Intelligent energy management systems for buildings in smart cities using PSO-driven optimization and penalty function constraints

Abstract

The growing need for energy conservation and sustainable development in smart urban areas demands the implementation of sophisticated Energy Management Systems (EMS) in modern buildings. This study presents a design for an intelligent EMS that merges Particle Swarm Optimization (PSO) with constraint-driven penalty functions to dynamically optimize energy consumption while maintaining both operational effectiveness and occupant comfort. The system efficiently oversees diverse energy consuming subsystems, including lighting, HVAC, and ventilation, by utilizing real-time data from environmental conditions and human presence, using a distributed wireless sensor network. The hardware setup features smart sensors and controllable nodes that monitor occupancy, temperature, ambient light, and air quality, enabling the coordinated management of energy-demanding systems. Central to the software architecture is a PSO-based control algorithm, augmented with penalty functions to impose crucial constraints, such as thermal comfort, minimum light levels, and ventilation standards. This hybrid optimization structure ensures that every control decision supports both energy conservation goals and human-centered needs. The novelty of the proposed EMS lies in its low-cost deployment, simplified integration with existing infrastructures, and enhanced interoperability across building systems, addressing key limitations of current solutions such as high complexity, inflexible architectures, and poor scalability. Experimental assessments in a simulated smart building setting reveal the system’s adaptability, reliability, and prompt responses in real-time. The EMS successfully detected varying occupancy and environmental patterns, finely tuning energy allocation across subsystems. Numerical analyses indicate up to 30% overall energy savings, confirming the efficacy of the PSO-penalty optimization method. The proposed EMS provides a scalable, interoperable, and cost-effective solution, making it apt for integration into current infrastructures and supportive of future-ready smart city environments.

Keywords

particle swarm optimization intelligent building management energy saving penalty function building automation optimization algorithm

Introduction

Concerns about resource depletion and climate change are making energy saving and environmental protection more and more crucial.¹ Intelligent building management systems are a significant source of energy consumption, so energy-saving measures must focus on improving them.² Conventional frameworks frequently use excessive energy because they cannot adapt to changing user behavior and environmental circumstances in real time.³ Their hindrance to optimizing energy according to actual demands arises from their lack of adaptive technology. Thus, developing intelligent solutions that can react instantly to environmental cues and user presence is imperative to minimize wasteful energy use and further sustainability objectives.

Because intelligent building systems use cutting-edge technology that enables dynamic changes, they present a viable solution to the problem of energy inefficiency. They can adjust various building parameters according to several parameters, including occupancy, ambient light levels, and user demands. By putting such technologies in place, energy usage and expenses can significantly decrease while preserving or increasing consumer comfort and satisfaction.^4,5 By ensuring that appliances and energy-hungry devices are only given when and where it is needed, this adaptive strategy maximizes the overall efficacy and efficiency of building management systems. Because they provide ideal settings based on user preferences and ambient factors, such intelligent systems enhance user experience and conserve energy.

Lighting accounts for a substantial portion of energy consumption in buildings, particularly in commercial and industrial settings.⁶ Efficient energy management systems (EMS) are crucial for reducing energy costs and environmental impact.⁷ Existing solutions, such as manual control, often fail to adapt to changing environmental conditions and user needs, leading to energy waste.⁸ Because of their inability to be adjusted, they may run longer or harder than necessary, wasting a significant amount of energy. As a result, there is a pressing need for increasingly sophisticated, automated building management systems that can react quickly to real-time changes and maximize energy efficiency while enhancing sustainability.

Intelligent building management systems have emerged as a promising solution to address these challenges. These systems leverage advanced algorithms and sensor networks to adjust various building parameters based on occupancy, daylight availability, and user preferences.⁹ Intelligent building management systems can significantly increase energy efficiency by continuously monitoring these elements and providing the ideal solution at any given time. In addition to lowering expenses and energy consumption, this real-time adaptability guarantees these settings are customized to the individual requirements and comfort of the occupants.

One such intelligent optimization strategy for energy management in smart buildings is based on the control algorithms that use Particle Swarm Optimization (PSO). This algorithm has excellent potential in optimizing energy consumption while maintaining user comfort and visual quality.^10,11 PSO is a nature-inspired optimization technique miming the collective behavior of bird flocks or fish schools. By incorporating a penalty function, the algorithm can enhance energy savings by penalizing inefficient lighting conditions and guiding the system towards optimal solutions.

The proposed system in this paper combines the PSO algorithm with a penalty function to create an intelligent energy-saving system in smart buildings. The system utilizes a wireless sensor network to gather environmental data and track user movements, enabling dynamic control and energy optimization.

This study aims to create and assess an intelligent energy-saving system that maximizes building automation and control by utilizing penalty functions and the PSO algorithm. More particular goals of the proposed work are:

• Develop a hardware architecture that seamlessly combines various nodes in smart buildings and wireless sensor networks without interfering with the current frameworks.

• Provide a software architecture with a robust energy control algorithm that can be adjusted in real-time according to user behavior and environmental factors.

• Deploy the PSO algorithm into practice while combining penalty functions for the best possible energy efficiency.

• Assess the system’s efficiency in lowering energy usage compared to conventional lighting systems and validate its performance through experimental setups.

