A systematic review of recent developments in IoT-based demand side management for PV power generation

2.1 Energy conservation techniques

Figure 3 provides insights into energy conservation techniques. In a study focused on controlling non-flexible power demand in smart grids for academic buildings (Ghorbanian et al. 2021), an optimization technique employing Integer linear programming and a fuzzy controller was proposed. This method utilizes human-like reasoning to regulate non-flexible loads based on the convenience level of one appliance affecting others (Ghosh and Saha 2021). Maximizing non-flexible loads reduces overall power usage while maintaining user comfort. The intention is to reduce the equation for energy consumption in academic buildings.

where p _i and c _i stand for an appliance’s power consumption at any given moment and its level of convenience, respectively, in a set of appliances, i denotes a specific appliance, and s _i denotes the switch control for that appliance (Gomez-Quiles and Valdivieso-Sarabia 2019).

Figure 3

The tips for conserving energy (Li et al. 2020).

The potential for energy-saving techniques as a DSM approach in economically disadvantaged nations is explored as a financially viable and successful option (Guo et al. 2020). The research indicates a substantial 50.7% reduction in energy demand when consumer behavior, contributing to decreased energy consumption, is integrated as a demand response strategy. The challenge of demand-side management (DSM) in home area networks (HANs) is addressed, offering a solution based on "multi-agent deep reinforcement learning in real-time." This technique is designed to effectively navigate the complexities of DSM in HANs (Hammad et al. 2021).

The optimal regulation of temperature, energy supply, and price is the aim of the suggested energy management (EM) system. A thorough analysis is carried out to investigate the possible effects of price adjustments on the proposed algorithm. The DSM approach used in this study encourages a decrease in home energy use over a predetermined time frame. We use deep reinforcement learning to tackle this problem. The user’s decision-making dilemma is modeled by means of a “Markov decision process (MDP).” The goal is to choose the MDP’s best course of action, as shown by the following equation, in order to maximize the expected total reward, or E (Table 2).

To mitigate peak load demand and overall electricity costs across three primary client categories residential, commercial, and industrial, an available solution is the Back Tracking Search Algorithm (Hossain and Fotouhi 2020). The proposed method represents an enhancement over the differential evolution (DE) process, specifically addressing the day-ahead scheduling of load DSM problems. The authors introduce a technique for multi-objective optimization (MOO) in DSM problems (Huang and Liu 2019). The cost of energy for a single day is computed through a discrete time-changing model. Schedules with manageable loads are generated using a greedy search approach, considering loads with varying average power density and limited on-time windows (Hussain et al. 2021). When the proposed approach is implemented successfully, all loads are scheduled, which lowers the cost and peak-to-average power ratio (PAPR). The following equation may be used to formulate the MOO for DSM problem:

where (u) can be seen as a matrix reflecting a home EM system-controlled equipment schedule, and the goal functions. The term “f(u) minimization” denotes the simultaneous minimization of each vector element while using the metrics “f1” for the energy cost and “f2” for the PAPR (Jakkala et al. 2020).

A unique home DSM system is shown in Figure 4. It is based on a clever load control technique that is applied on a half-hourly basis and has access to a variety of power consumption limits (PCLs) (Jayaraman et al. 2021). Real-time loads, user preferences, and comfort are all taken into account in the design. Time-of-use (ToU) pricing is used to evaluate how user choices affect demand response. The results of a traditional load management strategy that depends on predetermined PCLs are contrasted with those of the proposed control technique. The following equation serves as a mathematical representation of the optimization problem:

Figure 4

The concept of load shaping (Li et al. 2021): (a) Peak clipping, (b) Valley filling, (c) load shifting, (d) flexible load shaping, (e) strategic growth, and (f) conservation.

For scheduling loads throughout the day, a DSM strategy based on the latent Dirichlet allocation method is provided. Using a cost function and the probability density function of consumer appliance usage patterns, this method shifts loads. The primary objective of demand-side management (DSM) is achieved through the implementation of load-shifting strategies aimed at reducing costs (Kaddoum et al. 2021). Test findings indicate that employing the load-shifting approach effectively reduces the cost of power consumption, and utilizing a MATLAB-created neural network further diminishes the mean square error. Another application of the DSM approach focuses on a distribution network in three phases for flexible load scheduling. Beyond other DSM objectives, the primary aim is to efficiently address operational limit breaches (Karim et al. 2020). The effectiveness of the recommended strategy is evaluated through a case study on market scheduling, illustrating the concept’s application to mitigate high and over voltages. Market planning for minimizing both low and high voltages is exemplified through a case study aimed at reducing costs and peak load demand (Khademi et al. 2021).

