A frequency assessment expert system of piezoelectric transducers in paucity of data

摘要

It is difficult to measure in the laboratory the operating frequency of a piezoelectric transducer that is a kind of electric equipment. Some simulation methods were developed for this frequency calculation, but were still complex to perform. With high learning accuracy, a Frequency Assessment Expert System of Piezoelectric Transducers (FAESPT) is developed in this study to assess the frequency easily. FAESPT is based on mega-fuzzification that is a method to increase the learning accuracy with fuzzy neural network for the small dataset environment. In this article, FAESPT is established and its assessment accuracy is compared with traditional neural network and neuro-fuzzy method. The results indicate that FAESPT can easy to assess the frequency of a transducer and get good learning accuracy in the environment of the paucity of data.