Container high and low temperature fatigue detection method, electronic equipment and medium

Abstract

The invention relates to the technical field of container detection, in particular to a container high and low temperature fatigue detection method, electronic equipment and a medium, and the method comprises the steps: obtaining a first optimal fluid parameter and a second optimal fluid parameter based on a first preset strategy; obtaining a first data set and a second data set based on a measuring device arranged in the to-be-measured container; determining a first accumulative fatigue damage based on the first data set according to a first model, and determining a second accumulative fatigue damage based on the second data set according to a second model; based on the first accumulative fatigue damage and the second accumulative fatigue damage, total accumulative fatigue damage is determined.

Description

Technical Field

[0001] This invention relates to the field of container testing technology, and in particular to a method, electronic device and medium for testing container high and low temperature fatigue. Background Technology

[0002] Gaseous energy sources such as hydrogen and natural gas have wide applications in industry and daily life, with operating pressures ranging from atmospheric pressure to tens of megapascals. During use, the cylinders and valves storing these gases are subject to repeated filling and discharging, leading to repeated stress and temperature changes, which can cause damage, deteriorate the performance of the cylinders and valves, and even cause cracks. Furthermore, under high pressure, hydrogen can also cause the mechanical properties of some metallic materials to deteriorate. Therefore, standards such as UN-GTR13 and ANSI V2-2007 stipulate the need for cyclic testing with the stored gas to ensure the safe and stable operation of the product within its design life.

[0003] In high and low temperature environmental chambers, when conducting high and low temperature fatigue tests on containers under test, a fluid medium needs to be introduced into the container. During the test, it is essential to ensure uniform temperature distribution of the fluid throughout the container to ensure more accurate high and low temperature fatigue testing.

[0004] When there is a temperature difference within an object or when two objects at different temperatures come into contact, the thermal motion of the material's particles (molecules, atoms, or free electrons) transfers heat. When a high-pressure fluid is input from the inlet, because the fluid temperature at the inlet is higher, this thermal motion causes heat to transfer from the high-temperature region to the low-temperature region, thus raising the local temperature at the inlet.

[0005] 2. When a fluid flows over a solid wall, the combined effects of heat convection and conduction create surface convective heat transfer. When a high-pressure fluid is introduced from the inlet, the fluid flow carries away heat, but because the fluid velocity is higher at the inlet, the convective heat transfer is more significant, leading to an increase in the inlet temperature.

[0006] 3. An object continuously emits thermal radiation energy into the surrounding space, which is absorbed by surrounding objects. Simultaneously, the object also continuously receives heat energy radiated from surrounding objects. Under high-pressure fluid input, radiative heat transfer occurs due to the temperature difference between the fluid and the workpiece, causing a localized temperature increase at the inlet end.

[0007] When high-pressure fluid is input from the inlet end of the workpiece, it will cause the temperature difference between various temperature measuring points in the container to be too large, which will lead to inaccurate test results. Summary of the Invention

[0008] In view of this, one of the objectives of the present invention is to provide a method for high and low temperature fatigue testing of containers, which can solve the problem of inaccurate test results caused by the high local temperature at the inlet end in the prior art.

[0009] The present invention solves the above-mentioned technical problems through the following technical means:

[0010] In a first aspect, embodiments of this application propose a method for high and low temperature fatigue testing of a container, the method comprising:

[0011] Based on the first preset strategy, the first optimal fluid parameters and the second optimal fluid parameters are obtained. After the standard gas of the first optimal fluid parameters is introduced into the container to be tested, and the detection gas of the second optimal fluid parameters is introduced into the container to be tested, the standard deviation of the fluid temperature at all specified locations in the container to be tested is less than a specified value.

[0012] Based on the measuring device, a first dataset and a second dataset are obtained, wherein the first dataset and the second dataset respectively include N first fatigue parameters and N second fatigue parameters. The first fatigue parameters and the second fatigue parameters are both acquired by the measuring device. The N first fatigue parameters and the N second fatigue parameters are obtained by a first specified cycle rule and a second specified cycle rule, respectively. The first specified cycle rule represents that, based on the first optimal fluid parameters, the standard gas with the maximum temperature and the standard gas with the minimum temperature are alternately introduced into the container to be tested. The second specified cycle rule represents that, based on the second optimal fluid parameters, the detection gas with the maximum temperature and the detection gas with the minimum temperature are alternately introduced into the container to be tested. N represents a natural number greater than 1.

[0013] Based on the first dataset and according to the first model, a first cumulative fatigue damage is determined, and based on the second dataset and according to the second model, a second cumulative fatigue damage is determined.

[0014] The total cumulative fatigue damage is determined based on the first cumulative fatigue damage and the second cumulative fatigue damage.

[0015] Further, determining the first cumulative fatigue damage based on the first dataset, and determining the second cumulative fatigue damage based on the second dataset, includes:

[0016] The first dataset and the type of the standard gas are input into the trained deep learning model, and the first initial cumulative fatigue damage is obtained through the trained deep learning.

[0017] The second dataset and the type of the detected gas are input into the trained deep learning model, and the second initial cumulative fatigue damage is obtained through the trained deep learning model.

[0018] Based on the second preset strategy, a first cumulative fatigue damage is determined according to the first initial cumulative fatigue damage, and a second cumulative fatigue damage is determined according to the second initial cumulative fatigue damage.

[0019] Furthermore, the step of determining the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the second cumulative fatigue damage based on the second initial cumulative fatigue damage, according to the second preset strategy, includes:

[0020] The first specified cyclic rule is converted into a first instruction and input into a preset virtual scene. The first specified cyclic rule is simulated N times through the virtual container to obtain a first virtual cumulative fatigue damage. The second specified cyclic rule is converted into a second instruction and input into a preset virtual scene. The second specified cyclic rule is simulated N times through the virtual container to obtain a second virtual cumulative fatigue damage.

