Supporting Sustainable Development Goals with Second-Life Electric Vehicle Battery: A Case Study

Abstract

To alleviate the impact of economic and environmental detriments caused by the increased demands of electric vehicle battery production and disposal, the use of spent batteries in second-life stationary applications such as energy storage for renewable sources or backup power systems, offers many benefits. This paper focuses on reducing the energy consumption cost and greenhouse gas emissions of Internet-of-Things-enabled campus microgrids by installing solar photovoltaic panels on rooftops alongside energy storage systems that leverage second-life batteries, a gas-fired campus power plant, and a wind turbine while considering the potential loads of a prosumer microgrid. A linear optimization problem is derived from the system by scheduling energy exchanges with the Ontario grid through net metering and solved by using Python 3.11. The aim of this work is to support Sustainable Development Goals, namely 7 (Affordable and Clean Energy), 11 (Sustainable Cities and Communities), 12 (Responsible Consumption and Production), and 13 (Climate Action). A comparison between a base case scenario and the results achieved with the proposed scenarios shows a significant reduction in electricity cost and greenhouse gas emissions and an increase in self-consumption rate and renewable fraction. This research work provides valuable insights and guidelines to policymakers.

1. Introduction

Greenhouse gas (GHG) emissions are a significant contributor to global environmental challenges, with the energy production and transportation sectors being among the primary sources of pollution (Dawoud et al., 2024) [1]. The combustion of fossil fuels, including natural gas, crude oil, and biomass, continues to drive harmful emissions. Simultaneously, the global demand for energy is steadily increasing, exacerbating environmental concerns. In this context, the integration of renewable energy sources presents a viable and sustainable solution (Hosseini et al., 2024) [2]. By shifting towards cleaner energy alternatives, such as solar and wind, we can mitigate the adverse environmental impacts of conventional energy generation, thereby safeguarding both the environment and the future of the planet. The Sustainable Development Goals (SDGs) are a collection of 17 goals and 169 targets agreed upon by the United Nations to position global development on a path of greater prosperity. Of the many topics covered, energy is a major driver for progress in these goals. We have now seen research that integrates SDG consideration into energy system models.

Microgrids offer a robust solution for enhancing energy efficiency, reliability, and addressing key techno-economic and environmental challenges. One of the primary reasons for utilizing second-life batteries (SLBs) is their significant environmental advantage. Manufacturing a new 1 kWh battery consumes approximately 60–65 kWh of energy, while emitting around 55 kg of carbon dioxide (Musa et al., 2023) [3]. By repurposing spent batteries for second-life applications, we can drastically reduce the need for energy-intensive production and associated emissions, contributing to a more sustainable energy future. So, in this way, SLBs are an optimal and environmentally friendly solution for stationary applications.

A prosumer is an agent that can both consume and produce clean energy, of which any surplus energy produced and sold to the national grid will eventually help alleviate the additional demand of energy needed to power the gradually and surely increasing number of EVs. On the other hand, the key role of a prosumer agent is to conserve energy during off-peak times when tariffs are low and export energy during peak times when tariffs are high. This aligns with a focus on economic efficiency to support SDGs. This research adopts a techno-econo-environmental (TEE) approach, directly contributing to the advancement of SDGs 7, 11, 12, and 13, ensuring that the work is both economically viable and environmentally sustainable.

It is estimated that lithium-ion batteries (LIBs) used in EVs have a lifespan of 8 years, but an additional 10 years can be added to their lifespan by using them in stationary second-life applications (Walker et al., 2015) [4]. An estimation given by Akram and Abdul-Kader (2021) [5] says that these batteries can be used in EVs for up to 12.5 years. Casals et al. (2019) [6] assess and model the lifespan of the second life of EV batteries. The model considered repurposing the spent batteries in applications such as self-consumption, EV fast charging, area regulation, and transmission upgrade deferral. Their results suggest that these batteries could operate for approximately 12 years in self-consumption, 6 years in area regulation, up to 30 years in EV fast charging stations, and 12 years in transmission upgrade deferral applications.

A common scenario in which such spent LIBs are then used is in energy storage systems (ESSs). Hu et al. (2022) [7] present a comprehensive analysis of the technicalities involved with an ESS based on spent batteries. They highlight that, for such an ESS to be safe and viable, their specifications and requirements should be stricter than an ESS using fresh LIBs. That being said, they are far cheaper than their new counterparts and are thus good candidates for applications that demand less power and must stay within cost constraints. Two such scenarios are: within a framework that substitutes grid energy for renewable energy (RE), where the ESS is tasked with keeping output smooth to counterbalance the variability of renewable resources, and within remote locations that cannot feasibly connect to the grid without significant costs, where the ESS would then store RE in bulk for use when those resources are not available. This is a relatively new technology that can be widely applicable and beneficial to the economy and the environment. Tehrani et al. (2022) [8] design a simulation for a hybrid energy farm that considers various factors and incorporates a multiphysics real-time approach. They mention how such models can be highly sought after by planning authorities and companies looking to move towards RE and smart cities. This is a concrete example of where an SLB ESS could benefit the model, since it can provide smoother energy output while remaining cost-effective, which incentivizes planners, companies, and policymakers to investigate and promote such systems.

In a prosumer system, batteries play a critical role in energy partitioning. For example, an EV can operate in vehicle to grid and grid to vehicle situations. The electricity cost of prosumers can be reduced using optimal adoption of the aforementioned ESS. Additionally, the life of a battery is particularly important for optimal operation as the battery is dependent on several factors such as constant current, constant voltage, internal resistance, frequency of charging/discharging usage, its surrounding temperature and depth of discharge (DOD) (Bordin et al., 2017) [9]. An ESS has been used in conjunction with RE to store surplus renewable energy and the stored energy can be consumed to support the national grid during peak periods (Gulzar et al., 2022) [10]. Figure 1 shows the life cycle of a battery and its various repurposed applications.

Moreover, the aim of this research is to increase/promote knowledge regarding the utilization of SLBs from EVs in stationary applications by considering some of the changes these batteries undergo and assess their effects on sustainable development. The minimization of energy costs and GHG emissions through battery use in stationary applications would progress the attainment of SDGs 7, 11, 12, and 13.

The remainder of this paper is organized as follows: Section 2 presents the literature review. Section 3 provides the proposed system architecture, Section 4 covers the modeling approach, Section 5 shows the analysis of various scenarios and their related results, environmental benefits, and contribution to SDGs. Section 6 presents the sensitivity analysis, and finally, the conclusions and directions for future research are presented.

2. Literature Review

Recent technologies that consider EV batteries for applications other than in-vehicle use have been on the rise (Terkes et al., 2023) [11]. Fernández and Campo (2025) [12] present a case study focused on the Spanish market, examining electricity savings for end consumers due to SLB use, along with the projected extended second-life estimation for a battery pack. Their findings indicate that, given the current tariff structures, a smart energy storage system (SESS) comprising repurposed batteries could emerge as a viable energy use component for households within a span of fourteen years. Kang et al. (2025) [13] quantitatively address the environmental impacts, specifically carbon dioxide (CO₂) emissions, across the lifetime of an ESS that utilizes used batteries as opposed to new batteries. Their findings indicate that, when reused batteries are employed, annual carbon emissions are reduced by 2.8% as a lower bound and 18.5% as an upper bound, depending on the specific purposes for which the batteries are used. Das (2025) [14] investigates the technical challenges and safety issues associated with battery degradation and the advancement of effective repurposing methods. Additionally, they address regulatory and economic obstacles, which encompass the establishment of standards for battery reuse, the infrastructure for recycling, and the dynamics of the market. Spindlegger et al. (2025) [15] investigate the impact of various repurposing strategies on the environmental performance of SLB energy storage systems. Kebir et al. (2023) [16] address the potential of using SLB and solar PVs to deliver inexpensive energy to primary schools in Kenya. Sharma et al. (2022) [17] provide an analysis of a utility grid that is connected to solar, wind, and hydroelectric systems. Al-Alawi et al. (2022) [18] review and assess the latest modeling and experimental reports that considered cyclic life applications for EV batteries. Motjoadi et al. (2022) [19] look at how RE sources like solar and wind can be used with battery systems to help power the grid during power outages and interruptions. Bhatt et al. (2022) [20] utilize the HOMER Pro software version 3.14.2 to evaluate the cost savings associated with the use of SLBs in the grid. Elmorshedy et al. (2021) [21] present a techno-economic design and dynamic rules-based power management system for a solar–wind hybrid off-grid RE system. Fallah and Fitzpatrick (2022) [22] study the financial viability of life applications in EV batteries. Steckel et al. (2021) [23] present an approximate levelized cost for lifetime storage of battery storage systems considering spent and new batteries. Schulz-Mönninghof et al. (2021) [24] suggest a new life cycle analysis (LCA) framework that manufacturers can use to evaluate various instances of LIB reutilization. Husein and Chung (2018) [25] introduced financial aspects and technical feasibility of the campus microgrid in the case of Seoul National University, South Korea. Gao et al. (2018) [26] propose a microgrid control scheme with respect to system constraints to efficiently schedule energy storage. Rodríguez-Gallegos et al. (2018) [27] introduce an optimization model for a PV-storage diesel distribution generation system. Ahmad and Alam (2018) [28] show a hybrid system for Aligarh University in India. Bobba et al. (2018) [29] evaluated the environmental benefits of repurposing batteries for second-life applications for homes powered by photovoltaic self-consumption and grid systems. Ahmadi et al. (2014) [30] address the reuse of EV batteries at the end of their lives in stationary applications.

