| With the maturity of wind power technology,the application of distributed wind power generation solutions has become increasingly widespread.Distributed wind power generation has the advantages of flexible construction and effective improvement of energy utilization,but there are also some problems,such as the lack of wind tower data support,the lack of wind energy prediction system and so on.These problems,combined with the uncertainty of wind resources,aggravate the uncertainty of wind power supply and bring hidden dangers to the power system.In order to improve the above situation,the wind power prediction technology of distributed wind power generation systems is studied.According to the characteristics of distributed wind power generation,taking wind speed prediction as the core,the corresponding work of data processing,wind speed prediction algorithm,scheduling application of prediction results and design and development of remote monitoring software is carried out in stages.In this paper,the multi-step wind speed prediction method is researched and improved,combined with the characteristics and limitations of distributed wind power generation,a new multi-step wind speed prediction scheme is proposed.On the basis of wavelet decomposition data processing,a variety of algorithm models including neural networks and support vector machines are effectively combined,and a newly proposed dual-time sequence multi-step forecasting strategy is introduced,and a point forecasting scheme is designed,which is proved by simulation experiments.This method effectively improves prediction accuracy.Afterward,based on the point prediction,combined with Gaussian process regression to achieve interval prediction,it further improves the prediction accuracy and the guiding significance of actual energy dispatch.In terms of energy scheduling,this paper proposes an algorithm for formulating energy scheduling schemes for multi-energy complementary systems based on the predicted results.The algorithm aims to reduce the operating cost of the system,combined with the optimization ability of the particle swarm algorithm,to search for the optimal system scheduling plan.This part establishes the mathematical model of the multi-energy complementary system.Through the comparison of the traditional dispatching scheme and the new system scheme,it is verified that the wind speed prediction is of positive significance for improving the system dispatching ability.Finally,the remote monitoring software is designed and developed by using Python and QT.In this part,the relevant model algorithm proposed in the research is integrated into the form of software,and the corresponding user interface of each module is designed.The software realizes the data record reading,wind speed prediction,system energy scheduling and other related functions.In general,the research puts forward a new multi-step wind speed prediction algorithm,and explores the practical application of the algorithm in energy dispatch,which is of positive significance for the development of distributed wind power. |
