Unraveling the Economic Uncertainty of Urban Rooftop Photovoltaic Coupled Electric Vehicle Systems: A Scenario-Based Simulation Study

1. Introduction and Research Rationale

My research focuses on the economic viability of an urban rooftop photovoltaic system coupled with electric vehicles (PV+EV). Given the global imperative to achieve the Paris Agreement’s 1.5°C target, urban centers, which house over half of the world’s population, are under immense pressure to decarbonize. In China, where urbanization has reached approximately 66%, the critical significance of urban decarbonization is especially pronounced. The primary sources of urban carbon emissions are the power and transportation sectors. In Shenzhen, a pioneer in electric mobility, the transportation sector is increasingly relying on electric vehicles, while the local power grid still largely depends on external coal-based electricity. Thus, urban rooftop photovoltaics (PV) represent a potent solution for degregating the power supply at the local level. However, the inherent intermittency of solar energy creates a significant demand for energy storage.

My objective is to assess the economic performance of integrating electric vehicles into an urban rooftop PV system. The system leverages electric vehicle batteries as dynamic, distributed storage units through Vehicle-to-Grid and Vehicle-to-Home technologies. Yet, this system’s economic viability is riddled with uncertainty. The system’s financial viability is consequently affected by user behavior (such as electric vehicle driving patterns), technical progress (including changing battery costs and costs of degradation), and socio-economic developments (such as the expansion of urban electricity demand). Therefore, my research seeks to solve the following problem: to identify the key factors influencing the viability of urban community shared PV+EV systems, and to analyze their influence mechanisms.

To begin my analysis, I have developed a comprehensive techno-economic assessment framework based on the actual 3MW community PV project located in the Shangmeilin Community, Futian District, Shenzhen. This selected location is particularly noteworthy due to its designation as a “near-zero carbon” pilot, and its status as a high-density urban village. The project, developed in conjunction with the Shenzhen Gas Group, serves as the basis for the entire study. My primary goal is to use this framework to quantify the impacts and mechanisms of multiple uncertainty factors on viability. Studying its economic outcomes in scenarios marked by uncertainty allows me to derive conditions required for economic feasibility, thereby providing a scientific foundation for policy design and strategy for district energy system planning.

2. Literature Review and Research Gap

Existing literature on complex integrated photovoltaic and electric vehicle systems has extensively examined techno-economic factors including feed-in tariffs (FIT), retail electricity prices, self-consumption rates, and investment costs. Research on urban scale rooftop PV systems has shown a realization of grid parity in some regions, thereby enhancing production profitability based on local conditions. However, several critical research gaps still exist. First, many analyses for PV+EV systems are based on the assumption of fixed parameters, without systematically accounting for a comprehensive spectrum of interacting uncertainties related to behavioral and technical aspects. Second, existing studies seldom analyze the economic implications of variations in the *availability rate* of electric vehicles used for storage, and how these variations interact with pace of technological progress and energy storage degradation mechanisms. Third, relatively few studies examine PV+EV from the vantage point of a high-density, highly populated, non-western urban neighborhood. This research aims to fill these gaps by proposing a scenario-based analytical framework to treat a dynamically complex system with an explicit focus on an urban village context.

3. Methodology: An Integrated Framework for PV+EV Economic Assessment

To systematically evaluate the PV+EV system in question, I built a four-module framework that simulates its technical performance and economic viability. This is shown in the following methodology schematic (pseudo-conceptual representation in Table 1):

**Table 1: Core Components of the PV+EV Economic Assessment Framework**

| Module | Core Components | Implementation Method / Key Parameters | Output |
| :— | :— | :— | :— |
| **1. Module for Electricity Supply-Demand Matching** | PV Output Modeling; Building Load Modeling; Electric Vehicle/Battery Storage Modeling. | System Advisor Model (SAM) simulation based on hourly weather data; hourly grid data; system parameters. | Hourly PV generation curve; Hourly residential load curve; Battery charge/discharge data; Grid import/export profiles. |
| **2. Economic Analysis Module** | Cost Modeling; Revenue Modeling; Discounted Cash Flow Analysis. | Initial investment (PV & V2H); Operation & Maintenance (O&M); Battery replacement cost; Revenue (savings + feed-in tariffs). | LCOE, Net Present Value (NPV), Internal Rate of Return (IRR), Discounted Payback Period (DPBP). |
| **3. Factor Identification Module** | Identification of key uncertainty drivers. | Literature review; Model-based policy analysis; Data forecasting. | Set of key influencing factors: Electric vehicle availability; Technological progress; Battery degradation, etc. |
| **4. Economic Optimal Conditions & Policy Recommendations module** | Definition of optimal scenarios. | Cross-comparison of scenarios; Assessment of factor impacts. | Policy recommendations and optimal development paths. |

