STATISTICAL MODELLING FOR SIMULATING ELECTRIC ENERGY CONSUMPTION OF RESIDENTIAL USERS IN MENDOZA (AR)

1.1. SUMMARY The annual electrical consumption billing records show variability among users, which are clustered into five types. In addition, the difference of billing system between the two departments, Mendoza Capital, and Godoy Cruz, the 12-month and 6-bimonth billing with the missing of many bills, especially in the Godoy Cruz department. The missing bills are expected to refer to temporary use of a dwelling, which mainly refers to active rental activity and/or temporary accommodation. While most billing records and annual aggregation in case of Mendoza Capital department give more stable insight for the users that inhabit the area. About 70,300 users selected from a total of 113,000 recorded bills in year 2016 for the modelling phase, where the data is divided into two groups, one for the aggregated bimonthly-related processing that considers seasonal temperature effects, while the same data aggregated in annual basis to share in the modelling that takes into account the socio-demographic variables on census section scale. The results of the two models show significant correlation to the dwelling of rental activity with the room crowding ratio for each census section in the Multiple Linear Regression model of socio-demographic variables. While on the linear Regression model for seasonal change, the consumption is highly correlated to the change of mean temperature. 1.2.??CONCLUSION Analysing electricity consumption data using LR and MLR models according to the type of variables that share in the model, as the proposed model gives evidence for the impact of rental activity with home crowding, that show how electrical consumption is related to the ownership of dwelling and low room crowding, when related in terms of annual consumption per person, and on the other hand for the LR model, the effect of seasonal temperature change on about 60% of users that mainly use cooling in summer times, where the change of mean temperature has effect on seasonal consumption of electricity. 1.3.??RECOMMENDATIONS This study can be used by the municipality of Mendoza, especially the planning and infrastructure departments, to better estimate and plan future extensions and/or renovation processes to efficiently predict resources needed for the future demand of electricity supply for housing sector and residential projects. The prediction of demand is essential for network design and future energy production. In addition, policymakers can benefit from the results by setting rules that regulate and better moderate electricity consumption, one way by using the power of tariffs and subsidies and better targeting prices to users according to dwelling activity, or another way by offering incentives for self-production of energy by solar power and/or compensating consumption by encouraging the use of more electricity-efficient appliances. 1.4.??FUTURE WORKS The better the input data, the more efficient the results of prediction with minimal prediction error. Right from this point, I would recommend conducting field survey for random set of population and household to collect more information about the direct consumption of electricity in terms of devices and appliances. The study of load profile, along with accurate metering for home appliances on a daily and hourly interval basis, gives more accurate modelling for electrical consumption, where the variability of consumption appears during the whole day between day and night times, and shows how the consumption varies between weekdays and weekends.