Reliability and survivability of electric power systems (EPS) depend on transient stability assessment (TSA). One of the most effective way to TSA is time-domain simulation. However, large-scale EPS mathematical model contains a stiff nonlinear system of high-order differential equations. Such system cannot be solved analytically. At the same time, numerical methods are imperfectly applied for such system due to limitation conditions. To make it appropriate, the EPS mathematical model is simplified and additional limitations are used. These simplifications and limitations reduce reliability of simulation results. Consequently, their validation is needed. The most reliable approach to provide it is to compare the simulation results with the field data. However, in practice, there are not enough data for such validation. This paper proposes an alternative approach for validation - the application of a reference model instead of field data. A hardware-software system HRTSim was used as a reference model. This power system simulator has all the necessary properties and capabilities to obtain reliable information required for comprehensive validation of transient stability calculations in EPSs. Main disturbances leading to instability in EPSs are investigated to conduct the validation (processes in cases of faults, single-phase auto-reclosing operation and power system interconnection). Fragments of corresponding experimental studies illustrate the efficiency of the proposed approach. Obtained results confirmed the possibility of the developed approach to identify the causes of numerical calculation errors and to determine disturbances calculated with the significant error. In addition, experimental studies have revealed that numerical calculations error depends on disturbances intensity.