Information Analysis And Software Design For Wind Power Prediction

The accelerated development of wind power generation technology has reduced carbon dioxide emission and realized the significant sustainable development of energy and economy. At present, an increasing number of wind power equipment have been in sub-healthy operation in China. Although these wind power equipment are under real-time monitoring, the status information gathered from monitoring hasn’t been effectively analyzed or forecasted. The work of this paper includes the processing of monitoring data of wind power equipment and the forecasting of status information. The work is a process of gathering, processing and analyzing the monitoring data. The monitoring data will be processed through wind power data analysis software, and interference data that will probably affect the experimental results will be eliminated. Next, an experimental simulation will be conducted on the screened parameter data through MATLAB software. And some operation status information of wind power equipment will be forecasted. See the detailed content below:Firstly, the physical property of wind power generation equipment and the principle of power generation are elaborated. On this basis, mechanical and core components in wind power generation equipment that will easily be faulted are briefly analyzed. In light of the practical monitoring data, the relationship among the generated power, the capacity and wind speed of wind power generation equipment is analyzed.Secondly, wind power data analysis software is designed. This software can input original data of wind power generation equipment in a certain format and pre-process the data. Besides, the data screening function required in this research is also available in this software.. Thirdly, BP neural network model, which is the most effective learning method that can explain complicated sensor data, is adopted in the experiment. Also, MATLAB software tool is applied in short-term prediction of the wind power equipment status. The time interval selected in modeling is 10 s. In the process of simplifying the data, the resource consumption of BP neural network model in data sample learning is reduced, and the learning procedure is accelerated. The experiment includes two parts. In Experiment One, the wind speed, and the power and capacity of the draught fan of the cabin meteorological station are predicted. In Experiment Two, Data of different time intervals are adopted to carry out short-term prediction of wind speed, wind motor power and output capability of weather station of engine room. After that, relevant parameters will be calculated on the predicted data results and actual value. The absolute value is taken from the calculated parameters and the relevant indexes will be compared with each other. If the value is 0.81-1, there is a strong correlation; 0.31-0.8 indicates a moderate correlation and 0.1-0.3 indicates the correlation is weak. This experiment takes the data for the time intervals of 10 s and 20 s as the effective forecast results and the shorter the time interval, the closer the forecast result is to the actual datum.