Optimization design method for disc spring special for ultrasonic motor

Abstract

The invention relates to the field of computer numerical calculation, in particular to an optimal design method of a special disc spring for an ultrasonic motor, which comprises the following steps of: firstly, in finite element software, performing parametric modeling on a disc spring stress model, setting according to related material attributes, and setting and constraining the characteristics of a contact surface; boundary conditions are set according to parameter characteristics of the disc spring, then grids are divided and submitted for finite element software calculation, and initial data are collected; and then modifying parameters, running a Python script to control finite element software to perform batch operation, and outputting a disc spring characteristic curve. And an optimization algorithm model combining a particle swarm optimization (PSO) and a convolutional neural network (CNN) is constructed, the optimization algorithm model realizes calculation design optimization through iterative interaction, and optimal structure parameters and characteristic curves are obtained. According to the method, the structure parameters of the disc spring are calculated, designed and optimized, a convenient and efficient calculation mode is provided for the ultrasonic motor disc spring, and the specific working condition requirements of the ultrasonic motor are met.

Description

Technical Field

[0001] The present invention relates to the field of computer numerical calculation, and particularly relates to an optimization design method for a special disc spring of an ultrasonic motor. Background Art

[0002] Based on the inverse piezoelectric effect, ultrasonic motors have advantages such as high precision, high response, and low-speed large-torque output, and are widely used in frontier fields such as precision machinery, intelligent robots, and high-end optical instruments. It relies on the friction between the stator and the rotor to convert the microscopic vibration in the ultrasonic frequency domain into the macroscopic rotation of the rotor. When assembling an ultrasonic motor, applying an appropriate pre-pressure is extremely crucial for ensuring frictional force transmission. The magnitude of the pre-pressure directly determines the frictional effect between the stator and the rotor, and thus affects the overall performance of the motor such as the rotational speed stability and torque output accuracy. However, during actual operation, the high-frequency vibration of the stator and rotor and the continuous consumption of the friction medium cause relative extrusion deformation and fine movement of the internal structure of the motor, resulting in pre-pressure fluctuations, and further making the output performance of the motor unstable. In severe cases, it affects the normal operation of the equipment.

[0003] The disc spring is the core component that provides the pre-tightening force for the ultrasonic motor, and its performance is related to the stability and service life of the motor operation. However, the traditional disc spring design relies on empirical formulas and is difficult to adapt to complex working conditions. This often leads to uneven distribution of the pre-tightening force, affecting the overall performance of the motor, and even shortening the maintenance cycle and service life of the ultrasonic motor. Therefore, when designing a special disc spring, it is necessary to determine the optimal pre-pressure between the stator and the rotor, and then design the structural parameters to make the quasi-zero stiffness of the disc spring match the pre-pressure, expand the quasi-zero stiffness interval, and stabilize the pre-tightening force.

[0004] Currently, there is an urgent need to comprehensively analyze and optimize the geometric parameters and material properties of the disc spring by means of simulation technology, improve its comprehensive performance and reliability, meet the specific working condition requirements of the ultrasonic motor, and ensure its efficient and stable operation. Summary of the Invention

[0005] The purpose of the present invention is to provide an optimization design method for a special disc spring of an ultrasonic motor, aiming to overcome the design dilemma of the traditional disc spring of the ultrasonic motor, and use advanced simulation calculation technology and intelligent optimization algorithms to construct a complete and efficient disc spring optimization system to achieve precise optimization of the structural parameters of the disc spring.

[0006] To achieve the above purpose, the present invention provides a simulation calculation optimization method for a special disc spring of an ultrasonic motor, including the following steps:

[0007] Step 1: In the finite element software, model the parametric structure of the disc spring, draw and generate the three-dimensional models of the disc spring and the pressure plate, and assemble the force model; [[ID=!27]]

[0008] Step 2: Set the reference point of the pressure plate, process and set the contact surface characteristics of the disc spring, confirm the mesh element properties, mesh the disc spring and the pressure plate, and apply displacement boundary conditions according to the disc spring parameters;

[0009] Step 3: Perform finite element software calculations, output the relevant numerical curves of the reference point, synthesize the quasi-zero stiffness curve information of the disc spring, obtain the initial structure parameters, and perform batch calculations;

[0010] Step 4: Starting from the initial structure parameters, set variable parameters and input them into the optimization algorithm model;

[0011] Step 5: The optimization algorithm model performs optimization calculations to obtain the optimal ratio of variable parameters and the precise influence parameter factors;

[0012] Step 6: After obtaining the precise parameter influence factor values, determine the optimization parameter target range, upgrade and adjust the disc spring mesh, reduce the historical output frequency of the reference point, and perform finite element software calculations again to finally output the optimal disc spring quasi-zero stiffness characteristic curve that fits the ultrasonic motor working conditions.