### 3.1 Electricity Supply-Demand Matching Simulation

My analytical core involved the use of energy system modeling software, specifically the National Renewable Energy Laboratory’s System Advisor Model (SAM). I created a simulation over 8760 hours for three distinct elements: the generation from PV, consumption by residential buildings, and EV power storage/discharge to simulate each integrated PV+EV scenario.

(1) **PV Output Modeling**: I utilized actual contract data for a 3MW rooftop PV installation. I collected hourly meteorological data for Shenzhen, including solar irradiation (DNI, DHI and GHI), temperature, and wind speed, to simulate the precise output of the PV array for the location. The model simulates power production based on the solar data and specific technical design (component type, azimuth, tilt) incorporating all possible losses of the system.

(2) **Residential Load Modeling**: I first extrapolated the historical increase in electricity demand in Shenzhen to forecast the future demand in 2030. Since 2023, a linear regression model based on historical year demand data (2010-2023) provides the following forecast: Predicted 2030 electricity consumption for Shenzhen is **1386.26×10^8 kWh**, a **23% increase** from the 2023 baseline. After scaling this data to the target community, I obtained the annual electricity load for the Shangmeilin residents. Then, I used the SAM software to model a high-resolution hourly load profile of the community, taking into account the daily behavioral patterns of residential consumers.

(3) **EV Storage Model**: To model PV+EV storage, I calculated the quantity of electric vehicles that might be connected to the grid at the community level, based on the number of households that could host the required PV equipment. For this model, each household EV was assumed to have a **battery capacity of 50 kWh**, and the V2H charging/discharging unit had an assumed **power rating of 7 kW**. The model considered different percentages of EV availability for grid services. The core purpose of integrating an EV is to store surplus photovoltaic energy, which otherwise would be exported to the grid. The simulation logic follows an “PV first” scheduling: PV electricity is used as follows: (1) to power building loads directly, (2) to charge EV batteries, (3) to be exported to the grid only when demand is fully met and batteries are full.

**Table 2: Key Technical Specifications Used for Modeling**

From the annual simulation output, key technical performance metrics were derived, where direct and storage-based consumption of PV energy is of great interest to my analysis due to its economic significance. The difference in feeding energy to the grid vs. using it locally to offset retail purchases is central to system economics.

### 3.2 Economic Modelling Approach

After designing an appropriate technical model for each scenario, I used a comprehensive cost-benefit analysis. My focus is on the costs and benefits at the prosumer level. I incorporated several interacting costs for each scenario.

**Total system cost** includes:

\[
C_{\text{system}} = C_{\text{initial investment}} + C_{\text{O\&M}} + C_{\text{battery replacement}}
\]

Here, \( C_{\text{initial investment}} \) includes the purchase and installation cost of PV modules and V2H systems. Equation details are as follows:

\[
\begin{aligned}
C_{\text{PV+EV initial}} &= P_{\text{PV}} \times IC_{\text{PV}} + P_{\text{V2H}} \times IC_{\text{V2H}} \\
\text{Where: } \\
P_{\text{PV}} &= \text{Installed PV capacity (kW)} \\
IC_{\text{PV}} &= \text{Unit cost of PV (yuan/W)} \\
P_{\text{V2H}} &= \text{Installed V2H capacity (kW)} \\
IC_{\text{V2H}} &= \text{Unit cost of V2H system (yuan/W)}
\end{aligned}
\]

The operation and maintenance (O&M) cost (\( C_{\text{O\&M}} \)) is the unit cost per kW of PV every year. \( C_{\text{battery replacement}} \) costs are linked to the inevitable degradation of EV batteries that occurs over time.

