Study On Fusion Modeling And Application Theory Of Assembly Variation Sconsidering Multi-source Information In Auto Body Assembly Processes

The dimensional variation of auto bodies is one of the most important factors that affect the vehicle quality,manufacturing cost and market performance.Therefore,the quality control and diagnosis methods of body assembly have been highly valued by academics and industry.Predictive control and diagnosis of body assembly dimensional variation is an important way to improve product quality.However,the complexity of the assembly system and the incompleteness of the observation information make it difficult to accurately describe the relationship between the various sources of deviation and the quality of the body assembly,which limits the effectiveness of the model prediction control.Besides,the assembly stations.characteristic of multi-fault source coupling in multi-stations assembly process restricts the application of traditional statistical-based diagnostic methods under small sample conditions.Multi-source information fusion technology is an effective method for data mining under limited information conditions.On the one hand,through fusing a variety of assembly deviation prediction models can solve the problem that the traditional single model prediction accuracy is not high due to the technology used and the limited information extracted;On the other hand,the current measurement data and process knowledge can be fused to achieve fault diagnosis under small samples.In this paper,multi-source information fusion technology is applied to the quality control of vehicle body assembly,and the assembly deviation prediction control and real-time diagnosis method based on multi-source information fusion are proposed.Firstly,the combination model of variable weights considers both the static theoretical mechanic variation propagation mode and the dynamic variation relationships from the regression model based on data collections.Then,the qualification rate prediction method and prediction control method based on the model are studied.After that,the knowledge-based diagnosis method is studied.Finally,the effectiveness of the method proposed in this paper is verified by the actual engineering case.The main research contents are shown as follows:(1)Modeling of vehicle body assembly deviation based on multi-source information fusionIn order to improve the accuracy of vehicle body assembly quality model prediction,a method for establishing a variable weight combination forecasting model based on variation propagation theory and multi-source measurement data is studied.Firstly,the variation propagation relationship in the actual assembly process is analyzed,and the input and output of the assembly deviation model are defined.Then two kinds of single prediction models are established based on the variation propagation theory and multi-source regression theory.On this basis,the modeling method of the combined prediction model,the method for solving the weight coefficient,and the extrapolation method for the weight coefficient are studied.(2)Model predictive control(MPC)method based on fusion modelAiming at the problem that the lack of efficiency of traditional experience-driven quality control methods,this paper proposed the method of quality prediction control based on fusion model.Firstly,through the estimation of the distribution of influencing factors,Monte Carlo simulation is used to establish a qualified rate prediction method based on fusion model.For the unqualified situation,combined with the actual process knowledge and variable contribution rate,a comprehensive cost-driven predictive control strategy is established based on the fusion model.(3)Knowledge-based online diagnosis methodAiming at the problem that the traditional fault diagnosis method is difficult to realize real-time diagnosis,this paper studies the diagnosis method of assembly fault source under small sample based on assembly process knowledge and high-dimensional end measurement data.Based on the assembly process knowledge,including multi-station assembly hierarchy,fixture scheme,measurement characteristics and tolerances etc in the multi-station,a knowledge-based diagnostic methodology and procedures are proposed with the measurements of each Body in White(BIW)for part/component defections and faulty assembly station identification.For the station involved with defective parts/components,the sub-coordinate system of the part/component is established reflecting its position and pose in the space,and then the relative pose matrix to the “normally build” pose is calculated based on the deviations of sub-coordinates of the parts in this station.Finally,the assembly process malfunctions are determined by a proposed rule-based strategy with the relative pose matrix in real time.The proposed multi-information fusion modeling method and quality control methods will provide references for quality strategies in the production process in the mode of intelligent manufacturing and is probable to be used in the real assembly shops.