基于TiDE-PatchTST模型的柑橘冷藏效率时序预测模型优化

柑橘低温贮藏过程中果实温度波动是引发果品品质安全风险与增加制冷能耗的关键因素,同时果品品质与制冷能耗也是评判柑橘冷藏效率的重要评价指标,实现两者动态预测可为科学预知与精准优化柑橘冷藏效率提供可靠支持。本文提出一种基于PatchTST的柑橘冷藏效率时序预测模型。首先,基于自注意力机制和独立预测方法(Channel independent,CI)构建基础PatchTST模型;其次,通过融合基础PatchTST模型与TiDE模型中的协变量特征提取模块,实现对多元时序数据集中全部序列的特征提取,并有效改进模型预测精度;最后,基于皮尔森相关性分析方法定量分析冷库制冷参数与能耗、柑橘温度的相关性,确定TiDE-PatchTST模型输入参数,并基于5000组实验数据实现多种模型训练与测试,对比验证TiDE-PatchTST模型的准确性与优越性。结果表明,基于TiDE-PatchTST模型的冷库能耗预测值与实验值平均绝对误差(MAE)和均方根误差(RMSE)分别为3.645 W·h和10.421 W·h,柑橘温度预测值与实验值的MAE和RMSE分别为0.034℃和0.042℃,相比Transformer模型,能耗预测的MAE和RMSE最高分别下降41.43%和39.27%,柑橘温度预测的MAE和RMSE最高分别下降46.03%和28.81%。本研究可为柑橘冷藏过程温度波动与能耗动态感知与优化调控等提供可靠方法支持与参考。

Preservation efficiency of new cryoprotectant used for Acidithiobacillus ferrooxidans in liquid nitrogen

The efficiency of a new cryoprotectant,GP,for the preservation of Acidithiobacillus ferrooxidans（A.ferrooxidans） strain DC in liquid nitrogen was investigated.The optimal concentration of this new cryoprotectant for the maximal viable cell recovery and the highest ferrous ion oxidation activity was determined.The results show that 30%（volume fraction） GP is optimal for the cryopreservation with 84.4% of cells surviving,completely oxidizing ferrous ions within 120 h,and growing to a final density of 5.8×107 cell/mL after 6 d in the culture.Furthermore,the optimal residual GP concentration for viable cell recovery after culture of thawed cells in 9K medium for 6 d is 0.6%（volume fraction）.At this concentration,strain DC completely oxidizes ferrous ions within 108 h and grows to a final cell density of 6.8×107 mL-1.Thus,GP is a simple,effective cryoprotectant for the preservation of A.ferrooxidans strain DC in liquid nitrogen.