The Modeling And Parameter Identification Of Hydro Turbine Governing System Based On A Hybrid Ant Colony Algorithm

After 60 years of development, China power system has achieved great achievements. Both of the speed of electricty conctrction and grid size are in the forefront of the world, and the total installed capacity of hydropower is ranked first in the world. Following the situation, the safety of the power system becomes increasingly important and the high performance of the electrical equipment controlling is required. As a stable guarantee of the hydro-turbine operating, the hydroturbine governing system decides the quality of the hydropower. Good models of hydroturbine governing systems provide the power system stability calculations with a reliable base. So the modeling and parameter identifiaction of the hydroturbine governing system which is one of the three major modeling problems are research tops for the scholars.In this paper, we studies the current situation of the modeling and simulating of the hydroturbine governing system, analyses the characteristics of the comupter controller and the controlled system, and then build up the model of the hydroturbine governing system. Parameters of the hydroturbine governor are determined by parameter measureing and that of controlled system are obtained by the identification with recursive least square (RLS) method, genetic algorithm (GA), particle swarm optimization (PSO) algorithm and Ant colony optimization (ACO) algorithm.A hybrid algorithm is proposed in this paper integrating the positive feedback mechanism of ACO and the global searching strategy of AFSA to escape from prematurity. The dispersion factor is designed to control the ants'movements and to maintain the solution diversity. The Hooke-Jeeves method was embedded to the local search method to improve the convergence rate. The comparison of numerical results with LS, GA, PSO, and GACO shows the validity of the proposed algorithm.Finally, the builded model and the improved algorithm are tested with the field data of seven hydraulic turbine units. The hybrid algorithm proposed in this paper outperforms other algorithms in fitness and geometry errors. The simulation results of defined model in the paper are much better than that of the commonly used software PSASP and BPA in the power system stability calculations. The PSASP and BPA models are not consistent with the existing units.