Abstract: The mechanized picking of fresh tea leaves gains the advantages of high efficiency, low cost and laborsaving, which is favorable to large scaled and mechanized tea production in the future.However, due to the limitations of existing picking equipment, machine picked fresh leaves are often the mixtures of old and tender leaves with different grades, which seriously affects the sensory quality and economic benefits of tea products.Therefore, the classification of collected fresh leaves is required before processing.In order to avoid uneven quality of fresh leaves and thereby achieve mechanized plucking and processing for high quality tea leaves, the working settings of a novel fresh leaves classifier (ZL 201320064553.4) were systematically investigated in this study.Using mechanically plucked fresh leaves of Longjing 43 tea variety as experimental material, we compared the classification efficiency using different combinations of grading sieve plates (i.e., a variety of hole types such as round and ellipse holes) and vibration frequency (40, 50 and 60 Hz).We applied the tea processing workflow including spreading out fresh leaves, fixation, cooling, rolling, initial baking, cooling resurgence, first frying, cooling, second frying, and baking sufficiently.Hand picked leaves(tenderness from one bud with one leaf to one bud with 2 leaves), mechanically picked leaves and mechanically picked leaves with classification were processed to the coiled type green teas, respectively according to the steps described above.Then we systematically evaluated the sensory quality and comprehensive benefit of obtained tea products.The results showed that when the novel fresh leaves classifier was applied to classify the mechanically plucked fresh leaves of Longjing 43 tea variety, the sieve plate #1 with ellipse holes (the length of the short axis was 0.6cm, and the length of the long axis was 2.0 cm) resulted in better grading for small or elongated fresh leaves such as one leaf and debris, and the average proportion of such fresh leaves in the outlet was 75.7%.The sieve plate #2 with round holes(the diameter was 2.0 cm) was suitable for the classification of the fresh leavesthat were“one bud with one leaf”(one of the target products of grading operation), and the average proportion of this kind of fresh leaves in the outlet was 70.22%.The sieve plate #3 with round hole(the diameter was 2.6 cm) was suitable to classify the fresh leaves that were “one bud with 2 leaves”(one of the target products of grading operation), and the average ratio of such leaves in the outlet was 67.89%.Moreover, by combination of these sieve plates, the finally obtained fresh leaves after grading showed the tenderness predominantly equal to or higher than“one bud with 3 leaves”, and the average percentage of this kind of fresh leaves in the outlet was 79.53%.The selection of vibration frequency was dependent on the plate material and classification effect.When the vibration frequency was low, it not only affected the operation efficiency of classification, but also weakened the effect of grading since the hanging leaf would block the meshes.When the vibration frequency was too high, despite good fluency, the“flow” rate of fresh leaves was so high that some of the fresh leaves would be thrown out, resulting in classification efficiency reduction.Furthermore, the high vibration frequency was more likely to form plate resonance and break the cam bearing.So a vibration frequency of 50 Hz was appropriate to achieve the satisfactory grading and smooth operation with less resonance, and thus avoid the damage to sieve plates.Sensory evaluation results showed that the tea products made from the graded leaves had significantly improved sensory quality compared with those without classification, and the average sensory score was improved by 2.55, especially in the flavor.Compared with the products made from hand picked leaves, the tea sample of Grade 1 was similar in appearance, aroma and taste.In this study, by machine grading of fresh leaves and processing into coiled type green tea, the finished tea production was increased by 18.1%, and the average yield was increased by 2.74%; in addition, the comprehensive profit was remarkably improved by 1.22 yuan per kg fresh leaves, compared with that obtained by direct processing of machine picked leaves without grading.