Predicting Energy Requirement for Cooling the Building Using Artificial Neural Network

Abstract: This paper explores total cooling load during summers and total carbon emissions of a six storey building by using artificial neural network (ANN). Parameters used for the calculation were conduction losses, ventilation losses, solar heat gain and internal gain. The standard back-propagation learning algorithm has been used in the network. The energy performance in buildings is influenced by many factors, such as ambient weather conditions, building structure and characteristics, the operation of sub-level components like lighting and HVAC systems, occupancy and their behavior. This complex situation makes it very difficult to accurately implement the prediction of building energy consumption. The calculated cooling load was 0.87 million kW per year. ANN application showed that data was best fit for the regression coefficient of 0.9955 with best validation performance of 0.41231 in case of conduction losses. To meet out this energy demand various fuel options are presented along with their cost and carbon emission.


INTRODUCTION
degree north the elevation of the district ranges from 300 to 3,000 meter above sea level. During six month's summers (April to September) people use electricity (provided on subsidized rates) and other conventional fuels (diesel/petrol) to lower down the temperature. These result in burden on already depleting conventional fuels and same time causing emission of CO 2 and global warming. The other option to meet out energy requirement is solar passive technologies. This requires measured data of solar radiation which is not available in the state. This can be estimated by using various models on the basis of sunshine hour or temperature. The mean hourly values of such data for various places in India are available in the handbook by Mani [1]. The major problem is to calculate the energy demand of a building during summers. ANNs are the most widely used artificial intelligence models in the application of building energy prediction. In the past  [2] did a brief review of the ANNs in energy applications in buildings, including solar water heating systems, solar radiation, wind speed, air flow distribution inside a room, prediction of energy consumption, indoor air temperature, and HVAC system analysis. In [3], Yokoyama et al. used a back propagation neural network to predict cooling demand in a building. In their work, a global optimization method called modal trimming method was proposed for identifying model parameters. Kreider et al. [4] reported results of a recurrent neural network on hourly energy consumption data to predict building heating and cooling energy needs in the future, knowing only the weather and time stamp. Based on the same recurrent neural network, Ben-Nakhi and Mahmoud [5] predicted the cooling load of three office buildings. Considering the influence of weather on the energy consumption in different regions, Yan and Yao [6] used a back propagation neural network to predict building's heating and cooling load in different climate zones represented by heating degree day and cooling degree day. The neural network was trained with these two energy measurements as parts of input variables. In the application of building electricity usage prediction, an early study [7] has successfully used neural networks for predicting hourly electricity consumption as well as chilled and hot water for an engineering center building. Nizami and Al-Garni [8] tried a simple feed-forward neural network to relate the electric energy consumption to the number of occupancy and weather data. Wong et al. [9] used a neural network to predict energy consumption for office buildings with day-lighting controls in subtropical climates. The outputs of the model include daily electricity usage for cooling, heating, electric lighting and total building. Hou et al. [10] predicted airconditioning load in a building, which is a key to the optimal control of the HVAC system. Lee et al. [11] used a general regression neural network to detect and diagnose faults in a building's air-handling unit. Aydinalp et al. [12] showed that the neural network can be used to estimate appliance, lighting and space cooling energy consumption and it is also a good model to estimate the effects of the socio-economic factors on this consumption in the Canadian residential sector. Gouda et al. [13] used a multi-layered feedforward neural network to predict internal temperature with easily measurable inputs which include outdoor temperature, solar irradiance, heating valve position and the building indoor temperature. Kreider et al. [4] reported results of recurrent neural networks on hourly energy consumption data. Karatasou et al. [14] studied how statistical procedures can improve neural network models in the prediction of hourly energy loads. Azadeh et al. [15] showed that the neural network was very applicable to the annual electricity consumption prediction in manufacturing industries where energy consumption has high fluctuation. It is superior to the conventional non-linear regression model through Analysis of Variance (ANOVA). We have taken a university building having six storey which works for seven hours during a day time. The dimensions are length 45 m, 15 m wide and 18 m in height.

METHOD AND MATERIAL
Under the steady state approach (which does not account the effect of heat capacity of building materials), the heat balance for room air can be written as [16]: where Q total is total energy requirement if it is -ve then heating is required and if it is +ve then cooling is required. Q c is conduction losses in a building Q s is solar gain in a building Q i is internal gain in a building Q v is ventilation losses in a building

Conduction
The rate of heat conduction (Q c ) through any element such as roof, wall or floor under steady state can be written as If the surface is also exposed to solar radiation then where T i is the indoor temperature; T so is the solar air temperature, calculated using the expression: = emissivity of the surface R = difference between the long wavelength radiation incident on the surface from the sky and the surroundings, and the radiation emitted by a black body at ambient temperature

Solar Heat Gain
The solar gain through transparent elements can be written as:

Ventilation
The heat flow rate due to ventilation of air between the interior of a building and the outside depends on the rate of air exchange. It is given by:

Internal Gain
The heat generated by occupants is a heat gain for the building; its magnitude depends on the level of activity of a person. Table 1 shows the heat output rate of human bodies for various activities [17].
The total rate of energy emission by electric lamps is also taken as internal heat gain. Table 2 shows the heat gain due to appliances (televisions, refrigerators, etc.) should also be added to the Qi [17].  Mean hourly values of data shown in Table 3 for various places in India are available in the handbook by Mani [1].

RESULTS
The total conduction losses in a building are calculated as Table 3.
Q c = 22.6 kW = 97632 kW per annum whose ANN graphs are shown in Figures 1 and 2.
The total solar gain in a building is calculated as Table 4.        The total ventilation losses in a building are calculated as Table 5.
Q v = 136.1 kW = 587952 kW per annum whose ANN graphs are shown in Figures 5 and 6.
The total internal gain in a building is calculated as Table 6.
Q i = 108.5 kW = 104160kW per annum whose ANN graphs are shown in Figures 7 and 8.
The total energy requirement during winter is calculated as Table 7.

DISCUSSION
The neural network model was used with 10 hidden neurons. Figures 1, 3, 5, 7 and 9 didn't indicate any major problem with the training. The validation and test curves were very similar. The evaluation and validation of an artificial neural network prediction model were based upon one or more selected error metrics.
Generally, neural network models which perform a function approximation task will use a continuous error metric such as mean absolute error (MAE), mean squared error (MSE) or root mean squared error (RMSE). The errors will be summed over the validation set of inputs and outputs, and then normalized by the size of the validation set [19]. Here we had used mean      squared error (MSE) for the best validation performance. The next step in validating the network was to create a regression plot, which showed the relationship between the outputs of the network and the targets. If the training were perfect, the network outputs and the targets would be exactly equal, but the relationship was rarely perfect in practice. The result was shown in the Figures 2, 4, 6, 8 and 10. The three axes represented the training, validation and testing data. The dashed line in each axis represented the perfect result -outputs = targets. The solid line represented the best fit linear regression line between outputs and targets. The R value was an indication of the relationship between the outputs and targets. If R = 1, this indicated that there was an exact linear relationship between outputs and targets. If R was close to zero, then there was no linear relationship between outputs and targets.

CONCLUSIONS
The study reveals that the total cooling load of a six storey building is 0.87 million kW thus, cooling is required to meet out this energy demand. If we use electricity it will produce 3.5 ton carbon per annum and the cost of electricity used will be $ 47,709.16 as depicted in Table 8.
If we use diesel to meet out this energy requirement then 236.2 ton of carbon will be emitted and it will cost $57,250.9. The above results necessitate the use of solar passive technologies to meet out this energy requirement during summers.