军事知识服务平台
露天煤矿破碎站选型本地知识库系统构建方法
Construction method of local knowledge base systems for crusher station selection in open-pit coal mines
摘要 为解决露天矿破碎站选型中知识碎片化、经验依赖性强及决策效率低等问题,构建了基于本地化部署的露天矿破碎站智能选型知识库系统.通过系统梳理选型流程与多源异构资料,提出了包含数据预处理、语义向量化、混合检索与答案生成的构建方法.该方法依托 Python 与LangChain 工具实现选型参数标准化处理与文本精准分割;采用 BGE-M3 模型生成文本语义向量并存入 Chroma 数据库;同时基于 MaxKB 框架集成语义检索与关键词检索的混合检索机制;最终通过DeepSeek-R1 大模型在检索增强生成框架下生成专业可解释的选型决策建议.应用结果表明,该系统能精准响应参数查询、阐释专业概念、完成多约束方案筛选及技术经济分析,显著提升了露天矿破碎站选型工作的科学性与实用性.
Abstract:
To address issues such as fragmented knowledge,strong reliance on experience,and low decision-making efficiency in the selection of crusher stations in open-pit mines,an intelligent selection knowledge base system based on local deployment was constructed.By systematically sorting out the selection workflow and multi-source heterogeneous materials,a construction method comprising data preprocessing,semantic vectorization,hybrid retrieval,and answer generation was proposed.This method uses Python and LangChain tools to achieve standardized processing of selection parameters and precise text segmentation.The BGE-M3 model is used to generate text semantic vectors,which are stored in a Chroma database.Meanwhile,a hybrid retrieval mechanism integrating semantic retrieval and keyword retrieval is built based on the MaxKB framework.Finally,the DeepSeek-R1 large model generates professional and interpretable selection decision recommendations under the retrieval-augmented generation framework.Application results show that the system can accurately respond to parameter queries,explain professional concepts,complete multi-constraint scheme screening and technical-economic analysis,and significantly improve the scientificity and practicality of crusher station selection in open-pit mines.
刊 名 煤炭工程 ISTICPKU
年,卷(期): 2026, 58(9)
分类号: TD824
机标分类号: TP391 TD824 F270.7
在线出版日期: 2026年10月9日