Optimization design method for prepressing process of aircraft engine rotor bolt

By training the agent model in Comsol simulation software, the bolt prepressing process simulation app is generated, and the problem of difficult to determine the prepressing process parameters of the aero engine rotor bolt is solved, efficient bolt prepressing process parameters are achieved, and assembly efficiency and quality are improved. .

CN118965983APending Publication Date: 2024-11-15DALIAN UNIV OF TECH
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411019130.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The pre-pressing process parameters of the aero engine rotor bolt (pre-pressing window period, first tightening position, tightening step length and tightening force) are difficult to determine, resulting in low assembly efficiency and quality, and lack of research on the rotor structure of domestic core engines.

Method used

Use Comsol simulation software to create and train the agent model to generate a bolt pre-pressing process simulation app. By inputting the stop interference amount, surface morphology and temperature distribution data, the optimized pre-pressing process parameters are output, including the pre-pressing window period, The pre-pressure position, tightening force and tightening step length of a piece can realize the prediction of parameters and the formulation of process regulations.

Benefits of technology

The Aeroengine rotor assembly efficiency and qualification rate are improved, operations are simplified, repeated assembly and adjustment are reduced, assembly quality and personnel labor efficiency are enhanced, and the agent model is continuously corrected through testing to improve subsequent assembly efficiency and quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118965983A_ABST
    Figure CN118965983A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of aero-engine production, and discloses an aero-engine rotor bolt pre-pressing process optimization design method which comprises the following steps: establishing an aero-engine rotor parameterized three-dimensional model; generating a bolt pre-pressing process agent model; the value range of each input parameter is determined according to the actual working condition and the assembly condition, and preparation is made for training of a subsequent agent model; collecting input parameters in the value range; training an agent model on the basis of collecting the input parameters, and determining the number of times of model training according to the number of sampling points; industrial software used for guiding the pre-pressing process operation is generated; and measuring each input parameter data, and importing the data into industrial software. According to the method, on the basis of the requirements of actual assembly working conditions, optimized bolt pre-pressing assembly process parameters are input and output according to specific rotor parameters, the problem that pre-pressing process steps are not clear and clear in the assembly process is solved, and a basis is provided for design of aircraft engine rotor pre-pressing process procedures.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of aero-engine production and relates to an aero-engine rotor bolt pre-pressing process optimization design method. Background Art

[0002] As the heart of the aircraft, the aircraft engine is known as the "crown jewel of industry" and is a significant symbol of the country's advanced manufacturing level. The level of rotor assembly technology has a close impact on the performance of the whole machine, and is also an important way to reduce production costs and improve production efficiency. The core engine rotor is mainly composed of the compressor rotor and the turbine rotor. The multi-stage rotors are connected by a drum-disc structure and are an important part of the engine power system. The core engine needs compressed air to complete energy conversion at high temperature to provide power to the engine. Its purpose determines the extreme working conditions that the core engine rotor needs to withstand. Bolt connection is a typical connection method commonly used in the multi-stage assembly of aero-engine rotors. When the circumferential connection stiffness of the installation edge is uneven, the high-speed rotation of the rotor can easily lead to a sudden increase in the vibration of the whole machine, which will pose a great threat to the safe operation of the engine.

[0003] The multi-stage rotors of aircraft engines are often connected by a stop-bolt structure with interference fit. Unlike conventional bolt connection structures, the interference fit will have a huge impact on the assembly effect of the bolts. According to the "Overview and Prospect of Key Assembly Technologies for Aircraft Engines" written by Zhao Gang et al., the typical processes of aircraft engine assembly include bolt tightening process, rotating body / near rotating body stacking assembly process, balancing process, temperature difference assembly process, etc., but there is a lack of description of the pre-pressing assembly process.

