基于改进BP-GA方法的FPSO舷侧结构耐撞性能优化设计

基于遗传算法和ABAQUS参数化有限元仿真技术,对传统的BP-GA优化方法进行改进,并采用改进的BP-GA方法对浮式生产储油卸油装置(FPSO)舷侧结构的耐撞性能进行优化,以验证其可行性和准确性。结果表明,与传统的BP神经网络相比,经遗传算法优化的BP神经网络具有更高的预测精度和更强的泛化能力;改进的BP-GA优化方法可在结构减重的基础上进一步提高结构的耐撞性能,能较好地适用于复杂的FPSO舷侧结构耐撞性优化设计。采用的优化方法具有通用性,可为抗爆性能的优化设计提供参考。

Optimized Design of Crashworthiness of FPSO Side Structure Based on Improved BP-GA Method

Due to the complexity of ship collision optimization, traditional optimization methods are difficult to be used effectively. The traditional BP-GA optimization method is improved based on the genetic algorithm and ABAQUS parameterized FEM simulation technology. The crashworthiness of a FPSO side structure is optimized by using improved BP-GA to verify its feasibility and precision. The results show that the BP neural network optimized by genetic algorithm has higher prediction accuracy and generalization ability than the traditional BP neural network. The improved BP-GA optimization method can further improve the crashworthiness of structures based on structural weight reduction, which is more suitable for the complex ship structure crashworthiness optimization. The optimization method proposed in this paper is versatile and can provide a reference for the optimization of other properties, such as the anti-explosion capacity.