【摘要】 针对芯片固晶机高速高频运动导致其键合头振动冲击大、振动衰减时间长，同时原有控制系统无法检测和抑制键合头末端振动，键合精度难以长时间保持的难题，提出基于优化输入整形的抑振方法。分析固晶机键合头的运动和振动特征，提出基于输入整形的振动抑制方法；构建二阶系统仿真分析模型，基于粒子群算法对输入整形器的种类、脉冲幅值和时滞等参数寻优，简化构建流程，提升抑振效果；搭建固晶机单轴运动系统，对基于粒子群算法优化的输入整形抑振方法进行试验验证，结果表明该方法将使系统残余振动最大幅值降低了71.3%,运动开始到加速度整定至稳态时间减少了34.0%。更多还原

【Abstract】 High-speed and high-frequency motion in die bonding machines causes the bonding head intensive vibration. Existing control systems cannot detect or suppress the bonding head vibration, leading to difficulty in maintaining long-term bonding precision. An optimized input shaping method for vibration suppression was proposed, which utilized particle swarm optimization(PSO) to automate the adjustment of input shaper(IS) parameters. These parameters, including the pulse type, number, amplitude, and time delay, were optimized based on a second-order simulation dynamic model of the bonding head. The approach significantly reduces design complexity compared to the traditional manual tuning. Experimental results on a single-axis system demonstrate a 71.3% reduction in residual vibration amplitude and a 34% decrease in stabilization time.更多还原