The rest of the article is organized as follows: Section 2 summarizes the related works in the chosen field of study. Section 3 elaborates on the intelligent energy management system’s design and architecture. Section 4 focuses on methodological aspects of the energy-saving system, and it provides the experimental platform for the study, where the obtained results and discussions were summarized in Section 5. Finally, Section 6 concludes the article with this research’s key findings and outcomes.

Related works

Overview of intelligent Energy Management Systems (EMS)

Intelligent energy systems are a significant development in energy management and building automation. These systems use cutting-edge technologies, sophisticated algorithms, sensor networks, and automated controls to maximize the efficiency.¹² Intelligent EMS aims to minimize energy usage, improve user experience, and minimize environmental impact by delivering the appropriate amount of energy where and when needed. Sensors, sensor networks, lighting control devices, and control algorithms are essential to these systems.¹³ To create real-time energy usage modifications, control algorithms process real-time data from sensors that constantly monitor the surrounding environment and the presence of humans.

Intelligent EMS for smart buildings function by seamlessly combining hardware and software components. Sensors can detect changes in the environment, such as the presence or absence of people, variations in the amount of natural settings, and movements inside a space.¹⁴ The control algorithms process this data and then identify the best-fit configurations to satisfy the present needs. For example, the system can automatically dim or switch off the lights to save energy if a room gets empty. In contrast, the system can boost the illumination level when natural sunshine declines. Because of its flexibility, the illumination is always ideal for the occupants, and no energy is wasted.¹⁵

Intelligent EMS have many advantages, such as increased comfort, lower energy consumption, financial savings, and less adverse environmental effects. These systems help minimize carbon emissions and electricity costs by modifying lighting depending on real-time data. Additionally, they ensure the appliance usage suits the tasks, improving occupant comfort and productivity.¹⁶ Technological developments are lowering the initial costs and making these systems more accessible, despite their integration hurdles. Future developments will drive the transition to more intelligent and sustainable living and working environments. These developments include deeper interaction with other building management systems and improved predictive capacities through artificial intelligence (AI) and machine learning (ML).

Particle Swarm Optimization (PSO) in energy management

The social interactions of bird flocks and fish swarms inspired PSO. Its widespread use in energy management is due to its efficiency in resolving optimization issues. PSO is beneficial when the goal is to find the best possible solution within a complex, multi-dimensional search space. It can optimize several parameters in energy management to improve energy efficiency, cut expenses, and reduce consumption.¹⁷

For PSO to function, a population of possible solutions to the optimization issue needs to be initialized. Every particle modifies its location inside the search space according to its experiences and those of other particles. Particle velocity is related to their mobility and is dynamically changed based on social and cognitive factors.¹⁸ The social component represents the best-known position of the entire swarm, while the cognitive component represents the most prominent position of each particle. Particles move toward the best answer through iterative modifications.

PSO can be used in energy management to optimize resource distribution, energy storage system configuration, and device scheduling that consumes energy.¹⁹ For instance, PSO can optimize battery charging and discharging cycles in an intelligent grid to guarantee effective energy utilization. PSO can dynamically modify heating, ventilation, and air conditioning (HVAC) systems, and lighting systems in building automation to maximize energy efficiency and comfort levels.²⁰

PSO has been used with success in several energy management situations. PSO helps maximize the grid integration of wind turbines and solar panels in renewable energy systems. PSO ensures the maximization of energy output and the minimization of waste through the strategic placement and combination of these resources.²¹ They help to optimize machine schedules in industrial settings to lower peak demand fees and overall energy expenditures. There are numerous advantages to employing PSO in energy management. It offers a versatile and robust optimization framework that manages various goals and constraints. Second, because PSO is computationally efficient, it can be used in real- time scenarios where prompt decision-making is essential. Third, PSO improves the overall sustainability of energy systems by reducing costs and energy consumption by identifying practical solutions.²²

However, there are drawbacks to energy management when using PSO. One significant challenge is the possibility of becoming stuck in local optima, particularly in more intricate and non-linear search environments.²³ To solve this problem, more sophisticated PSO variations have been created, including hybrid and adaptive PSO. The scalability of PSO when handling large-scale situations presents another difficulty. Solutions for distributed and parallel computing are being investigated to improve the scalability of PSO.

There is great potential for integrating PSO with other cutting-edge technologies like ML and AI. These integrations may improve the EMS’ capacity for prediction and flexibility.²³ Furthermore, more precise and dynamic PSO optimization will be possible thanks to the growing availability of real-time data from smart meters and other Internet of Things (IoT) devices. PSO will play an increasingly important role in ensuring effective and sustainable energy management as energy systems become more complex and interconnected.

Penalty function methodologies

In optimization, penalty function approaches are crucial, particularly for solving limited optimization problems. These approaches address restrictions by converting a constrained optimization problem into an unconstrained one.²⁴ These methods optimize the intended aim while guaranteeing that limitations are met by adding penalty terms to the objective function.

Adding a penalty term to the objective function for each time a constraint is broken is the fundamental idea behind penalty function techniques. When requirements are not met, this penalty term raises the value of the objective function, discouraging solutions that do not adhere to the constraints. The overall objective function can be written using a penalty technique, as shown in equation (1).

F (x) = f (x) + P (x)

(1)

where f (x) is the original objective function, and P (x) is the penalty function. The penalty function P (x) is typically a function of the degree to which constraints are violated. For instance, if g (x) ≤ 0 represents a constraint, the penalty function can be formulated as

P (x) = \sum_{i = 0}^{n} \max (0, g_{i} {(x)}^{2})

which penalizes any positive violation of constraints. The two main categories of penalty functions are interior and external penalty functions.