Table 1. Summary of the reviewed research works.

Author	Objective	Main Components	Optimization Methods Used	Software Used	Types of Load/Regions
Mahdi et al., 2024 [31]	Minimize the energy cost and emission	Wind, solar PV, ESS, utility grid	Non-dominated genetic algorithm II (NSGA-II)	MATLAB R2023a	Residential load
Hussain et al., 2024 [32]	Maximize the facility to exchange the surplus energy within the community for energy trading	Solar PV, ESS, utility grid	Alternating direction method of multipliers (ADMM)	Simulation	Prosumer residential community load
Meng et al., 2024 [33]	Effective energy management of charging stations is proposed	EV, wind turbine, solar PV, BESS	MILP	Simulation	China
Lieskoski et al., 2024 [34]	Potential analysis of SLBs in Finland	SLB and repurposing	Tesla Model S/X material flow analysis	MFA Tools	Ostrobothnia region, Finland
Abo-zahhad et al., 2024 [35]	Feasibility analysis of solar PV indesert weather conditions	Solar PV, microgrids	Multi-criteria decision making	PVsyst ANSYS	Egypt, remote region
Wrålsen and Born, 2023 [36]	Use the life cycle assessment (LCA) of batteries to check the circular economy and GHG benefits	Battery reuse and repurposing LIBs	Inventory model	SimaPro 9.3.0.2	Norway-based analysis
Terkes et al., 2023 [11]	Feasibility analysis, especially considering the uncertainties in SLB usage	SLB, GHG, PV, ESS degradation	Arrhenius-based sub-models	HOMER Pro 3.14.2	Hybrid power system in Türkiye
Musa et al. 2023 [3]	Shared energy storage systems are used for self-consumption and GHG benefits	ESS, solar PV, GHG, economic analysis, sizing, and feasibility	Simulation-based analysis	HOMER Pro 3.14.2	Community prosumer microgrid
Colarullo and Thakur, 2022 [37]	Self-consumption, load shifting, and based on PV and SLB-ESS	Solar PV, ESS, EV, used meter. Front and back-end services	Simulation-based analysis	Tools for techno-economic analysis	Local energy community
Obrecht, 2022 [38]	Circular-economy-based solution using SLBs	ESS, solar PV	Forecasting analysis	Exponential triple smoothing	System in Slovenia
Schulz-Mönninghof et al. (2021) [24]	Repurposing LIBs forLCA and circular economy model	UPS, PV, ESS	Cost and revenue analysis	TOP energy	Industrial load (Germany)
Kamath et al., 2020a [39]	Fast charging infrastructure and optimal usage of SLBs	Life cycle assessment, ESS, global warming potential	Levelized cost of energy and sensitivity analysis	LCOE method	USA-based cities were considered
Kamath et al., 2020b [40]	Analysis ofnew and SLBs to check the effect in various scenarios, i.e., utility peak shaving, residential application, and PV farming	Solar PV, new and used LIBs	Simulation	HOMER Pro 3.14.2	Residential load
Cusenza et al., 2019 [41]	Reused battery impact on the application of SLBs in an institutional building considering the system constraints	PV, ESS	Load match analysis	Simulation	Building load
		Solar PV, ESS, utility grid, degradation	Linear programming	MATLAB R2019a	Ontario
Sofia Gonçalves, 2018 [42]	Minimize energy cost using V2B and scheduling	ESS, EV, solar PV	MILP	MATLAB /GAMS	Communal load (Swedish)
Richa et al., 2017 [43]	Comparison of SLBs of LIBs and new PbA batteries and impacts on the environment	PbA, LIBs GHG	Eco invent database	SimaPro 8.5.2	USA

summarizes the reviewed works.

Table 1 shows that existing works consider various components that were solved by different platforms and software packages. In these studies, many important parameters such as SDGs (7, 11, 12, and 13), SLBs, ultra-low overnight (ULO) tariff, time-of-use (TOU), renewable fraction (RF), self-consumption rate (SCR), and GHG emission are not collectively addressed, which we will incorporate in our model. In this research work, the microgrid of the University of Windsor in Canada is considered so that the most efficient energy exchange between the Centre for Engineering Innovation (CEI) building and the Ontario power grid can be achieved. The campus covers 51 contiguous hectares (130 acres) and is surrounded by a residential neighborhood. The university has nine major buildings. It had a gas-fired power plant (PP) installed in the mid-1990s that generates 4 MW of combined heat and power (CHP); see Conservation and Demand Management Plan, 2018 [44]. The primary contributions of this study are outlined below:

• This research introduces an innovative techno-econo-environmental (TEE) framework for energy scheduling within IoT-enabled campus microgrids, integrating SLBs alongside RE sources like solar PV and wind turbines.
• The model addresses energy costs, GHG emissions, and long-term sustainability, contributing to SDGs 7, 11, 12, and 13.
• A detailed cost analysis of SLBs is performed by using an optimization model, providing insights into the economic feasibility of repurposing EV batteries for stationary applications.
• The proposed model optimizes energy exchanges with the Ontario grid using ultra-low overnight (ULO) price schemes, employing net metering to ensure cost-effective and environmentally friendly energy management.
• The research assesses the effects of different SLB configurations on operational costs and GHG emissions, offering valuable insights into the cost-effectiveness and environmental advantages of various energy storage setups.
• Through comparative analysis with a baseline scenario, the proposed model shows significant reductions in electricity costs and GHG emissions, while improving the self-consumption rate and RF of the campus microgrid, promoting sustainable energy practices.
• A sensitivity analysis is performed to check costs and GHG emissions by varying several parameters of the system, including energy source capacities, LCOS, and other energy costs.
• A demand response strategy is presented to see the impact on costs by considering both ULO and TOU Canadian tariffs for the sensitivity analysis.

These contributions provide deeper insights into the integration of SLBs for optimizing microgrid energy and presenting a practical solution for sustainable energy management.

3. Energy System Architecture

There have been significant improvements in operational costs, installation costs, and load automation through changes in the current energy system brought about by novel power system technologies. At the core of the smart grid lies the energy management system (EMS), which integrates cutting-edge devices and sensors to optimize energy flow and enhance overall grid efficiency. For the energy system, it is expected that customers will have their role expanded beyond simply being buyers of electricity. Optimizing local distributed generation (DG) and their efficient deployment has the potential to significantly enhance the energy system performance and decrease operational expenses.

This study uses the University of Windsor’s campus as a pilot site to assess the model for a smart prosumer connected to the grid. The prosumer smart microgrid may provide year-round solar energy by employing the model that uses photovoltaic (PV) and wind turbine output power as renewable energy sources (RESs) and second-life EV battery integration, while the solar PV is the primary source in the daytime. For data communication in the building, supervisory control and data acquisition (SCADA) and local area network (LAN) systems are used along with Internet of Things (IoT)-enabled devices with sensors. The data stored on the server controls the whole system. The IoT-based devices are used to communicate with the various components of the building for energy management. There are three layers in our model’s prosumer control module. Layer 1 receives data and, through communication protocols, sends information to the net meter (i.e., it sends surplus energy from proprietary grids to the Ontario or utilities grid). Layer 2 determines how to use available resources efficiently via proposed programming. Layer 3 manages the various kinds of distributed generation and loads with respect to instructions in a dispatch signal from Layer 2. For this study, the Centre for Engineering Innovation (CEI) building at the University of Windsor, which covers 300,000 ft² (over 91,000 m²) of dedicated space (Bed Rock for a Bright Future) [45], is taken as a case study. Panday and Channi (2020) [46] state that the area requirement to generate 1 KW of power through a solar PV system is 100 ft² (shadow-free area). Using this data, it can be determined that approximately 3 MW of energy can be produced from the solar PV panels spanning the top of the CEI building, while the wind turbine maximum rating power is 2 MW. We set our maximum storage capacity to 2000 kWh with the intention of collecting and selling excess generated energy (from the mentioned renewable resources) to the grid or using it locally for partial self-sufficiency and supplementing the current stability of the power supply. For battery reliability, instead of considering 100% charge or 0% discharge, the states of charge of 85% maximum and 15% minimum are considered. Finally, we take the maximum power of the grid to be 50,000 kW. Figure 2 below shows the proposed energy system architecture.

Table 2. Electricity prices in 2025 as per ULO (Source: Ontario Energy Board) [47].

