IntelliGrid AI: A Blockchain and Deep-Learning Framework for Optimized Home Energy Management with V2H and H2V Integration

Featured Application

IntelliGrid AI revolutionizes smart home energy management by integrating blockchain, deep learning, and vehicle-to-home (V2H) technology, enabling optimized energy consumption, secure peer-to-peer energy trading, and adaptive scheduling. It demonstrated that 20% reduction in energy costs and scalability makes it ideal for renewable-powered communities and smart city applications.

Abstract

The integration of renewable energy sources and electric vehicles has become a focal point for industries and academia due to its profound economic, environmental, and technological implications. These developments require the development of a robust intelligent home energy management system (IHEMS) to optimize energy utilization, enhance transaction security, and ensure grid stability. For this reason, this paper develops an IntelliGrid AI, an advanced system that integrates blockchain technology, deep learning (DL), and dual-energy transmission capabilities—vehicle to home (V2H) and home to vehicle (H2V). The proposed approach can dynamically optimize household energy flows, deploying real-time data and adaptive algorithms to balance energy demand and supply. Blockchain technology ensures the security and integrity of energy transactions while facilitating decentralized peer-to-peer (P2P) energy trading. The core of IntelliGrid AI is an advanced Q-learning algorithm that intelligently allocates energy resources. V2H enables electric vehicles to power households during peak periods, reducing the strain on the grid. Conversely, H2V technology facilitates the efficient charging of electric cars during peak hours, contributing to grid stability and efficient energy utilization. Case studies conducted in Tunisia validate the system’s performance, showing a 20% reduction in energy costs and significant improvements in transaction efficiency. These results highlight the practical benefits of integrating V2H and H2V technologies into innovative energy management frameworks.

1. Introduction

The necessity for IHEMS is rising due to the growing adoption of renewable energy sources and EVs. These systems are crucial for reducing energy expenditure and improving domestic efficiency. As energy consumption patterns evolve, maintaining the balance between supply and demand becomes essential. Traditional systems lack the capabilities to address the complexities of modern energy management, emphasizing the need for advanced solutions [1]. With more households adopting solar panels and EVs, there is an urgent demand for decentralized systems that are capable of intelligent energy distribution and efficient usage. These systems must also adapt to fluctuations in supply and demand, ensuring minimal energy waste and reduced costs. Future homes must incorporate secure and decentralized energy management frameworks to meet these needs, paving the way for sustainable smart grids [2].

Blockchain technology plays a pivotal role in enabling secure energy transactions. It supports P2P trading among households, empowering prosumers to sell surplus energy directly. By removing intermediaries, blockchain facilitates faster, more transparent, and cost-effective transactions. Additionally, blockchain ensures data security and integrity, which are vital for managing decentralized energy systems. Such capabilities make blockchain an indispensable component of IHEMS, fostering trust and reliability in energy markets [3,4].

The integration of DL algorithms is essential for optimizing energy utilization in smart homes. DL models analyze extensive datasets, including historical and real-time energy consumption patterns, to make precise predictions about future energy needs. These insights enable the system to dynamically adjust the operation of household appliances, allocate energy efficiently, and prioritize essential loads during peak periods. By reducing unnecessary energy consumption, DL minimizes costs and enhances the sustainability of modern homes [5,6]. Furthermore, DL-powered algorithms play a critical role in forecasting renewable energy generation, ensuring intelligent scheduling and better integration within sustainable smart grids. V2H and H2V technologies significantly enhance the flexibility and reliability of energy systems. V2H allows EVs to function as mobile energy storage units, supplying power back to homes during peak demand or when renewable energy is insufficient [7,8].

This reduces the reliance on the grid and provides households with a resilient energy backup. Conversely, H2V enables EVs to be charged efficiently during off-peak hours, utilizing excess renewable energy and preventing waste [9,10]. Together, V2H and H2V technologies ensure a dynamic, two-way energy flow, stabilizing the grid and optimizing energy distribution. Their integration into IHEMS empowers homeowners by providing greater control over energy resources and supporting cost savings and environmental sustainability. Despite its benefits, blockchain technology faces challenges in scalability when applied to energy management. As networks expand, validating transactions can become time-consuming and energy-intensive, creating bottlenecks that reduce system efficiency. Exploring alternative consensus mechanisms like proof of stake (PoS) offers potential solutions to these issues. Optimizing blockchain architectures can further accelerate and streamline transaction processes, enabling efficient integration within IHEMS [11,12].

IntelliGrid AI is a sophisticated smart home energy management system designed to address these challenges. By integrating blockchain technology and DL, it enhances energy efficiency and security. The system supports secure peer-to-peer energy trading and incorporates V2H technology to store and share energy from electric vehicles. Advanced Q-learning algorithms facilitate real-time energy adjustments, ensuring optimal usage. Case studies in Tunisia highlight the system’s effectiveness, demonstrating a 20% reduction in energy costs. IntelliGrid AI represents a secure, efficient, and scalable solution for modern homes, contributing to the development of sustainable smart grids.

1.1. Related Works

The prevalence of IHEMS is increasing as they facilitate the administration of energy utilization within domestic environments. These systems are instrumental in reducing energy expenditure and enhancing efficiency. Modern IHEMSs frequently incorporate renewable energy sources, such as solar panels, and prioritize real-time optimization for enhanced control over energy usage. The primary objective of these systems is to reduce waste and ensure efficient energy utilization in smart homes. As a key component of sustainable smart grids, an IHEMS manages energy distribution and contributes to environmental sustainability by minimizing reliance on non-renewable resources [13,14].

Blockchain technology is essential to ensuring the security and transparency of energy trading within IHEMSs. It facilitates peer-to-peer (P2P) energy trading, enabling households to sell surplus energy directly to other users without intermediaries. This decentralized approach promotes faster, more economical transactions, benefiting prosumers and consumers alike. Blockchain also guarantees the integrity of energy transactions, protecting them from tampering or manipulation. Its ability to secure data makes blockchain an optimal solution for decentralized energy management and a critical enabler of sustainable smart grids [15,16].

Integrating DL further enhances energy management within IHEMSs. By leveraging DL algorithms, systems can analyze extensive datasets, predict household energy requirements, and make real-time adjustments to energy usage. This capability facilitates the efficient operation of home appliances, reducing overall energy consumption and lowering costs. DL also enables predictive modeling for renewable energy generation, ensuring seamless integration with household energy needs. These advancements significantly enhance the performance of modern IHEMS and contribute to the optimization of sustainable smart grids [17,18].

V2H technology is revolutionizing energy flexibility in smart homes. V2H enables electric vehicles (EVs) to store energy and share it with domestic premises during periods of high demand. This reduces reliance on the power grid and provides homeowners with a cost-effective alternative energy source. As EV adoption increases, V2H becomes a critical component of modern IHEMS, offering enhanced resilience during energy shortages. Similarly, H2V technology complements V2H by enabling EVs to charge during off-peak hours, ensuring optimal energy distribution and supporting grid stability. V2H and H2V technologies are essential for achieving energy efficiency and flexibility within sustainable smart grids [19,20].

