空调产品绿色设计指标冲突识别及分析  

Conflict Identification and Analysis of Green Design Indicators for Air Conditioning Products

作　　者：苏金 胡嘉琦 尤莎莉 

机构地区：[1]合肥工业大学机械工程学院,合肥230041 [2]中国电器科学研究院股份有限公司,广州510000

出　　处：《日用电器》2023年第12期7-15,共9页ELECTRICAL APPLIANCES

摘　　要：为了解决绿色产品设计指标间存在的潜在冲突问题,提出了一种基于主成分分析-径向基神经网络(PCA-RBF)模型的设计指标冲突识别方法。首先,利用修正余弦相似性分析和1-9标度法(Analytic Hierarchy Process,AHP)构建目标判定矩阵,计算指标组冲突等级。然后利用冲突解决图表分析(CRD分析)框架识别制约设计指标协调评价的冲突指标组,并确定出各指标组的冲突等级范围。PCA主成分分析法对指标组降维后的低维数据为模型的输入,冲突等级数据为模型的输出,构建出输入与输出的映射关系,训练可得到能够预测指标组是否发生冲突的RBF神经网络预测模型。以空调产品为例,基于此方法得到的空调类产品指标冲突的预测值与真实值基本契合,通过对空调产品设计指标的分析,验证了该方法的科学性与可操作性。A method of green product design index conflict recognition based on PCA-RBF neural network model is proposed to solve the potential conflict among green product design indicators due to the correlation phenomenon.Firstly,the modified cosine similarity analysis and the 1-9 scale method(Analytic Hierarchy Process,AHP)were used to construct the target decision matrix and calculate the indicator group conflict rank.Then,the CRD analysis framework was used to analyze the constraints that lead to conflicts,and then the conflict index groups that restrict the coordination and evaluation of design indexes were identified through this framework,and the conflict grade range of each index group was determined.The PCA principal component analysis method was used to reduce the dimensionality of the indicator group dataset.The PCA-processed low-dimensional dataset is then collected.The reduced dimensionality low-dimensional dataset was used as the input to the model,and the conflict rank dataset was used as the output of the model.In turn,the mapping relationship between the input and output of the model can be constructed.Through learning training,we can obtain an RBF radial basis neural network prediction model that can predict whether there is a conflict between new indicator groups.Taking air conditioning products as an example,the predicted values of the conflicting indicators of air-conditioning products obtained based on this method basically match with the real values,and the operability of the model is verified through the analysis of the design indicators of air conditioning products.

关 键 词：冲突识别 CRD分析 PCA主成分分析 RBF径向基神经网络 

分 类 号：TM925.12[电气工程—电力电子与电力传动]