ULO Period (Summer)	ULO Price Period	ULO Prices (¢/kWh)
ULO	11 p.m. to 7 a.m.	2.8
Mid-Peak	7 a.m. to 4 p.m. 9 p.m. to 11 p.m.	12.2
On-Peak	4 p.m. to 9 p.m.	28.6

below shows the ULO pricing of the Ontario utility grid. The Independent Electricity Supply Operator (IESO) is responsible for monitoring energy requirements in the province of Ontario incessantly. The IESO is also responsible for balancing the supply and demand of the system and directing the distribution of energy across the system in the event of excess demand. Following that,

Table 3. Electricity prices in 2025 as per TOU (Source: Ontario Energy Board) [47].

TOU Period (Summer)	TOU Price Period	TOU Prices (¢/kWh)
Off-Peak	7 p.m. to 7 a.m.	7.6
Mid-Peak	7 a.m. to 11 a.m. and 5 p.m. to 7 p.m.	12.2
On-Peak	11 a.m. to 5 p.m.	15.8

shows the time-of-use (TOU) pricing of the Ontario grid. These values are relevant for when we conduct a sensitivity analysis.

Given the system architecture and energy prices, the following section proposes a solution approach to decrease the energy costs as well as the GHG emission.

4. Mathematical Modeling

Prior to presenting the mathematical model, a flowchart of the tasks performed in the methodology is shown below in Figure 3. The list of abbreviations and notations of constants and variables are presented in the Abbreviations sections.

4.1. Objective Function

A linear optimization problem is formulated with the objective function to minimize the cost of daily energy consumption by tackling the scheduling of power sources and addressing the environmental aspects.

$$\min {\mathrm{C}}_{\mathrm{o}\mathrm{p}}=\sum _{\mathrm{t}=1}^{\mathrm{T}=24}\left[{(\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}){\mathsf{\lambda }}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d}}+({\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v},\mathrm{t}})\times {\mathsf{\lambda }}_{\mathrm{p}\mathrm{v}}+{(\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}})\times {\mathsf{\lambda }}_{\mathrm{w}\mathrm{d}}{+(\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}){\mathsf{\lambda }}_{\mathrm{e}\mathrm{s}\mathrm{s}}+({\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}})\times {\mathsf{\lambda }}_{\mathrm{p}\mathrm{p}}\right]$$

(1)

Equation (1) is the objective function considering the various sources of energy such as utility grid ${\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}$, solar PV output power ${\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v},\mathrm{t}}$, power of battery ESS ${\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}$, power of wind turbine ${\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}}$, and power plant ${\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}$. The associated values of λ are the operational costs (or levelized cost of energy) of each component during the operation of the whole day. The total duration is the 24 h scheduling for the available resources. In this model, a stationary SLB storage system is considered. The constraints related to this LP model are presented with detailed explanations in the following paragraphs.

4.2. Power Balance Constraints

This scenario relates to the energy exchange through the grid and is known as a two-way energy flow prosumer-based system. ${\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}$ is the total load of the building, which is calculated by adding the power from the utility grid, solar PV, ESS, wind turbine, and power plant (PP), as described in Equation (2) below. It is also known as the economic dispatch constraint.

$${\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}={\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}+{\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r},\mathrm{t}}+{\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}+{\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}}+{\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}$$

(2)

4.3. Limitation Constraints

The limitations and the range of the other parameters are expressed in Equations (3)–(6).

$${{\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}\le {{\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(3)

$${{\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v},\mathrm{t}}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v},\mathrm{t}}\le {{\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v},\mathrm{t}}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(4)

$${{\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}}\le {{\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(5)

$${{\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}\le {{\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(6)

The minimum and maximum of the above-mentioned parameters control the usage of the available resources.

4.4. Energy Storage System Output Constraints

An energy storage system (ESS) is a necessary component of an energy management system (EMS). In the event of grid outage, an ESS can help maintain power supply. For this research work, SLB of EVs are selected to store the surplus energy. So, an ESS can be interchangeably labeled a BESS in this paper. In Equation (7) below, a reserve factor R_t is introduced. It accounts for the need to maintain a certain reserve level of power in the BESS. This reserve factor provides the facility for emergency operations, i.e., sudden increases in energy demand, or grid outage. It plays the role of buffer capacity for the storage system to avoid complete depletion. It depends on the requirement of reliability. In the case of fluctuations in demand as well as variability in grid reliability, the system reserves a significant margin of its total capacity. The demand factor D_t in Equation (8) ensures the controlled lower bound of the energy storage for the operational strategies and depends on future demands. It provides support to the campus grid retaining sufficient energy in the battery during periods of low renewable generation.

$$\left(\mathrm{s}\mathrm{o}{\mathrm{c}}_{\left(\mathrm{t}-1\right)}\left(1-{\varnothing }_{\mathrm{e}\mathrm{s}\mathrm{s}}\right)-\mathrm{s}\mathrm{o}{\mathrm{c}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\right)\times \left({\displaystyle \frac{{\mathrm{C}\mathrm{a}\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s}}}{100}}\right)+{\mathrm{R}}_{\mathrm{t}}\le {\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}$$

(7)

$$\left(\mathrm{s}\mathrm{o}{\mathrm{c}}_{\left(\mathrm{t}-1\right)}\left(1-{\varnothing }_{\mathrm{e}\mathrm{s}\mathrm{s}}\right)-\mathrm{s}\mathrm{o}{\mathrm{c}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\right)\times \left({\displaystyle \frac{{\mathrm{C}\mathrm{a}\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s}}}{100}}\right)-{\mathrm{D}}_{\mathrm{t}}\le {\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}$$

(8)

The available energy in the battery is based on the state of charge (SOC) at that time. The above inequalities, (7) and (8), are used to express the battery parameters, which control the output power of the ESS (Choi and Min, 2018) [48]. For the lower limit $\mathrm{S}\mathrm{O}{\mathrm{C}}_{\mathrm{m}\mathrm{i}\mathrm{n}}$ and upper limit $\mathrm{S}\mathrm{O}{\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$ of the battery state of charge, 15% and 85% are taken, respectively. The power of the storage battery ${\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}$ is also related to other parameters such as the current SOC at time “t” and its storage capacity. The rated ESS capacity is illustrated as ${\mathrm{C}\mathrm{a}\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s}}$(kWh). Equation (9) expresses the SOC level of the ESS output power at any time during the whole day (Choi and Min, 2018) [48]. Also, ${\varnothing }_{ess}$ is the self-discharge rate of the battery, namely 0.01 for both new and second-life batteries, which reduces the SOC over time even when the battery is not in use (Seong et al., 2018) [49].

$${\mathrm{S}\mathrm{O}\mathrm{C}}_t={\mathsf{\alpha }\mathrm{S}\mathrm{O}\mathrm{C}}_{t-1}-\left({\displaystyle \frac{{\mathrm{P}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}\times \left(1-\left(\mu \times {\varnothing }_{ess}\right) \right)}{{\mathrm{C}\mathrm{a}\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s}}}}\right)\times 100$$

(9)

The battery’s SOC can be determined at any time as shown in Equation (10). To plan for optimal operation, the battery needs to be charged for the next working day. For example, the next day begins from a fixed state of charge. In this research, ${\mathrm{S}\mathrm{O}\mathrm{C}}_0=$ 25% is considered. The charge at the start and the charge at the end of the scheduling horizon (i.e., one day or twenty-four hours) are assumed to be the same as in Equation (10).

$$\mathrm{S}\mathrm{O}\mathrm{C}\left(\mathrm{t}=24\right)=\mathrm{S}\mathrm{O}\mathrm{C}\left(\mathrm{t}=0\right)$$

(10)

In the context of a campus microgrid, various storage technologies can be used.

Table 4. Comparison of SLBs and other storage technologies.

Energy Storage Technology	Life in Years	Cycles	Efficiency (%)	Cost	Environmental Impact	Source
EV Second-Life Batteries	6–10	2000	70–80	Low	$\mathrm{Low} {\mathrm{C}\mathrm{O}}_2$ Emission	Casals et al., (2019) [6]
New LIBs	15	2500	77–85	High	$\mathrm{High} {\mathrm{C}\mathrm{O}}_2$ Emission	May et al., 2018 [50]
Fly Wheel	20	Unlimited	70–80	High	$\mathrm{Low} {\mathrm{C}\mathrm{O}}_2$ Emission	Aquino et al., (2017) [51]
Compressed Air Energy Storage	25	10,000	65	High	$\mathrm{Low} {\mathrm{C}\mathrm{O}}_2$ Emission	May et al., 2018 [50]
Redox Flow Batteries	10- 20		80	High	Environmentally Friendly	Mongird et al., (2019) [52]
Lead–Acid Batteries	15	2000	79–84	Low	Environmentally Friendly	May et al., (2018) [50]

presents a comprehensive comparison of a variety of such storage technologies based on the parameters we are focusing on in this study.

The focus of our paper is repurposed EV batteries, because of their low cost, long lifespan, and low environmental impact. While the technologies presented in Table 4 are viable options in a microgrid, we find that EV SLBs can optimize many of these factors in low-budget constraints, and they can prioritize environmental benefits and circular economy objectives.

We assume identical lithium-ion-based SLBs with a rated capacity of 64 kWh. With this, the modeling framework stays consistent, and campus microgrid storage performance is simplified. As such, we can assess system behavior in isolation, without concern about the complexities of variabilities present in various SLBs.