**Total System Revenue** is primarily captured in two ways. The revenue stream is calculated as:

\[
\begin{aligned}
R_{\text{PV+EV}} &= R_{savings} + R_{\text{FIT}} \\
R_{\text{PV+EV}} &= (E_{self} \times r_{\text{retail}}) + (E_{export} \times r_{\text{FIT}})
\end{aligned}
\]

Where \(`E_{self}`\) is the kilowatt-hours produced by PV and used to offset the building’s electricity consumption that would otherwise be bought from the grid (at the retail rate, \( r_{\text{retail}} \)). \( E_{export} \) is the remaining PV electricity sent to the grid, which generates revenue at the feed-in tariff rate or FIT (\( r_{\text{FIT}} \)).

For this project, I used the following key economic parameters:

**Table 3: Economic and Policy Parameters Utilized in this Study**

| Parameter | 2023 Base Case | 2030 Slower Tech. Progress | 2030 Faster Tech Progress |
| :— | :— | :— | :— |
| PV Cost (yuan/W) | 5.79 | 3.45 | 2.57 |
| O&M Cost (yuan/W-year) | 0.055 | 0.05 | 0.041 |
| EV Battery Replacement Cost (yuan/kWh) | 1200 | 550 | 360 |
| V2H System Cost (yuan/kWh) | 1200 | 550 | 360 |
| Residential Retail Electricity Price (yuan/kWh) | 0.6628 | – | – |
| FIT (yuan/kWh) | 0.463 | – | – |
| Discount Rate | 8% | – | – |
| PV System Lifetime | 25 years | – | – |

The major economic indicators used to compare scenarios were the levelized cost of energy, and the internal rate of return. The **Levelized Cost of Energy (LCOE)** is calculated as follows:

\[
\text{LCOE} = \frac{\sum_{n=1}^{N} \frac{C_{\text{capital},n} + C_{O\&M,n}}{(1+r)^n}}{\sum_{n=1}^{N} \frac{E_{\text{generated},n}}{(1+r)^n}}
\]

It represents the average cost per unit of electricity generated, accounting for the time-value of money investments. **Net Present Value** (NPV) and **Internal Rate of Return** (IRR) are standard and were also computed:

\[
NPV = \sum_{n=0}^{N} \frac{CF_n}{(1+r)^n}
\]

Where \( CF_n \) is the cash flow in year *n*. The IRR is the discount rate at which NPV = 0. A threshold of 8% was set as the investment benchmark to determine whether the PV+EV project was economically attractive.

4. Identification of Influencing Factors and Scenario Design

Through an extensive literature review, I have identified the key factors influencing this system’s economics. Factors are broadly classified into three categories: user behavioral factors, technological progress factors, and socio-economic factors. From these, I defined four uncertainty-laden variables to use across multiple simulation scenarios.

1. **Electric Vehicle Availability Rate**: EV utilization is based on the percentage of EVs connected and available to the PV system for storing or discharging energy, when accounted. Given that some vehicles will be on the road, I defined three availability rates to encode this variability:
* 50% (often known as average daily commuting rate)
* 35%
* 20%

2. **Technological Progress Rate**: The pace of cost declines for critical components (e.g., PV panels, batteries) will fundamentally change overall project economics. I modeled three distinct base levels of technology costs:
* Baseline (2023)
* 2030 – Slower Technological Progress
* 2030 – Faster Technological Progress

3. **Battery Degradation Rate**: The duration during which EV batteries can sustain full performance, after which replacement is required. I considered accelerated degradation as following the battery standards after 750 cycles, whereas low degradation is defined as only needing replacement after 900 cycles.

4. **Building Electricity Demand**: The demand increases according to socio-economic development. Forecasts were made via linear regression, with demand for 2030 as 23% greater than that in 2023.

The structure of my factors and their derived scenario combinations are illustrated in the following table.

**Table 4: Variables and Their Values Defining the Scenario Matrix**

The ‘2023’ and ‘2030’ demand scenarios were crossed with the technology cost scenarios, battery degradation and EV availability scenarios. This generated a set of 54 distinct scenarios to assess. I also created a reference case (“PV-only”) by only utilizing the PV component and an EV availability rate of zero to use as a comparative baseline.