The interior penalty functions are applied to problems with bounded viable regions. The solution is discouraged from approaching the boundary of the feasible region by the penalty applied within it. The logarithmic barrier function is an example of an interior punishment function, with the penalty term having the form shown in equation (2).

- \sum_{i = 0}^{n} \ln (- g_{i} (x))

(2)

The exterior penalty functions are used outside of the practical area. Constraint violations result in severe penalties that force the solution back into the usable zone. Common examples of exterior punishment functions include quadratic penalty functions.

\sum_{i = 0}^{n} \max (0, g_{i} {(x)}^{2})

(3)

Approaches based on penalty functions are frequently employed in engineering design, economics, and operational research, among other optimization challenges. Energy management also employs penalty functions to ensure solutions adhere to significant limitations, including energy demand, supply constraints, and regulatory requirements.²⁵

For example, complying with safety rules, restricting energy usage to avoid peak demand charges, and maintaining a specific degree of indoor comfort are restrictions in an energy management system for a building. The optimization algorithm can efficiently balance these limitations while reducing energy expenditures using penalty functions.

Penalty function approaches have the main benefit of being straightforward and simple to use. They allow limited issues to be solved using conventional optimization methods without requiring specialist algorithms. Additionally, by adjusting the penalty terms appropriately, they offer flexibility in managing different limitations.

Nevertheless, penalty function approaches have drawbacks. One major issue is the sensitivity to the selection of penalty parameters. Inappropriately selected penalty parameters could cause the optimization process to fail to identify a workable solution or converge slowly. Penalty functions can also significantly complicate optimization by making the objective function substantially non-linear and non-smooth.

Existing solutions and gaps

Numerous methods have been created and implemented to improve the energy efficiency of building EMS. These options range from basic manual controls to sophisticated automated systems.²⁶ Although each has advantages, significant gaps prevent energy optimization and savings from reaching their full potential.

Manual control has always been the simplest method of energy management for buildings. This entails people physically turning on and off the appliances or modifying their parameters to suit their needs. Despite being straightforward and affordable, this approach relies heavily on user behavior. It frequently leads to appliances being left on when not in use, which wastes energy.

To increase productivity, basic automated systems included occupancy sensors and timers. Timers reduce the need for human involvement by turning lights on and off at preset periods. Make sure that appliances are only on when spaces are inhabited by using occupancy sensors to detect people’s presence accordingly.⁶ These systems are nevertheless somewhat rigid, even if they provide advantages over human controls. Timers don’t consider occupancy patterns, and simple occupancy sensors can’t distinguish between different occupancy and activity levels, resulting in less-than-ideal energy consumption.

Advanced appliances and control systems use more complex technology, such as networked sensors, daylight harvesting, and programmable logic controllers (PLCs).²⁷ These systems can dynamically adjust system parameters based on data collected in real-time from various sensors, such as motion detectors and ambient light sensors. For example, daylight harvesting systems minimize energy use and maintain a consistent lighting environment by adjusting artificial lighting levels based on the amount of natural light available.

Intelligent EMS represent the next step forward in energy conservation. These systems use ML, AI, and the IoT to adjust their settings automatically.⁸ These systems anticipate and proactively modify the parameters based on user behavior and ambient factors. AI-driven systems, for instance, can use past data analysis to predict occupancy trends and make proactive lighting adjustments, assuring ideal conditions while consuming the least energy.

Although EMS for smart buildings promise significant energy savings and improved user experience, they also have drawbacks. Implementation can come with a substantial upfront cost, and integrating it with the current building infrastructure can be challenging. Additionally, as these systems depend on ongoing data collection and analysis, data security and privacy issues must be addressed.

The main shortcomings of the current systems are their high costs, complexity, and problems with integration and interoperability. To address these research gaps, future research and development should concentrate on lowering the cost of sophisticated EMS through advancements in hardware and software. If the installation and maintenance procedures are simplified, these technologies will be more accessible.

System design and architecture

This section provides an in-depth overview of the system design, focusing on its hardware and software architectures, as well as integration with existing infrastructure. Figure 1 shows the proposed architecture for an intelligent energy management system in smart buildings. It encompasses the following hardware and software modules integrated into the framework.

Figure 1.

Illustration of overall architecture of the PSO-based intelligent energy management system for smart buildings.

Hardware modules

The hardware architecture serves as the essential framework for an intelligent energy management system in smart buildings, equipping it with the requisite components to gather environmental data, process that information, and implement appropriate adjustments. The primary elements include a wireless sensor network, intelligent appliances, and a central control unit (CCU).

Wireless sensor network

The hardware design prominently features a WSN, incorporating a range of sensors, including ambient light, occupancy, and motion detectors. The deployment of these sensors is executed with precision to achieve extensive coverage across the building, facilitating the acquisition of real-time data regarding lighting conditions and occupancy levels.

Ambient light sensors assess the intensity of natural light in the environment, enabling the system to diminish artificial lighting when adequate daylight is available. Occupancy sensors are designed to identify the presence of individuals within designated areas, thereby ensuring that lighting is activated solely when these spaces are occupied. Motion detectors monitor movement within the building, enhancing the granularity of occupancy detection and facilitating appropriate adjustments to lighting systems. The sensor network’s wireless architecture facilitates a high degree of flexibility in sensor positioning, eliminating the necessity for intricate cabling systems.

