Q: Hello, My name is Sharon and I am from Dubai. My question is with regard to the M&V Plan. I have a project that started in Aug 2017. I selected Option C for the Chiller MDB savings with dB and wB as independent variables. Recently, I came across this article https://evo-world.org/en/news-media/m-v-focus/868-m-v-focus-issue-5/1164-why-r2-doesn-t-matter, in which they advocate "the independent variables need to be independent of one another. For example, Tdry bulb and Twet bulb are not independent – they are highly correlated. Use Tdry bulb and Tdew point instead. Or better yet, use enthalpy H." I tried dB and H as my new independent variables over the same 2017 dependent variable(kWh) and got better results. My question is, can I change/revise the regression model/equation in between the project as the EPC is for 6 years. And if I can, please guide me to an article that says that.

A: Regarding IPMVP adherence and changing the model, I think you are fine to change the model if all parties to the EPC agree.  For the modeling approach and the linked article stating, "the independent variables need to be independent of one another. For example, Tdry bulb and Twet bulb are not independent – they are highly correlated. Use Tdry bulb and Tdew point instead. Or better yet, use enthalpy H." It appears to me that the author's intent in writing that was to use enthalpy as a single independent variable, not combined with Tdb. I haven't checked, but I believe enthalpy is more highly correlated with Tdb than Twb is correlated with Tdb. That said, I think using Tdb and Tdp might be the best approach. It should depend upon the system and type of chiller, and perhaps the control sequence.  So I suggest trying one model using only enthalpy as an independent variable, and one model using Tdb and Tdp, and see if the latter provides a notable improvement, and that the Tdp variable is statistically significant.