[0021] Calculate a first difference between the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and calculate a second difference between the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage. When either the first difference or the second difference is greater than a difference threshold, determine the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and determine the second cumulative fatigue damage based on the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage. When either the first difference or the second difference is less than or equal to a difference threshold, the first initial cumulative fatigue damage is the first cumulative fatigue damage, and the second initial cumulative fatigue damage is the second cumulative fatigue damage.

[0022] Further, determining the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and determining the second cumulative fatigue damage based on the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage, includes:

[0023] Based on the first difference and the second difference, a first initial weight and a second initial weight are determined according to a preset weight table. The first initial weight is negatively correlated with the first difference, and the second initial weight is negatively correlated with the second difference.

[0024] Based on the first initial weight, the first initial cumulative fatigue damage, and the first virtual cumulative fatigue damage, the first cumulative fatigue damage is determined according to the weighted fusion method.

[0025] Based on the second initial weight, the second initial cumulative fatigue damage, and the second virtual cumulative fatigue damage, the second cumulative fatigue damage is determined according to the weighted fusion method.

[0026] Furthermore, the first fatigue parameter includes the first fluid temperature, the first stress, and the first wall temperature at all fluid temperature measurement locations within the container under test;

[0027] The first model is:

[0028]

[0029] in, This indicates the number of times the first specified cycle rule is completed when standard gas is introduced into the container under test;

[0030] This represents the material fatigue life of the container under test when standard gas is introduced into the container under test, and the first specified cycle rule is followed for the i-th time.

[0031] This represents the basic damage to the container under test when the first specified loop rule is applied for the i-th time.

[0032] Indicates the first temperature standard deviation;

[0033] This represents the average value of all the first stresses after N cycles of the first specified rule.

[0034] The first stress obtained after completing the i-th first specified loop rule;

[0035] This represents the first cumulative fatigue damage;

[0036] Indicates the first temperature coefficient;

[0037] This represents the first stress correction factor.

[0038] Furthermore, the second fatigue parameter includes the second fluid temperature, the second stress, and the second wall temperature at all fluid temperature measurement locations within the container under test.

[0039] The second model is:

[0040]

[0041] This indicates the number of times the second specified cycle rule is completed when the detection gas is introduced into the container under test;

[0042] This indicates the material fatigue life of the container under test when the test gas is introduced into the container under test and the second specified cycle rule is performed for the i-th time.

[0043] This represents the basic damage to the container under test when the second specified loop rule is applied for the i-th time.

[0044] Indicates the second temperature standard deviation;

[0045] This represents the average value of all second stresses after N cycles of the second specified rule.

[0046] The second stress obtained after completing the i-th second specified cycle rule;

[0047] This is the second cumulative fatigue damage;

[0048] Indicates the hydrogen concentration in the detected gas;

[0049] Indicates wall temperature;

[0050] Indicates the second temperature coefficient;

[0051] This represents the second stress correction factor;

[0052] This represents the hydrogen embrittlement correction factor.

[0053] Furthermore, determining the total cumulative fatigue damage based on the first cumulative fatigue damage and the second cumulative fatigue damage includes:

[0054] Based on the third preset strategy, the first final weight of the first cumulative fatigue damage and the second final weight of the second cumulative fatigue damage are obtained.

[0055] Based on the first cumulative fatigue damage, the first final weight, the second cumulative fatigue damage, and the second final weight, the total cumulative fatigue damage is obtained;

[0056] The third preset strategy is:

[0057] Obtain the rated number of operations for the container under test;

[0058] A third initial weight is obtained based on the first cumulative fatigue damage, and a fourth initial weight is obtained based on the second cumulative fatigue damage;

[0059] Based on the third and fourth initial weights, the initial total cumulative fatigue damage is calculated.

[0060] Based on the third initial weight and the rated number of runs, the first number is obtained, and based on the fourth initial weight and the rated number of runs, the second number is obtained;

[0061] In the virtual scene, the first specified cycle rule for the first number of times and the second specified cycle rule for the second number of times are simulated on the virtual container to obtain the total virtual cumulative fatigue damage;

[0062] When the difference between the virtual total cumulative fatigue damage and the initial total cumulative fatigue damage is greater than a preset difference, the third initial weight and the fourth initial weight are adjusted until the difference between the virtual total cumulative fatigue damage and the initial total cumulative fatigue damage is less than or equal to the preset difference, at which point the third initial weight becomes the first final weight and the fourth initial weight becomes the second final weight.

[0063] Furthermore, the first optimal fluid parameters include a first optimal flow rate and a first optimal air intake angle, and the second optimal fluid parameters include a second optimal flow rate and a second optimal air intake angle.

[0064] The process of obtaining the first optimal fluid parameters and the second optimal fluid parameters based on the first preset strategy includes:

[0065] Based on user-issued instructions, a preset virtual scene is established in the specified finite element software, and a virtual container is built in the preset virtual scene.

[0066] Obtain the range of fluid inlet angle and the range of fluid flow rate;

[0067] Based on the range of fluid inlet angle and fluid velocity, the first optimal velocity, the first optimal inlet angle, the second optimal velocity, and the second optimal inlet angle are obtained according to the particle swarm optimization algorithm model. The discrete point values ​​of the fitness function of the particle swarm optimization algorithm model are obtained by performing finite element simulation in the virtual scene.

[0068] Secondly, embodiments of this application provide an electronic device, which includes a processor and a memory coupled to each other. The memory stores a computer program, and when the computer program is executed by the processor, the electronic device performs the above-described method.

[0069] Thirdly, embodiments of this application provide a computer storage medium storing a computer program that, when run on the computer, can execute the methods described above.

[0070] The beneficial effects of this invention are:

[0071] According to the technical solution proposed by the present invention, through a first preset strategy, in a preset virtual environment, with the aim of minimizing the temperature standard deviation, the first optimal fluid parameters and the second optimal fluid parameters are determined, thereby avoiding the problem of low fatigue detection accuracy due to the large temperature difference between the inlet end of the container under test and other locations during actual testing.

