Enhancing industrial liquid processing through intelligent pipeline mixing

该项目的目的是提供一种从间歇式工业搅拌罐工艺到可扩展管道混合的转换策略。流水线混合提供了提高效率、降低成本和提高产量的机会，以及通过实时快速测量改进质量控制的潜力。这一总体目标将通过应用先进的测量技术来实现，这些技术利用现场粒度测量、显微镜和其他快速质量测量方法，结合人工智能学习和机器控制，所有这些都应用于首次中试和实验室规模实验，检查非牛顿混合、液体分散和反应系统，然后快速扩大到工业管道测试和验证。该项目的合作伙伴将各自为其整体成功做出重要贡献;伯明翰大学的Federico Alberini博士将与英国合作伙伴Calgavin和4 t2 Sensors密切合作，在小规模管道回路中表征静态混合器，同时使用PLIF，ERT和4 t2的定制传感器收集数据并表征结果。萨斯喀彻温大学（湍流多相混合实验室）的Suzanne Kresta博士和William坎贝尔博士也将使用他们实验室的小规模管道回路和现有仪器来检查混合能量和粒度分布，同时与当地合作伙伴萨斯喀彻温研究理事会在其最先进的管流技术中心协调放大测试和验证。阿尔伯塔大学的Alexandra Komrakova博士将协调在线混合计算流体动力学建模的开发，同时与当地合作伙伴AltaML合作开发可应用于在线混合的机器学习方法。然后，这种机器学习方法将与NRC在人工智能数据收集和在线混合仪表编译方面的知识相结合。这种基于数据的学习方法将应用于自动化系统的开发，该系统将与其他研究中心的成果一起用于SRC管流中心的全尺寸测试系统的开发，集成和演示工作的各个方面。拟议项目的具体目标包括：·开发瞬时测量技术，·基于涉及粒子数平衡的直接数值模拟，使用机器学习和人工智能来模拟管道过程反应、粒子数平衡和混合能量·使用原位显微镜、粒度分析，颗粒跟踪和断层扫描技术与机器学习相结合，用于在线过程优化·集成仪表、人工智能和机器学习，用于管道混合与工业过程自动化系统。工业管道回路中仪器/机器学习算法和自动化的全面验证和测试

The purpose of this project will be to provide a conversion strategy from batch-style industrial stirred tank processes to scalable pipeline mixing. Pipeline mixing offers the opportunity for greater efficiency, reduced cost and higher throughput as well as the potential for improved quality control through real-time rapid measurement. This overall goal will be approached through the application of advanced measurement techniques utilizing in-situ particle size measurement, microscopy and other rapid quality measurement methods combined with AI learning and machine control, all applied to first pilot and bench scale experiments examining non-Newtonian mixing, liquid dispersion and reacting systems, followed by rapid scale-up to industrial pipeline testing and validation. The partners to this project will each contribute vital elements to its overall success; Dr. Federico Alberini at the University of Birmingham will work closely with UK partners Calgavin and 4t2 Sensors in characterizing static mixers in small scale pipe-loops while collecting data and characterizing the results using PLIF, ERT and 4t2’s custom sensor. Dr. Suzanne Kresta and Dr. William Campbell at the University of Saskatchewan (Turbulent Multi-phase Mixing Laboratory) will likewise examine mixing energy and particle size distribution using their lab’s small scale pipe-loop and existing instrumentation, while coordinating scale-up testing and validation with local partner Saskatchewan Research Council at their state-of-the-art Pipe Flow Technology Centre. Dr. Alexandra Komrakova at the University of Alberta will coordinate development of computational fluid dynamic modelling for in-line mixing while working with local partner AltaML in developing machine learning methodology that can be applied to in-line mixing. This machine learning method will then be integrated with NRC’s knowledge in AI data gathering and compiling from the in-line mixing instrumentation. This data-based learning method will then be applied to the development of an automation system that, together with the results of the other research centres will be used in developing a full scale test system at SRC’s pipeflow centre, integrating and demonstrating all aspects of the work.The specific goals of the proposed project include• Develop technique for instantaneous measurement, correlation and prediction of pipeline mixing energy (J) from advanced instrumentation• Simulation of pipeline process reaction, population balance and mixing energy using machine learning and artificial intelligence based on direct numerical simulations involving population balances• Use of in-situ microscopy, particle size analysis, particle tracking and tomography techniques combined with machine learning for in-line process optimization• Integration of instrumentation, AI and machine learning for pipeline mixing with industrial process automation systems.• Full-scale validation & testing of instrumentation/machine learning algorithm and automation in an industrial pipe-loop