Spectro-tomography - an electrical sensing method for integrated estimation of component identification and distribution mapping in industrial processes

Process tomography offers non-invasive sensing of the spatial distribution in complex processes, e.g. in pharmaceuticals manufacture. In one method electrical currents are injected via a set of peripheral electrodes. Response data are acquired and typically processed using an inverse method to estimate a pseudo-image of the operating process. Data rate is clearly critical to capture process dynamics. Current single frequency systems have limited application in multi-component processes, where the conductivity distribution is ill-posed and depends upon a range of physical or chemical states. The novel development described removes this limitation, with minor real-time data cost, by exploiting an added spectroscopic dimension of direct material discrimination; either from a look-up table of known process components; or identification via a process model. The method presented utilizes a compressed, spread spectrum excitation signal; sensed response data are pre-processed to give frequency-banded results. Feasibility trial data demonstrates the ability to simultaneously estimate in both spatial and spectral dimensions to realize the sensing capability of spectro-tomography.