Applying Fuzzy Inference System and CUSUM Algorithm to Process Control of Chip Corner Arcs of Multi-Layer Ceramic Capacitors

摘要

The increasing demand for smaller, lighter, thinner, and cheaper electronic products has made MLCC (Multi-Layer Ceramic Capacitors) an indispensable component. However, the production of MLCC could easily fail owing to many reasons, one of which being the difficult inspection and process control of tiny chip comer arc segments. Since the chip chamfer operations are processed in lots, the whole lot of products could be easily damaged if operators do not pay enough attention. The current human eye inspection method can only sift out defective products after production; it cannot prevent their occurrence beforehand. In this research, a computer vision system is proposed to inspect and control the chip corner arcs of MLCC automatically. The break points of the capacitor chip corner arcs are determined by the CUSUM (Cumulative Sum) control chart method and the features of comer arcs are extracted based on the geometric properties of the chips in the computer vision system. Then, the ANFIS (Adaptive Neuro-Fuzzy Inference System) is applied to find quality evaluation values of the chip corner arcs and proceed to the process control in the chip chamfer operations for ensuring the chip quality in the manufacturing process.