Emerging Intelligent Bidirectional Charging Strategy Based on Recurrent Neural Network Accosting EMI and Temperature Effects for Electric Vehicle

Transportation is currently advancing towards Electric Vehicle (EV) technology. This paper presents a brief and systematic analysis of the real-time issues obtained in Electric Vehicles (EVs) due to the various ranges of energy storage devices. In general, EV energy management system is integrated with power electronic circuits for effective power conversion and reliable operation. Some issues are addressed while using the batteries in EV systems such as charging time, efficiency of battery, and raw materials. Not only battery issues but also real-time non-technical issues and operational issues are also discussed in this paper. During energy conversion with power electronic circuits, the system attains extreme temperature levels which in turn reduces the performance of the system. To maintain an optimum temperature level, it is important to study the temperature effect of the system at the most prior levels. Due to the adaption of power electronic components, some extent of noise is generated, technically treated as Electromagnetic Interference (EMI), as system capacity increases the EMI content also improved proportionately. Therefore, to mitigate the EMI effect, the low pass filter-based EMI filter is introduced in the system such that the noise level is suppressed. Bidirectional Charging System (BCS) is one of the emerging technologies in EV to obtain autonomous power supply systems in the form of Vehicle to Grid (V2G), Grid to Vehicle (G2V), and Vehicle to Load (V2L). To know the behaviour of BCS the proposed RNN controller is employed and is compared with ANN bidirectional charging model. BCS charging system with RNN controller has better dynamic response to exchange the power via DC/DC converter and AC/DC converter as compared to ANN controller.


