The prediction of heating energy consumption in a model house by using artificial neural networks in Denizli–Turkey

https://doi.org/10.1016/j.advengsoft.2009.09.012Get rights and content

Abstract

Turkey does not have petrol and natural gas reserves on a large scale. National energy resources are lignite and hydropower. Together with increasing environmental problems and diminishing fossil resources, studies focusing on energy reduction as well as usage of renewable energy resources have accelerated. However, taking the technological and economical impossibilities into account, the most logical solution is energy saving by providing energy efficiency in households. In this study, an artificial neural network (ANN) model is developed in order to predict hourly heating energy consumption of a model house designed in Denizli which is located in Central Aegean Region of Turkey. Hourly heating energy consumption of the model house is calculated by degree-hour method. ANN model is trained with heating energy consumption values of years 2004–2007 and tested with heating energy consumption values of year 2008. The training and test figures were depicted for February month of these years. Best estimate is found with 29 neurons and a good coherence is observed between calculated and predicted values. According to the results obtained, root-mean-squared error (RMSE), absolute fraction (R2) and mean absolute percentage error (MAPE) values are 1.2575, 0.9907, and 0.2091 for training phase and 1.2125, 0.9880, and 0.2081 for testing phase respectively.

Introduction

After the 1970 energy crisis, energy has become the most important issue for countries which meet great part of their energy needs through importation. Considering the technological improvements, it is evident that energy will be of primary importance in these countries [1].

Located at the junction point of Asia, Europe, and Africa continents, Turkey is a bridge between Asia and Europe continents. Its area is 779,452 km2, population is 704,87,917, gross national income per capita is 9233 USD, and economical growth in the last 5 years is approximately 7%. Seventy percent of the population lives in cities, because of economical and social progress the need for energy is rapidly increasing. Turkey meets almost all of its need for petrol and natural gas by importation. Turkey’s national energy resources are lignite and hydropower. Approximately 50% of electrical energy need is met by thermal power plants operating on natural gas, while the rest is met by coal-fired thermal power plants and hydroelectric power plants [2].

Turkey is a country of energy markets which grow rapidly. Annual energy need had increased 6.6% between years 1995 and 2004 and it is estimated to increase by 8.5% between 2005 and 2015. While electricity consumption was 150 billion kW h in 2004, it is estimated that increasing by a factor of 4; this consumption will reach 499 billion kW h by 2020. In that case, it will be necessary to increase the established capacity, which is 38,500 MW as of 2005, to 96,000 MW until 2020. Since Turkey has a dynamic economical growth and rapid increase of population, the need for energy is increasing each year [3]. However, serious economical crises occurring in certain periods since 1994 have prevented the investments that could meet the growing energy need.

In Turkey, 80% of the energy consumed in households is utilized for heating purposes. This amount is luxury for a country like Turkey which imports almost all of the energy it is consuming. Together with the increase of environmental problems and decrease of fossil resources, studies focusing on energy reduction as well as usage of renewable energy resources have accelerated. Because of the shortage in necessary infrastructure and expensive setup costs, in Turkey, there are only a few projects applied utilizing renewable resources such as the sun, wind, water, etc. Taking into account technological and pecuniary impossibilities, the most reasonable solution is energy saving in energy-efficient households [1].

Seasonal energy consumption calculations have an important role in calculating heating and cooling loads of any household [4]. Precipitation and temperature records provide important information about a region’s weather conditions. In addition to the effects of changes in temperature on agriculture, architecture, energy production and consumption, and melting of snow; icing and freezing affect transportation systems, flowering and harvest days, heating–cooling electric power at households, and air-conditioning systems. All of these are dependent on daily and hourly outside temperatures and system design values [5].

Exact information on hourly outside temperature is critical for environmental, agricultural, and industrial applications [6]; information on change of outside temperature is very useful in estimating solar radiation [7], [8], forecasting hourly energy consumption and cooling loads of households [9], [10], and predicting room temperature [11]. Analyses of annual energy consumption and cost have an important role in heating load calculation of design and heating–cooling system selection [12]; because of the fact that air-conditioning and refrigeration engineers need system design data.

In constituting thermal accommodation quality of households, it is necessary to appropriately analyze heating and cooling conditions together with meteorological parameters. In addition to heating degree-day (HDD) and cooling degree-day (CDD) values, heating degree-hour (HDH) and cooling degree-hour (CDH) values are very frequently used in energy planning and heating–cooling loads calculation. Although there are various methods for calculating yearly, monthly, or hourly energy consumption, the most comprehensible and simple method is degree-day or degree-hour method which is a continuous state approach [13]; utilizing the hourly outside temperature values of a whole year, very precise results can be obtained for the energy analysis of a household [14].