Smart building and other associated nodes

The energy consumption nodes represent advanced fixtures that can modulate their energy usage and engage or disengage in response to instantaneous directives from the central control unit. Each node is integrated with relays, dimmers, and smart switches, facilitating precise regulation of power to the appliances.

Integrating sensor data with optimization algorithms enables the system to dynamically modulate energy levels across various zones and appliances within the building. The appliances could be the conventional electric appliances, smart lighting systems, HVAC systems and other building automation systems. Using the PSO-enabled approach enhances energy efficiency and prioritizes user comfort. The system effectively reduces reliance on robust means of energy usage in scenarios with elevated demand levels, leading to significant energy savings.

Central Control Unit (CCU)

The Central Control Unit (CCU) is at the system’s core, responsible for processing all incoming sensor data and dispatching commands to the associated nodes in the smart buildings. The CCU incorporates a microcontroller or microprocessor to oversee the system’s operations, including implementing control algorithms for regulating lighting adjustments. The PSO-based control algorithms and other functional modules are integrated in the CCU.

The implementation of wireless communication between sensors and appliances, as well as nodes, facilitates efficient system operation, eliminating the need for extensive rewiring. This makes it particularly suitable for retrofitting within existing structures, as it directly interfaces through low-power wireless communication protocols from the CCU module.

Software architecture

The system’s software architecture facilitates its adaptability, scalability, and efficiency. It is engineered to handle sensor data in real time, reduce the energy usage levels through sophisticated algorithms, and offer an intuitive interface for control and monitoring purposes.

Building energy consumption control algorithm

The PSO and penalty function-based energy consumption control algorithm are fundamental to the system’s software architecture. It determines the optimal configuration, aiming to balance energy efficiency, user comfort, and compliance with regulatory standards. The algorithm incorporates various essential factors, including optimal levels, occupancy and movement patterns, user preferences, and the minimum energy required levels.

The PSO algorithm systematically modifies the nodes associated with the smart buildings, optimizing the power conditions to reduce energy consumption while preserving user comfort. The PSO technique demonstrates significant efficacy in addressing intricate, multi-dimensional optimization challenges, such as those encountered in intelligent energy consumption control systems. In this context, the fitness function is formulated to minimize energy consumption, considering constraints associated with the comfort and compliance with regulatory standards. The fitness function is defined as shown in equation (4):

f (x) = E (x) + P (x)

(4)

Where:

• E(x) represents the energy consumption of the system,

• P (x) is the penalty function that enforces constraints such as minimum power consumption levels and user comfort.

Penalty function

A penalty function guarantees the system adheres to all requisite constraints, such as the minimum power requirement. This function imposes penalties on solutions that contradict the established constraints, thereby guiding the PSO algorithm towards feasible solutions. The penalty function P (x) is incorporated into the fitness function to deter solutions that fail to meet the established constraints. The penalty function can be defined as shown in equation (5):

P (x) = \sum_{i = 1}^{n} λ_{i} \max (0, g_{i} (x))

(5)

Where:

• g_i (x) represents the constraint function for each node in the smart buildings i,

• n is the number of nodes,

• λ_i is the penalty coefficient for the i-th constraint,

• max (0, g_i (x)) ensures that the penalty only applies if the constraint is violated (i.e., if g_i (x) exceeds 0).

For example, when the minimum power level mandated for a specific area is not met, the penalty function increases the total cost (fitness function value) associated with that solution. This mechanism prompts the PSO algorithm to explore alternative solutions that comply with the lighting criteria.

Fitness function with constraints

The final fitness function that the PSO algorithm optimizes can be expressed as shown in equation (6):

F (x) = E (x) + \sum_{i = 1}^{n} λ_{i} \max (0, g_{i} (x))

(6)

Where:

• E (x) minimizes energy consumption

• g_i (x) enforces user comfort and minimum power constraints,

• λ_i is chosen based on the importance of each constraint.

The PSO algorithm operates through an iterative process, with each particle serving as a distinct candidate solution that encapsulates a particular arrangement of power intensities distributed across various zones. The algorithm refines these particles through successive iterations to identify the configuration that optimizes energy consumption while adhering to all specified constraints. The penalty function is critical in guiding the swarm towards feasible and optimal solutions by imposing penalties on those that infringe upon user comfort or regulatory restrictions.

This framework enhances the system’s capacity for dynamic, real-time adjustments, optimizing energy savings and ensuring high user satisfaction and adherence to standards, particularly in large-scale implementations. This methodology guarantees that the intelligent energy management system enhances its performance while adhering to all operational constraints.

Building Management Systems (BMS)

The software architecture incorporates an integration module to facilitate seamless interfacing of the intelligent building energy management system with various Building Management Systems (BMS), including HVAC and security systems. Implementing standard protocols such as BACnet, Modbus, or KNX enables the system to adapt to changes in other systems, exemplified by adjusting lighting in correlation with HVAC settings or security notifications. The lighting system’s interoperability is crucial for integrating within a comprehensive building automation framework.

