Smart substation is an important part in the cyber-physical power system, whose information flow in the secondary system possesses the key function of reflecting and regulating the energy flow in the primary system. In order to achieve a further refined characterization of network traffic, information source models of devices at the bay level and process level are constructed. On this basis, a set of mathematical models about network injection traffic considering different substation operating conditions is formed by the source models of all the devices connected to the process level network. And a substation performance evaluation method based on process level network traffic is proposed, which enables the determination of the operating conditions experienced by a substation in a certain time period based on the injection traffic, providing a new method for real-time monitoring of the operating performance of smart substations.