Model training method, evaluation method, electronic device and computer program product

Abstract

The invention provides a model training method, an evaluation method, electronic equipment and a computer program product, and relates to the technical field of anti-seismic safety. The model training method comprises the following steps: screening historical seismic data to obtain target seismic oscillation data meeting low-frequency conditions and long-period characteristics; coupling according to the structural model of the tank body and the liquid storage model to obtain an analysis model; and training the initial prediction model according to the target seismic oscillation data and the analysis model to obtain a target prediction model. The evaluation method comprises the following steps: according to geographic data of a to-be-detected tank body, determining to-be-detected seismic oscillation data and a seismic grade index of the to-be-detected tank body; to-be-detected seismic oscillation data and target structure parameters of a to-be-detected tank are substituted into the target prediction model for prediction, and predicted damage levels of the tank under the multiple seismic level indexes are obtained; and determining an evaluation result of the to-be-detected tank according to the plurality of predicted damage grades.

Description

Technical Field

[0001] The present application relates to the field of earthquake-resistant safety technology, and in particular to a model training method, an evaluation method, an electronic device, and a computer program product. Background Art

[0002] Horizontal oil and gas storage tanks are one of the important equipment for storing oil and gas in the petrochemical industry. They are a complex system consisting of a cylinder, a head, a saddle steel support, auxiliary components, and liquid storage, and are connected to external pipelines through threads, flanges, welding, etc. Horizontal oil and gas storage tanks are widely used in oil and gas stations, petrochemical parks, refineries, etc. due to their advantages of good storage effect, space saving, and convenient maintenance.

[0003] At present, a total of 23 seismic belts are distributed in southwest, northwest, north China and southeast coastal areas of my country. The coexistence of oil and gas resources and seismic belts has led to the inevitable installation of horizontal oil and gas storage tanks in the above areas. There are two types of vibration modes in horizontal oil and gas storage tanks with free liquid surfaces, namely: the coupled vibration of partial liquid storage and tank body, which belongs to short-period vibration, and the sloshing of liquid storage, which belongs to long-period vibration. Long-period earthquake motions mostly occur in areas near active faults, and have the characteristics of long vibration periods and high vibration amplitudes. Once horizontal oil and gas storage tanks built near active faults encounter the effects of long-period earthquake motions, they may cause large sloshing of the free liquid surface, causing the tanks to suffer greater inertia and aggravate damage. Tank damage will lead to liquid leakage and trigger a series of secondary disasters. Therefore, it is of great significance to predict and evaluate the damage and consequences of horizontal oil and gas storage tanks located near active faults under the action of long-period earthquake motions.

[0004] At present, the study of the seismic performance of horizontal oil and gas storage tanks is generally carried out using the following methods: establishing an analysis model of the horizontal oil and gas storage tank based on the design parameters, selecting a certain number of seismic motions to conduct seismic response analysis of the horizontal oil and gas storage tanks, conducting statistical analysis on the seismic response results of the horizontal oil and gas storage tanks, and conducting seismic verification of the horizontal oil and gas storage tanks under different failure modes based on the statistical data, and evaluating their safety under earthquakes of different intensities.

[0005] In the existing seismic performance research methods, the stored liquid in the horizontal oil and gas storage tank is simplified as an additional mass, without taking into account the shaking of the stored liquid. In addition, during the evaluation, the damage of the tank body, saddle damage, wall foundation damage and anchor bolt damage are judged one by one, and the overall condition of the tank body cannot be evaluated. As a result, the accuracy of the seismic performance evaluation of the tank body is low, and the seismic performance of the tank body cannot be effectively evaluated, which cannot meet the current tank installation evaluation needs. Summary of the invention

[0006] In view of this, the purpose of the embodiments of the present application is to provide a model training method, an evaluation method, an electronic device and a computer program product to improve the problem of low accuracy in the seismic evaluation of oil and gas storage tanks in the prior art.

[0007] In order to solve the above problems, in a first aspect, an embodiment of the present application provides a model training method, the method comprising:

[0008] Screen historical earthquake data to obtain target earthquake data that meets low-frequency conditions and long-period characteristics;

[0009] The analysis model is obtained by coupling the tank structure model and the liquid storage model;

[0010] The initial prediction model is trained according to the target seismic data and the analysis model to obtain a target prediction model.

[0011] In the above implementation process, in order to improve the accuracy of the prediction results of the prediction model, the historical earthquake data can be screened and processed to obtain the target seismic data that meets the low-frequency conditions and long-period characteristics as the corresponding input data to improve the effectiveness of the simulated earthquake situation in the prediction model. In addition, considering the sloshing of the liquid inside the tank, the structural model of the tank and the liquid storage model inside the tank can be constructed separately, and the structural model and the liquid storage model are coupled to obtain the corresponding analysis model, so that the initial prediction model is trained according to the target seismic data and the analysis model to obtain the corresponding target prediction model that can evaluate and predict the seismic performance of the tank. It is possible to select low-frequency and long-period seismic data to simulate earthquake conditions, and take into account the impact of the sloshing of the liquid inside the tank on the overall seismic performance of the tank, which effectively improves the accuracy of the target prediction model in evaluating the seismic performance of the tank, and meets the installation evaluation requirements of tanks of various types and locations.

[0012] Optionally, the initial prediction model is constructed in the following manner:

[0013] Determine multiple failure modes of the tank body under earthquake conditions; wherein the failure modes include: at least one of tank body tensile failure, saddle support compressive failure, and anchor bolt tensile and shear failure;

[0014] Based on the material parameters of the tank corresponding to the multiple failure modes, the damage index function is determined in combination with the weight coefficient of each failure mode; wherein the material parameters include: the ultimate strength of the steel of the tank and the seismic allowable stress of the steel at multiple locations; the weight coefficient is determined based on historical tank seismic data;

[0015] Based on a multiple regression algorithm, the initial prediction model is determined in combination with the structural parameters of the tank body and the damage index function.

[0016] In the above implementation process, considering that there are many different types of failure modes when the tank is damaged, therefore, when constructing the initial prediction model, we can first determine the various failure modes of the tank under earthquake conditions, and determine the corresponding material parameters of the tank under various failure modes, and calculate the weight coefficient of each failure mode determined by combining the historical tank seismic data to obtain the corresponding damage index function. In order to differentiate and characterize the results of the prediction and evaluation, the multi-regression algorithm can be used to calculate the corresponding initial prediction model in combination with the structural parameters and damage index function of the tank. According to the different types of failure conditions on the tank, the corresponding initial prediction model with prediction and evaluation functions can be constructed in combination with the multivariate regression algorithm, and the overall damage condition of the tank can be evaluated, which effectively improves the comprehensiveness of the initial prediction model when performing prediction and evaluation.