[0013] Optionally, the execution process of Step 1 is specifically to draw a sketch according to the ultrasonic motor based on the geometric characteristics of the disc spring, generate a 3D model, simultaneously model the pressure plate, perform stiffness treatment, construct a force model framework, and set the relevant model material parameters according to the material selection characteristics of the disc spring in the material parameter setting link.

[0014] Optionally, for the reference point set in advance in Step 2, output the motion parameter values of the upper and lower pressure plates. During the process of setting the displacement boundary conditions, fix the force model and apply the displacement conditions, so that the displacement value is in proportion to the disc spring height value, restoring the actual force scenario of the disc spring.

[0015] Optionally, in Step 3, start the finite element software program for calculation. The initial structure parameters include the inner diameter, outer diameter, height, thickness, and inner cone angle, and initially analyze the performance of the disc spring.

[0016] Optionally, during the execution of Step 4, determine the inner diameter of the disc spring as the reference parameter, list the outer diameter, height, thickness, and inner cone angle as variables, set the ratio range of relevant parameter factors, and use the py script to control the finite element software for batch operations to obtain a data set and input it into the optimization algorithm model.

[0017] Optionally, the optimization algorithm model is composed of a combination of a particle swarm optimization algorithm and a convolutional neural network algorithm. The particle swarm optimization algorithm is responsible for global search, quickly locating the parameter influencing factors and the preliminary optimal solution; the convolutional neural network algorithm accurately identifies the local feature differences between the simulation and the real curve through deep learning, and further calibrates the parameter details; the particle swarm optimization algorithm and the convolutional neural network algorithm interact through iteration, and finally output the optimization result.

[0018] The present invention provides an optimization method for simulating and calculating a special disc spring for an ultrasonic motor. First, a deformable entity is selected, a three-dimensional geometric model of the disc spring is established, and the material parameters of the disc spring are set; the discrete stiffness treatment is performed on the upper and lower rigid pressure plates of the disc spring; the force contact surface setting and constraint of the disc spring are carried out, and the displacement and boundary constraint conditions are applied; the mesh model is determined; the Python script is run for batch operation, and the characteristic curve of the disc spring is calculated and output. An optimization algorithm model combining a particle swarm optimization (PSO) algorithm and a convolutional neural network (CNN) is constructed. Among them, PSO globally searches to locate the parameter influencing factors and the preliminary optimal solution, and CNN uses deep learning to identify the curve differences and calibrate the parameter details. The two achieve simulation calculation optimization through iterative interaction, and obtain the optimal structural parameters and characteristic curves. The present invention uses finite element software for calculation, calculates and optimizes the structural parameters of the disc spring, provides a convenient and efficient calculation method for the disc spring of the ultrasonic motor, and meets the specific working condition requirements of the ultrasonic motor. Description of the Drawings

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 It is a schematic flow chart of the steps of an optimization design method for a special disc spring of an ultrasonic motor according to the present invention.

[0021] Figure 2 It is a schematic model diagram of the disc spring modeling in a specific embodiment of the present invention.

[0022] Figure 3 It is a nephogram of the disc spring model before force application in the disc spring modeling of a specific embodiment of the present invention.

[0023] Figure 4 It is a flow chart of an optimization algorithm model combining a particle swarm (PSO) and a convolutional neural network (CNN) algorithm according to the present invention.

[0024] Figure 5 It is a comparison chart of the characteristic curve of the optimized disc spring and the real characteristic curve according to the present invention. Specific Embodiments

[0025] Embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention, and should not be construed as limiting the present invention.

