A New Tera-Hertz THz Imaging Device for Multiphase Flow Metering

Autor: Esra Al Hosani, Mahmoud Meribout, Imran Saied, Mohit Kalra
Rok vydání: 2016
Předmět:
Zdroj: Day 2 Tue, November 08, 2016.
DOI: 10.2118/183475-ms
Popis: In this paper, a new FPGA-based THz imaging device for real-time multiphase flow metering is proposed. The overall system consists of a THz source and a THz camera within which a stainless steel-flanged Teflon-made cylindrical probe is provided to carry the multiphase flow. Hence, the system acts like an X-ray-based device but has the advantage of having non-ionizing waves which makes it a safer alternative. In addition, the system provides a much informative indication about the flow by capturing its two-dimensional image in real-time and without the requirement of performing the time-consuming and errorless-free image tomography techniques. The THz camera consists of 64 × 64 pixels which represent the actual accumulated dielectric of objects value within a projected line in the space. The camera provides the pixel values via a digital video bus (USB port) to an FPGA board for real-time video processing and display. The video processing consists of a cascade of consecutive tasks which include image filtering and histogram, feature extraction and counting (for phase fraction measurement), in addition to block-based motion estimation (for flow rate measurement). This constitutes a breakthrough in the field of multiphase flow metering since this task can take full advantage of the mature and advanced techniques achieved in video processing to achieve the most accurate and informative multiphase flow measurement. Extensive experimental tests were successfully carried out on the developed device using various samples with different concentrations of water and air. Hence, an accuracy of 90.2% for the multiphase flow measurement, with a total processing time of 110 ms/frame were achieved. This can be improved even further if a more advanced FPGA featuring high speed clock, in addition to more advanced video processing algorithms are simultaneously used.
Databáze: OpenAIRE