**1. Introduction**

Fuel constraints are a relevant issue in both industrialized and developing countries and are related to energy prices and accessibility of energy services [1]. Public buildings such as universities, schools. and hospitals are challenged to manage the exponential growth of their energy demand and transform their buildings into energy efficient ones. The design of buildings should logically be adapted to the lowest energy consumption levels, but in most cases, it is necessary to focus on existing buildings [2]. Therefore, the reduction of both energy consumption and CO2 emissions from buildings is one of society's main targets today [3]. In Spain, there is a climatic classification according to the technical code of the building that contemplates these issues, which has been mandatory since 2006 [4]. Acting as models for communities, universities are supposed to provide innovative solutions through research in order to support the sustainability and reduce the carbon footprint [5]. One of the key operating aspect for universities is related to enhance students and teachers comfort levels, which may have a significant effect on their performance [6]. Visual, acoustic, and thermal comforts should not be considered as luxuries but rather as basic standard for schools [7]. However, maintaining indoor quality will eventually lead to a significant growth of electricity consumption; therefore, transforming university locals into energy efficient ones is a necessity. To ensure that these locals have optimal energy performance, researchers and professionals have developed managemen<sup>t</sup> systems such as energy benchmarking and energy audit [8].

The energy benchmarking technique allows us to compare the energy consumption of buildings by dividing the key performance metric by gross floor area [9]; this index is usually expressed in (kBtu/ft<sup>2</sup>/yr or kWh/m<sup>2</sup>/yr), and it is labeled as Energy Use Intensity (EUI) or Energy Intensity (EI). This gives the opportunity to the portfolio manager to track the key performance metric overtime [8]. EUI is expressed as energy per square meter per year. It is calculated by dividing the total energy consumed by the building in one year by the total gross floor area of the building. The main benefit of using EUI is that the performance of a building can be compared with similar buildings across the country. EUI can vary significantly depending on building type; therefore, it is necessary to calculate it in buildings used in which there is no data so far. Energy audit is a tool that allows building owners and managers to determine which energy efficiency measures meet their sustainability goals and their investment return criteria [10]. The energy efficiency directive (201/27/EU) requires the auditing of the energy performance of old schools to assess them and propose future retrofitting if necessary [11]. In Italy, over 28% of schools are energy inefficient [12]. A previous experiment executed an energy managemen<sup>t</sup> program in a high school located in Dubai, UAE [13], and its results show that energy performance can be basically improved by 35%. Many evaluation programs for green schools have been designed to assess managers towards sustainable solutions, like the program whole-school approaches, this initiative integrated different elements of school life such as governance, pedagogical methods, curriculum, resource management, school operations and grounds [14]. In the case of the University in Spain, particular studies have been carried out for the Universidad Politecnica de Valencia (UPV) in order to predict electricity consumption patterns in buildings [15] or the use of algorithms using demand and generation forecasts and costs of the available resources, so the benefit obtained in a whole year is five times higher, with a percentage of participation in demand response programs (DRPs), which is accepted as 60.27% or higher [16]. At this same university, with the use of energy efficiency measures (EEMs), in three different types of buildings (a research building (Building 8G), a teaching and staff building (School of Telecom Engineering building 4P), and a greenhouse building 8I-8J), the savings representing about 10% of total annual energy consumption [17].

HVAC and lighting systems have drastically changed in the last decade. Today, they incorporate sophisticated sensors and computer networking programs to monitor and adjust building systems and energy usage. These new technologies are called building automation systems, and they control, monitor, and collect data on the buildings performance technology [18,19]. University campuses serve different functions by providing spaces such as teaching rooms, academic offices, laboratories, restaurants, and sport facilities. This research outlines the classification of categories by their ECs and EUIs. The building category that influences substantially the overall EC of the University by 47% even though it covers only 27% of its total GFA. This category is the science and research category, and it is also the most energy intense by an average EUI 119.5 kWh. Similar results were reported by a study that was ran to support the ASHRAE standard 100. It has determined the EUI median for 18 major categories by climate zone in the USA, according to CBECS 2003. The national median of the laboratory has the highest energy intensity on a university campus (98 kWh/m2). Our case study provides an opportunity to treat a diverse dataset of buildings. A study carried out in Australia reported that laboratory energy intensity was the highest among other categories, and it was three times higher than non-laboratory buildings [20]. In addition, another study divided laboratories into different classes of science, applied science, and intervention, and the results show that the HVAC and electric appliances load, as well as the long operating hours, are the main reasons behind the high energy consumption of this category [21].