[0017] Optionally, the screening of historical earthquake data to obtain target earthquake data satisfying low-frequency conditions and long-period characteristics includes:

[0018] Determining the historical earthquake data according to a preset fault distance;

[0019] Performing empirical mode decomposition processing on the historical seismic data to obtain initial components corresponding to a plurality of eigenmode functions;

[0020] Screening the initial components based on a preset low-frequency threshold to obtain a plurality of low-frequency components that are less than or equal to the low-frequency threshold;

[0021] Each of the historical earthquake data is screened based on the proportion value of the low-frequency component and the low-frequency proportion threshold to obtain the target seismic data.

[0022] In the above implementation process, when screening historical earthquake data, considering the validity of the distance range, the corresponding historical earthquake data can be selected according to the preset fault distance, and then the historical earthquake data can be subjected to empirical mode decomposition processing to decompose the historical earthquake data into multiple initial components corresponding to the eigenmode functions. In order to select low-frequency effective data, multiple initial components can be screened according to the preset low-frequency threshold to obtain multiple low-frequency components less than or equal to the low-frequency threshold, and considering the continuous distribution of low-frequency components in the overall historical earthquake data, the proportion of low-frequency components can be compared with the preset low-frequency proportion threshold to screen the historical earthquake data, and the high proportion of low-frequency components can be used as a long-period feature to obtain the target seismic data that meets the low-frequency conditions and long-period features. The ability to perform multiple processes such as decomposition, frequency screening and period analysis on historical earthquake data effectively improves the validity of the obtained target seismic data.

[0023] Optionally, coupling the structural model of the tank and the liquid storage model to obtain an analysis model includes:

[0024] Constructing the structural model according to the structural characteristics of the tank;

[0025] Constructing the liquid storage model according to the fluid characteristics of the liquid stored in the tank;

[0026] The analysis model is obtained by performing coupling processing on the structural model and the liquid storage model.

[0027] In the above implementation process, considering the differences in the shapes, sizes and other structures of different types of oil and gas tanks, a corresponding structural model can be constructed for the structural characteristics of each tank that needs to be predicted and evaluated. In addition, considering that there is some liquid inside the tank such as oil and gas tanks, and the liquid has corresponding shaking conditions under earthquake conditions, in order to simulate the shaking of the liquid, a corresponding liquid storage model can be constructed according to the fluid properties of the liquid stored in the tank. By coupling the structural model and the liquid storage model, an analysis model that can simulate the interaction between the liquid storage and the tank structure can be obtained. The present application does not simply treat the liquid inside the tank as mass, but by constructing a liquid storage model and coupling the liquid storage model with the structural model, the actual situation between the tank structure and the liquid inside the tank under earthquake conditions is effectively simulated, which effectively improves the effectiveness of the analysis model, thereby improving the effectiveness of training the initial detection model.

[0028] Optionally, the output result of the target prediction model includes the damage prediction value of the tank body and the damage level corresponding to the numerical range of each damage prediction value;

[0029] The damage levels are determined based on industry standards.

[0030] In the above implementation process, the output result of the target prediction result model includes the calculated damage prediction value of the tank body, and in order to associate the value with the damage condition to predict the damage condition of the tank body under earthquake conditions, each damage prediction value has a corresponding damage level formulated according to industry standards, so as to determine the subsequent processing method based on the output damage level. The damage level corresponding to the damage prediction value can be output to characterize the evaluation result of the tank body, so that the user can determine whether the tank body meets the seismic requirements.

[0031] In a second aspect, an embodiment of the present application provides an evaluation method, the method comprising:

[0032] Determining the seismic data and earthquake level index of the tank to be tested according to the geographical data of the tank to be tested;

[0033] The seismic data to be tested and the target structural parameters of the tank body to be tested are brought into the target prediction model for prediction, so as to obtain the predicted damage level of the tank body to be tested under the multiple earthquake level indicators; wherein the target prediction model is a model trained by any one of the above-mentioned model training methods;

[0034] An evaluation result of the tank to be tested is determined according to the multiple predicted damage levels.

[0035] In the above implementation process, due to the differences in the distribution areas of the tanks to be evaluated and tested, and different areas have different earthquake conditions, the corresponding local earthquake data and earthquake level index to be tested can be determined according to the geographic data of each tank to be tested, and considering the structural differences of each tank to be tested, the earthquake data to be tested and the target structural parameters of the tank to be tested can be brought into the target prediction model obtained in the model training method for prediction and evaluation, and the predicted damage level of the tank to be tested under multiple earthquake level indicators can be obtained, so as to determine the evaluation result of the seismic performance of the tank to be tested according to the predicted damage level. According to the actual structural and geographical conditions of the tank to be tested, the target prediction model can be used to predict and evaluate the damage of the tank to be tested under different types of earthquake conditions, thereby improving the effectiveness of the evaluation results of the seismic performance of the tank to be tested, meeting the installation evaluation requirements of various tanks, and can also determine whether the tank to be tested meets the corresponding seismic safety requirements before it is put into use according to the evaluation results, thereby reducing the use of tanks that do not meet the seismic safety requirements, and further reducing the damage caused to the tanks put into use in earthquakes.

[0036] Optionally, the earthquake level index includes: frequent earthquakes, fortification earthquakes and rare earthquakes; the predicted damage level includes: a first predicted damage level of the tank under the frequent earthquake condition, a second predicted damage level under the fortification earthquake condition and a third predicted damage level under the rare earthquake condition;

[0037] Determining the evaluation result of the tank to be tested according to the plurality of predicted damage levels includes:

[0038] Based on historical tank seismic data, multiple performance indicators associated with damage levels are determined;

[0039] Determining a first target performance indicator corresponding to the second predicted damage level and a second target performance indicator corresponding to the third predicted damage level;

[0040] If it is determined that the first target performance indicator represents the normal operation of the tank body to be tested, and the second target performance indicator represents the restricted operation of the tank body to be tested, then it is determined that the evaluation result is that the seismic performance of the tank body to be tested meets the seismic requirements.

