FA info icon.svg Angle down icon.svg Page data
Authors Nima Asgari
License CC-BY-SA-4.0
Language English (en)
Related 0 subpages, 1 pages link here
Impact 41 page views (more)
Created August 23, 2023 by Nima Asgari
Last modified January 3, 2024 by Joshua M. Pearce

1. Introduction[edit | edit source]

Solar Photovoltaics (PV) is among the most widely used renewable energy technologies[1] since it has become the least expensive renewable energy provision system.[2][3][4] The global power capacity produced by PV systems has become close to 940 GW in 2021, with the largest annual increment (175 GW) ever.[5] Although the transportation sector has a substantially large contribution to global electricity consumption (30% in 2021), it has the lowest renewable energy share compared with the other sectors (3.3% biofuels and 0.3% renewable electricity).[5] Nevertheless, electric vehicle (EVs) sales have reached 6.6 million in 2021, more than double its growth in 2020.[5] Therefore, electrification of the transportation sector using solar PV technology can play a pivotal role in reducing fossil fuel consumption. For this, some research has taken place in the literature to explore the feasibility of the integration of PV systems and EVs.

2. General Studies[edit | edit source]

Evaluation of On-Board Photovoltaic Modules Options for Electric Vehicles[6]

Abdelhamid, M., Singh, R., Qattawi, A., Omar, M., & Haque, I. (2014). Evaluation of On-Board Photovoltaic Modules Options for Electric Vehicles. IEEE Journal of Photovoltaics, 4(6), 1576–1584. https://doi.org/10.1109/JPHOTOV.2014.2347799

  • Impact of commercial options for photovoltaic (PV) modules on on-board EV charging.
  • Driving hours’ range of 0-5 - low and high-temperature operating conditions for PV modules.
  • Reduction in CO2 emissions by approximately 3 to 6.5 kg compared to internal combustion vehicles of a similar kind.
  • Mono-Si PV modulesː well-suited for powering low-speed, lightweight, and aerodynamically efficient EVs.

Integrating electric vehicles and residential solar PV[7]

Coffman, M., Bernstein, P., & Wee, S. (2017). Integrating electric vehicles and residential solar PV. Transport Policy, 53, 30–38. https://doi.org/https://doi.org/10.1016/j.tranpol.2016.08.008

  • The life cycle cost and the GHG emissions of EVs vs. those of the other popular and similar cars in Hawaii.
  • Total cost of ownership of EVs > similar non-electric vehicles and hybrid electric vehicles (HEVs).
  • But, with federal tax credits of EVs, the life cycle costs of Nissan Leaf = similar internal combustion engine vehicles (ICEVs) and HEVs.
  • 79% of the generated electricity in Hawaiiː from fossil fuels. Therefore, the most efficient EVs emit GHG 7% more than the best-performing HEV counterparts.
  • Using residential PVsː life cycle cost of Nissan Leafː about $1200 lower than the Toyota Corolla (the next lowest cost ICEV) over a lifetime of 150,000 miles.
  • Using residential PVs, just for weekend applicationsː all EVs perform better than Toyota Prius (GHG emissions).

3. Small Scale EV & PV[edit | edit source]

Urban Scale Photovoltaic Charging Stations for Electric Vehicles[8]

Brenna, M., Dolara, A., Foiadelli, F., Leva, S., & Longo, M. (2014). Urban Scale Photovoltaic Charging Stations for Electric Vehicles. IEEE Transactions on Sustainable Energy, 5(4), 1234–1241. https://doi.org/10.1109/TSTE.2014.2341954

  • A grid-connected PV carport with a charging station of two EVs in three Northern, Central, and Southern locations in Italy.
  • The maximum and minimum charging power of 22 kW, and 11 kW.
  • Two scenariosː contemporary and sequential charging patterns.
  • A small fraction of the energy generated by PV systemsː charges EV batteries (1-72% depending on the energy required).
  • By employing sequential chargingː higher self-consumption.
  • The most unfavorable scenarioː two vehicles with a maximum charging power of 22 kW are combined.
  • The most favorable situationː a car and a quadricycle with the highest ΔSOC and the lowest charging power are together. Recharging method is not important.

Integration between electric vehicle charging and PV system to increase self-consumption of an office application[9]

Roselli, C., & Sasso, M. (2016). Integration between electric vehicle charging and PV system to increase self-consumption of an office application. Energy Conversion and Management, 130, 130–140. https://doi.org/https://doi.org/10.1016/j.enconman.2016.10.040

  • A PV system → EV stations, heating and cooling facilities and electric devices in an office located in southern Italy.
  • The dynamic energy performance and the environmental impacts.
  • 3 chargers for constant charging | 3 DCC chargers for variable charging.
  • The self-consumption of 69.3% for a 9 kW PV in DCC mode for charging a vehicle for a daily distance of 40 km. For 4.5 kW, the SC was below 17%.
  • Using this systemː mitigation in CO2 emission rate by 40% compared with the conventional system, a natural gas-fired boiler, an electric chiller, and a diesel car.