Despite its advantages, blockchain technology faces challenges in scalability. Transaction times may slow as networks grow, and energy consumption may increase. Researchers are exploring innovative solutions, such as PoS, to address these issues and improve efficiency. PoS reduces the energy-intensive nature of traditional blockchain processes while accommodating more transactions. Optimizing blockchain architectures can enhance scalability, making it suitable for integration into extensive energy networks. These advancements are crucial for ensuring the long-term viability of blockchain-based IHEMS [21,22].

AI integration is transforming IHEMS by introducing automation and predictive capabilities. AI-driven systems analyze real-time data from connected devices to adjust energy usage dynamically. This ensures optimal system efficiency while streamlining energy management processes. Additionally, AI facilitates predictive maintenance, identifying potential system failures before they occur. This reduces operational costs and guarantees uninterrupted functionality, making AI a vital component of modern IHEMS and sustainable smart grids [23,24].

DL and AI also work together to enhance the adaptive capabilities of IHEMS. For instance, AI-powered DL models predict energy demand and renewable energy availability, optimizing the operation of V2H and H2V technologies. This integration ensures that energy is distributed efficiently, minimizing waste and stabilizing the grid. Adapting to fluctuating energy conditions is vital for households seeking to maximize the benefits of renewable energy sources while reducing costs [25,26].

Combining V2H and H2V technologies within IHEMS promotes a bi-directional energy flow, ensuring that households can access flexible energy solutions. During peak demand, V2H allows homes to draw energy from EVs, while H2V ensures that EVs are efficiently charged during off-peak hours. These technologies empower homeowners to manage their energy needs dynamically, aligning with the principles of sustainable smart grids [27,28].

The IHEMS extends beyond individual households. As these systems become more interconnected, they contribute to the broader energy ecosystem by improving grid reliability and supporting the integration of renewable energy sources. By combining blockchain technology, deep learning, and V2H/H2V capabilities, IHEMSs create a secure, adaptive, and efficient framework for modern energy management. This integration enables households to participate actively in energy trading and resource optimization, driving the transition to more resilient and sustainable smart grids [29,30].

Table 1. Key developments and comparative analysis of approaches in energy management.

Feature	IntelliGrid AI	Related Works	Ref.
Energy Optimization	Real-time optimization using adaptive algorithms	Basic real-time optimization	[13,14]
Transaction Security	Blockchain ensures secure transactions	Blockchain ensures transparency and security	[15,16]
Decentralized Energy Trading	P2P trading supported via blockchain	P2P trading with intermediaries	[15,16]
Deep Learning Integration	Advanced Q-learning for energy allocation	Basic DL for energy prediction	[17,18]
V2H and H2V Technologies	Full V2H and H2V integration	Partial V2H, minimal H2V use	[19,20]
Scalability	PoS for scalable blockchain	Scalability challenges with PoS research	[21,22]
Real-Time Data Usage	Extensive use for demand-supply balance	Limited real-time adaptive features	[23,24]
Predictive Maintenance	AI-enabled predictive maintenance	Limited AI use for automation	[23,24]
Impact on Energy Costs	20% cost reduction in Tunisia case studies	Cost reduction without specific metrics	[25,26]

compares IntelliGrid AI’s features with related works oon IHEMSs. IntelliGrid AI distinguishes itself through an advanced integration of blockchain, deep learning, and bidirectional V2H and H2V energy transfer technologies, ensuring dynamic optimization, enhanced transaction security, and scalability. While related works emphasize the importance of real-time optimization and energy trading, IntelliGrid AI provides quantifiable benefits, including a 20% reduction in energy costs, that are validated through case studies. This highlights IntelliGrid AI’s innovative approach to achieving efficient, secure, and sustainable energy management.

1.2. Contributions

Compared to previous works, our research introduces IntelliGrid AI, a novel IHEMS that integrates blockchain technology, DL, and bidirectional V2H and H2V energy transfer capabilities. Unlike prior systems, IntelliGrid AI ensures secure and transparent P2P energy trading using blockchain, eliminating intermediaries and safeguarding transaction integrity. It employs advanced Q-learning algorithms for dynamic, real-time optimization of household energy flows, balancing demand and supply more efficiently. The complete integration of V2H and H2V technologies enhances energy flexibility by enabling EVs to provide energy during peak demand and charge efficiently during off-peak hours, stabilizing the grid. Additionally, our approach addresses blockchain scalability challenges through PoS mechanisms and optimized architectures, making it suitable for extensive energy networks. By leveraging DL, IntelliGrid AI predicts energy consumption and renewable energy generation, facilitating efficient energy allocation and reducing waste. Validated through case studies in Tunisia, the system demonstrates a 20% reduction in energy costs and significant improvements in transaction efficiency, showcasing its superiority in advancing sustainable smart grids and modern energy management solutions.

1.3. Challenges and Gaps

IntelliGrid AI brings many advancements, but some challenges remain. Blockchain technology, especially energy-intensive mechanisms like PoW, is not very efficient and needs a shift to better methods like PoS. Using data from many different sources in real-time and handling changes in the renewable energy supply is still a difficult task. The adoption of V2H and H2V technologies faces issues like limited infrastructure and lack of compatibility. Decentralized systems and devices often struggle to work smoothly together, creating interoperability problems. Additionally, the high cost of implementing these systems and unclear regulations make it harder to adopt them widely. Solving these issues is key to building energy management systems that are scalable, efficient, and sustainable.

1.4. Paper Outline

This paper is organized as follows. Section 1—Introduction discusses the growing need for IHEMS and introduces IntelliGrid AI, highlighting its importance in improving energy use and integrating renewable sources. Section 2—Methodology outlines the development of IntelliGrid AI, including integrating blockchain, DL, and bidirectional energy flow (V2H and H2V). It explains how these technologies can improve energy management in smart homes. Section 3—AI for the Smart Grid: A Smart Home Energy Management System Using Blockchain and AI describes the core architecture and algorithm of IntelliGrid AI, detailing how it uses machine learning for real-time energy optimization, secure peer-to-peer trading using blockchain, and efficient energy flow management using V2H and H2V technologies. Section 4—Results and Discussion presents the results from Tunisia, which demonstrate a 20% reduction in energy costs and improved transaction efficiency while analyzing the system’s scalability, flexibility, and implications for innovative grid applications. Section 5—Conclusion summarizes the contributions of IntelliGrid AI, discusses its potential for future applications, and suggests further research to enhance its capabilities and scalability in innovative grid systems.

2. Methodology

The development of IntelliGrid AI involves a comprehensive integration of blockchain technology, DL, and energy engineering to create an advanced IHEMS. The system consists of three primary units: the energy management unit (EMU), which monitors real-time energy usage and controls household appliances for optimal efficiency; the blockchain-based transaction layer, which facilitates secure and transparent P2P energy trading using smart contracts; and the DRL module, powered by the deep deterministic policy gradient (DDPG) algorithm, which dynamically optimizes energy flows based on real-time and historical data. The system incorporates V2H and H2V technologies to enable bidirectional energy transfer, allowing EVs to supply energy to homes during peak demand and charge efficiently during off-peak periods. Data from intelligent meters, meteorological sources, and energy tariffs support real-time decision-making and dynamic pricing adjustments. Blockchain scalability challenges are addressed using energy-efficient PoS mechanisms. The system is trained and validated through simulations that evaluate performance metrics such as cost savings, grid dependency reduction, and transaction security. This approach ensures that IntelliGrid AI provides a secure, adaptive, and efficient energy management solution, promoting sustainability and scalability within modern smart grids.

