研究目的
To design and implement an optical monitoring system for assessing the amount of fouling material remaining in process tanks and predicting the required cleaning time in clean-in-place (CIP) processes.
研究成果
The developed optical monitoring system and image processing procedure effectively assessed fouling levels and predicted cleaning times. The NARX neural network demonstrated high accuracy in cleaning time prediction, suggesting potential for resource and time efficiency improvements in food manufacturing.
研究不足
The study was conducted at a laboratory scale, and the optical hardware needs further development for industrial applications. The system's performance with different types of fouling materials was not extensively tested.
1：Experimental Design and Method Selection:
The study utilized a two-tank system to simulate industrial cleaning processes, employing UV-induced fluorescence for monitoring. Image processing techniques were developed to assess fouling levels.
2：Sample Selection and Data Sources:
White chocolate was used as a fouling medium due to its representative properties for food fouling. Digital images were collected during cleaning cycles.
3：List of Experimental Equipment and Materials:
A Nikon D330 DSLR camera, a 10–20 mm F4-5.6 EX DC HSM wide-angle zoom lens, and a dual 18 W 370 nm fluorescent lamp were used for image acquisition and UV illumination.
4：6 EX DC HSM wide-angle zoom lens, and a dual 18 W 370 nm fluorescent lamp were used for image acquisition and UV illumination.
Experimental Procedures and Operational Workflow:
4. Experimental Procedures and Operational Workflow: Fouling was manually applied to the process tank, and cleaning cycles were conducted using heated water. Images were captured at 5-second intervals during cleaning.
5：Data Analysis Methods:
Image processing involved baseline subtraction, green channel extraction, and thresholding to assess fouling. A NARX neural network was used for cleaning time prediction.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
