机理与数据融合驱动的磷酸反应过程建模
Modeling of Phosphoric Acid Reaction Process Driven by Mechanism and Data Fusion
李显军 1赵小平 1余德靖 2赵亮2
作者信息
- 1. 贵州磷化(集团)有限责任公司,贵州贵阳 550005
- 2. 华东理工大学,上海 200237
- 折叠
摘要
为了探究磷矿粒径、液固比等因素对湿法磷酸反应槽中磷矿溶解反应的影响,提出了一种机器学习数据融合的建模方法.首先采用集总参数法建立能够表征磷酸生产特性的机理模型,其次利用工业运行数据和深度置信网络,获得不同工况下模型中未知的状态参数.并利用Python平台进行建模仿真,分析不同液固比以及对磷矿浆溶液转换率的影响.仿真表明,在反应槽液固比为10∶1以及磷矿粒径为100 μm时,磷矿的转化率最高,该模型对湿法磷酸生产和磷收率的优化调控能起到支撑作用.
Abstract
In order to investigate the influence of factors such as phosphorus ore particle size and liquid-solid ratio on the dissolution reaction of phosphorus ore in wet process phosphoric acid reaction tanks,this paper proposes a modeling method based on machine learning data fusion.Firstly,a mechanism model that can characterize the production characteristics of phosphoric acid is established using the lumped parameter method.Secondly,industrial operation data and deep confidence networks are used to obtain unknown state parameters in the model under different operating conditions.And use the Python platform for modeling and simulation,analyze the impact of different liquid solid ratios on the conversion rate of phosphate slurry solution.Simulation shows that the conversion rate of phosphate ore is highest when the liquid-solid ratio in the reaction tank is 10∶1 and the particle size of phosphate ore is 100.This model can provide support for the optimization and control of wet process phosphoric acid production and phosphorus yield.
关键词
湿法磷酸/反应槽/机理模型/机器学习/数据融合Key words
wet process phosphoric acid/reaction tank/mechanism model/machine learning/data fusion引用本文复制引用
出版年
2024
NETL
NSTL
万方数据