2.1. System Design and Integration

IntelliGrid AI integrates blockchain technology, DL, and bidirectional energy transfer capabilities (V2H and H2V) to create an advanced IHEMS. The system consists of three main components: the energy management EMU, which monitors real-time energy usage, prioritizes loads, and manages bidirectional energy flows; the blockchain-based transaction layer, which facilitates secure and decentralized P2P energy trading using smart contracts and a PoS mechanism; and the DRL module, which employs the DDPG algorithm for dynamic energy optimization and adaptive scheduling.

The system integrates data from intelligent meters, renewable energy forecasts, energy tariffs, and EV state of charge (SoCEV) to enable efficient energy allocation, reduce costs, and enhance grid stability, providing a scalable solution for sustainable smart grids.

Figure 1 illustrates the IntelliGrid AI system architecture, highlighting its three primary components: the EMU, the blockchain-based transaction layer, and the DRL Module. The EMU interacts with data sources, including intelligent meters, renewable energy inputs, and EV SoC, to monitor and manage energy flows. The blockchain layer secures P2P energy trading through smart contracts and PoS mechanisms. The DRL module dynamically optimizes energy allocation and scheduling based on real-time predictions. Bidirectional energy flows via V2H and H2V are integrated seamlessly, ensuring efficient energy utilization and reduced grid dependency. This architecture demonstrates a robust framework for modern smart home energy management.

2.2. Problem Statement

Power systems today face significant challenges, such as inefficiencies in centralized management, poor use of renewable energy, and insecure P2P energy trading. Centralized systems often allocate power inefficiently and incur high operating costs. Grid stability is also a significant issue due to the unpredictable nature of renewable energy sources. Additionally, current P2P trading methods lack strong security and scalability, which raises concerns about transparency and reliability. If these traditional systems continue, energy costs will rise, grid instability will worsen, and more energy will be wasted.

To tackle these problems, IntelliGrid AI offers a solution by integrating blockchain for secure and transparent P2P trading, deep learning for intelligent and adaptive energy management, and V2H and H2V technologies for two-way energy flow. This system improves energy efficiency, reduces costs, enhances grid stability, and ensures a better integration of renewable energy, providing a sustainable and flexible approach to meet todays and tomorrow’s energy needs (See Figure 2).

2.3. Power Consumption Calculation for Household Appliances

Equation (1) calculates the total energy consumption (E) for household appliances by multiplying the power rating (P) of each appliance by its operating time (T) across specified intervals. This matrix-based approach provides a clear and scalable framework for evaluating energy usage.

2.4. Energy Demand Prediction

The energy demand prediction equation is pivotal in the IntelliGrid AI framework. It forecasts the future energy demand (Ed(t + 1)) by analyzing historical energy consumption (Eℎ(t)), weather conditions (W(t)), and appliance usage patterns (P(t)). Each factor is assigned a weight (α, β, γ) to reflect its relative importance in shaping energy demand. This prediction enables the system to anticipate household energy needs, facilitating dynamic scheduling and efficient energy allocation. By incorporating real-time weather data and historical consumption patterns, the prediction model adapts to fluctuations in renewable energy availability and changing usage behaviors, ensuring optimized resource utilization and cost savings [31]. The equation is the foundation for proactive energy management, supporting the seamless integration of renewable energy sources and smart appliances into modern home energy systems (See Equation (2)).

$$\left(\begin{array}{l}\mathrm{E}=\mathrm{P}.\mathrm{T}\\ \mathrm{P}=\left[\begin{array}{c}{\mathrm{p}}_1\\ {\mathrm{p}}_2\\ {\mathrm{p}}_3\\ \\ {\mathrm{p}}_{\mathrm{N}}\end{array}\right]\\ \mathrm{T}=\left[\begin{array}{cccc}{\mathrm{T}}_{11}& {\mathrm{T}}_{11}& & {\mathrm{T}}_{1\mathrm{M}}\\ {\mathrm{T}}_{21}& {\mathrm{T}}_{22}& & {\mathrm{T}}_{2\mathrm{M}}\\ \mathrm{\dots }& \mathrm{\dots }& \mathrm{\dots }& \mathrm{\dots }\\ {\mathrm{T}}_{\mathrm{N}1}& {\mathrm{T}}_{\mathrm{N}2}& & {\mathrm{T}}_{\mathrm{NM}}\end{array}\right]\end{array}\right.$$

(1)

where P₁, P₂, …, P_N is the power consumption (kW) of appliances 1 through N. T_ij is the duration (hours) that appliance i operates in time interval j.

$$\left|\begin{array}{l}{\mathrm{E}}_{\mathrm{d}}(\mathrm{t}+1)=\left(\mathsf{\alpha }.{\mathrm{E}}_{\mathrm{h}}\left(\mathrm{t}\right)\right)+\left(\mathsf{\beta }.\mathrm{W}\left(\mathrm{t}\right)\right)+\left(\gamma .\mathrm{P}\left(\mathrm{t}\right)\right)\\ {\mathrm{E}}_{\mathrm{d}}=\left[\begin{array}{ccc}\mathsf{\alpha }& \mathsf{\beta }& \mathsf{\gamma }\end{array}\right].\left[\begin{array}{c}{\mathrm{E}}_{\mathrm{h}}\left(\mathrm{t}\right)\\ \mathrm{W}\left(\mathrm{t}\right)\\ \mathrm{P}\left(\mathrm{t}\right)\end{array}\right]\end{array}\right.$$

(2)

2.5. Renewable Energy Prediction

The renewable energy prediction equation calculates the energy generated by renewable sources, particularly solar panels, at a given time (Er(t)). This is achieved by factoring in the efficiency of the solar panels (ηs), the total panel surface area (A), and the solar irradiance (I(t)) at that moment. By accurately estimating the renewable energy output, the equation helps align energy generation with household demand, enabling better planning and allocation. It supports the integration of solar energy into the overall energy management system, reducing reliance on grid electricity and promoting sustainability [32]. Additionally, the predictive capabilities of this equation play a critical role in optimizing energy storage and usage, ensuring the maximum (See Equation (3))

$$\left\{\begin{array}{l}{E}_r(t)={\eta }_s\cdot A\cdot I(t)\\ E_r(t)=\left[\begin{array}{ccc}{\eta }_s& A& I(t)\end{array}\right]\end{array}\right.$$

(3)

2.6. Dynamic Energy Allocation with Cost Optimization

Equation (4) ensures that the total allocated energy (Ea) effectively meets household energy demand (Ed), efficiently utilizing available resources. Initially, renewable energy (Er) is prioritized to satisfy demand, leveraging the system’s reliance on sustainable resources. If renewable energy is insufficient, the equation supplements it with energy from V2H transfers (E_V2H), ensuring a reliable energy supply. This approach optimizes energy usage while minimizing dependence on the grid, promoting cost savings and grid stability [33]. Including the primary function ensures that energy allocation does not exceed the actual demand, thereby preventing waste. This equation is integral to the IntelliGrid AI framework, enabling dynamic energy distribution that aligns with sustainability goals and real-time system needs.