[0026] The present invention provides an optimized design method for a disc spring dedicated to an ultrasonic motor, including the following steps:

[0027] Step 1: In the finite element software, model the parametric structure of the disc spring, draw and generate the 3D models of the disc spring and the pressure plate, and assemble the force model;

[0028] Step 2: Set the reference point of the pressure plate, process and set the contact surface characteristics of the disc spring, confirm the mesh element properties, mesh the disc spring and the pressure plate, and apply displacement boundary conditions according to the disc spring parameters;

[0029] Step 3: Perform calculations in the finite element software, output the relevant numerical curves of the reference point, synthesize the quasi-zero stiffness curve information of the disc spring, obtain the initial structure parameters, modify the parameters, run the script, and perform batch calculations;

[0030] Step 4: Starting from the initial structure parameters, set the variable parameters and input the optimization algorithm model combining PSO and CNN;

[0031] Step 5: The optimization algorithm model performs optimization calculations to obtain the optimal ratio of the variable parameters and the accurate influencing parameter factors;

[0032] Step 6: After obtaining the accurate parameter influencing factor values, determine the optimization parameter target range, upgrade and adjust the disc spring mesh, and reduce the historical output frequency of the reference point. Then, perform calculations in the finite element software again, and finally output the optimal disc spring quasi-zero stiffness characteristic curve that fits the working conditions of the ultrasonic motor.

[0033] The specific implementation steps are as Figure 1 shown. [[ID=X]]

[0034] The following will be described in detail with specific embodiments (please refer to Figures 2 to 5 ):

[0035] The steps of Step 1 are as follows: Establish as Figure 2 and Figure 3For the shown three-dimensional model of the disc spring of the ultrasonic motor, first draw a sketch, including geometric parameters such as an inner diameter of 15.10 mm, an outer diameter of 51.80 mm, a height of 1.8 mm, and a thickness of 0.8 mm. Generate the three-dimensional geometric model of the disc spring through a rotation operation. The upper and lower pressure plates are also modeled and discretely stiffened, and processed into a shell structure. In the material parameter setting section, according to the material selection characteristics of the disc spring, set the elastic modulus to 186 GPa and the Poisson's ratio value to 0.3.

[0036] Step 2: Set the positioning reference points for the upper and lower rigid pressure plates, that is, the geometric center position of the force-bearing surface, and set the contact pair of the force-bearing surface of the disc spring. Select linear elements to divide the mesh, set the displacement boundary conditions, fix the lower pressure plate, and set a downward displacement for the upper pressure plate. The displacement value accounts for 70% of the height value of the disc spring, that is, 1.3 mm, to restore the actual force-bearing scenario of the disc spring. Submit the job for simulation calculation.

[0037] Step 3: Observe the calculation results and strain nephogram of the finite element software, and the deformation process and trend of the disc spring can be observed. Output the load-displacement curve of the reference points of the upper and lower pressure plates, and extract the quasi-zero stiffness curve information of the synthesized disc spring. Specifically, run a Python script for batch calculation to calculate and output the characteristic curve of the disc spring.

[0038] Output model parameters such as the outer diameter D, height H, thickness t, inner conical surface angle α, etc., and initially analyze the performance of the disc spring.

[0039] Step 4: Starting from the initial structural parameters, obtain the original quasi-zero stiffness curve and model parameters to anchor the benchmark for subsequent optimization. Combining with the actual working conditions of the ultrasonic motor, determine the inner diameter of the disc spring as the reference parameter, and list the outer diameter, height, thickness, and inner conical surface angle as variables. Set the ratio range of the height to the thickness, and use the batch calculation function of the finite element software script to input the calculation results into the optimization algorithm model.

[0040] Step 5: As Figure 4 shown, in the optimization algorithm model combining the particle swarm optimization algorithm (PSO) and the convolutional neural network algorithm (CNN), the particle swarm optimization algorithm (PSO) is used for efficient optimization to obtain the optimal ratio of the height to the thickness and the maximum influence parameter factor.

[0041] The specific implementation steps of the optimization algorithm are as follows;

[0042] PSO optimization parameter initialization:

[0043] Map the outer diameter D, height H, thickness t, and inner conical surface angle α of the disc spring obtained in step 4 to the particle position vector X = [D, H, t, α], and set the particle swarm size, search space dimension to four dimensions, and the number of iterations. Within the set range of disc spring parameters, randomly initialize the positions and velocities of the particles. Each particle represents a set of structural parameters of the disc spring (outer diameter, height, thickness, inner conical surface angle), and set the objective function as the sum of the squared errors between the quasi-zero stiffness curve obtained from the simulation calculation and the target curve.

