Transient Stability Analysis Of Integrated Power System

The transient stability analysis of integrated power system based on numerical integration method is used for off-line analysis because of more calculation and time-consuming. For no requirement of real-time, numerical integration method has a great room for improvement in accuracy of calculation, rate of convergence and applicability. Transient stability analysis based on numerical integration method is applied by solving the dynamic equations, so models of power system are required. Models of generator, excitation system, diesel engine and load are established. A rigid differential equations is made up of mathematical equations of each module.Implicit trapezoidal integration method is selected for ensuring A-stability. By Implicit trapezoidal integration method, analysis results are made by solving nonlinear equations which made by difference equations. Newton method, quasi-Newton method and steepest descent algorithm are analyzed for comparing advantages and disadvantages in the solution of nonlinear equations. And a combinatorial algorithm is made in which initial value of Newton method is provided by steepest descent algorithm to solve difference equations.The steady-state voltage and power are computed in power flow calculation. The power flow computing based on rapid decoupling method is applied in Python programming language. Analysis showed that amount of calculation decreased, computing speed increased, because that sparse matrix, nodes and branches data were expressed by list and dictionary in Python language node numbering optimization of power system is applied by an improved ant colony algorithm. Ant colony algorithm has fast convergence and ability of finding more than one optimal solution by distributed parallel computing. An improving of ant colony algorithm is applied by adding 'initial nodes switch operator' and 'ants out' strategy to the MAX-MIN ant colony algorithm. The probability of finding optimal solutions increased in that switch operator stops ants choosing initial nodes from which optimal solution shouldn't be gotten, and calculation decreased with 'ants out' strategy stopping ants which couldn't get optimal solve keeping finding. Analysis show that the convergence rate is improved and more optimal solves are found with both 'switch operator' and 'ants out' strategy.