Advanced monitoring and evaluation of social housing renovations
Dutch social housing associations own 2.4 million dwellings, which is one third of the housing stock in the Netherlands. Their renovation strategies towards a CO2 neutral built environment in 2050 have a large impact on the lives of millions of inhabitants, which usually are low-income households. Housing associations want to do renovations that have a high impact on actual energy savings, related CO2-emissions, and related energy cost for their tenants.
We set up an advanced framework to monitor and evaluate the impact of renovations of social housing associations. Annually we gather building characteristics of over 2.0 million dwellings of social housing associations, combine them with actual energy consumption, and build a model that is able to assess the effects of renovations on actual energy savings. This is challenging, firstly, because housing characteristics and renovations vary a lot (we consider 25 characteristics with 90 parameters). Secondly, a huge variation in actual consumption exists due to the influence of inhabitant behavior. Thirdly, the search for a good modelling technique is challenging (we consider linear regression, non-linear regression and a machine learning technique (GBM). Fourthly, inconsistencies in (the availability of) data can have a large effect on model predictions and fifthly, validating research results is challenging.
We would like to present and discuss insights and challenges, from the data collection, the building characteristics we assess, the pros and cons of different modelling techniques, conceptual challenges arising from inconsistencies in the data and, the search for methods to validate model predictions. We are confident this research will provide models that will enable social housing associations to monitor, evaluate ánd predict the effects of future building renovations, contributing to reduced actual CO2 emissions and reduced energy costs for low-income households.
