Fiber Bragg grating accelerometer-based nonintrusive flow rate measurements and leak detection.

We aim to develop a nonintrusive, reliable, convenient flow rate measurement and leak detection method. The flow rate is known to be related to the vibration excited by the turbulent internal flow, but there exists some scatter in the vibration data at low flow speeds (<2/) due to the resolution limits of the sensor. A customized fiber Bragg grating acceleration sensor is designed and applied to measure the low flow rate and detect leakage in the pipeline. Experimental tests show the relationship between the standard deviation of vibration data and flow rate at low velocity is quadratic; the of quadratic curve fitting is 0.990. Moreover, for leak detection, selected features are used as the input of support vector machine classifiers to identify the leak in the pipeline. The results demonstrate that the proposed method achieves a high accuracy of 99.00% to determine the leak state and nonleak state.