Energy Management Systems in Higher Education Institutions’ Buildings

Abstract

This study reviews the methods used to implement energy management systems (EnMS) in higher education institutions (HEIs) and their impact on improving energy performance considering their relationship with the requirements for an EnMS according to ISO 50001. From 2310 articles, 136 articles and 5 technical reports related to EnMS and energy efficiency were selected and analyzed. A synthesis of the major actions taken by HEIs to enhance their energy performance is presented, including energy management strategies, methods for measuring and estimating consumption, occupant behavior models that influence energy use, barriers to energy efficiency in HEIs buildings, and future challenges. It was found that studies on building energy management systems often do not incorporate an analysis of CO₂ emissions reduction. Funding for this research is driven by directives and policies related to energy performance. These results should assist HEIs seeking to implement an EnMS to improve their energy performance and reduce CO₂ emissions, thereby contributing to energy security, climate change mitigation, and fostering a new culture of energy use and consumption. It was also found that, although most studies do not explicitly mention the ISO 50001 standard, all of them comply with at least one of its requirements. Additionally, 27% of energy management strategies focus on operational aspects, while 26% involve energy audits, primarily through measurement, estimation, forecasting, energy reviews, and the establishment of an energy baseline (EnBL).

1. Introduction

By 2023, buildings accounted for 32% of global energy demand and 34% of energy-related CO₂ emissions [1], with most of this energy being used by building subsystems consisting of cooling or heating systems [2], lighting, automation systems, office, and laboratory equipment such as computers, televisions, printers, video beams, etc. Moreover, energy consumption in this sector is expected to increase by more than 40% in the next 20 years [3].

The integration of energy efficiency in buildings contributes to the reduction in energy consumption and is a key area that can achieve future sustainable goals as it offers opportunities to build an environmentally sustainable and high-impact educational culture focused on successful transformation towards a climate-neutral society [4]. Therefore, retrofitting existing buildings could contribute more to reducing energy consumption and improving sustainability compared to constructing new green buildings, as the constructions only add to or replace a small portion of existing buildings, which are minor compared to the many currently inefficient buildings [5].

Currently, efforts are focused on meeting energy efficiency requirements in buildings, ensuring the operation of subsystems at the lowest possible energy cost, thus contributing to a sustainable environment [3]. This does not necessarily imply a change to new, more efficient technologies; instead, it is proposed to implement or improve an energy management system (EnMS). As an example, in cooling or heating systems to achieve greater energy efficiency, it is more cost-effective and sustainable to manage these systems than to replace them with more efficient modern technologies since, if these new technologies operate inadequately, it may result in excessive energy consumption [6].

Consequently, in recent decades, the evaluation of the energy performance of any organization has become an extremely important topic, given the urgent need to strengthen the path towards energy sustainability. An EnMS makes it possible to determine energy performance by assessing an organization’s energy use and consumption. Considering this, the ISO 50001 standard [7] provides guidelines for planning, implementing, monitoring, and controlling the energy performance of an organization through the adoption of an EnMS [8]. Implementing an EnMS under ISO 50001 requirements is a contribution to the objectives and strategies aiming to improve energy efficiency of buildings [9], as well as the reduction in energy cost, with savings potentials between 20% and 30% [10]. An EnMS can provide a solution to increasing energy consumption by promoting energy security, climate change mitigation and transition to new energy sources, as well as reducing the environmental impact [4].

Strategies to implement an EnMS and improve energy efficiency are not a new subject; in fact, case studies have been presented in which EnMS were applied to buildings since 1976, and reports indicate that, until 2016, 10.5% of these cases have occurred in universities worldwide, mainly in North America, Europe and Asia [11]. Since university campuses represent specific groups of diverse buildings with significant energy consumption, they are an excellent test bed for characterizing and understanding the energy consumption of a group of mixed buildings [12].

In this context, the Educa-RUE program was developed in Europe, which seeks to improve the energy performance of buildings, with special emphasis on educational buildings, promoting guidance and orientation that will lead to initiatives that promote energy savings. The program led to the development of actions to promote the sustainable use of energy, the definition of an environmentally friendly energy management model, and a commitment by local authorities to comply with European legislation, raising the awareness of communities and public institutions, and increasing the adoption of environmentally friendly behavior [13]. In China, for example, a sustainable campus project was developed, including an energy management system as an initial step, to implement and maintain a sustainable campus [14].

Electricity costs represent the largest expense in the budget of higher education institutions (HEIs) [15]. This is because HEIs are significant energy consumers, making them highly vulnerable to issues stemming from the scarcity of non-renewable resources and the effects of climate change. Additionally, as educational institutions, they play a crucial role in shaping future generations; however, rising energy costs have become a major obstacle to their growth and development. For these reasons, HEIs worldwide are increasingly committed to energy conservation efforts [16]. One way to address this challenge is through the implementation of an EnMS. This system can be established by utilizing schedules and event management systems to take corrective actions aimed at eliminating or mitigating nonconformities. By improving energy performance and employing block-based systems, HEIs can raise awareness among students and the broader educational community about energy use and consumption [15]. Such initiatives not only contribute to more efficient energy management but also foster a culture of sustainability, ultimately leading to a more sustainable society.

Ensuring effective compliance with ISO 50001 requires each organization to develop an EnMS that is aligned with its own strategy, mission, processes, or systems, and, in this area, the choice of representative energy performance indicators is critical. For a factory, an energy manager can easily correlate energy performance with production output, and effective and repeatable measures can be obtained. On the other hand, these conditions are not suitable for organizations that offer intangible services, such as universities, public administrations, or government agencies, which represents a challenge to obtaining real-time, reliable, and effective data on energy performance [8].

To implement an EnMS in higher education buildings, five primary factors must be considered: economic feasibility, environmental impact, institutional characteristics, impact because of occupancy, and technical practicality [17]. For example, in Colombia, these factors must be considered according to the energy transition policies of this country [18], which have established mandatory laws to have energy-efficient buildings. Likewise, the European Commission estimates that implementing energy efficiency policies can lead to 25% savings in energy demand. In addition, energy efficiency also encompasses the variables of affordability, energy security and environmental sustainability, which require the implementation of policies by governments. These policies are mainly focused on sectors with high energy demand, such as energy, industry and transportation, and studies usually consider energy efficiency within the sector policies and measures to reduce energy consumption and related emissions [4].

Implementing energy management systems (EnMS) has proven to be effective in increasing the energy performance of organizations [8,19,20]. This does not imply changing to new or more efficient technologies, but controlling energy procedures through EnMS.

This document aims to provide an overview of cases in which an energy management system (EnMS) has been implemented in higher education institutions, demonstrating how such an implementation leads to an increase in energy performance and contributes to sustainability. It is hoped that this document will encourage higher education institutions to implement an EnMS to reduce the environmental impact of buildings in terms of CO₂ emissions, contributing to energy security, mitigating climate change, and setting an example of behavior in terms of energy use and consumption, which is why more and more institutions are joining this action. One contribution of higher education institutions is the carbon footprint (CF) report, which provides policies and management systems to establish strategies to reduce the emissions of polluting gases from their campuses. Since the ISO 50001 standard does not include the topic of carbon emissions, this article does not include the calculation of carbon footprint reduction, which is covered in other articles [21].

We consider that higher education buildings have their own characteristics that differentiate them from government buildings because educational buildings are focused on providing training services; so, their buildings can provide services for classrooms, laboratories, cafeteria, and offices, as well as integrate two or more of these services, which makes them a tested, since it allows for studying energy performance in different buildings, which offer multiple or a single service, mainly educational [22,23].

According to the commercial building energy consumption survey CBECS [24], education buildings in the U.S represent 13% of total energy consumption in commercial buildings and, according to the European Union’s building stock observatory [25], the education building stock in Italy is 17.4%, which represents approximately 2.88 Mtoe of energy.

This paper reviews the methods applied to implement energy management systems (EnMS) in higher education institutions and the impact EnMS has had on increasing energy performance. This section highlights the EnMS used in buildings related to ISO 50001 and those considering its requirements.

This article is divided into three sections. First, it describes the use of energy in buildings of higher education institutions and the systems associated with its use, and it explains what an energy management system is. Second, it describes the methodology used to search and analyze the articles, as this research conducts a systematic review to understand energy consumption and environmental impacts on the implementation of BEMS on HEIs, using the ISO 50001 standard as a guide for energy performance improvement on buildings of organizations. Third, a discussion of the literature found is carried out regarding the energy management strategies used in university campuses, energy consumption estimation and measurement models, management models are demanded, accompanied by the presentation and explanation of optimization models, response models and energy management systems for air-conditioning systems are demanded, risk detection diagnostics and opportunities when implementing EnMS are established, and, finally, conclusions and future challenges are presented.

2. Energy Management Systems and Higher Education Buildings

2.1. Energy Use in Higher Education Institution Buildings

For implementing an EnMS, it is necessary to understand that there are several types of buildings according to their energy use: intelligent buildings, green buildings, and near-zero-energy buildings.

