中文摘要：

山梨酸钾是一种常用防腐剂,应用非常广泛,但食用过量会严重危害人体健康。研究了山梨酸钾在水溶液和橙汁中的荧光特性,山梨酸钾水溶液荧光特征峰为λex/λem=375 nm/485 nm,山梨酸钾和橙汁的混合溶液除了存在此荧光特征峰,还有一个侧峰λex/λem=470 nm/540 nm。在混合溶液中,橙汁和山梨酸钾的荧光特性相互干扰,加大了山梨酸钾浓度检测的难度。为准确测定混合溶液中山梨酸钾的浓度,采用微粒群算法优化的误差逆向传播（PSO-BP）神经网络对其进行检测。3组预测样本的平均回收率为98.97%,PSO-BP神经网络能够精确测定混合溶液中山梨酸钾的质量浓度范围为0.1~2.0 g/L。预测结果表明荧光光谱法和PSO-BP神经网络相结合的方法能有效地检测山梨酸钾在橙汁中的浓度。

英文摘要：

Potassium sorbate, one of preservatives, has been used widely, but it will do harm to human health if it is overtaken. Fluorescence spectrum properties of potassium sorbate in aqueous solution and orange juice are studied. The results show that the fluorescence characteristic peak of potassium sorbate in aqueous solution exists at λex/λem = 375 nm/485 nm, the mixture of potassium sorbate and orange juice has a side peak at λex/λem=470 nm/540nm besides the fluorescence characteristic peak. In the mixture, there is mutual interference of fluorescence characteristic between potassium sorbate and orange juice, which makes the concentration detection of potassium sorbate more difficult. To determine the concentration of potassium sorbate in the mixture, back propagation neural network optimized by particle swarm optimization （PSO-BP） is applied. The average recovery rate of the 3 prediction samples is 98.97%, and the range in which the PSO-BP neural network can accurately measure the concentration of potassium sorbate in the mixture is 0.1-2.0 g/L. The prediction results indicate that the method combining fluorescence spectrum and PSO-BP neural network can effectively detect the concentration of potassium sorbate in orange juice.