| Canada is one of the countries that signed the Kyoto Protocol . If Canada ratifies the Protocol, it will be legally bound to reduce its greenhouse gas (GHG) emissions. One of the effective means of reducing the GHG emissions is reducing the end-use energy consumption and the associated emissions from the residential sector. This requires that the energy consumption characteristics of the residential sector, as well as the complex interrelated effects of energy saving measures that can be adopted to reduce energy consumption are well understood. To this end, detailed mathematical models of residential end-use energy consumption are required. So far, two types of models have been used to model residential end-use energy consumption. These are the Engineering Model, used to model the energy consumption at the national and regional levels, and the Conditional Demand Analysis (CDA) Model used at the regional level.; This thesis investigates the use of Neural Network (NN) and CDA methods for modeling residential end-use energy consumption at the national and regional levels. In this work, end-use energy consumption models were developed for the Canadian residential sector using the NN and CDA methods and the extensive data available in the 1993 Survey of Household Energy Use database of Statistics Canada. Although NN's have characteristics suitable for modeling residential energy consumption at the national and regional levels, no NN based model had been reported in the literature before the current work. Similarly, the CDA method had not been used to model residential energy consumption at the national level, although there are several studies where CDA was used to model energy consumption at the regional level. Thus, the NN and CDA models developed in this work are the first of their kind, and represent original contributions to the state-of-the-knowledge in energy modeling.; The prediction performance and the ability to characterize the residential end-use energy consumption of the NN Model and the CDA model are compared with those of an Engineering Model developed earlier by others. The effects of socio-economic factors and impacts of energy saving measures on the end-use energy consumption were estimated using the NN and the CDA Models, and, where possible, the results are compared with those of the Engineering Model.; A comparison of the estimates of the models showed that the NN Model has a higher prediction performance than the CDA and the Engineering Models. The NN Model was able to successfully estimate the impact of socio-economic factors and energy saving measures on the end-use energy consumption in the residential sector. Due to the limited number of variables the CDA Model can accommodate, its capability to evaluate these effects is significantly lower than the NN Model. While the Engineering Model has the highest level of flexibility amongst the three models to evaluate energy saving measures, it cannot deal with socio-economic factors since these factors are not accommodated by the thermodynamics/heat transfer based model.; The results of this work show that the NN Model can be used to estimate the end-use energy consumption in the residential sector, to categorize the household and end-use energy consumption, and to evaluate the effects of a large number of socio-economic factors and the impacts of energy saving scenarios on end-use energy consumption. The CDA model, while simpler and easier to use than the NN model, has limitations that limit its usefulness. |
