The Untapped Potential of Smart Charging: How EV Owners Can Save Money and Reduce Emissions Without Behavioral Change

Basic Information

• Paper ID: 2503.03167
• Title: The Untapped Potential of Smart Charging: How EV Owners Can Save Money and Reduce Emissions Without Behavioral Change
• Authors: Yash Gupta, William Vreeland, Andrew Peterman, Coley Girouard, Brian Wang (Rivian Automotive)
• Classification: eess.SY (Systems and Control), cs.SY (Systems and Control), physics.soc-ph (Physics and Society)
• Publication Year: 2025
• Paper Link: https://arxiv.org/abs/2503.03167

Abstract

The transportation sector represents the largest source of emissions in the United States and the second-largest globally. Electric vehicles (EVs) are projected to comprise half of global vehicle sales by 2035, serving as a critical solution for reducing emissions and enhancing grid flexibility. Based on research utilizing de-identified EV telemetry data from Rivian Automotive in 2023, the study reveals that 72% of home charging sessions commence immediately after vehicle connection, regardless of utility time-of-use (TOU) rates or managed charging programs. Fewer than 26% of charging sessions in the sample involve EV owners actively scheduling charging times to align with utility rates or programs. By optimizing home charging schedules within existing plug-in/plug-out time windows, the research demonstrates that EV owners can save an average of $140 annually and reduce charging-related carbon emissions by up to 28%.

Research Background and Motivation

Problem Context

1. Environmental Challenges: The transportation sector accounts for 28% of U.S. CO₂ emissions and 16.2% of global emissions. By 2035, EV electricity demand could reach 14% of total U.S. electricity demand, compared to the current 0.6%.
2. Grid Stress: Uncoordinated EV charging may increase peak loads, straining grid infrastructure and leading to higher electricity costs. Grid planners project that peak electricity demand will increase by 38GW within five years due to cross-sector electrification and AI service growth.
3. User Behavior Issues: Research indicates that 72% of home charging begins immediately after plug-in, disregarding TOU rates or incentives. Only 51% of EV owners are aware of residential EV utility incentive programs, and merely 35% regularly schedule charging times.

Research Motivation

Existing smart charging research often assumes unrealistic vehicle availability and charging flexibility, overlooking users' daily habits. This study aims to quantify the potential cost and emission savings achievable through smart charging without requiring behavioral change.

Core Contributions

1. Real Data-Based Analysis: Utilizes authentic telemetry data from 5,195 Rivian vehicles encompassing over 432,900 charging sessions throughout 2023.
2. Optimization Framework Without Behavioral Change: Proposes a constrained optimization scenario that optimizes charging only within actual user plug-in/plug-out time windows, without requiring habit modification.
3. Quantified Savings Potential: Demonstrates average savings of 21.5% in charging costs and 3.9% in carbon emissions under TOU rates, with annual savings of $140 per vehicle.
4. Regional Differentiation Analysis: Analyzes differentiated savings potential across 12 major utility service areas, revealing complex relationships between rate structures and grid carbon intensity.

Methodology Details

Task Definition

Input: EV charging session data (plug-in/plug-out times, state of charge, location), utility rate information, marginal emission data

Output: Optimized charging schedule minimizing cost and emissions

Constraints: Complete required charging within actual user plug-in/plug-out time windows

Optimization Framework

1. Constrained Optimization Scenario

• Optimizes charging only within actual user plug-in time windows
• Quantifies potential benefits without behavioral change
• Handles multiple charging sessions within single plug-in windows

2. Unconstrained Optimization Scenario

• Optimizes charging across full day with no time restrictions
• Establishes upper bound baseline for potential savings
• Enables comparative analysis of maximum possible benefits

Optimization Objectives

1. Primary Objective: Minimize electricity costs
2. Secondary Objective: Minimize carbon emissions

For TOU rate structures, selects lowest marginal emission periods within lowest-cost time periods; for flat-rate structures, optimizes emissions only.

Technical Implementation Details

Rate Processing

• Baseline Calculation: Assumes users utilize standard residential rates
• Optimization Calculation: Assumes users select EV-specific rates (if available)
• Savings Calculation: Integrates comprehensive benefits of rate conversion and charging schedule optimization

Emission Data Processing

• Utilizes marginal operational emission rates (MOER) provided by WattTime
• Averages 5-minute signals into hourly intervals to reduce computational complexity
• Employs marginal emission factors rather than average factors for accuracy

Experimental Setup

Dataset

1. Vehicle Data: 5,195 Rivian vehicles distributed across 13 states
2. Charging Data: 432,900+ home charging sessions (January 1 - December 31, 2023)
3. Rate Data: Rate information from 12 major utility companies obtained via Arcadia Signal platform
4. Emission Data: Marginal carbon emission data from WattTime covering 210+ countries and regions

Filtering Criteria

• Ownership of Rivian-specific home charging station (Wall Charger)
• Enrollment in data sharing for product development
• Service by utility companies with at least 100 Rivian customers
• Coverage of major U.S. utility service areas

Evaluation Metrics

1. Cost Savings: Percentage difference between optimized and baseline costs
2. Emission Reduction: Percentage difference between optimized and baseline emissions
3. Annual Savings: Absolute annual cost savings and emission reductions per vehicle

Experimental Results

Primary Results

TOU Rate Regions (58% of vehicles, 3,010 units)

