The HOMER Pro Optimization Plot and Optimization Surface Plot helps you understand how your optimization variables affect the overall costs of your system. Once you have calculated results for a HOMER model, you can access these plots by clicking on the Results view and then choosing Graphical from the Tabular/Graphical radio buttons in the upper right of the HOMER Pro window. A drop-down menu below the Graphical radio button gives you several plotting choices, including the Optimization Plot and Optimization Surface Plot. The Optimization Plot and and Optimization Surface Plot will show you relationships between output parameters and optimization variables.
Optimization Plot example
Using the sample file “African Mini Grid for Energy Access,” we plotted PV array capacity versus Net Present Cost (NPC) for the 8% discount rate and 100 ₦/L cost of fuel. We right-clicked on the plot, chose “Properties” and changed the plot range to 0 to 250 kW for PV capacity (x-axis) and 190 to 240 million ₦ for NPC (y-axis). Each dot on this plot represents a simulation run in the HOMER engine. If we hover the mouse over a dot, HOMER will display the variable values for the simulation. We can see the optimum NPC occurs when there is 85 kW of PV. We can also see the NPC of simulations near the optimal system.
Now, we’ve optimized the same system using the search space for the PV component. This allows us to manually specify the PV sizes that we want to model in HOMER Pro. For the PV search space, we entered 0 through 100 in 10 kW increments, as well as 120 and 150 kW. We made the Optimization Plot again, pictured below.
>We can see each value in the PV search space corresponds to a column of dots. All the points within a column have the same value of PV capacity, but different values of converter, battery bank capacity, and dispatch. The dot at the bottom of each column represents the most optimal combination of converter size, battery bank quantity, and controller for that PV capacity. This gives us a feel for how the optimization space is covered very differently when using the search space instead of the optimizer. The Optimizer is very good at finding and targeting the optimal NPC simulation, whereas you can simulate a much broader range -- and specific sizes -- using the search space.
Optimization Surface Plot example
The Optimization Surface Plot adds a third dimension, so you can see how a resulting value changes with two optimization variables. This example also uses the sample file “African Mini Grid for Energy Access.” Below we’ve plotted net present cost (color) over 1 kWh LI battery quantity (x-axis) and PV size (y-axis). HOMER uses an interpolation algorithm to draw the colors. Each diamond represents a simulation. From the colors, I can see that the lowest NPC systems have 30 to 40 batteries and 80 to 90 kW of PV. I can also infer that the NPC of the system increases sharply with less than 20 kW of PV.
Because of the need to interpolate the colors between simulations--- and potentially over additional optimization variables --- this plot provides a limited snapshot of the results. For example, in the surface plot above, we had to specify the converter size, generator capacities, and the dispatch strategy. The actual HOMER results may not include the specified sizes, so HOMER will perform one interpolation to determine the NPC at each of the diamonds. HOMER will also interpolate between diamonds/ simulations to draw the colors. One option to reduce interpolation (and increase accuracy of the plot) is to specify a specific size in the search space of the converter and calculate again.