Energetic, economic and environmental viability of off-grid PV-BESS for charging electric vehicles: Case study of Spain[10]

Grande, L. S. A., Yahyaoui, I., & Gómez, S. A. (2018). Energetic, economic and environmental viability of off-grid PV-BESS for charging electric vehicles: Case study of Spain. Sustainable Cities and Society, 37, 519–529. https://doi.org/https://doi.org/10.1016/j.scs.2017.12.009

  • The techno-economic and environmentalː an off-grid PV-powered EV charging station with a battery energy storage system (BESS) in Madrid, Spain.
  • A 281.52 kW PV system and a 442.854 kWh BESSː required for the base case.
  • The energy priceː 0.4 Euro/kWh, similar to the ICE.
  • The PPː 7 years
  • For a life span of 13 years for batteries and 25 years for PVsː the final cash flow = 1,200,000.00 Euro (considerably high).
  • Two additional chargersː match the excess capacity and avoid overproduction.

Transforming a residential building cluster into electricity prosumers in Sweden: Optimal design of a coupled PV-heat pump-thermal storage-electric vehicle system[11]

Huang, P., Lovati, M., Zhang, X., Bales, C., Hallbeck, S., Becker, A., Bergqvist, H., Hedberg, J., & Maturi, L. (2019). Transforming a residential building cluster into electricity prosumers in Sweden: Optimal design of a coupled PV-heat pump-thermal storage-electric vehicle system. Applied Energy, 255, 113864. https://doi.org/https://doi.org/10.1016/j.apenergy.2019.113864

  • A residential cluster with 3 buildings in Sweden to an electricity prosumerː How? byː an optimized integration of PV, heat pump, thermal storage, and EV.
  • The objective functionː maximizing self-consumption under a non-negative NPV.
  • In the baseline scenarioː self-consumption up to 77% - self-sufficiency level of 20%.
  • Number of EVs from 2 to 48 → self-consumption from 79.4% to 80.4% and a slight decrease in LCOE.

PV assisted electric vehicle charging station considering the integration of stationary first- or second-life battery storage[12]

Bartolucci, L., Cordiner, S., Mulone, V., Santarelli, M., Ortenzi, F., & Pasquali, M. (2023). PV assisted electric vehicle charging station considering the integration of stationary first- or second-life battery storage. Journal of Cleaner Production, 383, 135426. https://doi.org/https://doi.org/10.1016/j.jclepro.2022.135426

  • A PV-powered EV charging station with batteriesː optimization to minimize the annual CO2 emissions and to maximize the self-consumption.
  • 3 scenarios: without any battery, with a first-life battery (FLB), and with a second-life battery (SLB).
  • In the SLB scenarioː the accumulative CO2 emission = decreased by 15% and 10%, compared with the FLB scenario and the one without any battery.

Building-centric investigation into electric vehicle behavior: A survey-based simulation method for charging system design[13]

Liu, X., Fu, Z., Qiu, S., Li, S., Zhang, T., Liu, X., & Jiang, Y. (2023). Building-centric investigation into electric vehicle behavior: A survey-based simulation method for charging system design. Energy, 271, 127010. https://doi.org/https://doi.org/10.1016/j.energy.2023.127010

  • EV behavior was mirrored using survey data from private vehicle fleets in two office buildings in Beijing and Ruicheng China.
  • Purposeː How well EV charging demands align with power generated by distributed PV systems.
  • The size of a cityː impact on the behavior of private EVs and their charging demands.
  • Despite their similar average annual driven distances (around 1.32 × 10^4 km/year in Beijing and 1.29 × 10^4 km/year in Ruicheng), the nature of the trips in the two cities was different.
  • Workplace slow chargingː can meet the daily city travel needs of private EVs.
  • With a single peak in charging demand, high power chargers → discrepancies between EV and PV power outputs → low solar self-consumption rate (annual load factors between 18% and 42%). Reasonː excessive PV power during weekends and summertime, highlighting weekly and seasonal imbalances.

4. Large Scale EV & PV[edit | edit source]

Co-benefits of large scale plug-in hybrid electric vehicle and solar PV deployment[14]

Denholm, P., Kuss, M., & Margolis, R. M. (2013). Co-benefits of large scale plug-in hybrid electric vehicle and solar PV deployment. Journal of Power Sources, 236, 350–356. https://doi.org/https://doi.org/10.1016/j.jpowsour.2012.10.007

  • The benefits of the deployment of solar PV for charging plug-in hybrid EVs (PHEVs)
  • PV and EV = potentially complements of each other.
  • Shifting from overnight charging to mid-day charging = a decrease in battery size - an increase in electrification → PHEVs can travel for longer distances - a reduction in petroleum displacement. However, an augmentation in peak load.
  • PVː meets all the increased mid-day peak loads corresponding to PHEV charging.