The following is the formulation of the optimization problem: when the appliance is in the on state, X i j t it is 1, and when it is off, it is 0. The genetic algorithm (GA) is used to solve this issue. The emphasis (Khedkar et al. 2021) is on reducing the maximum load in residential areas by utilizing the load-shifting DSM technique. Hourly monitoring is done on four main categories of loads: total load, cooling and heating, white appliances, and entertainment. Each of these loads has a particular pattern of energy use. Peak load is significantly reduced when the load-shifting DSM technique is used (Kim et al. 2021). Addressing the DSM challenge to lower peak energy consumption and electricity costs, the load-shifting approach is employed in conjunction with the MATLAB GUI. The suggested software allows users to compute their total expenses based on energy usage. This is achieved by scheduling household appliances for off-peak and peak times. Utilizing DSM modeling and the daily load-shifting approach, a minimization problem that accounts for user conveniences is formulated (Krishna Rao et al. 2023). The optimization process involves the use of the “Symbiotic Organisms Search (SOS) algorithm.” Simulation results demonstrate that, compared to various other techniques, the combination of DSM with the SOS algorithm achieves greater precision in reducing peak load. Equation (6) expresses the objective problem, while equation (7) describes the daily electricity bill (Kong et al. 2019).

The objective curves value at time t is known as T of t.

District heating systems undergo DSM through substation regulatory plan control or load shifting. A substation controlling technique called “Differential of Return Temperatures (DRT)” is used to reduce thermal peaks without sacrificing interior comfort levels. Peak load reduction, entails load transfer to decrease peak times, flatten the load curve, and minimize the requirement for peak load. Research on the impacts of the Integrated Electricity and Gas System (IEGS), employing DSM measurements was conducted by Kwon and Kim (2020). The optimal scheduling of IEGS leads to an overall cost reduction, accounting for DSM compensation, operational costs, carbon emission costs, and investment costs. The study also provides typical operating constraints for the safe operation of gas and electrical networks. The specific problem is outlined in

As proposed by Lee et al. (2020), DSM is used to regulate power electronic loads (PELs) in order to lower the system’s running expenses. The bus voltage is determined by modeling the PELs as variable resistors in this study. The best voltage solution for managing the PELs and reducing system expenses is then determined using a DSM procedure. The goal of the objective issue, which is to determine the optimal bus voltage V that lowers overall system costs is characterized as a minimization problem (Lee et al. 2020).

3.1 SEM systems used for PV generation

With the decrease in the usage of fossil fuels diminishing the world's energy production capabilities, renewable sources have gained more significance in the power sector. They are increasingly vital to meet the escalating demands of customers and the necessity for dependable, economically feasible energy supply (Nardelli et al. 2020). A new grid-connected system needs a small amount of PV storage capacity. Due to their discontinuous and extremely impulsive nature, RERs are frequently difficult to understand in terms of their potential advantages. To address this issue, an intelligent system of daily planning and precise energy accessibility forecasting is being planned and developed (Nguyen et al. 2020). An ISEMS was developed as part of this endeavor with a focus on renewable energy to control energy demand in a smart grid environment (Niu et al. 2020). The ISEMS’’ demand-side EM architecture makes it possible to employ RERs (Niknam et al. 2021). ISEMS is made up of a forecast-based ISEMS, an IoT environment for customer data collection on energy usage and consumption, and solar energy generation and data collection. Power scheduling, considering variable loads, is employed to minimize power consumption fluctuations. The proposed DSM approach prioritizes simplicity over complex algorithms. Pal et al. (2021) introduced an electric spring to control demand and eliminate voltage swings in microgrids due to RERs, maintaining constant voltage for critical loads. Three control scenarios demonstrate the effectiveness of the proposed approach. A previous study utilized direct load control and automated generation control for microgrid frequency changes.

