The Studies On Techniques And Software Of Power System Full Dynamic (Electric-Mechemical Transient, Mid-Term And Long-Term Dynamic) Simulation

This dissertation presents the stUdy achievement of developing thetechniques and softWare of power system full dynaAnc simulation, i.e.,unified simulation of power system electric-mechanical transient,mid-term and long-term dynandcs. The main points of the research workof this thesis are as followPower system full dynandc simulation is such a method in which thesimulations of e1ectric-mechanical trallsient, ndd-term and long-termdguamic phenomena are unified and the main feathes of this method areto develop a unified simulation of fast trallsient process and slow ndd-and long-term dynamc process. This kind of simulation is a tyPical stiffsystem and the stiff integration method with variable order and step sizeshould be aPplied.On the basis of analyZing the solution of stiff problems of the fulldynandc simulation in essence, the Gear Method, a numerical integrationmethod that is suitable to solve stiff differential equations is studied. It isrecognized that the Gear Method is strict theoretically, a practicalalgorithm with easy prograedng, and flexible fOr vallable order and stepsize. The Gear Method is one of the most efficient methods to solve stiffproblems. So, Gear Method is adopted as the main numerical solutionalgorithIns for power system full dynandc simulation. The teclinique ofvarying order and step size of integration is researched according to thetrUncation error controI. The study results have established the theorybases for the numerical integral methods of the full dynandc simulation.As the require of v~ng order and step size for the numericalintegration, a differential-algebraic equations simultaneous solutionmethod based on Gear Method are adopted as the main numericalsolution algorithIns fOr network solution. The advantage of this method isthat it is mostly suits to vary order and step size with the Newton Methodor pseud Newton Method, and to eliIninate the interface errors of thesolution of differential equations and algebraic equations. Furthermore,the Newton Method can relax the lindt of step size due to convergentl[ ~condition, llhP* fI] < l, for the simPle iteration method that is used inll lyI]<,, for the simPle iteration method that is used in,some general electric-mechanical transient simulation programs. Andlager step size could be used in the integral process.Based on the Gear Method and the differential-algebraic equationsimultaneous solution method, the Newton iteration formula is derived inthis thesis. The Jacobin Matrix and its elements for differential-algebraicequation simultaneous solution with Newton Method are formed. Thenetwork injection current and its partial differential coefficient are alsoderived. So, the basic calculation formulas for the full dynandcsimulation are established. And the theoretical stUdies for the numericalmethods and algorithms of the full dynandc simulation are carried out.Fault simulation is one of the main conshtuents of power systemdynandc simulation. In power system full dynandc simulation software,the adndttance matrixes of fault branches are adopted in the fault modelsto simulate arbitrary symmetrical and unsymmetrical faults, by use ofcontraction technique for adndttance matrix the multi-faults and therandomly occurred faults can be processed. The feature of this method isthat the arbitrary multi-fault occurred in a branch can be simulatedwithout any additional branches or buses to be added, thereby, theca1culation efficiency and the expedience of usage are imProved.ln power system full dynandc simulation, the electric-mechanicaltransient processes of dynamic elements of power system should besimulated in detail, and the detailed dynandc models of mid-term andlong-term processes are necessary Based on the research and analysis fOrthe ndd-term and long-term models, the dynandc models, such as thermalpower plant, hydraulic power plant, nuclear power plant and theautoma...