$$\left|\begin{array}{l}{E}_a=min(E_r+E_{V2H},E_d)\cdot (1-{\displaystyle \frac{C_{peak}}{C_{total}}})\\ E_a=min(\left[\begin{array}{c}{E}_r+\\ EV2H\end{array}\right],[Ed]\cdot (1-{\displaystyle \frac{C_{peak}}{C_{total}}})\end{array}\right.$$

(4)

2.7. Energy Surplus for Trading Equation

The energy surplus for trading equation calculates the excess energy (E_s) available after meeting the household energy demand (Ed) using renewable energy (Er) and V2H energy (EV2H). This equation determines the surplus energy that can be traded in a peer-to-peer (P2P) energy market, supporting decentralized energy systems. By accounting for renewable generation and V2H contributions, the model ensures that only excess energy beyond the household’s immediate needs is allocated for trading. This approach minimizes wastage while promoting efficient energy use. In the IntelliGrid AI framework, the surplus energy calculation is vital for enabling secure and scalable P2P trading through blockchain technology [34]. The equation ensures that trading activities do not compromise household energy requirements, enhancing energy self-sufficiency and grid stability. All appendices are provided in Appendix A 

Table A1. Parameters and values used in IntelliGrid AI models and equations.

Parameter	Symbol	Description	Example Value	Unit
Energy Demand	Ed	Total energy demand at time t.	5	kWh
Renewable Energy	Er	Energy is available from renewable sources.	3	kWh
Energy Surplus	Es	Excess energy is available after demand is met.	2	kWh
Energy Waste	Ew	Unused renewable energy due to system limitations.	0.5	kWh
V2H Energy	EV2H	Energy supplied from EV to the home.	1.5	kWh
H2V Energy	EH2V	Energy is used to charge EV from the home.	2	kWh
State of Charge	SoCt	Battery state of charge at time t.	80	%
Peak Cost	Cpeak	Cost of energy during peak hours.	0.25	USD/kWh
Dynamic Cost	C(t)	Cost of energy at time t.	0.15	USD/kWh
Allocated Energy	Ea	Energy is allocated to meet demand.	5	kWh
Energy Efficiency	η	Efficiency factor for energy utilization.	9	%
Discount Factor	γ	Weighting of future rewards in optimization.	0.95	-
Policy Function	pi(at)	Probability of taking action at in state st.		0.8
Action-Value Function	Q(st,at)	Value of taking action at in state st.	10	-
Energy Prediction Input	Xt	Input feature vector for energy prediction (e.g., weather, historical data).	-	-
Weighting Factors	w1, w2, w3	Weights for surplus, waste, and cost in the reward function.	0.5, 0.3, 0.2	-
Battery Capacity	battery_max	Maximum EV battery capacity.	50	kWh
Time Horizon	T	Duration for which optimization is performed.	24	Hours
Transition Probability	(P(st, at))	Probability of transitioning to a new state.	-	-

.

$$\left|\begin{array}{l}{E}_s=E_r+E_{V2H}-E_d\\ E_s=\left[\begin{array}{l}Er\\ +\\ EV2H\end{array}\right]-[Ed]\end{array}\right.$$

(5)

2.8. Blockchain Energy-Trading Efficiency

Equation (6) quantifies the blockchain energy-trading efficiency (Te), a critical metric for assessing the performance of decentralized energy transactions within the IntelliGrid AI system. The equation evaluates the adequate energy traded per second by considering the surplus energy available for trading (Es), the transaction efficiency (η), and the transaction time (t). This metric ensures that energy-trading processes are optimized for security, scalability, and cost-effectiveness. By factoring in the blockchain efficiency (η), the equation highlights the importance of minimizing energy losses during transaction validation. Including the transaction time emphasizes the need for faster blockchain operations to enhance trading efficiency. In IntelliGrid AI, this equation ensures reliable peer-to-peer energy trading, supporting a decentralized and resilient energy ecosystem while promoting sustainable energy utilization [35]. PoW consensus mechanisms require significant computational resources. PoS mechanisms are more energy-efficient. The energy consumed (E_PoS) is based on the number of validators and the power consumption of each validator.

$$\left|\begin{array}{l}Te=\eta \cdot {\displaystyle \frac{Es}{t}}\\ E_{PoW}=H\cdot P_{comp}\cdot t_{comp}\\ E_{PoS}=N_v\cdot P_{node}\cdot t_{valid}\end{array}\right.$$

(6)

2.9. Dynamic Cost Optimization

Equation (7) is a critical component of IntelliGrid AI, designed to minimize energy costs while ensuring efficient energy allocation. This process involves real-time adjustments to energy usage based on dynamic pricing models and renewable energy availability. This approach prioritizes energy usage during off-peak hours when costs are lower, utilizing renewable energy and V2H resources to reduce dependence on the grid further. IntelliGrid AI dynamically schedules appliances and EV charging to achieve cost-efficient operations by integrating predictive models for demand and pricing [36]. This methodology ensures financial savings for users while contributing to grid stability and the sustainable management of energy resources.

$$\left|\begin{array}{l}{C}_{opt}={\displaystyle \sum _1^T(C(t)\cdot E_a(t))}\\ Copt=C(t)·{Ea}^T\\ C=[C_{(1)},C_{(2)},\dots ,C_{(T)}]\\ E_a=[E_{a(1)},E_{a(2)},\dots ,E_{a(T)}]\end{array}\right.$$

(7)

2.10. Bidirectional Energy Flow for V2H and H2V and State of Charge for EV

Bidirectional energy flow, encompassing V2H and H2V technologies, is a cornerstone of IntelliGrid AI’s dynamic energy management. V2H enables EVs to supply energy to households during peak demand, reducing reliance on the grid and supporting cost-effective energy utilization. Conversely, H2V facilitates the efficient charging of EVs during off-peak hours, utilizing surplus renewable energy and ensuring minimal waste. These bidirectional flows are governed by Equation (8). The state of charge (SoC) of EV batteries further ensures efficient energy transfer and optimal battery health. This equation updates the SoC based on the net energy exchange and represents the maximum battery capacity. These equations enable IntelliGrid AI to dynamically balance energy flows, maintain EV readiness, and optimize grid stability [37]. This seamless integration of V2H and H2V technologies empowers households with flexible and sustainable energy solutions while ensuring an effective utilization of renewable resources and an enhanced resilience against energy fluctuations.

$$\left|\begin{array}{l}{E}_{bidirectional}=E_{V2H}-E_{H2V}\\ SoC\left(t\right)=SoC\left(t-1\right)+{\displaystyle \frac{E_{H2V}-E_{V2H}}{E_{battery\_max}}}\end{array}\right.$$

(8)

2.11. Deep-Learning Model for Energy Prediction

The deep-learning (DL) model in IntelliGrid AI is a cornerstone for the accurate forecasting of energy demand. This model leverages historical consumption patterns, weather data, and appliance usage behaviors to predict future energy demand dynamically. By minimizing the mean squared error (MSE) loss function during training, the model refines its predictions to align closely with real-world energy requirements. These forecasts guide the energy allocation and cost optimization processes, enabling the proactive scheduling of appliances and an efficient utilization of renewable energy sources. Integrated within IntelliGrid AI, the DL model ensures a robust, adaptive, and predictive framework for managing energy resources in smart homes, reducing costs, and enhancing grid stability [38].

