Optimization of control strategy of microwave drying for apple based on online volatile measurement with electronic nose

Abstract: Apple is widely cultivated in China for its characteristic flavor. In order to prolong the shelf life of harvested apple, the use of the method of dehydration drying to preserve product is a common way. Since the early study in 1920s, more than 300 volatile compounds have been identified as aroma substances from various cultivars of apples. Within these numerous compounds, only a few have been determined to have a decisive impact on the sensory quality. However, only a few exhaustive studies of apple aroma during the drying process have been published; moreover, the present analyses are all restricted to offline detection of apple aroma volatiles and no online measurement has been reported. Although the offline volatiles detection can provide some useful information, it is inconvenient and time-consuming. In addition, the methods are all impossible to provide control parameters to optimize drying conditions in real time and in situ to improve volatiles control with online measurement. Compared with other drying methods, microwave drying is used more extensively due to its advantages of high efficiency and short time-consumption. However, microstructure of the object may be destroyed and the surface may be burnt under improper drying parameters, leading to the emission of corresponding volatiles. During the drying process, an electronic nose (zNoseTM) was used to carry out online detection of the smell spectrum, where four significant peaks could be extracted. Peak 1 and Peak 2, i.e. "natural peak" could be detected from both fresh and burnt apples. Peak 3 and Peak 4, i.e. "burning peak" appeared only in the spectrum of burnt apple. Apple cubes were first dried at three fixed temperatures (60, 70 and 80℃). At 60℃, "natural peak" and "burning peak" were kept small in the whole drying process. At 80℃, higher "natural peak" indicated more loss of aroma volatiles, and higher "burning peak" indicated that more burnt appeared. Fuzzy logic controller was designed with the input of Peak 2, Peak 3 and the output of the magnetron power, so as to adjust the object temperature immediately and properly. Linear temperature control scheme was designed to omit the electronic nose in drying process based on the temperature curve generated under fuzzy logic control. Several parameters (including drying time, energy consumption, sensory evaluation, vitamin C and color) were measured to evaluate the drying strategies. The results showed that the highest drying temperature produced the worst product quality, but the best quality was not achieved at the lowest temperature due to volatiles loss over the longer drying time. Instead, the middle drying temperature resulted in a better product quality. Drying at the lowest temperature cost the most time and energy, although the mean power rate was low. Hence, high and low temperatures should be avoided in microwave drying when volatiles retention, burning avoidance, time and energy efficiency are all considered. Based on these considerations, a fuzzy logic controller was developed. This new control strategy successfully improved the volatiles retention and burning avoidance with acceptable time and energy consumption. A linear control method was further attempted to imitate the fuzzy logic control without the assistance of electronic nose. The control effect of linear control was comparable to that of fuzzy logic control. However, the relationship of aroma with respect to volatiles was not investigated in this study. A volatile may be or may not be an aroma compound, and hence further research is recommended to illustrate their relationships. Moreover, as the electronic nose adopts a short column, the separation of volatile compounds might not be complete. One peak may represent many chemicals. The compromise is that the speed is dramatically increased, making the fast and online detection of volatiles possible. Performance of sensory evaluation, vitamin C and color under linear temperature control are the best. Moreover, the omission of the expensive electronic nose makes the strategy convenient for industrial application.