A variation between the peak load for a typical day and hourly time-series peak load of the entire year.

[adnanalakori]

Dear @Bachibouzouk , there is a variation between the peak load for a typical day and other peak load of the hourly time series for the whole year for a fridge.
Peak load for a typical day is 3000W which is fine and fits the input load, while the peak load of the hourly time series generated for the whole year is only 410 W.

I have tested another load e.g. lamps and the results are fine.

input_file_1.xlsx

Attached an input file.

[Bachibouzouk]

Thanks for the bug report @adnanalakori, I will look at it when I have a moment!

[a-linke]

Dear @Bachibouzouk and @adnanalakori we observed the same issue. In our case we had 4 different yearly load profiles and in every case the peak value of the minute resolution is twice as high as the peak values of the hourly resolution

[a-linke]

@Bachibouzouk and @adnanalakori: What do you think about solving this via using the max() resampling instead of the mean() resampling for the hourly resolution?
df_chps = pd.read_csv("yearly_profile_min_resolution_chps.csv", index_col=0) df_chps.index = pd.to_datetime(df_chps.index, infer_datetime_format=True) chps_peak=df_chps.resample("H").max() chps_peak.to_csv('demand_hourly_peak_chps.csv')

[Bachibouzouk]

@a-linke - I agree with you that for peak demand it makes more sense to use .max() rather than .mean(), the averaging smoothes the peak. Every resampling is a loss of information, maybe we should resample and keep the average, median, standard deviation, min and max values of the load profile during one hour, thus we go from 3600 data points to 5 characteristic numbers.

[Bachibouzouk]

One could do this by adding columns for each of these 5 numbers (so the csv would have the hourly timesteps, then 5 columns with those values)

[adnanalakori]

@Bachibouzouk and @adnanalakori: What do you think about solving this via using the max() resampling instead of the mean() resampling for the hourly resolution? df_chps = pd.read_csv("yearly_profile_min_resolution_chps.csv", index_col=0) df_chps.index = pd.to_datetime(df_chps.index, infer_datetime_format=True) chps_peak=df_chps.resample("H").max() chps_peak.to_csv('demand_hourly_peak_chps.csv')

Make sense. Will try it and check the results.

[adnanalakori]

@a-linke and @Bachibouzouk, Yes, .max() seems ok.

[Bachibouzouk]

Ok so I will add the column to the output csv before I close the issue