FIGURE 1. Basic Components of EV.
are suitable alternatives to replace combustion vehicles. EVs are not just a method of transportation, but also referred as new generation of electric equipment, allowing for clean and efficient transport [3], [7].
The EV architecture consists of power electronic converter, control unit, electric motor, and battery pack. EV is used as convenient load to stabilize the grid when there is significant proportion of fluctuating renewable energy supply [8]. The components of electric vehicles are shown in Fig.1. The specific EV design considerations are described as listed below [9].
1. Recognizing the EV environment and industry trends. 2. Evaluating the infrastructure needed for design, as well as integrating battery recycling. 3. Defining its power source for various scenarios such as production or reserve, standalone or hybrid. 4. Identifying the essential components of an electric vehicle propulsion system, such as single or multiple motors, converters, transmission systems, and installation methodologies.
In [10], developed an integrated controller for charging and driving the EVs. The controller permits the controller to concurrently accomplish motor drive and onboard charging for the vehicle. Authors used TMS320F28335-based experimental platform. A resistive load replicated the power battery to validate charging mode, and a 5 kW PMSM verified driving mode. Li et al. [11], examined the lifetime design of Plug-in hybrid electric vehicle, a hybrid electric vehicle like Battery electric vehicle batteries in domestic applications. The author experimented on second-use batteries based on residential load, and outdoor temperature. The quality of battery is tested for nearly 150Hours in software simulation and experimental ways.
In [12], developed a novel Game theory-based theoretical approach for congestion problem control. The authors used a sophisticated method of pricing, as well as the development of a plan to alleviate congestion in the network between Plugin EVs and the utilized grid. Awami et al. [13], proposed the self-governing EV charge controller for mitigating the negative impacts during load fluctuations. Low-voltage nodes are more vulnerable to load variations than high-voltage nodes. The author addressed the estimation of load fluctuations depending upon the input provided to the EV source also considering various loading scenarios, distributed generators, system reconfiguration, and voltage control devices.
Zhao et al. [14], proposed the voltage-source converter for an onboard Electric Vehicle (EV) charger that is compatible with both single-phase and 3-phase grids. LCL filters are specifically designed to eliminate the harmonics absorbed by power converters. The author experimented on a 3KW prototype. The harmonics are reduced by transforming the source from star to delta pattern. Therefore, effective utilization of DC link voltage and filter capacity increases. Park et al. [15], implemented the estimated controller and applied it to the battery cooling controller for an electric vehicle. The battery life is degraded when the battery temperature is too low or high, so the author developed the above controller in proper stochastic prediction manner [16]. The proposed controller has robust performance as compared with other thermal management controllers.
Naseri et al. [17], proposed the regenerative braking system in EVs based on the balancing features of batteries and supercapacitors. Brush Less DC (BLDC) motor acts as a generator during the braking operation. The EV configuration is associated with hybrid energy storage systems, so the appropriate switching operation is performed i.e., the effective energy is transferred between super capacitor and battery through inverter. An Artificial Neural Network (ANN) is used to balance the braking force. PI controller adjusts the brake current at the same time for constant torque braking.
Fu et al. [18], proposed an output voltage control with harmonic correction and a virtual impedance term to achieve high power quality and sufficient damping for the EV. To achieve proper evaluation, the author introduced the time delay for the controller. The efficiency of the design process is validated by simulation results and experimental data on an EV battery charger prototype comprised of silicon carbide devices that includes one voltage source inverter and one isolated DC/DC converter.
In [19], developed a battery based on ultracapacitor hybrid energy storage system for the rippled load. Ultracapacitor energy storage is clearly understood by using the MATLAB/Simulink platform, after examining all the results the authors decided that the energy storage capacity of developed method is effective as compared to Lead-acid storage battery.
Stirban et al. [20], introduced a Current and frequencybased sensor less control technique for a 3-phase, 4 pole permanent magnet BLDC motor. By considering the prediction of a line-to-line permanent magnetic flux linkage, the control strategy investigation using a real-time offline finite element technique based on supported position and speed observer. A digital system is used to represent the simulation process.
Mishra et al. [21], investigated a sigma-modified power control and gain adaptive implementation for EV charging applications. Sigma-mod control is effective for the analysis of a wide range of supply voltage distortion. Lyapunov candidate analytic function is used for convergence checking process. The author successfully demonstrated the gain of adaptive control through experimental validation.
In [22], developed a novel hybrid forecasting model for electric vehicle fleet charging that is based on numerous decompositions. The Swarm Decomposition, often known as SWD, is included in the Complete Ensemble Empirical Mode Decomposition Adaptive Noise (CEEMDAN) method through the utilization of the presented strategy. Authors compared both traditional and hybrid models and concluded as the developed model performed significantly better.
Baszynski et al. [23], implemented a digital algorithm in a Field Programmable Gate Array (FPGA) approach for monitoring the high-speed functioning of a BLDC motor using the PI control mechanism. The proposed method frequency of measurement is six for two pole motor and twelve for four pole motor. The FPGA measuring system is highly efficient, and there is no need to use any additional external devices for high-speed measurements, such as sensors and tachometers, etc., which decreases the cost of BLDC motor production.
Electromagnetic (EM) field interferes with power electronic equipment, which happens when the electronic device is near the EM field and causes interference in the radio frequency band. Electric vehicle has power electronic devices, which results in developing EMI issues. These issues are clearly addressed in Table 1. This paper is organized as the real-time issues of advanced automobile technology based on Electric Vehicles (EVs). Section II described generally realtime issues in EV. The drive mechanism of EV based on Power electronic converters is depicted in Section III. The influence of various parameters to change the performance of EV is prescribed in the form of technical issues, which are reported in Section IV. Section V, concludes the main reviews/ findings of this paper.

II. MODELING OF EV
The basic dynamics of a conventional vehicle are used for the formulation of an electric vehicle. In EV traction power is supplied to the electric motor using battery as shown in Fig.2. The traction motor generates torque, which causes the power train to move. The clutch, gearbox, differential, and driveshaft comprise the EV power train transmission. Power electronic converters are helped to control the speed and torque of the motor. The auxiliary loads are incorporated with systems for comfort, lighting, and safety. Maximum voltage of the battery is checked by using braking resistor during regenerative braking [42]. From the figure, thick lines represent the high voltage DC bus and dotted lines represented the low voltage DC bus.