The pure PV-EV energy system – A conceptual study of a nationwide energy system based solely on photovoltaics and electric vehicles[15]

Boström, T., Babar, B., Hansen, J. B., & Good, C. (2021). The pure PV-EV energy system – A conceptual study of a nationwide energy system based solely on photovoltaics and electric vehicles. Smart Energy, 1, 100001. https://doi.org/https://doi.org/10.1016/j.segy.2021.100001

  • The pure PV-EV nationwide power provisionː PV and EVs’ batteries → all the electric demands of Spain.
  • Assumptionː all facilities (including vehicles) are electrified.
  • Vehicle-to-grid (V2G) technology while parking → regulate the PV grid.
  • The 100% self-reliance on the PV gridː with 3.45 billion m2 of PVs and 29.4 million EVs following V2G.
  • Insteadː if 2 billion m2 for PV → self-reliance would be 79%.

A New Framework for Plug-In Electric Vehicle Charging Models Supported by Solar Photovoltaic Energy Resources[16]

Assolami, Y. O., Gaouda, A., & El-Shatshat, R. (2021). A New Framework for Plug-In Electric Vehicle Charging Models Supported by Solar Photovoltaic Energy Resources Un nouveau cadre pour les modèles de recharge du véhicule électrique rechargeable supporté par des ressources d’énergie solaire photovoltaïque. IEEE Canadian Journal of Electrical and Computer Engineering, 44(2), 118–129. https://doi.org/10.1109/ICJECE.2020.3008689

  • A model for temporal-spatial characteristics of plug-in EV (PEV) drivers’ behavior
  • National Household Travel Survey (NHTS) global data as well as Buffalo and New York State.
  • Enhanced by the Markov Chain Monte Carlo process.
  • Model elementsː parking duration (PDT) - driven distance (DDM) - driving duration (DDT) - uncertainty model of solar radiation.
  • Purposeː estimating the probability of PEV charging load, conventional loads, and PV generation.
  • A sensitivity analysis onː location - PEV penetration level (number of PEVs supplied by PV-powered charging stations) - PEV battery specifications.
  • Drivers plug their EV into the fast charging station (FCS)ː when they leave their workplace by 3 pm.
  • Drivers plug their EV into FCSː when the remaining EV battery SOC level doesn’t support the remaining daily activities.
  • Penetration level of 30%ː the load demand for PVs = increased by 75%
  • Penetration level of 50%, the load demand = doubled.

A multi-objective optimization model for fast electric vehicle charging stations with wind, PV power and energy storage[17]

Sun, B. (2021). A multi-objective optimization model for fast electric vehicle charging stations with wind, PV power and energy storage. Journal of Cleaner Production, 288, 125564. https://doi.org/https://doi.org/10.1016/j.jclepro.2020.125564

  • Fast EV charging stations powered by a hybrid system of PV, wind turbine (WT), and battery.
  • Optimization to minimize the costs of electricity and emissions.
  • Dynamic model for EV charging expectations according to the electricity price policies.
  • Will be run for the case of Inner Mangolia, China.
  • The optimal capacity of WT, PV, and batteriesː 330 kW, 280.75 kW, and 750 kWh with an electricity cost of 0.306 yuan/kWh and a total pollution emission of 472.38 g.

5. Machine Learning Approaches (Small Scale)[edit | edit source]

Electric vehicle charging profile prediction for efficient energy management in buildings[18]

Kumar, K. N., Cheah, P. H., Sivaneasan, B., So, P. L., & Wang, D. Z. W. (2012). Electric vehicle charging profile prediction for efficient energy management in buildings. 2012 10th International Power & Energy Conference (IPEC), 480–485. https://doi.org/10.1109/ASSCC.2012.6523315

  • Artificial Neural Network (ANN) for past charging patterns, initial State of Charge (SOC), and final SOC → predict the charging profiles of EVs connected to a building.
  • Mean Absolute Percentage Error (MAPE)ː about 4.5%.
  • In urban areas like Singapore, where EVs are typically charged in batches of 50-100ː the variations in power consumption are impactful.

Household EV Charging Demand Prediction Using Machine and Ensemble Learning[19]

Ai, S., Chakravorty, A., & Rong, C. (2018). Household EV Charging Demand Prediction Using Machine and Ensemble Learning. 2018 IEEE International Conference on Energy Internet (ICEI), 163–168. https://doi.org/10.1109/ICEI.2018.00037

  • Various machine learning algorithms → predict the day-ahead EV charging time for households & forecast days with no charging.
  • Additionally, a two-layer hybrid stacking ensemble learning approach integrating multiple machine learning techniques.
  • Prediction of day-ahead of chargingː the accuracy of Random Forest (RF) and Naive Bayes (NB) was high.
  • Prediction of "No Charge" dayː the accuracy of AdaBoost (AdaB) and GBoost (GB) was high.
  • Above all, the ensemble approach showcased the finest performance.

