Technology for Design and Quality Control of Microchannel-Based Hemodialyzers and Separators
Why is this technology needed?
The use of microchannel architectures for the bulk production of materials and energy has hinged on an assumed inherent ease of scale-up. While this assumption may hold true for laboratory scale devices with a small number (hundreds) of microchannels, as the number of channels increases, so too does the possibility of uneven and variable flow performance.
In membrane separation processes performed in microchannel-based devices, the equal and even distribution of flow among many parallel microchannels is the determining factor in overall device performance. Directly measuring the velocities in each channel independently is rarely feasible or practical, but a satisfactory measurement of the fluid exit age distribution obtained from a tracer impulse response test is feasible. The distribution of probable device performance characteristics can be inferred from that measurement.
OSU and MBI have developed both an experimental method to evaluate the efficiency of microchannel-based membrane separation processes and a numerical method and software package, Trace-Now®, to analyze data and to design new microtechnology-based membrane separation processes. These methods were developed in connection with our research and development effort on a microchannel-based hemodialyzer. Pertinent fundamentals and the experimental basis for the developed methods are presented in Anderson (2009) and Warner-Tuhy (2009).
The measurement of the fluid exit age distribution, or curve, is a straightforward process if one can induce an impulse bolus of a convenient compound-tracer, function, in the inlet stream of the whole device. The distribution of the tracer in the outlet stream is the transfer function of the device or response curve. Since a perfect function cannot be easily generated in a real physical system, we correct the observed response tracer curve for the error associated with the approximate function. Figure 1 illustrates the experimental method developed to test the separation efficiency of microchannel hemodialyzers.
A numerical code is also developed that uses a generalized deconvolution algorithm to correct the observed tracer output function. Once the system response function is recovered, a framework is provided for the reduction in the error, and an accurate response function is developed. The response curve is then used to generate a set of possible microchannel velocities in the array. A Monte Carlo method was used to determine the overall performance of the device given a randomly generated set of the possible internal states provided by the system response function. This assumes that one has knowledge of the mass transfer performance (theoretical or empirical, see Warner-Tuhy 2009) of a pair of microchannels at any possible set of microchannel-microchannel velocities.
A three-dimensional finite-volume code predicts the mass transfer performance of a randomly generated pair of microchannel velocities. To reduce computational workload, a surface of possible microchannel states was constructed and their performances evaluated numerically. Trace-Now® software then applies a two-dimensional interpolation technique that approximates the values of each microchannel pair. Because of the large number of finite volume simulations that would need to be evaluated for each internal state, this approach significantly shortens computational time. Figure 2 illustrates a performance space of the microchannel hemodialyzer tested in this project.
For additional information . . .
To learn more about this technology, please contact Goran Jovanovic.