Inverse measurement method for detecting bubbles in a fluidized bed reactor—toward the development of an intelligent temperature sensor

The objective of this work is to contribute to the development of an intelligent temperature sensor capable of reconstructing the actual process temperature from the indicated temperature signal through numerical solution of the inverse transduction problem. In addition, it is required that the reconstruction algorithm results in a small computational code, suitable for implementation in a commercial microcontroller for on-line monitoring. The necessary regularization technique was based on extracting derivatives and current indicated values from a low order polynomial fitted to the last few temperature readings. Numerical and experimental results show that the proposed technique allows the reconstruction of the process temperature under realistic experimental conditions at relatively high noise levels.