Monthly average hourly load profiles displaying incorrect information #32

shorowit · 2017-09-25T16:16:26Z

Posted at https://beopt.nrel.gov/node/935:

I love the 'Profile' function in Dview that allows quick creation of the average hourly load profile for each month of the year.

However I have sometimes had issues where I'm not sure it is displaying the correct average values. I have a good example of this for a project I'm working on currently. I have a client's hourly energy consumption data and want to look at the monthly average load profiles. I noticed particularly in July, August & September, the 2016 profile is significantly depressed compared to the 2015 profile. However looking at the overall average daily consumption for each month/year, this didn't make sense, as the months had similarly high overall usage each year.

I took the same data and used Excel's PivotTable function to create average hourly values for each month of each year and plotted them using Excel's graphs. As I thought, when I did this, the profiles track much more closely, I trust these numbers as I've done a lot of checking of them in Excel. Particularly in July and September, where the values are very close to each other for both years of data.

Whereas the values in the profile from DView seem to indicate that the consumption for 2016 would be roughly one-third of that in 2015. This is clearly incorrect. Interestingly, the shape of the graph is the same as the correct profile from Excel, it just seems to be somehow the magnitude of the values are scaled down, to approximately one-third of what they should be.

I've attached relevant spreadsheets, graph screenshots and csv files. Any ideas? Looks to me as if Dview is calculating the values as one-third of what they should be for one of the data series. Annoying as it's otherwise an incredibly useful and fast function - while I can do the same thing in Excel with Pivot Tables, it's a lot slower to do so.