Scholarlink

<p>Hydrogen plays an increasingly important role in the world’s carbon neutrality, but due to the high cost of storage, underground hydrogen storage (UHS) especially in depleted natural gas fields is considered. An important factor for UHS is the ability to predict gas loss during the cycles of injection and production. The use of reservoir simulation can be computationally exhaustive. Alternatively, we can use semi-analytical data-tuned methods such as the hybrid capacitance resistance model and long short-term memory model. We apply this model for the first time to UHS. The hybrid model closely aligns with actual data, reducing the maximum error rate from 4.32% to 2.37% and increasing computational time from 3.5 s to over 4.5 s. The study also highlights the unique challenges of storing hydrogen, which has a lower density than methane and a smaller molecular size with risks of escaping or leakage. In 2030, hydrogen production is set to rise significantly, with three key areas of strategic development expected to contribute over 70% of the national output in China, emphasizing the role of three key areas in bolstering global energy sustainability. Predictions indicate substantial potential hydrogen loss rates, particularly in these key areas, with projections showing losses exceeding 0.4 million tons/year in one of the key areas alone.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Gas loss prediction in underground hydrogen storage using an improved capacitance–resistance model for carbon neutrality of China

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年3月25日

查看全文
  • Zhengguang Liu,
  • Liu Lu,
  • Lin Ma,
  • Haizhi Luo,
  • Xiaohu Yang,
  • Masoud Babaei

摘要

Hydrogen plays an increasingly important role in the world’s carbon neutrality, but due to the high cost of storage, underground hydrogen storage (UHS) especially in depleted natural gas fields is considered. An important factor for UHS is the ability to predict gas loss during the cycles of injection and production. The use of reservoir simulation can be computationally exhaustive. Alternatively, we can use semi-analytical data-tuned methods such as the hybrid capacitance resistance model and long short-term memory model. We apply this model for the first time to UHS. The hybrid model closely aligns with actual data, reducing the maximum error rate from 4.32% to 2.37% and increasing computational time from 3.5 s to over 4.5 s. The study also highlights the unique challenges of storing hydrogen, which has a lower density than methane and a smaller molecular size with risks of escaping or leakage. In 2030, hydrogen production is set to rise significantly, with three key areas of strategic development expected to contribute over 70% of the national output in China, emphasizing the role of three key areas in bolstering global energy sustainability. Predictions indicate substantial potential hydrogen loss rates, particularly in these key areas, with projections showing losses exceeding 0.4 million tons/year in one of the key areas alone.