Research on Data Model Fusion Driven Dual End Uncertainty Source Load Probability Prediction and Its Application in New Energy Power Systems

다국어 초록 (Multilingual Abstract)

In view of the strong randomness of supply and demand of new energy system, a new energy power system model with the uncertainty of source and load is established. Considering the output power of the power side, use probability modelling, and analyze the generating power of new energy in one day. Considering the strong impact of the load side, the operation mechanism of time-domain modelling is described. On this basis, a 30-node power system model with uncertain source charge is built on the MATLAB platform, and the power current is calculated to obtain the observed value of each node. The mean of IACKF algorithm was 1.6342103,1.6614103,1.6824103,103,1.7103 103,1.7019 103 and 1.7205 103,92%, 90%, 81%, 81%, 81%, 59%, 4% compared with ACKF algorithm, respectively.
The IACKF algorithm is greatly improved compared with the CKF algorithm. Aiming at the problem of the transient failure time and bad data injection, the combined algorithm is proposed to improve the prediction performance.
Selected the low voltage through, high voltage through, continuous high voltage through and bad data four classic fault cases, on the basis of the new energy power system, analyses the four kinds of fault node voltage amplitude change rule, in view of the traditional kalman filter algorithm in mutation accuracy is not high, put forward the long short memory neural network and adaptive volume kalman filter algorithm combining method, improve the mutation data prediction accuracy. The simulation results show that the improved combined algorithm has higher robustness compared with the traditional algorithm, and can effectively improve the overall prediction accuracy of the new energy power system.