[0072] According to the technical solution proposed by this invention, by using a trained deep learning model, the first dataset, the second dataset, and the gas type are taken as input to initially obtain a relatively accurate initial cumulative fatigue damage. Subsequently, based on a second preset strategy, the actual cycle is simulated through a virtual scene, and the virtual and initial calculation results are compared to further correct the fatigue damage value, thereby improving the accuracy of fatigue damage determination, reducing errors caused by purely theoretical calculations or single model predictions, and improving calculation precision. Attached Figure Description

[0073] Figure 1 This is a flowchart of the method proposed in this invention;

[0074] Figure 2 This is a sub-flowchart of S110 of the method proposed in the embodiments of the present invention;

[0075] Figure 3 This is a sub-flowchart of S130 of the method proposed in the embodiments of the present invention;

[0076] Figure 4 This is a sub-flowchart of S140 of the method proposed in the embodiments of the present invention. Detailed Implementation

[0077] 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.

[0078] In the figures of this invention, the same or similar reference numerals correspond to the same or similar components. In the description of this invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the figure, they are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the figures are only for illustrative purposes and should not be construed as limiting this invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0079] This application provides an electronic device that may include a processing module and a storage module. The storage module stores a computer program, which, when executed by the processing module, enables the electronic device to perform the corresponding steps in the following container high and low temperature fatigue testing method.

[0080] Please refer to Figure 1 This application also provides a method for high and low temperature fatigue testing of containers. The method may include the following steps:

[0081] S110, based on the first preset strategy, obtain the first optimal fluid parameters and the second optimal fluid parameters, wherein after the standard gas of the first optimal fluid parameters is introduced into the container to be tested, and the detection gas of the second optimal fluid parameters is introduced into the container to be tested, the standard deviation of the fluid temperature at all specified locations in the container to be tested is less than a specified value.

[0082] S120, based on the measuring device, a first dataset and a second dataset are obtained, wherein the first dataset and the second dataset respectively include N first fatigue parameters and N second fatigue parameters. The first fatigue parameters and the second fatigue parameters are both acquired by the measuring device. The N first fatigue parameters and the N second fatigue parameters are obtained by a first specified cycle rule and a second specified cycle rule, respectively. The first specified cycle rule represents that, based on the first optimal fluid parameters, the standard gas with the maximum temperature and the standard gas with the minimum temperature are alternately introduced into the container to be tested. The second specified cycle rule represents that, based on the second optimal fluid parameters, the detection gas with the maximum temperature and the detection gas with the minimum temperature are alternately introduced into the container to be tested.

[0083] S130, based on the first dataset and according to the first model, determine the first cumulative fatigue damage, and based on the second dataset and according to the second model, determine the second cumulative fatigue damage;

[0084] S140, based on the first cumulative fatigue damage and the second cumulative fatigue damage, determine the total cumulative fatigue damage.

[0085] The following is a detailed explanation of each step in the high and low temperature fatigue testing method for containers:

[0086] In S110, the first step is to build a finite element model based on the user's instructions to obtain a preset virtual scene and a preset virtual container.

[0087] To achieve virtual pre-optimization of fluid parameters, a high-precision finite element model is constructed based on the actual structure and material properties of the container under test, which is then used to replace the physical container under test for initial parameter iteration.

[0088] First, geometric and mesh modeling is performed. Specifically, based on the container design drawings (e.g., a cylindrical container with a diameter of 1m and a length of 3m, including inlet pipes and end caps), a 1:1 geometric model is created using ANSYS or ABAQUS software. A hybrid structured and unstructured mesh is used: a fine structured mesh (2mm element size) is used for the inlet pipe and key stress-bearing parts of the container wall (welds, transition zones), while an unstructured mesh (5mm element size) is used for the fluid region inside the container. The total number of elements is ≥500,000 to ensure the accuracy of temperature and stress field calculations.

[0089] Definitions of materials and physical fields:

[0090] Container Material: Enter the actual material parameters (e.g., 304 stainless steel, elastic modulus 200 GPa, Poisson's ratio 0.3, thermal conductivity 16.3 W / (m·K), coefficient of linear expansion 17.2 × 10⁻⁶). -6 / ℃), and introduce temperature dependence (such as the thermal conductivity change curve in the range of -60℃ to 120℃).

[0091] Fluid media: The physical parameters of the standard gas (high-purity nitrogen, density 1.165 kg / m³, specific heat capacity 1039 J / (kg·K), thermal conductivity 0.026 W / (m·K)) and the detection gas (hydrogen, density 0.0899 kg / m³, specific heat capacity 14304 J / (kg·K), thermal conductivity 0.1815 W / (m·K)) are defined to simulate the characteristics of real fluids.

[0092] Physics field coupling: Enable "fluid-thermal-solid multi-physics field coupling" analysis, consider the influence of fluid flow on temperature distribution (convective heat transfer), the influence of temperature change on container wall stress (thermal stress), and the mechanical stress of fluid pressure on the container, to ensure that the virtual model can accurately reflect the physical response of the physical container.

[0093] Based on this, the first preset strategy mentioned above can be understood as a preset virtual scene built on finite element software, conducting virtual experiments on a virtual container, and determining the optimal fluid parameters and the second optimal fluid parameters based on the particle swarm optimization algorithm model, with the goal of minimizing the temperature standard deviation of several fluid temperature measurement positions inside the virtual container.

[0094] Therefore, according to Figure 2 As shown, S110 may specifically include the following steps:

[0095] S111: Based on the instructions issued by the user, a preset virtual scene is established in the specified finite element software, and the virtual container is built in the preset virtual scene;

[0096] S112: Obtain the range of fluid inlet angle and the range of fluid flow rate;

[0097] S113: Based on the range of the fluid inlet angle and the range of the fluid velocity, according to the particle swarm optimization (PSO) algorithm model, obtain the first optimal velocity, the first optimal inlet angle, the second optimal velocity, and the second optimal inlet angle, wherein the discrete point values ​​of the fitness function of the PSO algorithm model are obtained based on finite element simulation performed in the virtual scene. The first optimal fluid parameters include the first optimal velocity and the first optimal inlet angle, and the second optimal fluid parameters include the second optimal velocity and the second optimal inlet angle.

