracy.

[0049] In addition, the arrangement of the five temperature sensors on the outer surface of the coke tower of the high-temperature tower equipment ensures that the real-time temperature data collected by the monitoring device is accurate, complete, and reliable. The temperature monitoring device on the outer surface of the high-temperature tower equipment has good stability and can continuously collect the temperature status of the outer surface of the high-temperature tower equipment in real time, which can further provide data support for the safety status assessment of the high-temperature tower equipment. Attached Figure Description

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart of the temperature field reconstruction method for high-temperature tower equipment according to the present invention;

[0052] Figure 2 This is a block diagram of the temperature data acquisition system of the present invention;

[0053] Figure 3 This is a schematic diagram of the on-site arrangement of the temperature sensor of the present invention; Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] The purpose of this invention is to provide a method for reconstructing the temperature field of high-temperature tower equipment, which improves computational efficiency, enables real-time monitoring of the equipment's temperature field, and allows for structural health detection and diagnostic assessment of the equipment, thereby ensuring the safety of workers and the property of the enterprise in a timely manner.

[0056] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0057] Figure 1 This is a flowchart of the temperature field reconstruction method for high-temperature tower equipment according to the present invention, as shown below. Figure 1 As shown, the method in this invention includes:

[0058] S1: Install temperature monitoring equipment on the tower equipment, and determine the outer surface temperature of the tower equipment based on the temperature monitoring equipment.

[0059] Specifically, this invention provides a real-time online monitoring device for the external surface temperature of petrochemical high-temperature tower equipment, to achieve real-time temperature monitoring, such as... Figure 2As shown, the device includes: a temperature data acquisition system for collecting monitoring data; a thermocouple temperature sensor installed on the outer surface of the coke tower to be measured; the output of the thermocouple temperature sensor is connected to the temperature data acquisition system for real-time data acquisition via a transmission cable; and the output of the temperature data acquisition system is connected to a temperature data processing system. The temperature data processing system is used for post-processing tasks such as data recording, curve analysis, monitoring and early warning, data review, and report generation. Furthermore, the output of the temperature data acquisition system is connected to the temperature data processing system via RS232 or Ethernet. Furthermore, the real-time online monitoring device for the outer surface temperature of the high-temperature tower equipment also includes a 4G or 5G wireless communication module and a cloud platform; the temperature data processing system is connected to the cloud platform via the 4G or 5G wireless communication module.

[0060] Step S1 above specifically includes the following steps:

[0061] 1) Based on the structure and damage mode of the coke tower, a temperature sensor is installed at the weld joint between the coke tower head and the shell. Another temperature sensor is installed below the weld joint on the head section, and at the weld joints of the second, third, and fourth layers of the shell, for a total of five temperature sensors. This ensures that the real-time temperature data collected by the monitoring device is accurate, complete, and reliable. Figure 3 As shown.

[0062] 2) Remove the insulation layer from the location where the temperature sensor will be installed and sand the surface to remove paint, oxide layer and dirt. The sanding area is about 30*30cm.

[0063] 3) Wait for the coke tower process to enter the hydraulic decoking stage, and for the tower temperature to drop to around 50℃.

[0064] 4) Thoroughly mix the high-temperature adhesive according to the required ratio. Use the high-temperature adhesive as soon as possible after mixing.

[0065] 5) Apply high-temperature adhesive to the temperature sensor, place the sensor in the installation position, and fix the sensor with high-temperature resistant Teflon tape. Wait for the high-temperature adhesive to cure.

[0066] 6) Connect the temperature sensor to the transmission cable and initially fix it to the tower body with Teflon tape. After the tower insulation layer is restored, fix the sensor cable to the on-site support and the original cable.

[0067] 7) Then connect the temperature sensor to the temperature data acquisition system via a transmission cable.

[0068] 8) The output of the temperature data acquisition system is connected to the temperature data processing system via RS232 or Ethernet.

[0069] 9) The temperature data processing system connects to the cloud platform via a 4G or 5G wireless communication module.