$$\left\{\begin{array}{l}{E}_d(t)=f_{\theta }(X_t)\\ L(\theta )={\displaystyle \frac{1}{N}}{\displaystyle \sum _1^N{(E_d^{True}d(t_i)-E_d^{pred}(t_i))}^2}\\ \theta =\theta -\alpha \cdot {\nabla }_{\theta }L(\theta )\end{array}\right.$$

(9)

2.12. Markov Decision Process (MDP) for Energy Management

The MDP is a foundational framework used in IntelliGrid AI to model the sequential decision-making process in energy management. The MDP captures the stochastic nature of energy supply and demand, enabling the system to make adaptive and optimal decisions [39,40].

$$\left|\begin{array}{l}\left|\begin{array}{l}State:\left|\begin{array}{l}P(s_{t+1}\mid s_t,a_t)=\pi (a_t\mid s_t)\\ s_t=[E_d(t),E_r(t),SoC(t)]\end{array}\right.\\ Reward/Action:\left|\begin{array}{l}{R}_t=w_1\cdot (E_s-E_w)-w_2\cdot C_{peak}+w_3\cdot \eta \\ a_t\in \{Allocate,Charge,Discharge,Trade\}\end{array}\right.\end{array}\right.\\ StateValue:\left|\begin{array}{l}V(s_t)=E_{\pi }[k={\displaystyle \sum _0^{\infty }{\gamma }^k{R}_{t+k}}]\\ Q(s_t,a_t)=R_t+\gamma E[V(s_{t+1})]\\ {\pi }^{\ast }(s)=arg\underset{a}{max}Q(s,a)\end{array}\right.\end{array}\right.$$

(10)

3. Smart Grid AI: Intelligent Home Energy Management System Using Blockchain and AI

3.1. The Proposed Intelligent Home Energy Management Algorithm

The Smart Grid AI algorithm is the core of an IHEMS, leveraging blockchain technology and AI to optimize energy allocation and usage in modern smart homes. This algorithm integrates V2H and H2V technologies, dynamic energy-forecasting models, and a secure decentralized transaction framework. The algorithm ensures efficient energy allocation and reduced costs by combining DL for energy demand forecasting and DRL for adaptive scheduling. Blockchain technology adds a strong layer of security and transparency, enabling P2P energy trading and secure data handling. The algorithm’s design emphasizes scalability, sustainability, and interoperability, addressing the complexities of renewable energy integration, grid stability, and electric vehicle management. This comprehensive approach makes Smart Grid AI a transformative solution for developing innovative grid systems and promoting sustainable energy practices. Algorithm 1 outlines the core computational framework of the Smart Grid AI system, designed to optimize energy management in intelligent homes. The algorithm integrates dynamic energy demand prediction, renewable energy forecasting, and bidirectional energy flow management through V2H and H2V technologies. It employs a DRL approach for adaptive energy scheduling and blockchain technology to ensure secure P2P energy trading. By incorporating real-time data inputs, such as energy demand, renewable energy availability, and EV state of charge (SoC), the algorithm dynamically adjusts energy allocation to minimize costs, enhance grid stability, and improve overall efficiency. This robust framework ensures the seamless operation of IntelliGrid AI, enabling smart homes to achieve sustainable energy goals while reducing dependency on traditional grid systems. We proposed two algorithms. The first is the dynamic energy cost optimization (DECO) algorithm, and the second is adaptive energy management algorithm, which serves complementary roles in modern innovative energy systems, with each targeting specific efficiency and sustainability aspects.

The DECO algorithm primarily focuses on minimizing energy costs by leveraging dynamic pricing, renewable energy utilization, and bidirectional energy flows (V2H and H2V). It emphasizes cost-effective scheduling, prioritizing renewable energy for critical appliances while shifting non-critical loads to off-peak hours. In contrast, the adaptive energy management algorithm provides a comprehensive framework integrating DL, blockchain technology, and bidirectional energy flows.

Algorithm 1: Smart Grid AI Optimization for Intelligent Home Energy Management
1	Initialization  • Load initial system parameters (Ed; Er, SoC).  • Configure blockchain network and deep learning (DL) model.  • Define optimization goals and constraints.
2	Input Data  • Collect real-time energy demand (Ed).  • Gather renewable energy availability (Er) and EV (SoC).
3	Predict Energy Demand  • Use the DL model to predict future energy demand:                E d (t) = f θ (X t)      IF prediction error > threshold:       ○ Update DL model with new data.
4	Renewable Energy Forecast  • Estimate renewable energy output:         Er = I⋅A⋅η pv     IF Er = 0:       ○ Trigger fallback energy sources (grid or V2H).
5	Energy Allocation  • Calculate allocated energy:       Ea = min(Er + EV2H,Ed) IF Ea < Ed       ○ Allocate remaining demand to grid supply.
6	Bidirectional Energy Flow  • Manage V2H and H2V operations: Ebidirectional = EV2H − EH2V  IF SoC < threshold:     ○ Disable V2H to preserve EV battery health.   ELSE:     ○ Enable V2H for peak demand.
7	Dynamic Cost Optimization  • $\mathrm{Optimize}\mathrm{energy}\cos \mathrm{ts}:=$  IF cost > budget threshold:     ○ Shift non-critical appliances to off-peak hours.
8	Blockchain for P2P Energy Trading  • Verify energy surplus for trading: Et = max (0, Es − δ)  • Initiate secure transactions using PoS mechanism.
9	Reinforcement Learning Optimization  • Update Q-value for action taken: Q(st, at) = Rt + γ⋅E[V(st + 1)]   IF reward (Rt) < target:     ○ Explore alternative actions.   ELSE:     ○ Continue with current policy.
10	Iterative Feedback Loop  • Monitor performance metrics (energy savings, grid stability).  • Update models (DL, DRL) with new data.  • Repeat until optimization goals are met.

This algorithm dynamically forecasts energy demand and renewable availability, ensuring precise energy allocation while incorporating secure P2P trading via blockchain and leveraging V2H and H2V technologies for energy flexibility. While DECO is cost-centric, adaptive energy management algorithm adopts a holistic approach, addressing scalability, grid stability, transaction security, and cost optimization. Together, these algorithms create a robust and intelligent energy management system, with DECO offering immediate cost-saving measures and adaptive energy management algorithm delivering a scalable and sustainable solution for modern smart grids.

