A method for optimizing the flange structure of a tire dismounting machine using a multi-objective genetic algorithm
Through the multi-objective genetic algorithm, the flange structure of the tire unloader is optimized, which solves the problem of unreasonable design of the existing flange structure, and achieves a more balanced force and a longer service life under complex loads.
Patent Information
- Application Number
- CN202111369742.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-11-16
- Publication Date
- 2025-05-30
- Estimated Expiration
- 2041-11-16
AI Technical Summary
The flange structure design of the existing tire unloader is unreasonable, resulting in uneven stress under complex loads and reducing service life.
The flange structure is optimized by multi-objective genetic algorithm, and the size parameters and structural design of the flange are optimized by establishing virtual prototypes, simulation analysis, Bp neural network prediction and direct optimization methods.
The optimized flange structure has more reasonable stress under complex loads, significantly improved fatigue strength, extended service life and lighter mass.
Smart Images
Figure CN114492155B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of optimizing the flange structure of a tire unloading machine, and particularly to a method for optimizing the flange structure of a tire unloading machine by using a multi-objective genetic algorithm. Background Art
[0002] With the rapid development of large-scale mining machinery in China, the output of large-scale mining machinery tires has also been continuously increasing. The tire unloading machine involved in the present invention is a mechanical device specifically used for specific mining tires, and the weight of the transported object for unloading is 6 tons. During the tire production process, the tire unloading machine is required to clamp the tire and perform a 90° flip. During the flipping process, the huge inertial force and collision caused by the huge weight of the tire will cause strong vibrations in the system. The flange, as a connecting part on the tire unloading machine, is responsible for connecting the driving device and the entire tire unloading device. In addition to bearing the flipping torque brought by the driving device and the weight of the tire unloading system, it also needs to bear the random loads generated during the contact and flipping process between the tire unloading machine and the tire. The overall stress situation is relatively complex. Unreasonable flange structure design often leads to unreasonable stress on the flange and reduces its service life. Therefore, reasonable structure design is very important for giving full play to the mechanical properties of the flange material and improving its fatigue strength. Summary of the Invention
[0003] In order to overcome the deficiencies in the prior art, the purpose of the present invention is to provide a method for optimizing the flange structure of a tire unloading machine by using a multi-objective genetic algorithm, so that the structure of the flange is more reasonable in stress, has a longer service life, and at the same time, the optimal flange structure is obtained.
[0004] To achieve the above purpose, the present invention adopts the following technical solutions:
[0005] A method for optimizing the flange structure of a tire unloading machine by using a multi-objective genetic algorithm, characterized in that it includes the following steps:
[0006] 1) Establish a virtual prototype of the overall tire unloading machine, perform simulations according to the actual working conditions, and obtain the load time history of the connecting flange at the fixed ring rotation;
[0007] 2) Establish a three-dimensional model of the flange, import the three-dimensional model of the flange into Ansys, set the material parameters of the flange, apply unit loads to the three-dimensional model respectively according to different working conditions, obtain the stress results under different working condition unit loads, combine the stress nephograms under different working conditions with the dynamic load spectra in each direction, and use the quasi-static method to convert the force load spectrum into a stress history spectrum; use the Ansys nCode fatigue analysis software to predict the service life of the flange in combination with the stress history data;
[0008] 3) Change the size of the flange, and simulate and predict the life of the flange under different sizes;
[0009] 4) Take the life structure and each dimension parameter predicted by the above simulation as the training and test data of the Bp neural network, and use matlab to construct a Bp neural network life prediction model;
[0010] 5) On the basis of the optimized neural network, use the multi-objective genetic algorithm to optimize the structural parameters of the flange. Take the dimension parameters as individuals, and the predicted flange life and mass under different dimensions by the neural network as the objective functions, so that the flange has the longest life and the smallest mass under the optimal parameters;
[0011] 6) Use the direct optimization method in Ansys for multi-objective optimization to obtain the optimal dimensions, obtain the optimal structure of the flange, and establish a three-dimensional model of the optimal structure of the flange.
[0012] Furthermore, in step 4), take the weights and thresholds as individuals, and the reciprocal of the sum of the squared errors between the predicted value and the simulation value as the fitness function. The greater the fitness of the individual, the better the individual value. Optimize the weights and thresholds of the neural network with the individual with the maximum fitness as the optimal individual to improve the accuracy of the neural network prediction.
