A smart sensing system to analyze piping vibrations in industrial installations

An important issue that deserves a special attention in almost all industrial installations is related with piping vibrations. This paper presents a measurement system that includes several measuring nodes to acquire vibration data from a set of locations of an industrial piping system. Each measuring node includes basically a tri-axial accelerometer, a vibration sensor, a temperature sensor, a microcontroller and a ZigBee wireless communication unit. Concerning the software component of the measurement system, a particular attention was dedicated to measurement accuracy and errors' compensation caused by influence variables. Artificial neural networks, namely Kohonen maps, are proposed to classify vibration patterns associated with the data collected from the distributed measurement network. A flow meter test bench was used to perform static and dynamic tests and to evaluate measurement system performance.