A. DYNAMIC EV MODEL
To know the behaviour of EV is mathematically expressed by applying the fundamental dynamic and rotational physics, especially the concepts of Newton's Law of Motion [43]. The forces acting on the vehicle body represent the state of EV as depicted in Fig. 3. Vehicle has mass m moving with a velocity ( − → V ), then the equation of motion is obtained as Vehicle bodies are influenced by external forces, which are forward tractive force − → F t rolling resistance moment − → F r , aerodynamic drag − → F u , grading resistance − → F g tractive forces of front and rear wheels are represented as − → F tF and − → F tR , vehicle velocity is positive when the vehicle is in forward direction. When the vehicle's velocity is reversed, the result is negative. Rolling reaction force is expressed as where f r is the coefficient of rolling resistance coefficient, α denotes the road angle in radians, and g is the acceleration due to gravity. Aerodynamic drag is acted upon the vehicle, vehicle velocity − → V , vehicle anterior area A f , air density ρ, and force due to wind − → F W is expressed as Velocity of wind (V W ), Aerodynamic drag coefficient (C d ). Maximum extent of dragging force (F tmax ) is required to maintain the pulling force (W ), and the road adherence coefficient is denoted as µ. The maximum tractive effort obtained as The downward opposing tractive force with grading resistance is given as The resultant force obtained from the load acting on the front and rear wheels (W F , W R ) can be given as VOLUME 10, 2022  where, corresponding lengths and heights are L a , L b , L, h g fluctuating resistance r d Y of the tire is shown in Fig.3.
After examining the traction force of various types of roads, results that the sliding of the vehicle-driven wheel is VOLUME 10, 2022  crucial for the drive wheel. Frictional movement between wheels and surface is The above equations represent the driver slip (S D ), Braking slip (S B ), speed translation at the centre of the tire (V t ), angular speed (ω) of the tire, rolling radius (r), and effective radius (r e ) indicate the above Eq. 8(a)-8(d).

B. FOUR-WHEEL PLANAR DYNAMIC MODEL
The uphill motion of the kinematic model is discussed in previous section. The horizontal or planar road, during control design problems [44]. The longitudinal, lateral, and vertical movements of the vehicle's centre of gravity are defined by x, y, and z. The yaw and rolling angles are represented by φ Y and φ R . According to Newton's law of motion the longitudinal and lateral accelerations (a X , a Y ) are formulated as Newton's law-based yaw speed on the initial instant is calculated as

C. MULTIBODY DYNAMIC MODEL
The multi-body dynamic model from Fig.5 represents various subsystem features of the couplings in the vehicle, as well as many kinematic and non-linear constraints [45]. Differential equations describing the motion of the vehicle and the applied forces, such as bushing reactions and tire forces are found to capitulate the system position variables and the Lagrange model for the joint's reactions. These equations are developed using a genetic method based on the constrained system Lagrange's equations. Legrangeous multiplier (λ), coordinate vector (q), the matrix equation is given as where, VOLUME 10, 2022

III. TECHNICAL AND NON-TECHNICAL IMPACTS OF EV
Even after producing an environmentally friendly product, this industry continues to face several barriers which are depicted in Fig. 6.
The infrastructure for EV charging to hasten the electric vehicle revolution. The energy experts stated that there are no operating license requirements for EV charging stations worldwide, which might aid in the development of a larger global network of EV charging stations. The government offers discounts to buyers of EVs, lowers the applicable rate of taxes on Li-ion batteries, and offers incentives to the public transit sector to switch to electric cars [46].

2) DEMAND RESPONSE
To ensure that all customers have access to power during times of peak energy demand, utilities may better control the electrical grid with the use of demand response (DR) and load management technologies. Although it occasionally also refers to changeable pricing based on demand, the phrase ''demand response'' is most often used to explain reducing power to certain equipment.

B. INFRASTRUCTURE ISSUES 1) BATTERY REPROCESSING
Although electric car batteries are made to last for a certain period, they will eventually degrade. Manufacturers do not disclose actual pricing for battery replacement nevertheless, if replacement is necessary after the warranty has expired, the price will rise since a new battery must be purchased. Batteries' chemical components, including lithium, nickel, cobalt, manganese, and titanium, not only increase the efficiency of the supply chain but also pose a threat to the environment [47].

