Removing copper from nickel electrolysis anode solution has been a major keypoint in the nickel metallurgy industry.In this study,we proposed a novel process flow to promote removing copper from nickel electrolysis an...Removing copper from nickel electrolysis anode solution has been a major keypoint in the nickel metallurgy industry.In this study,we proposed a novel process flow to promote removing copper from nickel electrolysis anode solution.A simulated nickel anode solution was designed,and static and dynamic adsorption experiments were conducted to determine the best of solution pH,adsorption time and temperature,resin dosage and particle size,and stirring speed.The optimal conditions were explored for copper removal from nickel electrolysis anode solution.Based on the optimal experimental conditions and the relevant experimental data,a novel process for copper removal from nickel electrolysis anodes was designed and verified.This novel process of copper removal from nickel electrolysis anodes was confirmed with nickel anolyte solution with nickel 50−60 g/L and copper 0.5 g/L.After finishing the novel process of copper removal,the nickel in the purified nickel anolyte became undetectable and copper concentration was 3 mg/L,the novel process of resin adsorption to remove copper from nickel anode solution through static and dynamic adsorptions has an efficacious copper removal.It is a beneficial supplement to traditional methods.展开更多
生成式人工智能(AIGC)的快速崛起,掀起“人工智能+”赋能各行各业的浪潮,对会计领域的研究提供了新思路。由于财会行业要求高精确性与可解释性,而大语言模型存在着模型幻觉与决策黑箱,使得大模型应用于财会领域仍存在较大阻碍,构建财会...生成式人工智能(AIGC)的快速崛起,掀起“人工智能+”赋能各行各业的浪潮,对会计领域的研究提供了新思路。由于财会行业要求高精确性与可解释性,而大语言模型存在着模型幻觉与决策黑箱,使得大模型应用于财会领域仍存在较大阻碍,构建财会知识图谱成为弥补大语言模型不足的有力工具。基于此,文章调用GPT-4o mini API,利用提示词工程(Prompt)与智能体(Agent)思路方法进行自动化知识抽取,构建中国会计准则知识图谱,从模式层提取财会实体与实体间复杂的勾稽关系,为大模型优化微调提供数据层支持。图谱构建结果表明,GPT-4o mini能够成功从大量财务数据中提取丰富的知识五元组,并基于此构建财会知识图谱。最终,通过Neo4j技术实现知识图谱的可视化和查询功能。本研究证明大语言模型构建财会知识图谱具备可行性,能够显著提高知识图谱构建效率,为知识图谱的优化构建提供新思路,也为未来知识图谱融入大模型、优化模型性能提供基底数据支撑。展开更多
基金Project(2019yff0216502)supported by the National Key Research&Development Plan of Ministry of Science and Technology of ChinaProject(2021SK1020-4)supported by the Major Science and Technological Innovation Project of Hunan Province,China。
文摘Removing copper from nickel electrolysis anode solution has been a major keypoint in the nickel metallurgy industry.In this study,we proposed a novel process flow to promote removing copper from nickel electrolysis anode solution.A simulated nickel anode solution was designed,and static and dynamic adsorption experiments were conducted to determine the best of solution pH,adsorption time and temperature,resin dosage and particle size,and stirring speed.The optimal conditions were explored for copper removal from nickel electrolysis anode solution.Based on the optimal experimental conditions and the relevant experimental data,a novel process for copper removal from nickel electrolysis anodes was designed and verified.This novel process of copper removal from nickel electrolysis anodes was confirmed with nickel anolyte solution with nickel 50−60 g/L and copper 0.5 g/L.After finishing the novel process of copper removal,the nickel in the purified nickel anolyte became undetectable and copper concentration was 3 mg/L,the novel process of resin adsorption to remove copper from nickel anode solution through static and dynamic adsorptions has an efficacious copper removal.It is a beneficial supplement to traditional methods.
文摘生成式人工智能(AIGC)的快速崛起,掀起“人工智能+”赋能各行各业的浪潮,对会计领域的研究提供了新思路。由于财会行业要求高精确性与可解释性,而大语言模型存在着模型幻觉与决策黑箱,使得大模型应用于财会领域仍存在较大阻碍,构建财会知识图谱成为弥补大语言模型不足的有力工具。基于此,文章调用GPT-4o mini API,利用提示词工程(Prompt)与智能体(Agent)思路方法进行自动化知识抽取,构建中国会计准则知识图谱,从模式层提取财会实体与实体间复杂的勾稽关系,为大模型优化微调提供数据层支持。图谱构建结果表明,GPT-4o mini能够成功从大量财务数据中提取丰富的知识五元组,并基于此构建财会知识图谱。最终,通过Neo4j技术实现知识图谱的可视化和查询功能。本研究证明大语言模型构建财会知识图谱具备可行性,能够显著提高知识图谱构建效率,为知识图谱的优化构建提供新思路,也为未来知识图谱融入大模型、优化模型性能提供基底数据支撑。
