Design and Implementation of a Portable Digital Energy Measurement Field Diagnosis Instrument Based on DBN

A portable digital energy metering on-site diagnostic instrument based on Deep Belief Network (DBN) is designed to address the issue of current on-site digital energy meters being unable to provide a complete sampling value model and accurately conduct optical power testing. Firstly, the hardware architecture and appearance structure plan of the diagnostic unit were developed through requirement analysis, and corresponding hardware equipment was developed. Then, a dual core (APU&RPU) software structure suitable for on-site diagnostic units was designed and developed separately. Finally, a deeper analysis and feature extraction of the relationship between metric data were carried out by introducing the DBN model, and high accuracy fault diagnosis was achieved. A comparative analysis was conducted between the designed diagnostic instrument and the other two methods under the same conditions through simulation experiments. The results show that the proposed method has the highest accuracy and Fl score in diagnosing 5 different types of faults, reaching 96.54% and 95.75% respectively, which is about 3% and 5% higher than the other two methods, respectively. Its performance is superior to the other three comparative models.