2) CHARGING INFRASTRUCTURE
Based on a design perspective, EV manufacturers recognized the Chargeable batteries, so that discharge batteries can be replaced with fully charged ones. The Charging station can arrange to charge their batteries at off-peak hours when the power tariff is lower [48].

C. MARKETING ISSUES 1) HIGH CAPITAL COST
Electric vehicle battery packs are costly, and they are replaced by many times throughout their lifespan. When compared to electric automobiles, gas-powered automobiles are less expensive [49].

2) CONSUMER PERCEPTION
Consumer reputation is critical in acquiring qualified customers and maintaining existing models. Although the increasing variety of electrical vehicles, the selection of importing an EV leftover controlled and is predictable to remain in the future. According to studies, lack of awareness, financial benefits, and knowledge of vehicular technology contains a direct impact on the approval of EVs [50].

3) RAW MATERIALS FOR THE BATTERIES
EV batteries consist of Lithium (Li), nickel (Ni), phosphate (P), and manganese (Mn), along with graphite (Gr) and cobalt (Co). To filter destructive gases, catalyzers for ignition cars need platinum, rhodium, and palladium. Lithium-ion batteries alone need 7 million tons of Ni per year, which leads to an increase in lithium and cobalt use from 15-22 times [51].

4) VEHICLE SERVICING
To secure the EV, an expert is necessary to repair, troubleshoot and preserve it. They are able to use their expertise to solve the problem as soon as feasible.

D. CLASSIFICATION OF EV
The EVs provide convenient features like noise less travel, availability of required torque, less energy utilization, and preservation price minimum, thus all features are much attracted to the automobile industry to switch towards EVs. All the above-mentioned qualities are much better than IC engine-based vehicles [52]. EVs are mainly classified into two types, one is autonomous EVs (AEV) and another one is Plug-in Electric Vehicles (PEV). The classification of EV is represented as shown in Fig.7.

1) FUEL CELL ELECTRIC VEHICLE (FCEV)
FCEV layout is the same as the BEVs exchanging batteries with a hydrogen (H 2 ) fuel arrangement depicted in Fig.7.
In FCEVs, hydrogen is used as a fuel, and batteries are used as sole source of energy. The fuel cell, which converts hydrogen to electric energy, provides the electric energy required for vehicle propulsion. In FCEVs, the proton exchange membrane is a sophisticated model. Additionally, the nil notions are largely practical for FCEVs. Because the technology is still in its infancy, hydrogen fuel is expensive, and fueling the car is a laborious process [53].

2) FUSION/HYBRID ELECTRICAL VEHICLE (HEV)
To increase the driving range of BEVs, HEVs are created by fusing an ICE and an EV power train. The main parts of an HEV are the ICE, electric motor battery, fuel tank, and control system [54]. The classification of HEV is depicted in Fig.7.  iii. Control system: Oversees the electric drivetrain and internal combustion engine while effectively optimizing energy storage systems.

3) BATTERY ELECTRIC VEHICLE (BEV)
A battery serves as the only energy source used for vehicle propulsion in a BEV, or pure electric vehicle. BEVs are powered by huge battery packs that emit no pollutants. The battery is charged through an external outlet or the grid. Motors with high torque are used in BEVs to provide the needed torque, resulting in a loss in powertrain performance. loss in mechanical transmission and vehicle weight is quite low. The architecture is fairly basic, with a restricted driving range that is determined by the battery pack. These are ideal for short-distance city journeys with frequent halt driving patterns [55].

4) PLUG IN HYBRID ELECTRIC VEHICLE (PHEV)
PHEVs have the same design as HEVs, excluding highefficiency onboard battery packs (see Fig. 8). Driving range adapters are another name for PHEV devices. Charge depletion mode is used in PHEVs until the current battery SOC limit was reached After that, leave the vehicle in charge maintaining mode until it is fully charged. The combination of charge depletion and sustaining modes leads to more complicated control methods as compared to HEVs. PHEVs produce less pollution and use less fuel than HEVs [56].