[0044] PSO iterative optimization:

[0045] First, calculate the fitness: In each iteration, calculate its fitness (objective function value) according to the current position of the particle. Call the finite element software script to batch calculate the parameter combinations of each particle, output the disc spring characteristic curve, and calculate the error value between it and the target curve as the fitness.

[0046] Update the individual and global optima: Record the historical optimal solution (pbest) of each particle and the global optimal solution (gbest) of the group.

[0047] Then, update the particle state: According to the particle velocity update formula Adjust the particle position, generate a new parameter combination, and enter the next iteration.

[0048] Finally, perform the termination condition judgment: If the maximum number of iterations or the error threshold (such as Minmum Error ≤ 0.03) is reached and the target requirements are met, output the optimal parameter combination (such as Height = 2.30mm, Thickness = 1.76mm, Angle = 7.87°, MinmumError(PSO) = -0.0266); otherwise, continue the iteration.

[0049] CNN data generation and model training:

[0050] First, construct the dataset: After the py script controls the abaqus calculation, parse the.odb file, and use all the particle parameter combinations and their corresponding simulation characteristic curves obtained from the PSO optimization iteration as the input layer input, and the true disc spring quasi-zero stiffness curve as the label to construct the training dataset.

[0051] Then, build the CNN network model architecture: The intermediate layer design includes a deep learning network with a convolutional layer (extracting local features of the curve), a ReLU activation function (enhancing non-linearity), a pooling layer (dimensionality reduction), and a fully connected layer (parameter mapping). The output layer is the four-dimensional parameter adjustment amount ΔX = [ΔD, ΔH, Δt, Δα].

[0052] Finally, perform model training: Use the backpropagation algorithm to optimize the CNN weights and minimize the prediction error loss function (such as mean square error).

[0053] CNN iterative calibration parameters:

[0054] First, perform error prediction and parameter screening. Input the optimal parameters output by PSO into the trained CNN model, input the simulation curve corresponding to the PSO optimal parameters, the CNN outputs the adjustment amount ΔX, and perform parameter correction X cnn = X PSO + η·ΔX, where η is the learning rate (such as 0.1), predict the simulation curve error, and screen out the parameter combination with the smallest error. Then perform parameter fine-tuning. If the CNN prediction error does not meet the standard, reverse-adjust the disc spring parameters according to the gradient information, and take X cnn as the new PSO initial population, restart the loop, perform simulation calculations and iteratively update the CNN model until the error converges.

[0055] Step 6: Input the PSO optimal parameters, the CNN predicts the adjustment amounts ΔD = -0.1 mm and Δt = +0.03 mm, and the parameter error after correction drops to 0.018 (1.8%), meeting the target threshold. Using the accurate parameter influence factor values obtained from the optimization algorithm model, optimize the simulation calculation. First, adjust the disc spring mesh type from linear elements to quadratic elements (such as C3D20), then increase and refine the number of steps of the historical output target value, enhance the calculation accuracy and stability, and refine the displacement application process to ensure that the deformation simulation is more in line with the actual working conditions. Perform the finite element software calculation again, and finally output the optimal disc spring quasi-zero stiffness characteristic curve that fits the ultrasonic motor working conditions.

[0056] Furthermore, compare the optimized disc spring characteristic curve with the true characteristic curve to verify the optimization effect. Through comparison, the deformation of the disc spring is significantly larger, nearly reaching 0.002 m, and the error is still a bit large, but the zero stiffness values are basically the same; Figure 5 The two simulation calculation curves in

[0057] Compared with the prior art, the present invention has the following beneficial effects:

[0058] 1. Rapidly analyze the structural characteristics of the disc spring: Through the simulation optimization method, it is possible to rapidly analyze the structural characteristics of the disc spring dedicated to the ultrasonic motor, providing strong guarantee for the operation of the ultrasonic motor under specific working conditions.

[0059] 2. Improved Optimization Efficiency and Accuracy: This innovative approach integrates the particle swarm optimization (PSO) algorithm and the convolutional neural network (CNN). PSO is used to identify the core influencing factors of disc spring model parameters, while the CNN is used to accurately calibrate the deviation between simulation results and actual values. The two algorithms work together to iterate and significantly improve the efficiency and accuracy of disc spring optimization.

[0060] 3. Filling the technical gaps in traditional design methods: This method effectively fills the technical gaps in the application of traditional design methods in ultrasonic motor disc springs, and promotes the in-depth application and innovative development of ultrasonic motors in high-end manufacturing, intelligent equipment and other fields.

