Optimizing UPVC profile production using adaptive neuro-fuzzy inference system

This paper explores the implementation of an Adaptive Neuro-Fuzzy Inference System to optimize Unplasticized Polyvinyl Chloride profile production. Given the intrinsic complexities of polymer extrusion, such as maintaining consistent quality amidst varying raw material properties and environmental conditions, traditional control methods like Proportional-Integral-Derivative (PID) controllers often need to catch up in adapting to non-linear production environments. This study proposes an ANFIS-based control system that leverages the adaptability of neural networks and the human-like reasoning of fuzzy logic to address these challenges effectively. The system uses real-world data from expert operators to refine its control parameters, ensuring accurate regulation of critical variables such as extrusion temperature, feed material composition, and ambient conditions. By testing the system against traditional control methods, the research demonstrates significant improvements in quality control, efficiency, and waste reduction. The ANFIS system’s adaptability allows it to minimize errors and adjust dynamically to variations in production conditions, outperforming traditional PID systems in maintaining consistent quality. Furthermore, the study outlines the implications of this research for intelligent manufacturing within Industry 4.0 frameworks, highlighting how ANFIS can integrate seamlessly with existing manufacturing systems to optimize production. This paper ultimately contributes to the growing literature advocating for intelligent control systems in polymer extrusion. It underscores their potential to transform manufacturing efficiency and quality in Unplasticized Polyvinyl Chloride production.