基于赋存状态多维语义特征向量化的锂矿石相似性评价方法研究

杨佳伟 ,  张康辉 ,  高希宇 ,  张晓亮 ,  余瑾彦 ,  朱阳戈

有色金属(选矿部分) ›› 2026, Vol. 0 ›› Issue (9) : 122 -131.

PDF (1725KB)
有色金属(选矿部分) ›› 2026, Vol. 0 ›› Issue (9) : 122 -131. DOI: 10.20239/j.issn.1671-9492.2026.09.014
矿石特征智能解析专栏

基于赋存状态多维语义特征向量化的锂矿石相似性评价方法研究

作者信息 +

A Method for Ore Similarity Evaluation Based on Multidimensional Semantic Feature Vectorization of Occurrence States

Author information +
文章历史 +
PDF (1766K)

摘要

随着矿产资源品位下降与矿石类型日益复杂,矿石赋存状态差异显著、矿物学特征维度多样,传统经验驱动的选矿工艺设计方法已难以满足资源高效利用的需求。针对复杂矿石赋存特征难以统一表达和定量比较的问题,本文提出基于赋存状态多维语义特征向量化的矿石相似性评价方法。首先,构建包含矿石类型、化学成分、矿物组成、嵌布特征及解离度等多维属性的结构化特征体系,并引入语义特征向量化方法,实现复杂矿物学信息的数字化与语义化表达。其次,基于词频-逆文本频率指数(Term frequency-inverse document frequency,TF-IDF)特征向量化与余弦相似度算法,建立矿石赋存特征的相似性评价模型,并通过人工设定的特征重要性权重调节机制提升关键赋存属性的区分能力。通过对待分析矿石样本与典型矿石样本库的语义相似度对比,实现不同赋存状态矿石在特征层面的相似性度量与筛选。研究结果表明,该方法可实现复杂矿石赋存特征的统一表征,能够从矿物学特征维度量化矿石间的相似程度;需注意矿石赋存特征相似并不直接等价于选矿工艺及分选指标一致,本研究可为后续开展选矿工艺知识迁移、工艺候选方案匹配与智能决策研究奠定数据与方法基础。

Abstract

With the continuous decline in mineral resource grades and the increasing complexity of ore types, significant differences in ore occurrence states and the high dimensionality of mineralogical characteristics have made traditional experience-driven beneficiation process design methods insufficient to meet the requirements of efficient resource utilization. To address the difficulty of unified representation and quantitative comparison of complex ore occurrence features, this study proposes an ore similarity evaluation method based on multidimensional semantic feature vectorization of occurrence states. Firstly, a structured feature system is constructed by integrating multiple attributes, including ore type, chemical composition, mineral assemblage, textural characteristics, and degree of liberation. Semantic feature vectorization techniques are then introduced to achieve digitalized and semantic representations of complex mineralogical information. Subsequently, an ore similarity evaluation model is established using term frequency-inverse document frequency (TF-IDF) feature vectorization and cosine similarity, and an empirically-defined feature importance weighting adjustment mechanism is incorporated to enhance the discriminative capability of key occurrence attributes. By comparing the semantic similarity between target ore samples and a database of representative ore samples, similarity measurement and screening of ores with different occurrence states at the feature level are realized. The results demonstrate that the proposed method can realize unified representation of complex ore occurrence features and quantitatively evaluate ore similarity from mineralogical perspectives. It should be noted that similar mineralogical occurrence features do not guarantee consistent beneficiation performance. This work provides a data and methodological foundation for subsequent research on beneficiation process knowledge transfer, candidate-scheme matching, and intelligent decision-making.

关键词

矿石赋存状态 / 多维语义特征 / 特征向量化 / 相似度分析

Key words

ore occurrence state / multidimensional semantic feature / feature vectorization / similarity analysis

引用本文

引用格式 ▾
杨佳伟,张康辉,高希宇,张晓亮,余瑾彦,朱阳戈. 基于赋存状态多维语义特征向量化的锂矿石相似性评价方法研究[J]. 有色金属(选矿部分), 2026, 0(9): 122-131 DOI:10.20239/j.issn.1671-9492.2026.09.014

登录浏览全文

4963

注册一个新账户 忘记密码

参考文献

[1]

俞柏. 战略性矿产资源供需形势分析与判断[J]. 中国金属通报, 2025(8): 11-13.