[0061] 4. Accurately optimize structural parameters: Through simulation calculations and intelligent optimization algorithms, the structural parameters of the disc spring can be accurately optimized to ensure that it matches the preload of the ultrasonic motor, expand the quasi-zero stiffness range, and stabilize the preload force.

[0062] In summary, the present invention proposes a disc spring optimization design method for ultrasonic motors. By combining a highly efficient optimization algorithm model with a particle swarm optimization (PSO) algorithm and a convolutional neural network (CNN) algorithm, the two algorithms work together and iteratively interact to achieve a combination of "coarse adjustment" and "fine calibration," significantly improving optimization efficiency and accuracy. This method achieves precise optimization of the disc spring structure for ultrasonic motors. This method not only improves optimization efficiency and accuracy but also fills a technical gap in traditional design methods, providing strong support for the application of ultrasonic motors in high-end manufacturing, intelligent equipment, and other fields.

[0063] The above disclosure is only one or more preferred embodiments of the present invention, and certainly cannot be used to limit the scope of the present invention. Those skilled in the art can understand that the above embodiments can be realized.

[0064] All or part of the processes of the embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. An optimization design method for a special disc spring of an ultrasonic motor, characterized in that, The following steps are involved: Step 1: In the finite element software, perform parametric structural modeling of the disc spring, draw and generate a three-dimensional model of the disc spring and pressure plate, and assemble the force model; Step 2: Set the reference point of the pressure plate, process and set the contact surface characteristics of the disc spring, confirm the mesh element properties, mesh the disc spring and pressure plate, and apply displacement boundary conditions based on the disc spring parameters; Step 3: Perform finite element software calculations, output relevant numerical curves of reference points, synthesize the quasi-zero stiffness curve information of the disc spring, obtain initial structural parameters, and perform batch calculations; Step 4: Taking the initial structural parameters as the starting point, set the variable parameters and input the optimization algorithm model combining PSO and CNN; Step 5: The optimization algorithm model performs optimization calculations to obtain the optimal ratio of variable parameters and accurate influencing parameter factors; Step 6: After obtaining the precise parameter influencing factor values, determine the target range of the optimized parameters, upgrade and adjust the disc spring grid, reduce the historical output frequency of the reference point, perform the finite element software calculation again, and finally output the optimal disc spring quasi-zero stiffness characteristic curve that matches the ultrasonic motor operating conditions.

2. The method for optimizing the design of disc springs for ultrasonic motors according to claim 1, wherein: The execution process of step 1 is to draw a sketch based on the ultrasonic motor and the geometric characteristics of the disc spring, generate a three-dimensional model, model the pressure plate, implement stiffness processing, build a force model framework, and set relevant model material parameters according to the material selection characteristics of the disc spring in the material parameter setting link.

3. The method for optimizing the design of disc springs for ultrasonic motors according to claim 2, wherein: The reference point set in advance in step 2 outputs the motion parameter values of the upper and lower pressure plates. In the process of setting the displacement boundary conditions, the force model is fixed and the displacement conditions are applied so that the displacement value is proportional to the disc spring height value, restoring the actual force scenario of the disc spring.

4. The method for optimizing the design of disc springs for ultrasonic motors according to claim 3, wherein: In step 3, the finite element software calculation program is started to perform simulation calculations. The initial structural parameters include inner diameter, outer diameter, height, thickness and inner cone angle, and the performance of the disc spring is preliminarily analyzed.

5. The method for optimizing the design of disc springs for ultrasonic motors according to claim 4, wherein: During the execution of step 4, the inner diameter of the disc spring is determined as the reference parameter, the outer diameter, height, thickness, and inner cone angle are listed as variables, the ratio range of relevant parameter factors is set, and the finite element software is controlled by py script to perform batch calculations to obtain a data set and input it into the optimization algorithm model.

6. The method for optimizing the design of disc springs for ultrasonic motors according to claim 5, wherein: The optimization algorithm model is based on the combination of particle swarm optimization and convolutional neural network algorithm. The particle swarm optimization is responsible for global search, quickly locating parameter influencing factors and preliminary optimal solutions; the convolutional neural network algorithm accurately identifies the local feature differences between simulation and real curves through deep learning, and further calibrates parameter details; the particle swarm optimization and convolutional neural network algorithm interact iteratively and finally output the optimization results.

Images