Short-term load forecasting using system-type neural network architecture.

This thesis presents a methodology for short-term load forecasting using a system-type neural network based on semigroup theory. A technique referred to as algebraic decomposition is used to decompose a distributed parameter system into a semigroup channel made of coefficient vectors and a function channel made of basis vectors. The actual load data is preprocessed by regression to become better correlated to daily time and temperatures. A rearrangement method based on the hourly temperature is developed to solve the problem of the roughness of the coefficient vector in the seimigroup channel. Interpolation or extrapolation of coefficient vector can be achieved for each hour using the historical temperatures and the temperature forecast. Recombination of the basis vector and predicted coefficient vector will give the next-day load forecasting. Load data from New England Independent System Operator is used to verify the capability of the proposed approach.