摘要
介绍了多光谱辐射测温原理、遗传算法(genetic algorithm,GA)以及第二代非支配排序进化算法(non-dominated sorting genetic algorithm-Ⅱ,NSGA-Ⅱ)的应用。从自适应角度出发对NSGA_Ⅱ算法加以改进,使其交叉变异算子具有一定动态调整能力,并把差分进化算法融合到NSGA_II变异算子进化中使其进化方向得以优化。作者从计算精度,稳定度,计算速度角度出发对传统的GA遗传算法、经典的NSGA-Ⅱ算法和改进后的NSGA-Ⅱ算法进行仿真实验并且利用真实四路测温实验验证,结果表明改进的NSGA-II算法具有单次计算精度高、稳定性好、计算速度略快、最后一代种群特性良好等优点,适合应用在温度范围为700~1 000℃下涡轮叶片多光谱辐射测温中。
This paper briefly introduces the principle of multi-spectral radiation temperature measurement and the application of genetic algorithm （GA） and the second generation of non-dominated sorting evolutionary algorithm （NSGA_Ⅱ） in this circumstance. In this paper, the NSGA II algorithm was improved from the adaptive angle, so that the cross-mutation operator has a certain dynamic adjustment ability, the differential evolution algorithm was integrated into the evolution of NSGA_Ⅱ mutation operator for optimizing the evolution direction. The traditional GA, NSGA_Ⅱ algorithm and the improved NSGA_Ⅱ algorithm were simulated on such aspects as calculation precision, stability and calculation speed, in addition, a true four-way temperature measurement experiment was used for veri- fication. The results show that the improved NSGA-Ⅱ algorithm has such advantages as high single-time calculation accuracy, more stable, a little fast calculation speed and excellent population characteristics of the last generation. Therefore, the NSGA_Ⅱ algorithm is applicable for the turbine blade multi-spectral radiation temperature measurement within the temperature range of 700 - 1 000 ℃.
出处
《应用科技》
CAS
2018年第1期89-95,共7页
Applied Science and Technology
关键词
多光谱测温
NSGA-Ⅱ算法
GA算法
涡轮叶片
差分进化
种群特性
稳定度
精度
muhispectral temperature measurement
NSGA-Ⅱ evolution
population characteristics
stability
precision algorithm
GA algorithm
turbine blade
differential
