Progress of digital-intelligent coal preparation technology based on big data and artificial intelligence
基于大数据与人工智能的数智化选煤技术进展研究

摘要

为梳理大数据与人工智能技术推动选煤厂由自动化向数智化转型的技术路径，结合近年公开研究和智慧选煤数智管控工程资料，从多源数据采集、数据治理与融合、数字孪生、机器学习、专家知识、深度学习及智能控制等方面分析技术进展。工程案例表明，通过构建基础信息模型、专家算法库和场景应用系统，可实现设备健康诊断、原煤分选、磁选液位、压滤集控及辅助系统智能管理；项目最终实现精煤回收率提升1.1%、介耗降低16%、日均停机时间减少0.55 h、故障诊断准确率93%，年直接经济效益341.9万元。研究认为，数智化选煤正由数据汇聚与可视化向知识图谱、时序深度学习、数字孪生闭环、强化学习和行业大模型演进，数据标准、模型泛化、控制安全、系统开放性及全生命周期运维是规模化推广的关键。

Abstract

To review the technological pathway by which big data and artificial intelligence promote the transformation of coal preparation plants from automation to digital-intelligence, recent public studies and an engineering smart coal-preparation project are analyzed from the perspectives of multi-source data acquisition, data governance and fusion, digital twins, machine learning, expert knowledge, deep learning and intelligent control. The engineering case establishes a basic information model, an expert algorithm library and multiple scenario applications for equipment health diagnosis, raw-coal separation, magnetic-separator level control, filter-press control and auxiliary-system management. Final acceptance reported a 1.1% increase in clean-coal recovery, a 16% reduction in medium consumption, a 0.55 h/day reduction in downtime, 93% fault-diagnosis accuracy and annual direct economic benefits of CNY 3.419 million. Digital-intelligent coal preparation is evolving toward knowledge graphs, time-series deep learning, closed-loop digital twins, reinforcement learning and domain foundation models; data standards, model generalization, control safety, system openness and lifecycle operation remain key issues.