Instructing building occupants and maintenance personnel on how to operate and maintain the system is a critical element of successful integration. Comprehensive user manuals, training programs, and support services are provided to ensure effective implementation and sustained operation.²⁸

Methodology

The methodology employed in this study encompasses a comprehensive strategy for data collection, utilizing particle swarm optimization (PSO) for optimization, implementing penalty functions for effective constraint management, and system calibration to ensure optimal performance. The methodology was meticulously crafted to guarantee real- time adaptability, operational efficiency, and scalability within the intelligent building energy-saving system.

Data collection

Establishing an energy-optimized building automation framework depends on the precision and thoroughness of data collection processes. The system integrates multiple sensor types, such as ambient light sensors, occupancy detectors, and motion detectors. The placement of these sensors was executed with precision across the experimental setup to facilitate continuous monitoring and the transmission of real-time data.

Robust sensors played a crucial role in quantifying the influx of natural parameters of the appliances into the environment, enabling the system to modulate power usage in response. Occupancy sensors detected the presence of individuals across various areas, ensuring that the systems were activated exclusively in occupied spaces. Motion detectors have enhanced occupancy detection by monitoring movement. This yields comprehensive insights into user activity levels, enabling the system to adjust lighting conditions in real-time.

To maintain effective power management conditions, optimization algorithms continuously process the sensor data at a central controller. Data collection was crucial in facilitating real-time adjustments, reducing energy usage, and ensuring that user comfort was maintained under diverse environmental conditions.

Implementation of PSO algorithm

The optimization process of framing an intelligent automation of energy management in smart buildings fundamentally relies on the PSO algorithm. It demonstrates a significant advantage in navigating multidimensional search spaces, effectively optimizing energy consumption through the dynamic adjustment of various appliance parameters informed by real-time sensor data. In the context of the PSO algorithm, each particle signifies a potential solution, specifically a configuration of the energy intensities distributed across various zones within the building.

The fitness function was formulated to reduce energy consumption while maintaining optimal comfort. The analysis considered various factors, including the power usage of different appliances, the presence of people in different areas, and the necessary usage of applications for specific activities, which are critical factors to consider. The particles modified their positions through an iterative process, guided by their optimal solutions and the swarm’s collective best- known solutions, achieving convergence towards the most energy-efficient configurations of the power consumption nodes.

Application of penalty functions

The application of penalty functions was used to address constraints related to minimum power levels, user comfort, and regulatory requirements. These penalty functions facilitated the operation of the PSO algorithm within the established parameters by imposing penalties on solutions that contravened the constraints.

The penalty function P (x) has been incorporated into the fitness function to address any violations of constraints, including inadequate lighting or excessive energy consumption. This approach guaranteed that the optimization process complied with the required constraints while simultaneously exploring the most effective lighting solutions.

The energy constraint was implemented to guarantee that the power usage levels in each designated zone remained above the established minimum threshold, particularly in locations designated for specific activities, such as reading or working.

System calibration and setup

The calibration process constituted an essential component of the methodology to guarantee the system’s efficient and accurate operation in real-time scenarios. The system underwent an initial calibration phase, during which baseline sensor readings were systematically collected across various conditions to determine default settings applicable to different scenarios.

The calibration process for sensors involved adjusting the sensitivity of environment and occupancy sensors to align with typical conditions and anticipated movement patterns. Particle size, inertia weight, cognitive and social coefficients, and other components of the PSO algorithm were carefully adjusted to achieve fast convergence and accurate optimization results.

After the calibration phase, the system was configured to function autonomously, modulating power management conditions in response to real-time data inputs. Establishing continuous maintenance involved regular recalibration and system evaluations to ensure peak performance.

Experimental environment

The experiments were conducted in an environment that closely represents the standard environmental settings, characterized by variable power requirements influenced by occupancy levels and the availability of electrical appliances in smart buildings. The experimental zone covered roughly 500 square meters, segmented into multiple rooms and open spaces, which facilitated various lighting conditions and occupancy behaviors in the smart building.

The environmental conditions were systematically monitored over several weeks to encompass both daytime and nighttime scenarios, ensuring that the system was rigorously tested across various power load conditions, including sunny, cloudy, and low-light environments. The experimental setup included windows and glass doors, which facilitated power usage and were essential for evaluating the system’s capacity to adapt to fluctuating conditions.

Figure 2 illustrates the spatial configuration of the experimental environment, depicting the layout of Room 1, Room 2, and a central Open Hall. Each area is annotated with its physical dimensions in meters, collectively accounting for approximately 500 m². The positions of sensor nodes (S-) and lighting nodes (L-) are marked to demonstrate full spatial coverage for adaptive control and monitoring.

Figure 2.

Schematic floorplan of experimental setup (500 m²) with sensor and lighting node placement.

Sensor and power consuming nodes arrangement

The configuration included installing numerous wireless sensor nodes alongside intelligent building nodes throughout the experimental setting. The sensors and smart building power-consuming nodes were placed precisely to guarantee thorough coverage of the test area.

Optimal sensor positioning is crucial for accurate data collection and analysis. Sensors were strategically placed near windows, doors, and associated appliances in the buildings to quantify power usage in the environment. Occupancy sensors were strategically deployed in high-traffic zones, including entrances, hallways, and communal areas, to ensure precise detection of space utilization in the smart buildings. Motion detectors were strategically installed in smaller rooms and office spaces to capture detailed movement data, facilitating more precise adjustments to load conditions.

