Cooling Energy Increase after Reducing Nominal Lighting Power
Dear Community,
I hope everyone is doing well. I'm currently trying to replicate this study: "Im, P., R New, J., & Bae, Y. (2019, September). Updated OpenStudio (OS) Small and Medium Office Prototype Models. In Building Simulation 2019 (Vol. 16, pp. 1311-1317). IBPSA.". I'm using the DOE reference model for medium office building for Tampa 2A, 2013.
After I developed the baseline model (updated the spaces and corresponding model components based on the DOE modle), I started to iteratively update the four model parameters mentioned in this study based on this figure, 
I first redistributed the lighting power density using a space-specific scheme—let's call this model v1. In the baseline model, the lighting power density is uniform across all zones at 10.19587 W/m². I calculated the total nominal lighting power by multiplying each space’s area by its lighting power density. The baseline model has a total lighting power of 50,797.35 W, while model v1 has a slightly lower total of 49,947.00 W—a 1.67% decrease after the update. However, I noticed something counterintuitive when I reviewed the end-use results: cooling electricity consumption increased slightly, from 622.99 GJ to 623.15 GJ, despite the reduction in nominal lighting power. This building has three floors, each served by a dedicated air handling unit (AHU), with each AHU containing a DX cooling coil to meet the cooling load for all zones on that floor.
To understand the reason behind this observation, I first used the ideal load air system to examine the building dynamics after redistributing the lighting power density. The results made sense to me: after redistribution, district cooling energy decreased from 1993.59 GJ to 1983.08 GJ, while district heating energy increased slightly from 42.89 GJ to 42.94 GJ. Based on this, I assumed that the issue lies with the air handling unit models. Please note, nothing else was changed except for the lighting power density between baseline and v1.
I then checked the Cooling Coil Total Cooling Rate and Cooling Coil Electricity Energy and Cooling Coil Runtime Fraction . I noticed for all three DX coils, the average runtime fraction increases: ground floor - 0.5318 to 0.5324, middle floor - 0.4325 to 0.4331, top floor - 0.4233 to 0.4236. For Cooling Coil Total Cooling Rate and Cooling Coil Electricity Energy, I noticed that: the average cooling coil cooling rate and cooling coil electricity energy for ground floor slightly decreases, but for both middle and top floor, they increase. I'm stuck here wondering what should I further do to justify why the middle and top floor consumes more cooling energy. I also found that the cooling unmet hour decreases from 381.17 to 374.67.
I would greatly appreciate any insights you can offer on how to address this problem and which additional output variables might be helpful. Thank you so much for your time and support!
Does the baseline model have different envelope specs? If so, its possible that the thickness of the walls has reduced the floor area, which will reduce the lighting power density and also reduce the required cooling.
Hi Greg, thank you for your response. But the building envelop specs are the same and the floor area is also the same, and my cooling energy actually increases after the nominal lighting power decreases.
why is the updated model showing less lighting when table 5 shows that it should increase by 13%. Further, Table 5 in your post, shows that all category internal gains have increases except small office electric equipment. These would certainly increase the cooling energy.
If you're running simulations after each change and all things are held constant with lighting being the only changed variable, then I agree the results don't make sense. However, I'm sure something is a miss and once found, the cooling energy will make sense. Check the energy end use table for both models and ensure match.
Hi Greg, I didn't fully followed this study so there is a mismatch between the lighting power density defined in the DOE model and the model the author used. I assume that the authors used the OS measure to create the baseline model, which has the LPD of 8.83W/m2, I am using the DOE prototype EP model, which has the LPD of 10.19587W/m2. That's why after updating LPD, the total lighting decreases in my model.
What I did was after I got the baseline, I did nothing but change the LPD according to the table, so I am very confused now.
The redistribution puts more lighting in occupied zones, which have more ventilation needs and thus higher cooling loads. A total reduction of lighting power of 1.67% is not very significant, while the increase in the zones that already had the highest cooling load (occupied zones) can easily offset it. It also could be anything else, as redistributing internal loads should not result in the exact same cooling and heating load, rather small changes in any direction are expected.