3.2. Adaptive Energy Management Algorithm Using DL, Blockchain, and Bidirectional Energy Flow

Figure 3 presents the “Enhanced Dynamic Energy Allocation Algorithm (Enhanced DEA),” which integrates DL, bidirectional energy flow technologies (V2H and H2V), and blockchain to form a secure, intelligent, and adaptive energy management framework. The workflow begins with data collection and monitoring, where real-time data from smart meters, renewable energy sources, and SoCEV are combined with historical and meteorological data to provide a comprehensive input for energy forecasting. The DL module dynamically predicts energy demand and renewable energy availability, enabling precise resource allocation. Predictions are continuously refined through feedback loops, ensuring adaptive and accurate decision-making. The bidirectional energy flow system employs V2H and H2V technologies to manage energy flow effectively. During peak demand, V2H supplies energy from EVs to households, while H2V charges EVs efficiently during off-peak hours using surplus renewable energy. The blockchain-based P2P energy-trading module facilitates the secure and transparent trading of surplus energy among prosumers through decentralized mechanisms, leveraging smart contracts and an energy-efficient PoS consensus algorithm. The dynamic cost optimization module minimizes energy costs by scheduling non-critical appliances during off-peak hours and prioritizing renewable energy utilization. The system incorporates a continuous feedback mechanism to monitor energy efficiency, cost savings, and grid dependency. This feedback is used to retrain the DL model and optimize scheduling and allocation parameters, creating a robust, scalable, and sustainable energy management solution for modern innovative grid applications (See Figure 3).

3.3. Dynamic Energy Cost Optimization Algorithm (DECO Algorithm)

The DECO algorithm is a cutting-edge approach designed to minimize household energy costs by leveraging dynamic pricing models, renewable energy integration, and bidirectional energy flows through V2H and H2V technologies. By utilizing real-time energy pricing data and predictive deep-learning models, DECO identifies optimal energy usage schedules for critical and non-critical appliances, ensuring maximum efficiency. Renewable energy sources are prioritized, reducing reliance on grid electricity during peak pricing periods. Incorporating V2H and H2V technologies, DECO enables electric vehicles to function as flexible energy assets—supplying energy to the home during high-demand, high-cost periods and recharging efficiently during low-demand, off-peak hours. The algorithm also facilitates secure P2P energy trading via blockchain, allowing households to sell surplus renewable energy, further offsetting costs. This adaptive, data-driven approach continuously refines energy allocation strategies, ensuring the system responds dynamically to pricing fluctuations, energy availability, and user preferences. By combining cost-effective energy scheduling with robust optimization techniques, the DECO algorithm enhances household energy management, reduces overall expenses, and contributes to sustainable smart grid ecosystems (See Figure 4).

The DECO algorithm minimizes energy costs by optimizing consumption schedules through dynamic pricing, renewable energy use, and bidirectional energy flows (V2H and H2V). It is particularly effective for households seeking cost savings without extensive energy trading or scalability. However, its feedback mechanism and adaptability are more limited.

In contrast, adaptive energy management algorithm provides a comprehensive energy management framework. It integrates deep learning for precise demand forecasting, blockchain for secure P2P energy trading, and dynamic bidirectional energy flows. With a robust feedback system, it continuously refines predictions and operations, adapting to real-time conditions and user needs. Beyond cost efficiency, it ensures grid stability and energy optimization, making it suitable for modern, scalable smart grids. While DECO is ideal for cost-focused applications, the adaptive algorithm delivers balanced energy efficiency, scalability, and secure trading solutions, meeting the demands of advanced grid systems.

4. Results and Discussion

4.1. System Performance Test

The IntelliGrid AI system’s performance is optimized using various datasets, including weather data from the Tunisian Meteorological Services. These data help to predict solar power generation more accurately and improve the efficiency of PV systems.

Table 2. Weather data and solar irradiance trends for PV system optimization.

Day	Solar Irradiance (W/m2)	Temperature (°C)	Predicted PV Output (kWh)
Monday	800	30	5.5
Tuesday	850	32	5.8
Wednesday	780	29	5.3
Thursday	900	34	6.0
Friday	850	33	5.7
Saturday	800	31	5.5
Sunday	750	28	5.0

shows comprehensive datasets supporting predictive and real-time energy optimization in an IntelliGrid AI system. The weather data from Tunisia’s meteorological services is crucial for predicting PV system output. Table 2 demonstrates solar irradiance and temperature data over a week, with irradiance values ranging from 750 W/m² to 900 W/m² and temperatures between 28 °C and 34 °C. These parameters influence PV output, which peaks at 6.0 kWh on days with high irradiance. This integration ensures accurate solar energy predictions, enabling the IntelliGrid AI system to manage energy effectively. By integrating these datasets, the system ensures real-time decision-making and optimized energy management. Table 2 shows the primary datasets used in the system. These include weather data for solar power forecasting, home appliance data for efficient scheduling, electric vehicle data for energy flow management, PV system data for energy production forecasting, blockchain transaction data for secure trading, energy-pricing data for cost optimization, and IoT data for real-time insights.

Table 3. Household appliance energy consumption patterns for enhanced scheduling and efficiency.

Appliance	Time of Peak Usage	Energy Consumption (kWh)
HVAC System	Midday	2.0
Lighting	Morning, Evening	0.5–1.5
Kitchen Appliances	Morning, Afternoon	1.0–2.0

shows energy usage patterns in a home. The HVAC system uses the most energy during midday (2.0 kWh), while lighting peaks in the morning and evening (0.5–1.5 kWh). This data helps the system prioritizes critical devices during peak hours, reducing unnecessary energy use.

Table 4. Electric vehicle energy cycles for optimized grid independence.

Time of Day	Charging Status (%)	Energy Usage (kWh)
Early Morning (6–9 AM)	40–70	Charging
Midday (12–2 PM)	100	Idle
Evening (5–9 PM)	40–60	Discharging for household use

shows how EVs charge and discharge. EVs are charged during the early morning using low-cost energy and are fully charged by midday. The stored energy is then used to power homes in the evening, reducing reliance on the grid.

Table 5. Dynamic energy pricing patterns for optimized cost management.

Time of Day	Price (USD/kWh)	Energy Management Actions
Morning (6–9 AM)	0.08–0.10	Prioritize appliance operation
Midday (12–2 PM)	0.15–0.17	Use stored EV energy (V2H)
Evening (5–9 PM)	0.08–0.10	Optimize appliance usage

shows the trends in energy prices. Prices are lower in the morning and evening (USD 0.08–0.10/kWh) and higher during midday (USD 0.15–0.17/kWh). These data help the system schedule non-critical devices during peak hours to save costs. Indeed, the table highlights dynamic electricity pricing trends over 24 h. Prices are lowest during off-peak hours (USD 0.08–USD 0.10/kWh) and peak at midday (USD 0.15–USD 0.17/kWh). These data help IntelliGrid AI schedule non-critical appliances during low-cost hours, maximizing cost savings while maintaining efficient energy use.

Table 6. Performance metrics for blockchain transaction times in energy trading.

Consensus Mechanism	Demand Level	Transaction Time (Seconds)	Description
Proof of Stake (PoS)	Low Demand	3–4	Faster transaction processing with energy-efficient consensus during reduced network activity.
Proof of Stake (PoS)	High Demand	5–6	Moderate transaction times, even with increased network congestion due to efficient validation.
Proof of Work (PoW)	Low Demand	4–5	Relatively slower processing due to computational overhead, even during low network activity.
Proof of Work (PoW)	High Demand	7–8	Longer transaction times due to intensive validation processes and high network congestion.

details blockchain transaction times ranging from 3 to 8 s. Transactions are faster during low energy demand, ensuring energy is traded smoothly and securely. Using the weather data from Tunisia and integrating these datasets, the IntelliGrid AI system effectively reduces costs, improves energy efficiency, and ensures reliable household energy management. This approach makes the system practical and sustainable. Blockchain technology ensures secure energy trading, and the Table compares transaction times for PoW and PoS mechanisms. Under low demand, PoS achieves faster transaction times (3 s) compared to PoW (6 s). PoS transactions take 8 s during high demand, significantly faster than PoW’s 12 s. This efficiency enhances energy-trading reliability, particularly in high-load scenarios.