[0004] The preloading process is an assembly process between the circumferential positioning installation of the rotor assembly and the tightening of the bolt group. Its function is to apply a certain preloading pressure to the rotor to ensure that the two-stage rotor is in a seated state during the process of the interference fit temperature difference assembly of the stopper to restore to room temperature, give full play to the centering effect of the interference fit, and improve the end face fit of the rotor installation edge and the uniformity of the circumferential contact stress of the stopper. The ultimate goal of the preloading process is to reduce the dispersion of the preload force of the bolt group, improve the rotor assembly quality and the qualified rate of one-time assembly. According to Li Xiaodong's "Research on the Stop-Bolt Assembly Process of Aero-Engine Rotors", a special pressure equalization equipment was designed. Since the core engine of the aero-engine rotor is connected by multiple stages of rotors, and the stop connection position between the rotors is in a deep cavity structure, it is difficult to compress the contact interface by pressurizing the rotor as a whole. Therefore, it is chosen to apply a clamping force to the rotor stop installation edge during the temperature recovery process by bolt preloading. The main parameters of the bolt preloading process are: preloading window period, first tightening position, tightening step length, and tightening force. The preloading window period refers to the time between the temperature difference assembly process and the first bolt preloading, and the tightening step length refers to the time required to tighten each bolt. In China, rotors are generally preloaded in the form of preloading bolts. Unlike other mature aero-engine core engine assemblies, there are few studies on the preloading process of domestic core engine rotor structures, and the mechanism of the influence of the stop fit after temperature difference assembly on the bolt preloading process is unclear. Summary of the invention

[0005] In view of the current immaturity of the prestressing process and the difficulty in determining the bolt prestressing process parameters including the prestressing window period, the first tightening position, the tightening step and the tightening force, the present invention proposes an optimization design method for the prestressing process of aircraft engine rotor bolts. This method is based on Comsol to create and train an agent model to generate a simulation App suitable for the bolt prestressing process, so that it can input a set of parameters (interference, surface morphology, temperature distribution data) to obtain a set of prestressing process parameters (prestressing window period, the first prestressing position, tightening force, tightening step) output. A new and easy-to-operate method is provided for the prediction of aircraft engine rotor bolt prestressing process parameters and the formulation of process regulations. The method has the characteristics of simple operation, controllable all process parameters and high accuracy.

[0006] The technical solution of the present invention:

[0007] An optimization design method for aeroengine rotor bolt preloading process comprises the following steps:

[0008] Step 1: Create a parametric 3D model of the engine rotor

[0009] 1.1. Preliminarily establish a 3D model of the aero-engine rotor in the 3D modeling software;

[0010] 1.2. Import the 3D model into the simulation software and perform parametric modeling on the geometric parameters (interference of the stopper and surface morphology) to obtain the parametric 3D model of the aero-engine rotor;

[0011] Step 2: Define the proxy model for the bolt preloading process

[0012] 2.1. Set the materials, physical fields, and meshes of the imported model in Comsol simulation software to ensure high fidelity of the simulation results;

[0013] 2.2. Add the "Proxy Model Training" node in the "Research" of the Comsol simulation software, and define the stop interference, surface morphology, and temperature distribution as the input of the bolt preloading process proxy model, and the preloading window period, the first tightening bolt position, the bolt tightening force, and the bolt tightening step length as the output of the bolt preloading process proxy model;

[0014] Step 3: Determine the value range of each parameter in the bolt preloading process proxy model input according to the actual working conditions;

[0015] Step 4: Collect the parameters of the bolt preloading process proxy model input within the value range

[0016] 4.1. Determine the number of sampling points according to actual working conditions and accuracy requirements;

[0017] 4.2. Use the Latin hypercube method to collect the parameters of the bolt preloading process proxy model input within the value range;

[0018] Step 5: Based on the collected parameters of the bolt preloading process proxy model input, the deep neural network (DNN) module of Comsol is used to train the bolt preloading process proxy model, and the number of bolt preloading process proxy model training times is determined according to the number of sampling points;

[0019] Step 6: Encapsulate the trained proxy model into the Comsol simulation App module to facilitate subsequent calls and updates of the proxy model;

[0020] Step 7: Measure each input parameter data and import the input parameter data into the Comsol simulation App module generated in step 6;

[0021] 7.1. Measure the interference data and surface morphology data of the actual aircraft engine stop;

