基于不确定度和敏感度分析的LCA数据质量评估与控制方法

LCA data quality assessment and control based on uncertainty and sensitivity analysis

通过提出定量评估并控制LCA数据质量的系统化方法(称为CLCD-Q方法),从LCA案例的原始数据和清单数据算法开始评估不确定度;然后通过两次蒙特卡罗模拟,先后得出单元过程清单数据及LCA结果的不确定度;最后结合敏感度分析,辨识出LCA模型中具有高不确定度和高敏感度的关键数据,从而指出控制和改进数据质量的关键点.结果发现,上述方法可在eBalance软件和CLCD数据库中实现.同时,对中国电网电力生命周期的示例研究表明,上述方法将传统的LCA数据质量评估延伸到了原始数据层面,从而为数据收集过程中的原始数据与算法选择提供了直接的支持,同时也可以针对数据质量不达标的LCA结果,指出最有效的改进方向.

This paper presents a systematic approach, named as CLCD-Q method, to assess and control data quality of LCA studies. The method starts with the uncertainty assessment of raw data and mathematical relations based on pedigree matrix. Afterwards, the uncertainties of process data and LCA results can be derived from two Monte Carlo simulations. For each LCA result, key process data and raw data with high uncertainty and high sensitivity in LCA model can be identified, which indicates the ＂hot spot＂ for data quality improvement. CLCD-Q is supported by LCA software （eBalance） and CLCD database. The case study of Chinese grid power shows that this method can guide the selection of raw data and the mathematical relations with the uncertainty assessment extending on the raw data. It also provides a guide for efficient data quality improvement by revealing the most relevant data in the life cycle model.