[0041] In the above implementation process, since the earthquake level index includes earthquakes of various probabilities and severity, the target prediction model can obtain multiple predicted damage levels for different earthquake level indexes. When evaluating, multiple performance indicators associated with the damage level can be determined based on the earthquake data of the historical tank, and multiple predicted damage levels can be associated with the performance indicators to determine multiple target performance indicators of the tank to be tested under different types of earthquake conditions. Considering the severity of the two earthquake level indicators of fortified earthquakes and rare earthquakes, the performance conditions actually represented by the target performance indicators corresponding to the tank to be tested under these two earthquake conditions can be used to determine whether the seismic performance of the tank to be tested meets the seismic requirements and obtain the corresponding evaluation results. It is possible to classify and predict various types of earthquake conditions, and associate the predicted damage level with the performance indicators that characterize whether the tank to be tested can be used, thereby determining the corresponding evaluation results according to the performance of the tank to be tested under different levels of earthquakes, effectively improving the validity and correctness of the evaluation results.

[0042] Optionally, the method further comprises:

[0043] When the evaluation result is that the seismic performance of the tank body to be tested does not meet the seismic requirements, optimization prompt data for the tank body to be tested is generated based on the predicted damage level.

[0044] In the above implementation process, when the evaluation result shows that the seismic performance of the tank to be tested does not meet the seismic requirements, in order to improve the safety and stability of the tank put into use under earthquake conditions, analysis can be performed based on the predicted damage level to generate optimization prompt data for the tank to be tested, so that users can refer to the optimization prompt data and optimize the structure, material, installation location and other related parameters of the tank until the evaluation result shows that the seismic performance of the tank to be tested meets the seismic requirements before it can be officially put into use. The tank to be tested can be iteratively optimized according to the evaluation results to ensure that the seismic performance of the tank to be tested can meet the seismic requirements.

[0045] In a second aspect, an embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores program instructions, and when the processor reads and runs the program instructions, it executes the steps in any implementation of the above-mentioned model training method and / or evaluation method.

[0046] In a third aspect, an embodiment of the present application further provides a computer program product, which includes a computer program / instructions, and when the computer program / instructions are executed by a processor, the steps in any implementation of the above-mentioned model training method and / or evaluation method are implemented.

[0047] In summary, the embodiments of the present application provide a model training method, an evaluation method, an electronic device, and a computer program product, which can use a target prediction model to predict and evaluate the damage of the tank to be tested under different types of earthquake conditions according to the actual structural conditions and geographical conditions of the tank to be tested, thereby improving the effectiveness of the evaluation results of the seismic performance of the tank to be tested, meeting the installation evaluation needs of various tanks, and can also determine whether the tank to be tested meets the corresponding seismic safety requirements before it is put into use based on the evaluation results, thereby reducing the use of tanks that do not meet the seismic safety requirements, and further reducing the damage caused to the tanks in use during earthquakes. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A block diagram of an electronic device provided in an embodiment of the present application;

[0050] Figure 2A flowchart of a model training method provided in an embodiment of the present application;

[0051] Figure 3 A flowchart of another model training method provided in an embodiment of the present application;

[0052] Figure 4 A detailed flow chart of step S200 provided in an embodiment of the present application;

[0053] Figure 5 A detailed flowchart of step S300 provided in an embodiment of the present application;

[0054] Figure 6 A schematic diagram of a flow chart of an evaluation method provided in an embodiment of the present application;

[0055] Figure 7 A detailed flowchart of step S800 provided in an embodiment of the present application.

[0056] Icon: 100 - electronic device; 111 - memory; 112 - storage controller; 113 - processor; 114 - peripheral interface; 115 - input and output unit; 116 - display unit. DETAILED DESCRIPTION

[0057] The technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments of the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the embodiments of the present application.

[0058] At present, the seismic performance research of horizontal oil and gas storage tanks is generally carried out by the following methods: establish an analysis model of horizontal oil and gas storage tanks based on design parameters, select a certain number of earthquake motions to conduct seismic response analysis of horizontal oil and gas storage tanks, conduct statistical analysis on the seismic response results of horizontal oil and gas storage tanks, and conduct seismic verification of horizontal oil and gas storage tanks under different damage modes one by one based on statistical data, and evaluate their safety under earthquakes of different intensities. However, in the existing seismic performance research methods, the stored liquid in the horizontal oil and gas storage tank is usually simplified as the additional mass of the tank body, without considering the shaking of the stored liquid under earthquake conditions. In addition, during the evaluation, the damage of the tank body, saddle damage, wall foundation damage and anchor bolt damage are judged one by one, and the overall situation of the tank body cannot be evaluated, resulting in low accuracy in the seismic performance evaluation of the tank body, and the seismic performance of the tank body cannot be effectively evaluated, which cannot meet the current installation evaluation requirements of the tank body.

[0059] In order to solve the above problems, the embodiments of the present application provide a model training method and an evaluation method, which are applied to electronic devices. The electronic devices can be servers, personal computers (PCs), tablet computers, smart phones, personal digital assistants (PDAs), and other electronic devices with logical computing functions. The corresponding target prediction models can be constructed and trained to evaluate the seismic performance of the tank body, thereby effectively improving the effectiveness and accuracy of the prediction and evaluation of the seismic performance of the tank body.

[0060] Optionally, see Figure 1 , Figure 1 The block diagram of an electronic device provided in an embodiment of the present application. The electronic device 100 may include a memory 111, a storage controller 112, a processor 113, a peripheral interface 114, an input and output unit 115, and a display unit 116. A person skilled in the art may understand that Figure 1 The structure shown is only for illustration and does not limit the structure of the electronic device 100. For example, the electronic device 100 may further include Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0061] The memory 111, storage controller 112, processor 113, peripheral interface 114, input / output unit 115 and display unit 116 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines. The processor 113 is used to execute the executable module stored in the memory.

[0062] The memory 111 may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), etc. The memory 111 is used to store programs, and the processor 113 executes the program after receiving the execution instruction. The method executed by the electronic device 100 defined by the process disclosed in any embodiment of the present application can be applied to the processor 113, or implemented by the processor 113.

[0063] The processor 113 may be an integrated circuit chip with signal processing capability. The processor 113 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. The general-purpose processor may be a microprocessor or any conventional processor.

[0064] The peripheral interface 114 couples various input / output devices to the processor 113 and the memory 111. In some embodiments, the peripheral interface 114, the processor 113 and the memory controller 112 can be implemented in a single chip. In other embodiments, they can be implemented by separate chips.

[0065] The input / output unit 115 is used to provide input data to the user. The input / output unit 115 may be, but is not limited to, a mouse and a keyboard.

