Artificial Intelligence-based Battery Life Estimation of Electric Vehicle-Presentation-1.pptx

Artificial Intelligence-based Battery Life Estimation of Electric Vehicle-Presentation-1.pptx

1. 1 Final Year Project Progress Presentation 2K20 – Session “Second Presentation”

2. “ Artificial Intelligence-based Battery Life Estimation of Electric Vehicle ” Supervised by: Your Supervisor Name 2

3. Group Members XYZ – CGPA ( x.xx ) (Group Leader) ABC – CGPA ( x.xx ) ABC – CGPA ( x.xx ) ABC – CGPA ( x.xx ) 3

4. Table of Contents Chapter 1: Abstract/Introduction Objectives Chapter 2: Literature Review Chapter 3: Proposed Methodology and Modeling Chapter 4: Software Advantages and Disadvantages/Applications Energy And environmental concerns Proposed System/Project Planning System Configurations/System Components Chapter 5: Conclusion References 4

5. Research Area Research Domain Electric Power Distribution: Electric power distribution is the final stage in the delivery of electric power; it carries electricity from the distribution system to individual consumers i.e. EVs. Area of Interest Artificial intelligence, or AI: Artificial intelligence, or AI, has the potential to cut energy waste, lower energy costs, and facilitate and accelerate the use of clean renewable energy sources in power grids worldwide. AI can also improve the planning, operation, and control of power systems. 5

6. introduction When many people consider the environmental damage caused by automobiles, the first thing that comes to mind is pollution. The use of conventional automotive gasoline contributes significantly to environmental challenges such as global warming and smog. People’s attention is once again being directed to the increasing trend of new energy vehicles, notably electric-powered ones, to achieve this goal [1]. Currently, every in the world is researching EV, and some of them have noted that issues with Battery Management systems (BMS) are stifling progress. Advanced nations have implemented identical legislation to stimulate battery development [2]. China’s EV industry has been strengthened by the passage of the nation’s, and Five-Year Plans. The safety of , the longevity of their batteries, cost savings, and increased driving range are all determined by how efficiently the batteries are handled [3]. 6

7. introduction Because of the many advantages of lithium-ion battery packs, including their high discharge rate and lack of memory effect, many people are interested in EVs that use them as their primary power source [4]. Aged batteries limit the performance of EVs in terms of energy storage and power generation, as well as cost and lifespan [5]. The EV industry is battling with battery life management and asset appraisal. 7

8. introduction Electric vehicles (EVs) are the most useful by-product of renewable technologies in the transportation domain due to their eco-friendly and user-friendly nature. EVs that run on batteries rely on Lithium-ion (Li-ion) cells due to their high capacity for energy and longer lifecycle. Over time or with continued use, any system, no matter how intricately constructed, the batteries will degrade. The expense associated with a failed Li-ion battery is considerable. For these uses, determining battery health and estimating battery life is the sole essential feature required to improve EV reliability and protection. Because of the nonlinear battery capacity fading with small fluctuations in early cycles, most previous research yielded unreliable predictions for the future. 8

9. Motivation In this Final Year Project, we present a comprehensive Artificial Intelligence. Framework for making reliable predictions about battery life cycles. This is accomplished by acquiring data and then removing unwanted features and extracting eight of the most salient features. Two recommended types of AI models employed here. Finding an accurate battery pack life prediction method is crucial for improved performance detection and evaluation after EV installation. 9

10. Scope & aim Provides an electric vehicle situation where battery operating data from vehicle application is collected. Battery data is collected but it is not necessarily obvious how the data can be used in order to improve future battery operation. Use of the available field data trying to condensate information on the battery from the raw signals and in this way render the data meaningful information. Measure battery performance evaluated, and it became apparent that in order to reach the initial goal of evaluating battery usage on-board electric vehicles, a data-driven method to access battery performance measures from the raw data is needed. To develop a method for on-board SOH estimation. 10

11. Objectives A battery’s capacity and internal resistance are both variables in deciding how long it will survive. Estimating the SOH, or state of health, of the battery, is critical for evaluating how much the battery has declined with age. When a battery’s capacity falls below a certain threshold, it reaches end-of-life (EOL) and ceases to function correctly. All experimental model-based and data-driven methods exist for making such predictions. The experimental model-based technique evaluates and predicts performance using experimental information. Because the generic empirical model is an open-loop model, the estimation results can be incorrect. As a result, the goal of this work is to provide an AI-based estimation of battery life. 11

