Research On Analytical Aggregation Method In Dynamic Equivalents Based On Structure Preserving Technique

Dynamic equivalence of power system can reduce the system in a large scale, while the dynamic performances are remained unchanged. Dynamic equivalence can emphasize the main characteristics of system and reduce the computation time, thus it has become an important tool in the area of large power system study. Generator units aggregation is the core work in dynamic equivalence, and generator units aggregation have an significant impact on the dynamic characteristic of the equivalent system. Aggregation method such as weighted average, frequency-domain fitting both showed clear defect, thus they can be hardly implied to complex large power system.This paper first the summarize weighted average, frequency-domain fitting and leading generator parameter method, and analyze their advantages and disadvantages. In this paper, an improved analytical method of coherency generating unit aggregation is presented. Traditional methods, which are based on transfer function fitting in frequency domain or time domain simulation, rely on the setting of working point and optimization direction, while the presented method makes up for these shortages. The presented method first eliminates the difference derived from moving coherent generators to the common bus via adding phase shift transformer. Based on the coefficient matrices structure preservation in time domain of generating units with detailed models, as well as the sum of current phasors on every axis of synchronous plane is equal, the parameters of the equivalent generating units then can be directly derived.This paper proposes a generator excitation system aggregation method based on a unified model. It can be used to aggregate generator excitation systems with different models and parameters of significant difference. The key idea of this paper is to convert various excitation models into a unified one. Then models and parameters of equivalent generator excitation system can be directly derived from the structure preserving characteristic of coefficient matrix. Compared with traditional methods which are based on transfer function fitting in frequency domain and time domain simulation, the proposed method requires less calculation, and does not rely on setting of working point and optimization direction. Simulation results show good aggregation effect, thus the validity of this method was demonstrated.The analytic method presented in this paper is more adaptive to different structures and complexities of generator and its control system, and can be applied to on-line calculation. Various simulation results demonstrated the availability of this method. The south grid power grid is taken as an example, where the generator and its excitation system are aggregated based on the presented method. The presented method is evaluated by comparing with the traditional methods Simulation results show good aggregation effect, thus the validity of the presented method is demonstrated.This method is supported by the National High Technology Research and Development Program(“863” Program) of China(2012AA050209). Some of the research results have been applied in China South Grid, and the economic benefits and technical advantages verify the correctness and feasibility of the work.