[0066] The above-mentioned display unit 116 provides an interactive interface (such as a user operation interface) between the electronic device 100 and the user or is used to display image data for the user's reference. In this embodiment, the display unit can be a liquid crystal display or a touch display. If it is a touch display, it can be a capacitive touch screen or a resistive touch screen that supports single-point and multi-point touch operations. Supporting single-point and multi-point touch operations means that the touch display can sense touch operations generated simultaneously from one or more positions on the touch display, and the sensed touch operations are handed over to the processor for calculation and processing. In an embodiment of the present application, the display unit 116 can display a variety of data such as target ground vibration data and analysis models, and can also display data such as the predicted damage level and evaluation results of the tank to be tested.

[0067] The electronic device in this embodiment can be used to execute each step in each model training method and evaluation method provided in the embodiment of the present application. The implementation process of the model training method and evaluation method is described in detail below through several embodiments.

[0068] See also Figure 2 , Figure 2 A flow chart of a model training method provided in an embodiment of the present application, the method may include steps S200-S400.

[0069] Step S200, screening historical earthquake data to obtain target earthquake data that meets low-frequency conditions and long-period characteristics.

[0070] Among them, historical earthquake data can include relevant record data of multiple earthquake activities in multiple locations, and the corresponding historical earthquake data can be obtained from databases such as the Earthquake Science Data Center. Considering the effectiveness of simulating earthquake activities, historical earthquake data can be screened according to low-frequency conditions and whether they have long-period characteristics, so as to obtain target seismic data that meet low-frequency conditions and long-period characteristics as parameter data for input prediction models, so as to improve the effectiveness of simulating earthquake conditions in prediction models.

[0071] Optionally, the low-frequency condition can be a condition that limits the vibration frequency, for example, the low-frequency condition is considered to be met when the vibration frequency is lower than 2 Hz, and the long-period characteristic can be a characteristic that limits the arm force occupied by continuous low-frequency vibrations, for example, selecting a record in which the energy ratio of at least multiple consecutive low-frequency vibrations is cumulatively greater than 80% of the energy of the entire earthquake record as near-fault seismic motion data with significant long-period characteristics.

[0072] Step S300, coupling the tank structure model and the liquid storage model to obtain an analysis model.

[0073] Among them, taking into account the sloshing of the liquid inside the tank, the structural model of the tank and the liquid storage model inside the tank can be constructed separately, and the structural model and the liquid storage model are coupled to obtain the corresponding analysis model.

[0074] Optionally, multiple types of models can be selected according to actual needs. For example, the structural model can include a FEM (Finite Element Method) model, the liquid storage model can include a SPH (Smooth particle hydrodynamics) model, and the SPH-FEM coupling algorithm can be used to simulate the interaction between the tank wall and the liquid storage interface of the tank body to obtain a corresponding analysis model.

[0075] Step S400 , training the initial prediction model according to the target earthquake data and the analysis model to obtain the target prediction model.

[0076] Among them, the target seismic data can be used as parameter data for simulating earthquake conditions, and the analysis model can be used as a parameter for simulating tank conditions. The initial prediction model is trained according to the target seismic data and the analysis model to obtain the corresponding target prediction model that can evaluate and predict the seismic performance of the tank.

[0077] Optionally, the initial prediction model is a model that determines the damage to the tank under earthquake conditions based on the simulated earthquake conditions and the tank model. The output result of the target prediction model may include a damage prediction value for the tank. The damage prediction value may be a normalized value, such as a value of 0-1, or a value of 0%-100%, etc.

[0078] It should be noted that in order to associate the numerical value with the damage condition to predict the damage condition of the tank body in an earthquake, each numerical range of the damage prediction value has a corresponding damage level formulated according to industry standards, so as to determine the subsequent processing method based on the output damage level. The damage level corresponding to the damage prediction value can be output to characterize the assessment result of the tank body, so that the user can determine whether the tank body meets the earthquake resistance requirements.

[0079] For example, the numerical range of each damage prediction value has a corresponding damage level. For example, taking the damage prediction value D as 0-1 as an example, multiple damage levels can be determined according to industry standards. For example, according to the relevant provisions of "Petrochemical Engineering Earthquake Damage Assessment Standard" (GB50992-2014), "Petrochemical Steel Equipment Seismic Assessment Standard" (GB / T 51273-2018) and "Petrochemical Steel Equipment Seismic Design Standard" (GB / T 50761-2018), 5 different damage levels are set, and the corresponding D can include 5 numerical ranges. Accordingly, the multiple numerical ranges correspond to the five damage levels of basically intact, slightly damaged, moderately damaged, severely damaged, and destroyed. The corresponding relationship between the damage prediction value and the damage level can be referred to the following table:

[0080] Damage level Basically intact Minor damage Moderate damage Severe damage destruction Damage prediction value 0≤D<0.05 0.05≤D<0.15 0.15≤D<0.4 0.4≤D<0.7 0.7≤D≤1.0

[0081] The macroscopic performance of each damage level on the tank can be found in the following table:

[0082]

[0083] exist Figure 2 In the embodiment shown, low-frequency and long-period seismic data can be selected to simulate earthquake conditions, and the impact of the shaking of the liquid inside the tank on the overall seismic performance of the tank is taken into account, thereby effectively improving the accuracy of the target prediction model in evaluating the seismic performance of the tank, and meeting the installation evaluation needs of tanks of various types and locations.

[0084] Optionally, see Figure 3 , Figure 3 A flowchart of another model training method provided in an embodiment of the present application, the method may further include steps S510-S530.

[0085] Step S510, determining multiple failure modes of the tank body under earthquake conditions.

[0086] Among them, considering that there are many different types of failure modes when the tank is damaged, therefore, when constructing the initial prediction model, the various failure modes of the tank under earthquake conditions can be determined first.

[0087] For example, multiple failure modes of the tank can be determined based on industry standards or historical conditions, and the failure modes may include one or more different types of failure conditions, such as failure of the tank under tension, failure of the saddle support under compression, and failure of the anchor bolts under tension and shear. The failure of the tank under tension is usually due to the excessive tensile load on the tank, which causes the material to yield or break. The saddle support is an important component that supports the horizontal tank container, and its failure under compression is usually due to the pressure on the support exceeding its bearing capacity. The failure of the anchor bolts under tension and shear may be due to the loosening or shearing of the anchor bolts under the impact of an earthquake.

