基于EBF神经网络的复合材料耐压壳性能研究

Research on a composite pressure hull based on an EBF neural network

为了对水下航行器非均匀内部环肋复合材料耐压壳进行高效率结构性能设计,基于复合材料可设计性特点,应用复合材料细观力学刚度的材料力学理论与有限元分析方法对耐压壳结构进行力学仿真分析,结合EBF椭圆基神经网络近似模型技术以及拉丁超立方设计试验方法从细观层面对组分材料属性在非均匀环肋复合材料耐压壳性能中的影响进行研究。结果表明:纤维和基体的弹性模量对耐压壳结构材料力学性能影响最大,剪切模量对各力学性能影响都很小。在组分材料属性一定的情况下,随着纤维体积分数增加,Tsai-Wu失效指数和临界失稳压力有所提高,而相邻肋骨中点处壳板周向应力、肋骨处壳板轴向应力和肋骨应力随之降低。因此,在非均匀环肋复合材料耐压壳设计过程中应重点考虑较大弹性模量以及适当纤维体积分数的组分材料以达到结构性能最优化的目的。

In order to implement a structural design with high efficiency for a non-uniform inner ring-stiffened composite pressure hull of an underwater vehicle,material mechanics theory of composite material mesomechanical rigidity and finite element analysis were adopted for simulation analysis of a pressure hull on the basis of the designability of composites. Combined with EBF neural network approximation technology,a Latin hypercube design test method was utilized to study the effect of the properties of the component material on mechanical performance of the non-uniform ring-stiffened composite pressure hull. The results illustrate that the elasticity modulus of the fiber and matrix mostly influence the mechanical properties of the pressure hull, while the influence of shear modulus is small. Given that the properties of component material are certain,the Tsai - Wu failure index and critical buckling pressure will increase with the increase of fiber volume fraction,while the circumferential stress of midpoint of adjacent frames,the axial stress of the shell at frames and the frame stress decrease. Therefore,in the design progress of the non-uniform ring-stiffened composite pressure hull,a large elasticity modulus and proper fiber volume fraction of component material should be taken into consideration to obtain optimal structural performances.