微流控芯片中电渗流动拓扑优化设计方法研究

发布时间：2018-03-18 16:48

本文选题：微流控芯片　切入点：拓扑优化　出处：《中国科学院长春光学精密机械与物理研究所》2017年博士论文　论文类型：学位论文

【摘要】：微流控技术是近三十年发展起来的一项高新技术,已经在医学、生物和环境领域得到了广泛应用,具有高效、便携、微尺度效应等特点。微流控芯片是微流控技术的载体,主要功能部件包括微阀、微泵、微混合器以及管道网络等。因此,微流控芯片中这些功能部件的设计及优化决定了微流控芯片的可靠性、稳定性及集成化程度。拓扑优化方法是工程结构设计的重要工具,尺寸及形状优化只是在已有拓扑结构上进行,不仅对设计者的经验依赖性大且对结构的性能提升有限,而拓扑优化方法能够在设计者缺乏经验的情况下寻找到结构的最优拓扑,并能够同时进行一定程度的尺寸及形状优化,有助于实现创新设计,在保证结构功能实现的前提下最大程度的提升性能。因此本文采用拓扑优化方法对微流控芯片进行结构优化设计。压力驱动和电渗驱动是目前微流控技术中最为常见的两种驱动方式,基于压力驱动的流体流动拓扑优化方法已经非常成熟,包括传统的变密度法和水平集方法等。然而基于电渗驱动的流体流动拓扑优化方法发展缓慢,一方面是由于从压力驱动到电渗驱动,较为单一的流场问题变成了流场、电场和离子浓度场的耦合问题,增加了优化模型的复杂程度;另一方面,双电层的厚度相对于微管道的尺度通常要小两到三个数量级以上,此时电渗驱动的流体流动拓扑优化问题也是多尺度优化问题,增大了计算难度。因此,本文从改进拓扑优化方法和简化物理模型两个方面解决上述问题。变密度法和水平集方法是应用最为广泛的两种拓扑优化方法,两种优化方法各有优缺点,为了克服这些缺点并充分利用它们的优点,近年来出现了拓扑描述函数方法和参数化水平集方法,两者的共同点都是引入基函数对水平集函数进行插值显式表达。由于电渗驱动的驱动力沿管壁切线方向分布,所以本文采用可以描述边界演化的参数化水平集方法,并对其进行了拓展,与传统变密度法相比,该方法能够有效减少设计变量的数量,并且具有清晰的演化边界;与传统的水平集方法相比,该方法迭代步数少,不需要进行重新初始化,并且可以使用成熟的优化算法。本文先将拓展参数化水平集方法应用到悬臂梁结构刚度最大化的优化问题中,验证了算法的有效性和鲁棒性。本文还将拓展参数化水平集方法应用到无体力定常Navier-Stokes流体流动拓扑优化问题中,推导了连续伴随方程和伴随敏度,并利用经典算例进行验证,发现不同大小磨光参数下的拓扑优化呈现出完全不同的特点。基于以上拓展参数化水平集方法理论和数值实验的经验,本文使用拓展参数化水平集方法对诱导电渗流微泵进行了拓扑优化,与传统变密度法相比,最终结果中结构边界更加清晰,并且中间密度值更少,提高了结果的准确性;和传统水平集方法相比,不同交变电压频率下的诱导电渗流流量都有不同程度的提升。为了将拓扑优化方法应用到管道尺寸较大情况下的电渗流动中,本文使用Helmholtz-Smoluchowski电渗流滑移边界条件代替电渗流模型中的双电层,简化了物理模型,然后使用和Brinkman惩罚模型类似的惩罚项,将隐式边界上的滑移边界速度条件添加到Navier-Stokes方程中,推导了问题的连续伴随方程和伴随敏度,并利用拓展参数化水平集方法进行优化。最后本文在前人微混合器拓扑优化的基础上提出了一种新的直流电渗微混合器,该混合器能够有效减少电极数量和电极上的电压,从而降低制造难度和成本。
[Abstract]:Microfluidic technology is a new technology developed in recent thirty years, has been in the medical, biological and environmental fields has been widely used, which is efficient, portable, micro scale effect and so on. The microfluidic chip is the carrier of microfluidic technology, the main features include micro valve, micro pump, micro mixer and pipe network etc. Therefore, determines the reliability of microfluidic chip design and optimization of these functional components in microfluidic chip, stability and integration degree. The topology optimization method is an important tool for engineering design of the structure, size and shape optimization was done in existing topology, not only on the experience of the designer and enhance the dependence of Co. the performance of the structure and topology optimization method can find the optimal topology in the designers lack of experience, and also the size and shape of a certain degree of optimization, Contribute to the realization of innovative design, improve the performance to ensure the greatest degree of structure under the premise to realize the function. So this paper uses the method of topology optimization for structure optimization design of the microfluidic chip. The pressure driven and electroosmotic driven microfluidic technology is currently the two most common drive mode, the pressure driven fluid flow topology optimization based on the method has been very mature, including the traditional variable density method and level set method. However, electroosmotic flow topology optimization method based on the slow development, one is driven by pressure from the driver to the electroosmotic flow problem, is a single into a flow field, the coupling problem of electric field and ion concentration field, increase the complexity optimization model; on the other hand, the thickness of the double layer relative to the micro channel are