[0088] Step S520: determining a damage index function based on the material parameters of the tank corresponding to the various failure modes and in combination with the weight coefficient of each failure mode.

[0089] Among them, the corresponding material parameters of the tank under various failure modes can be determined, and the weight coefficient of each failure mode determined by combining historical tank seismic data can be calculated to obtain the corresponding damage index function.

[0090] Optionally, the material parameters may include: the ultimate strength of the steel of the tank and the allowable seismic stress of the steel at multiple locations. The weight coefficient is determined based on historical tank earthquake data. The damage index function may be:

[0091]

[0092] Where D1 is the damage index of the horizontal oil and gas storage tank, and 0≤D≤1. b is the ultimate strength of the steel material of the tank, [σ] sck is the seismic allowable stress of the steel material of the saddle support, [σ b ] is the allowable tensile stress of the steel material for earthquake resistance of the anchor bolt, [τ b ] is the allowable seismic shear stress of the steel material of the anchor bolt, which is generally taken as [τ b ]=0.8[σ b ]. α1, α2, α3 and α4 are weight coefficients of failure of tank body under tension, failure of saddle support under compression, and failure of anchor bolt under tension and shear, respectively. The weight coefficients can be determined based on historical tank body seismic data. The historical tank body seismic data can be data related to the tank body collected under historical earthquake conditions. For example, the analytic hierarchy process can be used to analyze the historical tank body seismic data, and the weight coefficients corresponding to multiple failure modes can be determined by combining the experience of experts in the fields of petrochemicals, process equipment, earthquake engineering, emergency rescue, etc. For example, α1, α2, α3 and α4 can be 0.15, 0.20, 0.35 and 0.30, respectively.

[0093] Optionally, historical earthquake data along the X-axis, Y-axis and Z-axis can be simultaneously input at the bottom of the two saddle supports, and the seismic dynamic response of the tank body at each moment can be calculated and analyzed under the action of earthquake acceleration. The data such as the Mises stress σ1 of the tank wall of the horizontal oil and gas storage tank, the vertical compressive stress σ2 of the saddle support, the tensile stress σ3 of the anchor bolts and the shear stress τ of the anchor bolts at each moment of the earthquake can be obtained through calculation.

[0094] Step S530: Based on a multiple regression algorithm, an initial prediction model is determined in combination with the structural parameters of the tank and the damage index function.

[0095] In order to differentiate and characterize the results of the prediction and evaluation, the multi-distance regression algorithm can be used to combine the structural parameters of the tank and the damage index function to calculate and obtain the corresponding initial prediction model.

[0096] Optionally, the structural parameters of the tank body may include multiple parameters such as the tank size, saddle type, anchor bolt size, and filling coefficient of the tank body. The damage index function can be simplified based on the multi-distance regression algorithm and the structural parameters of the tank body. The tank structure size, saddle support type, anchor bolt size, filling coefficient of the tank, long-period characteristics of near-fault seismic motion in the target seismic data, and peak acceleration (PGA) are used as the influencing factors of seismic damage of horizontal liquid storage tanks. Different factor levels are set for each factor to construct an initial prediction model. The initial prediction model is used to perform seismic response analysis, extract the response result data at each moment (such as σ3, σ2, σ3, and τ), and calculate the damage index at each moment based on the damage index function. The results of the seismic response analysis of horizontal oil and gas storage tanks are used as the sample space, and the initial prediction model is established by the multivariate regression analysis method. The initial prediction model D may include:

[0097] D = f (structural parameters, target ground motion data);

[0098] The initial prediction model can calculate the damage prediction value of the tank under different tank structure sizes, saddle support types, anchor bolt sizes, tank filling coefficients, long-period characteristics of seismic motion, and peak acceleration.

[0099] exist Figure 3 In the embodiment shown, according to different types of failure conditions on the tank body, a corresponding initial prediction model with predictive evaluation function can be constructed in combination with a multivariate regression algorithm, and the overall damage condition of the tank body can be evaluated, effectively improving the comprehensiveness of the initial prediction model in predictive evaluation.

[0100] Optionally, see Figure 4 , Figure 4 A detailed flowchart of step S200 is provided for an embodiment of the present application. Step S200 may include steps S210-S240.

[0101] Step S210, determining historical earthquake data according to a preset fault distance.

[0102] When screening historical earthquake data, considering the validity of the distance range, corresponding historical earthquake data can be selected according to the preset fault distance.

[0103] Optionally, the fault throw refers to the relative distance between the corresponding rock layers of the faulted rock layers on the two plates. The fault throw value is different on the profiles in different orientations. In geology, the fault distance is sometimes understood as a concept related to the fault throw, or the vertical distance from the building to the seismic fault, or the shortest distance from the ground observation station to the projection of the fault on the ground. The preset fault distance can be set to 20km to screen out multiple historical earthquake acceleration records with a fault distance less than 20km from the database as historical earthquake data.

[0104] Step S220, performing empirical mode decomposition processing on the historical seismic data to obtain initial components corresponding to a plurality of intrinsic mode functions.

[0105] Among them, the historical earthquake data can be subjected to empirical mode decomposition processing to decompose the historical earthquake data into initial components corresponding to a plurality of eigenmode functions.

[0106] Optionally, an empirical mode decomposition (EMD) can be performed on the selected seismic acceleration record, and the record can be decomposed into a number of intrinsic mode functions (IMFs). EMD decomposition can be divided into the following three steps:

[0107] (1) Determine all local extreme points of the historical earthquake acceleration records x(t) in the historical earthquake data, and connect all local maxima and local minima with cubic spline curves to form upper envelopes x max and the lower envelope x min .

[0108] (2) Find the average value of the upper and lower envelopes, denoted as m 11 (t), subtract the average value from the original record x(t) to get the new data sequence h 11 (t):

[0109] m 11 (t) = 0.5[xmin max ];

[0110] h 11 (t) = x(t) - m 11 (t);

[0111] Judge h 11 (t) Two conditions for whether IMF is met may include: over the entire duration of the record, the number of zero crossings and the number of extreme points must be equal or differ by at most one, and at any time, the average value between the upper envelope defined by the local maximum and the lower envelope defined by the local minimum is zero.

[0112] If it is not satisfied, h 11 (t) Repeat the above process as the original data until a new data sequence is obtained.