E. POWER ELECTRONIC DRIVER
Inside an electric propulsion system, power electronic devices are the most important components for improving system efficiency, a power switching device with closed-loop control and a switching strategy is used. The switching operation of thyristors is replaced with new power devices such as bipolar junction transistors (BJT), metal oxide semiconductor field effect transistors (MOSFET), gate turn-off thyristors (GTO), and insulated gate bipolar transistors (IGBT). For electric propulsion vehicles, there are three main types of semiconductor devices often employed [57], [60]. After examining the properties of all power devices, IGBTs are gaining attention for use in electric vehicles. IGBTs have several benefits over other devices, including higher BJTs having excellent conductivity and high-power density, as well as great efficiency, compactness, and a cheap cost of useable power. Each module has six thyristors, and the driving circuit is included inside the same package [61].

F. RELIABILITY OF POWER ELECTRONICS
The fundamental enabler is power electronics technology to improve EV/HEV self-sustained technology in the form of drives. Different power levels and voltage levels i.e., 12V and 450 V are involved in EV/HEV conversion. Therefore, excellent reliability of power electronic components is essential for EV/HEV applications to meet the expected acceptable failure rate, operational behaviour, and safety levels [62]. The reliability of the power electronic converter is depicted in Fig.8. The important power electronic sources used in EVs are DC-DC converters, Rectifiers, and Inverters. In the case of PEVs used in residential settings, the grid voltage that is taken through the plug is either 110V/60Hz or 220V/50Hz. single-phase ac input.

G. CHARGING TIME
The driving range is a problem that is strongly related to the difficulty of charging time. The EV can take up to 8 hours to completely charge from nothing using a 7 kW charging station and a slow charger [63]. The duration of time required to charge the battery is mostly influenced by its size. In addition,  the rate of the charging station is directly proportional to the battery charging time. The comparison of various switching methods is described in Table 2.
To maintain less amount of charging time rapid chargers are developed. Commercially available electric vehicles can use charge stations with a greater maximum charge rate. This means that the battery is being charged at its maximum rate without any problems. As the temperature drops, the charging rate of the battery with a quick charger decreases [64], [66]. The charging speed of an EV battery is used to classify the chargers observed in Table 3.

H. EV SECURITY NECESSITIES
The electric vehicle must meet the safety criteria of the area regulations. Temperature, short circuit, fire impact, vibration, humidity, and water immersion are just a few of the conditions that the batteries must survive. Not only the above safety requirements but also collision detection, short circuit detection, and insulation should all be included in high-voltage transmission lines [67].

I. ENVIRONMENTAL IMPACT
Electric vehicles do not ruin the environment in general, but the components of the batteries are quarried or harvested from saline in the desert. This extraction has a low environmental impact on mining [68].

J. DRIVING RANGE
EVs have a lower driving range than comparable IC engine vehicles. Furthermore, the driving range is acknowledged as VOLUME 10, 2022 the most important factor. The range of an EV on a full charge or tank is seen as a significant disadvantage when compared to IC engine vehicles [69]. The above-mentioned technical barriers are represented as shown in Fig.9.

K. BATTERY PERFORMANCE
Electric vehicles are manufactured with electric motors, batteries, chargers and controllers by replacing energy tanks and gasoline engines of usual vehicles. At present, most manufacturers are offering eight years/120,600-mile warranty for their batteries [70], [71].