[0098] In S111, based on the above method, a preset virtual scene is built using finite element software, and a virtual container is built in the preset virtual scene.

[0099] In S112, the range of fluid inlet angle and the range of fluid flow rate are obtained from the instruction manual of the device for applying fluid and input to the user.

[0100] In S113, based on the particle swarm optimization model, the technical solution for obtaining the first optimal flow velocity, the first optimal inlet angle, the second optimal flow velocity, and the second optimal inlet angle can be as follows:

[0101] The optimization variables are set as "fluid velocity v (0.2-1.0m / s)" and "inlet angle θ (10°-50°)", the objective function is "minimize the maximum temperature deviation ΔT_max inside the container", the constraint condition is "pressure fluctuation ΔP of the physical container ≤ 5% of the design pressure", and the number of iterations is set to 60 to ensure the accuracy of virtual optimization.

[0102] The IA-PSO algorithm is used to iteratively optimize the finite element virtual container model, determining the optimal fluid input parameters under different temperature conditions for both standard and test gases, thus avoiding repeated adjustments during physical testing.

[0103] To address the detection requirements of standard gases within the typical temperature range (-40℃ to 80℃), fluid parameters are optimized using the IA-PSO algorithm:

[0104] Optimization goals and constraints:

[0105] Objective function: minΔT_max (the maximum temperature deviation of 7 virtual temperature measurement points inside the container, each virtual temperature measurement point is used to obtain the fluid temperature at the corresponding position inside the virtual container), while also taking into account "the maximum equivalent stress of the container wall σ_eq ≤ 80% of the material yield strength" (to avoid excessive stress on the container in the virtual model);

[0106] Constraints: Flow velocity v∈[0.3,1.0]m / s, inlet angle θ∈[15°,45°], pressure inside the container P∈[18,22]MPa (design pressure 20MPa, fluctuation ±2MPa);

[0107] Algorithm Iterative Optimization:

[0108] Particle initialization: Generate 60 particles, each corresponding to a set of (v,θ) parameter combinations. Input the finite element model, perform fluid-thermal-solid coupling analysis, and output ΔT_max and σ_eq for each set of parameters.

[0109] Adaptive Iteration:

[0110] In the early stages of iteration (1-20 times): a larger inertia weight (0.7-0.9) is used to expand the search range and avoid getting trapped in local optima;

[0111] Mid-cycle (21-40 iterations): Weights are linearly reduced to 0.5, balancing global and local searches;

[0112] In the later stages of iteration (41-60 times): the weights are reduced to 0.3-0.5, a particle mutation mechanism with a 10% probability is introduced, and the optimal parameters are searched more precisely.

[0113] Optimal parameter output: After the iteration, the optimal parameter combination that satisfies "ΔT_max≤0.7℃ and σ_eq≤240MPa (80% of the yield strength of 304 stainless steel of 300MPa)" is selected. For example, the optimal parameters for conventional high temperature conditions (80℃) are: v_opt1_h=0.6m / s, θ_opt1_h=30°; the optimal parameters for conventional low temperature conditions (-40℃) are: v_opt1_l=0.7m / s, θ_opt1_l=28°, forming a fluid parameter optimization table for standard gas conventional conditions.

[0114] To address the characteristics of the detection gas (which can be hydrogen) in the extreme temperature range (-60℃-120℃) (high thermal conductivity, prone to hydrogen embrittlement), the IA-PSO algorithm parameters are adjusted for virtual optimization.

[0115] In S120, the measuring device includes seven fluid temperature sensors, one wall temperature sensor, and one set of stress sensors. The seven fluid temperature sensors are arranged with three downstream of the inlet, two in the middle, and two at the far end. To ensure a one-to-one correspondence with the "virtual temperature measurement points" in the virtual model, the error is ≤5mm. In this embodiment, the positions of the fluid temperature sensors correspond to the temperature measurement positions within the virtual container in the preset virtual scene. Multiple stress sensors can be placed at different locations within the container to be measured, and the stress is obtained by averaging the results.

[0116] In this embodiment, the optimal flow rate v_opt and the optimal air intake angle θ_opt obtained in the previous step are input into the servo valve control system. No pre-ventilation debugging is required; they are directly set as the initial parameters for detection.

[0117] By executing the first specified loop rule N times, a first dataset is obtained. The first dataset includes N first fatigue parameters. Each time the first specified loop rule is completed, a first fatigue parameter is collected once. The first dataset is the set of all first fatigue parameters. In this embodiment, a method of collecting one first fatigue parameter and storing one first fatigue parameter is adopted so that the first fatigue parameters in the first dataset are arranged in chronological order.

[0118] In this embodiment, the first dataset is represented as (AIW1, AIW2, ..., AIW). i ...AIW N ), i represents the number of times the first specified loop rule is executed.

[0119] AIWi= , These represent the temperature values ​​measured by the seven fluid temperature sensors after the i-th iteration of the first specified loop rule. This represents the temperature value measured by the wall temperature sensor after the i-th completion of the first specified loop rule. This represents the stress value measured after the i-th completion of the first specified loop rule.

[0120] In this embodiment, the combination of all the first specified loop rules is to alternately introduce a standard gas with a first maximum temperature into the test container and a standard gas with a first minimum temperature into the test container. After each alternation is completed, the first fatigue parameter is obtained. After all the first specified loop rules are executed, the first dataset is obtained.

