掘進機截割頭智能控制系統(tǒng)及其關鍵技術研究

打開文本圖片集

中圖分類號：TD421.5 文獻標志碼：A 文章編號：1003-5168（2025）08-0053-04

DOI：10.19968/j.cnki.hnkj.1003-5168.2025.08.009

Abstract： [Purposes] This paper aims to solve the problem of low effciency and high energy consumption caused by insufficient intelligent control of the roadheader cutting head，thus improving its control accuracy.[Methods] A combined system of reinforcement learning and model predictive control was built.The Q-learning algorithm was used to optimize cutting parameters，a Lagrange-Euler dynamics predictive controllr was designed，and a multi-objective model with adjustable weights （O.2＼～O.8） was created.[Findings] In 55 MPa rock conditions，the system's tunneling efficiency increased by 1 8 . 1 % （to （202 3 . 8 5 m / h ）， energy consumption decreased by 1 7 . 4 % （to ）， trajectory error was reduced to 2.8 mm，and control signal fluctuations dropped by 4 9 . 2 % .[Conclusions] This method improves the energy efficiency of the cutting head through algorithm optimization and provides key technical support for smart tunneling equipment.

Keywords： roadheader; cutting head; reinforcement learning

0 引言

掘進機是煤礦開采、隧道施工等領域的重要裝備，其工作性能直接影響掘進效率、施工進度與成本[]。（剩余5201字）