I. INTRODUCTION
Power System is considered to be stable, if generation and the load are equal. This state of equilibrium gets disturbed every time, as load changes. Excitation control at generators keeps generator terminal voltage constant, if the load changes. Turbine governor keeps the speed constant, in case of load changes. Larger changes in load or generation make power system more unbalance thus the chances of system instability increase. Loss of any major equipment or transmission line may create a generation-load unbalance which in turn may lead the system towards instability or blackout. Separation of power system into different balanced islands is one of the possible solutions to avoid wide area blackouts. Synchrophasor technology has made it possible to get the phasor values of voltage and current along with frequency, this information can be used to find the groups of coherent generators, which can be very helpful for proper system islanding. In case of a wide area out-of-step (OOS) instability, where power grid may have different groups of coherent generators, difference in rate of change of generator bus voltage phase angles can indicate the level of wide area OOS instability. Moreover, the information of groups of coherent generators can be very important to initiate islanding or separating the power grid by tripping the tie-lines, into separate systems, which may achieve a better load-generation balance. A new sychrophasor data arrangement is being proposed, which can be used for different applications. A matrix-based data collection method is used to analyze and detect the groups of coherent buses. It has been assumed that the synchrophasor measurements are available from whole power system. Generator bus voltage phase angles have been used for the analysis. This paper presents the detection of coherent groups of generators and the need of system separation using synchrophasor data. The groups of coherent generators are found using Hierarchical Clustering Algorithm and the need for system separation is detected by analyzing the rate of change of mean of the bus voltage phase angles. These two techniques, which use a new synchrophasor data arrangement, provide important information to the power grid operators, the level of wide area OOS instability and the groups of coherent generators in the system, which make these techniques unique as compared to other techniques.