[0022] 7.2. When using the temperature difference method for interference fit, evenly select points in the radial direction of the containing part, evenly select points in the axial direction and on the mounting edge of the contained part, measure the temperature data of the above points, and use digital twin technology to generate virtual temperature distribution data;

[0023] 7.3. Import the input data obtained from the above measurements into the simulation App module generated in step 6 for parameter optimization;

[0024] Step 8: Implement the bolt preloading process according to the process parameters output by the software;

[0025] Step 9: Test the bolt preload results;

[0026] 9.1. If the preloading result meets the requirements, proceed to the subsequent tightening process;

[0027] 9.2. If the preloading result does not meet the requirements, the bolt preloading process proxy model shall be corrected according to the test data, and the temperature difference assembly and bolt preloading process shall be carried out again.

[0028] Beneficial effects of the present invention: The bolt preloading process provided by the present invention can effectively improve the assembly efficiency and qualified rate of aeroengine rotors. The method can output optimized bolt preloading process parameters based on the provided input parameters, avoid repeated assembly and adjustment, and reduce the workload of operators. At the same time, after assembly, the proxy model can be continuously corrected according to the test results to improve the subsequent assembly efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A flow chart of the aero-engine rotor bolt pre-stressing process optimization design method provided by the present invention. DETAILED DESCRIPTION

[0030] The specific implementation of the present invention is further described below in conjunction with the accompanying drawings and technical solutions.

[0031] A method for optimizing the design of the preloading process of rotor bolts, such as Figure 1 As shown, the specific steps include:

[0032] Step 1: Establish a parametric 3D model of the aero-engine rotor;

[0033] 1.1. Create a 3D model of the relevant parts of the aero-engine rotor stop in the 3D modeling software, and import the model into the geometry module of Comsol;

[0034] 1.2. Use offset surface, custom function modeling and other methods to parametrically model the stop part of the aero-engine rotor 3D model, and define control parameters for subsequent parameter sampling;

[0035] 1.3. Generate an assembly and define the contact pairs of each contact surface of the stopper;

[0036] 1.4. Define the material properties of the 3D model for subsequent simulation;

[0037] Step 2: Define the bolt preloading process proxy model;

[0038] 2.1. Establish boundary conditions in the solid heat transfer and solid mechanics modules according to the actual assembly conditions of the aero-engine rotor, and perform parameterization processing on the flange cylindrical contact, surface temperature distribution, etc.

[0039] 2.2. To ensure the accuracy of the model, the three-dimensional model of the aero-engine rotor is divided into hexahedral meshes;

[0040] 2.3. Establish a deep neural network model in the global definition node, define input parameters and output parameters, and set the learning rate, loss function, number of epochs, etc.;

[0041] Step 3: Determine the value range of each input parameter according to the actual working conditions and assembly conditions to prepare for the subsequent training of the agent model;

[0042] Step 4: Train the proxy model;

[0043] 4.1. Establish agent model training under the research node and select the parameters that need to be input and the parameters that need to be output;

[0044] 4.2. Import the parameter range determined in step 3 into the input parameter description;

[0045] Step 5: Determine the input parameter collection method;

[0046] 5.1. Determine the number of input points and define the random seed type based on the actual computing power and accuracy requirements;

[0047] 5.2. Use the Latin hypercube method to collect input parameters within the value range;

[0048] Step 6: Encapsulate the trained proxy model into the Comsol simulation App module;

[0049] 6.1. Click App Developer to create an application and set the layout;

[0050] 6.2. Select each input parameter in the global definition node and select the result preview to be displayed;

[0051] 6.3. Define the output method of process parameters and finally test and debug the application;

[0052] Step 7: Measure the input parameter data and import the data into the simulation app generated in step 6;

[0053] 7.1. Use a three-coordinate measuring machine to measure and record the interference of the stop cylindrical surface of the aero-engine rotor, and the geometrical morphology of the cylindrical surface and the end surface;

[0054] 7.2. Use temperature sensors to measure the temperature data of specific points on the rotor, and use digital twin technology to generate virtual and real-time temperature distribution data;

