Table 3.
Summary of the ANN model.
| Neural network information | ||
|---|---|---|
| Input | Outdoor temperature [oC] | Hourly average value |
| City water temperature [oC] | Measured value | |
| Month | 1–12 | |
| Day | 1–31 | |
| Hour | 0–23 | |
| Day of the week | 1(Sunday) – 7(Saturday) | |
| Holiday | 0(non-holiday), 1(public holyday), 2(new-year's day, Thanksgiving Day) | |
| Consecutive holiday | The number of weekends or holidays out of the five days, including the day and two days before and after | |
| Active case in nationwide | log10(active case in nationwide area), 0 at no active case | |
| Active case in Seoul metropolitan | log10(active case in Seoul metropolitan area), 0 at no active case | |
| Output | DHW demand [MWh] | Measured value |
| Number of hidden layers | 3 | |
| Number of neurons in a hidden layer | 100 | |
| Activation function | ReLU | y = max (0, x) |
| Train-test data ratio | 7:3 | |
| Standardization | Z-score standardization | x' = (x − μ)/σ |