The quantity of energy used in universities can change from a country to another, as a recent study in Taiwan has demonstrated that gross domestic product (GDP) of the country has a positive correlation with the energy consumption [22]. Furthermore, a study carried out by Catherine and Byrne et al., (2014) summarized the major factors that significantly impact university buildings' energy use are as follows: occupancy rate, HVAC load and artificial lighting, number of computers and electric equipment, and weather conditions [23]. The influence of these various parameters on the energy use

and their correlated relationship to each other define the stochastic nature of the EC. This paper focused on two parameters—weather and the size of the active network inside the campus. The choice of those variables was made based on an energy survey that was conducted inside the campus and the analysis of energy consumption patterns over the last eight years. Unlike many previous studies that focused on modeling the occupant behavior and its influence on the EC, this study tested the impact of the network on a yearly basis. We gathered the number of occupants active inside the university, including the number of students, professors, and administrative sta ff, and since this parameter varies during the academic year, we had to line up the two other parameters in order to have them all on the same sequence. Then, we tested the correlation with both variables (the number of occupants and the sum of CDD and HDD) with macro-scale energy data. On the other hand, energy benchmarking seeks to give a reference value by defining reliable indicators, we split the buildings portfolio by the following six categories: research, administration o ffice, teaching and seminary room, library, sport facility, and restaurants. This will allow us to benchmark within the same category and identify the benchmark value of each category. These values could be used in the future to define a national baseline for universities or in the Mediterranean region. As the majority of studies have proven that outdoor conditions are the main variable that influences energy use, in this study, we will test the correlation of the total EC with the size of network from one side and with the sum total of the HDD and CDD during each academic year. Moreover, identifying the variable that has the strongest correlation is primarily in order to take suitable actions to achieve better energy management.

#### **2. Materials and Methods**

The University of Almeria is a Spanish Public University located in the south coast of Spain, with the coordinates of latitude 36◦4945"N and longitude −2◦2416"E, see Appendix A. The university campus spreads on a surface of 17 hectares and has 33 buildings (see Figure 1). In the 2018–2019 academic year, the university o ffered 38 di fferent degree programs, with 883 lecturers and 13,547 students.

**Figure 1.** The University of Almeria ground plan.

Almeria is a coastal city located on the southern region of Spain, the climate is particularly arid and semi-continental, with relatively dry warm winters with an average temperature at 16◦C (60 ◦F) and hot summer with an average temperature of 28◦C (80 ◦F). The most daily sunshine hours are scored in July and the wettest month is January with an average of 30mm of rain and an annual average percentage of humidity of 61.0% [24,25].

The dataset used in this analysis consists of (1) daily outdoor temperature scored every half hour during the last eight years, (2) Total energy consumption on a monthly basis and gross floor area data from 2011–2018 (Table 1), (3) campus buildings' energy consumption data during the last three years 2016–2018 (Table 2), (4) the average EUI within each category (Table 3), (5) Building energy performance classification of all the buildings by category (Table 4), the number of students, professors, and administration staff per academic year (Table 5). Figure 2 outlines the methodology flow chart, starting from defining objectives to collecting data to developing results.


**Table 1.** Monthly evolution in energy consumption of the university campus per year.

**Table 2.** Space category, energy consumption, and gross floor area.



**Table 3.** Average energy use index per year and average energy consumption of each space category.

**Table 4.** Building energy performance classification of all the buildings by category.


**Table 5.** Inputs and outputs used for the correlation analysis and regression model.


**Figure 2.** Methodology top-down chart.