[0070] 10) The cloud platform assesses and displays the monitoring results and the safety status of the high-temperature tower equipment.

[0071] S2: Determine the basis functions, i.e., the characteristic matrix, for each stage of the high-temperature tower equipment.

[0072] S2-1: Starting from the actual production platform, after making reasonable assumptions, a physical model of petrochemical high-temperature tower equipment is obtained, and a mathematical model of the heat conduction positive problem is constructed based on the heat conduction differential equation and thermal boundary conditions.

[0073] S2-2: Verify the mesh independence and heat transfer model of the mathematical model for the heat conduction problem in S2-1.

[0074] S2-3: Based on the validated model, the basis functions, i.e., the characteristic matrix, of each process stage of the petrochemical high-temperature tower equipment were obtained through a large number of numerical experiments.

[0075] S3: Determine the current process stage of the tower equipment, and call the basis function corresponding to the current process stage from S2.

[0076] Specifically, the following steps are included:

[0077] S3-1 uses a piecewise linear representation (PLR) algorithm to divide the time series data (i.e., the temperature data measured in step S1) based on feature points, and retains important local features of the time series data.

[0078] In this information age, with the rapid development of computer technology, the ability to generate, collect, and store information has been continuously enhanced, resulting in an increasing number of data types: images, text, audio, etc. Time series data is not only related in terms of time sequence; any data that has a logical sequential relationship and is immutable is time series data. Therefore, time series data is widely present in various data types. In order to reduce the amount of data while preserving the effective characteristics of time series data, data compression was proposed by Keogh, Chu, and Hart.

[0079] The main idea of ​​the PLR ​​algorithm is to extract important feature points from time-series data, and use the line segments connecting these feature points to represent the original sequence. This method uses the found key points to more intuitively demonstrate the characteristics of the original sequence data.

[0080] S3-2: Classify the locally important features in S3-1.

[0081] Decision tree classifiers classify time-series data based on feature subsequences. The k-neighbor classifier categorizes the measured sequence and key feature points into process segments. The k-nearest neighbor classifier first calculates the distance between time-series data points, using this distance as a feature, and selects the k time-series data points with the smallest distance to the data to be classified. The category of the data to be classified is determined by the categories of these k data points. When using the k-nearest neighbor classifier to classify time-series data, the value of the parameter k needs to be adjusted. If k is too large, it can easily lead to classification errors; if k is too small, the classification result is easily influenced by the data. Cross-validation is typically used to fine-tune the value of k. In time-series data classification tasks, sometimes k=1, in which case the k-nearest neighbor algorithm is also called the nearest neighbor algorithm. To avoid the influence of the k value, this invention sets k=1, i.e., the nearest neighbor classifier. By reading the temperature range data of the temperature measurement points, the current process segment of the equipment is confirmed, and the feature matrix corresponding to the current process segment is called. That is, basis functions.

[0082] S4: Based on the outer surface temperature of the tower equipment and the basis function corresponding to the current process stage, the basis coefficients are calculated.

[0083] In traditional numerical simulation research, whether using the finite difference method (FDM), finite element method (FEM), or finite volume method (FVM), the numerical solution methods often involve a high degree of freedom, which places high demands on the computing power. Because of the significant time consumption caused by the numerous iterative calculations in the finite element method, the use of the reduced-order model (POD) for solving the temperature field is employed. POD, as an effective dimensionality reduction method, can significantly reduce the amount of data required to describe the physical process, thereby shortening the simulation time.

[0084] The POD method, when faced with a physical problem, obtains a series of basis functions that best express a certain physical property. By comparing and selecting a few of these basis functions that contain the highest-order basis functions, the property information of the physical problem can be covered. By linearly combining these basis functions, a low-dimensional description of a high-dimensional physical field can be achieved, providing a basis for the establishment of low-order POD models.

