CN-121988861-A - Inertia friction welding process control method, control system and inertia friction welding machine

CN121988861ACN 121988861 ACN121988861 ACN 121988861ACN-121988861-A

Abstract

A control method, a control system and an inertia friction welding machine for an inertia friction welding process. The control method comprises the steps of S1, S2, S3 and S4. The method comprises the steps of S1, setting welding control parameters, S2, welding by the aid of the welding control parameters in the step S1, collecting welding process parameters and actual deformation parameters of a welding joint in the welding process, S3, predicting and obtaining predicted deformation parameters of the welding joint by using a cyclic neural network model, wherein input parameters of the cyclic neural network model comprise the welding process parameters in the step S2, and the hidden layer of the cyclic neural network model is obtained by carrying out iterative optimization with the aim of reducing errors between the predicted deformation parameters and the actual deformation parameters in the step S2, so that the cyclic neural network model is obtained, S4, predicting and obtaining the predicted deformation parameters by using the cyclic neural network model in the step S3, and executing the step S1 with the aim of reducing errors between the predicted deformation parameters and the design deformation parameters.

Inventors

Liao Zhongxiang
ZHANG XUAN
TAN BO
HAN XIUFENG
LI JINSHENG

Assignees

中国航发商用航空发动机有限责任公司

Dates

Publication Date: 20260508
Application Date: 20241105

Claims (12)

1. A method of controlling an inertia friction welding process, comprising: s1, setting welding control parameters; S2, welding is carried out according to the welding control parameters in the step S1, and welding process parameters and actual deformation parameters of a welding joint in the welding process are collected; s3, predicting and obtaining predicted deformation parameters of the welding joint by using a cyclic neural network model, wherein the input parameters of the cyclic neural network model comprise the welding process parameters of the step S2, iteratively optimizing a hidden layer of the cyclic neural network model with the aim of reducing errors between the predicted deformation parameters and the actual deformation parameters of the step S2 to obtain the hidden layer, thereby obtaining the cyclic neural network model, and And S4, predicting and obtaining the predicted deformation parameters by using the cyclic neural network model in the step S3, and executing the step S1 with the aim of reducing errors between the predicted deformation parameters and design deformation parameters.
2. The inertia friction welding process control method according to claim 1, characterized in that the step S3 further comprises: setting the activation function of the hidden layer as a nonlinear activation function.
3. The inertia friction welding process control method according to claim 1, characterized in that the step S3 further comprises: setting a mean square loss function to express the error between the predicted deformation parameter and the actual deformation parameter.
4. The inertia friction welding process control method according to claim 1, wherein the iterative optimization of the hidden layer of the step S3 includes: The parameters of the hidden layer are iteratively optimized using a gradient descent method.
5. The inertia friction welding process control method according to claim 1, wherein the iterative optimization of the hidden layer of the step S3 includes: Parameters of the hidden layer are iteratively optimized using newton's method or quasi-newton's method.
6. The inertia friction welding process control method according to claim 1, wherein the iterative optimization of the hidden layer of the step S3 further includes: Setting an end condition of iterative optimization of the hidden layer, wherein the end condition measures the discrete degree of the predicted deformation parameter obtained by multiple iterations and the discrete degree of the error between the predicted deformation parameter and the actual deformation parameter obtained by multiple iterations.
7. The inertia friction welding process control method according to claim 1, wherein: the welding control parameters of the step S1 include flywheel rotational speed, flywheel rotational inertia and friction pressure.
8. The inertia friction welding process control method according to claim 1, wherein: The welding process parameters of step S2 include flywheel speed, friction pressure and friction torque.
9. The inertia friction welding process control method according to claim 1, wherein: The actual deformation parameter of the step S2 includes an actual shortening amount, the predicted deformation parameter of the step S3 includes a predicted shortening amount, and the design deformation parameter of the step S4 includes a design shortening amount.
10. The inertia friction welding process control method according to claim 1, wherein: The welding control parameters of the step S1 include flywheel rotational speed; The step S1 further includes setting the flywheel rotational speed so that the predicted deformation parameter of the step S3 is larger than the design deformation parameter of the step S4 when the step S1 is performed for the first time; The step S1 further includes setting the flywheel rotational speed to be lower than the flywheel rotational speed set when the step S1 was performed the previous time when the step S1 was not performed the first time.
11. An inertia friction welding process control system configured to perform the inertia friction welding process control method of any one of claims 1 to 10.
12. An inertia friction welding machine comprising a controller for performing the inertia friction welding process control method of any one of claims 1 to 10.