Table 5. Battery parameters (Mongird et al. 2019 [52]; Casals et al., 2019 [6]; Neigum and Wang 2024) [53].

Parameter	Value	Unit/Remarks
Battery Type	Lithium Ion	-
Rated Capacity	64	kWh
Nominal Voltage	400	V
Depth of Discharge (DOD)	80	%
State of Health (SOH)	80	%
Round-trip Efficiency	80	%
Maximum Charge/Discharge Rate	0.5 C	Relative to Rated Capacity
Remaining Cycle Life	~2000	Cycles
Battery Configuration	Identical	Consistent across Simulation
Management System	Passive	Assumed in Modeling

shows the specifications of our assumed SLBs.

Additionally, in this study, SLBs are modeled with consideration of their distinct characteristics compared to new LIBs.

Table 6. Comparison of new battery and SLB efficiency.

Parameter	Li-ion Battery (New)	Second-Life Battery (SLB)	Source
Chemistry	Li-ion	Li-ion	Same for both
Rated Capacity	64 kWh	64 kWh	Assumed
SOH	~100%	~80%	(Neigum and Wang 2024) [53]
Round-trip Efficiency Loss	0.5%/year	0.5%/year	(Mongird et al., 2019) [52]
Aging/1000 Cycles	6–12%	5%	(Gao et al., 2024) [54]
Aging/Year	1.5–2.5%	1.5%	(Gao et al., 2024) [54]
Impedance (mΩ)	1 mΩ	1.5–3.0 mΩ	(Gao et al., 2024) [54]
Aging Cycle	1500–2500	2000	(Gao et al., 2024) [54]
Service Time	10–15 years	11 years	(Gao et al., 2024) [54]

presents a comparative analysis of key characteristics between SLBs and new LIBs.

Inequality (11) is used to lemmatize the gradient of storage power. The constraints of the maximum and minimum energy storage power also control the battery to fully charge it and discharge it.

$${\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},{\mathrm{t}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}\le {\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}\le {\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},{\mathrm{t}}_{\mathrm{m}\mathrm{a}\mathrm{x}}}$$

(11)

Overcharging is harmful for the battery storage system, so there is a need to control charging and discharging processes, so the limits are expressed in Equations (12) and (13).

$${\mathrm{C}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}\le {\mathrm{C}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(12)

$${\mathrm{D}\mathrm{i}\mathrm{s}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{e}\mathrm{s}\mathrm{s},\mathrm{t}}\le {\mathrm{D}\mathrm{i}\mathrm{s}\mathrm{c}\mathrm{h}\mathrm{a}\mathrm{r}\mathrm{g}\mathrm{e}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(13)

Performance Degradation Modeling for SLBs

Degradation is the process by which all LIBs decrease in lifespan. Every charge and discharge deteriorate battery components through an SEI layer that grows on the electrode surfaces, which increases internal resistance, thus hastening the overall decrease in capacity. This degradation in turn affects a battery’s energy efficiency, power, and capacity (Ramoni et al., 2017) [55]. The values given in

Table 7. Second-life battery specification.

Quantity	Value	Source
Initial Capacity/SOH of SLB	80%	(Neigum and Wang 2024) [53]
Capacity Fade per Year	1.8%	(Argue, 2025) [56]
Round-trip Efficiency	80%	(Casals et al., 2019) [6]
Round-trip Efficiency Loss	0.5%	(Mongird et al. 2019) [52]
Number of Cycles of SLB	2000 cycles	(Gao et al., 2024) [54]
Cycle Loss per Year	1.5%	(Gao et al., 2024) [54]

are taken from the literature on battery lifespans, and we use these to calculate trends in battery degradation, presented in

Table 8. Analysis of degradation of SLB capacity and efficiency.

Year	Capacity (%)	Number of Cycles (Years)	Round-Trip Efficiency (%)
0	80.0	2000	80
1	78.2	1970	79.5
2	76.4	1940	79
3	74.6	1910	78.5
4	72.8	1880	78
5	71.0	1850	77.5
6	69.2	1820	77
7	67.4	1790	76.5
8	65.6	1760	76
9	63.8	1730	75.5
10	62.0	1700	75

.

The effective capacity and efficiency at year (t) can be calculated by Equations (14) and (15), respectively. Figure 4 present the degradation pattern of capacity, round-trip efficiency, and number of cycles remaining of the SLB.

$$Ca{p}_{ess}\left(t\right)=80-(1.8\times t)$$

(14)

$$\mathrm{ɳ}\left(t\right)=80-\left(0.5\times t\right)$$

(15)

4.5. Modeling of Solar PV

The output power of the solar PV is expressed in Equation (16), that is the function of area of solar PV panels (m²), solar panel efficiency, and solar irradiance (kW/m²).

$${\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v}}={\mathrm{A}}_{\mathrm{p}\mathrm{v}}{\mathrm{ɳ}}_{\mathrm{p}\mathrm{v}}{\mathrm{G}}_{\mathrm{t}}$$

(16)

Solar photovoltaic (PV) systems are based on the conversion system that converts sunlight directly into electricity. The output power varies significantly based on some factors like solar irradiance (G_t), weather conditions of specific locations, and geographic location. For this analysis, the output power is calculated for a hypothetical 3 MW solar PV system. The data for this calculation is based on a relative analysis of solar PV performance provided by Profile Solar (2024) [57], which gives seasonal data for a standard 1 kW PV panel in both summer and winter scenarios. For the calculation, we assume linear scalability, meaning that the output of the entire 3 MW system can be extrapolated by simply multiplying the 1 kW data by 3000. While this method simplifies the calculation, it does assume that the system’s overall efficiency and output characteristics are proportionally the same at both scales (1 kW and 3 MW), without accounting for potential system-level losses. In general, the solar PV systems experience significant variations during the summer and winter seasons, primarily due to changes in basic parameters such as sunlight duration, angle of incidence, and, as mentioned earlier, overall solar irradiance. In the summer months, longer daylight hours and higher solar angles lead to increased solar irradiance, while the winter days are shorter, leading to reduced sunlight availability and thus lower energy generation. Figure 5 shows both seasons, and the summer pattern has a pronounced peak during midday, with high power output extending for several hours. This solar PV output power pattern benefits from the extended daylight hours and higher irradiance. The solar PV power for summer and winter is shown in Figure 5.

4.6. Modeling of Wind Turbine

The output of wind energy depends upon many factors along with the speed of wind in a specific area, as given below in Equation (17), see (Faraz, 2024) [58]. We base the wind capacity of our model on a 100-kW-rated wind turbine for residential use (Chengdu EagleTech Energy, 2024) [59].

$${\mathrm{P}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d}}={\displaystyle \frac{1}{2}}\rho \mathrm{A}{\mathrm{V}}^3$$

(17)

where:

• $P_{wind}$ = power in Watts;
• $\rho$ = density of the air in kg/m³;
• A = cross-sectional area of the wind in m²;
• V = velocity of the wind m/s.

The wind output power depends on the density $\rho$, area characteristics of wind turbines A, and the velocity of the wind V in meters per second. The output pattern for a 2 MW wind turbine system was determined with respect to analyses performed by Windy (2024) [60] and Kasper (2024) [61] that focused on the Windsor area, providing average wind speed and an expected power output curve as a function of wind speed, respectively. These resources allowed us to estimate the expected seasonal performance (summer and winter) of the wind turbine. In particular, the first source was used to retrieve the monthly average wind speeds specific to the city of Windsor, capturing seasonal trends over the year. This data is crucial, as wind turbine output is highly sensitive to wind speed, which tends to vary seasonally. For instance, wind speeds in Windsor generally show higher averages during winter months compared to summer months, potentially due to regional weather patterns. Then, the second source provides a performance curve, indicating the expected power output of a 2 MW wind turbine at various wind speeds. This curve accounts for the turbine’s design thresholds, such as cut-in speed, rated speed, and cut-out speed, which are instrumental in calculating realistic output values. The resulting seasonal output patterns are presented in Figure 6, illustrating distinct power generation profiles for summer and winter. The winter output pattern shows a generally higher power generation capacity due to increased average wind speeds. With more frequent periods of high wind, the turbine operates closer to its rated power output, especially during peak wind hours. In summer, the turbine output is comparatively lower. This is due to reduced average wind speeds, which fall closer to the turbine’s cut-in threshold. As a result, the turbine experiences more periods of sub-optimal performance, with lower overall energy production. The output power of a wind turbine for summer and winter is presented in Figure 6, while

Table 9. Environmental parameters for renewable energy modeling.

Parameters	Unit	Data Source	Seasonal Variation	Role in Modeling
Solar irradiance (Gt)	kW/m2	Profile Solar (2024) [57]	High in summer	Calculate solar PV output power
Wind speed (V)	m/s	Windy (2024) [60]	High in winter	Impacts wind turbine power curve
Air density (ρ)	kg/m3	Assumed standard: 1.225 (UBC, 2019) [62]	Minor variation	Used for the calculation of wind output power
Ambient temperature	°C	Not directly included in the model	Moderate variation	Affects solar PV efficiency (not included)

shows the environmental parameters.