usually small scale of two to three orders of magnitude, the electroosmotic flow of fluid Is the multi-scale optimization problem of dynamic topology optimization problem, increases the difficulty of calculation. Therefore, the improved topology optimization method and solve the problem in two aspects. The simplified physical model of variable density method and level set method is used most widely two kinds of topology optimization methods, two kinds of optimization methods have advantages and disadvantages, in order to overcome these shortcomings and make full use of their advantages, in recent years the topology description function method and parametric level set method, the similarities are the basic functions of the level set function are explicit expression. Due to the driving force electroosmotically driven distribution along the tangent direction of tube wall, so the parametric level can be described the evolution of the boundary set method, and analyzes its development, compared with the traditional variable density method, this method can effectively reduce the number of design variables, and with clear boundary evolution; Compared with the traditional level set method, the method of minimum iterations, without re initialization, and can use the mature optimization algorithms. This paper first will expand the parametric level set method is applied to the cantilever beam structure stiffness maximization optimization problems, to verify the effectiveness and robustness of the algorithm. This paper also will expand the parametric level set method is applied to the unsteady flow of a Navier-Stokes fluid physical topology optimization problem, deduces the continuous adjoint equation and adjoint sensitivity, and the use of the classic examples validate that different size of polishing parameters under the topology optimization has completely different characteristics. The above extended parametric level set theory and method numerical experiments based on the experience of topology optimization on the induced electroosmotic flow micropump using extended parametric level set method, compared with the traditional variable density method, the structure of the final results The boundary more clearly, and the intermediate density value, improve the accuracy of the results; and compared to the traditional level set method, the different frequency of the alternating voltage induced electroosmotic flow have different degrees of improvement. For the topology optimization method is applied to the case of large size pipeline electroosmotic flow, this paper uses Helmholtz-Smoluchowski electric double layer seepage slip boundary condition instead of electric seepage model, simplified physical model, and then use the Brinkman model similar to the penalty punishment, add the slip boundary conditions on the boundary velocity implicit to Navier-Stokes equation, continuous adjoint equation and the adjoint sensitivity problem is derived, and the use of extended parametric level optimization set methods. Finally, this paper proposes a new DC electroosmosis micromixer based on previous micro mixer topology optimization, the mixer can effectively Reduce the number of electrodes and the voltage on the electrode, thereby reducing the difficulty and cost of manufacturing.

【学位授予单位】：中国科学院长春光学精密机械与物理研究所
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TN402

【相似文献】