热力发电
熱力髮電
열력발전
THERMAL POWER GENERATION
2014年
9期
108-112
,共5页
配煤%多目标%均匀设计%BP网络%遗传算法
配煤%多目標%均勻設計%BP網絡%遺傳算法
배매%다목표%균균설계%BP망락%유전산법
coal blending%multiple obj ects%uniform design%BP neural network%genetic algorithm
针对试验法动力配煤存在操作繁琐、信息滞后较大以及配煤方案粗略等缺陷,将BP 网络、均匀设计、遗传算法等方法相结合,建立了多目标遗传算法优化模型,快速得到优化配煤方案。其中,BP算法用于实现混煤与单煤的发热量等煤质信息之间的非线性映射关系；采用均匀设计建立适应度函数；利用遗传算法进行寻优。研究结果表明：预测模型具有较高的可靠性和置信度。
針對試驗法動力配煤存在操作繁瑣、信息滯後較大以及配煤方案粗略等缺陷,將BP 網絡、均勻設計、遺傳算法等方法相結閤,建立瞭多目標遺傳算法優化模型,快速得到優化配煤方案。其中,BP算法用于實現混煤與單煤的髮熱量等煤質信息之間的非線性映射關繫；採用均勻設計建立適應度函數；利用遺傳算法進行尋優。研究結果錶明：預測模型具有較高的可靠性和置信度。
침대시험법동력배매존재조작번쇄、신식체후교대이급배매방안조략등결함,장BP 망락、균균설계、유전산법등방법상결합,건립료다목표유전산법우화모형,쾌속득도우화배매방안。기중,BP산법용우실현혼매여단매적발열량등매질신식지간적비선성영사관계；채용균균설계건립괄응도함수；이용유전산법진행심우。연구결과표명：예측모형구유교고적가고성화치신도。
There are several disadvantages when using the determination method for coal blending,such as complicated operation,lag information and rough blending scheme.Therefore,by combiring the BP net-work,uniform design and genetic algorithm,a multi-obj ective genetic algorithm optimization model was es-tablished,to get coal blending optimization solution faster.The model employs the BP neural network to determine the nonlinear relationship between the blending coal and single coal during the learning process, the uniform design to establish fitness function,and the genetic algorithm to optimize.The results proved that the prediction model has high reliability and confidence.
