No abstract

No abstract

Abstract:Predicting the future has always been one of mankind's desires. In recent years, artificial intelligent techniques such as Neural Networks, Fuzzy Logic, and Genetic Algorithms have gained popularity for this kind of applications. Much research effort has been made to improve the prediction accuracy and computational efficiency. In this paper, a hybridized neural networks and fuzzy logic system, namely the FeedForward NeuroFuzzy (FFNF) model, is proposed to tackle a financial forecasting problem. It is found that, by breaking down a large problem into manageable "chunks", the proposed FFNF model yields better performance in terms of computational efficiency, prediction accuracy and generalization ability. It also overcomes the black art approach in conventional NNs by incorporating "transparency" into the system.

No abstract

Dynamic reactive power compensation equipment typically requires a fast response to output the necessary reactive power. The term"dynamic response time of reactive power"is often used but has never been clearly defined. This paper summarizes the reactive power calculations under different definitions and algorithms and considers these calculations in terms of signal processing to simulate and analyze the step response. This paper subsequently focuses on the widely used instantaneous reactive power algorithm and finally concludes that the dynamic reactive power response time closely depends on the reactive power calculation method itself. The single-phase instantaneous reactive power algorithm has the fastest response time. The reactive power response time of dynamic reactive devices in power systems is a minimum of a quarter of one cycle time for the well-known and widely used single-phase reactive power algorithms.