混合液体火灾爆炸危险性——闪点预测与实验研究

支持向量机应用核函数技术,已经成为当前国际上一个研究的热点,由于支持向量机具有良好的理论基础和泛化性能,可将其引入到混合液体闪点预测的研究之中,以期建立准确、高效的预测模型。本文建立了一个基于支持向量机的理论模型,用于预测二元互溶混合液体的闪点。根据所研究混合液体的物理性质,选择了纯物质的粘度、表面张力、配比、燃烧下限等物理参数来表征闪点,以这些参数作为输入参数,二元混合液体的闪点作为输出值,应用支持向量机方法对两者之间的内在定量关系进行模拟。结果表明,闪点预测值与实验值符合良好。本方法的提出为工程上提出了一种预测二元互溶液体闪点的有效方法,可应用于评估混合溶液的火灾爆炸危害性及本质较安全设计。

Fire and Explosion risk of mixture-flash point prediction and experimental study

Kernal function used by support vector machine which is hot spot of international study.Support vector machine has a good theoretical foundation and generalization capability,so it can be uesd for predicting the flash points of mixed-liquid.As a result,we can produce an accurate and efficient predictional model.A model based on support vector machine was established to predict the flash points of binary liquid.Based on the physical properties of binary liquid,such as viscosity、surface tension、mole fraction and lower flammable limit which were chosen to represent flash point.Also,these physical parameters were used as input arguments,flash points of binary liquid were used as output one.Support vector machine was used for simulating the quality relationship between input arguments and output one.The result shows that predicted flash points are in good agreement with the experimental value.The method proposed can be used to predict the flash points of binary liquid for chemical engineering.We can use this method to assess the dangers of fire and explosion of mixed-liquid and a safer design.