4.2. Validation of Bidirectional Energy Flow in IntelliGrid AI

The bidirectional energy flow capabilities of IntelliGrid AI exhibit significant adaptability to varying weather conditions, ensuring efficient energy management under both sunny and cloudy scenarios. Figure 5a demonstrates that, on sunny days, EV charging status peaks at 100% during midday, with consistent energy discharges of 2 kWh during evening peaks (5–9 PM), reducing grid dependency by 35%. In contrast, cloudy days show reduced charging status (up to 85%) and lower evening energy discharges (1.5 kWh), highlighting the impact of diminished renewable energy availability.

Table 7. Comparative metrics of V2H and H2V performance under sunny and cloudy conditions.

Metric	Sunny Conditions	Cloudy Conditions
Total Energy Supply (kWh)	450	320
Supply Peaks (Time of Day)	6 AM–9 AM, 5 PM–9 PM	6 AM–9 AM, 5 PM–9 PM
Blockchain Efficiency (%)	98	92
Average Transaction Time (seconds)	3	5
Energy Availability Impact	High	Moderate
Metric	Sunny Conditions	Cloudy Conditions

summarizes key V2H and H2V performance metrics under sunny and cloudy conditions. It highlights peak-charging status, discharging energy, grid dependency reduction, and usage and associated costs. The table emphasizes how sunny conditions enable higher efficiency and cost savings through the better utilization of renewable energy, while cloudy conditions result in reduced performance and higher costs due to limited energy availability.

Table 8. Energy supply and blockchain efficiency metrics for sunny and cloudy days.

Metric	Sunny Conditions	Cloudy Conditions
Peak-Charging Status (%)	100	85
Discharging Energy (kWh)	2	1.5
Grid Dependency Reduction (%)	35	25
Peak Energy Usage (kWh)	2.3	1.5
Charging Cost Range (USD/kWh)	0.10–0.14	0.15–0.18
Cost Reduction (%)	18	10

comprehensively compares the total energy supply and blockchain efficiency metrics for sunny and cloudy days. It shows that sunny conditions facilitate a more significant energy supply, aligning with household demands and enhancing blockchain efficiency for faster and more secure transactions. However, cloudy days exhibit reduced energy availability, impacting transaction performance and increasing reliance on the grid. These findings highlight IntelliGrid AI’s ability to optimize energy flow and blockchain transactions dynamically, ensuring sustainable and efficient energy utilization under varying weather scenarios. Similarly, Figure 5b illustrates that H2V charging schedules leverage surplus renewable energy during sunny mornings (6–9 AM), achieving higher energy usage (up to 2.3 kWh) and lower costs (USD 0.10–USD 0.14 per kWh). On cloudy mornings, energy usage decreases to 1.5 kWh, with costs rising to USD 0.15–USD 0.18 per kWh. The synergy of the V2H and H2V technologies is depicted in Figure 5c, showing superior total energy supply during sunny days, with peaks in the morning and evening aligning with household demands. Cloudy days exhibit a lower total energy supply, necessitating a greater reliance on grid energy. Figure 5d underscores the impact of weather on blockchain transaction efficiency, with sunny days achieving up to 98% efficiency due to abundant renewable energy, compared to 92% on cloudy days.

4.3. Scalable and Adaptive Energy Management

IntelliGrid AI leverages its DECO to adapt energy schedules dynamically, ensuring efficient alignment with renewable energy generation and pricing trends. As illustrated in Figure 6a, EV charging status peaks at 100% during midday on sunny days, with evening discharges of 2 kWh reducing grid dependency by 35%. On cloudy days, EV charging status peaks at 85%, with lower evening discharges of 1.5 kWh, as renewable energy availability decreases. This dynamic scheduling aligns with real-time conditions, ensuring a stable energy supply. The same figure further highlights these metrics, including significant reductions in grid dependency. IntelliGrid AI’s DL model accurately predicts energy demand, with a mean absolute error (MAE) of 3.2%, ensuring adaptive scheduling and optimal resource allocation. Figure 6b demonstrates the DL model’s ability to closely match the predicted energy demand with the actual demand, enabling effective household energy flow management. During sunny days, energy demand peaks are met using renewable energy, while cloudy conditions necessitate increased reliance on grid energy. IntelliGrid AI ensures a resilient energy management framework by integrating blockchain technology for secure transactions. As shown in Figure 6c, the total energy supply peaks at 450 kWh on sunny days, with lower supply levels of 320 kWh on cloudy days. Blockchain efficiency complements this dynamic energy management, achieving 98% efficiency on sunny days and 92% on cloudy days, as illustrated in Figure 6d. Indeed, it summarizes these metrics, including transaction times of 3 s for sunny days and 5 s for overcast conditions. The figures and tables demonstrate IntelliGrid AI’s robust, scalable, and adaptive energy management capabilities. Dynamic scheduling aligns demand with renewable energy availability, while DL-driven predictions ensure resource optimization. Blockchain integration enhances transaction security and efficiency, promoting sustainability and reliability across diverse weather scenarios.

4.4. Comparative Analysis of IntelliGrid AI and Traditional Systems

IntelliGrid AI demonstrates superior performance in energy cost reduction, transaction security, and scalability compared to traditional systems. As shown in Figure 7a, IntelliGrid AI achieves a 20% energy cost reduction on sunny days, capitalizing on peak renewable energy availability during the midday and evening hours. The system maintains a 15% cost reduction on cloudy days, leveraging dynamic scheduling to optimize energy use. In contrast, traditional systems achieve only 10–12% reductions due to limited adaptability and reliance on grid energy. Figure 7b illustrates the blockchain efficiency trends, with IntelliGrid AI reaching 98% efficiency on sunny days and 92% on cloudy days, ensuring secure and transparent energy transactions. In contrast, traditional systems lack integrated blockchain mechanisms, making them prone to inefficiencies and security vulnerabilities. Scalability is another critical advantage of IntelliGrid AI, as depicted in Figure 7c. The system achieves a response time of 3 s on sunny days and 4 s on cloudy days due to its efficient proof-of-stake (PoS) consensus mechanisms. Traditional systems, with a response time of 5 s, struggle with scalability during peak demand periods, highlighting their limitations.

Figure 7d illustrates IntelliGrid AI’s integrated performance across energy cost reduction and blockchain efficiency over 24 h. On sunny days, the combined efficiency peaks above 50% during the midday (12 PM–3 PM) and evening hours (5 PM–9 PM), driven by optimal renewable energy utilization and effective V2H discharges and supported by the blockchain efficiency reaching 98%. Cloudy conditions exhibit slightly reduced efficiency, averaging 40% during peak hours and sustained by H2V optimization and a scaled-down blockchain efficiency of 92%. These trends underscore IntelliGrid AI’s adaptability to varying renewable energy availability, ensuring secure and cost-effective energy management across diverse weather scenarios. This demonstrates its scalability and reliability in delivering consistent performance for real-world applications.

