Introduction to Neural Network Applications Seminar Topic:
This paper discussed about the importance of chaos theory and how to use the chaos theory in system load forecasting. Short-term load forecasting is completed by using the load data of Shaanxi province power grid of China. Artificial Neuron Network (ANN) structure is determined by embedding dimension. ANN is established based on chaos theory and applied into the power system Train and forecast can be done with the help of improved back propagation algorithm (BP) which developed based on genetic algorithm (GA).
Overview:
Chaos thery is used to analyze the characters of load time series and employed in power system forecasting. Many non-linear time series forecasting models are developed based on chaos theory and those are used for power short-term load forecasting. Few of the popular models which are applied to power system load forecasting are regression model, fuzzy logic model, expert system model, time series model and grey-theory model.
Application and Analysis:
Standard BP neural network method uses uncertain structure of BP neural network in order to forecast the next-hour load. The results yield by BP is not very accurate and also not so quick. But by using the chaotic characters of the power load time series, the structure of BP network will get determined then the accuracy in BP network results will get improved. BP neural network model using chaotic characters of the power load time series is more effective in short-term power load forecasting when compared to that of classical standard BP neutral network model.
Conclusion:
All the characters of chaos are exists in power load time series and these are chaotic and nonlinear. Precision and accuracy of a power load forecasting system is increased greatly by introduced chaos theory in power load forecasting system. BP neural network, developed based on chaos theory which is used for forecasting to in order to increase the accuracy and the training rate.
Download Electrical seminar topic and Report on Neural Network Applications .