A microgrid EM system (G) suggests optimized storage management in the presence of DGs, reducing grid load and consumption prices (Pan et al. 2020). A DRM system involves service providers at various energy grid levels for mutual benefit (Perumal et al. 2021). In the integration of RERs on consumer properties, proposed a resource scheduling system, optimizing energy usage cost-effectively and suggesting a battery scheduling technique. The effects of wind and solar energy on DSM and DSO were explored, emphasizing increased DSM value and reduced DSO cost. In the literature, introduced a stochastic multi-objective linear programming model for optimal scheduling of DERs was introduced. Energy hubs aim to interconnect energy carriers, using renewable sources like solar PV and windmills, reducing greenhouse gas output. Priyakumari and Mishra (2021) suggest strategies for residential applications with integrated renewable energy, considering loads, costs, and flexible generation. The incorporation of renewable energy in electricity grids increases operating costs prompting a technique using demand response for electricity changes. Rajasekaran et al. (2019) emphasized production planning and reserve market participation for economic potential. Ramachandran et al. (2021) provided architecture for EM, employing ToU pricing and DSM to cap energy use. The focus is on maximizing customer satisfaction and minimizing energy costs through efficient scheduling strategies. Rao et al. (2020) proposed RTES in smart homes using evolutionary algorithms. A DSM methodology for smart grid-connected homes with distributed generation was introduced, utilizing AI for battery management to lower customer power costs.

In SEM systems for PV generation, renewable sources gain significance due to reduced fossil fuel usage. An intelligent system for daily planning and precise energy forecasting was developed by Rao et al. (2021a,b). The ISEMS employs forecast-based DSM, IoT for data collection, and solar energy generation in a smart grid environment (Rao et al. 2021a,b). IoT was created to use devices in microchips, controls, transceivers, data, and communication protocols to enable rapid and easy interaction among commonplace parts like computers, smartphones, sensors, and actuators over the internet. Therefore, compared to human inspection, the system interacting with the IoT can monitor and regulate the solar system in a vast and isolated area more efficiently (Rao et al. 2021a,b). The system outlined in the article was designed to track the values of PV cells for each of these parameters as well as the voltage, current, temperature, and amount of sunshine received by solar cells (Rao et al. 2021a,b). Data collected by Arduino was uploaded to the internet using the NodeMCU wireless transmitter (Salehahmadi et al. 2021). In the presence of an internet connection, all sensor data are compiled and presented to the user as an analyzable image using ThingSpeak, an open-source IoT cloud platform (Sardari and Sardari 2020). A schematic of the system block is shown in Figure 5. The open-source IoT cloud network application ThinkSpeak is used in theory. Data from the sensor were uploaded to the cloud by the Arduino kit that was connected to the Wi-Fi device. The user receives a status app, and all information records created by the sensors are refreshed. To utilize this service, the user first needs to register an account or IP address that includes several channels for measuring different network parameters. Users can access information, including photos, through this platform. The data is readily available via an online interface, and monitoring is conducted over the internet (Rao et al. 2022). The benefit of the system is that it can be used to conveniently monitor solar cell production data from any location with an internet connection. The gateway information system diagram is illustrated in Figure 6 (SeyedShenava and Niazi 2020).

Figure 5

The EVs power flow diagram (Sardari and Sardari 2020).

Figure 6

The drive train-specific long-term passenger vehicle scales globally (Jayaraman et al. 2021).

4.1 RER, DER, and storage device integration in systems using DSM techniques

Integration of RER on consumer properties is quickly gaining acceptance. Xu et al. (2019) proposed a resource scheduling system to lower the cost of each user’s energy usage through effective utilization of the energy generated. To choose the right size of battery to maximize savings, a battery scheduling technique is suggested. The study demonstrates the effects of installing wind and solar energy on DSM and DSO. The benefits include a higher DSM value and lower DSO costs. The limitations associated with loads and storage systems that are thermostatically regulated are taken into account by the authors (Xu et al. 2021). The DER system’s ideal operating schedule was utilized by Yang et al. (2019) to lessen the negative economic and environmental effects of DER systems with diverse energy sources. The suggested stochastic multi-objective linear programming model handles the best operation scheduling for DER (Yao et al. 2019).

