Applying regression analysis to optimize the length of components of a six bar mechanism

6th ModTech International Conference - Modern Technologies in Industrial Engineering

Gheorghita, V.^1 ; Gheorghita, C.^1

Politehnica University of Bucharest-Romania, Department of Manufacturing Engineering, Splaiul Independenei nr. 313, Bucureti, Bucharest

060042, Romania^1

关键词: Correlation coefficient;    Direction of motion;    Kinematic Analysis;    Planar mechanism;    Regression model;    Regression modelling;    Six-bar linkages;    Specialized software;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/400/4/042024/pdf
DOI  :  10.1088/1757-899X/400/4/042024

来源: IOP

【 摘 要 】

A mechanism can be defined as several rigid bodies connected to produce constrained relative motion between them and transmit forces and couples from the source of input power to result in motion. The movement of a body is studied using geometry so the linksare considered to be rigid. The connections between links are modeled being ideal movement, pure rotation or sliding. The design of mechanism presupposes knowledge of the direction of motion, velocity, accelerations and the forces of each component at various positions during the operating cycle. The necessary informations can be obtained from the kinematic analysis of the motion of the mechanism. Because the purpose of a planar mechanism is to move one of the links or a point of an element through a specific motion, it is important to verify that the mechanism performs the established function. This paper presents the problem of modelling and analyzing the six bar linkages mechanism using specialized software. Regression analysis will be performed to find the optimum domain lenghts of the parameters bars taking into account the angles of the mechanism to ensure the stability of the entire equipment. The results of the model were used to correct and optimize the design of the product. Effectiveness of regression models has been determined by analyzing correlation coefficient and by comparing with the experimental results. It has been concluded that regression modelling can be used to predict the datas accurately.