چكيده به لاتين
In recent years, the technology related to microfluidic systems has been expanding rapidly and has found wide applications in the chemical and biotechnology industries. This equipment is in the form of a lab on a chip or on a disk and are used for various processes such as mixing, chemical reaction, separation and extraction. Due to the small dimensions, the Reynolds numbers of this equipment are very low; Therefore, in a process such as mixing, the flow is never turbulent and the mixing efficiency is very weak. One of the ways to increase mixing in this situation is to use micromixers. In this research, at the beginning, the micromixer is designed on a new disk, in which, in addition to improving the mixing, after the end of the process, the sample fluid does not remain in the micromixer. It is worth mentioning that addressing how to design the micromixer so that the sample fluid does not remain in it at the end of the process is an issue that has not been addressed in previous studies. The new micromixer is made in the laboratory and the results of the mixing flow are photographed in the conditions of the rotational speed of 600 rpm. After that, the tested disc, which included tanks containing sample fluids, micromixer and collection tank, is simulated. In the following, after simulating the micromixer in the transient state and rotating speed of 600 rpm and comparing it with the laboratory results to ensure the accuracy of the simulation, it was found that the simulation can be performed stably. In all the mentioned cases, the amount of mixing in the outlet of the designed micromixer is about 99%. Then, the mixing flow of the micromixer at different rotation speeds and its effect on the mixing rate are investigated. In the continuation of the project process for the optimization of the designed micromixer, a geometric model with three parameters of radius of curvature of the path, the angle between the initial simple zigzag lines and the length of each initial simple zigzag section is defined according to the reduced primary micromixer and according to the obtained model. With the intervals and limitations defined for the mentioned parameters, 125 models will be created by random numbers code in MATLAB and their mixing flow will be simulated in steady state. In the end, a network with the input of the three mentioned parameters and the output of the mixing rate in the outlet is built by means of the neural network toolbox in MATLAB, and finally, with the help of the genetic algorithm in MATLAB, the network is optimized; The obtained result leads to a mixing of over 98% in the outlet.