[2]

Yu Bai. Analysis and judgment of supply and demand situation of strategic mineral resources[J]. China Metal Bulletin, 2025(8): 11-13.

[3]

陈从喜. 战略性矿产资源开发利用研究进展与展望[J]. 中国矿业, 2025, 34(2): 4-19.

[4]

Chen Congxi. Progress and prospects in the development and utilization of strategic mineral resources[J]. China Mining Magazine, 2025, 34(2): 4-19.

[5]

李鹏, 刘艳, 唐晨宇, 等. 中国高科技矿产资源供应安全现状与保障路径[J]. 工程科学学报, 2025, 47(11): 1985-1998.

[6]

Li Peng, Liu Yan, Tang Chenyu, et al. Current status and safeguarding pathways of supply security for high-tech mineral resources in China[J]. Chinese Journal of Engineering, 2025, 47(11): 1985-1998.

[7]

王登红. 关键矿产的研究意义、矿种厘定、资源属性、找矿进展、存在问题及主攻方向[J]. 地质学报, 2019, 93(6): 1189-1209.

[8]

Wang Denghong. Research significance, mineral classification, resource attributes, exploration progress, existing problems, and principal directions of critical minerals[J]. Acta Geologica Sinica, 2019, 93(6): 1189-1209.

[9]

Zhironkina O, Zhironkin S. Technological and intellectual transition to mining 4. 0: a review[J]. Energies, 2023, 16(5): 1427.

[10]

苏妤芸, 苏秀珠, 刘全军, 等. 新疆某难处理金矿的金赋存状态研究[J]. 有色金属(选矿部分), 2025(10): 20-27.

[11]

Su Yuyun, Su Xiuzhu, Liu Quanjun, et al. Study on the occurrence state of gold in a refractory gold ore from Xinjiang[J]. Nonferrous Metals(Mineral Processing Section), 2025(10): 20-27.

[12]

孙传尧, 周俊武. 流程工业选矿过程智能优化制造发展战略[J]. 有色金属(选矿部分), 2019(5): 1-5.

[13]

Sun Chuanyao, Zhou Junwu. Development strategy of intelligent optimization manufacturing for process-industry mineral processing[J]. Nonferrous Metals(Mineral Processing Section), 2019(5): 1-5.

[14]

吕燕, 周家云, 惠博, 等. 吉林某铌钽多金属矿石中元素的赋存状态及对选矿工艺的影响[J]. 矿产综合利用, 2019(2): 83-88.

[15]

Lv Yan, Zhou Jiayun, Hui Bo, et al. Occurrence state of elements in a Nb-Ta polymetallic ore from Jilin and its influence on beneficiation technology[J]. Multipurpose Utilization of Mineral Resources, 2019(2): 83-88.

[16]

叶超, 罗政. 某低品位铌矿石工艺矿物学特征及可选性研究[J]. 中国矿山工程, 2021, 50(6): 87-91, 96.

[17]

Ye Chao, Luo Zheng. Study on technological mineralogical characteristics and processability of a low-grade niobium ore[J]. China Mine Engineering, 2021, 50(6): 87-91, 96.

[18]

张愿宁, 郭园园, 逯文辉, 等. 甘肃敦煌五一山钒矿矿石特征及钒赋存状态[J]. 甘肃科技, 2025, 41(7): 37-43.

[19]

Zhang Yuanning, Guo Yuanyuan, Lu Wenhui, et al. Ore characteristics and vanadium occurrence state of the Wuyishan vanadium deposit in Dunhuang, Gansu[J]. Gansu Science and Technology, 2025, 41(7): 37-43.

