Due to the complex digital mapping relationships within secondary power equipment and the dynamic influence of external environmental factors

the equipment states present strong nonlinearity and uncertainty

making accurate condition assessment difficult.In this study

the covariance matrix is decomposed

and the cumulative contribution rate is calculated to determine the number of principal components.The original data are then projected onto a new coordinate system formed by the selected principal components to extract key features.By combining neural network algorithms

a mapping model from device states to digital space is constructed

thereby generating the corresponding state vector matrix.The autoregressive integrated moving average (ARIMA) model is further used to predict the future operating states of the equipment.Combining these predictions with the abnormality degree calculation results enables intelligent condition assessment.Experimental validation demonstrates that the proposed method achieves higher values of area under curve compared to conventional approaches

indicating superior accuracy and reliability in condition assessment of secondary power equipment.