Research On Algorithms For Power System Transient Stability Analysis And Control

Transient stability emergency controls are important means for maintaining power systems and the efficiency of the algorithms for the transient stability emergency control has been a restriction its online application.On the basis of all the power system model parameters obtained,it is very meaningful and important to develop high-efficiency algorithms for transient staility emergency control.The paper focuses on the development of fast algorthms for the transient stability assessment and the emergency control.Several novel algorithms are proposed and verified.The main work is presented as follows:(1)A transient stability numerical integration algorithm for variable step sizes based on virtual input is proposed.The method for fully constructing the nonhomogeneous virtual input for a certain integration scheme is given,and the calculation method for the local truncation error of the power angle for the corresponding integration scheme is derived.A step size control strategy based on the predictor corrector variable step size method is proposed,which performs an adaptive control of the step size in the numerical integration process.The proposed algorithm demonstrates a high level of accuracy and efficiency in practical simulations compared to the conventional numerical integration algorithm.(2)A fast transient stability simulation algorithm based on dynamic equivalence is proposed.Generator models are grouped and aggregated according to a short-time numerical integration.A fast calculation method of the admittance matrix is then proposed to calculate the parameters of an equivalent system,and numerical integration is performed using the obtained equivalent system.Then,based on integral sensitivity,a new fast emergency control strategy for the equivalent system is proposed.The final emergency control strategy is obtained by mapping the control strategy for the equivalent system back to the original system.The suggested algorithm can output an accurate transient stability simulation result and form an effective emergency control strategy.The proposed algorithm is also much faster than the existing solutions.(3)A novel algorithm for the transient stability emergency control based on the deep feedfoward neural network is proposed.The double-level deep feedfoward neural network consists of the 'Classification Neural Network' and the 'Fitting Neural Network'.The transient stability performance index is extacted as the input of the neural network to obtain the result of the transient stability assessment and the value of transient stability constraint function.Then,the gradients of transient stability performance indexes related to transient stability constraint function and the gradients of emergency control variables related to transient stability performance indexes are calculated to obtain the final gradients of emergency control variables related to transient stability constraint function.The final emergency control strategy is obtained according the gradients and a certain optimizing approach.The proposed algorithm transfers most of the numerical integration using the original system to the calculation using the neural network.It reserves the character of transient stability calculation of the original system and significantly reduces the calculation amout,which dosen't significantly increase along with the scale of power system.The suggested algorithm can form an effective emergency control strategy and it is suitable for online application.