论文部分内容阅读
为了满足电站锅炉“高效低排”的运行要求,结合RBF神经网络,根据工况数据,分析了电站锅炉燃烧效率与NOx排放的矛盾关系,建立了锅炉NOx排放与热效率的混合模型。以此为基础,针对现有粒子群优化算法研究成果,引入了适应度与随机数值比较选择的思想和相似度函数的概念,并对算法的惯性权重进行了相应设计,使之随迭代次数逐渐减小,通过对测试函数的效果检验,表明算法的有效性。最后将其应用于锅炉混合模型中,进行某工况多目标优化仿真研究,得到了不同目标要求下的燃烧组合,为电站锅炉多目标优化提供了技术支持。
In order to meet the operating requirements of power plant boiler “high efficiency and low emission ”, combined with RBF neural network, according to the working condition data, the contradiction between combustion efficiency of boiler and NOx emission was analyzed and a hybrid model of boiler NOx emission and thermal efficiency was established. Based on this, aiming at the existing research results of particle swarm optimization (PSO), the concept of similarity and similarity function between fitness and random number is introduced, and the inertia weight of the algorithm is designed correspondingly so as to gradually increase with the number of iterations Decrease, and test the effectiveness of the test function to show the effectiveness of the algorithm. Finally, it is applied to the boiler mixing model to carry out a multi-objective optimization simulation of a working condition. The combustion combinations under different target requirements are obtained, which provides technical support for multi-objective optimization of power station boilers.