建筑节能
建築節能
건축절능
CONSTRUCTION CONSERVES ENERGY
2013年
11期
67-69
,共3页
季文娟%顾永松%冯乐
季文娟%顧永鬆%馮樂
계문연%고영송%풍악
T-S模糊神经网络%改进FCM聚类%能耗评估模型
T-S模糊神經網絡%改進FCM聚類%能耗評估模型
T-S모호신경망락%개진FCM취류%능모평고모형
T-S fuzzy neural network%improved FCM clustering%assessment model of energy consumption
通过研究公共建筑耗能特点,结合建筑节能标准,提取了建筑能耗主要影响因素。在研究T-S模糊神经网络结构和算法的基础上,提出了基于T-S模糊神经网络的建筑能耗评估模型,用改进FCM聚类方法确定网络结构和参数初值,运用混合学习算法训练网络模型。将模型运用到评估实例中,结果表明基于改进FCM聚类的T-S模糊神经网络评估模型结构简单,学习和泛化能力强。
通過研究公共建築耗能特點,結閤建築節能標準,提取瞭建築能耗主要影響因素。在研究T-S模糊神經網絡結構和算法的基礎上,提齣瞭基于T-S模糊神經網絡的建築能耗評估模型,用改進FCM聚類方法確定網絡結構和參數初值,運用混閤學習算法訓練網絡模型。將模型運用到評估實例中,結果錶明基于改進FCM聚類的T-S模糊神經網絡評估模型結構簡單,學習和汎化能力彊。
통과연구공공건축모능특점,결합건축절능표준,제취료건축능모주요영향인소。재연구T-S모호신경망락결구화산법적기출상,제출료기우T-S모호신경망락적건축능모평고모형,용개진FCM취류방법학정망락결구화삼수초치,운용혼합학습산법훈련망락모형。장모형운용도평고실례중,결과표명기우개진FCM취류적T-S모호신경망락평고모형결구간단,학습화범화능력강。
Main influence factors are extracted by analyzing characteristics of energy consumption with energy efficiency standards for building. The model for energy consumption assessment is built after researching structure and algorithm of TS fuzzy neural network. Improved FCM clustering is used to determine the structure of network and the initial values of former parameters. The model is trained by hybrid learning algorithm. The model is used to assess the energy consumption situation of public buildings. The structure of T-S fuzzy neural net-work assessment model based on improved FCM clustering is simple, and it has strong learning and generalization ability.