Development of a Pattern Recognition Methodology with Thermography and Implementation in an Experimental Study of a Boiler for a WHRS-ORC.

Waste heat dissipated in the exhaust system in a combustion engine represents a major source of energy to be recovered and converted into useful work. A waste heat recovery system (WHRS) based on an Organic Rankine Cycle (ORC) is a promising approach, and it gained interest in the last few years in an automotive industry interested in reducing fuel consumption and exhaust emissions. Understanding the thermodynamic response of the boiler employed in an ORC plays an important role in steam cycle performance prediction and control system design. The aim of this study is, therefore, to present a methodology to study these devices by means of pattern recognition with infrared thermography. In addition, the experimental test bench and its operating conditions are described. The methodology proposed identifies the wall coordinates, traces the paths, and tracks the wall temperature along them in a way that can be exported for subsequent post-processing and analysis. As for the results, through the wall temperature paths on both sides (exhaust gas and working fluid), it was possible to quantitatively estimate the temperature evolution along the boiler and, in particular, the beginning and end of evaporation.

On the Application of Image Processing Methods for Bubble Recognition to the Study of Subcooled Flow Boiling of Water in Rectangular Channels.

This work introduces the use of machine vision in the massive bubble recognition process, which supports the validation of boiling models involving bubble dynamics, as well as nucleation frequency, active site density and size of the bubbles. The two algorithms presented are meant to be run employing quite standard images of the bubbling process, recorded in general-purpose boiling facilities. The recognition routines are easily adaptable to other facilities if a minimum number of precautions are taken in the setup and in the treatment of the information. Both the side and front projections of subcooled flow-boiling phenomenon over a plain plate are covered. Once all of the intended bubbles have been located in space and time, the proper post-process of the recorded data become capable of tracking each of the recognized bubbles, sketching their trajectories and size evolution, locating the nucleation sites, computing their diameters, and so on. After validating the algorithm's output against the human eye and data from other researchers, machine vision systems have been demonstrated to be a very valuable option to successfully perform the recognition process, even though the optical analysis of bubbles has not been set as the main goal of the experimental facility.