Smart buildings are those with a monitoring and control system that perform actions through the automation of their subsystems, based on the incorporation of technology and energy systems within the buildings. They focus on automation, resource management, occupant comfort, and energy savings [26]. Green buildings aim to be environmentally friendly throughout the building’s life cycle, from planning, development, operation, and activities to refurbishment and destruction [27]. Zero-energy buildings are intended to be powered by sustainable generators and energy sources to achieve near-zero net energy use, thereby reducing carbon dioxide emissions [28].

This work focused on the application of energy management systems to construct buildings with the purpose of turning them into zero-energy buildings by implementing some distributed self-generation models, such as photovoltaic energy or wind turbines [29].

Figure 1 shows how buildings are classified according to their energy objectives.

Major energy reduction efforts come in various forms, ranging from the use of more energy-efficient lighting and equipment, insulation, sustainable building design, night ventilation, phase change materials (PCMs), intelligent automation and control, photosensitive windows, adaptative comfort, load shifting, development of legislation and qualification procedures, and the use of renewable energies, to mention a few [5].

The building structure consists of energy generation, energy storage, energy demand, and automation and control.

2.1.1. Energy Supply

Power generation systems in buildings of higher education institutions are the main subsystem in EnMS. These systems are mainly composed of energy consumed by the grid and non-conventional renewable energy sources, such as river micro-turbines, small-scale wind turbines, and solar panels. Solar panels are currently the most studied source of renewable energy generation as a self-generation system in Colombia [30]. Usually, when one wants to convert an old building to a near-zero-energy building, one thinks of using renewable energy sources to self-generate [28].

According to Filho et al. [31], in a survey of 50 higher education institutions, more than half produce between 1% and 20% of their energy from renewable energies, with solar and photovoltaic energies being the most used sources, accounting for approximately 70%. Another significant source of power generation in buildings is backup systems, such as fossil fuel or renewable energy generators, which can be diesel engines, natural gas engines or electric power plants that use biomass as a source of energy; this biomass can be biofuels from vegetable oil, wood chips or pellets [32].

2.1.2. Energy Storage

There are two types of energy storage systems: battery and thermal, as shown in Figure 2. The use of renewable energy has driven the development of storage systems because of the unreliability of these systems, since the energy resource has a variable behavior.

Thermal energy storage systems have proven to be a sustainable solution, making heating, ventilating, and air-conditioning (HVAC) systems more flexible in their time of use, improving system demand capabilities. Battery storage systems improve the flexibility of photovoltaic systems, allowing for their use when there is no sunlight or when the energy price is higher, reducing the operating costs of the building’s electrical system [33].

Buildings can also operate as a thermal energy storage system through the materials used in their structure since walls can use materials allowing for heat storage in their structure or materials that allow for rapid heat extraction; also, thermal energy storage can be further divided into regular storage in water (buffer) tanks and materials (cold and heat) in the building shell. Phase change materials (PCMs) fulfill this function because they are characterized by having a high energy density, which allows for storing or releasing heat at a constant temperature, maintaining the comfort temperature of the buildings with a lower energy consumption [34,35].

From a grid operator’s perspective, a building can be used as a thermal energy storage system to run various heating and cooling devices. This includes not only heating and air conditioning but also appliances such as refrigerators, freezers and water heaters. This thermal storage capacity can be used to bring flexibility to grid demand and reduce energy consumption in buildings [36]. Thus, energy storage system’s separate energy generation and consumption provide flexibility in the energy system [37], allowing for energy demand, as well as reducing costs, by providing energy consumption at appropriate times or taking actions from an energy management system, to supply the demand, as well as improving comfort of users in buildings. For example, Gallardo and Berardi [38], performed a building simulation using a radiant ceiling panel with a PCM.

2.1.3. Energy Demand Management

An energy-demand-management-based system can be used to improve energy distribution among users, promoting efficient energy use while improving grid stability and reliability [39]. Users play an important role in energy use and consumption, according to their behavior and awareness of the end use of energy resources, their different sources of generation, and their contribution to energy consumption reduction. Load shifting by users is often referred to as a demand response (DR).

Energy management systems, based on demand response, focus on user behavior and seek to ensure user awareness in energy use and rational consumption through communication strategies allowing them to understand the importance of their behavior on the impact of energy consumption; part of this is also related to the human need to have control over systems; therefore, considering the desire to control the systems by the users should be established in an EnMS, as well as cooperation and interaction between occupants, and these systems should not be limited to define only set points and some states of the cooling system; therefore, an EnMS should encourage users to cooperate with it to improve energy performance and occupant comfort [40].

Modeling of user behavior in buildings can be achieved through the following approaches: (1) the theory of planned behavior, (2) the MODE model of attitude–behavior process, (3) the modified norm activation model, and (4) the knowledge–desire–skill–action model [41].

2.1.4. Automation and Control

Monitoring is essential for understanding the energy consumption behavior in buildings [41], and new metering systems, with appropriate communication and information technologies, are improving data quality and frequency. An appropriate metering system allows for a more effective development of an EnMS, complementing energy data with data on climatic factors. Online metering systems require five principal components: measurement, external data collection, data transfer, analysis, and reporting. These systems provide accurate, real-time information to stakeholders. It is important to properly locate meters to obtain reliable and accurate data on variables that impact energy consumption [42].

A control system must be able to apply variable factors to measurements, allow for the replacement of meters and installation of new instruments, and automatically detect them. These systems are known as building automation systems (BASs), which perform the actions described above, and have recently become popular with growing interest in obtaining data that allow for the management of consumption and energy use. This control system allows us to achieve optimum performance by coordinating the operating conditions of all subsystems.

An advance in measurement and control systems for supervision is the possibility of including predictions of any type of variable that affects energy consumption to improve the optimization of the operation of the subsystems of a building in a space of time. This is known as model predictive control (MPC) [43]. Model predictive control needs improvement in areas such as optimization strategies, disturbance prediction effects, and building characteristics where MPC is to be implemented [44]. Model predictive control systems can consider constraints, disturbance prediction, and multivariable targets [45].

A predictive control strategic in a BEMS for boilers in buildings based on a neural network that turns the boiler at the optimum time was implemented at Universitat Politècnica de Catalunya (UPC), estimating an energy consumption reduction of 20% [46].

An artificial neural network was developed to predict the energy generation patterns of renewable energies, such as photovoltaic systems and wind turbines. A predictive control model was adapted to this neural network to use sensors to manage the energy of a residence as the system is responsible for maintaining good occupant comfort and, in addition, maintains energy consumption under renewable energy generation [29].

2.2. Energy Management Systems and Higher Education Buildings

2.2.1. A Standardized Energy Management System—ISO 50001

International experience over the last 40 years has shown that implementing and operating an EnMS in an organization can reduce energy consumption through a continuous improvement in energy performance. Thus, pioneering countries like Japan, Korea, Australia, Canada, China, the USA, Denmark, Sweden, Holland, etc. have adopted national energy management standards. The impact of EnMS on reducing consumption and therefore CO₂ emissions made it an important point in the fight against climate change.

In this context, the United Nations Industrial Development Organization (UNIDO) and the International Standards Organization (ISO) agreed to develop an international standard specifying the minimum requirements for an EnMS to be effective in improving energy performance. Therefore, in 2011, the ISO published the standard “ISO 50001: 2011, “Energy Management Systems. Requirements with guidance for use”; this standard is currently the benchmark for an EnMS, and some call it a standardized EnMS. The second version of this standard was published in 2018. It is important to note that ISO 50001 is focused on energy performance, a concept that relates to energy efficiency, energy use and energy consumption, which is key to ensure effective and measurable results over time [7]. ISO 50001 establishes that the EnMS is based on the “plan-do-check-act” (PDCA) continuous improvement cycle and incorporates energy management into the organization’s standard practices, as shown in Figure 3.

Energy planning and its elements are considered the core of an energy management system. It consists of understanding energy performance to establish objectives and action plans to improve the energy performance of an organization [47].

2.2.2. Energy Management Systems in Higher Education Buildings

An energy management system allows us to monitor, control, and take actions to improve energy performance in a production or service process, as well as improve thermal comfort and create a safer and healthier environment [48] (Majid Jamil & Sonam Mittal, 2017). It is applied through management strategies, which are understood as a set of techniques used in the EnMS which interact dynamically with the building to improve energy efficiency [3].

Building energy management systems (BEMSs) can be described as a combination of strategies and methods needed to improve building performance, efficiency, and energy use [49]. These measures are mostly technological rather than behavioral. These measures are measures that are and have been covered extensively in the literature, as seen in this paper. However, because technology is operated by humans, human behavior can increase energy consumption. This shows that occupant behavior is one indicator to be considered in energy consumption, and in the development and implementation of an appropriate EnMS [50,51].

