Chiller Load Forecasting Using Hyper-Gaussian Nets

Energy load forecasting for optimization of chiller operation is a topic that has been receiving increasing attention in recent years. From an engineering perspective, the methodology designing and deploying system should take into account several issues regarding prediction horizon, available data, selection variables, model adaptation. In this paper these are parsed to develop neural forecaster. The method combines previous ideas such as basis expansions local models. particular, hyper-gaussians proposed provide spatial support (in input space) models can use auto-regressive, exogenous past errors constituting thus particular case NARMAX modelling. Tests using real data from different world locations given showing expected performance proposal with respect objectives allowing comparison other approaches.