The interplay of many energy carriers through the grid is the goal of these energy hubs. Additionally, employing electricity from renewable sources like throughout the process, energy hubs use solar PV and windmills rather than gas-fired systems of scheduling horizon has reduced greenhouse gas output (Yasin et al. 2021). For stochastic producers, to lessen load shedding, as a fallback, combined cooling, heating, and power systems are utilized. Surplus energy is also kept in energy-storage devices. Ye et al. (2021) depicted the idea for the hub of renewable energy. Integrated with distributed energy systems, a demand response programmer, the rapid forward selection method, and Monte Carlo simulation are utilized. Simulation results demonstrate the effectiveness of the recommended technique in managing energy within energy hubs (You et al. 2019). For residential applications with integrated renewable energy resources (RERs), temporary reductions in system costs can be achieved (Yuan et al. 2021). They consider loads, energy costs, and flexible renewable generation. They have taken into account the battery’s cost and operating limitations, the power needs and life expectancy of each load, as well as the minimum and average delay standards (Zhang et al. 2021).

Increased power system operating costs are a result of the enormous magnitude of swings and uncertainties brought on by the wide-scale incorporation of RERs into electricity grids. Zhang et al. (2019) introduced a method utilizing demand response from multiple domestic appliances to mitigate fluctuations in electricity consumption. Increased activation of flexible loads is required due to higher energy price variations brought on by the inclusion of RERs. According to Zhang et al. (2019), harnessing the economic potential of emerging energy markets requires attention to patterns, production planning, and active participation in reserve markets.

For optimal outcomes, Zhang et al. (2019) provided architecture for EM for controlling the PV power outputs that consist of systems for managing supply and demand that use ToU pricing. To individually cap energy use, the DSM system employs PV power outputs in the distribution network hence lowering residential prices and any resulting discomfort to homeowners. The supply-side management system strives to preserve the quality of the electricity by utilizing the PV power outputs as fully as is practical (Zhang et al. 2019). To maximize customer pleasure while minimizing user energy costs through the most effective scheduling of household appliances, the EM by using energy storage technologies and RER in modern dwellings was examined by Zhang et al. (2019). In RTES, a DSM program is suggested for managing energy in smart homes. The evolutionary algorithm is used to solve the optimization issue at predetermined intervals, and the scheduling system inputs are changed before each optimization computation (Zhang et al. 2021).

Zhao et al. (2021) developed a methodology to conduct DSM for homes connected to a smart grid and equipped with distributed generation and energy storage devices. The battery management and optimization strategies used in this method are managed by artificial intelligence. The goal of the research is to lower customer power costs (Zhao et al. 2020).

6.1 Optimization algorithms

Electricity consumption varies across locations, making it essential to strike a balance between user comfort and power utilization. In a bid to minimize residential demand, a combination of GAs and Pigeon-inspired optimization (PIO) is proposed (Kim et al. 2021). The study evaluates GA, PIO, and HGP under a real-time pricing system, demonstrating their potential to reduce power expenses and consumption while enhancing customer satisfaction. Hybrid bio-inspired optimization techniques, specifically Bacterial foraging optimization, are employed in a residential EM method to reduce peak loads and power costs. This approach considers real-world equipment limits and user perspectives. Additionally, an optimized control strategy is proposed, combining dynamic pricing, user response, and equipment operating capabilities to lower power costs. To enhance energy use and DSM, an EM controller is introduced, utilizing heuristic optimization approaches such as "BAT inspired and flower pollination" for shiftable appliances and "hybrid BAT pollination optimization algorithm (HFBA)" for residential appliances (Khedkar et al. 2021). Fuzzy logic is incorporated to regulate interruptible appliances, addressing the unpredictability associated with manually operated appliances.

6.2 Pricing based on ToU

ToU pricing structures, where rates vary across predetermined periods, prove effective in lowering costs for both utilities and consumers. Studies in emerging nations like China explore the impact of ToU pricing at the residential level (Hossain and Fotouhi 2020). ToU pricing is incorporated in a decentralized DSM framework, optimizing residential and commercial loads with RER (Haider et al. 2021). The implementation of a household energy management system is recommended to schedule the use of RERs based on ToU pricing (Ghosh and Saha 2021). Multi-objective residential DSM with batteries is proposed to reduce power expenses and peak loads for residential consumers using ToU pricing. A mechanism for managing supply and demand sides based on ToU pricing is suggested for effective use of residential PV production (Gomez-Quiles and Valdivieso-Sarabia 2019).