To include behavior as a variable that influences consumption, methodologies have been established to develop EnMS on university campuses. These methodologies include establishing a baseline of energy consumption in buildings, which can be established from a linear regression analysis in which significant variables that influence energy consumption on campus are considered. A graphical model used is consumption control diagrams, in which the consumption pattern of the building or equipment evaluated can be observed. Initially, to carry out any energy management methodology, it is necessary to have historical consumption data of the building, since, from these, it is possible to estimate consumption patterns or baseline consumption in such a way to apply the PDCA methodology (Plan–Do–Check–Act) proposed in ISO 50001 to implement an efficient and appropriate EnMS for an organization.

For effective energy management in buildings, it is important to measure the historical consumption of equipment, areas of the building and its subsystems. An EnMS can be very complex or simple according to the technologies used, management strategies, and its scope within the organization. Note that no matter how simple an EnMS is, it allows for considerable savings in the energy consumption of a building [52].

3. Materials and Methods

This research conducts a systematic review to understand the energy consumption and environmental effects on the implementation of BEMSs on HEIs, using the ISO 50001 standard as a guide for energy performance improvement on the buildings of organizations. This systematic review analysis tackled this objective by performing a literature review in several scientific databases, international agencies’ websites for energy development, government directives and non-governmental organizations (NGOs) using the search equations shown in Figure 4. The selection criteria used for article review was Energy Management Systems implemented in Higher Education Institutions. Based on the articles’ analysis, it is identified that methods usually used for the analysis of building energy performance are qualitative, quantitative and hybrid, and related topics to energy management found are energy policy, energy performance indicators, barriers to EnMS implementation, smart grids, occupants behavior, demand side management and ISO 50001.

3.1. Database and Data Collection

A literature research was conducted in the Science direct, Springer Link, Taylor & Francis, Dimensions and Google Scholar databases. The following keywords were used in this search: energy management systems, buildings, higher education institutions, energy efficiency, and sustainability. Within the search, we included 5 reports on energy performance in buildings and laws related to energy efficiency, demonstrating the importance of energy efficiency for the development of any kind of organization (public or private).

A database of 2310 articles was obtained, which were filtered according to the year of publication, title, keywords, abstract and their applicability to higher education institutions (HEIs) and non-residential buildings, from which 136 articles were selected for analysis and the extraction of information, such as EnMS models, barriers to energy efficiency, energy efficiency strategies, energy performance indicators, energy policies, measuring and forecasting of energy consumption, optimization models, demand for response, demand for side management, occupants behavior, fault detection, and diagnosis and reviews, was carried out.

3.2. Papers Analysis

It was found that out of the 136 articles analyzed, the highest publication density per continent was in Europe, with 40%, and Asia, with 29%, in contrast to Oceania and Africa, as the continents with the lowest publication density in these topics, being 2% and 4%. In Figure 5, the countries with the highest publication density are Italy and China, with more than 16 publications on the topic by country.

This shows that governmental laws affect funding research in energy efficiency for improving energy performance; so, this analysis shows the synergy between governmental laws and the research conducted by countries. Based on this, we can determine that major research on energy management in countries such as Italy is because of the countries’ public policies, which are focused on improving their energy performance to have a more reliable and secure energy system.

Also, the directives of the European Union for improving energy performance in buildings have raised funding related to energy management, energy efficiency, and energy security, which is shown in Figure 6, the continent being the one with the largest publishing density, followed by Asia, because of all the new policies related to energy efficiency issued in China.

Knowing the advances made by regions worldwide, energy management models in different buildings are analyzed, and energy consumption reduction data are collected, as well as the model applied to each one, considering that, depending on the location, the EnMS changes because of climatic and cultural conditions. In addition, the information is analyzed and discussed to identify patterns of energy performance improvement in buildings of higher education institutions or by implementing actions appropriate to the stages presented in the ISO 50001 standard.

4. Results

4.1. Energy Management Strategies in Higher Education Institutions

The analysis of energy consumption in buildings is a fundamental component of energy management systems (EnMS). This analysis provides critical insights into the equipment or systems that exhibit high energy usage within a building. By identifying these energy-intensive components, it becomes possible to implement targeted actions to optimize their performance and improve overall energy efficiency.

This consumption analysis can be performed in different ways, using smart energy meters, simulating energy consumption using software, implementing mathematical models, analyzing occupant behavior, and monitoring equipment efficiency, among others. This chapter presents original works that have performed the measurement and/or estimation of energy consumption in several higher education buildings, as well as in different types of buildings [53,54,55,56,57,58].

Acevedo, Noreña and Escobar [59] presented a model to determine, through mathematical estimations, electrical load profiles and consumption patterns of buildings, which can be applied to an EnMS. Initially, the architectural and operational information of the building was collected to integrate the actual behavior in the model. It was categorized into general and independent loads according to its location, service, and number of users. The influence of hourly occupancy was verified, and it can be used to identify the percentage of participation of different areas and loads in electricity consumption.

X. Zhou et al., 2013 [60], conducted a questionnaire survey to investigate energy consumption and measures adopted by higher education institutions in Guangdong Province, which provides relevant information for government policy to encourage energy efficiency improvement.

Energy management strategies allow for consumption reduction, according to a study conducted in Korea [61], where the energy consumption patterns of 12 university buildings were determined, and energy-saving strategies were determined by analyzing the potential of each building to reduce its consumption. Energy efficiency strategies (EESs) were proposed, as shown in

Table 1. Energy efficiency strategies table sourced from the work of Chung and Rhee [61].

EES	Description
1	Automatic Standby Switches
2	Reduction of the Power of the Lights
3	Internal Temperature Setting
4	Window Replacement
5	Replacement of Wall and Ceiling Insulation

, showing an energy potential conservation between 6% and 30%.

An excellent understanding of energy consumption or energy awareness is a key prerequisite for making timely decisions to reduce energy consumption [62]. Gormally et al. [63], conducted a survey to analyze policies that are not necessarily focused on improving energy efficiency but have an impact on reducing energy consumption on university campuses.

Yoshida et al. [64], developed a strategy for implementing an energy management system according to the sustainable campus program of the Osaka University in Japan; so, the campus buildings were classified into three categories, and, from each building category, actions were determined to be considered. Category 1 buildings are low-consumption buildings that can be converted to zero-energy buildings; category 2 buildings are medium consumption buildings, where consumption can be maintained; and category three buildings are high-consumption buildings, where it is decided to outsource the energy management of these buildings.

Sharma et al. [65] proposed a multi-time-scale building energy management system integrated with renewable energy sources. The aim is to dispatch building components on different time scales to counteract uncertain variations associated with renewable energy sources. On the longest schedules, fuel cells based on micro-combined heat and power cycle with energy storage systems are dispatched before uncertainty exists. The robustness and effectiveness of the proposed scheme were demonstrated in simulations to avoid generation and load uncertainties.

Guerrieri et al. [22], started from the idea that a university campus can be considered a small-scale model of a city; based on this premise, a quantitative method was used to estimate the changes a municipality would have to carry out to achieve a city with a profile of nearly zero energy. To meet this task, the study considers the low-carbon transition of the campus of the University of Palermo and applies a simple method used by Yoshida et al. [64], which provides a graphical and effective representation of the level of success of the campus in meeting the targets to reach near-zero energy.

Becoming a low-carbon campus is based on three pillars: (1) actions towards the buildings, (2) actions towards the exterior lighting of the campus, and (3) actions towards mobility. Based on the actions performed, a reduction in consumption of 18% was estimated.

Kolokotsa et al. [66], based on the premise that a university campus can be considered a small-scale city, propose an efficient energy management system at the University of Crete based on an internet connection using the IP protocol, with a network of sensors monitoring energy consumption in the building, which achieved an energy efficiency of 20%.

At the University of Troyes in France [67], a post-occupancy study of the building management system was used to analyze energy consumption over a period of 3 years. Initially, an energy audit was performed, followed by an air quality analysis with CO₂ sensors, and an analysis of the impact on energy consumption because of the occupation of the building’s lobby.

C. W. Chen et al. [68], proposed three dimensions to set energy performance indicators in educational institutions, (1) management policies, (2) buildings and equipment, (3) educational activities, and 9 more dimensions are added to these three dimensions, which are efficient procurement, management policy, daily behavior, reduction in energy consumption, reduction in energy resources use, carbon sequestration, educational and awareness activities, and teaching models for efficient energy use. Abu Bakar et al. [69], conducted a bibliographic review presenting methods for calculating energy performance indicators in buildings.

Sonetti & Cottafava [70], proposed different energy clusters, allowing for a meaningful energy performance comparison among campuses within international rankings to enhance the comparability of different campuses’ energy profiles, contributing to better energy performance assessment on university campuses and energy policies for improving energy performance on university campuses.

Ascione et al. [71], proposed a multi-objective methodology to retrofit a building at the University of Sannio in Italy to bring it to near-zero energy. Likewise, Pereira, Tavares and Soares [72] established a framework for the energy retrofitting of old cultural heritage university buildings.

Toub et al. [73], proposed a model-based predictive control for optimal operation of a micro-scale concentrated solar power (MicroCSP) for heat and electricity cogeneration in an HVAC system for an office building at Michigan Technological University.