Description

Inertia friction welding process control method, control system and inertia friction welding machine Technical Field The invention relates to an inertia friction welding technology, in particular to a control method and a control system for an inertia friction welding process and an inertia friction welding machine. Background Turbine engines, particularly aviation turbojet engines and aviation turbofan engines, generate thrust by the combustion and work of compressed air. The efficiency of compressed air has a significant impact on engine thrust and efficiency. The compressed air of the aero-engine is mainly realized by the high-speed rotation of the multistage precisely matched blades in the compressor. In order to improve efficiency and performance, the multistage blisks of advanced aeroengines are connected by inertia friction welding. In the welding process, the inertia friction welding machine applies larger pressure to the blade disc connecting surface, and the flywheel drives the blade discs to rotate mutually to generate friction heat so as to soften the connecting surface, and finally stable connection is realized. Compared with the traditional fusion welding, the inertia friction welding has the advantages that the workpiece is subjected to high-temperature plastic deformation by applying pressure, but the metal material is not melted in the process, so that the high-temperature and long-time performance of the welding seam can be obviously improved, and the problems of microcracks and the like generated in the fusion welding process of high-temperature alloy, powder alloy and other difficult-to-weld materials can be avoided. However, inertia friction welding generates heat and deforms through the friction of a workpiece, and materials, equipment, a temperature field and a pressure field are mutually coupled in the welding process, so that the process control is extremely complex. Many scholars build a theoretical model for the welding process to calculate, but because friction welding is a process with large deformation, high temperature and high deformation rate, the friction welding process cannot be accurately calculated through analysis and solution due to the influences of phase transformation, plastic deformation and the like of materials in the welding process. In the inertia friction welding process of the multistage impeller, factors such as the difference of the multi-batch performance of the impeller raw materials and forgings, the difference of the multi-batch stability of welding equipment, tooling replacement, maintenance, production line combination and the like can cause deviation of the post-welding dimensional accuracy. In addition, as the advanced aeroengine is compact in structure and high in precision requirement, the dimensional deviation is difficult to repair through traditional machining after welding, and very strict requirements are put on the dimensional precision after welding, so that the great influence on the engine performance caused by the accumulation of the dimensional errors in the multi-stage blade disc welding process is avoided. In order to solve the problem of dimensional accuracy of multi-stage blade disc inertia friction welding, advanced aviation manufacturers at home and abroad often strictly control the production process of the whole blade disc, design a special machine for the high-accuracy and high-stability inertia friction welding machine, and stably control the whole process by performing multi-wheel welding tests so as to realize high-quality delivery of parts. By tightly controlling all influencing factors, the dispersity of the size after welding can be reduced to a certain extent. However, the method can only reduce the dispersity to a certain level, and cannot further improve the welding quality because the fluctuation accumulation of multiple influencing factors cannot be corrected, and the method cannot timely compensate the size fluctuation caused by sporadic factors, and only can discard parts when the size is shortened and changed due to emergency conditions (such as undiscovered welding machine pressure fluctuation), so that the economic loss of millions of yuan is caused. Disclosure of Invention The invention aims to provide a control method and a control system for an inertia friction welding process and an inertia friction welding machine, which are used for controlling the dimensional accuracy of inertia friction welding in real time. In a first aspect, the present invention provides a method for controlling an inertia friction welding process, and according to an embodiment of the present invention, the method includes step S1, step S2, step S3, and step S4. Step S1, setting welding control parameters, step S2, welding the welding control parameters in step S1, collecting welding process parameters and actual deformation parameters of a welding joint in the welding process, step S3, obtaining predicted deformation parameters of