[0085] Let the numerical solution of the POD temperature field be T′, which is a linear combination of the POD basis and M numbers, expressed as:

[0086]

[0087] The terms in the formula are represented as follows: For the i-th POD base, d iLet be the coefficients of the i-th POD basis. In the POD reduction model based on the temperature field gradient, the most important thing is to find the POD basis with the highest information content. Most of the information in the POD basis is reflected in the entire temperature field of the model in a projective manner.

[0088] Sample data were obtained using the finite element method or field testing. The formula for generating the correlation matrix A is as follows:

[0089]

[0090] After generating the correlation matrix A, solve for the correlation matrix A. N×N Given the eigenvalues ​​and eigenvectors Av = πv, the POD basis can be expressed as:

[0091]

[0092] in, -The i-th element of the j-th feature vector

[0093] Gradient of POD base gradient of POD basis This better reflects the orthogonality of the POD basis. The generated POD basis functions satisfy the following relationship:

[0094]

[0095] π j - The j-th eigenvalue of the correlation matrix, i.e., the expression for the approximate solution error of the POD method. Based on the mathematical derivation, we select M (1≤M≤N) POD bases and use the dimensionality reduction method to calculate the known temperature field, obtaining the gradient error of the POD solution as follows:

[0096]

[0097] It can be seen that after selecting a certain number of M (1≤M≤N) POD bases, the temperature field gradient truncation error obtained by the POD dimensionality reduction method is the sum of the eigenvalues ​​corresponding to the remaining POD bases. The significance of POD bases corresponding to individual eigenvalues ​​lies in the fact that the POD base reduction method improves the sharing accuracy in the process. Therefore, to reduce the truncation error of the dimensionality reduction method, it is necessary to select POD basis functions with larger eigenvalues, and arranging all eigenvalues ​​in descending order of size is meaningful.

[0098] In principle, the more basis functions a POD (Potentially Optical Distributed) basis function set, the better, as it leads to more accurate solutions and more refined information. However, in practical applications, excessive POD basis information can cause ill-posed problems in the algebraic equations. In this invention, the basis functions are arranged in descending order of eigenvalues ​​(information content). If the preceding m basis functions contain enough information to accurately describe the overall distribution and details of the temperature field, the smaller basis functions are omitted.

[0099] First, the temperature field of the equipment is represented by the vector T∈R. N Let N represent the number of hot spots in the temperature field, and let it be expressed as the product of the characteristic matrix and the coefficient matrix:

[0100]

[0101] in, It is an N-dimensional feature matrix, where α is the corresponding... The coefficient matrix, i.e., the basic coefficients.

[0102] Obtain the corresponding feature matrix Then, the temperature field is approximated using K-dimensional features to obtain a reduced-order representation of the temperature field. We can obtain it using its first K columns. Feature matrix The order of the column vectors indicates the importance of the eigenvectors. This approximation is equivalent to projecting the original N-dimensional temperature field onto a K-dimensional subspace, which is composed of K-order eigenma matrices. The columns can be used to represent the temperature at the measuring point as...

[0103]

[0104] Among them, T mea This represents the predicted temperature of l sensors. It is based on the sensor position. The row vector at the corresponding position is obtained. Then, solving the following formula will give α. k :

[0105]

[0106] S5: Determine the preliminary reconstructed temperature field based on the basic coefficients.

[0107] α will be obtained k Substituting the product formula of the characteristic matrix and the coefficient matrix, we can obtain the preliminary reconstructed temperature field as follows:

[0108]

[0109] in, This indicates the initial reconstruction of the temperature field. Represents a k-dimensional characteristic matrix. T represents the transpose of the characteristic matrix. mea This represents the predicted temperature of l sensors.

[0110] S6: Correct the preliminary reconstructed temperature field to obtain the final reconstructed temperature field.

[0111] Since the inverse problem is ill-posed, a Tikhonov regularization parameter is introduced to mitigate the impact of measurement errors.