It was found that just a few articles related to energy management strategies analyze the impact of the EnMS on the carbon footprint in education buildings because these articles focus on energy use, energy efficiency and reduction in energy consumption but not on the reduction in CO₂ emissions, as this broadens their focus.

Chamorro et al. [74], developed a guideline for conducting an energy assessment on campus universities, under ISO 50001, with restrictions to energy measurements. This methodology was implemented at the campus of the Universidad Autónoma de Occidente, resulting in an energy consumption reduction of 14.2% and a carbon footprint reduction of 0.42%.

Although most articles do not mention the ISO 50001 standard, the majority follow one or more of the requirements of the standard, which are energy baseline (EnBL), energy performance indicators (EPIs), and significant energy uses (SEU). With the objective of implementing energy management strategies in higher education buildings and achieving energy savings, it is found that studies obtaining a greater reduction in energy consumption are those that establish an EnBL.

Table 2. Summary of energy management strategies applied to HEIs.

Ref.	Authors	Building Typology	HEI	ISO 50001	EnBL	EPI	SEU	Strategies	Energy Saving	CO2 Reduction
[22]	Guerrieri et al.	Education Buildings	University of Palermo		x	x	x	(a) Improvement of the thermal insulation of the exterior wall.	18%
								(b) CO2 lighting and ventilation control
								(c) Use of efficient air-conditioning systems
[53]	Sukri et al.	Education Buildings	University of Technology Malaysia (UTM)	x	x	x	x	(a) Establishment of a Faculty Energy Management Committee.	14%
								(b) Awareness and energy saving campaign through talks and online based survey.
[54]	AlFaris et al.	Education Buildings	Princess Haya School, Dubai		x	x	x	(a) Development of an energy management program that:	35%	35%
								* Establishes an energy management committee
								* Identify baseline and target
								* Develop an energy policy
								* Implements an action plan
								* Monitors performance and results.
								(c) Energy audit on electrical appliances and a study on occupant’s behavior.
[56]	Escobedo et al.	Education Buildings	National Autonomous University of Mexico (UNAM)		x	x	x	(a) Install efficient lighting and cooling for refurbishment and new buildings.	7.5%	11.3%
								(b) Adapt hybrid systems (solar-LPG-electricity) in the showers and the Olympic swimming pool.
								(c) Substitute Diesel for LPG.
[57]	Eskander y Nitschke	Education Buildings	United Kingdom Universities		x	x	x	(a) Investment in renewable energy.	25%	29%
								(b) Carbon efficiency of power generation from natural gas and grid electricity.
[58]	Jomoah et al.	Education Buildings	King Abdulaziz University (KAU),					(a) Monitoring and control of occupancy, temperature, and air quality in the cooling ducts of rooms and offices through sensors.	14.13%
								(b) Damper Actuators
								(c) Motors of the AHUs with variable frequency drive to allow speed adjustment as the thermostat settings vary.
								(d) Color indicator to show if a room is occupied and the lights are on.
[60]	Zhou et al.	Education Buildings	Guangdong province Schools and Universities					(a) Full-time executive agency for energy conservation.	29%	25%
								(b) Implementation of completed energy saving projects
								(c) Use of renewable energy
[61]	Chung et al.	Education Buildings	Soongsil University					(a) Reduction of the electricity demand of auxiliary equipment and lights.	6–29%
								(b) Interior design temperature control
								(c) improvement of the thermal performance of the envelope
[62]	Sučić et al.	Education Buildings	Primary schools					(a) Assessment of current energy performance	14%
								(b) Analysis of opportunities to improve energy and water efficiency.
								(c) Reduction of direct and indirect GHG emissions
								(d) Waste reduction
								(e) Individualization of building-level responsibilities
[64]	Yoshida et al.	Education Buildings	Osaka University		x	x		(a) Improvement of the thermal insulation of the exterior wall.	22%
								(b) CO2 lighting and ventilation control
								(c) Use of efficient air-conditioning systems
[66]	Kolokotsa et al.	Education Buildings	University of Crete		x			(a) Installation of 19 energy meters to monitor the electricity consumption of various blocks of the university.	30%
								(b) Efficient management of public spaces through the CAMP IT energy management system.
								(c) Online user questionnaires to improve user satisfaction.
								(d) Optimization algorithms to improve the energy performance of the equipment.
[67]	Merabtine et al.	Education Buildings	University of Technology of Troyes			x	x	(a) Energy Audit
								(b) Indoor environment quality evaluation
								(c) Indoor air warmth evaluation
								(d) User surveys for thermal comfort together with numerical analysis and simulations.
[68]	Chen et al.	Education Buildings	University Campuses, High Schools, and Elementary Schools			x		(a) Efficient procurement
								(b) Management policy
								(c) Daily performance
								(d) Reduction of energy consumption
								(e) Reduction in the use of energy resources
								(f) Carbon sequestration
								(g) Educational and awareness-raising activities.
								(h) Energy efficiency teaching models.
[70]	Sonetti & Coffafava	Education Buildings	Politecnico di Torino-Hokkaido University		x	x		(a) Energy cluster for energy profiles comparison among university campuses.
[71]	Ascione et al.	Education Buildings	University of Sannio		x	x	x	(a) Building envelope: Use of special plaster, innovative coatings, thermal insulation layers, use of thermal absorption, low emissivity windows and solar screens.
								(b) Heating and cooling system: selection of efficient systems, such as efficient heat pumps, water chillers with cooling towers.
								(c) Installation of photovoltaic panels.
[72]	Pereira et al.	Education Buildings						(a) Characterization in-situ
								(b) Development of monitoring campaigns.
								(c) Development of validated simulation models.
								(d) Evaluation of thermal and energetic performance.
[73]	Toub et al.	Education Buildings	Michigan Technological University		x			(a) MicroCSP system for heat and electricity cogeneration integrated in the HVAC system.	37%
								(b) TES system for storage thermal energy.
								(c) MPC for optimum operation of the MicroCSP system for thermal confort in the office building.
[74]	Diaz et al.	Education Buildings	Universidad Autonoma de Occidente	x	x		x	(a) Analysis of energy consumption history	14.2%
								(b) Identification of significant energy uses and relevant variables.
								(c) Identification of opportunities for improvement in the use of energy.
[75]	Yuan et al.	Education Buildings	Osaka City University		x	x	x	The Energy Management strategies used were:	3.5%
								(a) Use of webcams for monitoring gas consumption.
								(b) Use of web cameras to identify the status of lighting and air conditioning (ON/OFF).
								(c) Use of current transformers to sense energy from the distribution boards.
[76]	Tong et al.	Education Buildings	Tianjin University		x			(a) Change of the HVAC system temperature set point.	44.3%
[77]	Semprini et al.	Education Buildings	School of Engineering and Architecture in Bologna		x			(a) Improvement of the control system of the heating plant	15–30%
								(b) Improvement of the heat generation subsystem
								(c) Improvement of the thermal performance of the windows

The x indicates that the corresponding reference implemented the respective element.

summarizes the energy management strategies applied to HEIs.

4.2. Measuring and Forecasting Energy Consumption

For proper implementation and monitoring of an EnMS, it is necessary to have reliable measurement and control systems, as these collect consumption data, which are used to perform the analysis and calculate energy performance of the building. For example, in a university campus located in India, they decided to perform a measurement at the device level with digital meters communicating by Wi-Fi, where they used smart sensors to obtain more detailed data to define the demand patterns of the building and make comparisons with each device to take corrective actions on them [78].

It is necessary to have a good programmed of all meters as they may have some communication problems and information may be lost; however, consumption patterns can be determined, as happened in a study where an educational building belonging to a university in Australia through a process of data mining of its 64 sub-meters made a prediction of consumption in which the estimated consumption patterns are accurate. However, the actual electricity and gas consumption were found to be 2.4 and 3.1 times the predicted consumption, indicating a considerable gap in the building’s energy performance [5].

According to the literature, there are three types of models for measuring and estimating energy consumption in buildings [3,79]. The first is a physics-based method, commonly known as white box modeling, which uses a procedure based on physical calculations to describe the energy performance of buildings [80]. The second is black box or data-driven methods, which are mainly associated with statistical time series analysis and machine learning algorithms to evaluate and forecast the energy consumption of buildings [81]. The third is known as the gray box model, which is described as the combination of the black box model and gray box model, also known as hybrid method, which can be an improved method that includes statistical data analysis models with optimization models [82]. The three types of models for measuring and estimating energy consumption in buildings are shown in Figure 7.

Other work using data mining has estimated energy consumption patterns in office buildings, which can lead to greater energy conservation, identify potential energy waste patterns, and detect faulty electricity load profiles [83,84].

Data mining is also used to benchmark buildings based on data from different buildings. As carried out in Dias Pereira et al. [85], to obtain comparison indicators between different educational institutions, establishing a benchmark for energy performance. Panapakidis et al. [86], developed a methodology using the k-means machine learning algorithm to recognize patterns of load curves to predict the behavior of load profiles in the buildings of a university in Greece.