Table 9. Comparative metrics of IntelliGrid AI and traditional systems.

Metric	IntelliGrid AI (Sunny)	IntelliGrid AI (Cloudy)	Traditional Systems
Energy Cost Reduction (%)	20%	15%	10–12%
Blockchain Efficiency (%)	98%	92%	N/A
Transaction Time (Seconds)	3	4	5

comprehensively summarizes the comparative metrics, showcasing IntelliGrid AI’s clear advantages. These include a 20% energy cost reduction, up to 98% blockchain efficiency, and significantly lower response times than traditional systems. These findings underscore IntelliGrid AI’s ability to deliver secure, scalable, and cost-effective energy management.

4.5. Comprehensive Analysis of IntelliGrid AI: Cost Savings, Renewable Integration, and Blockchain Efficiency in Real-Time Applications

IntelliGrid AI demonstrated exceptional performance in case studies conducted in Tunisia, achieving a 20% reduction in energy costs on sunny days and 15% on cloudy days, as shown in Figure 8a, by leveraging dynamic energy pricing and efficient V2H discharges during peak hours. Renewable energy integration was optimized with PV systems generating up to 6.0 kWh on sunny days and 4.5 kWh on cloudy days, ensuring that surplus energy was utilized or stored, as depicted in Figure 8b. The system reduced grid dependency by up to 40% on sunny days and 30% on cloudy days, as illustrated in Figure 8c, with EV discharges peaking at 2 kWh during evening demand. Blockchain efficiency, shown in Figure 8d, remained high at 98% on sunny days and 92% on cloudy days, with transaction times averaging 3–5 s, ensuring secure and scalable energy trading.

Table 10. Comparative energy performance metrics under sunny and cloudy conditions.

Metric	Sunny Days	Cloudy Days
Energy Cost Reduction (%)	20	15
PV Energy Output (kWh)	6.0	4.5
Grid Dependency Reduction (%)	40	30

highlights key metrics, including energy cost reduction, PV energy output, and grid dependency reduction.

4.6. Discussion

IntelliGrid AI represents a transformative advancement in home energy management systems by integrating blockchain technology, deep learning (DL), and bidirectional energy flows (V2H and H2V). The system’s robust architecture addresses key challenges in energy optimization, transaction security, and scalability, as validated through comprehensive case studies in Tunisia. These studies highlight significant improvements in cost efficiency, renewable energy utilization, and grid independence.

4.6.1. Cost Savings and Energy Optimization

IntelliGrid AI achieved a 20% reduction in energy costs on sunny days and 15% on cloudy days, as outlined in Table 10. This is attributed to the system’s dynamic energy allocation, prioritizing renewable sources and V2H discharges during peak demand periods. The integration of DL models ensured accurate energy demand prediction, with a mean absolute error (MAE) of 3.2%, facilitating optimal resource allocation. For example, EVs supplied 2 kWh of energy to households during evening peaks, reducing grid dependency by up to 40% on sunny days and 30% on cloudy days.

4.6.2. Renewable Integration and Efficiency

PV systems demonstrated high efficiency, generating up to 6.0 kWh on sunny days, with slight reductions to 4.5 kWh under cloudy conditions. The synergy between V2H and H2V flows ensured the efficient utilization of renewable energy, minimizing waste and aligning supply with household demands. Figure 8b illustrates the system’s ability to maintain stable energy outputs across varying weather conditions, highlighting its adaptability.

4.6.3. Blockchain Security and Transaction Efficiency

Integrating blockchain technology ensured secure, transparent peer-to-peer (P2P) energy trading. Transaction times averaged 3 s on sunny days and 5 s on cloudy days, supported by a high blockchain efficiency of 98% and 92%, respectively (Table 10). These metrics underscore the system’s capability to maintain reliable operations even under fluctuating energy availability.

4.6.4. Scalability and Adaptability

The scalability of IntelliGrid AI is evident in its ability to dynamically adjust energy schedules based on real-time data inputs, including weather forecasts and energy pricing. Table 9 compares IntelliGrid AI to traditional systems, demonstrating its cost savings, blockchain efficiency, and transaction speed superiority. For instance, IntelliGrid AI achieved a 20% cost reduction compared to the 10–12% reductions typically observed in traditional systems.

4.6.5. Future Directions

While IntelliGrid AI shows immense potential, areas for further research include improving blockchain scalability, optimizing V2H and H2V infrastructure, and enhancing interoperability among decentralized energy networks. Future implementations could leverage predictive seasonal models and advanced consensus mechanisms like proof of stake (PoS) to streamline energy transactions further and reduce computational overhead. IntelliGrid AI provides a scalable, adaptive, and sustainable framework for modern energy management by integrating advanced algorithms, secure energy trading, and renewable energy optimization. Its application in Tunisia is a model for broader adoption in smart cities and renewable-powered communities worldwide.

5. Conclusions and Future Works

The development of IntelliGrid AI represents a significant leap forward in smart home energy management, combining blockchain technology, deep learning, and two-way energy flows (V2H and H2V) to achieve scalable, secure, and cost-effective energy systems. Case studies conducted in Tunisia demonstrated the system’s effectiveness, showing a 20% reduction in energy costs on sunny days and 15% on cloudy days. Renewable integration was improved, with photovoltaic systems generating up to 6.0 kWh on sunny days and 4.5 kWh on overcast days. Grid dependency was reduced by 40% during peak periods, demonstrating the system’s ability to enhance energy resilience and promote sustainability. Blockchain technology achieved 98% and 92% transaction efficiencies on sunny and cloudy days, respectively, with average transaction times of 3 to 5 s, ensuring secure and reliable peer-to-peer energy trading. Two innovative algorithms have been key to IntelliGrid AI’s success. The DECO algorithm reduces household energy costs by leveraging dynamic pricing, renewable energy integration, and bidirectional power flows. It schedules critical appliances during low-cost hours and prioritizes renewable energy and V2H offloading during peak periods, achieving significant cost savings. As for the adaptive power management algorithm, this comprehensive framework integrates deep learning for accurate energy demand forecasting, the blockchain for secure peer-to-peer trading, and bidirectional power flow technologies. It continually improves predictions and dynamically optimizes power allocation, ensuring grid stability, reducing costs, and enhancing scalability.

Despite its promising performance, IntelliGrid AI has room for improvement. Future research will address blockchain’s scalability challenges by exploring advanced consensus mechanisms, such as PoS. Priority will also be given to improving the V2H and H2V infrastructure to maximize energy transmission efficiency and expand compatibility with diverse EV models. Additionally, integrating predictive seasonal models will enable more accurate energy demand forecasting and adaptive scheduling, improving system performance under changing climate conditions. Expanding the system’s application to industrial and commercial environments will also be another focus, replicating the cost-saving benefits and renewable energy integration on a larger scale. By addressing these challenges and building on its strengths, IntelliGrid AI is poised to become a foundational technology for developing sustainable energy systems, enabling smart cities powered by renewable energy, and contributing to a greener and more resilient energy future.