Optimal Design Method Of Multi-stress Accelerated Degradation Test Scheme For Motorized Spindle Of Machining Center

As the main machine tool of the equipment manufacturing industry,the CNC machine tool is an important symbol of a country's comprehensive competitiveness.However,due to the problem of reliability,the current development of China's machine tool industry has fallen into a bottleneck.As the key functional component of the NC machine tool,the reliability of motorized spindles is the main aspect that ensures the reliability of NC machine tools,and the reliability test is the basic work of reliability engineering.At present,accelerated test technology has become the main method to improve the reliability of a motorized spindle.However,due to the complexity of the structure and the working environment of a motorized spindle in a machining center,no reliability test equipment has been able to simulate complex actual working conditions,such as the dynamic and static cutting force of motorized spindle in a machining center for a prolonged period of time.Therefore,it would be impossible to conduct a real reliability accelerated test.In recent years,the Key Laboratory of Numerical Control Equipment Reliability of the Ministry of Education has developed a test equipment that can realize the dynamic and static loading of each stress of a motorized spindle of the machining center.This test equipment can replicate the test conditions to restore the real working conditions of a motorized spindle.Based on actual test experience,the acceleration degradation test with high efficiency is selected as the technical approach to conducting the reliability test of the motorized spindle of the machining center.Although the accelerated degradation test is characterized by high test efficiency,a complete test still needs to consume a great deal of test time and cost due to two reasons: the characteristics of the motorized spindle of the machining center and and the accelerated degradation test cannot change the degradation mechanism of the motorized spindle.Thus,in order to improve the test efficiency and save on test cost,the test scheme is usually optimized according to the established test objectives before conducting the test.However,with the increase of test stress types in the accelerated degradation test and the expansion of test applications,some problems cannot be easily solved by the existing optimization design methods.Therefore,to solve the problems faced by the optimization design of the multistress accelerated degradation test of motorized spindles,this paper considers the motorized spindle of the machining center as the research object and conducts research on the optimization design of such a test.The main research contents and innovative achievements of this paper are as follows:(1)Establishes a three-stress accelerated degradation model of motorized spindle with stress coupling term.In this work,the importance of the bottom fault of the motorized spindle of the machining center is determined according to the structure and function of the motorized spindle of the machining center and the historical fault data of similar products,combined with the fault tree analysis method and potential value improvement method.Based on the importance and the influence of each stress on the reliability of the motorized spindle,the comprehensive importance of each stress is obtained.The cutting force of the spindle speed and cutting power are used as the stress types in the modeling of motorized spindle acceleration degradation model.At the same time,we determine the specific expression of the three stress accelerated degradation model of motorized spindle with stress coupling term according to the coupling between the selected stresses and the correlation between the stress and the degradation rate of motorized spindle.Based on the conjugate prior distribution theory and prior data,we are able to determine the expression form of prior distribution of each parameter in the model and the parameter value in each prior distribution.(2)Proposes the optimal design method of multi-stress accelerated degradation test scheme with poor prior information.To solve the optimization design problem of the multistress accelerated degradation test scheme of a motorized spindle with poor prior information,an optimization design method,with information entropy as the objective function,is thus proposed based on Bayes theory.Furthermore,to maximize the test efficiency,the optimization design is changed from the traditional single optimization before the test to multiple optimization before and during the test,thereby realizing the “dynamic”optimization of the test scheme.The optimization model before the test is established by taking the maximum information entropy of the test scheme as the objective function,the value space of the test scheme and the expected cost of the test as the constraints,and the number of samples and the detection times of each stress level as the optimization design variables.In this way,the rapid solution of the optimization model is realized by genetic algorithm.In the test,in which the actual test information is considered,the distance between the prior distribution and the posterior distribution of the model parameters under each stress level in the test and the actual test situation are used as the judgment criteria.Hence,the basis for the dynamic adjustment of the test scheme is established,and the optimal design and dynamic adjustment of the multistress accelerated degradation test scheme of motorized spindle with poor prior information are realized.(3)The optimal design method of the multistress accelerated degradation test scheme in a non-cube test area is proposedTo restore the real working condition in the test,the limit stress level is determined without changing the fault mechanism of the tested motorized spindle through the stress mapping test,thus establishing a non-cube test area based on the real working conditions and stress coupling relationship.Then,based on the uniform design method and the Escobar Meeker design method,the optimization method of test stress points in this test area is proposed.Through example calculation,we obtain the test scheme with the minimum asymptotic variance of the estimated mean time between failures of the normal stress level of the motorized spindle under different stress levels,after which we determine the optimal number of samples and stress levels of the test scheme.At the same time,we establish the evaluation index of the test scheme to provide a theoretical basis for the subsequent optimization method of multi stress accelerated degradation test scheme of motorized spindle in a non-cube test area.According to the importance of each index in the project and the analytic hierarchy process,the weight of each index is assigned,and then the comprehensive evaluation method of the test scheme is proposed.Based on this evaluation method,we evaluate the optimal test scheme under each stress level is comprehensively evaluated and then propose the optimization method of the multistress accelerated degradation test scheme of the motorized spindle in the test area according to the evaluation results.(4)Proposes an optimal design criterion of multistress accelerated degradation test scheme of motorized spindle considering parameter weight.Here,we define the importance sensitivity conversion ratio and determine the weight value of each model parameter to solve the problem of identifying the estimation accuracy of important parameters or non-main parameters of the model when using the existing optimization criteria.This solution can help optimize the multistress accelerated degradation test scheme of motorized spindles based on the qualitative results of sensitivity analysis under the established objective function and the quantitative index of comprehensive importance of stress mentioned above.Furthermore,by combining this weight value with the optimization design idea of Ds optimization criterion,an optimization criterion considering the weight of model parameters is proposed.At the same time,an optimization model is established with the objective function in the optimization criterion as the optimization objective,the expected test cost as the main constraint,and the number of test samples,the detection interval,and detection times of each stress level in the test as the optimization design variables.The Tabu particle swarm optimization algorithm is also introduced to realize the rapid solution of the model,thus obtaining the optimal test scheme under different expected test costs.The effectiveness of the proposed method is verified by comparing the optimal test scheme with each optimal test scheme under the D and Ds optimization criteria,respectively.(5)Proposes a multi-objective optimization design method of multistress accelerated degradation test scheme of motorized spindles based on determined preference and variable decomposition.According to engineering importance,the estimation accuracy of mean time between failures of normal stress level,the estimation accuracy of model parameters,and the robustness of the test scheme are taken as the objective functions of the multiobjective optimization design.Then,the information entropy is introduced into the multi-objective function as the judgment function to ensure the correctness of optimization results.To achieve the rapid solution of the optimal experimental scheme,the preference guidance direction is also determined according to the multi-objective function.Furthermore,the optimization of the experimental scheme is transformed from a high-dimensional search problem to a low-dimensional search problem based on the correlation of variables.Finally,based on the proposed multi-objective function and search method,a multi-objective optimization design model is established,in which the expected test cost is the constraint and the number of test samples,the detection times,and detection interval of each stress level in the test are the design variables.Through the solution of the model,the optimal test scheme under different expected test costs is obtained.The effectiveness of the proposed method compared to the optimal scheme of multi-objective particle swarm optimization algorithm is also proven.