论文部分内容阅读
针对超临界机组末级长叶片的设计特点,采用遗传算法和人工神经网络,提出对长叶片典型截面叶型进行分区优化设计思想,并对原型与改型进行了多工况点的数值计算,结果表明,将叶型吸力侧后半段由直线型改为内凹型,能够显著降低超声速叶型在超声速工况范围内的叶型损失。对叶型前缘以及压力侧的局部优化设计能够改善超声速叶型在临界马赫数工况下的气动性能。优化设计最大程度地减小了样本空间,提高了优化效率。
Aiming at the design characteristics of the long blade at the final stage of supercritical generating unit, a genetic optimization algorithm and artificial neural network are proposed to optimize the design of typical sectioned blade of long blade, and numerical calculation is made for prototype and retrofit. The results show that changing the second half of the suction side of the blade from linear to concave can significantly reduce the leaf loss of the supersonic leaf in the supersonic working condition. The local optimum design of the leading edge of the blade and the pressure side can improve the aerodynamic performance of the supersonic blade under the critical Mach number. Optimized design minimizes sample space and optimizes efficiency.