At Keio University in Japan, a model was developed to control and predict university energy consumption patterns with error rates between 0.88 and 0.1 using the KNIVES system (Keio University Network Oriented Intelligent and Versatile Energy saving System) [87]. This study illustrates Keio University’s Sustainable Campus Initiative that students have undertaken to make the campus more sustainable and resilient. Several ideas for maximizing energy efficiency and conservation are discussed [88].

Raza et al. [89], developed a load demand forecast for a PV energy system integrated in smart buildings of a university campus in Australia, they used five different predictors as:

• Backpropagation neural network (BPNN);
• Elman neural network (EN);
• Autoregressive Integrated Moving Average (ARIMA);
• Feed forward neural network (FNN);
• Radial basis function (RBF);
• Wavelet transform (WT).

Models to remove spikes and fluctuations: The FNN and RBF network were trained with particle swarm optimization (PSO) for higher forecast accuracy. The output of each predictor in the ensemble network was combined using Bayesian model averaging (BMA).

Other methods, such as equipment monitoring, can also be used to determine energy consumption. At the Sugimoto campus of Osaka City University, a metering system was developed that consists of web cameras and current transformers; the cameras identify the status of lighting, air conditioning and gas consumption by processing gray scale images, and current transformers conducts a census of consumption by time intervals in different distribution boards, thus obtaining data on the consumption of electricity, lighting, air conditioning and gas [75].

Besides monitoring, it can track the behavior of building occupants to estimate energy consumption based on occupancy schedules, building loads, nominal consumption, load states, space, equipment use, and operating conditions. In Diaz-Acevedo et al. [59], a study was carried out in a tertiary building in Medellín where energy consumption was estimated hour by hour, obtaining information through guided visits and interviews; the variation between the measurements and the estimated load profile was lower than 18%.

In various studies, software tools have been employed to simulate energy consumption. For instance, Tong and Zhao [76], used EnergyPlus to simulate a new campus under construction at Tianjin University. This simulation allowed them to analyze potential energy savings based on the buildings’ characteristics, providing detailed hourly energy consumption data. Additionally, other software tools, such as TRNSYS, IDA ICE, and EnergyPlus, have been widely recognized for analyzing energy consumption, costs, and thermal comfort [90,91].

Models have also been developed to evaluate the thermodynamic behavior of a building on the campus of the University of Genoa [92]; the model is based on a resistive–capacitive (RC) circuit, where resistors model the conductive and convective thermal resistances, while capacitors are used to consider accumulation of thermal energy inside the building; this is another alternative which reduces the high computational cost involved in the use of above-mentioned software.

However, simulation software allows for the integration of other factors, such as occupant behavior in the building. Hax et al. [93], defined three types of users: careless user, intermediate user and proactive user. Using EnergyPlus software (https://energyplus.net/), the authors performed a simulation to measure the influence of behavior patterns on the energy consumption of a building on the campus of the Federal University of Pelotas. The study concluded that the user’s attitude significantly influences the energy performance of a building. Specifically, energy consumption can vary by up to 84% for cooling, 35% for lighting, and 30% for total annual consumption between proactive and careless users.

The methods to measure energy consumption and propose scenarios can be summarized as follows: smart metering, data mining, simulations using software, mathematical models, load monitoring, and the development of a load profile using information from the building. From these methods, information is obtained to develop a scenario that can be optimized through demand management.

At the school of engineering and architecture of Bologna, the Building Energy Signature (BES) was implemented, which defines the thermal power required by the building from outside air temperature. In buildings, the BES curve is usually linear; based on data collected from energy consumption and meteorological data of the outside temperature in different periods of time, a linear regression is performed and the actual BES is obtained; with this model, one can identify if there are anomalies, faults, misconfiguration in the control system and/or the presence of anomalous ventilation rates in the heating plant system of the building [77].

Table 3. Summary of models for measuring and forecasting energy consumption.

Measurement, Forecasting Model of Consumption	Ref.	Authors	Building Typology	HEI	Analysis
Black Box Model	[83]	Liu et al.	Non-Residential Buildings		Density-based spatial clustering application with noise (DBSCAN)
					K-means algorithm
					Classification
					Regression tree (CART)
	[84]	Zhou et al.	Non-Residential Buildings		K-means clustering,
					Decision tree,
					2-dimension scatter diagram
					Outlier detection
	[86]	Panapakidis et al.	Education Building	Aristotle University of Thessaloniki (AUTH)	Load curve data are transformed to the frequency-domain using the Fast Fourier Transform (FFT) algorithm.
					Minimum variance criterion (MVM)
					Self-organizing Map (SOM)
					Fuzzy C-means (FCM),
					Knee-point criterion for the ratio of Within Cluster Sum of Squares to Between Cluster Variation (WCBCR) measure.
	[89]	Raza et al.	Education Building	The University of Queensland	Combination of five different predictors named as backpropagation neural network (BPNN), Elman neural network (EN), Autoregressive Integrated Moving Average (ARIMA), feed forward neural network (FNN), radial basis function (RBF) and their wavelet transform (WT) models. WT is applied to historical load data to remove the spikes and fluctuations. FNN and RBF network were trained with particle swarm optimization (PSO) for higher forecast accuracy. The output of each predictor in the ensemble network is combined using Bayesian model averaging (BMA).
	[94]	Alam et al.	Education Building	Swinburne University of Technology	K-means clustering
Gray Box Model	[78]	Birindha et al.	Education Building	Karunya University	Device Level Analysis
	[87]	Kamiyoshi et al.	Education Building	Keio University	Cluster Energy Management System (CEMS)
					KNIVES (Keio University Network Oriented Intelligent
					and Versatile Energy saving System)
	[91]	Sharif et al.	Education Building	Concordia University	Life Cycle Assessment (LCA)
					Simulation-based multi-objective optimization
White Box Model	[75]	Yuan et al.	Education Building	Osaka City University	Current transformers to measure electric consumption through distribution box.
					Web cameras monitor lighting in each room, as well as air conditioning status by recognizing the movement of a ribbon at the outlet.
					Web cameras were installed in several rooms in one section of the campus to record lighting and air conditioning use data.
	[76]	Tong et al.	Education Building	Tianjin University	Building load simulation
	[77]	Semprini et al.	Education Building	School of Engineering and Architecture in Bologna	Building Energy Signature (BES)
	[85]	Pereira et al.	Education Building		Literature Survey
	[90]	Mazzeo et al.	Building		Building Performance Simulation (BPS)
	[92]	Bianco et al.	Education Building	University of Genoa	Simulator for the evaluation of the thermal dynamic behavior of a building at room level has been developed basing on an equivalent electric model and by focusing on the operation of the HVAC system.
	[93]	Hax et al.	Education Building	Federal University of Pelotas	Computer simulations were performed in the EnergyPlus program, to measure the influence of behavior patterns on the building’s energy consumption.

summarizes the energy consumption measurement and forecasting models.

4.3. Demand Management Models

This chapter presents actions that have led to improvements in energy performance, either at the level of operation of the building loads or retrofitting of the building.

There have been several studies that seek to optimize energy consumption in buildings, carrying out measurement models, prediction, and control of subsystems of buildings such as energy generation systems, HVAC systems, storage systems, control systems, and automation and demand management systems; below are some control models, simulation and estimation aimed at making buildings more energy-efficient [95,96,97,98].

4.3.1. Optimization Models

As previously mentioned, energy generation is the most important subsystem in educational institutions. Energy generation leads to the reduction in operating costs, however, in places where energy rates are based on Time of Use (TOU) it is an important schedule time in which it is generated, since greater savings can be achieved. At the campus of the University of Engineering and Technology of Taxila, a microgrid proposal was made with the objective of reducing operating costs, with different generation scenarios between solar photovoltaic, energy storage systems, energy from grid, diesel generation, and wind turbines, it is obtained that the best combination for savings is the connection between the grid and solar photovoltaic energy with savings of 43.6% [99].

Ullah et al. [100], proposed a bio-inspired algorithm for energy scheduling of a university campus, combining two algorithms: lyon and firefly; its results show that the energy costs were greatly reduced. Seyedzadeh et al. [101], proposed an algorithm using machine learning, based on multi-objective optimization (MOO), to predict load parameters of building HVAC systems; these predictions were validated through simulations in Energy plus.

Gul and Patidar [17], presented a pilot study to analyze the relationship between electrical energy demand profiles and user activities in a university building where the shape and magnitude of energy demand profiles showed a significant trend that does not seem to be strongly related to occupancy patterns. Likewise, in the work of Degha et al. [102], an intelligent building simulator was developed, proposing an intelligent and context-aware energy management system. This system was tested, and it yielded savings of 40% of total energy consumption.

Di Piazza et al. [103], proposed an energy management system based on the prediction of generation profiles from renewable energies and load demands, using an artificial intelligence method based on neural networks. This energy management system was validated in a smart house with an annual energy consumption of 4500 kWh, where energy efficiency increased by 2.43%.

