nRTD: Determining Complex Residence Time Distributions from Experimental Data Using Convolutional Neural Networks

Obtaining quantitative information about residence time behavior (i.e., the residence time distribution function) in realistic experimental systems is oftentimes experimentally challenging and numerically complex. The conventional way is to conduct very simple pulse or step tracer experiments or construct elaborate compartment models. In this work a workflow is proposed and demonstrated to obtain accurate residence time distributions without the need for trivial step experiments or complex compartment models. The workflow consists of a systematic experimental program and a novel machine learning algorithm to fit residence time distributions via convolutional neural networks which the authors call nRTD – neural residence time distribution. nRTD is demonstrated for literature models and a dynamically operated test rig used for kinetic experiments.