The aim of this doctoral work is to decrease the risks of the power system which are generated from loss of its dynamic stability. To reach this goal methods and algorithms are analyzed for evaluation of risks in the power system. Chosen evaluation criteria unite and reflect possibility of appearing events which causes loosing of stability and its consequences. Process of dynamic stability disturbances in large power systems is elaborated and simulated on the base of automatics and use of simplified models of protection. Methodology of validation of models is developed. Synthesized new structure based on the use of global positioning for preventing asynchronous operation and automatics of liquidation. Justified algorithm of functioning of distance protection, capable to calculate a value of risk in moment of short circuit. Proved effectiveness of algorithm. Performed evaluation of influence of distributed generation on the stability of large power system.