| This thesis has carried out some new explore and try in several common algorithms of electric power flow calculation. For solving linear equations is an important part of power flow calculation, the thesis introduces detaily on Gaussian Elimination, Gaussian-Jordan Elimination and three Triangular Decomposition Algorithms firstly besides do some analysis and comparison; Secondly, this thesis has systematicly generalize the formation and modification of the bus-admittance matrix and equivalent circuit of the circuit elements. It has also put off the application of sparsity and symmetry in the bus-admittance matrix. A specific example verifies the effectiveness of the proposed approach for the bus-admittance matrix; For the model of power flow calculation, this thesis has divided the buses of power system into PQ node, PV node and Blance node from four aspects of variable, type, function and cross-border condition, As for the more numerous iteration times and its low convergency speed of Gauss-Seidel iteration method, the thesis has introduced the acceleration factors so that we can reduce the numerous iteration times and improve its convergency speed of Gauss-Seidel iteration method. In addition, a new Gauss-Seidel iteration method has been put forward which is based on the bus-admittance matrix of rectangle coordinate, and more detail derivation is represented of the node-impedance matrix, For the Newton-Raphson power flow calculation, the thesis focuses on the sparsity and symmetry in the Jacobian matrix and analysis them by examples; Finally, for the especial algorithms of electric power flow calculation Decoupled Power Flow Method and DC Power Flow Method, there is a detailed introduction for them and also verify the Simplifying Algorithm for Decoupled Power Flow Method, As a result,its revised equations will be even more simple.But its convergence,calculating accuracy and original characteristics still remain unchanged. When it requires more calculation speed but less accuracy, The DC Power Flow Method may be the best algorithm of all. |