Characteristic Prediction Of Power System Dynamic Frequency Considering Governor Nonlinearity

Recent years,two kinds of countermeasures have been applied in the power system of China in order to cope with the energy crisis.One is to strengthen the connection between the power supply and the load area to achieve the reasonable distribution of electric energy.The construction of high-voltage DC line is a representative project to transfer the surplus energy of western region to supply the demand of eastern coastal cities.The other is to develop the clean and renewable energy sources such as the construction of large-scale wind farms and their integration to the main grid.The firs mentioned of two increases the power exchange between different regional grids.Once the interconnecting tie line tripps,both the sending area and the receiving area are confronted with the imbalance of large power.The other kind meets the high load demand while reducing the system inertia and thereby decreasing the systematic capability of frequency modulation.In such situations,the frequency stability of power system is more threatened than in the past.The study of power system frequency characteristics refers to analyze the frequency response under the impact of power imbalance caused by contingencies.The emphasis is the extreme and steady values of frequency variation.The analysis of power system frequency characteristics and the frequency prediction algorithm are studied in this dissertation.Considering the influence of nonlinear segments on the system steady-state,the static security assessment method taking into account both the efficiency and the veracity is proposed.The dissertation contains the following several aspects:(1)Aiming at the deadband in the actual power systems,the method of calculation the minimum frequency of the nonlinear power system is put forward.In the case of the primary frequency modulation,the dynamic process of the disturbed system is analyzed,and the frequency response curve of the system with deadband is compared with that of non-deadband system.The relationship between state variables and the changing process of each state variable of the control system are analyzed in detail.The mathematical expressions of the relation between input and output are given.The describing function of the governor system is derived.By the step test and the curve fitting method the parameters of governor equivalent model is determined.The system frequency response model considering the deadband effect is established.The model decouples the control loop associated with governor feedback from system frequency and approximate the short term frequency behavior to simplify the swing equation in order to obtain an estimate of frequency nadir after a sudden loss of generation.The effectiveness and accuracy of the algorithm is compared with the time-domain simulation using the IEEE-9 standard system and a practical grid in China.(2)Regardless of the dynamic process,the static frequency-power characteristic of generator under the control of governor with deadband is studied using the frequency response model proposed.The effective regulation coefficient considering the deadband impact on the value of steady-state frequency is put forward.According to the principle of the last step of iterative algorithm to converge,the steady-state frequency value is calculated directly by the operation of algebraic equations.The instantaneous data after disturbance acquired from WAMS is applied to initialize the system state variables and the effective regulation coefficient is adopted.The proposed method considers the network and load characteristics.Compared with the existing methods it takes the influence of governor deadband into account.Using IEEE-9 system and a practical grid in China,the effectiveness of the algorithm is verified.It is of high sensitivity to the width of deadband and is superior to other predictive methods.(3)The input-output relationship of governor with amplitude-limit is analyzed.The mathematical description of the saturation is adopted to be equivalent to the nonlinear characteristics of governor system with amplitude-limit.The describing function method is applied to establish the simplified model of governor system.According to the parameters of the simplified governor model and the pre-disturbance state of system,an index indicating the possibility of governor being amplitude-constrained is proposed.The smaller value of index refers to higher possibility of being amplitude-constrained.The number of governors that are deemed to be amplitude-constrained in the algorithm is determined based on the ratio of the disturbance power and the average margin of governor regulation.In accordance with the index value this number of governors are preselected as the amplitude-constrained to be simplified using the describing function method,and the others are simplified using general model.The calculation algorithm of minimum frequency proposed is applied to implement the minimum frequency prediction considering the amplitude-limit.The validity of the method is verified by the comparison with the time-domain simulation of the test system.The accuracy of prediction is improved with the help of the preselecting program.(4)The dynamic power flow model considering the deadband and amplitude-limit is established.It is used to compute the steady-state power distribution and the steady-state value of voltage,current,phase angle and other parameters.On the basis of the model the static security assessment method is proposed for large-scale power system.The rapidity and precision of the dynamic power flow computation help to reduce the expenses of calculation and increase the accuracy of security margin.The contingencies with lower margin are analyzed using the time-domain simulation to provide comprehensive information for control measures.The effectiveness of the dynamic power flow model is verified by the time-domain simulation.The proposed static security assessment method is applied in China Southern Grid to achieve the objective of both efficiency and accuracy.