[0121] The second dataset is obtained in the same way as above, except that the second maximum temperature is greater than the first maximum temperature, and the second minimum temperature is less than the first minimum temperature. This is because the first specified cycle rule is used to test the collection of relevant data of the container under test under standard operating conditions. The standard gas has a fixed composition (e.g., nitrogen purity ≥ 99.999%) and stable physicochemical properties (no corrosiveness, no hydrogen embrittlement effect, and known and constant parameters such as thermal conductivity / specific heat capacity). After being introduced into the container, it only acts on the container through "high and low temperature cycles + pressure loads" and will not introduce additional chemical damage (e.g., hydrogen embrittlement caused by hydrogen or wall erosion caused by corrosive gases). The first dataset obtained through standard gas detection (e.g., strain amplitude Δε1 and temperature standard deviation σ_T1 under normal operating conditions) can accurately reflect the "fatigue response under mechanical load (temperature + pressure) alone", becoming the benchmark for subsequent comparison of the fatigue effect of the detection gas. For example, if the strain amplitude Δε2 under the detection gas scenario is 10% higher than Δε1, it can be clearly stated that this difference is caused by the characteristics of the detection gas (e.g., hydrogen embrittlement), rather than mechanical load or changes in the initial state of the container.

[0122] The gas being tested corresponds to "extreme operating conditions" (temperature -60℃ to 120℃, pressure 30MPa), which are the harsh service scenarios that containers may face (such as outdoor hydrogen storage in winter and high-temperature process gas). The "superimposed effect" of extreme operating conditions and the characteristics of the gas being tested can significantly amplify fatigue damage.

[0123] In some embodiments, S130 uses a trained deep learning model to calculate the first cumulative fatigue damage and the second cumulative fatigue damage.

[0124] Therefore, as Figure 3 As shown, S130 specifically includes the following steps:

[0125] S131: Input the first dataset and the type of the standard gas into the trained deep learning model, and obtain the first initial cumulative fatigue damage through the trained deep learning;

[0126] S132: Input the second dataset and the type of the detected gas into the trained deep learning model, and obtain the second initial cumulative fatigue damage through the trained deep learning model;

[0127] S133: Based on the second preset strategy, determine the first cumulative fatigue damage according to the first initial cumulative fatigue damage, and determine the second cumulative fatigue damage according to the second initial cumulative fatigue damage.

[0128] In this embodiment, since both the first and second datasets are time-series related data, the CNN-LSTM model is used as the deep learning model.

[0129] The architecture of a CNN-LSTM model can include:

[0130] Input layer: 9-dimensional features (8 temperature features: instantaneous temperatures from 7 fluid temperature sensors and 1 wall temperature sensor; 1 stress feature: instantaneous stress value collected by 1 stress sensor; gas type feature, where standard gas is represented by 0 and detected gas by 1).

[0131] CNN Feature Extraction Layer: 3 convolutional blocks (each block contains 3×3 convolutional kernels: capturing the local interactions of 3 consecutive time steps × 3 related features (such as fluid temperature-strain-pressure); ReLU activation function: introducing non-linearity to enhance feature expressiveness; max pooling layer: pooling kernel size 2×2, stride 2×2, compressing feature dimensions while retaining key local information), extracting local temperature-mechanical related features;

[0132] LSTM temporal modeling layer: 2-layer bidirectional LSTM (64 neurons per layer: forward LSTM captures the potential impact of subsequent cycles on the current damage, backward LSTM traces the damage accumulation process of historical cycles; 0.2 dropout is set for each layer to prevent overfitting), capturing the temporal accumulation law of fatigue damage during the cycle (such as "damage superposition of continuous fluid temperature fluctuation cycle" and "damage growth rate change after the strain peak appears" and other long temporal dependencies based on sensor time series data);

[0133] Transformer Attention Layer: 1 multi-head attention mechanism (8 attention heads: each head independently calculates feature weights, focusing on different key scenarios such as "high strain cycle", "sudden change in fluid temperature", and "abnormal flow rate"; scaled dot product attention calculates weights to avoid interference from differences in feature magnitude), strengthening the feature weights of key scenarios such as "high strain cycle" and "sudden change in fluid temperature" (e.g., the feature weight of high strain cycle is 2-3 times higher than that of regular cycle, highlighting the core impact of strain sensor data on fatigue damage);

[0134] Output layer: Fully connected layer (first layer with 32 neurons: performs dimensional transformation and nonlinear mapping on the fusion features output by the attention layer; second layer with 1 output neuron: restricts the output to the [0,1] range through the Sigmoid activation function), outputs the fatigue damage value D_i of a single cycle (value range 0-1, 0 represents no damage, 1 represents complete fatigue failure, the calculation logic is adapted to the correlation of measured data from 7 fluid temperatures, 1 wall temperature, and 1 strain sensor).

[0135] In S133, the second preset strategy can be understood as comparing the results of finite element analysis with the results of deep learning model output to determine the final first cumulative fatigue damage and second cumulative fatigue damage.

[0136] Therefore, S133 includes the following steps:

[0137] S1331: The first specified loop rule is converted into a first instruction and input into a preset virtual scene. The first specified loop rule is simulated N times through the virtual container to obtain a first virtual cumulative fatigue damage. The second specified loop rule is converted into a second instruction and input into a preset virtual scene. The second specified loop rule is simulated N times through the virtual container to obtain a second virtual cumulative fatigue damage.

[0138] S1332: Calculate a first difference between the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and calculate a second difference between the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage. When the first difference or the second difference is greater than a difference threshold, determine the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and determine the second cumulative fatigue damage based on the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage. When both the first difference and the second difference are less than or equal to the difference threshold, the first initial cumulative fatigue damage is the first cumulative fatigue damage, and the second initial cumulative fatigue damage is the second cumulative fatigue damage.

[0139] Specifically, in S1332, the method for determining the first cumulative fatigue damage and the second cumulative fatigue damage is as follows:

[0140] Based on the first difference and the second difference, a first initial weight and a second initial weight are determined according to a preset weight table. The first initial weight is negatively correlated with the first difference, and the second initial weight is negatively correlated with the second difference.

[0141] Based on the first initial weight, the first initial cumulative fatigue damage, and the first virtual cumulative fatigue damage, the first cumulative fatigue damage is determined according to the weighted fusion method.