IV. ELECTROMAGNETIC INTERFERENCE (EMI)
In addition, the operational technical barriers of EVs are mitigation of EMI, EMC, and switching losses produced by the power electronic converters. The harmonic currents are responsible for these issues [72]. PWM-based controllers and DSP based DC-DC signal controllers are used to reduce the above-mentioned operational issues, which are represented as shown in Fig. 10(a)-(b). Two power switches (S I 1 and S I 2 ) are available for managing the converter, C 1 and C 2 (two capacitors) are used to reduce ripple voltage, L is an interconnected inductor, and the DC/DC converter is used to convert between the DC source of the battery pack (V dc Battery ) and the DC link voltage (V dc link ) in a bidirectional manner.
The DC/DC converter acts as a boost type for charging the battery by appropriately managing S I 1 to regulate the voltage and switching off S I 2 to work as a freewheeling diode. The AC/DC converter is also utilized as a bidirectional converter for the DC link voltage and the AC voltage V ac of the power grid, The H-bridge architecture as illustrated in Fig. 10 (a). The switches can be grouped in pairs to transition from DC to AC power. The PWM-driven S 2 and S 4 as well as the switched-off S 1 and S 3 change the positive portion of the resulting AC power, but the negative component can be regulated in the reverse direction, with PWM pulses driving S 1 and S 3 and switching off S 2 and S 4 .
The converter operates as a full-bridge uncontrolled rectifier when all power switches are off, by converting power from one source to another [73]. The inductive and Capacitive filters i.e., L f and C f are single-stage EMI filters with damping branch circuits R d and C b , where R d is the damping resistor and C b is the dc-blocking capacitor, as shown in Fig. 10 (b). The battery impedance, the filter input impedance, the filter output impedance, and the converter input impedances are represented as Z Batt , Z IF , Z OF , and Z IC , respectively.
The frequency spectrum of the current disturbance is shown in Fig.11.The theoretical asymptotic envelope of amplitude spectrum is also shown in the same picture. The theoretical asymptotic envelope was determined by assuming rise time (t R ) = fall time (t F ) = 50 ns; it has a first cutoff frequency of f1 = 1/ (π t R ) = 636 kHz. The present frequency spectrum includes considerable frequency content up to f2. Electromagnetic interference (EMI) is one of the most difficult problems to solve in the manufacturing of any electronic equipment. To comprehend and quantify the electromagnetic compatibility (EMC) of the instrument under test, the influence of these inevitable interferences on the instrument performance is carefully monitored. If the System EMI profile fails to satisfy accepted criteria, measures must be taken to mitigate the impact of these undesirable interferences so that the equipment may be utilized in an actual way [74], [78]. The electromagnetic interference (EMI) associated with a device is evaluated in order to assess its immunity and the quantity of interference.
EMI causes electrical equipment performance is degraded by introducing unnecessary currents and voltages in the circuitry. The behavior of EMI/EMC with respect to voltage and current is depicted in Fig.12.

V. TEMPERATURE EFFECT
Extreme temperature causes severe limitations to EV performance and its charging behaviour, the battery management system plays important role in the thermal sensitivity of the battery [79].

A. BATTERY MODEL
An artificial neural network (ANN) heating model, an ampere-hour (Ah) counter model, and an empirical voltage model make up the battery model. There are four sub modules: such as charging, discharging, high temperature, and low temperature are included in the voltage model. The input to the model is current, battery temperature, and State of Charge (SOC). The output voltage (V cell ) of the battery depends on the number of cells in series (N S ), the battery temperature ( 0 C) and SOC is given as  The neural network model evolves the Battery temperature during typical usage it takes the temperature of the battery and the state of charge as inputs and returns the batteries rate of heating (in 0 C/ S). The internal heating of the battery is the product of heating rate (in 0 C/ S) given by ANN ψ with the battery thermal mass (m b ) in (J/K) is The SOC of battery is given by t is the length of time in step, η C columbic charging and discharging efficiency, I denote current, while C is the battery pack's capacity in Ah. The Ah model does not consider battery age or deterioration.

B. THERMAL MODEL
The thermal model for the EV is a lumped capacitance model. The energy balance equations with respect to battery VOLUME 10, 2022

temperature (T b ) is
Cabin temperature (T C ) is where, m b , m C is the thermal of element (i), K is the heat transfer coefficient, q HVAC is the actual heat flow due to HVAC operation. q BTMS is the actual heat flow due to BTMS operation. η inv+EM are the combined efficiency of inverter and electric motor. Battery thermal management system plays an important for getting the desired output. The initial battery temperature is determined in Table 4. In EV, BTMS fails if the battery overheats (TB exceeds +60 0 C). In practice, if BTMS fails, then the battery is overheated and lost thermal capacity results in degrading the performance of the battery. Therefore, BTMS maintains for battery at a temperature is less than +20 0 C.

