An AST-ELM Method for Eliminating the Influence of Charging Phenomenon on ECT

Abstract

Electrical capacitance tomography (ECT) is a promising imaging technology of permittivity distributions in multiphase flow. To reduce the effect of charging phenomenon on ECT measurement, an improved extreme learning machine method combined with adaptive soft-thresholding (AST-ELM) is presented and studied for image reconstruction. This method can provide a nonlinear mapping model between the capacitance values and medium distributions by using machine learning but not an electromagnetic-sensitive mechanism. Both simulation and experimental tests are carried out to validate the performance of the presented method, and reconstructed images are evaluated by relative error and correlation coefficient. The results have illustrated that the image reconstruction accuracy by the proposed AST-ELM method has greatly improved than that by the conventional methods under the condition with charging object.

A facile synthesis of CoMn2O4 nanosheets on reduced graphene oxide for non-enzymatic glucose sensing

“…Its unique two-dimensional monolayer atomic structure provides a large number of active sites for the detected substances, which is beneficial to electrocatalytic oxidation and charge transfer. Consequently, it is of great research value and application potential to hybridize binary transition metal oxides and graphene into a composite material with high sensitivity and reliability for glucose sensors [12,[16][17][18].…”

A Water Fraction Measurement Method Using Heuristic-Algorithm-Based Electrical Capacitance Tomography Images Post-Processing Technology

“…Thus, some researchers turn to find a new method which does not rely on the sensitivity matrix. Wang et al developed an AST-ELM method [9] which directly establishes the map between measured capacitances and images. This neural network based algorithm does not need a sensitivity matrix compared with traditional algorithms (such as the Landweber iterative and the linear back projection algorithm), so the ill-posed nature of ECT's inverse system can be solved.…”