| 摘要： 本研究是利用碰擊試驗的方式，對水果進行破壞性果肉碰擊損傷及非破壞性成熟度分級檢測，首先利用一改製的單擺式衝擊試驗機進行平板（曲率半徑無限大）衝錘探頭碰擊蘋果，改變不同的動量及儲藏溫度，由力量-時間波形圖得到碰擊時的最大撞擊力Fp及撞擊接觸時間tc，碰擊後量測蘋果的損傷深度及損傷體積，並計算蘋果損傷的吸收能量；接著使用保麗龍網、發泡棉及瓦楞紙三種常見的緩衝包裝材料，探討其緩衝效果以作為水果選擇緩衝包裝材料的參考依據；進一步探討不同曲率半徑的衝錘探頭，在不同的動量及儲藏溫度下對蘋果碰擊損傷的影響。其次自行設計組裝一套番石榴自由落體碰擊檢測裝置和一套鋼珠自由掉落碰擊檢測裝置，對番石榴進行成熟度檢測，藉由碰擊試驗得到力量-時間波形圖，選擇有顯著性的碰擊參數，使用鑑別分析及類神經網路分析來判別番石榴成熟度分級的準確率。 實驗分析結果顯示，使用平板衝錘探頭進行碰擊試驗對蘋果所造成的碰擊損傷可得以下結論，在無緩衝材料下，當動量越大時，最大碰擊力Fp就越高，蘋果損傷深度及損傷體積也隨之增大，且蘋果吸收能量E2隨之增加；在有發泡棉、保麗龍網和瓦楞紙緩衝材料下，以發泡棉作為緩衝材料時，蘋果吸收能量最高，緩衝材料吸收能量最低，緩衝效果較其他兩緩衝材料差。在4℃冷藏下，使用瓦楞紙做為緩衝材料時緩衝效果最好，在三種動量都找不到碰擊損傷；而在室溫下，在動量10°時，使用保麗龍網當緩衝材料時有較佳的緩衝效果。最大碰擊力Fp隨著衝錘探頭曲率半徑的增加而提高，而碰擊接觸時間tc隨著衝錘探頭曲率半徑的增大而減少，曲率半徑R32衝錘探頭的損傷體積比曲率半徑R24和R40衝錘探頭大，應是R32的曲率半徑和蘋果的半徑相近緣故。 在番石榴自由落體檢測番石榴成熟度試驗中，取採摘後時間第一、三、五、七天的前兩個彈跳回波資料做分析，以彈性常數Ki作為分級指標，使用三個分類顯著參數，進行鑑別分析後其分級準確率為77.4%。；以採摘後時間作為分級指標時，使用五個分類顯著參數，進行鑑別分析後其分級準確率為84.21%。取採摘後時間第一、三、五天的前三個彈跳回波資料做分析，以彈性常數Ki作為分級指標時，使用四個分類顯著指標，進行鑑別分析後其分級準確率為72.69%。；以採摘後時間作為分級指標時，使用五個分類顯著參數，進行鑑別分析後其分級準確率為93.76%。 在番石榴鋼珠彈跳檢測番石榴成熟度的試驗中，以彈性常數Ki作為分級指標時，使用四個分類顯著指標，進行類神經網路分級，當訓練1500次後其分級準確率為74.51%；訓練2000次後其分級準確率為75.49%。以採摘後時間第一、三、五天作為分級指標時，使用四個分類顯著指標，進行類神經網路分級，訓練1500次後其分級準確率為69.23%；訓練2000次後其分級準確率為74.34%。由壓縮試驗結果可驗證輕微碰擊檢測番石榴成熟度分級是一種非破壞性的實驗。A manner of impact test was used in this study, the volume of bruise damage was estimated in destructive test and grade the maturity of fruits in non-destructive test. First of all, the apples were impacted by a restructured simple pendulum device with plate hammer (infinite of curvature radius). The impact test for apple was conduct for different the rise angle and storage temperature. The peak force Fp and the contact time tc from the force-time waveform graph could be obtained, the depth and volume of bruise damage would be measured and absorbed energy of the apple bruise damage would be also calculated. Three kinds of packing cushion, styrofoam net, cotton and corrugated paper were used to investigate the effect of cushion. The result could be used as a basis of the selection of packing cushion for fruit. The effect of pendulum hammer with different curvature radius was explored further in different angles and temperatures. Furthermore, self-design a fruit a drop freely device and a steel ball drop freely device were estimated maturity classification for the guava. According to Step Regression Analysis, the significantly different of parameters from force-time waveform graph of the impact experiment were selected. The maturity classification of guava was determined by the Discriminate Analysis and Artificial Neural Network. The results from the impact experiment showed that the impact bruise of apple used plate pendulum hammer can get the conclusions as following. With the non-cushion material, when the rise angle was raised increase, the peak force would be more increase. Not only the bruise damage depth and bruise volume of apple increased, but also the absorbed energy of apple bruises damage volume increased. However, the impact time was not significantly different at room temperature, but the impact time was significantly different between 10° and 20° angle at 4 ℃ temperature. For three kinds of packing cushion material, there was the largest absorbed energy of apple and lowest absorbed energy of the cushion material for the cotton, the buffer effect of the cotton was poor than the other cushion material. It was not found any damage for the corrugated paper used in anyone rise angle and have the best buffer effect at 4 ℃ temperature, but there was better buffer effect for the styrofoam net used in 10° rise angle at room temperature. With the curvature radius of pendulum hammer increasesd, the peak force Fp increase and contact time tc decreased. The bruise damage volume of apple show that the R32 of curvature radius pendulum probe was higher than the R24 and the R40 of curvature radius, it is the reason that the R32 of curvature radius is similar with that of the apple. In the guava drop freely experiment of maturity classification, the data of former two rebound waves in the 1st 、3rd、5th and 7th of the postharvest time were obtained. The elasticity constant Ki was used as a classification index, three significantly different parameters were selected to conduct the discriminate analysis. The accuracy of maturity classification was 77.4 %. When the post-harvest time was used as a classification index, five significantly different parameters were selected to conduct the discriminate analysis. The accuracy of maturity classification was 84.21%. The data of former three rebound waves in the 1st、3rd and 5th of the postharvest time were obtained. The elasticity constant Ki was used as a classification index, four significantly different parameters were selected to conduct the discriminate analysis. The accuracy of maturity classification was 72.6 %. When the post-harvest time was used as a classification index, five significantly different parameters were selected to conduct the discriminate analysis. The accuracy of maturity classification was 93.76 %. In the experiment of the steel ball drop freely to estimate maturity of the guava, The elasticity constant Ki was used as a classification index, four significantly different parameters were selected to conduct the artificial neural network. The accuracy of maturity classification was 74.51 % when artificial neural network was trained 1500 times and 75.49 % when artificial neural network was trained 2000 times. While the post-harvest time , 1st、3rd and 5th times, were used as a classification index, , four significantly different parameters were selected to conduct the artificial neural network. The accuracy of maturity classification was 69.23 % when artificial neural network was trained 1500 times and 74.34 % when artificial neural network was trained 2000 times. It could be approved a non-destructive detection for slight impact test from the compression test of guava. |
頁面載入中請稍後...