[0013] The present invention adopts the above technical solutions and has the following beneficial technical effects:
[0014] The present invention uses intelligent algorithms and the screening method to optimize the structure of the connecting flange, and the optimization method is simple and reliable. The optimized flange structure is lighter in mass than the original flange structure, more reasonable in force under complex loads, and has a higher fatigue strength improvement than before, filling the gap in the current research on the flange of the giant tire changer. Brief Description of the Drawings
[0015] The following further describes the present invention in detail in conjunction with the drawings and specific embodiments;
[0016] Figure 1 It is a schematic diagram of the virtual prototype of the tire changer;
[0017] Figure 2 It is a schematic diagram of the three-direction force and resultant force curves at the flange connection of the tire changer;
[0018] Figure 3 It is a schematic diagram of the three-direction moment and resultant moment curves at the flange connection of the tire changer;
[0019] Figure 4 It is a schematic diagram of the three-dimensional model of the flange;
[0020] Figure 5 It is a schematic diagram of the load application position of the flange;
[0021] Figure 6 It is a stress nephogram under the unit force load in the x direction;
[0022] Figure 7 It is the stress nephogram under the unit moment load in the y direction;
[0023] Figure 8 It is the stress history diagram of the maximum node under the force conditions in different directions;
[0024] Figure 9 It is the stress history diagram of the maximum node under the bending moment conditions in different directions;
[0025] Figure 10 It is the flange life nephogram;
[0026] Figure 11 It is the schematic diagram of the flange life under different parameter combinations;
[0027] Figure 12 It is the schematic diagram of the flange mass under different parameter combinations;
[0028] Figure 13 It is the bar chart of the sensitivity coefficients of each parameter;
[0029] Figure 14 It is the trade-off schematic diagram of each parameter against the target value;
[0030] Figure 15 It is the curve diagram of the fitness value changing with the number of generations;
[0031] Figure 16 It is the curve diagram of the sum of squared errors changing;
[0032] Figure 17 It is the schematic diagram of the fitting degree of the neural network training data;
[0033] Figure 18 It is the schematic diagram of the life prediction fitting curve;
[0034] Figure 19 It is the fitting diagram of the flange mass prediction;
[0035] Figure 20 It is the distribution diagram of the first front individuals;
[0036] Figure 21 It is the schematic diagram of the optimal flange structure. Specific implementation manner
[0037] The tire unloading machine is a mechanical device specifically used for certain mine tires, and the transported object weighs 6 tons. During the tire production process, the tire unloading machine is required to clamp the tire and perform a 90° flip. During the flipping process, the huge inertial force and collision caused by the huge weight of the tire will cause strong vibrations in the system. The flange, as a connecting part on the tire unloading machine, is responsible for connecting the driving device and the entire tire unloading device. In addition to bearing the flipping torque brought by the driving device and the weight of the tire unloading system, it also needs to bear the random loads generated during the contact and flipping of the tire unloading machine and the tire, and the overall stress situation is relatively complex. Unreasonable flange structure design often makes the stress of the flange unreasonable and reduces its service life. Therefore, reasonable structure design is very important for giving full play to the mechanical properties of the flange material and improving its fatigue strength.
[0038] A method for optimizing the flange structure of a tire unloading machine using a multi-objective genetic algorithm includes the following steps:
[0039] The first step is to establish a virtual prototype of the overall tire unloading machine. The virtual prototype is as Figure 1 shown. According to the actual working conditions, perform simulations to obtain the load time history of the connecting flange at the fixed ring rotation position. Among them, the three-directional forces and three-directional torques at the fixed ring rotation position are as Figure 2 , Figure 3 shown.
[0040] The second step is to establish a three-dimensional model of the flange according to the flange size. The three-dimensional model is as Figure 4 shown.
[0041] Import the three-dimensional model of the flange into Ansys, set the material parameters of the flange. The material is ordinary carbon structural steel. Apply unit loads to the three-dimensional model according to the working conditions. Since the moment in the x direction is small and its influence on the flange life can be ignored, there are a total of five working conditions, namely the forces in three directions and the torques in two directions. The loading positions are as Figure 5 shown.
[0042] Obtain the stress results under unit loads in different working conditions. Some of the results are as Figures 6 - 7 shown.
[0043] According to the stress nephograms under different working conditions and the dynamic load spectra in each direction, use the quasi-static method to convert the force load spectrum into a stress history spectrum, as Figures 8 - 9 shown.
[0044] Use Ansys nCode fatigue analysis software to predict the flange life in combination with the stress history data. Set the survival rate to 90%. The life nephogram is as Figure 10 shown. According to the nephogram, the total number of cycles is 30,830 times, which is converted to a duration of 17 hours.
[0045] In the third step, change the size of the flange and simulate and predict the flange life at different sizes.
[0046] The range of flange size parameters is shown in Table 1 below.
[0047] Table 1 Range of flange size selection
[0048]
[0049] Simulate the flange under different size parameters and predict the flange life under each parameter combination. There are 190 groups in total, and the flange life and quality under each parameter combination are as Figure 11 、 12 shown.
[0050] The influence of each parameter on the flange life and quality is as Figures 13 - 14 shown.
[0051] It can be seen from the sensitivity coefficient histogram that among them, the thickness of the taper neck has a greater impact on the life, followed by the flange thickness, and the height of the large end of the flange has a negligible impact on the life. The flange thickness has the greatest impact on the flange quality, and the influence of the taper neck thickness and height and the large end height on the quality decreases successively. According to the simulation results, the stressed part is mainly the taper neck part, which bears the largest tensile, compressive and shear stresses, and its material thickness directly affects the flange life, indicating that the simulation results are relatively reasonable.
