论文部分内容阅读
首先推导出了燃气轮机可靠性和造价之间的拟合函数;然后利用惩罚函数法把有造价约束问题转化为无约束优化问题,基于Hopfield神经网络构造和最优可靠性分配相对应的网络优化计算的能量函数,建立了系统可靠性优化分配的数学模型;最后用适合刚性方程的ode15s算法给出了最优的数值解.得出了惩罚因子变化趋势对最优数值解的影响关系:当惩罚因子的值在一定范围内逐渐增加时,优化结果趋于稳定,即收敛于最优解.改变约束条件,得出了燃气轮机可靠性随总造价的优化后的变化趋势,为燃气轮机的优化设计和综合决策提供一定的参考.
Firstly, the fitting function between gas turbine reliability and construction cost is deduced. Then, penalty cost method is used to convert cost constrained problem into unconstrained optimization problem. Based on Hopfield neural network construction and network optimal calculation corresponding to optimal reliability allocation The mathematical model of the optimal distribution of system reliability is established. Finally, the optimal numerical solution is given by the ode15s algorithm which is suitable for the rigid equation, and the influence of the trend of the penalty factor on the optimal numerical solution is obtained: when the penalty When the value of the factor is gradually increased within a certain range, the optimization result tends to be stable, that is to say, it converges to the optimal solution.Changing the constraints, the trend of the reliability of the gas turbine with the total cost is obtained and the optimal design of the gas turbine Comprehensive decision-making to provide some reference.