[0142] Based on the second initial weight, the second initial cumulative fatigue damage, and the second virtual cumulative fatigue damage, the second cumulative fatigue damage is determined according to the weighted fusion method.

[0143] In S1331, under the virtual scene, the virtual container is simulated N times according to the first specified loop rule and N times according to the second specified loop rule to obtain the first virtual cumulative fatigue damage and the second virtual cumulative fatigue damage.

[0144] For example, based on the above process, the process of realizing the first cumulative fatigue damage and the second cumulative fatigue damage can be as follows:

[0145] Taking "high and low temperature fatigue testing of vehicle-mounted hydrogen storage cylinders" as an example, a single test needs to complete 50 cycles. The "threshold for the difference between initial cumulative fatigue damage and virtual cumulative fatigue damage" is set to 5% (based on engineering testing accuracy requirements). The preset weight table follows the principle that "the smaller the difference, the higher the initial damage weight output by the deep learning model." The specific process is as follows:

[0146] I. Basic Parameter Presets

[0147] Difference threshold: 5%, which needs to be dynamically adjusted based on the actual initial damage value (e.g., if the first initial damage is 0.3, the corresponding absolute threshold is 0.3 × 5% = 0.015).

[0148] Preset weight table (the sum of the initial weight and the virtual damage weight is 1):

[0149] Difference percentage ≤ 5%: Initial weight = 1.0, Virtual weight = 0.0;

[0150] For differences of 5% or less and ≤10% of the total difference: initial weight = 0.7, virtual weight = 0.3;

[0151] For differences of 10% or less and 15% or less: initial weight = 0.5, virtual weight = 0.5;

[0152] If the difference percentage is greater than 15%, the initial weight is 0.3 and the virtual weight is 0.7.

[0153] Scene parameters:

[0154] Standard gas (nitrogen): First specified cycle rule temperature -40℃ to 80℃, pressure 20MPa;

[0155] Detection gas (hydrogen): Second specified cycle rule temperature -60℃-120℃, pressure 30MPa.

[0156] II. Step S1331: Obtain virtual cumulative fatigue damage in a virtual scene

[0157] The first / second specified loop rule is converted into an instruction containing "optimal flow rate, inlet angle, and temperature and pressure parameters," which is then input into the finite element virtual container. After simulating 50 loops, the following results are obtained:

[0158] First virtual cumulative fatigue damage (standard gas): with the first optimal parameters applied (flow rate 0.62 m / s, angle 32°), the calculated value according to Miner's theory is 0.31;

[0159] Second virtual cumulative fatigue damage (detection gas): After loading the second optimal parameters (flow rate 0.52 m / s, angle 36°), considering the hydrogen embrittlement effect (acceleration coefficient 1.3), the result is 0.42.

[0160] III. S1332 Step: Difference Calculation and Final Damage Determination

[0161] 1. Obtain initial cumulative fatigue damage

[0162] Input the first / second dataset into the trained CNN-LSTM-Attention model, and the output will be:

[0163] First initial damage (standard gas): 0.3 (model auxiliary bias ≤3%);

[0164] Second initial damage (detection gas): 0.45 (the model has taken hydrogen characteristics into account, hydrogen embrittlement weight 0.4).

[0165] 2. Calculate the difference and determine the threshold.

[0166] Calculate the difference between "|Initial Damage - Virtual Damage|" and determine the result based on the percentage:

[0167] (1) Standard gas scenario

[0168] Difference: |0.3 - 0.31| = 0.01;

[0169] Percentage of difference: (0.01 / 0.3)×100%≈3.33%≤5%;

[0170] Weights: Initial weight = 1.0, virtual weight = 0.0.

[0171] (2) Gas detection scenario

[0172] Difference: |0.45-0.42|=0.03;

[0173] Percentage of difference: (0.03 / 0.45)×100%≈6.67% (5%<6.67%≤10%)

[0174] Weights: Initial weight = 0.7, virtual weight = 0.3.

[0175] 3. Weighted fusion yields the final damage.

[0176] Calculated as "Initial Damage × Initial Weight + Virtual Damage × Virtual Weight":

[0177] First final damage (standard gas): 0.3 × 1.0 + 0.31 × 0.0 = 0.3;

[0178] Second final damage (detection gas): 0.45×0.7+0.42×0.3=0.315+0.126=0.441.

[0179] IV. Special Cases (Difference Exceeding 3 Times the Threshold)

[0180] If the detection scenario involves a gaseous environment, with initial damage = 0.5 and virtual damage = 0.4:

[0181] The difference is 0.1, and the percentage is 20%, which is greater than 15%.

[0182] Weights: Initial weight = 0.3, virtual weight = 0.7;

[0183] Final damage: 0.5×0.3+0.4×0.7=0.15+0.28=0.43 (correction of model overestimation error).

[0184] In some other embodiments, the first cumulative fatigue damage and the second cumulative fatigue damage can be calculated using the formulas corresponding to the first model and the second model. Specifically:

[0185] The first model is:

[0186]

[0187] in, This indicates the number of times the first specified cycle rule is completed when standard gas is introduced into the container under test;

[0188] This represents the material fatigue life of the container under test when standard gas is introduced into the container under test, and the first specified cycle rule is followed for the i-th time.

[0189] This represents the basic damage to the container under test when the first specified loop rule is applied for the i-th time.

[0190] Indicates the first temperature standard deviation;

[0191] This represents the average value of all the first stresses after N cycles of the first specified rule.

[0192] The first stress obtained after completing the i-th first specified loop rule;

[0193] This represents the first cumulative fatigue damage;

[0194] Indicates the first temperature coefficient;

[0195] This represents the first stress correction factor.

[0196] The second model is:

[0197]

[0198] This indicates the number of times the second specified cycle rule is completed when the detection gas is introduced into the container under test;

[0199] This indicates the material fatigue life of the container under test when the test gas is introduced into the container under test and the second specified cycle rule is performed for the i-th time.

[0200] This represents the basic damage to the container under test when the second specified loop rule is applied for the i-th time.