The Application of Improved Random Forest Algorithm on the Prediction of Electric Vehicle Charging Load[20]

Lu, Y., Li, Y., Xie, D., Wei, E., Bao, X., Chen, H., & Zhong, X. (2018). The Application of Improved Random Forest Algorithm on the Prediction of Electric Vehicle Charging Load. Energies, 11(11). https://doi.org/10.3390/en11113207

  • A forecasting technique for EV charging loadː random forest (RF) algorithm.
  • Load data from a single charging station in Shenzhen, China.
  • Charging demand in the summer > the winter.
  • Distinct patterns observed during holidays.
  • Effective estimation of the station's charging capacity in 15-minute intervals or over a 24-hour period. The former prediction achieved an MAPE of 9.76% and an RMSE of 2.27 - the latter yielded an MAPE of 10.83% and an RMSE of 39.59.
  • A benchmark for various EV charging load management strategies.
  • Valuable insights for both electricity providers and consumers.

Data analysis of a monitored building using machine learning and optimization of integrated photovoltaic panel, battery and electric vehicles in a Central European climatic condition[21]

ur Rehman, H., Korvola, T., Abdurafikov, R., Laakko, T., Hasan, A., & Reda, F. (2020). Data analysis of a monitored building using machine learning and optimization of integrated photovoltaic panel, battery and electric vehicles in a Central European climatic condition. Energy Conversion and Management, 221, 113206. https://doi.org/https://doi.org/10.1016/j.enconman.2020.113206

  • A 160 m2 building area in Belgiumː the efficiency of an energy system (A heat pump and two EVs) supplied by an on-grid PV/T system with battery.
  • Reduction in the cost of imported electricity. How? optimizing the size of the electrical components & implementing control strategies.
  • The data were gathered from May 2017 to May 2018.
  • The gradient tree boosting ensemble methodː to forecast the PV production - electricity consumption of the house - power of EVs.
  • ML algorithmsː to identify the most influential features. Furthermore, to assess the impact of curtailment on PV production - to identify specific time periods within the data where curtailment of PV production may occur - to estimate the full PV production potential achievable without any curtailment.
  • 80% training and 20% testing. Afterward, a 5-fold cross-validation technique. The trained model's cross-validated MAE was 3%.
  • The performance of the energy complex was optimizedː variables → capacity of PV system (20, 10, and 5 kW), battery capacity (64.4, 32.2, 16.1, and 0 kWh), and EV charging scenarios (actual, fixed, and variable).
  • The PV system with a capacity of 20 kW demonstrated superior performanceː reduction in imported energy - a decrease in imported energy costs - an increase in the onsite energy fraction - favorable net-zero energy balance level.
  • For the same PV capacity, with a battery capacity of 16.1 kWh → annual cost savings of up to 31% compared with the base case building (with 10 kW PV panels and a battery capacity of 32.2 kWh).

Prediction of EV Charging Behavior Using Machine Learning[22]

Shahriar, S., Al-Ali, A. R., Osman, A. H., Dhou, S., & Nijim, M. (2021). Prediction of EV Charging Behavior Using Machine Learning. IEEE Access, 9, 111576–111586. https://doi.org/10.1109/ACCESS.2021.3103119

  • Historical charging records & data on weather, traffic, and eventsː to forecast the charging duration and energy usage of EVs.
  • ML techniques such as RF, SVM, XGBoost, and deep NN.
  • Charging records from two university campus stations: JPL and Caltech. The weather informationː NASA's Modern-Era Retrospective analysis for Research and Applications. Traffic detailsː Google Maps.
  • The predictions of energy consumptionː more precise than those for session duration, regarding improved R2 and SMAPE metrics. The discrepancyː users' reluctance to input their estimates each time they charge their EVs.
  • Deep ANNː the lowest accuracy in all scenarios.
  • Ensemble learning surpassed standalone ML models.

Deep reinforcement learning control of electric vehicle charging in the presence of photovoltaic generation[23]

Dorokhova, M., Martinson, Y., Ballif, C., & Wyrsch, N. (2021). Deep reinforcement learning control of electric vehicle charging in the presence of photovoltaic generation. Applied Energy, 301, 117504. https://doi.org/https://doi.org/10.1016/j.apenergy.2021.117504

  • Deep reinforcement learning (RL)ː for EV charging control in an energy system (utility grid, building, PV generation, and a single EV).
  • The dataset of hotels for a two-month period in 2020.
  • No data for EV → the EV charging patterns were simulated.
  • Objectiveː increasing PV self-consumption and EV SOC at departure.
  • Compared with the Naive and Rule-based control (RBC), the RLC method was effective in providing practical strategies for controlling EV charging.

