摘要
热耗率是汽轮机重要的热经济性指标之一。针对热耗率难以准确计算的问题,提出一种基于云粒子群算法(CPSO)优化快速学习网(FLN)的热耗率短期预测模型。在粒子群算法(PSO)中引入了云模型自适应权值策略,利用云滴的随机性和稳定倾向性特点自适应地调整粒子群算法的权值以提升PSO算法的全局优化性能。采用CPSO算法调整FLN的模型参数并建立CPSO-FLN热耗率预测模型。最后,将CPSOFLN模型应用于某汽轮机的热耗率预测,输入参数为12个强相关性的可控变量,将热耗率预测结果与标准的FLN模型和PSO-FLN模型预测结果进行对比。结果表明,CPSO-FLN模型具有更高的预测精度和泛化能力,是一种有效的预测方法。
The heat rate is one of the important indexes to assess the thermal economy of thermal power units. Regarding the difficulties in accurate heat rate calculation, this paper proposes a short-term heat rate forecast model based on optimized fast learning network (FLN) by cloud particle swarm optimization (CPSO). The cloud model based adaptive weight strategy is introduced into the particle swarm algorithm such that the weight of the particle swarm algorithm can be self-adaptively adjusted to improve the global optimization performance of the PSO algorithm by taking advantage of the randomness and stability of cloud droplets. Furthermore, the CPSO algorithm is used to tune the model parameters of FLN and establish the CPSO-FLN heat rate forecast model. Finally, the CPSO-FLN model is applied to predict the heat rate of a steam turbine with 12 controllable variables of strong correlation as the input parameters. Through the comparison of the results from the proposed model with those from standard FLN model and PSO-FLN model, the CPSO-FLN model has demonstrated higher accuracy and better generalization capacity, and hence is proved to be an effective forecast method.
作者
胡坚
刘超
HU Jian;LIU Chao(Information Technology Department, Zhejiang Institute of Economics and Trade, Hangzhou 310018, China;Guizhou Aerospace Electronics Co., Ltd., Guiyang 550009, China)
出处
《中国电力》
CSCD
北大核心
2019年第8期179-184,共6页
Electric Power
关键词
汽轮机
热耗率
快速学习网
粒子群算法
云模型
steam turbine
heat rate
fast learning network
particle swarm optimization algorithm
cloud model
