Application Research of Intelligent PID Parameter Optimization Algorithm for XY Linear Motor Motion Platform in Wire Bonder

In the domain of semiconductor packaging, the efficacy and efficiency of wire bonder hold paramount importance for product integrity and manufacturing productivity. The XY linear motor motion platform, constituting a pivotal element within wire bonder, intricately dictates the precision and stability of the bonding procedure. While conventional PID controllers offer a degree of control over XY linear motor motion platforms, their parameter optimization often necessitates extensive empirical iterations and frequently falls short of achieving optimal control outcomes. Confronted with sizable quantities of inconsistently performing motors and intricate production settings, this predicament substantially escalates production expenditures while concurrently diminishing operational efficiency. Herein, a novel approach utilizing BP neural networks for real-time PID parameter auto-tuning is proposed, markedly enhancing tuning efficiency vis-à-vis manual methodologies. Addressing the inherent challenge of initializing weights and thresholds in BP neural networks, this study advocates for the adoption of the artificial fish swarm algorithm to optimize the initial weight selection, culminating in the design of the AFSA-BP-PID controller. Subsequently, leveraging the dynamic model of Permanent Magnet Linear Synchronous Motors (PMLSM), a Matlab/Simulink simulation framework is constructed to juxtapose and scrutinize the dynamic error and tuning duration of PID, BP-PID, and AFSA-BP-PID controllers. The findings evince the efficacy of the AFSA-BP-PID control algorithm in augmenting the dynamic performance of permanent magnet synchronous linear motors while concurrently abbreviating system tuning intervals.