Power output optimization of electric vehicles smart charging hubs using deep reinforcement learning[24]

Bertolini, A., Martins, M. S. E., Vieira, S. M., & Sousa, J. M. C. (2022). Power output optimization of electric vehicles smart charging hubs using deep reinforcement learning. Expert Systems with Applications, 201, 116995. https://doi.org/https://doi.org/10.1016/j.eswa.2022.116995

  • Goalː charges are completed by redistributing the demand load away from the times of peak.
  • Deep Reinforcement Learning is used.
  • Dataː 120 charging transactions in the building’s garage.
  • An environmentː basic charging station operations - random EV arrival and departure timings.
  • Compared to baselineː A cut down in the load from EV charging by 80% during the peak periods.

Modelling community-scale renewable energy and electric vehicle management for cold-climate regions using machine learning[25]Zahedi, R., hasan Ghodusinejad, M., Aslani, A., & Hachem-Vermette, C. (2022). Modelling community-scale renewable energy and electric vehicle management for cold-climate regions using machine learning. Energy Strategy Reviews, 43, 100930. https://doi.org/https://doi.org/10.1016/j.esr.2022.100930

  • An on-grid PV power supplyː electrical, cooling and heating loads in a residential neighborhood in St. Albert, AL, Canada.
  • EV chargingː by the excess power of PV → The charging pattern to be predicted.
  • Recurrent Neural Network technique of GRU with 192 hidden layers.
  • Data of first 9 months of the year to predict the last three months of the year.
  • A 24 kW PVː provides 32% of the electrical, cooling and heating demands of the house.
  • The excess power for EV chargingː supplying the house demand became 29.23%.
  • The prediction accuracy was 88.6%.

A Robust Model Predictive Control-Based Scheduling Approach for Electric Vehicle Charging With Photovoltaic Systems[26]Yang, Y., Yeh, H.-G., & Nguyen, R. (2023). A Robust Model Predictive Control-Based Scheduling Approach for Electric Vehicle Charging With Photovoltaic Systems. IEEE Systems Journal, 17(1), 111–121. https://doi.org/10.1109/JSYST.2022.3183626

  • A MPC-based scheduling algorithmː charging services at an on-grid PV-powered campus parking lot.
  • Predictive models for EV charging needs and solar energy provision.
  • Maximizing the profitability of the charging station rather than optimizing the energy expense of the entire grid.
  • The one-step worst-case prediction was employed.
  • Additional optimization to boost the charging power without reducing profit → resilience against uncertainties.
  • Dynamic power load equations rather than static equivalents.
  • Improvementsː profitability, voltage stability, and the quality of service.

6. Machine Learning Approaches (Large Scale)[edit | edit source]

Prediction of electric vehicle charging-power demand in realistic urban traffic networks[27]

Arias, M. B., Kim, M., & Bae, S. (2017). Prediction of electric vehicle charging-power demand in realistic urban traffic networks. Applied Energy, 195, 738–753. https://doi.org/https://doi.org/10.1016/j.apenergy.2017.02.021

  • A time-spatial forecasting modelː EV charging-power demand in urban areas.
  • The arrival rates of discharged EVs at the stations using a Markov-chain method - traffic data (i.e., arrival rate) - EV battery specifics (e.g., SOC levels) -charging patterns are used.
  • Real-time closed-circuit television (CCTV) data from Seoul, South Korea's urban road network.
  • Traffic modelingː the teleportation method encompassing the distinctive traffic features.
  • The SOC level of the discharged EV arriving at the charging stationː by the probability density function of Gaussian distribution.
  • 3 distinct charging intervals: morning, afternoon, and evening.
  • Various parameters were effective on charging station profilesː the count of discharged vehicles arriving, the remaining SOC, charged SOC, and the charging rate of charging infrastructures.
  • Maximum charging-power demandː when all vehicles were fully charged.
  • The lowest demandː when EVs had varying SOC levels.
  • There was a potential for enhancing prediction precision for urban road networksː by leveraging detailed traffic data, accounting for disparities in charging locations and times.