[0113] h 1k (t) = h 1(k-1) (t)-m 11 (t);

[0114] The two conditions of IMF are met, and the first IMF component c1(t) is obtained:

[0115] c1(t)=h 1k (t);

[0116] (3) Remove c1(t) from the original record x(t) to obtain the remaining sequence r1(t):

[0117] r1(t)=x(t)-c1(t);

[0118] Take r1(t) as the new signal sequence and follow the above steps to obtain the second, third, ..., and nth intrinsic mode functions c n (t), when the residual quantity r n When (t) is a monotonic function of time or is less than a predetermined value, the EMD decomposition ends. The original record can be expressed as the sum of each IMF component and the final residual, that is:

[0119]

[0120] Perform Hilbert transform on each IMF and get the corresponding Hilber spectrum:

[0121]

[0122] Where ω is the original vibration frequency of the IMF component. By integrating H(ω,ω) with respect to time, we can obtain the Hilbert amplitude marginal spectrum:

[0123]

[0124] By integrating the square of H(ω,t) over time, we can obtain the Hilbert energy spectrum ES(ω):

[0125] ES(ω)=∫ ω H(ω,t)2 dt.

[0126] Step S230 , screening the initial components based on a preset low-frequency threshold to obtain a plurality of low-frequency components that are less than or equal to the low-frequency threshold.

[0127] In order to select the low-frequency effective data, the multiple initial components may be screened according to a preset low-frequency threshold to obtain multiple low-frequency components that are less than or equal to the low-frequency threshold.

[0128] Optionally, the IMF components of the historical earthquake data after EMD decomposition can be divided into vibration frequencies. Sort by high and low, and set the corresponding low frequency threshold, such as 2Hz, to filter each IMF component and obtain multiple f 0i The IMF components <2 Hz are regarded as low-frequency components.

[0129] Step S240 , screening each historical earthquake data based on the proportion value of the low-frequency component and the low-frequency proportion threshold to obtain target seismic data.

[0130] Among them, taking into account the continuous distribution of low-frequency components in the overall historical earthquake data, the historical earthquake data can be screened by comparing the proportion of low-frequency components with the preset low-frequency proportion threshold, and taking the high proportion of low-frequency components as the long-period feature to obtain the target seismic motion data that meets the low-frequency conditions and long-period characteristics.

[0131] Optionally, it is also possible to calculate and count the energy ratio of each IMF component to the original record, that is, the energy ratio of each IMF component in the original historical seismic data, set a corresponding low-frequency ratio threshold, for example 80%, and select records with at least multiple consecutive low-frequency components with an energy ratio cumulatively greater than 80% as target seismic data with significant long-period characteristics.

[0132] exist Figure 4 In the illustrated embodiment, various processes such as decomposition, frequency screening, and period analysis can be performed on historical earthquake data, thereby effectively improving the validity of the obtained target earthquake data.

[0133] Optionally, see Figure 5 , Figure 5 A detailed flowchart of step S300 is provided for an embodiment of the present application. Step S300 may include steps S310-S330.

[0134] Step S310: construct a structural model according to the structural characteristics of the tank.

[0135] Among them, taking into account the differences in the shapes, sizes and other structures of different types of oil and gas tanks and other tanks, a corresponding structural model can be constructed based on the structural characteristics of the tank that needs to be predicted and evaluated.

[0136] Optionally, a structural model of the tank can be established based on the structural characteristics of the tank, such as the tank structure size, the physical and mechanical performance parameters of the steel and the stored liquid, the filling coefficient of the tank, etc. The structural model can be a FEM model, using a four-node curved shell element to simulate the steel, and using a kinematic hardening bilinear elastoplastic model that obeys the Von Mises yield criterion to simulate the mechanical properties of the steel.

[0137] Step S320: construct a liquid storage model according to the fluid characteristics of the liquid stored in the tank.

[0138] Among them, considering that there is some liquid inside the tanks such as oil and gas tanks, and the liquid will shake accordingly during an earthquake, in order to simulate the shaking of the liquid, a corresponding liquid storage model can be constructed according to the fluid properties of the liquid stored in the tank.

[0139] Optionally, the liquid storage model may be an SPH model, which may use continuous particle units to simulate liquid storage with arbitrary filling coefficients, and use the Mie-Grüneisen equation in the form of Us-Up to simulate the mechanical properties of liquid storage with a free liquid surface.

[0140] Step S330, performing coupling processing according to the structural model and the liquid storage model to obtain an analysis model.

[0141] The structural model and the liquid storage model can be coupled to obtain an analysis model that can simulate the interaction between the liquid storage and the tank structure.

[0142] Optionally, the SPH-FEM coupling algorithm can be used to simulate the interaction between the tank wall and the interface between the tank and the liquid storage. The point-surface contact penalty function method can be used to apply the dynamic parameters of the SPH particles to the surface of the FEM unit. The FEM unit is regarded as the contact master surface, the SPH particle is regarded as the slave node, and the penalty function is simplified to a contact spring with normal stiffness connecting the FEM unit and the SPH particle. The penalty function detects the penetration state of the SPH particle and the surface of the FEM unit in each time step. If no penetration occurs, no processing is performed. If penetration occurs, contact force is introduced to limit the SPH particle from penetrating the FEM unit surface. Introducing the SPH-FEM coupling algorithm into the seismic analysis of oil and gas equipment can efficiently simulate the interaction between the interface between the liquid storage and the liquid storage.

[0143] exist Figure 5In the illustrated embodiment, by constructing a liquid storage model and coupling the liquid storage model with the structural model, the actual situation between the tank structure and the liquid inside the tank under earthquake conditions is effectively simulated, thereby effectively improving the effectiveness of the analysis model and thereby improving the effectiveness of training the initial detection model.

[0144] See also Figure 6 , Figure 6 A flowchart of an evaluation method provided in an embodiment of the present application, the method may include steps S600-S800.

[0145] Step S600, determining the seismic data and earthquake level index of the tank body to be tested according to the geographical data of the tank body to be tested.

[0146] Among them, due to the differences in the distribution areas of the tanks that need to be evaluated and tested, and different areas have different earthquake conditions, the corresponding local seismic data and earthquake level indicators to be tested can be determined based on the geographical data of each tank to be tested.

[0147] For example, the tank body to be tested may be various types of oil and gas storage tanks, such as a horizontal oil and gas storage tank.

