Abstract

The lack of mastery of housing problems in developing countries causes difficulties in the forecasting of the demand for potable water, which in turn causes waste and high cost of water. We modeled the forecasting of the demand for potable water in a city, taking into account the housing typology. The latter depends on parameters such as: household income, size of households, price of water, annual rainfall height, the average summer temperatures, the number of households and the number of subscribers to the existing water distribution network. The stratification of a town was done using a definition of the housing typology and using multiple linear regression, the total potable water consumption was estimated to depend on the aforementioned parameters. This was applied to Yaoundé in Cameroon for the years 2020 and 2035, and the results obtained yield forecasts for the demand for potable water of 115 778 781 m³ and 281 815 641 m³ respectively, with a distribution yield of 70%. This leads to the deduction that the demand for potable water, as well as investment requirements as previously assessed were overestimated, thus restricting new extensions in the districts where the drinking water problem of demand for drinking water was solvable.

Key words: Housing typology, forecast, consumption, modeling, multiple linear regressions.