Research On Key Techniques Of Wind Power Monitoring System For Very Large-scale Wind Power Bases

The wavy character of wind power is enlarged when a Very Large-scale Wind Power Base(VLWPB) connects to the power grid. The character breaks the steady of the power grid. The production of a VLWPB must be monitored and predicted so that it can be adjusted correctly in time. The output of the VLWPB is therefore smooth and the VLWPB and connect to the power safely.A Wind Power Monitoring System(WPMS) for VLWPB plays a necessary role when the VLWPB connects to the grid. It must output accurate data in time so that the adjustment is correct. No monitoring system for VLWPB has been found yet after studying the research status of the WPMS. Therefore, the paper, based on the problems encountered during the construction of the WPMS of the Jiuquan 10 GW wind power base in Gansu province, focuses on wind electricity data integration, processing performance and high availability. The research concentrates on:1. The features of a VLWPB are studied and the organization and principles ofthe WPMS are studied. The indices are constructed and the formulae forwind power analysis are provided. A real-time, accurate and reliable WPMSis built.2. Wind electricity information islands exist. So the characters of the originalwind power data are studied. A modular data integration system is presentedfor wind power data acquisition, transformation and push, which is based ondata extraction, transformation and load principle. The system solves theinput problem of the monitoring system. It also makes the wind powermonitoring system a uniform platform of wind power data for the VLWPB.3. To eliminate the performance bottleneck caused by processing on largeamount of data, a hybrid message passing interface and shared memoryparallel processing model is presented. It reduces the time cost and theblocking risk caused by the message passing. The parallel module is involvedonly when the calculation scale is large enough. The parallel service, togetherwith the client program and the parallel database, forms a flexible dataprocessing system. The system accelerates the statistical calculations on largeamount of data without raising the budget apparently.4. A Dynamic Availability(DA) concept is presented in order to get the highavailability. The DA can indicate the working possibility of a system atmoment scale. A Dynamic Early Warning and Scheduling(DEWS) strategyis presented based on DA. With the help of BP neural network andmulti-thread technology, the DEWS can war or switch the working nodebefore an error occurs by predicting the DA dynamically. The monitoringsystem therefore gets high availability.The monitoring system is based on the key technologies. It fulfills the proposal such as management convenience, smooth power, electricity security and high extensibility. The monitoring system is a helpful tool which makes the power system stable after the connection of large-scale wind power. It is also a great importance for spreading the wind power and making it more competitive in the power market.