[0148] Optionally, the seismic data to be measured may include simulated seismic recording data, and the seismic recording data may include long-period seismic characteristics and PGA. The earthquake level index is the corresponding earthquake fortification level, which can be set according to industry standards, and can characterize the magnitude and probability of the seismic effects that the tank may suffer during its life cycle. For example, according to my country's "Petrochemical Steel Equipment Seismic Design Standard" (GB / T 50761-2018) and "Building Seismic Code" (GB 50011-2010) and other standards, a three-level earthquake fortification level can be adopted, and the annual exceedance probability or earthquake recurrence period can be used to indicate the determination of the earthquake level index. Please refer to the following table:

[0149] Defense level Frequent earthquakes Earthquake protection Rare earthquake Annual Exceedance Probability % 63% in 50 years 10% within 50 years 2% in 50 years Return period (years) 50475 475 2475

[0150] Among them, a frequent earthquake refers to an earthquake with a probability of exceeding 63% (recurrence period of 50 years) that may be encountered within a 50-year period. This usually means that the area is likely to experience earthquakes of this magnitude multiple times over a long period of time. The magnitude of a frequent earthquake is relatively small, but the frequency of occurrence is high. Due to the frequency of such earthquakes, special consideration needs to be given to the seismic performance of building structures during design and construction to ensure that they can maintain good stability and safety under multiple earthquakes. The fortification earthquake is the earthquake intensity or acceleration value used to determine the seismic fortification standard of a building in the seismic design of a building. It is usually determined based on factors such as the use function and importance of the building and the level of seismic activity in the area where it is located. The purpose of the fortification earthquake is to ensure that the building can maintain structural integrity and not collapse when subjected to a certain degree of earthquake action, and to minimize damage to personnel and property. Through reasonable seismic fortification, the seismic performance of buildings can be improved and the safety of people's lives and property can be guaranteed. A rare earthquake refers to an earthquake with a probability of exceeding 2% that may be encountered within a 50-year period. The probability of this type of earthquake is very low, but once it occurs, its destructive power will be very huge. Rare earthquakes are usually larger in magnitude and accompanied by strong ground motion. Due to the rarity and destructiveness of this type of earthquake, more stringent earthquake-resistant measures need to be taken when designing and constructing building structures to ensure that they can still maintain good stability and safety under extreme conditions.

[0151] Step S700, bringing the to-be-tested seismic data and target structural parameters of the to-be-tested tank into a target prediction model for prediction, and obtaining predicted damage levels of the to-be-tested tank under multiple earthquake level indicators.

[0152] Among them, taking into account the structural differences of each tank body to be tested, the seismic data and target structural parameters of the tank body to be tested can be brought into the target prediction model obtained in the model training method for prediction and evaluation, so as to obtain the predicted damage level of the tank body to be tested under multiple earthquake level indicators.

[0153] Optionally, the target prediction model may output a corresponding damage prediction value. Since the damage prediction value has an associated damage level, the predicted damage level corresponding to the numerical range of the damage prediction value may be determined according to the size of the damage prediction value.

[0154] Step S800: determining an evaluation result of the tank to be tested according to a plurality of predicted damage levels.

[0155] Among them, the evaluation result of the seismic performance of the tank to be tested can be determined according to the actual situation of the predicted damage level.

[0156] Optionally, the evaluation result may include two results, namely, the seismic performance of the tank body to be tested meets the seismic requirements, and the seismic performance of the tank body to be tested does not meet the seismic requirements.

[0157] It should be noted that when the evaluation result is that the seismic performance of the tank to be tested does not meet the seismic requirements, optimization prompt data for the tank to be tested can be generated based on the predicted damage level. When the evaluation result is that the seismic performance of the tank to be tested does not meet the seismic requirements, in order to improve the safety and stability of the tank put into use under earthquake conditions, an analysis can be performed based on the predicted damage level to generate optimization prompt data for the tank to be tested, so that users can refer to the optimization prompt data and optimize various related parameters such as the structure, material, and installation location of the tank until the evaluation result of the tank to be tested is that the seismic performance meets the seismic requirements before it can be officially put into use. The tank to be tested can be iteratively optimized based on the evaluation results to ensure that the seismic performance of the tank to be tested can meet the seismic requirements.

[0158] For example, the areas where the seismic resistance of the tank body to be tested is relatively weak, such as the tank body, saddle and other areas, can be determined based on the optimization prompt data, so that the weak areas can be reinforced and structurally modified, for example, the material thickness of the tank body can be thickened, or the force-bearing structure of the saddle can be modified.

[0159] exist Figure 6 In the illustrated embodiment, the target prediction model can be used to predict and evaluate the damage of the tank body to be tested under different types of earthquakes according to the actual structural conditions and geographical conditions of the tank body to be tested, thereby improving the effectiveness of the evaluation results of the seismic performance of the tank body to be tested and meeting the installation evaluation requirements of various tank bodies. It is also possible to determine whether the tank body to be tested meets the corresponding seismic safety requirements before it is put into use based on the evaluation results, thereby reducing the use of tanks that do not meet the seismic safety requirements, and further reducing the damage to the tanks in use during earthquakes.

[0160] It should be noted that the earthquake level index may include: frequent earthquakes, fortification earthquakes and rare earthquakes. Since the earthquake level index includes earthquakes of various probabilities and severity, the target prediction model can obtain multiple predicted damage levels for different earthquake level indexes. Therefore, the predicted damage level may include: the first predicted damage level of the tank under frequent earthquakes, the second predicted damage level under fortification earthquakes and the third predicted damage level under rare earthquakes.

[0161] Optionally, see Figure 7 , Figure 7 A detailed flowchart of step S800 is provided for an embodiment of the present application. Step S800 may include steps S810-S830.

[0162] Step S810: determining a plurality of performance indicators associated with the damage level based on historical tank seismic data.

[0163] Among them, when conducting an assessment, multiple performance indicators associated with the damage level can be determined based on historical tank seismic data.

[0164] Optionally, multiple performance levels of the tank corresponding to various damage levels can be determined as performance indicators based on historical tank earthquake data. The performance level can characterize the expected operating state and maximum degree of damage of the tank under a specific fortification earthquake. The corresponding relationship between the performance indicator and the damage level can be referred to in the following table:

[0165] Performance Indicators Normal operation Restricted Operation Stop running tear down tear down Damage level Basically intact Minor damage Moderate damage Severe damage destruction

[0166] Step S820, determining a first target performance indicator corresponding to the second predicted damage level and a second target performance indicator corresponding to the third predicted damage level.

[0167] Among them, taking into account the severity of the two earthquake level indicators, the fortified earthquake and the Chinese earthquake, the second predicted damage level and the third predicted damage level corresponding to the two earthquake conditions can be associated with the performance indicators to determine the two target performance indicators of the tank under two more severe earthquake conditions.