VI. BIDIRECTIONAL CHARGING STRATEGY
Bidirectional Electric Vehicle (BEV) charging systems achieve fast and dynamic response due to effective usage of power transfer [80], [82]. BEV is not only collected power from the grid to charge the EV battery, but also supplies power back to the grid. Depending upon the type of charging application BEV is classified as Vehicle to Grid (V2G), Vehicle to Home (V2H), and Vehicle to Load (V2L) charging system. Therefore, consumers mostly prefer BEV chargingbased systems.

A. VEHICLE TO GRID (V2G)
In V2G charging scheme, the energy is absorbed from vehicle battery and simply delivered it to the utility grid. BEV-V2G charging system is depicted in Fig. 13(a).

B. VEHICLE TO HOME (V2H)
In V2H charging scheme, the energy from the battery is sent to the home or any other existing building. The conversion of energy is facilitated by DC to AC converter system which is internally linked with EV. Fig. 13(b), represents the V2H charging scheme in BEV.

C. VEHICLE TO LOAD (V2L)
V2L technology includes an onboard converter in the vehicle, which converts DC from the battery to AC. In V2L charging system the energy is transferred from EV battery to other vehicle or any other external electric loads. V2L charging scheme in BEV is represented as shown in Fig. 13 (c).
Bidirectional Charging System (BCS), completely monitors the active and reactive power flow between the grid and the EV battery and vice versa. BCS charger has two conversion stages one is AC to DC with the help of bridge rectifier and DC to DC using Buck-Boost Converter (BBC). The purpose of using controllers is to provide a proper power supply to the utility system. It is necessary to control the reactive power for regulating the grid voltage by incorporating the voltage-controlled strategy using AC to DC converter. The excess amount of reactive power is delivered to the DC link capacitor. In DC to DC converter the voltage is regulated and further involves in different modes of operation.

D. EXISTING BCS USING ARTIFICIAL NEURAL NETWORK
Bi-directional EV charging data is collected at regular intervals of time for ten days. These observations are considered in the form of the virtual system using MATLAB/ Simulink process. ANN has two outputs, first signal controls the voltage levels through rectifier and DC link capacitor. Second signal control the charging current from the BBC. The important parameter observed using ANN strategy in BCS is to calculate the error between DC link voltage and the preset DC link voltage. The calculated error is sent to ANN training system for further computations which are in the form of phase shift angle. ANN-based BCS setup is depicted as shown in Fig. 14. The device built here has two inputs and two outputs (reference DC-link voltage and charging current). The training parameters are listed in table 5. DC link voltage and charging current are provided for the input as the ANN controller. The controller equations of BBC are mentioned in Eq. (25)- (32). The key operating equations for buck mode are from eq. (25)- (28), whereas boost mode is from Eq. (29)- (32).
EV battery voltage across the dc link is obtained as The charging current is obtained by integrating Eq. (25), and the complete voltage discharging cycle is given by  For a complete charging cycle, I Charging is given by V DC,Link is input to the system, which is represented as From the above equations, a positive sign is representing charging and a negative sign for discharging the operation. ANN training process requires minimal error obtained between target result and the obtained result. The training process is based on classic Lavenberg-Marquardt (LM) based technique. The summative error is obtained as For any weight, the updated algorithm is given by Here, γ K = W P K ,N · β K · δ K is treated as LM parameter.

