Three-dimensional Winding Path Planning And Experimental Validation Based On Generalized Mandrel Contours And The Minimum Potential Energy Principle

Filament winding technology is considered to be one of the most promising composite molding technologies because of its design flexibility,high structural efficiency and low manufacturing cost.However,China’s winding technology started late,and the industrial software at this stage has deficiencies such as low simulation accuracy and narrow applicability,which lead to quality defects such as fiber slippage and large gaps between tows in actual production.The core of the winding industrial software lies in winding path planning(CAD)and machine trajectory planning(CAM),Although many scholars have researched this issue,there is still a lack of effective 3D winding path planning methods in CAD,and there is a problem of solution failure due to over-constraint in CAM.Based on the in-depth analysis of the research status,the following three key scientific and technical problems have been refined:(1)the contradictory relationship between the generalized mandrel contour(GMC)and highprecision winding;(2)the mechanical mechanism of the expansion of the generalized mandrel contour;(3)the kinematic modeling in the multi-degree-of-freedom redundant space and its inverse solution algorithm.In order to solve the above three problems,3D winding path planning,machine trajectory planning,six-axis winding machine development,and experimental verification of trajectory and path accuracy are carried out.The research contents and conclusions of this dissertation can be summarized in the following four aspects.(1)A 3D winding path modeling method has been established.To avoid the doffing deviation caused by the traditional 2D winding path planning methods,a 3D winding path modeling method is established in this dissertation.The "pseudo-3D path"is constructed by geodesic offset and normal offset based on the center path,and the backward search(BS)algorithm and potential energy minimization optimization(PEMO)algorithm are proposed to correct the doffing heights of the points on the "pseudo-3D path".The BS algorithm is used to construct the GMC and provide input for the PEMO algorithm,which optimizes the potential energy of the winding path to solve the actual doffing height.The BS and PEMO algorithms are combined to model the 3D winding path on the cylindrical surface.(2)A series of algorithms for constructing GMC and solving the minimum potential energy path has been proposed.To overcome the deficiencies of the BS and PEMO algorithms in terms of computational efficiency and accuracy,the outer-contour expanding(OCE)algorithm and the convex helix(CH)algorithm are proposed for constructing the GMC and solving the minimum potential energy path,respectively.The OCE algorithm effectively reduces the storage space requirement of the GMC by introducing height sampling points.The CH algorithm is inspired by the convex hull theory of computational geometry,which can calculate the actual doffing height of each point by only scanning the path points once and has the advantages of high efficiency and accuracy.The combination of OCE and CH algorithms realizes the 3D winding path modeling on the truncated cone.However,the OCE and CH algorithms still have limitations when used for mandrel surfaces with large curvature.Therefore,the mesh-based directional projection(MDP)algorithm and the normal adaptive convex helix(NACH)algorithm are proposed for constructing the GMC and solving the minimum potential energy path,respectively.The MDP algorithm is based on the collision detection theory in computer graphics,which transforms the task of finding the doffing position into a ray-surface intersection problem with high efficiency and wide applicability,while the NACH algorithm optimizes the CH algorithm according to the physical laws to adapt to the mandrel surface of arbitrary curvature.By combining the MDP with the NACH algorithm,the 3D winding path modeling of pressure vessels is realized.In addition,the MDP and NACH algorithms have good prospects for 3D winding path modeling of complex non-axisymmetric surfaces such as elbows,T-pipes and S-shaped air inlets.(3)A machine trajectory planning algorithm based on the optimization of the shortest hanging tow has been proposed.The conventional machine trajectory planning method suffers from solution failure due to over-constraint.In this dissertation,from the perspective of robot kinematics,two virtual axes are ingeniously introduced to delineate the motion of the hanging tow.The six-axis machine trajectory planning is transformed into an eight-degree-of-freedom kinematic inverse problem,thus enhancing the dimensionality of the inverse solution space.By means of kinematic decoupling,the analytical inverse solution is obtained in the eight-degree-of-freedom redundant space,avoiding the errors and inefficiencies caused by numerical solutions.With the shortest hanging tow as the optimization objective,this algorithm plans the motion trajectory of the winding machine in the eight-degree-of-freedom space,which not only avoids the problem of solution failure but also helps to improve the winding accuracy and stability.(4)A six-axis winding software and hardware system has been developed and the algorithms proposed in this dissertation have been verified through experiments.The experimental results show that the 3D winding path can significantly improve the doffing accuracy and effectively avoid fiber slippage and gaps between tows.In addition,the GMC can effectively predict the winding thickness,which is of practical value for winding path optimization and product shape optimization.Finally,the 3D winding path provides a basis for mesoscale finite element modeling of the wound parts,which is of great significance for studying mechanical phenomena such as stress concentration,interface delamination and crack propagation inside the wound parts.