Degree-day or degree-hour method is based on the principle that inside conditions are stable during heating and cooling period and it is assumed that heating or cooling equipment efficiency does not get affected by outside temperature changes. If all thermostats in the household are set to the same temperature value at the beginning of the heating or cooling period and kept absolutely unchanged, these circumstances can be obtained. By this way, instead of full load or design efficiency, seasonal average efficiency is used in heaters or coolers [12].

Artificial intelligence can be defined as the ability of computers to think, understand, and learn; however, computers can only think through the algorithms given to them. Artificial neural network (ANN) is an important branch of artificial intelligence and it is widely used in many engineering problems [1]. One of the most important characteristics of ANN models is their capability of anticipating emerging patterns in a complex natural system and adapting themselves according to repeated changes [15].

In this study, an ANN model was developed in order to estimate hourly energy consumption of a model house designed in Denizli which is located in Central Aegean Region of Turkey. Hourly energy consumption of the model house for the analysis period was calculated by degree-hour method utilizing hourly outside temperature values of 4 years (2005–2008) which belonged to the heating period of this city and obtained from Turkish State Meteorological Service [16]. Afterwards, the constructed ANN model was trained with heating energy consumption values of years 2004–2007 and tested with heating energy consumption values of 2008. February months of these years is taken as basis for the generation of training and test graphics. Learning algorithm and transfer function of ANN model were selected as Levenberg–Marquardt (LM) and tangent sigmoid (TANSIG) respectively. Output values of the network are hourly energy consumption values that are constructed by input values. The location of the city, which is selected in the study, is given on Turkey map in Fig. 1.

Section snippets

Model house

In this study, a model house with single-storey was designed. External walls of the model house are insulated and composed of external rendering, brick, reinforced concrete, and internal rendering. Ceiling of the house is built using reinforced concrete and top of the ceiling is covered by roof. In the insulated external walls also called as sandwich wall, double horizontal bricks stuffed with insulation material are used. Polypan is selected as insulation material (k = 0.035 W/m °C); total heat

Calculation of hourly heating energy consumption with heating degree-hour method

Degree-hour method is a very useful method to determine heating and cooling energy need as well as the effect of regional climate change. Method assumes that energy need of the house is proportional to the difference between hourly outside temperature and base temperature. In hourly heating, if base temperature is higher than outside temperature then need for heating arises; contrary to this, in hourly cooling, if base temperature is lower than outside temperature then need for cooling arises

Artificial neural network (ANN)

Artificial neural network, which is a branch of artificial intelligence developed in 1950s in order to imitate human brain’s biological structure [19], is a very frequently used method in recent years for modeling and prediction purposes [20]. Like non-linear regression, ANN is a model that can learn the relationship between input parameters and controllable or uncontrollable variables, and thus, does not need detailed information about the system. One of the most important advantages of ANN is

Results and discussion

The ANN model, which was developed to teach hourly heating energy consumption of a single-storey model house designed in Denizli city which is located in Central Aegean Region of Turkey, was trained. For training and testing, hourly energy consumption values, which were calculated by heating degree-hour method using hourly temperature data of years 2005–2008, were used. The patterns in the database of this data are clustered under two sets. The first set is the data composed of energy

Conclusions

The purpose of this study was to predict hourly energy consumption of a model house in Denizli utilizing ANN model. In the study a three-layer ANN including one hidden layer was used in order to predict hourly energy consumption of the model house. A total of 35,070 h temperature data were used to calculate hourly energy consumption values; 26,310 of this data were used for training and 8760 data were used for testing. In the ANN model utilizing LM learning algorithm, the best result was

Cited by (67)

  • Energy saving and CO<inf>2</inf> reduction potential of external building walls containing two layers of phase change material

    2022, Energy
    Citation Excerpt :

    In Turkey, about 80% of the energy consumed by buildings is used to create thermal comfort. This indicates a considerable savings potential by reducing energy use for space heating and cooling for Turkey, which imports a significant amount of the energy needed [2]. Recently, the energy consumption of buildings has been increased around the world due to people spending more time indoors due to the Covid-19 pandemic [3].

View all citing articles on Scopus
View full text