The Building Energy Balance

Picture the air in a room as an account that must always balance. On the gains side: solar gains (short-wave radiation coming through glazing and, more slowly, warming opaque surfaces), and internal gains - the waste heat from people (roughly 70-100 W each at rest, more when active), lights, and equipment (computers, screens, kitchen loads, servers). On the losses side: conduction through the envelope (walls, roof, glass, floor - governed by their U-values and the indoor-outdoor temperature difference), and air exchange, split into deliberate ventilation for fresh air and unwanted infiltration through cracks and gaps.

If gains exceed losses, the room heats up; if losses exceed gains, it cools down. To hold a setpoint - say 24 C - the HVAC must remove or add exactly the shortfall. In a warm climate that mostly means removing heat (cooling); in a cold one, adding it (heating). The engine writes this as a heat-balance equation on the zone air node and solves it at every timestep. That is literally all EnergyPlus is doing - just very fast, and with honest physics for each flow.

A useful way to hold the whole idea is a single accounting line: the change in the zone's stored heat over an hour equals everything that came in minus everything that left, minus whatever the HVAC removed. When the design and controls keep that line at zero at the setpoint, the occupants feel steady comfort; when it cannot be held - because the plant is off, or too small - the temperature swings, and the model reports it as an unmet or free-floating hour rather than as energy. Every output you will ever read traces back to this one bookkeeping statement, repeated hour after hour.

Heat comes in two flavours, and a model tracks both. Sensible heat changes air temperature - it is what a thermometer reads. Latent heat changes air moisture - the energy tied up in water vapour, which you feel as humidity rather than temperature. People give off both (we breathe out water and sweat); ventilation air in a humid climate drags in a large latent load; cooking and plants add moisture too.

This matters enormously in India. In warm-humid coastal cities - Chennai, Mumbai, Kochi - a big share of the cooling energy is spent not on lowering temperature but on wringing water out of the air (dehumidification). A model that ignored latent load would badly under-predict the cooling bill. When you read HVAC output, you will often see the cooling load split into sensible and latent parts; in humid zones the latent slice is stubbornly large, which is why simply opening windows can make comfort worse, not better. In hot-dry climates like Jaipur, latent loads are small and evaporative strategies suddenly make sense - the same balance, a different mix.

Solve the balance for one hour and you know the instantaneous load. Solve it for all 8760 hours of a typical year - each with its own outdoor temperature, sun position and cloud cover from the EPW weather file, and its own occupancy and equipment use from the schedules - and you can add up the energy the HVAC delivered. That annual total, divided by floor area, is the building's Energy Use Intensity (EUI), in kWh/m2 per year: the single most quoted number in energy modelling and the basis of most code and rating comparisons.

The shape of that annual total - how it splits across cooling, heating, fans, lighting, equipment and hot water - is the building's energy signature. Two identical buildings in two climates produce completely different signatures: a Chennai office is dominated by cooling and fans, a Shimla one by heating. Reading the split tells you where the energy actually goes, and therefore where design effort pays off. A model that says 'EUI 140' is far less useful than one that says '52% cooling, 12% fans' - the breakdown is the actionable part.

This is also why energy modellers care so much about the weather file behind the run. Swap a typical-year EPW for a hotter recent year, or a coastal file for an inland one, and the same building returns a different EUI - not because the design changed but because the climate driving the balance did. Reading an annual result therefore always means reading it against a stated location and weather file; an EUI with no climate attached is a number without a context.

You might ask why we cannot just size these flows by hand. The reason is that they are all coupled and time-varying. Solar gain depends on sun angle, which changes hourly and seasonally and interacts with your shading. Conduction depends on the temperature difference, which depends on how much the sun and internal gains have already warmed the space. Thermal mass delays and smears the peaks (a heavy building rides through the afternoon that a light one cannot). Ventilation helps at night and hurts at midday. These effects do not add up linearly - they chase each other through the day.

A simulation resolves this by stepping through time and re-solving the balance each hour, carrying the stored heat in the mass forward. That is what turns a pile of plausible assumptions into a defensible annual number. And it is why the same design decision - deeper shading, more insulation, a different glazing - can help in one climate and hurt in another: it changes one term in a balance whose other terms respond. Understanding the balance is what lets you predict the direction, and the model is what quantifies the size.

Not every building is air-conditioned, and the balance behaves differently depending on how the space is run. In a free-running building - no mechanical heating or cooling, just openable windows and fans - there is no HVAC term to close the balance, so the indoor temperature simply floats wherever the gains and losses leave it. The model's job then is not to report energy but to report comfort: how many hours the floating temperature stays inside the acceptable band. Much of India's residential stock, and a great deal of good passive design, runs this way for large parts of the year.

In a conditioned building, the HVAC term is always active, pinning the air to the setpoint and paying an energy price to do it. And in a mixed-mode building - increasingly common and sensible in India - the space free-runs when the outdoor conditions allow (windows open, no AC) and switches to conditioning only when they do not. Modelling mixed-mode is subtle because the balance keeps changing character through the year, but it is exactly where the biggest energy savings hide: every hour you can hold comfort without the compressor is an hour of near-zero cooling energy. Knowing which of these three regimes a space is in tells you what question to ask the model - energy, comfort hours, or the split between the two.

Recognising the regime early also shapes the honest deliverable. For a free-running house you promise the client comfort hours, not a kWh figure; for a conditioned office you promise an EUI and a comfort guarantee together; for a mixed-mode school you promise both, plus the fraction of the year the building can coast on openable windows. Naming the regime up front stops the commonest reporting mistake - quoting an energy number for a building that was never meant to be mechanically conditioned in the first place.