[20]

朱汇派, 费光春, 谭华, 等. 扫描电镜-能谱分析在四川西部某锂矿石铌钽元素赋存状态研究上的应用[J]. 有色金属(选矿部分), 2023(5): 16-21, 41.

[21]

Zhu Huipai, Fei Guangchun, Tan Hua, et al. Application of SEM-EDS to the study of Nb-Ta element occurrence in a lithium ore from western Sichuan[J]. Nonferrous Metals (Mineral Processing Section), 2023(5): 16-21, 41.

[22]

汤化明, 郎增瑞, 王宏铭, 等. 基于深度学习的金属矿物镜下嵌布粒度检测方法[J]. 地质科技通报, 2024, 43(3): 351-358.

[23]

Tang Huaming, Lang Zengrui, Wang Hongming, et al. Deep-learning-based method for measuring dissemination grain size of metallic minerals under microscope[J]. Bulletin of Geological Science and Technology, 2024, 43(3): 351-358.

[24]

蔡明明, 李光胜, 朱幸福, 等. 基于金矿岩芯的自动工艺矿物学[J]. 矿产综合利用, 2024, 45(3): 174-178.

[25]

Cai Mingming, Li Guangsheng, Zhu Xingfu, et al. Automated process mineralogy based on gold-ore drill core[J]. Multipurpose Utilization of Mineral Resources, 2024, 45(3): 174-178.

[26]

Yang Jun, Sun Chuanyao, Zhou Junwu, et al. Knowledge-inference-based intelligent decision making for nonferrous metal mineral-processing flowsheet design[J]. Minerals, 2025, 15(4): 374. DOI: 10.3390/min15040374.

[27]

王智宇, 王达, 邱昆峰, 等. 机器学习在矿物岩石地球化学大数据挖掘中的应用与展望[J]. 成都理工大学学报(自然科学版), 2025, 52(5): 844-858.

[28]

Wang Zhiyu, Wang Da, Qiu Kunfeng, et al. Application and prospect of machine learning in big-data mining of mineralogy and geochemistry[J]. Journal of Chengdu University of Technology (Natural Science Edition), 2025, 52(5): 844-858.

[29]

田妞, 黄宋魏, 和丽芳, 等. 深度学习在选矿领域的应用进展[J]. 化工矿物与加工, 2023, 52(8): 69-74, 82.

[30]

Tian Niu, Huang Songwei, He Lifang, et al. Application progress of deep learning in the field of mineral processing[J]. Industrial Minerals & Processing, 2023, 52(8): 69-74, 82.

[31]

Wang Qiang, Zhang Kai, Zou Guobing, et al. Advances in deep learning-based ore particle size detection: a review of methods, challenges, and trends[J]. MetaResource, 2025, 2(2): 83-104.

[32]

于潇然, 徐泉, 初延刚, 等. 基于相似性检测的选矿指标趋势监控系统[J]. 控制工程, 2020, 27(增刊1): 8.

[33]

Yu Xiaoran, Xu Quan, Chu Yangang, et al. A trend-monitoring system for mineral-processing indices based on similarity detection[J]. Control Engineering, 2020, 27(S1): 8.

[34]

于潇然. 选矿生产指标趋势监控与分析系统的设计与开发[D]. 沈阳: 东北大学, 2020.

[35]

Yu Xiaoran. Design and development of a trend-monitoring and analysis system for mineral-processing production indices[D]. Shenyang: Northeastern University, 2020.

[36]

Al-Obaydy W N I, Hashim H A, Najm Y A, et al. Document classification using term frequency-inverse document frequency and K-means clustering[J]. Indonesian Journal of Electrical Engineering and Computer Science, 2022, 27(3): 1517-1524.

[37]

Tang Xiao. A latent hidden Markov model for process data[J]. Psychometrika, 2024, 89(1): 205-240.

基金资助

地球深部探测与矿产资源勘查国家科技重大专项(2024ZD1002000)

AI Summary AI Mindmap
PDF (1725KB)

0

访问

0

被引

详细

导航
相关文章

AI思维导图

/

〈 〉