• Constrained Optimization: Average savings of 21.5% in charging costs, 3.9% reduction in carbon emissions
• Unconstrained Optimization: Average savings of 29% in charging costs, 10.1% reduction in carbon emissions
• Annual Savings: Average $140 per vehicle (constrained),$164 per vehicle (unconstrained)

Flat-Rate Regions (42% of vehicles, 2,185 units)

• Constrained Optimization: 7.4% reduction in carbon emissions
• Unconstrained Optimization: 15.3% reduction in carbon emissions
• Emission Reduction: Average 64 kg CO₂ per vehicle (constrained), 152 kg CO₂ per vehicle (unconstrained)

Regional Differentiation Analysis

High-Performance Regions

• PG&E Customers: Achieve 7.4% emission reduction under cost minimization objective
• Colorado Public Service Company: Achieve 17.4% emission reduction under flat-rate structure

Lower-Performance Regions

• Arizona Public Service Company: Only 0.6% emission reduction
• Ameren Illinois: Only 1.5% emission reduction

Cost Increase Scenarios

Approximately 1/6 of charging sessions experience cost increases after optimization, primarily when charging/plug-in windows overlap with rate peak periods. Despite individual session cost increases, overall optimization achieves 21.5% cost savings.

Key Findings

1. Efficiency: Constrained optimization achieves 74% of unconstrained optimization's cost savings (21.5% vs. 29%)
2. Emission Benefits: Constrained optimization achieves 39% of unconstrained optimization's emission reductions (3.9% vs. 10.1%)
3. Behavioral Patterns: Vehicles remain plugged in for approximately twice the actual charging duration needed, providing optimization opportunities

Related Work

Current State of Smart Charging Research

Existing research primarily focuses on several areas:

1. Grid Integration Optimization: Research demonstrates that smart charging can reduce charging costs by 30%, grid operational costs by 10%, and renewable energy curtailment by 40%
2. Solar Integration: Smart charging with solar PV integration can achieve cost savings exceeding 50%
3. Emission Optimization: Charging optimization based on emission signals can reduce carbon emissions by 18-32%

Differentiation of This Research

• Realistic Constraint Consideration: Most studies assume ideal vehicle availability, overlooking daily user habits
• Real Data Validation: Employs large-scale authentic charging data rather than simulated data
• No Behavioral Change: Focuses on optimization potential within existing user habits

Conclusions and Discussion

Main Conclusions

1. Significant Savings Potential: Smart charging can achieve substantial cost and emission savings even without behavioral change
2. Practical Feasibility Validation: Constrained optimization scenarios demonstrate the feasibility of achieving benefits under real-world conditions
3. Regional Variability: Savings potential varies significantly across regions, requiring tailored strategies

Limitations

1. Data Representativeness: Based solely on Rivian vehicle data, potentially not fully representative of all EV user populations
2. Geographic Limitations: As an emerging EV company, Rivian vehicles are highly concentrated in California
3. Static Analysis: Does not account for dynamic factors such as charging habit evolution and energy market structure changes
4. Implementation Mechanisms: Does not test specific methods for assisting customers in automatically shifting charging times

Future Directions

1. Expanded Data Scope: Research incorporating more EV brands, geographic locations, and charging patterns
2. Dynamic Pricing Models: Investigation of dynamic pricing impacts and evolving energy structures
3. Automated Implementation: Development of automated tools and strategies to help customers realize theoretical savings
4. Behavioral Interventions: Research on effective methods to encourage users to proactively adjust charging times

In-Depth Evaluation

Strengths

1. High Data Quality: Utilizes large-scale authentic EV telemetry data, providing reliable empirical foundation
2. Methodological Innovation: Proposes constrained optimization framework quantifying benefits under realistic constraints with strong practical value
3. Comprehensive Analysis: Simultaneously addresses cost and emission dimensions across multiple regions and utility companies
4. Credible Results: Findings align with prior research (70% immediate charging), enhancing result credibility

Limitations

1. Sample Bias: Rivian customer demographics may possess specific characteristics, limiting result generalizability
2. Geographic Concentration: High vehicle concentration in California may affect national representativeness
3. Lack of Implementation Validation: While quantifying theoretical potential, actual deployment effects remain unverified
4. Insufficient Cost-Benefit Analysis: Inadequate analysis of cost-benefit tradeoffs for smart charging system deployment

Impact

1. Policy Guidance: Provides data support for utility companies designing EV rates and incentive programs
2. Industry Application: Provides technical pathways for EV manufacturers developing smart charging features
3. Academic Contribution: Establishes empirical research exemplar based on real data for smart charging field
4. Environmental Benefits: Demonstrates feasibility of emission reduction without sacrificing user convenience

Applicable Scenarios

1. EV Manufacturers: Integration of smart charging platforms to enhance user value
2. Utility Companies: Design of more effective TOU rates and EV incentive programs
3. Policymakers: Development of policy frameworks promoting smart charging adoption
4. Charging Infrastructure Operators: Optimization of charging network operational strategies

References

The paper cites 31 important references covering:

• EV and grid integration research (Richardson et al., 2012)
• Smart charging technology surveys (Deb & Pihlatie, 2022)
• User behavior analysis (Brady & O'Mahony, 2016)
• Emission optimization methods (Huber et al., 2020)
• Industry reports (IEA Global EV Outlook 2024, J.D. Power research, etc.)

Summary: This paper validates the significant potential of smart charging through large-scale real data, demonstrating that substantial cost and emission savings are achievable even without requiring behavioral change. The research provides important empirical support for the EV industry and grid operations, though opportunities remain for improvement in data representativeness and implementation validation.