Design And Implementation Of Short-term Power Prediction System Of Small Hydropower Cluster

In recent years, the small hydropower as a clean energy has got a brand new opportunity to develop which has the advantages of the less investment, the quicker effect and the shorter development cycle etc. However, most of the hydropower stations are radial flow stations which are lack of regulation and in the actual operation as the small hydropower produces power as soon as the water coming, the output is volatile and irregular. When the small hydropower is connected into the grid, it will cause great load impact to the grid, which makes the main load more irregular and lower safety. Therefore, the short-term power prediction of the small hydropower is helpful to make reasonable generation schedules to ensure the safe and stable operation of power grid and the balance between supply and demand, which has theoretical and practical significance.In some mountainous area, the small hydropower's output is irregular, the installed capacity is small and it has a large number. In view of all feature, after analyzing the power characteristics of volatility, periodicity and probability distribution, a cluster has been proposed. Several hydropower stations access to the same substation which are called a cluster. It has better prediction results by using the method of cluster.As most of the small hydropower is located in the mountainous area separately where geographical environment is complex and it's unable to get accurate enough meteorological data of each site, in this article, we dig the information in the history output data, then calculate and forecast data according to finding the similar output sequence. Considering it's cumulative effect and hysteresis effect, the traditional concept of similar day is extended to the similar day interval which has more flexible time length. The coefficient of the normalized form is defined to measure the shape similarity of the two data sequences. According to analyzing the change trend of the power sequence after the similar day interval, the future power sequence can be calculated more accurately. The experiment results show that this method can get ideal results under the condition of complete historical power data.On the basis of the forecast model, short-term power prediction system of small hydropower cluster is designed and implemented by using SSH technology, J2EE framework, Tomcat as the application server, Oracle as the database. There are rights management, data management, power generator characteristics analysis, cluster management, power prediction and forecasting results analyzing in the system, and these function modules are able to satisfy multistage scheduling persons to forecast small hydropower under their jurisdiction. According to trial operation, it shows complete system function, stable operation, unified interface style, flexible interaction, reasonable forecasting time and rich chart shows. Therefore, the system has strong application value for raising the level of small hydropower Information management.