5. Results & Analysis

### 5.1 The ‘PV-only’ System Baseline

In the base case (PV only), the techno-economic simulation yields considerable insight. The first-year simulation results show a PV self-consumption rate of **91.6%** and a self-sufficiency rate of **32.78%**, due to the relatively high building load density in this high-population-density district. This means nearly all generated solar energy is used locally to offset electricity purchases, showcasing the existing synergy between the PV generation pattern (daytime) and the building load pattern (which also peaks at midday).

The LCOE for this PV-only system is **0.43 yuan/kWh**. Interestingly, that falls *below* the current thermal power benchmark rate used for FIT (0.463 yuan/kWh) and below the retail price of electricity (0.66 yuan/kWh). This indicates that this district has already achieved *both* generation-end and user-end grid parity with PV technology alone. Furthermore, with an NPV of **956.85×10^4 yuan** and an IRR of **14.64%**, this investment offers a higher return rate than the 8% benchmark and is deemed a comfortable profit.

### 5.2 PV+EV Analysis: Technological Progress as a Driver

The PV+EV scenarios include all cost data. This is shown in Table 5. There is a dramatic effect on cost metrics when the effect of technological progress is introduced. This effect can be quantified by analyzing total costs across different cost-decline scenarios.

**Table 5: Cost and LCOE values across various cost scenarios**

| Cost Scenario | Total Cost (million yuan) | LCOE (yuan/kWh) |
| :— | :— | :— |
| 2023 Baseline Costs | 19.6 – 25.0 | 0.49 – 0.57 |
| 2030 Slower Technology Improvement | 11.7 – 14.6 | 0.30 – 0.33 |
| 2030 Faster Technology Improvement | 8.8 – 10.9 | 0.20 – 0.25 |

The key result is that LCOE undergoes a **38–44% reduction** transitioning from the 2023 scenario to the ‘slow’ technological progress scenario (2030). This is coupled with an *additional* 22–26% reduction in LCOE from the slower to faster progress scenario. This is a critical analysis that demonstrates the significance of technology innovation by achieving a cumulative cost reduction total of 60%.

Under a fast technological progress scenario, results for **IRR** show an even more drastic improvement. IRR values of **32% and above** were observed in the scenario that combines fast technology progress with more favorable battery degradation conditions. This contrasts with IRR values of 15% or lower in 2023 baseline settings. It is obvious that large-scale technical progress could shift the economic viability of well-designed PV+EV systems from moderate to very high profitability.

**Figure 1: IRR values under different technology development paths and various availability rates. (Not shown due to the requirement to avoid referencing images.)**

Technological advancement, therefore, emerges as the most significant determinant of PV+EV system economics: its effect overrides the simultaneous cost burden potential of EV integration.

### 5.3 Impact of EV Availability on System Economics

The impact of electric vehicles on the system’s economics was not uniform and is nuanced. Table 6 depicts the effects.

**Table 6: Effect of EV Availability Rate on Technical and Economic Indicators in the 2023 scenario.**

| EV Availability Rate | Self-consumption rate increase | Total System Cost (Low Degr.) | Total System Cost (Fast Degr.) | IRR from PV+EV system |
| :— | :— | :— | :— | :— |
| 20% | Baseline | Low | Low | Highest |
| 35% | +0.03%-0.93% | Higher | Higher | Intermediate |
| 50% | + more (see fig) | Highest | Highest | Lower |

A rise in the EV availability rate from 20% to 50% is expected to increase the system’s self-consumption. However, the amount of additional power captured and consumed per unit of added storage is subject to diminishing marginal returns. Due to the shape of the solar curve being single-peaked, increasing the battery capacity from 35% to 50% yields only a slight increase in total battery cycling, as shown in Figure 1 analysis, contrary to what one would initially expect.

The economic implications of my study are compelling. Evaluating the costs: A higher availability rate inherently requires a larger number of V2H units and battery capacity. Consequently, this drives up initial capital expenditure and total system investment. Even though higher availability may reduce battery replacement frequency or eliminate need for replacement, added V2H installation costs, as seen in my simulation, often dominate system economics.