[0112] The Tikhonov regularization method, named after Andrey Tikhonov, is most commonly used for regularization of ill-posed problems. In statistics, this method is called "ridge regression," and in machine learning, it is known as "weight decay." The essence of Tikhonov regularization is to regularize the covariance matrix (Ai) of a non-full-rank matrix A. T Add a small perturbation λ to each diagonal element of A) to make the singular covariance matrix (A) T A) Inverse transformation into a nonsingular moment (A) T The inverse of A) + λI is used, which greatly improves the numerical stability of solving non-full-rank matrices Ax = b, thus reducing the size of the cond condition number. The added term imposes a penalty, and the resulting solution is better than that obtained by optimizing only ∑A T A is more practical.

[0113] In the inverse problem calculation process, measurement information is required as a boundary value condition. However, errors are unavoidable in the data provided by the measurement process. Amplified by the inverse operation of the compactness operator, measurement errors can make the inversion results unstable. The Tikhonov regularization method solves the existence and stability problems of the inversion process by stabilizing the functional and using a regularization coefficient that matches the error level. When using Tikhonov regularization to handle ill-conditioned equations, choosing the optimal regularization parameter becomes crucial to solving the problem. On the one hand, from the perspective of approximation, the regularization parameter should be as small as possible; on the other hand, from the perspective of numerical stability, the regularization parameter should be as large as possible. The usual approach is to use iterative solutions; however, the results are often sensitive to the initial values ​​of the parameters.

[0114] Regularization parameters are crucial in the inversion calculation process. They determine the approximation of the solution to the original problem constructed using regularization, its ability to resist the effects of measurement and rounding errors, and its capacity to extract the true solution. There are generally two types of strategies: prior and posterior. Many prior strategies have theoretical analytical value, but in practice, it is often difficult to verify the conditions under which they are applied. Therefore, posterior strategies are more prevalent and practical methods for determining regularization parameters. In practice, the error level δ in the original data b is often obtainable, but difficult to obtain in others. Therefore, posterior strategies for determining regularization parameters can be categorized into two cases: δ known and δ unknown. In this invention, the regularization parameter is obtained using the L-curve criterion.

[0115] The L-curve obtains suitable parameter values ​​by balancing the amplified error caused by the constraint value and the error between the approximate solution and the exact solution (ultimately, the two types of errors constrain each other, thus seeking the inflection point).

[0116] make The curvature, as a function of parameter λ, is defined as follows:

[0117]

[0118] Here, “′” represents the derivative with respect to λ.

[0119] The L-curve, proposed by Hansen, describes the modulus ||x| corresponding to different parameter canonical solutions within a log-log scale. λ ||Compared with other moduli||b-Ax λ The L-curve is a curve comparing the values ​​of ||. Its principle is to minimize the sum of the inner product of the residual vectors and the norm of the solution. When the given data vector b satisfies the Picard condition, the L-curve will show a sharp turn as the value of the regularization parameter λ changes, at which point the value of λ is close to the optimal value λ. pot This optimal value is the regularization parameter value. The optimal value is λ. pot This corresponds to the inflection point of the L-curve, which is also the point of maximum curvature. The regularization parameter chosen for the L-curve is not optimal, but approximately optimal.

[0120] By utilizing the automatic identification of the process stage and the basic coefficients obtained after calling the corresponding basis functions and back-calculating corrections, the temperature field of the equipment can be rapidly reconstructed. The specific formula is as follows:

[0121]

[0122] in, This indicates the initial reconstruction of the temperature field. Represents a k-dimensional characteristic matrix. T represents the transpose of the characteristic matrix.mea Let λ represent the predicted temperature of l sensors, λ represent the regularization parameter, and I represent the identity matrix.

[0123] The present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described high-temperature tower equipment temperature field reconstruction method.

[0124] The present invention also provides a computer-readable storage medium, characterized in that it stores a computer program, which, when executed by a processor, implements the temperature field reconstruction method for high-temperature tower equipment as described above.