6.3 Game approaches

Game approaches are employed to model interactions in the electricity market, considering electricity providers, local agents, end-users, and aggregators (Perumal et al. 2021). Different game strategies, including Bayesian games and cooperative games, are utilized to optimize day-ahead energy usage, especially in scenarios involving EVs (Ma et al. 2019). Non-cooperative game approaches characterize challenges in a smart grid with multiple energy suppliers, addressing peak demand scenarios (Eisen and Ribeiro 2020). A cooperative game technique is used to formulate a multi-objective residential side demand management issue, finding a compromise between reducing customer power costs and peak load demand (Dong et al. 2019). Cooperative games, based on super-criterion and Pareto-minimum solution, provide a balanced approach for various scenarios involving different numbers of home clients.

6.4 Research perspectives in IoT EM

EM is essential in microgrids for the effective and seamless transfer of energy among the numerous energy sources and storage innovations for a reliable, cost-effective, durable, and efficient operation. Because RERs are intermittent, a lot of expensive energy storage and a variety of energy sources are needed to ensure a safe and dependable power supply. EM is crucial in microgrids with RERs, dispatchable energy sources, storage systems, and loads to ensure the greatest possible power flow between the various components and preserve the system's optimal stability and cost-effectiveness. In EM studies, which are primarily simulation-based, there is a large category of publications using a range of algorithms and methodologies. However, further research is necessary to confirm the viability of applying the suggested methods in actual microgrids (Chen et al. 2020). The authors believe that simulation based on real-time systems can produce more credible results for practical application even when many methodologies are used in EM research since the results accurately mirror the behavior of the real system. More investigations using hardware in loop to assess the viability of practical implementation are essential for validating the results (Baneshi et al. 2020).

6.5 Future directions of demand side SEM system

There are several interesting possibilities in the fast-developing field of grid energy storage, including the research directions described here.

• Advanced control and optimization strategies: Modern control and optimization techniques are being developed by researchers to boost the effectiveness and performance of energy storage systems in microgrids (MGs). Innovative methods like artificial intelligence or machine learning could be used in these tactics (Ma et al. 2019).

• Hybrid energy storage systems: To address the issues with individual storage technologies and offer better overall performance, hybrid energy storage systems which integrate numerous storage technologies like batteries, capacitors, and super capacitors are being researched (Miao et al. 2019).

• Grid integration and management: In order to ensure the MG's dependable and successful functioning and interactions with the main grid, appropriate grid integration and management strategies are becoming increasingly important.

• New storage technologies: The effectiveness and efficacy of cutting-edge energy storage technologies are being investigated in contrast to those of the present, including flow batteries, thermal storage, and hydrogen storage.

• Smart grid integration: With the widespread adoption of smart grid technologies, researchers are focusing on enhancing the effectiveness and reliability of power transmission. Innovative methods for integrating energy storage devices into smart grids are being explored to address these needs.

• Life cycle analysis: In order to comprehend how MG energy storage systems affect the environment, and to pinpoint areas for improvement, researchers are performing life cycle assessments.

• Standardization: Standardization is required because of the proliferation of various MG configurations and energy storage technologies in order to guarantee system interoperability and compatibility.

• Energy trading: To maximize energy utilization and cut costs, researchers are looking into the potential for energy trade across other MGs and between microgrids and the main grid.

• Reliability and resiliency: Researchers are looking into redundant storage systems and cutting-edge control algorithms as potential improvements for the dependability and resilience of energy storage systems in MGs (Nardelli et al. 2020).

• Optimal sizing and placement: Effective techniques for designing and positioning these systems are required in order to maximize their performance and financial feasibility in light of the growing demand for MG energy storage systems.

• Aging and degradation: The capacity and performance of energy storage devices may decline with time, which might reduce their reliability and efficiency. Researchers can enhance the replacement and maintenance of MG energy storage devices by investigating modeling and predicting the aging and degradation of these systems.

• Environmental and social impact: In order to reduce adverse effects and advance sustainability, Energy storage systems in MGs are being studied for their environmental and social consequences, including how they affect social justice, resource consumption, and land use.

• Electromagnetic compatibility: Energy storage systems must be electromagnetically compatible in order to avoid interference and guarantee reliable operation in microgrids, where electronic gadgets and wireless communications are proliferating. Last but not least, as the sector develops, energy storage technologies and the uses for them in MG systems are expected to develop and innovate further (Niu et al. 2020).