An intelligent ultrasonic flowmeter for improved flow measurement and flow calibration facility

Abstract

An increasing need for reduced uncertainty has forced metrologists to look for novel ways to improve the calibration standards for flow. The NIST is experimenting with the use of an advanced ultrasonic flowmeter (AUFM) to improve flow measurement and to detect the dynamic properties of calibration facilities. Ultrasonic technology is evolving rapidly and technical advances have significantly improved flow measurement in continuous industrial processes. The AUFM couples multipath ultrasonic sensing capabilities with pattern recognition software to predict likely flow fields and their probability of existence. The knowledge encoded in the AUFM is derived from training exercises that use computational fluid dynamics and experimental results to teach a flow field recognizer (FFR) via a learning algorithm. A four-path ultrasonic flowmeter prototype has been used to demonstrate the AUFM operational principle. Results showed that the four-path meter can successfully identify flow patterns among several selected flow fields. The results also indicated that the ability of the FFR to identify flow patterns increases as the accuracy of the sensor increases, while decreases as the number of flow patterns considered increases. In addition to being used as a flow diagnostic tool, the AUFM could prove beneficial in field applications where installation effects can lead to gross errors when ultrasonic signals are evaluated using conventional integration techniques.