文摘
Proposed methodology for the modelling of the energy use of small appliances in households, categorised in four groups: audio-visual, computing, kitchen and other appliances. Predictions of time varying fractional power demands of aggregations of appliance. Comparison of seven different modelling approaches, based on two types of stochastic process: discrete time random (Markov) process (combined with clustering algorithms for matrix reduction) and continuous time random (Survival analysis) process. Models are validated using three types of evaluation measure: time series analysis, model accuracy and aggregated energy use. These validation measures suggest that the modelling of survival in discrete states performs well and is straightforward to integrate with energy simulation programs.