Abstract: The secondary circuit of the intelligent substation adopts fiber optic digital transmission, and its "virtual circuit" characteristics result in opaque physical connections, making fault diagnosis difficult. Therefore, this article proposes an intelligent warning and fault diagnosis method. This method first models physical entities such as process layer switches, IEDs, and optical cables to construct a complete physical link topology; Secondly, by parsing the SCD file, the logical virtual circuit is accurately mapped to the physical topology, forming a panoramic view of virtual real fusion; Furthermore, a diagnostic model based on multi-source information correlation analysis is established, which utilizes the link alarms and real-time communication status sent by the device, combined with the coupling relationship between virtual and real topology, to achieve probabilistic sorting and accurate positioning of faulty physical nodes. Engineering applications have shown that this method does not require additional hardware and can significantly improve operational efficiency.