Development Of Near Infrared On-line Detection System For Liquid Food Quality

Liquid food is one of the most important kinds of food and plays an important role in people's daily life.In the process of food processing,it is of great significance to obtain the information of key quality indicators of products in time for control of production process and product quality.This study designed a set of online near infrared liquid food quality testing system,which can be used as a laboratory liquid food quality online testing and grading platform.It can also be used in educating food science professional undergraduate and graduate students to understand modern intelligent online detection technology development trend of food processing production line,in designing test-analysis-feedback-control of closed loop process parameters adjustment simulation system,so as to maximize resource utilization,improve product quality and yield,and maximize production efficiency.The main contents of this study are as follows:(1)Through analyzing performance and function of the online near infrared detection system,overall hardware and software design of the system was carried out,several sampling methods of the liquid online detection were discussed,and transmission online detection scheme was finally determined.In order to verify feasibility of the design scheme,simulated dynamic spectra of four concentrations(20%,30%,40% and 50%)alcohol solution were collected and compared.The results show that the maximum standard deviation of the spectra at four concentrations is less than 0.002,which proves that the design scheme is reasonable and feasible.(2)Hardware design and selection of main components of the online near infrared liquid food quality testing system were completed.The structure forms including feeding mixing module,detection module and control module were designed,and the main non-standard parts were designed and outsourced.A near-infrared detection module including power supply,light source,optical fiber,quartz rod and circulation pool was designed to realize the on-line detection function of liquid food quality.Additionally,the selection of control platform,sensors and motion execution parts and the design of space layout,pipe docking and electrical wiring of the detection system were completed.(3)Design of the software part of the online near infrared liquid food quality inspection system was completed,which is divided into the upper computer software,the lower computer software and the upper and lower computer joint collaboration software.The upper computer software includes functions such as communication,spectrum acquisition,spectrum display and model prediction.The lower computer software includes communication setting,temperature calculation,flow calculation,liquid level detection and touch screen point control and other functions.The joint collaboration software includes lower computer interface,upper computer instruction,feeding mixing,online detection and control,animation simulation and other functions.It realized and completed the control and unified installation and debugging of execution components,sensors,near-infrared spectrometer and other components.(4)Performance of the online NIR liquid food quality detection system was tested.The light source stability,baseline stability and absorbance of the NIR spectrometer under static/dynamic detection state were tested respectively.The results show that the on-line NIR liquid food quality inspection system is stable and reliable,and has good continuous working ability.(5)Taking alcohol solution as the detection object,the application of alcohol concentration online detection was carried out.In the concentration range of 10% ?60%,26 test samples of 1 L were prepared and their near-infrared spectra were detected online.Without pretreatment,the correction set and prediction set were divided according to the ratio of 3:2,and three correction models were established by SWR,PCR and PLSR.The results showed that the three models all have good prediction accuracy,but the SWR model has a determination coefficient of the calibration set and the prediction set sample as high as 0.9991,which is suspected of over-fitting.PCR and PLSR models are more reliable and stable.The established calibration model was used to verify actual detection ability of the online NIR liquid food quality detection system under normal working conditions.The system automatically mixes 50%,40% and 30%alcohol solution by means of automatic feeding and mixing.The automatic detection results of the system showed that both PCR and PLSR models yield good prediction results,and the difference between the predicted concentration and the true value is less than 0.5%(V/V).