[0168] Step S830: If it is determined that the first target performance indicator represents the normal operation of the tank body under test, and the second target performance indicator represents the restricted operation of the tank body under test, then it is determined that the evaluation result is that the seismic performance of the tank body under test meets the seismic requirements.

[0169] Among them, according to the actual performance of the target performance indicators corresponding to the tank under test under these two earthquake conditions, it can be judged whether the seismic performance of the tank under test meets the seismic requirements and obtain the corresponding evaluation results. Considering that the tank under test near the active fault is damaged by the sloshing of the liquid storage due to the long-period characteristics of the earthquake motion, the evaluation method for evaluating the tank near the active fault may include: the tank under test can operate normally under the action of the fortified earthquake and remain basically intact during the earthquake; the operation of the tank under test is restricted under the action of rare earthquakes, and slight damage may occur during the earthquake, but no moderate or above damage may occur. If the above two requirements are met at the same time, it is determined that the evaluation result of the tank under test near the active fault is that the seismic resistance meets the seismic requirements; if any of the requirements are not met, the evaluation result of the tank under test is that the seismic resistance does not meet the seismic requirements, and measures can be taken to improve its seismic resistance, and then the seismic resistance evaluation is carried out again until its evaluation result is that the seismic resistance meets the seismic requirements.

[0170] exist Figure 7In the embodiment shown, a variety of different types of earthquake conditions can be classified and predicted, and the predicted damage level can be associated with a performance indicator that characterizes whether the tank to be tested can be used, thereby determining a corresponding evaluation result based on the performance of the tank to be tested under earthquake conditions of different levels, thereby effectively improving the validity and correctness of the evaluation result.

[0171] It should be noted that the failure modes, damage prediction values, damage levels, performance indicators, earthquake level indicators, etc. in the embodiments of the present application can be set and adjusted according to actual conditions, and other types of embodiments will not be described in detail.

[0172] An embodiment of the present application also provides a computer program product, which includes a computer program / instructions. When the computer program / instructions are executed by a processor, the steps in any implementation of the above-mentioned model training method and / or evaluation method are implemented.

[0173] In several embodiments provided in the present application, it should be understood that the disclosed device can also be implemented in other ways. The device embodiments described above are merely schematic, for example, the block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices according to the multiple embodiments of the present application. In this regard, each box in the block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram, and the combination of the block diagrams, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0174] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0175] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0176] The above description is only an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0177] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

[0178] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

Claims

1. A model training method, characterized in that: The method comprises: Screen historical earthquake data to obtain target earthquake data that meets low-frequency conditions and long-period characteristics; The analysis model is obtained by coupling the tank structure model and the liquid storage model; The initial prediction model is trained according to the target seismic data and the analysis model to obtain a target prediction model.

2. The method according to claim 1, characterized in that in, The initial prediction model is constructed in the following manner: Determine multiple failure modes of the tank body under earthquake conditions; wherein the failure modes include: at least one of tank body tensile failure, saddle support compressive failure, and anchor bolt tensile and shear failure; Based on the material parameters of the tank corresponding to the multiple failure modes, the damage index function is determined in combination with the weight coefficient of each failure mode; wherein the material parameters include: the ultimate strength of the steel of the tank and the seismic allowable stress of the steel at multiple locations; the weight coefficient is determined based on historical tank seismic data; Based on a multiple regression algorithm, the initial prediction model is determined in combination with the structural parameters of the tank body and the damage index function.

3. The method according to claim 1, characterized in that The screening of historical earthquake data to obtain target earthquake data that meets low-frequency conditions and long-period characteristics includes: Determining the historical earthquake data according to a preset fault distance; Performing empirical mode decomposition processing on the historical seismic data to obtain initial components corresponding to a plurality of eigenmode functions; Screening the initial components based on a preset low-frequency threshold to obtain a plurality of low-frequency components that are less than or equal to the low-frequency threshold; Each of the historical earthquake data is screened based on the proportion value of the low-frequency component and the low-frequency proportion threshold to obtain the target seismic data.

4. The method according to claim 1, characterized in that The analysis model is obtained by coupling the tank structure model and the liquid storage model, including: Constructing the structural model according to the structural characteristics of the tank; Constructing the liquid storage model according to the fluid characteristics of the liquid stored in the tank; The analysis model is obtained by performing coupling processing on the structural model and the liquid storage model.

5. The method according to claim 1, characterized in that in, The output result of the target prediction model includes the damage prediction value of the tank body and the damage level corresponding to the numerical range of each damage prediction value; The damage levels are determined based on industry standards.

6. An evaluation method, characterized in that The method comprises: Determining the seismic data and earthquake level index of the tank to be tested according to the geographical data of the tank to be tested; Bringing the seismic data to be tested and the target structural parameters of the tank body to be tested into the target prediction model for prediction, and obtaining the predicted damage level of the tank body to be tested under the multiple earthquake level indicators; wherein the target prediction model is a model trained by any one of the methods in claims 1-5; An evaluation result of the tank to be tested is determined according to the multiple predicted damage levels.

7. The method according to claim 6, characterized in that in, The earthquake level index includes: frequent earthquakes, fortification earthquakes and rare earthquakes; the predicted damage level includes: the first predicted damage level of the tank under the frequent earthquake condition, the second predicted damage level under the fortification earthquake and the third predicted damage level under the rare earthquake condition; Determining the evaluation result of the tank to be tested according to the plurality of predicted damage levels includes: Based on historical tank seismic data, multiple performance indicators associated with damage levels are determined; Determining a first target performance indicator corresponding to the second predicted damage level and a second target performance indicator corresponding to the third predicted damage level; If it is determined that the first target performance indicator represents the normal operation of the tank body to be tested, and the second target performance indicator represents the restricted operation of the tank body to be tested, then it is determined that the evaluation result is that the seismic performance of the tank body to be tested meets the seismic requirements.

8. The method according to claim 6, characterized in that The method further comprises: When the evaluation result is that the seismic performance of the tank body to be tested does not meet the seismic requirements, optimization prompt data for the tank body to be tested is generated based on the predicted damage level.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, wherein the memory stores program instructions, and when the processor runs the program instructions, the steps in the method according to any one of claims 1 to 8 are executed.

10. A computer program product, characterized in that The computer program product comprises a computer program / instruction, and when the computer program / instruction is executed by a processor, the steps in the method according to any one of claims 1 to 8 are implemented.

Images