Trial / Demo
Load tracking, stratification, heat gains, conduction, inter zonal air movements, plant specification or sizing, control strategy, computation restrictions. These are all reasons for zoning within any dynamic simulation model. How are you zoning your models?
The architectural design
Zoning is critical within any model. You cannot just replicate your architectural plan and expect it to be perfect for energy or regulation compliance analysis. Zoning will be entirely up to you, the analyst / engineer, to decide where and where not it is applicable to subdivide your architectural design. There is a fine line; too many zones and the model becomes over complex, too few and the detail and therefore accuracy of results is lost.
There are no set standards to building geometry creation - there is guidance out there but it’s generally different guidance for differing calculation methodologies. That’s exactly how it should be. Each calculation process uses the data from your model geometry in a slightly different way. Whether it is you completing energy efficiency, daylighting, building regulations, airflow, thermal comfort or HVAC design calculations; all models need to be sub-divided effectively.
Maybe there is not enough guidance out there. The following points are some of my thoughts on this area. I hope they help some of you maybe think about your model in a slightly different way.
Building zoning
Simulations are essentially a solution to a set of equations based on the variables you input. How you input the specification into the model will all be dependent on the zoning of your geometry, and how the equations take into consideration these inputs are therefore significantly influenced by the zoning.
We all know you should subdivide your building to take into consideration difference in specification. Let’s take for example a core building zone. We will subdivide this area to take into consideration a difference in specification through the WC’s, circulation and lift areas. But you may not subdivide all the toilets; these could be included as a single zone, but only if the specification says so. The thermal loss or gain between identical zones is negligible and therefore for an energy model you could model these areas as a single zone. However you may have a change in design. Maybe an architectural change such as an additional window, or maybe a difference within the M&E strategy, an extract system difference, different fan coil size or lighting control strategy. This change will in turn change the results. Then we need to have a different volume to assign or simulate this difference in information.
The same process is required for aligning your zoning to the controls or sensor locations. Increasing the accuracy of the design will mean a more representative output.
You also need to remember you only have a single computation point per volume, meaning we only have this single virtual sensor to take into consideration all these details accurately.
A single computation point
This is the main point people tend to forget. Once you consider your model and each of its zones to have a single computation point i.e a thermometer sitting in the centre of the zone recording temperature, it really makes you consider what detail the simulation tool and therefore equations are considering in a single instance.
CIBSE Guide B suggests when considering large open plan offices, the floor area should be zoned so that differences in zonal heating and cooling load density (W/m2 of floor area) can be identified and catered for in the designer’s layout of the HVAC system. We understand perimeter zones will obviously experience heat losses as well as solar heat gains whilst internal zones will be subject to internal heat gains only, so why would we have a single zone for the whole open plan space? We need to consider what consequences this can have on our results. Does this change your lighting control strategy? Does it affect your conduction loss or gain through to adjacent spaces? Will the loads and therefore capacities of plant within these adjacent rooms change if you modify the zoning within this open plan space? This may be just as simple as an L shaped room. We know orientation of a room can make a large difference in loads, so is it right we have a single point to consider multiple orientations at the same time?
I am not saying you need to subdivide all large zones into many smaller spaces, but you always need to consider what effect your zoning is having on your results.
Large model spaces
Maybe a better example is an atrium. I have seen plenty of geometry models where people have modelled the atrium as a single zone. How is a single virtual sensor meant to show you any accuracy if you have a single result for the whole space? You may have ventilation plant at ground floor level, lighting gain at the first floor level, maybe openings at roof level, stack effect, and different internal conduction gains from adjacent spaces as you travel up the internal volume. How can a single calculation or a single zone result show you all this detail?
CIBSE AM11 gives some guidance on the modelling and data assignment details within an atrium.
Anticipate the detail
Stepping away from the software and without running any simulations, you should expect a difference in condition as well as results as you go up through the atrium. I have heard people say: “well, we are not taking into consideration the landlord zones”, or “this area falls out of our demise”. This may be true, but this atrium will be influencing the conditions in the adjacent spaces that may be within your demise. Having a single result for the whole volume will mean a similar conduction loss through to the adjacent spaces, whether this is through balconies to the adjacent offices or through internal walls to the adjacent apartment. Using a single zone may not be taking into consideration the accumulation of warm air at the top of the atrium giving larger conduction gain to the surrounding adjacent rooms. Or maybe the lower level ventilation could be giving a benefit to all areas.
And so my advice is think about the operational building, visualise yourself in each area of your design: Would you feel different? What factors will cause that difference? All these differences need to be included within your model.
Voids
This subject stems from the question “When should I model voids?” When to model a separate volume or to include the void within the occupied volume? When to model detail into the roof? Again, this does all depend on what you are using the model for. What are you trying to calculate? I have reviewed models which incorporate a 1m void within their occupied rooms. This resulted in a building occupied volume difference of around 30%, leading to 30% more load which had to be catered for in the servicing strategy. I have also reviewed models which have included a serious amount of detail into the zoning, bulkheads, floor and ceiling voids as well as detailed roof areas and zoning, taking three hours to complete a simulation I expected would take under half an hour, with no real difference in end plant sizes or energy consumption.
Guidance
We plan to review zoning in one of our online Upskilling sessions which we will add to our on-demand series of video tutorials. But in the meantime, there is some guidance on geometry zoning out there, mostly driven by regulations compliance requirements:
VE Tools
The VE software includes tools to aid in the zoning process. The ModelIT perimeter zoning tool will enable some quick automated zoning, the void splitter and storey creator will enable fast input of voids and replication of similar geometry. The partition tool is essential for the dividing up of existing geometry. Look out for the new ModelIT tools within the VE 2019 Feature Pack 2 update. There are some great updates to enable healing and modification of existing geometry.
The Space Groups facility in ModeIT is an essential part of defining the model, assigning of zones and assigning difference in specification.
I hope my thoughts on zoning will help some of you to consider your models in more or even less detail, before creating them. Consider what effect you are having on the results. Remember, an Architectural or BIM model will not be defined for simulation. Think about the physics and the calculations being run in the background. Visualise yourself within each space and consider your surroundings.
I am sure we have all heard different sayings about the standard of inputs affecting the standard of outputs. Well, this is an area that may not be defined by known guidance, and it is very dependent on the design specification as well as the calculation type. It really is up to you to decide to best way to model your building and therefore you decide how these calculations are processed.
If you would like further information on the creation of building models for energy, daylight or architectural design, please contact sales@iesve.com for details.
If you would like to review any of my previous Top Tips click here.
If you have any ideas or requests for Top Tips to help engineers get the most of the IES Virtual Environment, please let Dave know: david.pierce@iesve.com.
US