Zhu et al. [104], proposed a systematic process of quantifying load patterns and identifying abnormal energy consumption. The model consists of three main stages: using simple and efficient algorithms to preprocess time series frequency of building loads of significant daily profiles; the second stage is the construction of prediction models by selecting appropriate data mining algorithm; and, in the third stage, the residuals are analyzed by a statistical control theory for each load profile and construction of a control chart with an appropriate upper boundary.

Dagdougui et al. [105], proposed a model which handles the intelligent and programmed charging and discharging of an electric vehicle connector, which is supplied by a microgrid in a smart building, to reduce the peak loads of the building.

Fan et al. [106], developed a methodology to determine the optimal COP in an HVAC system in a multi-purpose building in a university in Hong Kong; they clustered the COP in two levels, low and high, and used the k-means clustering model to obtain the optimal COP, using three types of variables: the first to describe the outdoor environment, the second were the operating parameters, and the third were time variables.

Biyik and Kahraman [107], developed a predictive control model to coordinate the operation of the HVAC system, battery storage system, and renewable energy generation in buildings. The control algorithm was tested in a building at Yasar University in Turkey, finding that the control system can reduce energy consumption by up to 23%.

Chen, Tan, and Berardi [95], established a hybrid model to predict hourly electrical demand in a hotel building, which was classified as a non-stationary operating building because of irregular temporal electrical characteristics.

The attention of the scientific community is shifting towards predicting the energy resilience of buildings to changes in local climatic conditions because of global warming. The scientific community also aims to identify the best energy efficiency improvements for buildings to mitigate the effects of global warming and reduce energy demand and greenhouse gas emissions to the atmosphere [108].

Table 4. Summary of technologies used for optimization models.

Ref.	Author	Building Type	HEI	Technology	Energy Saving
[95]	Chen et al.	Non-Residential Buildings		Fuzzy c-means (FCM) clustering
				An online modified predictor was established by the combination of support vector regression (SVR) and wavelet decomposition
[96]	Gruber et al.	Buildings		Two-stage optimization based on a receding horizon strategy that minimizes the energy costs under consideration of the physical system constraints
[97]	Reynolds et al.	Non-Residential Buildings		Artificial neural networks (ANN)	25%
[98]	Santamouris et al.	Educational Buildings		Data have been collected through energy surveys performed in 320 schools in Greece.
				Innovative energy rating scheme based on fuzzy clustering techniques
[99]	Javed et al.	Educational Buildings	University of Engineering and Technology (UET), Taxila	Microgrid Management Model (uG)
[100]	Ullah et al.	Educational Buildings	Suranaree University of Technology	Fire-Fly Lion (FLA) Bio-inspired Algorithm
[101]	Seyedzadeh et al.	Residential and Non-Residential Buildings		Machine Learning (ML)
[102]	Degha et al.	Smart Buildings		Intelligent Context-Awareness Building Energy Management System (ICA-BEMS)	40%
[103]	Piazza et al.	Smart Buildings		An optimization algorithm, that starts from profiles of renewable energy generation and load demand, which are obtained by a forecasting method based on suitably chosen and trained Artificial Neural Networks
[104]	Zhu et al.	Non-Residential Buildings		ARIMA (auto-regressive integrated moving average) model of daily consumption.
				Linear Regression, LASSO, SVM, CART, Neural Networks, kNN and Random Forests
[105]	Dagdougui et al.	Smart Buildings		Integrating microgrid
				Dual tracking control problem subject to the quadratic cost function
[106]	Fan et al.	Education Buildings	The Hong Kong Polytechnic University	K-means algorithm, The multi-layer
				Perceptron, the C-type support vector machine with a Gaussian radial basis kernel function, extreme gradient boosting trees and Random Forest.
[107]	Biyik et al.	Educational Buildings	Yasar University	Model of Predictive Control to coordinate HVAC system operation, battery storage system, and renewable energy generation in buildings.	23%

below shows a summary of the technologies used for the optimization models.

4.3.2. Demand Response

The response to demand is a change in the normal use of electricity by the consumer in response to, usually, monetary motivations [109]. Hence, the “human dimensions” of energy use in buildings refer to the energy-related behaviors of key stakeholders affecting energy use during the life cycle of the building. Stakeholders include building designers, operators, managers, engineers, occupants, industry, suppliers, and policy makers, who directly or indirectly influence the design, construction, housing, operation, management, and regulation of the built environment, from the individual building to the urban scale [110].

An existing literature review demonstrates that human factors can contribute substantially to this variation in energy use [110]. Martirano et al. [111], proposed a central automated control system, which, through demand management, allows for flexibility and self-awareness of users in energy consumption, which reduces the peak loads of the building, obtaining peak load reduction values between 25% and 30%.

Naylor et al. [112], described different types of technologies focused on characterizing the occupancy of a building and evaluated the accuracy of measuring occupancy and number of occupants, as well as energy savings for each technology implemented and changes in occupant behavior regarding energy consumption.

Rotger-Griful et al. [113], developed a new BEMS to strengthen the adoption of demand response by building residents. The building energy management system proposed monitors the building and its surroundings, can interact with residents, optimally controls distributed energy sources, and provides demand-side management.

Sala-Cardoso et al. [114], presented a hybrid methodology to estimate demand patterns to optimize HVAC systems in educational buildings in Spain. The methodology consists of using an activity indicator defined as a percentage of active spaces; the activity in the zone is monitored with presence detectors used in modern buildings.

Cotton et al. [115], studied student energy behavior at universities in Portugal and England; they found that different student responses reflect institutional priorities and national contexts, and show a collective capacity of occupants to change their energy use behavior.

Table 5. Demand response models.

Ref.	Authors	Building Typology	Country	Demand Response Model
[109]	Siano	Residential Buildings	Italy	Rate-based or price DR programs:
				Demand reduction bids
				Incentive or event-based DR programs
[110]	D’Oca et al.	Buildings	United States	Adaptive actions: opening/closing windows, lowering blinds, adjusting thermostats, turning lighting on/off, and using plug-ins (such as personal heaters, fans, and electrical systems for space heating/ cooling).
				Non-adaptive actions: occupant presence and operation of plug-ins and electrical equipment (such as office and home appliances), as well as building occupancy and movement through spaces.
[111]	Martirano et al.	Residential and Non-Residential Buildings	Italy	Exploits the flexibility of smart appliances and the thermal inertia of the structure, by imposing local and central set-points of heating and cooling systems according to actual global net load and generation at a given moment.
[112]	Naylor et al.	Buildings	United Kingdom	Highlight recurring patterns in sensor data, which can then be manually assigned to activity types.
				Entrance/exit people-counting sensors can be used to identify anomalous events in occupancy and their time span as these events have the greatest impact on the effectiveness of scheduling in building control systems.
				Optimization of localized lighting through visual detection of location and activities, desk lighting by combining sensor types to determine space use, and desk-level lighting, electrical and HVAC control based on presence and desktop computer use.
				Predict the occupant-based context of a space in the near future in order to pre-emptively condition the space to acceptable levels.
[113]	Rotger-Griful et al.	Residential Buildings	Denmark	Monitoring energy usage and relevant external information of the building (e.g., weather forecast).
				Controlling DERs within the building in an optimal manner (energy efficiency).
				Enabling demand response provision and load aggregation to a third party (aggregator)
				Empowering consumers by providing them with information and energy related recommendations.
[114]	Sala-Cardoso et al.	Education Buildings	Spain	Monitoring of demand response using an activity space indicator, monitored with presence detectors.
				The method used in this study for the implementation of the load forecasting is an Adaptive Neuro-Fuzzy Inference System (ANFIS).
[115]	Cotton et al.	Educational Buildings	Portugal	Provides a foundation to extend the comparison to other institutions and other countries, and to expand the research to encompass actual energy use, in relation to perceived energy use.
				Turn off lights when they are not in use.
				Turn down the heat.
				Try to save water.
				Walk or cycle short distances instead of going by car.
				Buy things that are likely to involve less energy or resource use.
				Pay a bit more for environmentally friendly products.
				Avoid charging mobile phones overnight.
				Turn off the stand-by button of the TV set or switch appliances off at the plug.
				Use rechargeable batteries.

shows the demand response models.

4.3.3. Energy Management in HVAC Systems

In recent decades, responding to the high impact of buildings on global energy consumption and greenhouse gas emissions, international directives and research have focused on the physical aspects of buildings such as building envelopes and management of heating, ventilation, and air-conditioning (HVAC) systems [116].

Since heating and cooling systems are the largest consumers of energy in buildings, they have been extensively studied in the literature. Some studies related to the improvement in energy performance in HVAC systems are presented and analyzed below.

Jiang et al. [117], conducted the energy management of an HVAC system in a university library in China; this management system is composed of a Siemens APOGEE automatic building control system, monitoring software, digital controllers, sensors and actuators. The remote monitoring function is performed from the central station to each substation through MODBUS-TCP/IP and Ethernet/IP protocol; this management system reduced energy consumption by 5.6% in an air-conditioning system.

