基于数字孪生的高保真推耙机虚拟调试系统

Virtual Commissioning System of a High-Fidelity Pusher Rake Based on Digital Twins

摘要:

为提前预知无人推耙机控制系统的风险,降低推耙机的无人化作业开发成本和调试周期,基于数字孪生技术虚实高度一致的特性,提出了一种针对全自动无人推耙机控制算法演练与仿真的虚拟调试方法。根据无人推耙机的实际运行控制机理,从机械传动,液压推动到电气控制建立了推耙机在物理层面多领域统一的数字孪生模型；根据无人推耙机清舱作业的感知机理,构建了激光雷达、惯导、里程计和RGB-D的多传感器融合数字孪生模型；基于虚幻引擎的实时渲染特性与AGX Dynamics多体动力学库,设计了高保真的数字孪生模型,并构建了孪生推耙机的虚拟控制系统。通过在港口岸桥应用于全自动无人推耙机的调试实例,展示了基于数字孪生的高保真推耙机虚拟调试方法的有效性和可行性,有效避免了调试开发周期过长的缺陷。

Abstract:

In order to predict the risk of the control system of the unmanned rake pushing machine in advance and reduce the development cost and debugging cycle of the unmanned operation of the pushing rake machine, based on the high consistency of virtual and real characteristics of digital twin technology, a virtual debugging method for the drill and simulation of the control algorithm of the fully automatic unmanned rake pushing machine was proposed. According to the actual operation and control mechanism of the unmanned rake pushing machine, a unified digital twin model of the rake pushing machine in multiple fields at the physical level was established from mechanical transmission, hydraulic propulsion to electrical control. According to the perception mechanism of the unmanned push rake machine cabin cleaning operation, a multi-sensor fusion digital twin model of lidar, inertial navigation, odometer and RGB-D was constructed. Based on the high real-time rendering characteristics of Unreal Engine and the AGX dynamics multi-body dynamics library, a high-fidelity virtual debugging system was designed, and a virtual control system of the twin rake pusher was constructed. Through the debugging example of the automatic unmanned rake pushing machine applied to the port quay crane, the effectiveness and feasibility of the virtual commissioning method of the high-fidelity unmanned rake pushing machine based on digital twin are demonstrated, and the defect of long debugging and development cycle is effectively avoided.