Ensemble machine learning-based algorithm for electric vehicle user behavior prediction[28]

Chung, Y.-W., Khaki, B., Li, T., Chu, C., & Gadh, R. (2019). Ensemble machine learning-based algorithm for electric vehicle user behavior prediction. Applied Energy, 254, 113732. https://doi.org/https://doi.org/10.1016/j.apenergy.2019.113732

  • Predicting EV user behavior (the length of stay and energy usage, based on the entropy/sparsity ratio (R)).
  • Data from SMERC stations at UCLA between 2015 and 2017, and residential data in the UK from the EA technology website from 2014 to 2015. The research considered data from 50 UCLA users and 202 from the UK, totaling 39,458 charging records.
  • For lower R valuesː for duration predictions → support vector regression (SVR) was more accurate - for energy consumptionː random forest (RF).
  • For higher R valuesː diffusion-based kernel density estimator (DKDE) excels in both predictions.
  • The Ensemble Predicting Algorithm (EPA) incorporates SVR, RF, and DKDE. This combination leverages the strengths of each.
  • The EPA forecasts are incorporated into the best EV charging schedule algorithm → decrease load fluctuations and minimize charging expenses.
  • More accurate predictions → improved EV charging load management → reduced load variability and costs.
  • Using EPA for predictionsː the scheduling approach can cut peak load by 27% - decrease load fluctuations by 10% - save 4% in costs compared to non-strategized charging.
  • EPA method can be used at any charging station size.

Data-Driven Charging Demand Prediction at Public Charging Stations Using Supervised Machine Learning Regression Methods[29]

Almaghrebi, A., Aljuheshi, F., Rafaie, M., James, K., & Alahmad, M. (2020). Data-Driven Charging Demand Prediction at Public Charging Stations Using Supervised Machine Learning Regression Methods. Energies, 13(16). https://doi.org/10.3390/en13164231

  • Estimation of energy consumption of PEV users within the charging process.
  • Data from public charging stations in Nebraska, USA.
  • The XGBoost regression algorithm outperformed others.
  • However, all techniques achieved moderate accuracy, about 50% of the variation in user actions.
  • A significant challengeː analyzing random data.

Data-driven framework for large-scale prediction of charging energy in electric vehicles[30]Zhao, Y., Wang, Z., Shen, Z.-J. M., & Sun, F. (2021). Data-driven framework for large-scale prediction of charging energy in electric vehicles. Applied Energy, 282, 116175. https://doi.org/https://doi.org/10.1016/j.apenergy.2020.116175

  • An innovative prediction systemː enhancing the adaptability of energy forecasting - enhancing prediction accuracy under intricate real-world circumstances.
  • Linear and nonlinear contributions (like Proportion, XGBoost, Random Forest, and Neural Network).
  • MAPEs fluctuates between 2.5% and 3.8% for the testing-to-training ratio variations from 0.1 to 1000.
  • The MAPEs of the newly introduced frameworkː 18.9% lower than the existing models.
  • Broad applications for upcoming studies on EV power needs, charging strategies, effects on urban electricity grids, and policy decisions.
  • The recommended models can be expanded to evaluate the benefits of emission reductions, the influence on air quality, and the health advantages of urban EVs.

Prediction of electric vehicle charging duration time using ensemble machine learning algorithm and Shapley additive explanations[31]

Ullah, I., Liu, K., Yamamoto, T., Zahid, M., & Jamal, A. (2022). Prediction of electric vehicle charging duration time using ensemble machine learning algorithm and Shapley additive explanations. International Journal of Energy Research, 46(11), 15211–15230. https://doi.org/https://doi.org/10.1002/er.8219

  • Forecasting the charging durations for different EV categories (private or commercial) and charging methods (standard or fast).
  • Dataː 500 EVs in Japan
  • Various EML modelsː RF, CatBoost, LightGBM, and XGBoost.
  • Several factors were affecting the charging time of EVsː starting SOC - ending SOC - ambient lighting conditions - day of the week - time, season - the use of A/C and heater.
  • XGBoostː the most effective EML algorithm.
  • Feature importance analysisː starting SOC, ending SOC, season, and ambient lighting.

Dynamic energy scheduling and routing of a large fleet of electric vehicles using multi-agent reinforcement learning[32]Alqahtani, M., Scott, M. J., & Hu, M. (2022). Dynamic energy scheduling and routing of a large fleet of electric vehicles using multi-agent reinforcement learning. Computers & Industrial Engineering, 169, 108180. https://doi.org/https://doi.org/10.1016/j.cie.2022.108180

  • Optimal operation of an EV fleet for V2G project.
  • A Decentralized Markov Decision Process Reformulation approach, leveraging Multi-Agent Reinforcement Learningː to enhance the efficiency of the EV fleet in V2G task.
  • The Multi-Agent Reinforcement Learning approach vs. three heuristic methods (Genetic, Particle Swarm, and Artificial Fish Swarm techniques)ː lower MAE rates, with savings of 54.32%, 14.94%, 16.85%, 1.22%, and 15% against the Genetic, Particle Swarm, Artificial Fish Swarm, Simulated Annealing, and Differential Evolution strategies.
  • The Multi-Agent Reinforcement Learning techniqueː simulation time < Deep Reinforcement Learning model, Genetic, Particle Swarm, Artificial Fish Swarm, and Differential Evolution.

