The transient stability limits of power systems are determined by assuming conditions to be constant over short intervals of time and using the results obtained at the end of one interval to vary the conditions at the beginning of the next interval. The analysis is tedious and it is easy to make numerical errors.If a large number of machines is involved, it is essential to use an a.c. network analyser for determining the values of power to be used for the different time intervals, but there is no reduction in the computation labour, and in practice the greater part of the time taken for a study is devoted to the latter.To reduce the computation time and the possibility of numerical errors, the “step” evaluations can be carried out semi-automatically by using the computer described. Furthermore, by reducing the time interval, greater accuracy can be obtained without an increase in overall time as compared with normal procedure.