4.7. Modeling of Campus Power Plant

The cost of energy produced from the PP can be calculated by Equation (18). The limitations and the range of the other parameters are expressed in Equation (19).

$${\mathrm{C}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}={\mathsf{\lambda }}_{\mathrm{f}\mathrm{u}\mathrm{e}\mathrm{l}}\times {\mathrm{p}}_{\mathrm{p}\mathrm{p},\mathrm{t}}$$

(18)

$${{\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}\le {{\mathrm{p}}_{\mathrm{p}\mathrm{o}\mathrm{w}\mathrm{e}\mathrm{r} \mathrm{p}\mathrm{l}\mathrm{a}\mathrm{n}\mathrm{t},\mathrm{t}}}_{\mathrm{m}\mathrm{a}\mathrm{x}}$$

(19)

4.8. Modeling of Demand Response

The demand response, DR, encourages the customer to reduce or shift the loads by offering some incentives.

$${\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}={\mathrm{p}}_{\mathrm{b}\mathrm{a}\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{e} \mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}-{\mathrm{D}\mathrm{R}}_{\mathrm{t}}$$

(20)

Equation (20) expresses the baseline load without considering the demand response, and the load after the demand response is employed (10%), which helps in balancing supply and demand, reducing energy costs, and enhancing grid reliability. The change in the energy demand is expressed in Equation (21), while the cost of the response is expressed in Equation (22). In Equation (23), the demand response limitation is expressed, which is necessary for controlling and ensuring a balance between the energy supply and the load, especially during peak hours.

$${\Delta p}_{load,t}=\sum _{\mathrm{j}=1}^{24}\mathrm{D}{\mathrm{R}}_{\mathrm{j},\mathrm{t}}$$

(21)

$${\mathrm{C}}_{\mathrm{D}\mathrm{R}}=\sum _{\mathrm{j}=1}^{24}{\mathsf{\lambda }}_{\mathrm{D}\mathrm{R},\mathrm{j}}.\mathrm{D}{\mathrm{R}}_{\mathrm{j},\mathrm{t}}$$

(22)

$$0\le \mathrm{D}{\mathrm{R}}_{\mathrm{j},\mathrm{t}}\le {\mathrm{p}}_{\mathrm{t}}^{\mathrm{L},\mathrm{m}\mathrm{a}\mathrm{x},\mathrm{j}}$$

(23)

Among many other types of demand response, time-of-use (TOU) and ultra-low overnight (ULO) price tariffs are used. They comprise peak, mid-peak, and off-peak prices. In this analysis, we consider only the ULO energy price tariff.

$${\mathrm{P}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}=\mathrm{B}\mathrm{a}\mathrm{s}\mathrm{e} \mathrm{p}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}+{\mathrm{U}\mathrm{L}\mathrm{O}}_{\mathrm{a}\mathrm{d}\mathrm{j},\mathrm{t}}$$

(24)

Equation (24) represents the ultra-low overnight tariff model. Subscript “adj,t” is the adjustment factor at time t. It represents how much the grid price changes from the base price at a specific time t. As indicated earlier, Table 2 contains the values of the ULO.

4.9. Renewable Energy Resources

In this work, the solar PV and wind turbine are considered as renewable energy resources. Figure 7 below shows the current installed wind, solar, and energy storing capacities in Canada.

Renewable Fraction

The renewable fraction (RF) can be described as the ratio of the total energy consumption that is fulfilled by the RE resources to the total loads.

$$\mathrm{R}\mathrm{F}={\displaystyle \frac{\sum _{\mathrm{t}=1}^{24}({\mathrm{p}}_{\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{a}\mathrm{r} \mathrm{p}\mathrm{v},\mathrm{t}}+{\mathrm{p}}_{\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{d},\mathrm{t}})}{\sum _{\mathrm{t}=1}^{24}{\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}}}$$

(25)

The total energy generated from RE resources, i.e., solar PV and wind turbines, over the course of the day is expressed in Equation (25), while the total energy demand or consumption is ${\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}$.

4.10. Self-Consumption Rate

Self-consumption rate (SCR) is the proportion of the generated renewable energy ${\mathrm{E}}_{\mathrm{g}\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}$ that is consumed by the campus (prosumer) rather than exported to the utility grid.

$$\mathrm{S}\mathrm{C}\mathrm{R}={\displaystyle \frac{{\mathrm{E}}_{\mathrm{s}\mathrm{e}\mathrm{l}\mathrm{f}-\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}}}{{\mathrm{E}}_{\mathrm{g}\mathrm{e}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{d}}}}$$

(26)

4.11. Power Flow Load Campus Demand Relation

The grid power ${\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}$ can be positive or negative at various time intervals in a day. If the ${\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}$ is positive, it shows that the local renewable energy resource output power (solar PV, wind) is greater than local consumption and vice versa. This factor is very important as exporting power reduces the operational cost of the campus microgrids by generating the revenue as presented in Equation (27). The upper and lower limits of the campus load show the variation between the maximum and minimum values as presented in Equation (28).

$${\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}={\mathrm{p}}_{\mathrm{E}\mathrm{x}\mathrm{p}\mathrm{o}\mathrm{r}{\mathrm{t}}_{\mathrm{t}}}-{\mathrm{p}}_{\mathrm{i}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{r}{\mathrm{t}}_{\mathrm{t}}}$$

(27)

$${\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{m}\mathrm{i}\mathrm{n}}\le {\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{t}}\le {\mathrm{p}}_{\mathrm{l}\mathrm{o}\mathrm{a}\mathrm{d},\mathrm{m}\mathrm{a}\mathrm{x}}$$

(28)

4.12. Net Metering

During a day of energy consumption and generation, the excess of energy locally generated is exported to the grid, while the energy needed is imported through the net metering. The maximum upper and lower limits are expressed in Equations (29) and (30).

$$\mathrm{E}\mathrm{x}\mathrm{p}\mathrm{o}\mathrm{r}{\mathrm{t}}_{\mathrm{t}}\le \mathrm{E}\mathrm{x}\mathrm{p}\mathrm{o}\mathrm{r}{\mathrm{t}}_{\mathrm{t},\mathrm{m}\mathrm{a}\mathrm{x}}$$

(29)

$$\mathrm{I}\mathrm{m}\mathrm{o}\mathrm{p}\mathrm{o}\mathrm{r}{\mathrm{t}}_{\mathrm{t}}\le \mathrm{I}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{r}{\mathrm{t}}_{\mathrm{t},\mathrm{m}\mathrm{a}\mathrm{x}}$$

(30)

4.13. GHG Emissions

The primary goal of this research is to reduce the GHG and operational cost, which are also related to Sustainable Development Goals 7, 11, 12, and 13. Equation (31) sums the CO₂ emissions from two sources: the utility grid and campus power plant (PP).

$${\mathrm{E}}_{\mathrm{G}\mathrm{H}\mathrm{G},\mathrm{t}}=\mathrm{C}{\mathrm{O}}_{2,\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d}}\times {\mathrm{p}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d},\mathrm{t}}+\mathrm{C}{\mathrm{O}}_{2,\mathrm{p}\mathrm{p}}\times {\mathrm{p}}_{\mathrm{p}\mathrm{p},\mathrm{t}}$$

(31)

where ${\mathrm{E}}_{\mathrm{G}\mathrm{H}\mathrm{G},\mathrm{t}}$ is the total GHG emission and the two parameters ${\mathrm{C}\mathrm{O}}_{2,\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{d}}$ and ${\mathrm{C}\mathrm{O}}_{2,\mathrm{p}\mathrm{p}}$ are the emission factors.

SLBs are a cost effective and environmentally friendly solution for stationary applications such as peak shaving, renewable integration, and small appliance usage. The SLBs are obtained from EVs after their efficiency has decreased from 100% to 80% (Casals et al., 2019) [6]. Dong et al. (2023) [64] estimated that a repurposed battery’s selling price as compared to a new battery’s price is CAD 44180/kWh and CAD 150250/kWh, respectively. Their reuse extends the environmental benefits by deferring the recycling process that also contributes to the circular economy. Although SLBs have lower efficiency, a battery management system (BMS) can ensure adequate performance and longevity for stationary applications (Akram and Abdul-Kader, 2024) [65].

According to LIB chemistry, cobalt, nickel, and lithium can be recovered from EOL EV batteries through the recycling process. In particular, we can recover 116 g/kWh, 400 g/kWh, and 73 g/kWh of each material, respectively. Furthermore, the cradle-to-grate global warming potential of EV batteries is about 13 tons of CO₂-eq (Akram and Abdul-Kader, 2021) [5]. Our study takes a maximum storing capacity of 2000 kWh, using a battery bank of about 31 SLBs, each with a capacity of 64 kWh and SOH of about 80%. Using this information, we can calculate that such an energy storage system is effectively still harnessing 232 kg of cobalt, 800 kg of nickel, and 146 kg of lithium, and it saves 403 tons of CO₂-eq emissions.

