摘要
为解决核电企业采购不统一带来的管理痛点,基于卓越绩效管理理念,提出基于大数据画像的采购单位识别方法,以采购单位为“客户”,探索供应链管理部门由被动“执行服务”到主动“创造价值”的转型模式。基于某集团核电电商平台的547万条数据,经过数据预处理、特征工程等步骤,构建了涵盖采购单位规模、效率、多样性、集中度等的13维指标体系,通过稀疏主成分分析对指标降维,通过参数寻优得到累计解释方差为80.36%的4个稀疏主成分,分别表征采购集中度、效率、高价值供应广度及采购活跃度。根据主成分得分,通过K-均值聚类分析,将34个采购单位分为5类:日常高效运作型、战略高价值寻源型、MRO分散长周期型、核心依赖保障型、低效分散支持型。通过Kruskal-Wallis分析显示,12个指标间存在显著差异,其中低效分散支持型呈现极短采购周期与高度分散特征。研究结论表明,该模式可以识别各单位的主要特征和风险,并针对性提出运营策略。为驱动持续改进,进一步设计了基于LeTCI框架的差异化关键绩效指标仪表盘,将绩效衡量与战略价值紧密关联。本研究为核电企业采购管理从“一刀切”的传统模式,向以客户为中心、数据驱动的智能化、精细化卓越绩效管理模式转型,提供了系统性的分析框架与可操作的决策支持方案及路径建议。
To address the management challenges caused by disagreements in procurement among internal departments,this study proposes a data-mining-based approach for profiling internal procurement units,guided by the excellent performance management framework.Treating procurement units as internal“customers,”this research aims to explore the transformation path for supply chain management departments from passive service executors to proactive value creators,leveraging the“management by fact”principle.Using 5.47 million transaction records from a group's nuclear power e-commerce platform,a 13-dimensional indicator system was constructed through data preprocessing and feature engineering,encompassing scale,efficiency,diversity,and concentration.Sparse principal component analysis(SPCA)was employed for dimensionality reduction.After parameter optimization,four sparse principal components(with a cumulative explained variance of 80.36%)were identified,representing the degree of procurement concentration,efficiency level,high-value supplier breadth,and procurement activity,respectively.Based on the principal component scores,K-means clustering was applied to segment 34 procurement units into five categories:daily high-efficiency operations type,strategic high-value sourcing type,MRO dispersed and long-cycle type,core dependency assurance type,and inefficient and dispersed support type.The Kruskal-Wallis test validated that 12 features differed significantly among the clusters,with the inefficient and dispersed support type exhibiting extreme procurement cycles and high dispersion.This study concludes that the profiling system effectively identifies the key characteristics and risks of each unit type,enabling the proposal of a series of differentiated operational excellence strategies.To drive continuous improvement,a differentiated KPI dashboard based on the LeTCl framework was further designed,closely linking performance measurement with strategic value.Ultimately,this research provides a systematic analytical framework and actionable recommendations for nuclear power enterprises to transition their procurement management from a traditional“one size-fits-all”approach to a customer-centric,data-driven,intelligent,and refined excellent performance management model.
作者
徐正
曹建成
方世杰
XU Zheng;CAO Jiancheng;FANG Shijie(China National Nuclear Supply Chain Operation Co.,Ltd.,Shanghai 200233,China)
出处
《中国核电》
2025年第4期483-492,共10页
China Nuclear Power
关键词
核电采购管理
稀疏主成分分析
K-均值聚类
采购单位画像
供应链风险管控
LeTCI框架
差异化策略
数据挖掘
procurement management
sparse principal component analysis
K-means clustering
procurement unit profiling
supply chain risk control
LeTCI framework
differentiated strategies
data mining