[0125] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0126] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for reconstructing the temperature field of a high-temperature tower device, characterized in that, The reconstruction method includes: S1: Install temperature monitoring equipment on the tower equipment, and determine the outer surface temperature of the tower equipment based on the temperature monitoring equipment; S2: Determine the basis functions, i.e., the characteristic matrix, for each stage of the high-temperature tower equipment; S3: Determine the current process stage of the tower equipment, and call the basis function corresponding to the current process stage from S2; S4: Based on the outer surface temperature of the tower equipment and the basis function corresponding to the current process stage, the basis coefficients are calculated. S5: Determine the preliminary reconstructed temperature field based on the basic coefficients; S6: Correct the preliminary reconstructed temperature field to obtain the final reconstructed temperature field; The process of determining the current technological stage of the tower equipment and calling the basis function corresponding to the current technological stage from step S2 specifically includes the following steps: A piecewise linear representation algorithm is used to extract important local features of the outer surface temperature data of the tower equipment. The important local features are classified using a decision tree classifier; Determine the basis function corresponding to the current process stage based on the important local features after classification; The expression for the preliminary reconstructed temperature field is as follows: ； in, This indicates the initial reconstruction of the temperature field. express 3D feature matrix, This represents the transpose of the characteristic matrix. Indicates prediction Temperature of each sensor; The expression for the final reconstructed temperature field is as follows: ； in, This indicates the final reconstructed temperature field. express 3D feature matrix, This represents the transpose of the characteristic matrix. Indicates prediction The temperature of each sensor, Represents the regular expression parameter. Represents the identity matrix.

2. The method for reconstructing the temperature field of a high-temperature tower device according to claim 1, characterized in that, Installing temperature monitoring equipment on the tower equipment and determining the outer surface temperature of the tower equipment based on the temperature monitoring equipment specifically includes the following steps: Determine the structure and damage mode of the tower equipment; A first temperature sensor is arranged at the weld joint between the head and the shell of the tower equipment. The second, third, fourth, and fifth temperature sensors are sequentially arranged at the head portion below the weld joint, and at the weld positions of the second, third, and fourth layers of the cylinder. Remove the insulation layer from all temperature sensor installation locations and sand off the surface paint, oxide layer, and dirt. Wait for the tower equipment process to enter the hydraulic decoking stage, and for the tower body temperature to drop to the preset temperature. Mix the high-temperature adhesive thoroughly according to the required ratio; Then apply the thoroughly mixed high-temperature adhesive to all temperature sensors; Place all temperature sensors coated with high-temperature adhesive in the installation position and secure them with high-temperature resistant Teflon tape, then wait for the high-temperature adhesive to cure. Connect all temperature sensors to the transmission cables and initially fix them to the tower body with Teflon tape. After the tower insulation layer is restored, fix the sensor cables to the on-site support and the original cables. Connect all temperature sensors to the temperature data acquisition system via transmission cables; The output of the temperature data acquisition system is connected to the temperature data processing system via RS232 or Ethernet. The temperature data processing system connects to the cloud platform via a 4G or 5G wireless communication module. The cloud platform assesses and displays the monitoring results and the safety status of the high-temperature tower equipment.

3. The method for reconstructing the temperature field of a high-temperature tower device according to claim 2, characterized in that, The preset temperature is 50℃.

4. The method for reconstructing the temperature field of a high-temperature tower device according to claim 2, characterized in that, The area to be sanded to remove paint, oxide layer, and dirt is 30. 30cm.

5. The method for reconstructing the temperature field of a high-temperature tower device according to claim 1, characterized in that, Determining the basis function for each stage of the high-temperature tower equipment specifically includes the following steps: Construct a mathematical model for the positive problem of heat conduction; The grid independence and heat transfer model were verified using the mathematical model of the heat conduction positive problem. The basis functions for each process stage of the tower equipment are determined based on the validated mathematical model of the thermal conductivity positive problem.

6. An electronic device, characterized in that, The device includes a memory and a processor, wherein the memory stores a computer program and the processor runs the computer program to enable the electronic device to perform the temperature field reconstruction method for a high-temperature tower device as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the temperature field reconstruction method for high-temperature tower equipment as described in any one of claims 1-5.

Images