In a nuanced finding from the 2023 baseline with slower degradation, a EV availability rate of 20% *often yields the highest profitability* compared to 35% and 50%. For instance, the total system cost increased by roughly 3% with a rise in availability from 20% to 35% when considering degradation. Although the 50% availability rate also resulted in no required battery replacements, the massive increase in V2H system installation costs relative to lower availability scenarios leads to a total system cost increase of 2-3%. Therefore, the revenue increase from self-consumption is insufficient to meet the increased infrastructure expenditure of setting up a V2H network.

Under the future low-cost technological scenarios for 2030, the consequence is that an EV availability rate higher yields a lower IRR (e.g., from 32% to 31%). Therefore, from an investment perspective, one should not arbitrarily increase EV participation rates *unless* the cost of V2H equipment and battery power declines at the rate predicted by technology progress scenarios. Otherwise, one faces a cost-benefit imbalance that undermines profitability.

### 5.4 The Role of Battery Degradation Rate

Battery degradation is another important financial driver in my research.

**Table 7: Battery Replacement Schedule and Impact (for an Example 2023 Cost Scenario).**

Battery replacement frequency profoundly influences the system’s financial analysis. A sustained high rate of battery degradation leads to multiple replacement events over a PV system’s 25-year lifespan. When battery life is shortened from 900 cycles (slow degradation) to 750 cycles (fast degradation), the incidence of required battery capital expenditures can increase by 50-100% depending on the EV availability scenario. When factoring these additional costs into the levelized cost of energy, there is a **7.45% to 57.02% increase in LCOE**, which further demonstrates its importance. The economic penalty of battery degradation is very sensitive to EV availability, affecting decisions on system sizing.

### 5.5 The Effect of Building Electricity Demand

As an evergrowing economy drives up building electricity use, the economic attractiveness of the system is further influenced. In my scenarios where future building electricity demand increases by 23%, the results are subtle. This is because the *marginal gain* in the rate of PV self-consumption is minimal when the baseline self-consumption is already high.

**Table 8: Financial Impact Based on a 23% Increase in Electricity Demand**

As the load grows, consumption of PV solar energy increases, thus offsetting more retail electricity expensive purchases. However, for PV+EV systems, the influence lies in the fact that a larger part of the solar power is consumed instantly and is less reliant on electric vehicle charging, thereby reducing the number (or depth) of battery cycles. This impacts total system degradation over time.

The analysis shows diminishing marginal benefits of improving load matching, especially in districts like Shangmeilin, with a high population density and resultant large baseline demand. Here I found that transitioning to 2030 with higher demand only improved IRR by **0.46 to 0.67 percent**, a relatively minor improvement compared to the value derived from technology progress.

### 5.6 PV+EV Economic Viability Under Multiple Uncertainties: The Road to Optimality

The complete scenario analysis shows that the levelized cost of energy of all PV+EV scenarios falls below the retail price for residents as of 2023, but *exceeds* the standard FIT. This demonstrates that PV+EV, as a system, has achieved grid parity only at the *user-end* level but not at the *generation end*. However, in all scenarios simulated for the year 2030, the LCOE of the system falls below *both* the feed-in tariff as well as the retail price, thus achieving total grid parity.

In alignment with the core research question of identifying economic conditions, I performed cross-parameter comparisons.

**Table 9: Comparison of Cost and Revenue Impacts for PV-only and PV+EV (2023 scenario with baseline cost)**

| System Set-up | Total Discounted Cost (×10^6 yuan) | Total Discounted Revenue (×10^6 yuan) | Cost-Benefit Ratio | Cost-Benefit Difference (×10^6 yuan) |
| :— | :— | :— | :— | :— |
| PV-only | 19.57 | 29.14 | 1.49 | 9.57 |
| PV+EV EV=50%, slow degradation | 22.06 | 29.32 | 1.33 | 7.26 |
| PV+EV EV=35%, fast degradation | 24.99 | 28.33 | 1.13 | 3.30 |

For high-density communities, adoption of PV+EV systems remains *less economically attractive* than PV-only because the primary benefit, increased self-consumption, is marginal given the base high self-consumption. In an urban village setting, the added capital expenditure required by EVs is not vital for the PV system’s economics if the PV is already making significant progress. It is, however, crucial in capturing renewable energy at *other times*.

