Capacity Planning Tutorial

This tutorial provides a step-by-step guide to include investment constraints for capacity planning in a simple energy system with Spine Toolbox for SpineOpt. There is more information to be found in the documentation on Investment Optimization. To get the most out of this tutorial, we suggest first completing the Simple system tutorial.

Overview

In this tutorial we will:

• start from the simple system tutorial,
• change the temporal structure from one day to seasons of four months (4M) over the years 2030-2035,
• add a temporal block for investments,
• and add investment related parameters for the units.

In the second part, we extend the model to multi-year investments, with investment decisions in both 2030 and 2035.

Starting point

1. Create a workflow in Spine Toolbox with an input database, a Run SpineOpt tool and an output database, as on the Verify installation page.
2. Double-click the input database to open it in the Spine DB editor.
3. Select File > Import... and import the simple system example file.
4. Press Commit in the toolbar.

In the steps below, enter all values in the Parameter value table with the alternative Base. For values marked (drop-down), double-click the value cell and pick the value from the drop-down list. For values marked with a type, such as (Duration), right-click the value cell, select Edit... and choose that parameter type.

Temporal structure

For the investment optimization, we use a longer time horizon, 2030-2035. To keep the number of time steps small, the operation is modelled in seasons of four months (4M). Investments are decided in a separate temporal block with a resolution of five years (5Y), i.e. once in 2030 and once in 2035.

1. In the Entity tree, right-click the temporal_block class and select Add entities...
2. Enter the name investment and press OK.
3. Right-click the model__default_investment_temporal_block class, select Add entities..., add simple | investment and press OK.
4. Right-click the model__default_investment_stochastic_structure class, select Add entities..., add simple | deterministic and press OK.
5. Enter or change the following values (values marked with * are changed from the simple system):

1. Press Commit.

Unit investment parameters

Next, we make both power plants investment candidates. In the Entity tree, select the unit class and enter the following values:

What the values do:

• existing_units 0 and investment_count_initial_cumulative 0: the units do not exist unless the model invests in them.
• investment_count_max_cumulative 1: the model can invest in at most one unit of each power plant.
• investment_variable_type integer: the investment is a whole number of units. With the default linear, the model would invest in 0.25 units of power_plant_b. The unit size comes from capacity_per_unit of the unit__to_node entity: 0.25 × 200 = 50, which is exactly the flow needed from power_plant_b.
• lifetime_technical 10Y: a new unit must exist for at least 10 years after the investment. lifetime_economic is used to adjust the investment cost; here it is equal to the technical lifetime.
• unit_investment_cost 1000: the cost of one new unit.
• units_on_cost 0: being online costs nothing.

Examine output

To see the investments in the results, add them to the report:

1. Right-click the report__output class and select Add entities...
2. Add report1 | units_invested and report1 | units_on and press OK (see also How to manage Output Data).
3. Press Commit and run the SpineOpt tool.

Expected result for units_invested:

Both power plants are needed to cover the demand of 150 (100 + 50), so the model invests in one unit of each in 2030. In 2035, they still exist, so no new investments are needed.

The units_on values of power_plant_b may vary between 0.25 and 1.0. units_on is a continuous variable here and being online costs nothing, so any value large enough to cover the flow gives the same total cost.

Multi-year investments tutorial

Multi-year investments refer to making investment decisions at different points in time, such that a pathway of investments can be modelled. This is useful when long-term scenarios are modelled, but modelling each year is not practical, or when investment decisions are made in different years, which affects the cash flow.

In this part, we consider two investment points: 2030 and 2035. The operation is modelled in seasons of four months, but only in 2030 and 2035. In other words, 2030 and 2035 are milestone years for the pathway 2030-2035. We continue from the model of the first part.

Temporal structure

We replace the flat temporal block with two operation blocks, one for 2030 and one for 2035.

1. In the Entity tree, right-click the temporal block flat and select Remove.... If the old block is left in the model, the temporal structure will be wrong.
2. Right-click the temporal_block class, select Add entities..., add operation2030 and operation2035 and press OK.
3. Right-click the model__default_temporal_block class, select Add entities..., add simple | operation2030 and simple | operation2035 and press OK.
4. Enter the following values:

(* already set in the first part)

What the values do:

• block_start and block_end leave a gap between the years 2031 and 2034, which are not modelled.
• operation2035 starts already in the last season of 2034 (2034-09-01). This extra time slice sets the boundary conditions for the block, see the info box below.
• weight 5 for operation2030: this block represents the operation in 2030-2034, so its operating costs are multiplied by five. operation2035 only represents 2035, so its weight is 1. Without weights, the operating costs in the input data would have to cover the whole period.

Demand

We assume that the demand grows from 100 in 2030 to 400 in 2035. The demand changes from a constant to a time series:

1. Select the node class, right-click the demand value of electricity_node and select Edit...
2. Choose the parameter type Time series variable resolution and enter:

The value 0 at 2034-09-01 is for the extra time slice that sets the boundary conditions of operation2035.

Investment limits

We allow investments in power_plant_a in both 2030 and 2035, and in power_plant_b only in 2035. For that, investment_count_max_cumulative becomes a time series:

1. Select the unit class, right-click the investment_count_max_cumulative value of power_plant_a and select Edit...
2. Choose the parameter type Time series variable resolution and enter the values below. Do the same for power_plant_b.

Note that the value is cumulative: it includes the units invested in earlier. For power_plant_a, 2.0 in 2035 means one more unit on top of the one from 2030, not two more.

1. Press Commit and run the SpineOpt tool.

Results

Expected result for units_invested:

In 2030, one unit of power_plant_a covers the demand of 100. In 2035, the demand of 400 needs a second unit of power_plant_a (2 × 100) and one unit of power_plant_b (200).

The units_on results show one unit of power_plant_a online in 2030 and two in 2035, and one unit of power_plant_b online in 2035. The value 0 at 2034-09-01 is the redundant boundary time slice and can be ignored.

Debugging

The input database at the end of this tutorial is available as the capacity planning example file. You can import it into an empty input database to compare it with your own.