Machine monitoring using fuzzy-neural networks

Abstract

In response to the rapid pace of technological change, many big manufacturers are increasingly looking towards solutions based on plant informatization and Industry 4.0 concepts. However, in the context of Taiwan, such options are off limits to many small and medium-sized firms due to limited scale and capital. This paper proposes a plant informatization approach which can be implemented by smaller manufacturers through using add-on sensor systems to monitor production equipment. An accelerometer is installed on existing machinery to collect vibration data, which is subjected to feature extraction to create a monitoring model through implementing the LDA algorithm and the fuzzy neural networks. Experimental results indicate the resulting model can be effectively used to detect abnormal machinery operations.