SUN Fujian, FENG Yuehai, LIU Siyu. Design of an intelligent prediction for arc length and process monitoring on the wire and plasma arc additive manufacturing[J]. Eletric Welding Machine2018, 48(6): 8-12.DOI:
WPAAM弧长智能预测与过程监控系统设计
摘要
在丝材等离子弧连续增材制造过程中
由于电弧加热过程中热积累导致实际堆敷层高度与设计堆敷层高度出现差异
造成弧长长度难以恒定;同时工艺参数的变化也会影响增材制造过程的稳定性。因此针对丝材等离子弧增材制造
设计了基于LabVIEW的等离子电弧弧长智能预测及过程监控系统。系统以USB4711A采集卡为核心
采用传感器实时采集增材过程中的堆敷电流和电弧电压等信息
实时显示工艺参数动态信息并分析数据
同时建立了基于BP神经网络的弧长智能预测模型
以实时监测增材制造过程中弧长的变化。试验表明
该过程监测系统能够实时监测堆敷过程
通过实时调用3-7-1结构的BP神经网络
能够实时预测电弧弧长变化
预测的误差范围为-1.6
1
.4 mm
达到预期要求。
Abstract
In the process of the wire and plasma arc additive manufacturing
due to the accumulation of heat in the heating process
the difference between the height of the actual compress layer and the designed compress layer makes the arc length difficult to be constant;at the same time
the change of process parameters will also affect the stability of the additive manufacturing process.Therefore
the intelligent prediction for arc length and monitoring system based on LabVIEW is designed for wire and plasma arc additive manufacturing process.The system takes the USB4711 A acquisition card as the core
and uses sensors to collect signals such as current and voltage on the wire and plasma arc additive manufacturing process
and displays real-time process parameters
process information and data analysis.Meanwhile
an intelligent prediction model of arc length based on BP neural network is established to monitor the change of arc length in the process of additive manufacturing.The experiments show that the process monitoring system can monitor the process of additive manufacturing process in real time.By real-time invoking the BP neural network of 3-7-1 structure
it can predict the change of the arc length in real time
and the error range of the prediction is from-1.6 mm to 1.4 mm