Smart nodes have been systematically installed in all rooms and common areas. The nodes have demonstrated the ability to adjust their power consumption levels dynamically based on sensor data inputs and demands. The nodes established a wireless connection to the central controller, facilitating real-time communication and allowing for seamless adjustments to power usage conditions across the building.

More specifically, the way the sensors and associated nodes were set up made it easy to monitor every part of the experimental space. This meant that the energy usage could be changed on the suit the needs of each area’s use and appliances in the building.

Results and discussion

This section analyzes and compares the proposed intelligent energy management system’s performance in smart buildings across various important parameters with existing systems. The outcomes are assessed by analyzing energy conservation, system dependability, and economic efficiency. Additionally, comparisons are made with conventional energy management systems to highlight the benefits of the suggested approach.

Performance metrics

The assessment of the intelligent energy management system is presented in four principal performance metrics: energy savings, user comfort Score, cost savings, and response time. The energy savings analysis for 100 nodes, as presented in Table 1, indicates that the proposed system realizes a 33% reduction in energy consumption, in contrast to the highest-performing alternative, which achieves merely 27% savings. The observed trend becomes increasingly evident with the implementation of 200 nodes (Table 2), indicating that the proposed system achieves a 35% savings, in contrast to the 29% savings realized by the nearest alternative. The analysis of user comfort scores reveals a consistent trend in which the proposed system achieves a superior comfort score, reaching up to 89 for 200 nodes, compared to existing studies that peak at 85.

Table 1.

Performance analysis of the intelligent energy management in smart buildings under various scenarios with 100 nodes.

Scenario	Metric	Proposed (100 nodes)	Parvin et al.²⁹ (100 nodes)	Qu et al.³⁰ (100 nodes)	Wu et al.³¹ (100 nodes)
Occupancy adjustment	Energy savings (%)	32	26	29	21
	User comfort score	86	81	83	72
	Cost savings (%)	26	21	22	17.5
	Response time (s)	0.75	1.3	1.05	1.35
Dynamic environment	Energy savings (%)	31	25	28	21
	User comfort score	85	80	82	71
	Cost savings (%)	26	20.5	21.5	17.5
	Response time (s)	0.8	1.35	1.1	1.4
Peak energy load	Energy savings (%)	32	26	29	22
	User comfort score	86	82	84	72
	Cost savings (%)	26.5	22.5	23.5	18
	Response time (s)	0.7	1.2	1.0	1.3

Table 2.

Performance analysis of the intelligent energy management in smart buildings under various scenarios with 200 nodes.

Scenario	Metric	Proposed (200 Nodes)	Parvin et al.²⁹ (200 nodes)	Qu et al.³⁰ (200 nodes)	Wu et al.³¹ (200 nodes)
Occupancy adjustment	Energy savings (%)	34	28	32	25
	User comfort score	88	83	85	74
	Cost savings (%)	28	23	24	19
	Response time (s)	0.65	1.2	0.95	1.3
Dynamic environment	Energy savings (%)	33	27	31	24
	User comfort score	87	82	84	73
	Cost savings (%)	27	22	23	18.5
	Response time (s)	0.7	1.3	1.0	1.4
Peak energy load	Energy savings (%)	34	28	32	25
	User comfort score	88	84	85	74
	Cost savings (%)	28.5	24	25	19.5
	Response time (s)	0.6	1.1	0.9	1.2

Figure 3 illustrates that energy savings exhibit a nearly linear increase about the number of nodes, underscoring the proposed system’s scalability. Similarly, Figure 3 illustrates that the proposed system consistently attains superior user comfort scores with the addition of more nodes, thereby enhancing user satisfaction while maintaining energy efficiency.

Figure 3.

Comparative analysis on energy savings vs number of nodes.

Energy Savings analysis

The evaluation of an intelligent energy management system in smart buildings heavily relies on quantifying energy savings. The results presented in Table 1 and Table 2 indicate that the proposed system exhibits enhanced energy efficiency across all scenarios and varying numbers of nodes. The proposed system demonstrates a 33% energy savings at 100 nodes within energy efficiency scenarios, in contrast to other systems that exhibit savings ranging from 22% to 30%. At 200 nodes, the observed savings escalate to 35%, whereas alternative systems remain comparatively lower, reaching a maximum of 29%.

Figure 3 effectively illustrates this concept, demonstrating a consistent enhancement in energy savings correlated with the increasing number of nodes. The findings indicate that the proposed system effectively performs at smaller scales and exhibits efficient scalability in larger environments, resulting in increased energy savings by deploying additional nodes.

System reliability and accuracy

The reliability of the system and its response time are essential factors in ensuring the effectiveness and efficiency of an intelligent energy management system. The proposed system demonstrates a response time of 0.7 seconds at 100 nodes, in contrast to the 1.0 to 1.4 seconds observed in alternative systems. The response time further improves to 0.6 seconds with the addition of 200 nodes, indicating the system’s robustness and capacity to manage larger-scale deployments effectively.

The user comfort score illustrated in Figure 4 demonstrates the system’s swift responsiveness without com promising user comfort, as the proposed system consistently achieves the highest scores across all node counts.

Figure 4.

User comfort score vs number of nodes assessment for the proposed framework.

Figure 5 presents the response times across all systems about the increasing number of nodes. The proposed system consistently demonstrates a performance advantage over other works, suggesting enhanced reliability and responsiveness.

Figure 5.

