摘要: To address the uncertainty in hydrogen demand, this study proposes a chance-constrained stochastic multi-objective optimization model for China"s green hydrogen supply chain. The model simultaneously minimizes the total system cost and maximizes renewable energy consumption. It is applied to optimize the spatial layout, technology selection, and long-term development pathways of provincial hydrogen production, transmission, and storage in China for 2022–2040. Results from a trade-off decision preference at a 99% confidence level indicate that China’s green hydrogen production will reach 41.94 million tons by 2040, with 58.8% produced by alkaline electrolysis (ALK). Hydrogen transmission increases from 22,400 tons in 2022 to 1.26 million tons, with 80.1% transported via liquid hydrogen tank trucks and pure hydrogen pipelines. Hydrogen storage expands from 10,300 tons to 56.35 million tons, with the annual growth rate exceeding 461.6% from 2038 onward. Spatially, the model identifies distinct regional roles: Inner Mongolia emerges as the largest hydrogen-producing province (20.3%), Anhui and Fujian as major hydrogen storage hubs (21.1% and 17.3%, respectively), northeastern and northwestern provinces as key exporters (56.9%), and central China as the main recipient region (36.7%). Further analysis shows that increasing the confidence level significantly raises both system costs and water consumption, peaking at CNY 12.07 trillion and 476 million m3, respectively. The findings suggest China should promote region-specific large-scale green hydrogen deployment. Key recommendations include expanding proton exchange membrane (PEM) and solid oxide electrolysis cell (SOEC) production capacity, establishing hydrogen production hubs in Shanxi, Inner Mongolia, Shandong, and Anhui, strengthening transmission corridors in the northeast and northwest, and developing storage facilities in Qinghai, Ningxia, and Shanxi to foster an integrated green hydrogen supply chain.
摘要: 倾向得分匹配-双重差分模型(PSM-DID)是政策评估及因果推断中最为流行的方法之一.但是在实际应用中,该方法面临着控制变量在处理组样本和控制组样本之间非平衡性的挑战.传统基于均值差异t检验的平衡性检验容易产生片面和误导性的结论,使得后续因果推断产生偏误.为克服上述问题,本文对传统的平衡性检验提出以下改进:一是推荐更全面的多维度的平衡性测度指标,便于在匹配后更严谨地比较处理组和控制组的平衡性;二是提出了适用于非平衡样本的新估计方法:倾向得分匹配-逆概率加权-双重差分(PSM-IPW-DID),该方法结合了倾向得分匹配(PSM)克服样本自选择内生性及对非平衡样本稳健的优势和逆概率加权(inverse probability weighting, IPW)利用全样本信息的长处,在不进一步删除样本的情况下得到一种更稳健的双重差分估计方法.数据模拟和应用实例显示,本文提出的新方法能更全面、客观地评价宏观、微观政策的作用,得到更为可信的因果推断.
摘要: 倾向得分匹配-双重差分模型(PSM-DID)是政策评估及因果推断中最为流行的方法之一.但是在实际应用中,该方法面临着控制变量在处理组样本和控制组样本之间非平衡性的挑战.传统基于均值差异t检验的平衡性检验容易产生片面和误导性的结论,使得后续因果推断产生偏误.为克服上述问题,本文对传统的平衡性检验提出以下改进:一是推荐更全面的多维度的平衡性测度指标,便于在匹配后更严谨地比较处理组和控制组的平衡性;二是提出了适用于非平衡样本的新估计方法:倾向得分匹配-逆概率加权-双重差分(PSM-IPW-DID),该方法结合了倾向得分匹配(PSM)克服样本自选择内生性及对非平衡样本稳健的优势和逆概率加权(inverse probability weighting, IPW)利用全样本信息的长处,在不进一步删除样本的情况下得到一种更稳健的双重差分估计方法.数据模拟和应用实例显示,本文提出的新方法能更全面、客观地评价宏观、微观政策的作用,得到更为可信的因果推断.