A DBN-BPNN Cascade-Model for Electrical Equipment Quality Assessment

The current power equipment quality assessment model is primarily nonlinear and involves a wide range of data, making it challenging to precisely determine the intrinsic connection between the equipment's operational data and its quality level. Challenges such as strong subjectivity and poor interpretability hinder the fine-grained management of power grid assets and equipment. A power equipment quality assessment model based on the DBN-BPNN cascade-model is proposed. Firstly, based on the observability analysis of complex physical systems, the power equipment quality assessment indicator system is constructed and relevant operation data are selected. Then, the two-layer DBN-BPNN cascade model is constructed to realize the power equipment quality assessment. The first layer of the model employs a deep belief network (DBN) to remove data noise and extract quality features of the power equipment, thereby achieving data dimensionality reduction. The second layer employs a quality grading back-propagation neural network (BPNN) to classify the power equipment into various quality grades. Based on the grade, the corresponding quality scoring BPNN determines the quality score of the power equipment. Simulation results demonstrate that the proposed model accurately evaluates the quality of power equipment, offering a scientific basis for the operation and maintenance management of power grid equipment.