[0201] Indicates the second temperature standard deviation;

[0202] This represents the average value of all second stresses after N cycles of the second specified rule.

[0203] The second stress obtained after completing the i-th second specified cycle rule;

[0204] This is the second cumulative fatigue damage;

[0205] Indicates the hydrogen concentration in the detected gas;

[0206] Indicates wall temperature;

[0207] Indicates the second temperature coefficient;

[0208] This represents the second stress correction factor;

[0209] This represents the hydrogen embrittlement correction factor.

[0210] For the second model, when the hydrogen concentration in the detected gas is 0, =1, which should be understood as meaning that when the hydrogen concentration is 0, the detected gas does not contain hydrogen, and this is the default value. =1, when the hydrogen concentration is greater than 0, then according to Calculate the hydrogen embrittlement correction factor.

[0211] For S140, the total cumulative fatigue damage needs to be determined based on the first and second cumulative fatigue damage obtained in the aforementioned steps. Based on this, as follows: Figure 4 As shown, S140 specifically includes the following steps:

[0212] S141: Based on the third preset strategy, obtain the first final weight of the first cumulative fatigue damage and the second final weight of the second cumulative fatigue damage;

[0213] S142: Based on the first cumulative fatigue damage, the first final weight, the second cumulative fatigue damage, and the second final weight, the total cumulative fatigue damage is obtained.

[0214] The third preset strategy in this embodiment is:

[0215] Obtain the rated number of operations for the container under test;

[0216] Based on the first cumulative fatigue damage and the second cumulative fatigue damage, a third initial weight and a fourth initial weight are obtained. In this step, the third initial weight is obtained by dividing the first cumulative fatigue damage by the sum of the first cumulative fatigue damage and the second cumulative fatigue damage, and the fourth initial weight is obtained by dividing the second cumulative fatigue damage by the sum of the first cumulative fatigue damage and the second cumulative fatigue damage.

[0217] Based on the third initial weight and the fourth initial weight, the initial total cumulative fatigue damage is calculated. In this step, the initial total cumulative fatigue damage is obtained by multiplying the fourth initial weight and the second cumulative fatigue damage, plus the product of the third initial weight and the first cumulative fatigue damage.

[0218] Based on the third initial weight and the rated number of runs, a first number is obtained, and based on the fourth initial weight and the rated number of runs, a second number is obtained. The first number is obtained by multiplying the third initial weight by the rated number of runs, and the second number is obtained by multiplying the fourth initial weight by the rated number of runs.

[0219] In the virtual scene, the first specified cycle rule for the first number of times and the second specified cycle rule for the second number of times are simulated on the virtual container to obtain the virtual total cumulative fatigue damage. This step is obtained through finite element simulation.

[0220] When the difference between the virtual total cumulative fatigue damage and the initial total cumulative fatigue damage is greater than a preset difference, the third initial weight and the fourth initial weight are adjusted until the difference between the virtual total cumulative fatigue damage and the initial total cumulative fatigue damage is less than or equal to the preset difference. At this point, the third initial weight becomes the first final weight, and the fourth initial weight becomes the second final weight. The adjustment method can be to adaptively increase or decrease the third initial weight and the fourth initial weight when the difference exceeds the preset difference.

[0221] Then, the total cumulative fatigue damage is obtained by multiplying the first cumulative fatigue damage by the first final weight, and adding the product of the second cumulative fatigue damage by the second final weight.

[0222] Thirdly, this application also provides a computer storage medium containing a computer program that, when run on a computer, can execute the above-mentioned intelligent detection method for container high and low temperature fatigue.

[0223] Based on the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by hardware or by using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application can be embodied in the form of a software product. The software product can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause a computer device (such as a personal computer, simulation device, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0224] In the embodiments provided in this application, it should be understood that the disclosed methods can also be implemented in other ways. The method embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or part of code, which includes one or more executable instructions for implementing a specified logical function. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions. Furthermore, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0225] The above are merely embodiments of this application and are not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for high and low temperature fatigue testing of containers, characterized in that, The method includes: Based on the first preset strategy, the first optimal fluid parameters and the second optimal fluid parameters are obtained. After the standard gas of the first optimal fluid parameters is introduced into the container to be tested, and the detection gas of the second optimal fluid parameters is introduced into the container to be tested, the standard deviation of the fluid temperature at all specified locations in the container to be tested is less than a specified value. Based on the measuring device, a first dataset and a second dataset are obtained, wherein the first dataset and the second dataset respectively include N first fatigue parameters and N second fatigue parameters. The first fatigue parameters and the second fatigue parameters are both acquired by the measuring device. The N first fatigue parameters and the N second fatigue parameters are obtained by a first specified cycle rule and a second specified cycle rule, respectively. The first specified cycle rule represents that, based on the first optimal fluid parameters, the standard gas with the maximum temperature and the standard gas with the minimum temperature are alternately introduced into the container to be tested. The second specified cycle rule represents that, based on the second optimal fluid parameters, the detection gas with the maximum temperature and the detection gas with the minimum temperature are alternately introduced into the container to be tested. N represents a natural number greater than 1. Based on the first dataset and according to the first model, a first cumulative fatigue damage is determined, and based on the second dataset and according to the second model, a second cumulative fatigue damage is determined. The total cumulative fatigue damage is determined based on the first cumulative fatigue damage and the second cumulative fatigue damage.

2. The method according to claim 1, characterized in that, The determination of a first cumulative fatigue damage based on the first dataset, and the determination of a second cumulative fatigue damage based on the second dataset, include: The first dataset and the type of the standard gas are input into the trained deep learning model, and the first initial cumulative fatigue damage is obtained through the trained deep learning. The second dataset and the type of the detected gas are input into the trained deep learning model, and the second initial cumulative fatigue damage is obtained through the trained deep learning model. Based on the second preset strategy, a first cumulative fatigue damage is determined according to the first initial cumulative fatigue damage, and a second cumulative fatigue damage is determined according to the second initial cumulative fatigue damage.