[0055] 7.3. Import the above three types of data into the simulation app for analysis;

[0056] 7.4 Based on the input data, the simulation app will output preloading process parameters including preloading window period, first bolt position, tightening force and tightening step length according to the built-in proxy model to guide the subsequent process;

[0057] Step 8: Perform a pre-pressing process for the aero-engine rotor bolts according to the output process parameters;

[0058] Step 9: Test the preloading results and provide feedback;

[0059] 9.1. After the bolt preloading process is completed and the rotor returns to room temperature, the bolt preloading results are tested and analyzed. The analysis mainly includes end face clearance, interface lamination uniformity, installation axis deflection, etc.;

[0060] 9.2. If the preloading result meets the process requirements, the subsequent bolt tightening process will be carried out on this basis;

[0061] 9.2. If the preloading result does not meet the process requirements, the proxy model shall be appropriately modified according to the test data, and the temperature difference assembly and bolt preloading process shall be re-performed on the basis of the modified proxy model, and the above operations shall be repeated.

[0062] This method can output a set of optimized bolt preloading process parameters through the input interference, surface morphology, and temperature distribution data. Because it is an optimization method based on the proxy model, the output result is very efficient and very suitable for the actual assembly scenario. The bolt preloading process guided by the software can effectively optimize the assembly results, thereby improving the rotor assembly quality.

Claims

1. A method for optimizing the design of aeroengine rotor bolt preloading process, characterized in that it comprises the following steps: Step 1: Create a parametric 3D model of the engine rotor 1.

1. Preliminarily establish a 3D model of the aero-engine rotor in the 3D modeling software; 1.

2. Import the 3D model into the simulation software and perform parametric modeling on the geometric parameters to obtain a parametric 3D model of the aero-engine rotor; Geometric parameters include stop interference and surface topography; Step 2: Define the bolt preloading process proxy model 2.

1. Set the materials, physical fields, and meshes of the imported model in Comsol simulation software to ensure high fidelity of the simulation results; 2.

2. Add the "Proxy Model Training" node in the "Research" of the Comsol simulation software, and define the stop interference, surface morphology, and temperature distribution as the input of the bolt preloading process proxy model, and the preloading window period, the first tightening bolt position, the bolt tightening force, and the bolt tightening step length as the output of the bolt preloading process proxy model; Step 3: Determine the value range of each parameter in the bolt preloading process proxy model input according to the actual working conditions; Step 4: Collect the parameters of the bolt preloading process proxy model input within the value range 4.

1. Determine the number of sampling points according to actual working conditions and accuracy requirements; 4.

2. Use the Latin hypercube method to collect the parameters of the bolt preloading process proxy model input within the value range; Step 5: Based on the collected parameters of the bolt preloading process proxy model input, the deep neural network module of Comsol is used to train the bolt preloading process proxy model, and the number of training times of the bolt preloading process proxy model is determined according to the number of sampling points; Step 6: Encapsulate the trained proxy model into the Comsol simulation App module to facilitate subsequent calls and updates of the proxy model; Step 7: Measure each input parameter data and import the input parameter data into the Comsol simulation App module generated in step 6; 7.

1. Measure the interference data and surface morphology data of the actual aircraft engine stop; 7.

2. When using the temperature difference method for interference fit, evenly select points in the radial direction of the containing part, evenly select points in the axial direction and on the mounting edge of the contained part, measure the temperature data of the above points, and use digital twin technology to generate virtual temperature distribution data; 7.

3. Import the input data obtained from the above measurements into the simulation App module generated in step 6 for parameter optimization; Step 8: Implement the bolt preloading process according to the process parameters output by the software; Step 9: Test the bolt preload results; 9.

1. If the preloading result meets the requirements, proceed to the subsequent tightening process; 9.

2. If the preloading result does not meet the requirements, the bolt preloading process proxy model shall be corrected according to the test data, and the temperature difference assembly and bolt preloading process shall be carried out again.

Citation Information

Cited By

  • Preparation process method of rare earth giant magnetostrictive transducer with double-rod structure

    CN119772517A