Research On Fault Diagnosis Technology Of Machine Tools Spindle Based On Digital Twin

As the core component of CNC machine tool,the working performance of spindle directly affects the machining state of CNC machine tool.It has great significance to study fault diagnosis method of spindle.Traditional fault diagnosis method has good application effect in fault pattern recognition and fault location,but the structure of machine tool spindle is complex and coupled,it is difficult to obtain effective and enough data to train data-driven model,and it is difficult to build a mechanism model,which reduces the accuracy of fault diagnosis.Therefore,this thesis studies fault diagnosis technology of spindle based on digital twin,and proposes a digital twindriven fault diagnosis method for machine tool spindle.Introduce fault injection and multi-domain modeling technology,and establish a machine tool spindle fault injection modeling technology based on digital twin.Finally introduced the idea of model and data fusion,and proposed an intelligent diagnosis method of machine tool spindle faults based on model and data fusion.The specific research contents are as follows:(1)A digital twin-driven fault diagnosis method of machine tool spindle is proposed.In view of the changing working conditions of the machine tool spindle system,the difficulty in obtaining sensor monitoring information,and the inaccurate fault diagnosis results,the digital twin technology is introduced to analyze the system structure and failure mode of the machine tool spindle,and combine the difficulties of machine tool spindle fault diagnosis and The digital twin function innovatively proposes a digital twin-driven machine tool spindle fault diagnosis method,and analyzes the key technologies used in the method in detail.Compared with the traditional fault diagnosis method,this method combines the digital twin model with data,and can achieve accurate fault diagnosis even in the absence of data samples.(2)A digital twin-based fault injection modeling technology for machine tool spindle is established.In view of the low degree of participation and the single validity of the twin model in the application process of the current digital twin theory;the complex coupling relationship is difficult,diverse failure modes of spindle systems,and difficulty in building twin model for real-time mapping,multi-domain modeling and fault injection are introduced.By studying the system structure of the machine tool spindle,the coupling relationship between the various subsystems and the failure mechanism,the machine tool spindle fault injection modeling technology based on digital twin is established,and the digital twin model of the machine tool spindle fault state is realized.Finally,a digital twin model was constructed by taking the CNC machine tool JSC-180 A motorized spindle as an example,the model error is within10%,and fault injection was performed to verify the feasibility of the proposed method.Compared with the traditional digital twin model,this model considers the working conditions of the machine tool spindle.The environment,working mechanism and the mutual coupling between various subsystems,and can reflect the fault state of the machine tool spindle,have the advantages of high fidelity and wide application range.(3)An intelligent diagnosis method for machine tool spindle faults based on model and data fusion is established.Aiming at the intricate and implicit relationship between the failure mode and the cause of the machine tool spindle,it is difficult to directly mine the failure information,and fault quantify diagnosis is difficult.Combined with the digital twin modeling of machine tool spindles and fault injection methods,a digital twin model library of machine tool spindles is constructed.By combining with interpretable machine learning algorithms,a model selector is established.By analyzing the model,the intelligent diagnosis of machine tool spindle faults is realized.It is feasible and effective to reduce the failure by the stiffness of the machine tool spindle during the operation,the diagnostic error of stiffness deterioration degree is within 5%Compared with the traditional failure diagnosis method,it can realize the diagnosis of the deterioration degree of the failure.