3. The method according to claim 2, characterized in that, The step of determining the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the second cumulative fatigue damage based on the second initial cumulative fatigue damage, according to the second preset strategy, includes: The first specified cyclic rule is converted into a first instruction and input into a preset virtual scene. The first specified cyclic rule is simulated N times through the virtual container to obtain a first virtual cumulative fatigue damage. The second specified cyclic rule is converted into a second instruction and input into a preset virtual scene. The second specified cyclic rule is simulated N times through the virtual container to obtain a second virtual cumulative fatigue damage. Calculate a first difference between the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and calculate a second difference between the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage. When either the first difference or the second difference is greater than a difference threshold, determine the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and determine the second cumulative fatigue damage based on the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage. When either the first difference or the second difference is less than or equal to a difference threshold, the first initial cumulative fatigue damage is the first cumulative fatigue damage, and the second initial cumulative fatigue damage is the second cumulative fatigue damage.

4. The method according to claim 3, characterized in that, The step of determining the first cumulative fatigue damage based on the first initial cumulative fatigue damage and the first virtual cumulative fatigue damage, and determining the second cumulative fatigue damage based on the second initial cumulative fatigue damage and the second virtual cumulative fatigue damage, includes: Based on the first difference and the second difference, a first initial weight and a second initial weight are determined according to a preset weight table. The first initial weight is negatively correlated with the first difference, and the second initial weight is negatively correlated with the second difference. Based on the first initial weight, the first initial cumulative fatigue damage, and the first virtual cumulative fatigue damage, the first cumulative fatigue damage is determined according to the weighted fusion method. Based on the second initial weight, the second initial cumulative fatigue damage, and the second virtual cumulative fatigue damage, the second cumulative fatigue damage is determined according to the weighted fusion method.

5. The method according to claim 1, characterized in that, The first fatigue parameters include the first fluid temperature, the first stress, and the first wall temperature at all fluid temperature measurement locations within the container under test. The first model is: ； in, This indicates the number of times the first specified cycle rule is completed when standard gas is introduced into the container under test; This represents the material fatigue life of the container under test when standard gas is introduced into the container under test, and the first specified cycle rule is followed for the i-th time. This represents the basic damage to the container under test when the first specified loop rule is applied for the i-th time. Indicates the first temperature standard deviation; This represents the average value of all the first stresses after N cycles of the first specified rule. The first stress obtained after completing the i-th first specified loop rule; This represents the first cumulative fatigue damage; Indicates the first temperature coefficient; This represents the first stress correction factor.

6. The method according to claim 1, characterized in that, The second fatigue parameters include the second fluid temperature, the second stress, and the second wall temperature at all fluid temperature measurement locations within the container under test. The second model is: ； This indicates the number of times the second specified cycle rule is completed when the detection gas is introduced into the container under test; This indicates the material fatigue life of the container under test when the test gas is introduced into the container under test and the second specified cycle rule is performed for the i-th time. This represents the basic damage to the container under test when the second specified loop rule is applied for the i-th time. Indicates the second temperature standard deviation; This represents the average value of all second stresses after N cycles of the second specified rule. The second stress obtained after completing the i-th second specified cycle rule; This is the second cumulative fatigue damage; Indicates the hydrogen concentration in the detected gas; Indicates wall temperature; Indicates the second temperature coefficient; This represents the second stress correction factor; This represents the hydrogen embrittlement correction factor.

7. The method according to claim 1, characterized in that, The determination of total cumulative fatigue damage based on the first cumulative fatigue damage and the second cumulative fatigue damage includes: Based on the third preset strategy, the first final weight of the first cumulative fatigue damage and the second final weight of the second cumulative fatigue damage are obtained. Based on the first cumulative fatigue damage, the first final weight, the second cumulative fatigue damage, and the second final weight, the total cumulative fatigue damage is obtained; The third preset strategy is: Obtain the rated number of operations for the container under test; A third initial weight is obtained based on the first cumulative fatigue damage, and a fourth initial weight is obtained based on the second cumulative fatigue damage; Based on the third and fourth initial weights, the initial total cumulative fatigue damage is calculated. Based on the third initial weight and the rated number of runs, the first number is obtained, and based on the fourth initial weight and the rated number of runs, the second number is obtained; In the virtual scene, the first specified cycle rule for the first number of times and the second specified cycle rule for the second number of times are simulated on the virtual container to obtain the total virtual cumulative fatigue damage; When the difference between the virtual total cumulative fatigue damage and the initial total cumulative fatigue damage is greater than a preset difference, the third initial weight and the fourth initial weight are adjusted until the difference between the virtual total cumulative fatigue damage and the initial total cumulative fatigue damage is less than or equal to the preset difference, at which point the third initial weight becomes the first final weight and the fourth initial weight becomes the second final weight.

8. The method according to claim 1, characterized in that, The first optimal fluid parameters include a first optimal flow rate and a first optimal air intake angle, and the second optimal fluid parameters include a second optimal flow rate and a second optimal air intake angle; The process of obtaining the first optimal fluid parameters and the second optimal fluid parameters based on the first preset strategy includes: Based on user-issued instructions, a preset virtual scene is established in the specified finite element software, and a virtual container is built in the preset virtual scene. Obtain the range of fluid inlet angle and the range of fluid flow rate; Based on the range of fluid inlet angle and fluid velocity, the first optimal velocity, the first optimal inlet angle, the second optimal velocity, and the second optimal inlet angle are obtained according to the particle swarm optimization algorithm model. The discrete point values ​​of the fitness function of the particle swarm optimization algorithm model are obtained by performing finite element simulation in the virtual scene.

9. An electronic device, characterized in that, The electronic device includes a processor and a memory coupled together, the memory storing a computer program that, when executed by the processor, causes the electronic device to perform the method as described in any one of claims 1-8.

10. A computer storage medium, characterized in that, The computer storage medium stores a computer program that, when run on the computer, can perform the method as described in any one of claims 1-8.

Images