Illustration on the observation of response time vs number of nodes for the proposed framework.

Comparison with traditional systems

Conventional EMS have limitations in their capacity for real-time adaptability compared to intelligent systems, often leading to increased energy consumption and suboptimal load conditions. Additionally, the cost savings metric highlights the economic advantages associated with the proposed system. The proposed system demonstrates a cost savings of 27% at 100 nodes, in contrast to the 18% to 23% savings reported in other studies. At 200 nodes, the observed savings escalate to 29%, in contrast to 24% for alternative systems, thereby illustrating the sustained economic advantages of implementing intelligent energy management solutions as opposed to conventional systems for smart buildings.

The quantitative information obtained from the performance metrics provides significant insights regarding the scalability and efficacy of the proposed system. The data presented in Table 1 indicates that, with 100 nodes, the proposed system exhibits superior performance across all evaluated metrics. Specifically, it achieves 33% energy savings, a user comfort score of 87, 27% cost savings, and a response time of 0.7 seconds. The observed trend becomes more pronounced at 200 nodes (refer to Table 2), with energy savings reaching 35% and a corresponding enhancement in response time to 0.6 seconds.

The data provides additional support for these conclusions. Figure 6 illustrates a positive correlation between cost savings and node count, demonstrating that the proposed system consistently exceeds the performance of alternative systems. Figure 5 illustrates the efficiency of the proposed system’s response time, showing a decrease as the node count increases, which highlights its scalability and reliability.

Figure 6.

Comparative analysis on the cost savings vs number of nodes.

As detailed in Figure 7, the proposed intelligent energy management system demonstrates that even with reduced node counts, it offers significant energy savings and enhanced user comfort compared to conventional systems. Conventional systems generally lack the capability for dynamic response, in contrast to the proposed system, which demonstrates adaptability in real-time to variations in occupancy and ambient conditions. This results in a reduction in energy usage and an improvement in comfort levels.

Figure 7.

Comparison of performance metrics between proposed system and other works.

Table 3 presents a comparative analysis of the proposed system against recent literature,²⁹,³⁰ and³¹ based on key dimensions, including implementation cost, along with scalability and interoperability. The proposed system offers significant benefits, including low-cost deployment ($4.80 per square meter), minimal hardware intervention, and support for standard wireless protocols, facilitating easy retrofitting into legacy systems.

Table 3.

Comparison of cost factors of the proposed system with related works.

System	Cost (USD/m²)	Integration effort	Interoperability	Scalability
Parvin et al.²⁹	$8.30	Medium	Zigbee only	Medium
Qu et al.³⁰	$6.75	High	Custom protocol	Low
Wu et al.³¹	$7.20	High	Wi-Fi only	Medium
Proposed	$4.80	Low	Wi-Fi, Zigbee, BMS-ready	High

Conclusion

This research paper outlines the creation and validation of an intelligent energy management system tailored for buildings. It uses a PSO algorithm enhanced with penalty function constraints to optimize the energy usage of critical subsystems. The system adjusts automatically to real-time changes in environmental conditions and human activities via a wireless sensor network, facilitating coordinated control over energy-demanding components like lighting and HVAC systems. By integrating penalty functions within the PSO framework, the system skillfully balances energy reduction with occupant comfort and adheres to operational constraints such as temperature limits, lighting requirements, and air quality standards. Experimental results showed that the system operates both reliably and adaptively under actual building conditions, achieving energy savings of up to 30% compared to standard energy management practices. Moreover, the implementation of wireless and modular hardware allows for easy integration into current infrastructure, thereby minimizing deployment challenges and renovation expenses.

The suggested system exhibits exceptional scalability and adaptability, rendering it a feasible solution for both small- and large-scale implementations in contemporary smart buildings. Future endeavors will focus on enhancing the system’s functionalities by integrating additional control parameters, including collaboration with other BMS, and exploring the feasibility of advanced AI and ML methodologies to improve the system’s optimization efficiency further.

As part of future work, the system architecture will be extended to support BIM-compatible data exchange protocols, subject to industry foundation classes for seamless integration with building design and management platforms. This integration would enable real-time synchronization between live sensor-actuator data and digital building models, enhancing energy management systems’ scalability, interoperability, and multi-system coordination within smart building ecosystems.

ORCID iD

Jialing Li https://orcid.org/0000-0003-3315-843X

Footnotes

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Ehsanullah

Tran

Sadiq

, et al.

How energy insecurity leads to energy poverty? Do environmental considerations and climate change concerns matter?

Environ Sci Pollut Res Int 2021; 28(39): 55041–55052.

Sadeghian

Moradzadeh

Mohammadi-Ivatloo

, et al. A comprehensive review on energy saving options and saving potential in low voltage electricity distribution networks: building and public lighting. Sustain Cities Soc 2021; 72: 103064.

Day

McIlvennie

Brackley

, et al. A review of select human-building interfaces and their relationship to human behavior, energy use and occupant comfort. Build Environ 2020; 178: 106920.

Aliero

Qureshi

Pasha

, et al. Smart home energy management systems in internet of things networks for green cities demands and services. Environmental Technology & Innovation 2021; 22: 101443.

Jagatheesaperumal

Bibri

Ganesan

, et al. Artificial intelligence for road quality assessment in smart cities: a machine learning approach to acoustic data analysis. Comput Urban Sci 2023; 3(1): 28.

