There is insufficient use and analysis of the operational and maintenance data of the pipe corridor power cabin. Also there is a failure to achieve intelligent operation and maintenance based on big data. Thus a new method for analyzing and judging the status of a pipe corridor power cabin is proposed to protect the integrated pipeline gallery power cabin system. This is intended to improve safety and enhance the level of operation and maintenance. First, the data from multiple distributed data sources is integrated using middleware technology for data layer fusion. Secondly, data with similar characteristics are allocated to the same subspace. Then, in each subspace, a novel Locality Global Projections (LGP) method that can maintain both the global structure of the dataset and the local structure of the dataset is proposed to extract features. The extracted features of each subspace are fused, and the Support Vector Data Description (SVDD) method is used to build a classification model. Finally, through testing on the operational and maintenance data of the power cabin of the integrated pipe corridor, the effectiveness and superiority over other methods are proved. This work is supported by the National Natural Science Foundation of China (No. 61703161), the Science and Technology Project of State Grid Corporation of China (No. KJGW2019-04-II), and the Natural Science Foundation of Shanghai (No. 19ZR1473200).