Buildings’ internal heat gains prediction using artificial intelligence methods

1-s2.0-S0378778821010781-main.pdf

2.75MB

1. Introduction

20% of total national energy and 50% of building energy are consumed by HVAC systems

Also, in public buildings with a large capacity, 40–60% of the total energy consumption is related to air conditioning

Hence, in terms of ecological balance and sustainable development, the efficient performance of HVAC systems is highly important

 

As for energy-efficient building and control optimization, load prediction is an essential factor and has drawn researchers’ attention

 

Loads of buildings are mostly categorized into two parts.

The first one is external loads that are affected by weather conditions and envelope structures of buildings,

and the second one is internal loads induced by indoor thermal sources such as lighting, residents and equipment

 

There are two important reasons that account for the significance of internal heat gains.

First, the progress in envelope structures of buildings and fenestration has led to improving building envelopes, which reduces air leakage and consequently decreases the heat transfer loss

Second, with the development of technology, people’s dependence on equipment has increased, which causes the diversifying and development of the equipment

 

As a result, it can be found that internal heat gain is an important factor in load prediction and building control.

 

For this reason, Firla˛g et al [12] obtained the difference of internal heat gain between a real measured value and an ordinance. Such a little difference had a considerable effect on the building’s energy performance.

 

As for the present situation of predicting the internal heat gains, general rules, schedule and design values are considered as general criteria. 

- 방법 

1. Researchers have widely adopted and employed ASHRAE standard 90.1 throughout the world

2.Samaan et al. [19] deployed templates of indoor sources utilizing Design Builder for simulating internal heat gains.

3. An et al. [20] used a new stochastic technique to predict the cooling loads considering occupancy schedules, occupants’ densities, equipment and lighting.

 

4. Wang et al. [21] employed two models including Long Short-Term Memory Networks and a specific form of deep neural network for predicting miscellaneous electric loads (MELs), occupant numbers, internal heat gains and lighting loads

*miscellaneous electric loads (MELs)

https://en.wikipedia.org/wiki/Miscellaneous_electric_load

Miscellaneous electric load - Wikipedia

5. Zhou et al. [11] extracted the data of lighting consumption from 15 office buildings and developed predictive mod-els for lighting power at various time intervals

 

6. Amasyali et al. [22] developed a model based on support vector machines to predict the energy consumption from lighting in office buildings

 

- 기대효과

It can be seen from the literature that predicting internal heat gains affected by parameters such as occupants, lighting and equipment is of great importance in predictive control, efficient operation, and energy conservation of HVAC systems.