SUMMARYThis paper proposes a novel method for identifying the faults that would impact the transient stability of a multimachine power system. In this method each synchronous machine's swing equation is formulated in the form of a closed loop transfer function in Laplace domain (s-plane). The duration of a fault under consideration, and the changes of maximum electrical power output of the machine during and after the fault are included in the same transfer function. Then the narrowest possible range containing the real part of the 'dominant' root of each machine's characteristic equation is identified on the real axis of s-plane. This is because it is the real part that decides whether the transient response would sustain or subside. If the average of the lower limits of the real parts of dominant roots for all the machines is in the left half but very close to the imaginary axis (poorly damped) or in the right half of the s-plane (absence of damping), the system as a whole is unstable under the considered fault. The proposed method is extremely fast. An extensive digital simulation test and comparison of the proposed method with a conventional step-by-step (SBS) numerical integration method has been done on a 25-machine practical system belonging to Bangladesh Power Development Board (BPDB), and good agreements observed for an array of fault scenarios. The faults for which the aforesaid average lower limit is in the left half plane but its distance from imaginary axis is marginally above or below an established minimum value, can be further investigated by a detailed time domain simulation (TDS) transient stability analysis technique. Thus the proposed method is useful for real-time screening of the contingencies (e.g., faults) before being subjected to a full-scale dynamic security analysis in a power system of any size with any number of machines.