  1. Tavakoli, A., Saha, S., Arif, M. T., Haque, M. E., Mendis, N., & Oo, A. M. (2020). Impacts of grid integration of solar PV and electric vehicle on grid stability, power quality and energy economics: A review. IET Energy Systems Integration, 2(3), 243-260.
  2. Barbose, G., Darghouth, N., Millstein, D., Cates, S., DiSanti, N., & Widiss, R. (2016). Tracking the sun IX: The installed price of residential and non-residential photovoltaic systems in the United States.
  3. Feldman, D., Barbose, G., Margolis, R., Bolinger, M., Chung, D., Fu, R., ... & Wiser, R. (2015). Photovoltaic system pricing trends: historical, recent, and near-term projections 2015 edition.
  4. Jamil, U., Bonnington, A., & Pearce, J. M. (2023). The Agrivoltaic Potential of Canada. Sustainability, 15(4), 3228.
  5. 5.0 5.1 5.2 https://www.ren21.net/gsr-2022/
  6. Abdelhamid, M., Singh, R., Qattawi, A., Omar, M., & Haque, I. (2014). Evaluation of On-Board Photovoltaic Modules Options for Electric Vehicles. IEEE Journal of Photovoltaics, 4(6), 1576–1584. https://doi.org/10.1109/JPHOTOV.2014.2347799
  7. Coffman, M., Bernstein, P., & Wee, S. (2017). Integrating electric vehicles and residential solar PV. Transport Policy, 53, 30–38. https://doi.org/https://doi.org/10.1016/j.tranpol.2016.08.008
  8. Brenna, M., Dolara, A., Foiadelli, F., Leva, S., & Longo, M. (2014). Urban Scale Photovoltaic Charging Stations for Electric Vehicles. IEEE Transactions on Sustainable Energy, 5(4), 1234–1241. https://doi.org/10.1109/TSTE.2014.2341954
  9. Roselli, C., & Sasso, M. (2016). Integration between electric vehicle charging and PV system to increase self-consumption of an office application. Energy Conversion and Management, 130, 130–140. https://doi.org/https://doi.org/10.1016/j.enconman.2016.10.040
  10. Grande, L. S. A., Yahyaoui, I., & Gómez, S. A. (2018). Energetic, economic and environmental viability of off-grid PV-BESS for charging electric vehicles: Case study of Spain. Sustainable Cities and Society, 37, 519–529. https://doi.org/https://doi.org/10.1016/j.scs.2017.12.009
  11. Huang, P., Lovati, M., Zhang, X., Bales, C., Hallbeck, S., Becker, A., Bergqvist, H., Hedberg, J., & Maturi, L. (2019). Transforming a residential building cluster into electricity prosumers in Sweden: Optimal design of a coupled PV-heat pump-thermal storage-electric vehicle system. Applied Energy, 255, 113864. https://doi.org/https://doi.org/10.1016/j.apenergy.2019.113864
  12. Bartolucci, L., Cordiner, S., Mulone, V., Santarelli, M., Ortenzi, F., & Pasquali, M. (2023). PV assisted electric vehicle charging station considering the integration of stationary first- or second-life battery storage. Journal of Cleaner Production, 383, 135426. https://doi.org/https://doi.org/10.1016/j.jclepro.2022.135426
  13. Liu, X., Fu, Z., Qiu, S., Li, S., Zhang, T., Liu, X., & Jiang, Y. (2023). Building-centric investigation into electric vehicle behavior: A survey-based simulation method for charging system design. Energy, 271, 127010. https://doi.org/https://doi.org/10.1016/j.energy.2023.127010
  14. Denholm, P., Kuss, M., & Margolis, R. M. (2013). Co-benefits of large scale plug-in hybrid electric vehicle and solar PV deployment. Journal of Power Sources, 236, 350-356.
  15. Boström, T., Babar, B., Hansen, J. B., & Good, C. (2021). The pure PV-EV energy system – A conceptual study of a nationwide energy system based solely on photovoltaics and electric vehicles. Smart Energy, 1, 100001. https://doi.org/https://doi.org/10.1016/j.segy.2021.100001
  16. Assolami, Y. O., Gaouda, A., & El-Shatshat, R. (2021). A New Framework for Plug-In Electric Vehicle Charging Models Supported by Solar Photovoltaic Energy Resources Un nouveau cadre pour les modèles de recharge du véhicule électrique rechargeable supporté par des ressources d’énergie solaire photovoltaïque. IEEE Canadian Journal of Electrical and Computer Engineering, 44(2), 118–129. https://doi.org/10.1109/ICJECE.2020.3008689
  17. Sun, B. (2021). A multi-objective optimization model for fast electric vehicle charging stations with wind, PV power and energy storage. Journal of Cleaner Production, 288, 125564. https://doi.org/https://doi.org/10.1016/j.jclepro.2020.125564
