Artificial Neural Network Modeling of Solvent-Free Extrusion Emulsification

含有微米或纳米级聚合物颗粒的液-固分散体用于加拿大公司的广泛产品，例如食品，药品，化妆品，石油钻探/回收，因此对我们的环境和经济产生重大影响。 所使用的聚合物中很少有可以在不使用对环境有害的溶剂的情况下分散。 无溶剂挤出乳化（SFEE）是一种基于更环保的制造理念的新方法，其中任何具有足够粘度的聚合物熔体都可以通过非常规使用双螺杆挤出机转化为水性分散体。该方法重新设计了整个生产工艺，消除了许多下游单元操作，降低了能源利用率，并避免了有害溶剂的必要性。然而，这种新工艺缺乏任何深入的研究和建模，迄今为止几乎没有成功的产品。它目前完全依赖于经验丰富的处理器和几年的试错技能。由于其对许多变量表现出高灵敏度，因此该过程难以控制和缩放。 我们正在提出一个为期两年的项目，研究人工智能（AI）技术的实施，以模拟制造过程，创建一个动态平台，以利用我们积累的知识进行增长，并使用新配方测试控制方案。人工智能技术在这方面很有价值，因为它们不需要物理机制来提供良好的预测，并且已知可以突出响应中通常不被解释为系统机制重要的细节。我们的目标是与公司赞助商分享这个模型，并提供足够的培训，以帮助他们满足生产需求，并为化学工程界开发将人工智能技术应用于流程的新信息。参与该项目的学生将参与人工智能技术和工业相关机械的研究，研究创新流程和建模方法，并与财富500强公司联系。

Liquid-solid dispersions containing micro- or nano-sized polymer particles are used for a broad range of products from Canadian companies, spanning foods, pharmaceuticals, cosmetics, oil drilling/ recovery, to name a few examples and thus have a major impact on our environment and economy. Few of the polymers used can be dispersed without use of environmentally-harmful solvents. Solvent Free Extrusion Emulsification (SFEE) is a new approach based on greener manufacturing philosophies where any polymer melt of sufficient viscosity can be converted into an aqueous dispersion by making unconventional use of a twin-screw extruder. The approach redesigns the entire manufacturing process, eliminating many downstream unit operations, lowering energy utilization, and avoiding the necessity of harmful solvents. This new process, however, lacks any deep investigation and modeling, with few successful products prepared to date. It currently relies entirely on the skills of highly experienced processors and several years of trial-and-error. Because of its high sensitivity displayed to a great many variables, the process is difficult to control and scale. We are proposing a two-year project to study the implementation of artificial intelligence (AI) techniques to model the manufacturing process, creating a dynamic platform to grow with our accumulated knowledge, and test control schemes with new formulations. AI techniques are valuable in this regards since they do not need a physical mechanism to give good predictions and are known to highlight details in the responses that might not be normally interpreted as significant of the system's mechanism. The goal is to share this model with the company sponsor, with sufficient training to assist in their production needs and to develop new information for the Chemical Engineering community on applying AI techniques to processes. The student working on the project will be involved in studies of AI techniques and industrially relevant machinery, looking at an innovative process and modeling method, and having contact with a Fortune 500 company.