Additionally, the second-life batteries (SLBs) support grid stability by participating in the demand response program and ancillary services, all at lower cost as compared to new batteries (Ahmadzadeh et al., 2021) [66]. These benefits and their affordability combined with their environmental and operational benefits make SLBs a highly advantageous option for building energy storage, especially where maximum battery performance is not critical. So, SLBs strike an excellent balance between affordability, sustainability, and functional reliability, particularly for stationary applications where their intermediate capacity does suffice.

The linear programming (LP) model is solved by using Python 3.11 with packages SciPy and matplotlib, where the computational/execution times are noted as 1.5 s. The specifications of the machine are as follows: processor—AMD Ryzen 5 7350U, memory—32 GB SODIMM with 3200 MT/s speed.

5. Results and Discussion

As per the system architecture discussed earlier in Section 3, the model is applied to the University of Windsor’s smart campus prosumer microgrid (SCPM). Currently, the campus has a 4 MW gas-fired power plant and a grid connection to feed its loads. In addition, it is anticipated that the utility grid connection has a net metering system, and surplus energy can be exported to the grid network to get financial benefits. Parts (a) and (b) of Figure 8 illustrate the cumulative winter and summer load patterns of the campus microgrid.

5.1. Case Studies

In this section, Case 1 (summer) and Case 2 (winter) are discussed. Every case considers the various demand response programs and components. The anticipated levelized costs for solar, wind, and gas-fired power plants in the province of Ontario for the year 2024 are CAD 0.08, CAD 0.05, and CAD 0.15 per kW, respectively (Dunsky, 2022) [68]. Steckel et al. (2021) [23] provides an estimation that the levelized cost of the energy storing system using SLBs in stationary applications is CAD 0.314 per kWh. The GHG emission is based on the usage of the power plant (PP) and utility-grid-based power. The carbon emission rate for energy produced by the gas power plant is 0.44 kg CO₂ per kW (EIA, 2022) [69], while the emission rate from the Ontario utility grid is 0.035 kg CO₂ per kW (Canada Energy Regulator, 2024) [70]. In each case, the costs associated with winter and summer energy consumption, as well as the energy exchange with the grid, are evaluated considering the environmental (GHG) impacts.

5.1.1. Case 1: Summer Load with 0% Demand Response (Base Case)

Case 1 considers demand fulfilled by the utility grid only. The cost observed is CAD 64,049. After the integration of the RE resources and power plant along with the ESS, the operational cost reduced to CAD 59,201 with a saving of 7.57%. The renewable fraction (RF) is 100% obtained by using Equation (25). Figure 9 shows the summer load results obtained using the values output by the Python script.

Figure 9b displays the load pattern of Case 1 with respect to renewable energy sources, the ESS, and the power plant. The sources are strategically used based on load demand and energy availability. In particular, the output solar PV energy significantly contributes to energy demands but only within the time interval of 5:00 and 20:00 (i.e., from 5 a.m. to 8 p.m.), which is during daylight, see Equation (16) solar power output.

Wind power provides a consistent but lower energy supply (between 120 kW and 250 kW) shown by Equation (17)wind output power—which is largely due to reduced average wind speeds that fall closer to the turbine’s cut-in threshold.

The campus power plant provides 4000 kW (Equation (6) lower and upper limit of PP) of energy during peak times (16:00 to 20:00), but due to its higher levelized cost (CAD 0.15 per kW), the use of this source is minimized to reduce costs and GHG emissions (see Equation (31) GHG emission calculations).

The ESS is charged/recharged during low demand and high renewable energy supply, especially during off-peak hours (2:00, 21:00, and 23:00) when grid energy cost is the lowest, and it is discharged during peak hours (Equation (24)—ULO price) or when renewable energy generation is low (i.e., 19:00, 22:00, and 24:00) to offset power plant energy and other demands. Figure 9b presents the energy used to charge the ESS as negative values. During periods when renewable energy and ESS stored energy are insufficient, such as night hours from 1:00 to 4:00 and 21:00 to 24:00, the grid supplies energy at a much higher amount, as shown in Figure 9a.

Figure 10 shows the SOC pattern (Equations (7) and (8)). Note that, since the model considers a maximum and minimum state of charge of 85% and 15%, respectively, each of the cases will exhibit similar patterns; as such, Figure 10 suffices for all the four cases describing the SOC.

5.1.2. Case 2: Summer Load with 10% Demand Response

In Case 2, a 10% DR is considered to reduce the operational costs and GHG emissions. The total per day cost observed in this scenario is CAD 52,862, showing a saving of 17.47%. The renewable fraction (RF) is 100% obtained from Equation (25). Summer load results are presented in Figure 11.

5.1.3. Cases 3 and 4: Winter Load with 0% and 10% Demand Response

As for Cases 3 and 4, we consider the same metrics applied for summer over the winter season.

Excess energy, primarily from wind generation, results in a surplus that is exported to the utility grid through net metering, which generates revenue and decreases net costs. Furthermore, we see a decrease in GHG emissions. The results of Cases 3 and 4 are shown in

Table 10. Summer and Winter Results.

Case Study	Scenario Description	Parameters	Results
Case 1: Summer Load (0% DR)	Summer load with 0% demand response	-PV output -Wind generation -Battery storage -Power plant usage -Utility grid exchange	Only utility ULO/day cost= (summer) CAD 64,049 Energy cost/day (CAD) = 59,201 Cost saving = 7.57% Renewable fraction = 4.3% Self-consumption rate = 100% GHG emission saving/day = 782 kg CO2-eq
Case 2: Summer Load (10% DR)	Summer load with 10% demand response. Load reduced by 10%	-PV output -Wind generation -Battery storage -Power plant usage -Utility grid exchange	Energy cost/day (CAD) = 52,862 Cost saving = 17.47% Renewable fraction = 4.8% Self-consumption rate = 100% GHG emission saving/day =782 kg CO2-eq
Case 3: Winter Load (0% DR)	Winter load with 0% demand response	-PV output -Wind generation -Battery storage -Power plant usage -Utility grid exchange	Only utility ULO/day cost (winter) = CAD 10,257 Energy cost/day (CAD) = 4173 Cost saving = 59.32% Renewable fraction (RF) = 45.36% Self-consumption rate = 72.0% GHG emission saving/day =1348 kg CO2-eq
Case 4: Winter Load (10% DR)	Winter load with 10% demand response. Load reduced by 10%	-PV output -Wind generation -Battery storage -Power plant usage -Utility grid exchange	Energy cost/day (CAD) = 3159 Cost saving = 69.21% Renewable fraction = 50.4% Self-consumption rate = 67.50% GHG emission saving/day =1348 kg CO2-eq

. Energy exchange with the grid is shown in Figure 12 and Figure 13 below.

Figure 12 and Figure 13 present similar information to Figure 9. The key difference is that surplus energy generation is higher in winter, which can then be sold to the grid (Equations (29) and (30)). This surplus energy is presented as negative values during the time interval (16:00 to 20:00) as shown in Figure 12a and Figure 13a.

All the above-mentioned results for summer and winter seasons are presented in Table 10 below.

Table 10 reveals various results, including the energy cost efficiency and the impact on the environment, in the proposed campus microgrid under the various seasonal loads and demand response program. The analysis explores the microgrid’s potential to integrate renewable energy (RE) resources along with the storage system to avoid the dependency on the utility grid and conventional power sources. The summer cases exhibited higher energy demand, thus the energy produced by the model leaves little to no surplus energy; furthermore, since we partly rely on energy from the Ontario utility grid and campus power plant, the energy costs and carbon taxes are high, and the RF is relatively low. The demand response program shows a noticeable reduction in energy costs and GHG emissions when optimally scheduling renewable resources. As such, the RF also increases, demonstrating energy independence. The program thus shows significant monetary and environmental benefits.

In winter, the energy demand is far lower, so the energy generated from renewable sources produces a high surplus of energy that can then be sold to the Ontario grid. Thus, when we apply the model to the summer and winter cases, the energy costs and GHG emissions decrease drastically, and RF increases significantly as compared to summer cases.

The objectives we sought out to meet were to explore a tangible means of benefiting the economy and the environment while considering the SDGs. As such, a system that leverages renewable energy, such as solar PV, wind turbines, and battery storage, is interesting. We have thus presented through this paper a configuration of energy supply that is practical, feasible, and widely applicable. We can see tangible benefits to the economy and the environment as well as supportive actions to the SDGs.

5.2. Contributions to Sustainable Development Goals (SDGs)

As mentioned at the beginning, the primary goal of this research is to support SDGs, particularly goals 7, 11, 12, and 13. From the above, we discussed four cases for the model, namely summer and winter scenarios with 0% and 10% demand response, all of which provide insights regarding costs, GHG emissions, and renewable fractions. The paragraphs below present the details related to the contributions of this research to these identified Sustainable Development Goals.

