大规模并行微反应器系统过程监控和故障诊断的研究

Microreactors, which can realize the quick reaction, have come into view in recent years. The low hold-up in microreactors can offer excellent controllability, reduce risk and result in cheaper and more environmentally friendly operation, however, commercially available microreactor products for the industry are still limited. In order to realize microreactors with high throughput, the parallelization of microreactors is an effective approach. Process monitoring and fault diagnosis are indispensable when the hign reaction yield is also needed. Due to the size limitation of microreactors, it is not practical to install sensors in each microreactor when parallelized structures are adopted. In a typical large scale parallelized microreactors system (LSPMS), which included the parallelized microreactors (PMR) and the flow distributors (FD), the PMR are connected to FD section. Mounting temperature sensors on the PMR and flow sensors in the FD can monitor process variables and detect faults. In PMR, a computational fluid dynamic (CFD) model, which can describe the relationship between the flow rate and the outer wall temperature of each microreactor, is first constructed. An optimal measurement location can be decided when minimizing the estimation accuracy for the flow rate by using these temperatures. The reaction dynamics can be described by using the temperature difference of neighboring microreator. There are three main faults in PMR: blockage, corrosion and catalytic deterioration. For the last two faults, the outer wall temperature is changed because the heat transfer efficiency of the outer wall is variable. The lower bound of the changed temperatures can build a fault area, it will detect the corrosion and catalytic deterioration. The blockage can be detected by temperature in PMR and flow rate in FD. In PMR, the model, which describes the flow rate and the outer wall temperature of each microreactor, can be used for fault diagnosis. When the flow rate is zero, the fault area is created by using the lower bound of changed temperatures. Meanwhile, in the FD, an optimal structure, optimal pipe size, also with optimal measurement location of flow sensors are designed to detect the blockage in PMR. Moreover, experimental results demonstrate the effectiveness of the proposed process monitoring and fault diagnosis method.

微反应器作为重要的化工过程强化设备，在提高化学反应转化率，保障安全生产，节能降耗等方面日益受到重视。但反应装置体积小，数量多，使得整个生产过程中的状态监控和故障诊断成为难题。本项目以并行微反应器和流体配送器构成的大规模并行微反应器系统作为研究对象，主要研究：并行微反应器内流体流量和反应器外壁温度之间的数学模型；以最少传感器配置为目标，利用温度分布相关关系，实现最优测量点配置；利用温度变化梯度和温度与流量之间的关系，判定反应器内阻塞，腐蚀和催化剂劣化等故障。同时，以阻塞故障检测为目标，研究流体配送器的最优结构，管径和流量传感器配置个数；结合并行微反应器内的温度检测，形成有效的阻塞故障判定。本研究通过CFD仿真模拟和实验验证，为大规模并行微反应器系统状态监控和故障诊断的实施奠定理论基础和技术支持。

本项目以并行微反应器和流体配送器构成的大规模并行微反应器系统作为研究对象，完成了以下研究内容：1) 建立了并行微反应器内流体流量和反应器外壁温度之间的数学模型；以4通道并行微反应器为研究对象，对在模型中配置2~4个温度传感器分别进行研究，以在模型中进行等间距配置时的方案为优化初始解，求得3个传感器的最优配置方案。2) 实现了微反应器内“部分阻塞”条件下，阻塞程度的定量判别。提出了阻塞形态的两种主要形态，第一类：沿流动方向分为粗糙增强型阻塞。描述这类阻塞的特征是：阻塞体体积和与反应流体接触的切向表面积。第二类：垂直于流动方向分为圆形截面、方形截面和三角形截面型阻塞。描述这类阻塞的特征是：阻塞体体积和与反应流体垂直的表面积。根据上述分类特征，建立了描述阻塞程度的阻塞指数。并根据阻塞指数，提出了消除阻塞的方法。3) 以阻塞故障检测为目标，在流体配送器中实现均一性运行的前提下，提出了一种基于流体分配器流量变化的微反应器单管故障检测法，并给出流量敏感度值的概念，用于评价单管式反应器阻塞时流体分配器中各个管路流量变化的程度大小，为配置流量传感器的安装个数和安装位置，提供可靠依据。4) 以并行微通道的温度检测和差压检测为手段，根据CFD计算结果，形成阻塞位置的有效判定。5) 在研究手段上，本研究建立了实验平台，实现了利用温利用加热点和制冷点实现阻塞验证实验：模拟了放热反应过程中，微通道发生部分阻塞时，温度的局部突增现象可以在微通道上缠绕加热线的方式来模拟。在吸热反应过程中，阻塞的发生温度变化曲线梯度增大，可以用放置制冷片的方法来模拟。按照研究计划，本研究完成的各项研究内容，可以为大规模并行微反应器系统状态监控和故障诊断的实施奠定理论基础和技术支持。