However, this outcome is not globally valid. I have found that for communities with lower building energy requirements, the situation inverts. In communities with significantly reduced loads (i.e., one-third of the Shangmeilin baseline), deficits arise on the PV output leading to low self-consumption in an PV-only context. In this case, I found that the IRR for an PV+EV system without replacements can *exceed* that of a PV-only system. This indicates that **proper selection of PV vs. PV+EV systems needs to be tailored to local socio-technical contexts**. As such, the rooftop PV systems and electric vehicle battery storage become integrated and essential, enabling cost-effective decarbonization in lower-density communities.

6. Optimal Conditions and Policy Recommendations

Synergistic factor interactions uncover clear conditions under which the system will achieve optimal economic performance. A shift towards **rapid technology advancement for photovoltaic/battery systems** is essential. Combined with **longer cycle life batteries**, the synergic effect could make urban PV+EV systems very profitable. The following are policy suggestions for areas such as Shangmeilin:

* **Accelerate Technology Innovation:** Governments must fund R&D that will advance core PV and EV technologies (e.g., high cycle-life battery chemistry, lower-cost power electronics for V2H), because technology advancement is a principal driver for economic performance. This could be achieved by subsidies and grants for relevant applied research.
* **Adapt strategies to local conditions:** There is no single universal optimal technology choice. In dense urban regions where self-consumption is high, focus should remain on PV-only systems or policy mechanisms that promote EVs as an additional form of storage may require more time before becoming fully cost-competitive. On the other hand, in low self-consumption areas, supporting the deployment of PV+EV systems could unlock significant untapped potential while being immensely more efficient than conventional grid fed solar PV without adequate EV/battery storage.
* **Implement smart electricity pricing:** Designing an electricity tariff structure with dynamic Time-Of-Use pricing is vital to incentivizing the correct behavior. Such strategies enable consumers to utilize electricity when surplus solar energy is available (charging EV) or when loads need to be met using PV-generated electricity (discharging) to ensure that consumers save most on their electricity costs from a holistic economic standpoint.

7. Limitations & Future Scope of Research

Although this study has produced useful insights, the limitations of my analysis are noteworthy. The model, for example, adopted a relatively simple assumption of ‘EV availability’, not considering the intricacies of actual travel behaviors (e.g., a day of the week, seasonal variations) that could modify certain modeling aspects. Future research should incorporate more comprehensive stochastic models for charging from various sources, beyond PV. Additionally, this study is based on a single local case study, which makes practical yet site-specific conclusions. To establish generalizable theories for smart city design, future studies need to expand this structured framework across many diverse cities, climates and sociotechnical contexts. This would enhance the translatability and application of conclusions on a global scale.

8. Conclusion

This dissertation comprehensively analyzes the urban rooftop photovoltaic coupling electric vehicle (PV+EV) system’s economic performance in a case study performed in Shenzhen. The methodology used dynamic system modeling to estimate the technical performance and then to assess the resulting economic implications under a variety of different unknown parameter values. Data shows that while solely PV systems have already reached full grid parity and are offering a strong degree of economic feasibility, the economic viability of integrating with EVs is heavily impacted by four key factors.

From a technical progression perspective: it emerges as the strongest driver. A faster technological development scenario could decrease LCOE by up to 60% by 2030 in comparison to the year 2023. Similarly, high EV availability rates present cost-benefit dilemmas. They raise self- consumption and operational benefits, but this is offset by increased capital costs related to the extensive installation of V2H charging units. Battery degradation is an economic critical variable due to replacement costs increasing LCOE significantly (from 7% to more than 57%). Finally, increasing building load has only a modest effect on improving the economic attractiveness of PV+EV due to fundamental diminishing returns in high density regions.

The research broadly concludes that urban PV+EV systems will be required for full decarbonization but will not be universally economical for high-density urban communities unless innovative technologies continuously evolve, and specifically must be appropriately tailored and based on community-specific characteristics.

In the future, by designing adaptive policies based on technological progress and local energy/lifestyle heterogeneity, cities can use such integrated PV+EV systems to create a low-carbon and economically efficient urban energy architecture.