HEAT LOAD CHARACTERISATION

Side project by Roberto Garay Martinez. https://github.com/robgaray/ https://robertogaray.com/

Overall inspection and modelling of head loads in buildings.

INSPIRATION. Over the years, I have worked together with several colleagues and students in experimental heat load characterisation. With a variety of approaches and datasets. This script tries to put some of this knowledge in a more coherent way. Obviously, this could not have been possible without my previous works with (in chronological order) Beñat Arregi, Mikel Lumbreras, Antonio Garrido, Markel Eguizabal, Ivan Flores, Olaia Eguiarte and Alaeddine Hajri. I have probably discussed these methods with several other researchers in the course of the last 4-5 years, but I sincerely can not remember with whom. To all of them, thank you.

CONTEXT EVOLUTION. The method proposed here takes its roots im methods such as PRISM (1970-80s)and ASHRAE changepoint (1990s). I initiated these kind of analysis around 2015. Most of the dataset I use now are less than 5 years old. In this period of time, the available data has increased exponentially, both in terms of availability, resolution and reliability.

DATA SOURCES. Based on data provided by GREN TARTU & used in several research papers & seminars. Climate data is taken from a a publicly available source at the University of Tartu. Special thanks to both of them.

PROCESS. Data from ONE building and local climate is used and the following process is performed.

• Data Inspection

• Development of a changepoint model

• Identification of outliers. Based on the residuals of changepoint model

• Data imputation for 1h & 5h gaps. Based on boundary of the data on the same day in the preceding and following week

• retraining of changepoint model

• Statistics