Optimisation Basic Tutorial

This tutorial will help you to learn the basic information needed to run an optimisation study in DesignBuilder in a few easy steps.

Step 1 - Create a base model

Create a new file located in London Gatwick and add a building to the site with a simple rectangular block having dimensions 30m x 20m as shown below. Use default Model options and template settings. DesignBuilder as supplied will use a Fan Coil Unit HVAC system as the default. For this example you should make sure that you have heating and cooling selected and no natural ventilation (this will be the case for a new model with a FCU HVAC system selected).

Step 2 - Run a standard simulation

Click on the Simulation tab and run a base annual simulation. Because it is a simple model you can select hourly results. Make sure to also choose Monthly results which are required by the Optimisation. Check the hourly results for the simulation period and make sure that the model is behaving as expected, including temperatures within the building, operations periods etc. If not fix the model and repeat this step until you are happy with the base model hourly results. Monthly results should something like the screenshot below.

 

Step 3 - Define optimisation problem on Optimisation Analysis Settings dialog

You are now ready to start the optimisation stage. To do this click on the Optimisation tab on the Simulation screen. Because you don't have any results yet the Optimisation Analysis Settings dialog is displayed. This allows you to define the optimisation problem, i.e. what it is that you want to achieve from the optimisation study. Use the default settings for this simple example which request an optimum set of solutions to minimise both operational construction emissions and occupant discomfort allowing for variations in window to wall ratio and heating and cooling setpoint temperatures.

 

Objectives:

 

  1. Minimise operational carbon emissions
  2. Minimise discomfort (ASHRAE 55)

Constraints:

 

None

 

Design variables:

Allow WWR and cooling and heating setpoints to vary as shown in the screenshot below. Variations are made at building level for all variables in this case as shown by the Target objects column.

 

Step 4 - Open Optimisation Calculation Options dialog

Having confirmed the Optimisation analysis options the next dialog to open will be the Optimisation calculation options. Use all default settings (which are generally suitable for small analyses like this).

Step 5 - Start optimisation analysis

Once you have completed your review of the options press the Start button in the bottom right of the screen. The optimisation process will involve running a lot of simulations. If you kept default settings there will be 200 generations with a population size of 10 in each. That means 200 x 10 = 2000 simulations! For our simple model and when using the Simulation manager to run these in parallel this shouldn't take too long. However if time is an issue it is usually worth keeping an eye on the solutions as they come in to check whether convergence has been achieved.

 

While preparing this tutorial we made the judgement that the solution was effectively converged after about 100 generations and pressed the Stop button at that point. If the optimisation had been left to continue the red Pareto front line would have been filled in further however this would not provide much extra information for the purposes of our example test. After pressing Stop the current generation continues to simulate and once that is finished....

Step 6 - Analyse results

 

Try using some of DesignBuilder's analysis tools to investigate the best settings in this case.

 

  1. Drag the Status header into the dark grey area above the grid where it says "Drag a column header here to group by that column"
  2. Sort the grid data by Discomfort by clicking on the "Discomfort" header.
  3. Scroll down to the bottom of the grid where you will find the optimal "Pareto Front" options coloured green. These are the design configurations which most successfully achieve the design objectives (minimise carbon and discomfort in this case). Because you sorted by discomfort you will see results either in ascending or descending order depending on how many times you clicked on the header. You will notice that when the discomfort results are in ascending order the carbon emission results in the next column are in descending order and vice-versa.
  4. All points on the Pareto front have 20% (or a little more) glazing . This is because it was the lowest allowed value in the analysis and higher values increase both discomfort and cooling loads (hence carbon emissions).
  5. Heating setpoint temperatures range from 18°C (lowest carbon emissions) to 23°C (lowest discomfort).
  6. Cooling setpoint temperatures range from 24°C (lowest discomfort) to 27°C (lowest carbon emissions).
  7. The easiest way to identify a balanced result is to click on a point on the Pareto Front which gives to right degree of emphasis to the 2 objectives. The corresponding item is highlighted in the grid giving the corresponding design settings.

 

Next steps to learning. Try adding more design variables. Glazing type, shading systems etc. Try using construction cost as a variable instead of discomfort.