In buildings, the HVAC system is one of the most energy-consuming because of the thermal load associated with it in different periods of the year; to reduce the consumption of this system, an important factor of the building is the envelope. A multidisciplinary energy audit was performed on the campus of the University of Architecture and Construction of Azerbaijan, where different factors that can be improved to reduce energy consumption were identified. The factor that had the greatest impact on consumption was the building envelope; its heat transfer coefficient leads to large heat losses. The improvement proposals that have the greatest impact on energy consumption are based on thermal insulation of the facade and installation of new facades [118].

Although modifying the building envelope has a great impact on energy consumption, there are historic buildings that must preserve their architectural value and therefore the possibility of modifying the envelope is limited, as is the case of the Bologna School of Engineering and Architecture. In this case, to improve energy savings, it is proposed to use thermostatic valves on each radiator to control the internal temperature and reduce heat dispersion, replace conventional gas boilers with condensing boilers, replace single glazed windows with ultra-thin vacuum insulating glass of the same thickness, and restore the frames [77].

Tong and Zhao [76], discussed two types of buildings on the campus of Tianjin University: residential buildings and public buildings. Using the Energy plus software, they performed consumption simulations of these buildings by keeping the temperature setting fixed and then varying the temperature setting; it was identified that annual energy consumption can be reduced with optimization strategies in HVAC systems. Moreover, active building envelopes have been developed, as presented by Luo et al. [119], where most studies related to building envelopes have focused on numerical simulation of these spaces.

In the development of air-conditioned spaces in the education sector, it is recommended that the internal temperature be dependent on external climatic conditions and investment should be dedicated to buildings with lighting and external ventilation. In addition, it is necessary to renovate old classrooms to improve energy performance and have adequate thermal comfort for education, and guidelines should be followed to have quiet educational environments [120].

Li et al. [121], developed a new method known as adjacent information retrieval (AIR) to eliminate problems related to fault detection and diagnosis (FDD), which are mainly associated with measurement errors and loss of measurement data. They found that, with this model, they can recover lost meter data. AIR is a machine learning model, which uses Bayesian approaches to retrieve data, and was demonstrated to retrieve data with real data taken from the ASHRAE 1312 research project.

In the work of Du et al. [122], distributed, robust optimization (DROA) was proposed to schedule the energy consumption of a heating, ventilation and air-conditioning (HVAC) system, considering weather prediction error. The maximum outdoor temperature interval was divided into sub-intervals and the proposed model increased the ambiguity because of probabilistic distribution based on historical information. It was shown that the proposed DROA helps to reduce electricity costs.

An energy management problem was studied, where the relationship between operating states and energy consumption of loads with prediction errors is evaluated. An energy management algorithm was developed to find the optimal operating states and trading strategies for energy consumers and suppliers. The algorithm was tested on an HVAC system through simulations, where it was shown that the algorithm converges to optimal operating values with a demand equal to the energy supplied with a reliability margin; it was observed that daily loads and costs were reduced using the energy management algorithm [123].

A probabilistic approach was proposed and applied to realistically simulate occupant behavior. The methodology was based on the probabilistic evaluation of input and output variables of building energy simulations. The objective was to compare the results obtained with the use of a deterministic approach (simulation). Three models of occupant interactions with heating control were determined, obtained and implemented in building simulation tools [124]. In the work of Yuan et al. [75], a correlation was found between the indoor temperature and electrical energy consumption by the air conditioner.

A methodology to assign HVAC operation planning schemes to connected buildings with the objective of energy saving and load leveling was proposed. The objective was to assign periodic temperature setpoints to a building cluster so that the aggregate cooling electric demand reduces with the minimum cost and a steady aggregate load shape. A 12.5% energy saving was achieved [125].

In the work of Gomez-Romero et al. [126], an MPC algorithm is proposed, which uses a highly complex grey-box simulation model to optimize the operation of an HVAC system. The system generates hundreds of operation plans, typically for the next day, and evaluates them in terms of comfort. It was implemented and tested in an office building in Helsinki, both in real and simulated buildings, and found energy savings between 35% and 20%.

Table 6. Summary of energy management systems for air-conditioning systems.

Ref.	Authors	Building Typology	Country	HEI	Energy Management System for HVAC System	Energy Saving
[116]	Jindal et al.	Educational Buildings	United States		A heuristic-based algorithm is proposed which optimally minimizes the use of HVAC without affecting user comfort. Moreover, it also minimizes the cost of rescheduling the classes on a given day.	19.75%
[117]	Jiang et al.	Educational Buildings	China	Xian Jiaotong University	Siemens APOGEE building control system including insight monitor software, direct digital controllers, sensors, and actuators.	5.6%
[118]	Samira et al.	Educational Buildings	Azerbaijan	Azerbaijan University of Architecture and Construction	Multi-disciplinary energy auditing	56%
[119]	Luo et al.	Buildings	China		Active building envelope (ABE)
[120]	Vijayan et al.	Buildings	India		Renovating old classrooms into the new classroom is essential to provide a satisfactory schooling environment
[121]	Li et al.	Buildings	China		Data-driven fault detection and diagnosis (FDD)
					Adjacent information recovery (AIR) filter
[122]	Du et al.	Buildings	United Kingdom		Distributionally robust optimization approach (DROA) is proposed to schedule the energy consumption of the heating, ventilation, and air conditioning (HVAC) system
[123]	Ma et al.	Buildings	Singapore		Energy management algorithm is developed to seek the optimal operating states and pricing strategies for the consumers and the energy provider.
[124]	Fabi et al.	Buildings	Denmark		A probabilistic approach is proposed and applied to simulate occupant behavior realistically logistic regression was used to infer the probability of adjusting the set-point of thermostatic radiator valves
[125]	Ghofrani et al.	Buildings	United States	Simulated University Campus	A periodic temperature setpoint to save energy by using building thermal inertia.	12.5%
[126]	Gomez-Romero et al.	Non-Residential Buildings	Finland		A novel MPC algorithm that uses a full-complexity grey-box simulation model to optimize HVAC operation in non-residential buildings.

shows a summary of energy management systems for air-conditioning systems.

4.4. Diagnosis and Detection of Risks and Opportunities

There are barriers to energy efficiency implementation. Nunayon et al. [127], discussed these barriers, conducting a questionnaire with 4 stakeholders, which are 361 respondents in a public university in Nigeria, and found that the main factors are lack of information, money, awareness and time [128]. Complex decision chains and lack of reinforcement are also major barriers to sustainability on campuses that are lost behind non-binding declarations rather than commitments to transformation [129].

A methodology was developed to characterize building energy time series and identify infrequent and unexpected energy patterns. The process is based on an enhanced version of the Symbolic Aggregate Approximation (SAX) approach. This improved method includes an optimized adjustment of the time window width and symbol intervals, tailored to the building’s energy behavior.

The SAX model was coupled with decision trees for anomaly detection and unexpected energy patterns. The robustness and flexibility of the system was evaluated by testing it in two buildings: the first one was in a town hall in Spain, and the second case was at the Polytechnic University Di Torino [130].

Deshmukh et al. [131], present analytical methods embodied within useful software tools to quickly identify and evaluate selected faults in air handling units (AHUs) that cause large building energy inefficiencies. Algorithms for faults like stuck dampers and leaking dampers were developed and tested. An example is presented of the potential energy savings in a large academic building that was monitored. Real data from an academic building in Boston were collected from the building energy management system (BEMS) for this study. Savings of around USD 3400/month or 18% were determined.

Li et al. [132], propose a novel expert knowledge-based unseen fault identification (EK-UFI) method to identify unseen faults by employing the similarities between known faults and unknown faults. The similarity is captured by incorporating essential expert knowledge that is encoded in the fault gene matrix.

Amaral et al. [133], through a systematic review, classified cases of ineffective factors in the application and implementation of sustainable campuses into four groups: technical, economic, climatic and behavioral factors. Within these groups, the major causes were inadequate planning, inappropriate system design, low return on investment, mismatch between actions and climatic conditions, or uncertainty of long-term sustainable behavior. Based on this, lessons learned are proposed, as shown in

Table 7. Lessons learned to implement an EnMS.

Lessons Learned

The need for a clearer understanding of the context of the technical options for the implemented actions chosen. The advantage of considering the campus community with shared needs and resources, as opposed to a sum of isolated buildings. The importance of monitoring the implemented actions and of including maintenance in both the economic and technical analysis to prevent unexpected costs or malfunctioning during the systems operation phase. The constricting impact of external financing programs which are conceived for specific type of measures, leaving out more appropriate or urgent ones. The robustness of solutions which consider local current and future climate data in their technical analysis. The need to promote long-standing sustainable behaviors by fostering an integrated culture of sustainability within HEIs.

.

Biresselioglu et al. [134], examine, through the literature, attitude, dimensions, motivators and barriers affecting energy behavior, and it was found that one barrier is the lack of focus between the user and their interaction with energy efficiency technologies.

Altan [135], poses the following barriers to energy efficiency: lack of a methodology for energy audit studies, lack of transparency in the complex problems associated with energy demand and use, difficulty in establishing boundary conditions for the evaluation of energy efficiency performance, indices and their usefulness as indicators of energy performance, hiding relevant information on actual energy performance, and the quality and reliability of aggregate data and results, which hide trends in the efficiency and effectiveness of certain projects.

