Electrical Capacitance Tomography with physical bound constraints

Electrical capacitance tomography (ECT) can be applied for the solution of complex measurement tasks in industrial process monitoring, e.g., multiphase flow imaging. ECT requires the solution of a challenging ill-posed inverse problem that usually needs regularization. A well-known method is regularized least squares minimization. Prior information that is often neglected in the minimization are physical bounds of parameters. Classical techniques for constraint handling in optimization can be computationally prohibitive for industrial real-time applications. To overcome this limitation a novel approach based on a nonlinear parameter transformation is proposed. The achievable improvement is demonstrated with measurement data from a prototype sensor.