Time slicing

NEMO provides a flexible mechanism for dividing years into sub-annual periods and modeling energy demand and supply in those periods. This approach can allow a more realistic simulation of fuels for which the timing of demand and supply is critical (e.g., electricity). The subannual periods are called time slices, and they're one of several dimensions in a NEMO model.

Time slice widths

The width of time slices as a fraction of a year is set with the YearSplit parameter. Within a given year, the sum of all time slice widths should be 1.

Enabling time-sliced modeling of a fuel

You can specify whether the modeling of a fuel is time-sliced through NEMO's demand parameters. If you define a demand with AccumulatedAnnualDemand, NEMO will model it at the annual level - that is, it will assume the demand must be met in the indicated year, but the timing of supply within the year doesn't matter. Conversely, if you set up a demand with SpecifiedAnnualDemand and SpecifiedDemandProfile, the demand is assigned to particular time slices and must be fulfilled in those slices.

Storage and time slice grouping

Time-sliced modeling is also crucial for simulating energy storage. In this case, both the width and the chronology, or ordering, of time slices are important. Chronology counts because the energy available in storage depends on the sequencing of prior charging and discharging.

NEMO addresses the issue of chronology by using three parameters to translate from time slices to an ordered set of hours within a year: TSGROUP1, TSGROUP2, and LTsGroup. TSGROUP1 and TSGROUP2 are hierarchical groupings of time slices, and LTsGroup assigns slices to TSGroup1 and TSGroup2.

  • TSGROUP1 - These groups are nested within years - i.e., they are major divisions of a year, such as seasons or months. The order field of TSGROUP1 defines their chronological order. The first TSGROUP1 in a year should have an order of 1, and the order should be incremented by 1 for each subsequent group.

  • TSGROUP2 - These groups are nested within TSGROUP1. They represent divisions of TSGROUP1 such as days of the week. TSGROUP2 also has an order field to define the chronological order within each TSGROUP1. The first TSGROUP2 should have an order of 1, and the order should be incremented by 1 for each subsequent group.

  • LTsGroup - This parameter maps time slices to TSGROUP1 and TSGROUP2. To enable storage modeling in NEMO, you must use LTsGroup to assign each time slice to one TSGROUP2 within one TSGROUP1. The lorder field defines the ordering of time slices within each combination of TSGROUP1 and TSGROUP2 (e.g., for a month and day of the week).

TSGROUP1 and TSGROUP2 also have a multiplier attribute that NEMO uses when constructing an ordered set of hours within a year. The application of the multipliers rests on an essential characteristic of NEMO: for any storage, the rate of net charging is constant within a time slice (i.e., over all hours of the time slice).[1] This design allows NEMO to build up a set of ordered hours from the group and timeslice orders and the group multipliers. In plain language, the process is as follows.

  1. Start with the first TSGROUP1 and the first TSGROUP2 within it.
  2. Take the first hour of each time slice in the TSGROUP1 and TSGROUP2 (in the order specified in LTsGroup).
  3. Assume this block of hours repeats TSGROUP2.multiplier times.
  4. Move to the next TSGROUP2 within the TSGROUP1, and repeat steps 2-3. Continue through all TSGROUP2 within the TSGROUP1.
  5. Assume the set of hours constructed in steps 1-4 repeats TSGROUP1.multiplier times.
  6. Repeat the preceding steps for each subsequent TSGROUP1.

This approach means the following identity should hold.

$\sum^{tg1}[\sum^{tg2}[[\sum^{l_{tg1,tg2}}1] \times multiplier_{tg2}] \times multiplier_{tg1}] = 8760$

The multipliers for TSGROUP1 and TSGROUP2 should be set accordingly. For example, suppose that:

  • TSGROUP1 represents two seasons (each covering half of the year).
  • TSGROUP2 represents weekend and weekday periods (in a calendar with two-day weekends).
  • There are 96 time slices, representing the 24 hours of a weekday and the 24 hours of a weekend day in each season.

In this case, the multiplier for the weekend period should be 2, the multiplier for the weekday period should be 5, and the multiplier for each season should be $\frac{8760}{2 \times 7 \times 24}$ ≈ 26.07.

  • 1This principle also holds for energy production, consumption, and demand, when time-sliced: the rate at which each of these occurs does not vary within a time slice.