E. BCS USING PROPOSED RECURRING NEURAL NETWORK TECHNIQUE
A recurrent neural network (RNN) is the modified form of ANN adapted for time series or sequential data. Purpose of  using RNN in BCS of EV is to introducing the huge number of data with accurate propagation, i.e., multiple set of data can be acceptable in RNN. The training process is based on many to many data exchanging process depending on the type of input is provided for the training or computation process. The input to the RNN system is DC link voltage and charging current (same as ANN). In ANN the system is limited to provide less quantity of error data, which results to increase the computation time and computation is non-uniform. RNN allows maximum number of error data and provide accurate computation levels to the system. During training process many hidden layers are participated in the error reduction process. The calculated error between actual battery current and reference battery current is minimum. Therefore, the battery system in EV is enhanced and improves the quality of the power conversion. The bidirectional charging system with RNN technique is depicted as shown in Fig. 15. The working parameters of RNN is observed in Table 6. VOLUME 10, 2022 For all 50 generations, the error value of the ANN has been held constant from 0.1 to 0.06. Error levels for the proposed RNN approach have been lowered consistently to between 0.1 and 0.02. As a result of this, it has been established that the proposed RNN controller has a shorter amount of time to charge the system. The Mean Square Error (MSE) values of controllers are shown in Fig. 16, represents the performance of charging system with 50 generations. The error value is continuously dropping from 0.5 to 0.1 for less than 20 generations.

VII. CONCLUSION
This paper presents about real time general issues, operational issues, Battery charging and discharging phenomena, power electronic converter internal issues such as EMI, temperature effect and BCS conversion in EVs. The classification of various EVs such as FCEV, HEV, BEV, and PHEV are depicted in Fig.7. The switching operation of each IGBT, BJT, and MOSFET are listed in Table 2. EMI, EMC, and harmonic currents are mainly responsible to minimizing the performance of EV. In circuitry, using the EMI filters and PWM based controllers are helpful to shave the peaky harmonics in the system. In this connection, it is benchmarked that the power electronic controllers mitigate the EMI/EMC and unnecessary currents in the circuitry also helpful to the other existing electronic devices. Advanced bidirectional charging system is incorporated to improve the performance of EV charging system. The power electronic components are using in the BCS, which lead to produce the error from the system, technically termed as Mean Square Error. The existing ANN controller acts as BCS charging system to reduce MSE value, but consumes much time due to uneven function of each generation. The results obtained in BCS using ANN controller is not optimum. Therefore, BCS is effectively improved using proposed RNN technique-based controller. The MSE error is eliminated in quick instant of time with a smaller number of generations. When compared to the ANN controller, the proposed RNN technique has a greater capacity to store information for each iteration.
GAURAV (Student Member, IEEE) received the B.Tech. degree in electrical engineering from Dr. A. P. J. Abdul Kalam Technical University (AKTU), Greater Noida, India, in 2018, and the M.Tech. degree in electrical power and energy systems from the National Institute of Technology, Srinagar (JK), India, in 2020. He is currently pursuing the Ph.D. degree with the National Institute of Technology Andhra Pradesh, India. His research interests include high-gain dc-dc converters, multilevel inverters, non-isolated bidirectional converters, power electronics, and ac motor drives and control. SUKANTA HALDER (Member, IEEE) received the B.Tech. degree in electrical engineering from the Jalpaiguri Government Engineering College, and the M.Tech. degree with the specialization in electric drives and power electronics and the Ph.D. degree in electrical engineering from the Indian Institute of Technology (IIT) Roorkee, India, in 2017. He is currently an Assistant Professor with the Department of Electrical Engineering, Sardar Vallabhbhai National Institute of Technology (SVNIT), Surat, India. After Ph.D., in 2017, he joined with the Rolls-Royce@NTU Corporate Laboratory, Nanyang Technological University, Singapore, as a Postdoctoral Fellow. He worked with Rolls-Royce Electrical, Singapore, Nanyang Technological University, which is holding QS world ranking 4th in electrical engineering. He joined the Charge Laboratory, University of Windsor, Windsor, ON, Canada, which is the one of the biggest EV Laboratory in North America under Tier 1 Canada Research Chair Professor in Electrified Vehicles. He was involved for developing WBG-based inverter and intelligent control for EV in collaboration with UofW, Magna International, and Concordia University. He works broadly on next-generation inverter development, intelligent ETDS for EV application. He is a reviewer of many international journals. His research interests include WBG (SiC and GaN) inverter development, gate driver design, electric vehicle, PMSM drives, multi-level inverters, BMS, and machine learningbased motor control. VOLUME 10, 2022