Study Of Optimization Method About Local Stiffened Carbon Fiber Reinforced Plastic Laminates

Due to the high stiffness ratio and strength ration, CFPR(carbon fiber reinforced plastic) has been widely used in many fields, such as aerospace and automotive industry. For long fiber composite structure, most of them are plate structure.Because of the low buckling ability and low resistance of the force along the thickness direction, it is urgent for local enhancement. Two of the common methods for local enhancement are to display ribs and add additional layers, namely change the thickness of the plates. Up to now, most of the optimization methods for ribs’ positions are conducted for cross-sectional dimensions assuming the unchanged layers conditions, which of course cannot fully explore the potentiality for better design. Therefore, this thesis conducts research on both single and multi-objective optimization for rib reinforced plates considering the cross-sectional shape and fibers sequential orders. Many researches have been conducted for the topology optimization of composite structure. However, these resultant structure cannot be easily manufactured. To overcome this, many scholars try to propose a topology method, and its resultant structure can be manufactured. However, due to the shortcomings of thickness interpolation model and too many constraints introduced, the algorithms are general too slow for convergence. Therefore, this thesis tries to propose a new optimization method for thickness parameterization.First, a problem decomposition method is introduced to the single optimization for composited plates with ribs. Design of experiment is introduced to ranking the design variables. Then surrogate models are built to improve the computational efficiency.Second, NSGA-II is used to solve this multi-objective optimization problem. In order to realize the change of numbers of the layers, permutation operator is introduced. At the same time, this paper conducts research on how to build surrogate model considering the variable number of layers, which leads to higher efficiency for optimization.Moreover, in the optimization of thickness for composite plates, this thesis proposed a density distributing curve method to realize thickness paramerization. Based on BCP, topology optimization is conducted for composite plates considering the variable thickness.