5.2.1. Contribution to Goal 7: Affordable and Clean Energy

A higher demand for energy production requires an optimized means of energy storage. Using second-life batteries through repurposed applications reduces the costs of energy storage and the demand for newly manufactured energy storage units (Akram and Abdul-Kader, 2024) [65]. The results shown in Table 10 provide evidence that leveraging repurposed batteries to store energy produced from RE sources is less costly than buying energy directly from the grid. In summer, buying electricity from the utility grid costs CAD 64,049 per day. In Case 1 (0% DR) when applying renewable energy, an ESS, grid, and power plant, our costs reduce to CAD 59,202, thus granting savings of 7.57%. In Case 2 (10% DR), our costs were reduced to CAD 52,862, thus a saving of 17.47%. In winter Cases 3 (0% DR) and 4 (10% DR), our savings are 59.32% and 69.21%, respectively. This inexpensive energy production and storage can then be used in remote regions and elsewhere. Thus, the need for environmentally harmful energy generation decreases. By reducing the demand for manufacturing new LIBs as well as the costs associated with energy storage, the energy produced can be more cost effective and clean, accelerating the progress for Goal 7.

5.2.2. Contribution to Goal 12: Responsible Consumption and Production

Repurposing EV batteries aids closing the loop of battery production, making the process far more circular. The product usage stream from production to consumption to disposal is inherently wasteful. But using spent batteries in secondary applications greatly reduces resource consumption and waste. As highlighted in Akram and Abdul-Kader (2023) [71], strategies to extend battery lifespans directly promote circularity and lead to lessened environmental harm caused by battery usage. Casals et al. (2019) [6] calculate approximately how much more battery lifespan can be increased through the use in stationary applications, namely 12 years in self-consumption, 6 years in area regulation, up to 30 years in fast EV charging stations, and 12 years in transmission upgrading deferral applications. Rather than sending batteries to landfills or recycling them immediately after their lives end, we can reuse them in other applications, thus partially closing the loop for a battery’s life cycle. As such, circularizing battery production aligns with responsible consumption and production, thus aiding the progress of Goal 12.

5.2.3. Contribution to Goal 13: Climate Action

The results from Table 10 show that buying energy from the grid is both costly and environmentally detrimental as compared to using RE and storing it via second-life batteries. Thus, if the latter option were to be widely adopted, we would see a drastic improvement in GHG emissions, moving forward to various worldwide climate goals, including Goal 13. Our investigation presents GHG emission savings of 782 kg CO₂-eq per day in summer (Cases 1 and 2), and 1348 kg CO₂-eq per day in winter (Cases 3 and 4). The results obtained by the model convey the pivotal role that second-life batteries and renewable energy sources play in supporting Goal 13.

5.2.4. Contribution to Goal 11: Sustainable Cities and Communities

Through this research work, we note contributions to both Goals 7 and 13. Due to the overlapping nature of the SDGs, it can be immediately seen that, through more cost effective and environmentally friendly energy production and storage, we can enhance sustainability and sustainable practices across wider regions, from smaller municipalities to wider metropolitan hubs. Due to the variability of wind power and the time constraints of solar power, these renewable energy sources are unfortunately not optimal as standalone means of energy. By building an energy storage system that manages surplus energy, we can stimulate continuous energy flow by offsetting deficits with stored surplus. Our findings show that the best option for an energy storage system is SLBs due to their lower costs compared to new batteries. These SLB-powered ESSs can then effectively harness renewable energy and increase sustainability across a variety of regions. The results presented above can motivate corporations to move towards these optimized means of energy production and storage and push lawmakers to pass legislation accordingly. This advances the progression of Goal 11 from multiple perspectives.

6. Sensitivity Analysis

6.1. Cost and GHG Emission Analysis

To further assess the feasibility of the model, a sensitivity analysis is conducted. The purpose of the sensitivity analysis is to examine the best solution through the modification of applied parameters.

Table 11. Sensitivity analysis results—costs (ULO).

Cases	SOC (min %)	SOC (max %)	LCOS (CAD)	Solar PV Capacity (KW)	Wind Turbine Capacity (KW)	Cost/Day Summer 0% DR (CAD)	Cost/Day Summer 10% DR (CAD)	Cost/Day Winter 0% DR (CAD)	Cost/Day Winter 10% DR (CAD)
Case 1	15	85	0.314	3000	2000	59,201	52,862	4173	3159
Case 2	10	90	0.314	3000	2000	59,201	52,862	4173	3159
Case 3	5	80	0.314	3000	2000	59,214	52,874	4186	3171
Case 4	15	85	0.2	3000	2000	59,190	52,851	4164	3150
Case 5	10	90	0.25	3000	2000	59,195	52,856	4168	3154
Case 6	15	85	0.314	1500	2000	59,535	53,197	4322	3308
Case 7	15	85	0.314	3000	1000	59,320	52,980	5367	4396
Case 8	5	80	0.314	3000	1000	59,332	52,993	5380	4408
Case 9	5	80	0.314	1500	2000	59,547	53,210	4334	3320
Case 10	10	90	0.314	1500	2000	59,535	53,197	4322	3308

and

Table 12. Sensitivity analysis results—GHG emissions (ULO).

Cases	SOC (min %)	SOC (max %)	LCOS (CAD)	Solar PV Capacity (KW)	Wind Turbine Capacity (KW)	GHG Emissions Summer 0% DR (CAD)	GHG Emissions Summer 10% DR (CAD)	GHG Emissions Winter 0% DR (CAD)	GHG Emissions Winter 10% DR (CAD)
Case 1	15	85	0.314	3000	2000	25.34	23.55	9.97	9.68
Case 2	10	90	0.314	3000	2000	25.34	23.55	9.97	9.68
Case 3	5	80	0.314	3000	2000	25.34	23.55	9.97	9.68
Case 4	15	85	0.2	3000	2000	25.34	23.55	9.97	9.68
Case 5	10	90	0.25	3000	2000	25.34	23.55	9.97	9.68
Case 6	15	85	0.314	1500	2000	25.75	23.96	10.16	9.89
Case 7	15	85	0.314	3000	1000	25.53	23.73	10.43	10.17
Case 8	5	80	0.314	3000	1000	25.53	23.73	10.43	10.17
Case 9	5	80	0.314	1500	2000	25.75	23.96	10.16	9.89
Case 10	10	90	0.314	1500	2000	25.75	23.96	10.16	9.89

present the costs and GHG emissions for each of the cases shown, respectively, for ULO tariffs.

Table 13. Sensitivity analysis results—costs (TOU).

Cases	SOC (min %)	SOC (max %)	LCOS (CAD)	Solar PV Capacity (KW)	Wind Turbine Capacity (KW)	Cost/day Summer 0% DR (CAD)	Cost/day Summer 10% DR (CAD)	Cost/day Winter 0% DR (CAD)	Cost/day Winter 10% DR (CAD)
Case 1	15	85	0.314	3000	2000	53,795	50,851	5976	5092
Case 2	10	90	0.314	3000	2000	53,795	50,851	5976	5092
Case 3	5	80	0.314	3000	2000	53,802	50,858	5983	5099
Case 4	15	85	0.2	3000	2000	53,782	50,839	5968	5081
Case 5	10	90	0.25	3000	2000	53,787	50,843	5972	5086
Case 6	15	85	0.314	1500	2000	54,153	51,029	6164	5280
Case 7	15	85	0.314	3000	1000	53,876	50,933	6870	5986
Case 8	5	80	0.314	3000	1000	53,883	50,940	6877	5993
Case 9	5	80	0.314	1500	2000	54,161	51,037	6171	5287
Case 10	10	90	0.314	2000	2000	54,153	51,030	6164	5280

and

Table 14. Sensitivity analysis results—GHG emissions (TOU).

Cases	SOC (min %)	SOC (max %)	LCOS (CAD)	Solar PV Capacity (KW)	Wind Turbine Capacity (KW)	GHG Emissions Summer 0% DR (CAD)	GHG Emissions Summer 10% DR (CAD)	GHG Emissions Winter 0% DR (CAD)	GHG Emissions Winter 10% DR (CAD)
Case 1	15	85	0.314	3000	2000	27.24	25.63	11.35	11.06
Case 2	10	90	0.314	3000	2000	27.24	25.63	11.35	11.06
Case 3	5	80	0.314	3000	2000	27.24	25.63	11.35	11.06
Case 4	15	85	0.2	3000	2000	27.24	25.63	11.35	11.06
Case 5	10	90	0.25	3000	2000	27.24	25.63	11.35	11.06
Case 6	15	85	0.314	1500	2000	27.43	25.72	11.45	11.16
Case 7	15	85	0.314	3000	1000	27.29	25.68	11.89	11.6
Case 8	5	80	0.314	3000	1000	27.29	25.68	11.89	11.6
Case 9	5	80	0.314	1500	2000	27.43	25.72	11.45	11.16
Case 10	10	90	0.314	1500	2000	27.43	25.72	11.45	11.16

do the same for TOU tariffs. For the sake of comparison, we take Case 1 to be the default case across all scenarios, and our comparisons are with respect to it.

Figure 14 and Figure 15 present the summer and winter per day cost and GHG emission results of ULO tariffs.

Figure 16 and Figure 17 present the summer and winter per day cost and GHG emission results of TOU tariffs.