Al-Ghaili

Kasim

Al-Hada

, et al. A review: buildings energy savings-lighting systems performance. IEEE Access 2020; 8: 76108–76119.

Petrinska

Ivanov

. Led retrofitting for fluorescent troffers with minimum enviromental impact–case study. In: 2024 Ninth Junior Conference on Lighting (Lighting). IEEE, 2024, pp. 1–3.

Putrada

Abdurohman

Perdana

, et al. Machine learning methods in smart lighting toward achieving user comfort: a survey. IEEE Access 2022; 10: 45137–45178.

Gou

Zhang

Calixto

, et al. Reliable data collection model and transmission framework in large-scale wireless medical sensor networks. Comput Model Eng Sci 2024; 140(1): 1077–1102.

10.

Khairul

Noor Abdullah

Hassan

, et al. A new metric for optimal visual comfort and energy efficiency of building lighting system considering daylight using multi-objective particle swarm optimization. J Build Eng 2021; 43: 102525.

11.

Bibri

Huang

Jagatheesaperumal

, et al. The synergistic interplay of artificial intelligence and digital twin in environmentally planning sustainable smart cities: a comprehensive systematic review. Environ Sci Ecotechnol 2024; 20: 100433.

12.

Kayode Olajiga

Chigozie Ani

Sikhakane

, et al. A comprehensive review of energy-efficient lighting technologies and trends. Eng sci technol j 2024; 5(3): 1097–1111.

13.

Chiesa

Di Vita

Ghadirzadeh

, et al. A fuzzy-logic IoT lighting and shading control system for smart buildings. Autom ConStruct 2020; 120: 103397.

14.

Chong

Ismail

, et al. A triz approach for designing a smart lighting and control system for classrooms based on counter application with dual pir sensors. Sensors 2024; 24(4): 1177.

15.

Chen

Guo

Liu

, et al. Therapeutic lighting in the elderly living spaces via a daylight and artificial lighting integrated scheme. Energy Build 2023; 285: 112886.

16.

Sadeghian

Mohammadi-Ivatloo

Oshnoei

, et al. Unveiling the potential of renewable energy and battery utilization in real-world public lighting systems: a review. Renew Sustain Energy Rev 2024; 192: 114241.

17.

T-H

Kumar Jagatheesaperumal

Nguyen

M-D

, et al. Applications of generative Ai (Gai) for mobile and wireless networking: a survey. IEEE Internet Things J 2024; 12: 1266–1290.

18.

Khan

Ahmad

Kumar Jagatheesaperumal

, et al. Social media informatics for sustainable cities and societies: an overview of the applications, associated challenges, and potential solutions. arXiv preprint arXiv:2412.03600, 2024.

19.

Kumar

Karthikeyan

. A multi-objective optimization solution for distributed generation energy management in microgrids with hybrid energy sources and battery storage system. J Energy Storage 2024; 75: 109702.

20.

Wang

Fang

, et al. Modeling and optimization of a passive building hvac system based on improved pso and hj algorithm. Energy Rep 2024; 11: 2415–2438.

21.

Thirunavukkarasu

SawleLala

Lala

. A comprehensive review on optimization of hybrid renewable energy systems using various optimization techniques. Renew Sustain Energy Rev 2023; 176: 113192.

22.

Karthick Raja

Shermatova

Ansari

MAL

, et al. Optimization of neural networks using swarm intelligence techniques for achieving energy efficiency in smart building architecture. In: 2024 International Conference on Cognitive Robotics and Intelligent Systems (ICC-ROBINS). IEEE, 2024, pp. 14–21.

23.

Zaini

Sulaima

Razak

IAWA

, et al. A review on the applications of pso-based algorithm in demand side management: challenges and opportunities. IEEE Access 2023; 11: 53373–53400.

24.

Chen

Wang

, et al. Suitability comparison of heuristic algorithms for the execution of lighting design objectives (lidos) procedure. Build Environ 2023; 242: 110539.

25.

Liang

Qin

, et al. Assessment and optimization of tunnel lighting quality based on drivers’ visual comfort: from methodology to application. Tunn Undergr Space Technol 2024; 143: 105487.

26.

Van Thillo

Verbeke

Audenaert

. The potential of building automation and control systems to lower the energy demand in residential buildings: a review of their performance and influencing parameters. Renew Sustain Energy Rev 2022; 158: 112099.

27.

Wagiman

Abdullah

Hassan

, et al. Lighting system control techniques in commercial buildings: current trends and future directions. J Build Eng 2020; 31: 101342.

28.

Ganesan

Jagatheesaperumal

. Revolutionizing emergency response: the transformative power of smart wearables through blockchain, federated learning, and beyond 5g/6g services. IT Prof 2024; 25(6): 54–61.

29.

ParvinLipu

KMSH

Hannan

Abdullah

, et al. Intelligent controllers and optimization algorithms for building energy management towards achieving sustainable development: challenges and prospects. IEEE Access 2021; 9: 41577–41602.

30.

QuXuSun

JQQLKX

Q-L

Sun

K-X

. Optimization of indoor luminaire layout for general lighting scheme using improved particle swarm optimization. Energies 2022; 15(4): 1482.

31.

WuZhang

IlminSui

Ilmin

, et al. Residential energy-saving lighting based on bioinspired algorithms. Math Probl Eng 2022; 2022(1): 7600021–7600029.