이차 시간 지연 모델 기반 시행착오 튜닝 규칙과 영상처리 기반 슬러리 체적 측정 방법 = Trial-and-Error Tuning Method Based on the Second-Order Plus Time Delay Model (SOPTD) and Image Processing-based Slurry Volume Measurement

다국어 초록 (Multilingual Abstract)

To achieve desired control performance, determining optimal tuning parameters is crucial. When the process model differs from actual dynamics, calculated PID tuning parameters often fail to achieve the desired performance. In such cases, the desired control performance can be achieved by manipulating the PID controller tuning parameters based on the current control results. However, in complex control structures (e.g., cascade structure), multiple PID controllers are involved, which increases the number of tuning parameters and makes the impact of each parameter on control performance difficult to understand.
This study proposes a method to enhance control performance by adjusting parameters of a Second Order Plus Time Delay (SOPTD) model. With this method, the number of parameters to be manipulated is fixed to the number of model parameters, regardless of the control structure’s complexity. Moreover, it links the current control results with the physical meanings of the SOPTD model parameters, making it possible to intuitively tune the controller.

Slurry is a pivotal intermediate in the production of phenolic foam, a core material for thermal insulation. Consequently, early defect detection during the slurry phase is critical for minimizing resource wastage and mitigating economic losses. However, there is currently a lack of standardized techniques for the quantitative measurement of slurry volume in industrial settings. Conventional methods relying on manual sampling suffer from low reproducibility and poor real-time capabilities. Furthermore, existing automated alternatives often face limitations such as high equipment costs and sensitivity to environmental factors like illumination changes and surface reflections, which make their on-site applicability difficult.
To address these limitations, this study proposes a slurry volume estimation method utilizing a simple laser source and camera-based image processing. The proposed method extracts the edges of the slurry surface by exploiting the intensity difference between the laser-illuminated line and non-illuminated regions of the slurry surface. Subsequently, a calibration model converts pixel coordinates into actual physical dimensions, and numerical integration is applied to quantitatively calculate the cross-sectional area and ultimately estimate the total volume. Since this approach enables precise volume estimation without the need for high-cost instrumentation, it significantly enhances applicability within manufacturing environments.

목차 (Table of Contents)

• 서론 1
• 1. 이차 시간 지연 모델 기반의 시행착오 튜닝 규칙 4
• 1.1 동기 4
• 1.2 이차 시간 지연 모델 기반의 튜닝 절차 및 파라미터 설정 5
• 1.3 이차 시간 지연 모델 기반의 시행착오 튜닝 12
• 1.3.1 시행착오 튜닝 절차 12
• 1.3.2 초기 튜닝 파라미터 14
• 1.3.3 파라미터 조작 범위 15
• 1.4 모사 연구 16
• 1.5 액위 제어 실험 29
• 1.6 결론 34
• 2. 영상처리 기반 슬러리 체적 측정 방법 35
• 2.1 동기 35
• 2.2 영상처리 기반 슬러리 체적 산출 절차 36
• 2.2.1 영상 획득 36
• 2.2.2 엣지 검출 36
• 2.2.3 상대 높이 및 폭 산출 38
• 2.2.4 보정 모델 개발 40
• 2.2.5 체적 산출 40
• 2.2.6 불량 판정 41
• 2.3 슬러리 체적 산출을 위한 영상처리 및 보정 모델 개발 42
• 2.4 결론 48
• 참고문헌 49
• 영어 초록 52