12. Flow Diagram 12

13. FLOW DIAGRAM 13

14. 14 Proposed System FLOW

15. 15 Proposed System

16. 16 Proposed System

17. 17 Proposed System

18. 18 Proposed System

19. 19 Simulations on MATLAB Simulink This Simulation shows how to use a feedforward deep learning network inside a Simulink® model to predict the state of charge (SOC) of a battery. We include the network inside the Simulink model by using a Predict block, which predicts the SOC at every simulation step. Battery SOC is the level of charge of an electric battery relative to its capacity measured as a percentage. SOC is critical information for the vehicle energy management system and must be accurately estimated to ensure reliable and affordable electrified vehicles.

20. 20 Simulations on MATLAB Simulink Methods based on the Kalman filter (EKF) algorithm are the traditional approaches to this problem but they usually require precise parameters and knowledge of the battery composition as well as its physical response. In contrast, using neural networks is a data-driven approach that requires minimal knowledge of the battery or its nonlinear behavior [1]. The FYP uses a trained feedforward neural network to predict the SOC of a Li-ion battery, given time series data representing various features of the battery such as voltage, current, temperature, average voltage, and average current. For an example showing how to train the network, see Predict Battery State of Charge Using AI.

21. 21 Simulations on MATLAB Simulink BatterySOCSimulinkEstimation_ini ; modelName = ' BatterySOCSimulinkEstimation ’; open_system ( modelName ); sim(' BatterySOCSimulinkEstimation ’); BatterySOCSimulinkEstimation_plot_inputs ; BatterySOCSimulinkEstimation_plot_outputs ;

22. 22 Simulations on MATLAB Simulink

23. 23 Simulations on MATLAB Simulink

24. 24 Simulations on MATLAB Simulink 12.8 V, 40 Ah, Lithium-Ion (LiFePO4) Battery Aging Model (1000 h Simulation)

25. 25 Simulations on MATLAB Simulink 12.8 V, 40 Ah, Lithium-Ion (LiFePO4) Battery Aging Model (1000 h Simulation) This FYP illustrates the effect of aging (due to cycling) on the performance of a 12.8 V, 40 Ah Lithium-Ion battery model. The battery is submitted for 1000 hours, to several discharge-charge cycles at ambient temperature of 25 degrees C , and at various depths of discharge (DOD) and discharge rates. As observed from the Scope, the impact of DOD and discharge rate on the battery life is as expected. As the DOD or discharge rate increases, the battery ages rapidly, which quickly reduces the battery capacity.

26. 26 Simulations on MATLAB Simulink (Plot Input)

27. 27 Simulations on MATLAB Simulink (Plot and Analyze Output)

28. 28 Simulations on MATLAB Simulink Simulation Results The plot below shows the real and estimated battery state-of-charge, estimated terminal resistance, and estimated state-of-health of the battery.

29. 29 Results (Result on the V Charging profiles)

30. 30 Results (Result on the I Charging profiles)

31. 31 Results (Result on the T Charging profiles)

32. Software 32 MATLAB SIMULINK 2021a

33. Conclusions Numerous methods have been proposed for health diagnostics and prognostics of Liion cells. There is no single method to solve all current issues. A trade-off between the accuracy, computational effort and generalizability is usually required for each particular application. To better understand these trade-offs, this section summarizes and compares the characteristics of the existing data-driven methods. Based on their comparison, the challenges of the up-and-coming technologies based on data-driven battery health diagnostics and prognostics 33

34. References 1. Sun, X.; Li, Z.;Wang , X.; Li, C. Technology development of electric vehicles: A review. Energies 2020, 13, 90. 2. IEA. Global EV Outlook 2022. IEA 2022. May 2022. Available online: https://www.iea.org/reports/global-ev-outlook-2022 (accessed on 5 October 2022). 3. Wang, Z.; Wang, S. Grid power peak shaving and valley filling using vehicle-to-grid systems. IEEE Trans. Power Deliv. 2013, 28, 1822–1829. 4. Rotering , N.; Ilic, M. Optimal charge control of plug-in hybrid electric vehicles in deregulated electricity markets. IEEE Trans. Power Syst. 2011, 26, 1021–1029. 5. Mohammad, A.; Zamora, R.; Lie, T.T. Integration of electric vehicles in the distribution network: A review of pv based electric vehicle modelling. Energies 2020, 13, 4541. 34

35. Any Questions? 35

36. Thank You! 36