Whitney et al. [136], suggest that the purpose of a commercial building—profit generation or provision of public service—is of fundamental importance to a system-level understanding of energy management and directly links to stakeholders’ views on decision-making in their roles and responsibilities. There are five major modes of behavior that can lead to increased energy management activity: (1) regulated building improvements; (2) voluntary building improvements; (3) voluntary operational/management improvements; (4) leveraging competitive advantage; and (5) mainstreaming green.

The paper above presents an analysis of the understanding and expectations of the interested parties, one aspect key to ensuring the implementation of the energy management system in an organization; for that reason, the ISO 50001 standard considers a requirement the understanding of the needs and expectations of the interested parties in the implementation of the management system (see Figure 8), which are:

• The owner controls the financing because this stakeholder can commit to the organization.
• The property manager is focused on investment value and is supportive of sustainability but is risk-averse.
• The facility manager requires equipment technology that can be maintained; so, the maintenance staff needs training and assurances. This stakeholder is the normal focus of EE.
• Users agree to comply with the constraints of the sustainability program if the program does not limit the mission goals of the organization.

4.5. Review Analysis

From the papers reviewed, 46% focus on educational institutions, 17% are reviews of the state of the art focused on energy management in buildings, 13% are on buildings, 9% are on non-residential buildings, 7% are on smart buildings, 3% of the papers deal with energy management systems in industry, 2% are on residential and non-residential buildings, 2% are exclusively focused on residential buildings, 1% focus on nearly zero-energy buildings, and 1% focus on non-residential and educational buildings.

Furthermore, 27% of the articles focus on energy efficiency strategies, where different actions to improve energy performance are discussed, and 11% present energy consumption measurement and prediction models, which show energy consumption measurement systems, and computational algorithms to predict energy consumption patterns and weather conditions prediction, to adapt lighting and equipment operation and maintenance.

In addition, 11% focus on energy management in HVAC systems, 8% of the articles focus on predictive control models, which integrate measurement, control and machine learning models, to improve energy efficiency in buildings, 5% focus on the impact on consumption because of user behavior, 5% talk about data analytics to predict energy patterns, 3% are models to optimize the operation of large energy consuming equipment and barriers when implementing an EnMS, 4% discuss energy demand response, energy policies and the façade of buildings as an energy storage or generation system, 2% report on the integration of the Internet of Things (IoT), energy performance indicators, renewable energy sources implementing an EnMS, retrofitting of buildings and simulation of energy consumption with different commercial software, and the remaining 1% focus on energy use, commercial energy management software, smart grids and load patterns, and anomaly detection.

Figure 9 divides the analysis into three key topics: ISO 50001, energy management, and technology change. It is observed that out of 136 papers, 27% are related to operation strategies, 26% perform energy audits, 21% are related to energy use planning and 4% implement maintenance strategies, according to energy management. Despite most papers not complying with the ISO 50001 requirements, all of them implement some of the ISO 50001 standard recommendations; 46% conduct measurements, estimations, and forecasting, 38% implement an EnBL, 21% apply and develop an EPI, 11% obtain an SEU, 26% perform an energy review, 14% propose actions to address risks and opportunities, 23% are related to design, communication, and activities needed for improving and maintaining the building’s energy performance (support), 8% plan and follow the requirements for planning an EnMS, and 1% of the papers are related to stakeholders.

It shows that despite just a few papers mentioning the ISO 50001 standard, all of them follow at least one requirement to improve the building’s energy performance, as shown in Figure 10. Most of them include EPI, and baselines, and a lot of energy management focuses on measuring, forecasting and simulating energy consumption to reduce the energy intensity of the buildings.

4.6. Future Challenges

In the literature, it was found that EnMS can be updated to have a continuous improvement, and to make them more efficient to reduce or maintain adequate energy consumption; therefore, several challenges have been identified for future improvement, as follows.

Digitizing EnMS is currently one of the most critical challenges, transforming how a building consumes, generates, and stores energy. The digitization of energy management in buildings will integrate digital technologies such as smart metering, real-time monitoring, IoT, cloud computing, blockchain, smart control, smart grids, energy management analysis software, communication networks, AI, and smart grids. The digitization of energy management in buildings will contribute to increasing building energy performance, making buildings more sustainable and comfortable.

In the work of Hannan et al. [137], the Internet of Energy, or IoE, which combines the characteristics of IoT and smart grids, was widely studied with distributed energy sources, as they enable reliable control and supply of electricity to consumers [138].

Smart grids will bring various advantages for energy management in buildings as they will allow for the optimization of energy flow while reducing losses. Smart grids allow to us monitor and manage energy consumption in real time, which will facilitate the integration of renewable energies and allow for more efficient demand management by adjusting energy consumption and generation, contributing to more efficient and reliable energy management.

According to Molina-Solana et al. [139], this can be achieved through appropriate synergy between energy and technology companies, to reduce meter costs, cloud computing, open data, etc. New developments focus on smart metering, which allows for real-time metering, enabling better visualization and user understanding of their energy consumption, with the advantage of increasing energy efficiency, encouraging users to change their energy consumption habits.

The connection of Big Data, which is understood as a large amount of data that can be analyzed to improve energy performance, and the Internet of Things (IoT), which is a term that refers to the communication and storage of energy consumption data and its visualization by users through an internet network, allows for the improvement of associated management tasks, with high precision and a fast response, saving energy, prioritizing uses, according to policies, and responding to outages.

Cloud computing, which aims to reduce computing costs and increase the flexibility and reliability of systems, can be defined, according to NIST, as a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of resources such as networks, servers, storage, applications, and services that can be rapidly provisioned and released with minimal management effort [140]. All this leads to security and privacy issues; therefore, methods must be developed to comply with legal frameworks, as well as to protect the security and privacy of users. In addition, metrics of accuracy and uncertainty about the predictions of the cases associated with energy consumption must be met.

A lack of systematic approaches to energy transition has been found for university campuses seeking to be a smart campus. The release of renewable energy and new storage systems are not being integrated with energy efficiency programs, and demand response schemes and bidirectional communication in the energy system [141].

The ISO 50001 standard has had a great impact on energy management in various types of organizations, but has been little applied in higher education buildings. The authors consider that future work should include the implementation of a management system in higher education buildings taking into account all the requirements of ISO 50001, such as methodologies to determine the EnBL and EPI that allow for the identification and control of relevant variables. In this regard, Castrillón-Mendoza et al. [142] presented a methodology to determine the EnBL in a university classroom.

5. Conclusions

A literature review was conducted to present strategies and methods used to develop, model, and validate an EnMS in buildings focused on HEIs; based on this energy management, strategies to reduce energy consumption in HEIs are presented. Understanding that HEIs can be considered a city-scale model, the application of an EnMS can develop models that lead to energy sustainability in cities and contribute to SDGs 7 and 11, proposed by the UN.

The results obtained from this literature review show that the ISO 50001 standard has had little impact on the implementation of EnMS in higher education institution buildings. Few studies consider it as a basis for implementing an EnMS. Only 3% of the articles reviewed mention the standard as a reference, 7% refer to significant energy use, and 15% propose energy performance indicators for buildings.

This research concludes that 46% of papers perform measurements, estimations and forecasting, 38% carry out an EnBL, 21% apply an EPI, 26% perform an energy review, and 23% of the studies are related to the support related to energy management systems. This means that most of the energy management strategies are under the ISO 50001 standard but do not follow its methodology. Also, it is of note that most of the energy management systems and strategies for buildings are focused on the operation; meanwhile, maintenance is not of major relevance for the stakeholders.

It was also found that a great part of the studies focused on energy management in air-conditioning systems, and very few studies conducted have focused on barriers to the implementation of energy management systems; many articles propose strategies, and some studies quantify the reduction in energy consumption by implementing an EnMS.

Chapter 7 (support) of ISO 50001 is related to the design, communications and activities necessary to improve and maintain energy performance, and, in this study, it was found that 23% of the studies are related to the activities of this chapter.

Most articles related to EnMS do not analyze the impact of a EnMS on the carbon footprint in HEIs because these articles focus on energy use, energy efficiency and the reduction in energy consumption but not on the reduction in CO₂ emissions, as this broadens their focus; so, only 4% of the articles reviewed include this analysis.

It is important to update the EnMS, through IoT, and integrate digital technologies such as smart metering, real-time monitoring, IoT, blockchain, smart control, smart grids, energy management analysis software, communication networks, cloud computing, AI, and smart grids. This integration enables a more efficient and responsive EnMS, allowing for timely decision-making to optimize energy performance in higher education institution buildings. By leveraging digitalization and real-time data analysis, institutions can improve energy efficiency and sustainability.

Finally, as future work, considering that ISO 50001 has had a great impact on the energy management of organizations and has been little applied in higher education buildings, the authors consider that future work should include the implementation of a management system in higher education buildings considering all the requirements of ISO 50001 and evaluate its impact on energy performance.