  18. Kumar, K. N., Cheah, P. H., Sivaneasan, B., So, P. L., & Wang, D. Z. W. (2012). Electric vehicle charging profile prediction for efficient energy management in buildings. 2012 10th International Power & Energy Conference (IPEC), 480–485. https://doi.org/10.1109/ASSCC.2012.6523315
  19. Ai, S., Chakravorty, A., & Rong, C. (2018). Household EV Charging Demand Prediction Using Machine and Ensemble Learning. 2018 IEEE International Conference on Energy Internet (ICEI), 163–168. https://doi.org/10.1109/ICEI.2018.00037
  20. Lu, Y., Li, Y., Xie, D., Wei, E., Bao, X., Chen, H., & Zhong, X. (2018). The Application of Improved Random Forest Algorithm on the Prediction of Electric Vehicle Charging Load. Energies, 11(11). https://doi.org/10.3390/en11113207
  21. ur Rehman, H., Korvola, T., Abdurafikov, R., Laakko, T., Hasan, A., & Reda, F. (2020). Data analysis of a monitored building using machine learning and optimization of integrated photovoltaic panel, battery and electric vehicles in a Central European climatic condition. Energy Conversion and Management, 221, 113206. https://doi.org/https://doi.org/10.1016/j.enconman.2020.113206
  22. Shahriar, S., Al-Ali, A. R., Osman, A. H., Dhou, S., & Nijim, M. (2021). Prediction of EV Charging Behavior Using Machine Learning. IEEE Access, 9, 111576–111586. https://doi.org/10.1109/ACCESS.2021.3103119
  23. Dorokhova, M., Martinson, Y., Ballif, C., & Wyrsch, N. (2021). Deep reinforcement learning control of electric vehicle charging in the presence of photovoltaic generation. Applied Energy, 301, 117504. https://doi.org/https://doi.org/10.1016/j.apenergy.2021.117504
  24. Bertolini, A., Martins, M. S. E., Vieira, S. M., & Sousa, J. M. C. (2022). Power output optimization of electric vehicles smart charging hubs using deep reinforcement learning. Expert Systems with Applications, 201, 116995. https://doi.org/https://doi.org/10.1016/j.eswa.2022.116995
  25. Zahedi, R., hasan Ghodusinejad, M., Aslani, A., & Hachem-Vermette, C. (2022). Modelling community-scale renewable energy and electric vehicle management for cold-climate regions using machine learning. Energy Strategy Reviews, 43, 100930. https://doi.org/https://doi.org/10.1016/j.esr.2022.100930
  26. Yang, Y., Yeh, H.-G., & Nguyen, R. (2023). A Robust Model Predictive Control-Based Scheduling Approach for Electric Vehicle Charging With Photovoltaic Systems. IEEE Systems Journal, 17(1), 111–121. https://doi.org/10.1109/JSYST.2022.3183626
  27. Arias, M. B., Kim, M., & Bae, S. (2017). Prediction of electric vehicle charging-power demand in realistic urban traffic networks. Applied Energy, 195, 738–753. https://doi.org/https://doi.org/10.1016/j.apenergy.2017.02.021
  28. Chung, Y.-W., Khaki, B., Li, T., Chu, C., & Gadh, R. (2019). Ensemble machine learning-based algorithm for electric vehicle user behavior prediction. Applied Energy, 254, 113732. https://doi.org/https://doi.org/10.1016/j.apenergy.2019.113732
  29. Almaghrebi, A., Aljuheshi, F., Rafaie, M., James, K., & Alahmad, M. (2020). Data-Driven Charging Demand Prediction at Public Charging Stations Using Supervised Machine Learning Regression Methods. Energies, 13(16). https://doi.org/10.3390/en13164231
  30. Zhao, Y., Wang, Z., Shen, Z.-J. M., & Sun, F. (2021). Data-driven framework for large-scale prediction of charging energy in electric vehicles. Applied Energy, 282, 116175. https://doi.org/https://doi.org/10.1016/j.apenergy.2020.116175
  31. Ullah, I., Liu, K., Yamamoto, T., Zahid, M., & Jamal, A. (2022). Prediction of electric vehicle charging duration time using ensemble machine learning algorithm and Shapley additive explanations. International Journal of Energy Research, 46(11), 15211–15230. https://doi.org/https://doi.org/10.1002/er.8219
  32. Alqahtani, M., Scott, M. J., & Hu, M. (2022). Dynamic energy scheduling and routing of a large fleet of electric vehicles using multi-agent reinforcement learning. Computers & Industrial Engineering, 169, 108180. https://doi.org/https://doi.org/10.1016/j.cie.2022.108180
Cookies help us deliver our services. By using our services, you agree to our use of cookies.