• Case 2 takes the state of charge minimum and maximum to be 10 and 90, respectively, which resulted in negligible change to costs and GHG emissions.
• Case 3 takes the state of charge minimum and maximum to be 5 and 80, respectively, which resulted in a minor increase in costs and no changes in GHG emissions.
• Case 4 takes LCOS to be 0.2, which results in a minor decrease in costs and no changes in GHG emissions.
• Case 5 takes state of charge minimum and maximum to be 10 and 90, respectively, and takes LCOS to be 0.25, which resulted in a minor decrease in costs and no changes in GHG emissions.
• Case 6 takes a lower solar capacity, which results in an increased cost of energy and an increase in GHG emissions.
• Case 7 has a lower wind capacity, which results in an increased cost of energy and an increase in GHG emissions.
• Case 8 takes the state of charge minimum and maximum to be 5 and 80, respectively, and takes a lower wind capacity which resulted in an increase in costs and an increase in GHG emissions.
• Case 9 takes the state of charge minimum and maximum to be 5 and 80, respectively, and takes a lower solar capacity which resulted in an increase in costs and an increase in GHG emissions.
• Case 10 takes the state of charge minimum and maximum to be 10 and 90, respectively, and takes a lower solar capacity which resulted in almost identical results to Case 6.

We see various patterns. If our sources of renewable energy, solar and wind, see a decrease in capacity, we see an increase in costs and GHG emissions. This is due to the system relying more on the grid and campus power plant, which are our sources of GHG emissions. When we decrease LCOS to 0.2, the changes to costs and GHG emissions are negligible. Since our ESS has higher LCOS as compared to other parts of our system (such as solar, wind, power plant, and grid), our system tries to minimize extracting energy from said source. When the state of charge is varied, there are negligible changes to costs and no changes to GHG emissions. This is also related to our ESS, and as we discussed prior, our system minimizes its use, thus showing us almost no impact when changing SOC.

Looking at the tables for TOU tariffs, we see an overall increase in costs and GHG emissions as compared to the ULO tariff tables. This is clear from the increase in the costs of tariffs using TOU, as seen in Table 13 and Table 14.

6.2. Battery Degradation Analysis

It is in the nature of batteries to degrade over time through repetitive use. The extent of this degradation is context-specific and depends primarily on charge and discharge cycles. As mentioned prior, Casals et al. (2019) [6] analyzed four applications in which spent LIBs can be used, of which use in a fast EV charging station can bring a spent battery’s lifespan to 30 years. Section 6.1 analyzes, among other factors, the impact on costs when varying LCOS, and we see that decreasing LCOS results in minute impacts. Here, we analyze the impact of varying the capacity of the ESS, and for comparison, we vary the capacities of solar and wind to more drastic measures to showcase a pattern.

Table 15. Sensitivity analysis of ESS.

Case	SOC (min %)	SOC (max %)	LCOS (CAD)	Energy Storage System Capacity (KW)	Solar PV Capacity (KW)	Wind Turbine Capacity (KW)	Cost Summer 0%DR (CAD)	Cost Winter 0%DR (CAD)
Case 1	15	85	0.314	2000	3000	2000	59,201	4173
Case 2	15	85	0.314	1500	3000	2000	59,201	4173
Case 3	15	85	0.314	1000	3000	2000	59,201	4173
Case 4	15	85	0.314	500	3000	2000	59,210	4182
Case 5	15	85	0.314	2000	1500	2000	59,535	4322
Case 6	15	85	0.314	2000	750	2000	59,704	4430
Case 7	15	85	0.314	2000	3000	1000	59,320	5385
Case 8	15	85	0.314	2000	3000	500	59,378	5973

presents the sensitivity analysis of the ESS. As in the prior section, we take Case 1 to be our base case. Across all cases, SOC minimum and maximum and LCOS are kept constant. The summer and winter cases with no demand response are shown.

Figure 18 presents the capacities of solar, wind, and ESS in different cases.

Figure 19 displays the summer and winter cost results of ESS analysis.

• Cases 2 and 3 decrease the ESS capacity to 1500 and 1000, respectively, which causes no increase in the costs.
• Case 4 decreases the ESS capacity to 500, which shows a minor increase in costs.
• Case 5 decreases the solar capacity, which causes a more substantial increase in costs.
• Case 6 decreases the solar capacity significantly, which causes a significant increase in costs.
• Case 7 decreases the wind capacity, which causes an increase in costs in summer and a substantial increase in costs in winter.
• Case 8 decreases the wind capacity significantly, which causes a more substantial increase in costs in summer and a significant increase in costs in winter.

We see a clear pattern that changes to ESS capacity and LCOS result in minute changes to costs as compared to changes in solar and wind capacities. As such, if the batteries in the ESS go through major degradation, it does not significantly impact the system, as most of the energy is harnessed through solar and wind, evident through the increased costs in Table 15. Furthermore, because the system relies primarily on solar and wind energy, and batteries degrade as a function of their charge and discharge cycles, it is clear that the ESS would degrade very slowly, as its reliance is lower. This further conveys the minor impact that ESS degradation would have on the system, hence we do not consider it in our model.

6.3. Model Validation and Reliability Assessment

With regard to verifying this system’s behavior, the sensitivity analysis grants insight into the parameters’ overall impacts, which assist in validating our results and trends. Further, we assess the system’s reactions to extreme cases, such as reducing solar, wind, and ESS capacities by 75%, and the results stay consistent with the model’s logical breakdown. To further ensure accuracy and credibility,

Table 16. Comparison with other research work.

Author	Year	Optimization Method	Application	Software	Cost Saving	GHG Saving	Remarks
Ali et al. [72]	2023	MIQCP	Microgrid	GAMS 44.1.1	2.60%/day 3.725/year	-	A new method is proposed to reduce the cost by considering demand response, distributed energy resource, and ESS.
Li et al. [73]	2022	MILP	Microgrid		36.6%	-	Analysis of SCR, ESS degradation, DR optimal sizing, and cost.
Muqeet et al. [74]	2020	MILP	Microgrid	MAITLAB R2020b	29–35%	730 kg - 750 kg per day	Reduce the operational cost of an institutional microgrid by applying energy management strategy.
Azimian et al. [75]	2020	MILP	Microgrid	GAMS 31.1.0	23%	-	Financial feasibility of microgrid projects with green resources explored and verified.
Nasir et al. [76]	2019	LP	Microgrid	MAITLAB R2019a	16%	-	Reviewed the key challenges in the integration of solar energy in residential power back-up units.
Gao et al. [26]	2018	MILP	Microgrid	MAITLAB R2018b	5.3%	-	Optimal scheduling of a microgrid, considering the energy cost.
Purpose Study		LP	Campus microgrid	Python 3.11	Summer 7.57–17.4% Winter 59.32–67.5%	Summer 728 kg/day Winter 1348 kg/day	Focuses on reducing the energy consumption cost and greenhouse gas emissions on a campus microgrid. In summer for 0% DR and 10% DR, SCR is 100% and RF is 4.3% and 4.8%, respectively. In winter for 0% DR and 10% DR, SCR and RF are 72%, 67% and 45.36%, 50.4%, respectively.

presents relevant studies to allow us to benchmark our model’s output for RF, self-consumption rate, costs, and GHG emissions. In particular, these studies present strategies and analyses to improve microgrid operation with cleaner and cheaper energy production, which is in line with our model’s overall purpose. As such the results we have found stay consistent with the current literature on the topic.

7. Conclusions

This paper assessed the impact of using SLBs from EVs in stationary applications by considering a case study on the University of Windsor campus microgrid. In this research work, we advocate for smart prosumers, those who consume and produce energy. Prosumers can take part in energy exchange through net metering. The study was divided into two scenarios (summer and winter seasons) each with two cases with respect to operational costs; then, various means of point-of-use energy generation, such as PV, wind, and energy storage systems, were considered under each case. The results show the potential progress towards the achievement of SDGs 7, 11, 12, and 13.

The sensitivity analysis presents an idea of the key system factors that impact costs and GHG emissions. With how our hybrid system is designed, we rely primarily on renewable energy whenever possible, leverage an ESS to smoothen output, and sell back any surplus energy to the Ontario grid to further decrease net costs.

The findings from the comparative analysis of summer and winter load scenarios reveal that demand response (DR) strategies can further improve the cost-effectiveness and sustainability of the system. One such example is presented in the sensitivity analysis, which applies both ULO and TOU Canadian tariffs to the system, thus giving an idea of how the system would operate under stricter time-oriented costs. The reduction in energy costs through net metering during winter months combined with higher renewable fraction emphasizes the potential of SLBs and renewable energy sources to reduce the dependence on grid power and minimize operational costs.

The integration of SLBs into campus microgrids presents a promising pathway toward achieving long-term energy sustainability. By reducing GHG emissions and operational costs and improving RE utilization, the proposed model offers a practical, eco-friendly solution for future energy management systems. The results of this work could be significant for energy managers/planners, policymakers, and environmental activists. These results could help attract relevant stakeholders. This study paves the way for further research on SLBs and RE integration in microgrids, particularly in the context of achieving SDGs and minimizing environmental impacts.