[0052] In the fourth step, take the flange life structure and each size parameter predicted by the above simulation as the training and test data of the Bp neural network, and use matlab to construct a Bp neural network life prediction model. The initial thresholds and weights of the Bp neural network are generally randomly assigned, and the initial weights and thresholds determine the prediction accuracy of the neural network. The present invention combines two intelligent algorithms of genetic algorithm and Bp neural network, uses the optimization ability of the genetic algorithm, takes the weights and thresholds as individuals, and takes the reciprocal of the sum of the squared errors between the predicted value and the simulated value as the fitness function. The greater the fitness of an individual, the better the individual value. Optimize the weights and thresholds of the neural network with the individual with the maximum fitness as the optimal individual to improve the prediction accuracy of the neural network. The population size is set to 200, the genetic iteration times are set to 100 generations, and the change of the genetic algorithm fitness is as Figure 15 shown, and the curve of the sum of the squared errors between the predicted value and the simulated value changing with the number of generations is as Figure 16 shown.
[0053] The optimal threshold and weights obtained from the optimization are used as the initial neural network parameters of the neural network. The 190 groups of simulation data are divided into two groups. 170 groups are used as the training data of the neural network. The four size parameters are used as the input data, and the life and quality are used as the output data. A three-layer BP neural network is adopted, the number of hidden layer nodes is set to 9, the training is iterated 3000 times, the learning rate is set to 0.001, and the gradient descent method is used for network training. The fitting degree of the training data is as Figure 17 shown, the fitting degree R = 0.98477, and the last 20 groups are used as test samples. Finally, the construction of the neural network is completed, and the prediction results of the test samples are as Figure 18 , 19 shown.
[0054] Step 5, on the basis of the optimized neural network completed above, use the multi-objective genetic algorithm to optimize the structural parameters of the flange. The size parameters are used as individuals, and the predicted flange life and quality at different sizes by the neural network are used as the objective functions, so that the flange has the longest life and the smallest quality under the optimal parameters. Set the population size to 300, the number of evolutionary generations to 200, the stop generation to 200, and the optimal front coefficient to 0.02. The prediction results are as Figure 20 shown.
[0055] List the first three solutions with the maximum life in Table 2 below.
[0056] Table 2 Optimal front points of the multi-objective genetic algorithm
[0057]
[0058] Step 6, use the direct optimization method in Ansys for multi-objective optimization, and select the optimal parameter points for the 190 groups of simulation data. The results are shown in Table 3 below.
[0059] Table 3 Optimal parameter selection points by the Screening method
[0060]
[0061] Comparing the results of the two methods, the optimal size parameters are relatively close, indicating that the optimization method of the present invention has a certain rationality.
[0062] Comparing the two results, rounding the size values, the optimal size is shown in Table 4.
[0063] Table 4 Optimal size
[0064]
[0065] According to the optimal size, the optimal flange structure is obtained, and a three-dimensional model of the flange is established. The optimal flange structure is as Figure 21 shown.
[0066] The implementation of the present invention has been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific implementation manners. The above specific implementation manners are illustrative rather than restrictive of the present invention. Those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and the description of the present invention.
Claims
1. A method for optimizing the structure of the flange of a tire dismounting machine using a multi-objective genetic algorithm, Characterized in that: It includes the following steps: 1) Establish a virtual prototype of the overall tire dismounting machine, perform simulations according to the actual working conditions, and obtain the load time history of the connecting flange at the fixed ring rotation; 2) Establish a three-dimensional model of the flange, import the three-dimensional model of the flange into Ansys, set the material parameters of the flange, apply unit loads to the three-dimensional model according to different working conditions respectively, obtain the stress results under different working conditions of unit loads, combine the stress nephograms under different working conditions with the dynamic load spectra in each direction, and use the quasi-static method to convert the force load spectrum into a stress history spectrum; use Ansys nCode fatigue analysis software to predict the service life of the flange in combination with the stress history data; 3) Change the size of the flange and simulate and predict the flange life at different sizes; 4) Use the flange life and each size parameter predicted by the above simulations as the training and test data of the Bp neural network, and use matlab to construct a Bp neural network life prediction model; in step 4), use the weights and thresholds as individuals, and the reciprocal of the sum of the squared errors between the predicted value and the simulated value as the fitness function. The greater the fitness of the individual, the better the individual value. Optimize the weights and thresholds of the neural network with the individual with the maximum fitness as the optimal individual to improve the accuracy of neural network prediction; 5) On the basis of the optimized neural network, use a multi-objective genetic algorithm to optimize the structural parameters of the flange. Use the size parameters as individuals, and the flange life and mass predicted by the neural network at different sizes as the objective functions, so that the flange has the longest life and the smallest mass under the optimal parameters; 6) Compare the two optimization results of the multi-objective genetic algorithm and the direct optimization method in Ansys, round up to obtain the optimal size parameters, thereby obtaining the optimal structure of the flange, and establish a three-dimensional model of the optimal structure of the flange.