Method and system for predicting optimal charging profiles

Abstract

The present invention describes a system for predicting an optimal charging profile for electric vehicles. The system (200) comprises an interface unit (202) that enables the selection of a departure time for an electric vehicle, and a processing unit (204) that is operatively connected to the interface unit (202) and a battery management unit of an electric vehicle. The processing unit is configured to determine real-time battery characteristic parameters of the electric vehicle from the battery management unit and real-time charging source characteristic parameters in order to estimate the time required for charging.Furthermore, the processing unit is configured to predict an optimal charging profile based on the determined real-time battery characteristic parameters and the selected departure time, using a pre-trained machine learning (ML) model (214).

Description

[0001] The following description particularly describes the invention and the manner in which it is to be carried out.

[0002] The present invention relates generally to the field of electric vehicle charging. More specifically, the present invention relates to a system and method for predicting optimal charging profiles for electric vehicles.

[0003] The following description includes information that may be helpful in understanding the present invention. This is not an admission that the information provided herein is prior art or relevant to the invention claimed herein, or that any publication expressly or impliedly mentioned is prior art.

[0004] In recent years, the electric vehicle industry has experienced significant growth and significant innovation, largely driven by the global need to reduce carbon emissions and transition to sustainable transportation solutions. Central to this transformation is the development of a robust electric vehicle charging infrastructure that meets the diverse needs of electric vehicle owners.

[0005] With conventional technology, the charging process is influenced by a variety of factors, including limitations imposed by direct current breakers (DCBs), limitations within the charging infrastructure, system-level limitations, thermal aspects, concerns regarding lithium plating (the formation of lithium metal on the battery anode), and state of charge (SoC) degradation, where charging can be adjusted to the current state of the battery.

[0006] The rate at which an electric vehicle is charged, often referred to as the charging current, is closely linked to these factors. The charging current depends on DCB limits, infrastructure capacity, system limitations, thermal conditions, SoC, and the need to consider lithium plating risks. Furthermore, a standardized battery charging profile and associated parameters are used to effectively control the charging process. Furthermore, auxiliary systems such as heaters and coolers are essential to maintain the appropriate temperature during charging, thus ensuring safety and efficiency.

[0007] Despite the straightforward charging process, there are cases where longer charging times are required due to specific battery and system limitations. These limitations can arise from factors such as the need for careful temperature management to avoid overheating, concerns about battery aging, or the desire to balance fast charging with preserving long-term battery life. This complex interplay between charging speed and EV battery integrity serves as an important backdrop for ongoing efforts to optimize EV charging processes and improve the sustainability and practicality of electromobility.

[0008] To solve the problem inherent in the conventional charging process, "departure time optimization" has come to the fore. Departure time optimization allows electric vehicle users to specify their desired departure time when charging their vehicles. In response, intelligent algorithms or systems adjust the charging process to ensure the battery is adequately charged at the user's scheduled departure time. This approach provides a tailored charging experience that not only meets scheduling requirements but also optimizes charging profiles to protect battery integrity.By taking into account factors such as charging speed, duration and battery temperature, departure time optimization represents a promising step in the effort to make electric vehicles more practical, efficient and sustainable, thus supporting the transition to a greener and more electrified transport future.

[0009] To solve this problem inherent in conventional technology, patent US9440547B2 discloses a method for charging all types of plug-in electric vehicles. Specifically, the method uses a charging profile and a scheduled departure time to develop a charging scheme for a plug-in electric vehicle that balances battery life with electricity costs.

[0010] While the cited patent documents disclose the generation of the charging profile taking into account the departure time using a complex process, there is room for improvement to provide a system and method for predicting an optimized charging profile for the electric vehicles using a simple and efficient process.

[0011] The present disclosure overcomes one or more disadvantages of the prior art and provides additional advantages discussed in the present disclosure. Additional features and advantages are realized through the techniques of the present disclosure. Other embodiments and aspects of the disclosure are described in detail herein and are considered part of the claimed disclosure.

[0012] In one aspect, the present disclosure describes a system for predicting an optimal charging profile for electric vehicles. The system includes an interface unit enabling the selection of a departure time for an electric vehicle, and a processing unit operably connected to the interface unit and a battery management unit of an electric vehicle. The processing unit is configured to determine real-time battery characteristic parameters of the electric vehicle from the battery management unit and real-time charging source characteristic parameters to estimate the time required for charging. Subsequently, the processing unit is configured to compare the selected departure time with the estimated time.Furthermore, the processing unit is configured to, if the selected departure time is greater than the estimated charging time, predict an optimal charging profile based on the determined real-time battery characteristic parameters and the selected departure time using a pre-trained machine learning (ML) model, wherein the pre-trained model corresponds to a model trained based on training data of the charging profile to generate an optimal charging profile.

[0013] In another aspect, the present disclosure describes that the real-time battery characteristic parameters correspond to the temperature and state of charge (SOC) of a battery of the electric vehicle.

[0014] In another aspect, the present disclosure describes a processing unit for training the ML model. The control unit is further configured to receive the charging profile training data over a specific period of time from at least one of a server and the battery management unit. The server is communicatively connected to the processing unit. The charging profile training data includes a relationship between battery characteristic parameters, departure time, and charging current. Subsequently, the processing unit is configured to determine an output based on an activation function, battery characteristic parameters, and departure time associated with the charging profile training data.Furthermore, the processing unit is configured to train the ML model to predict the optimal charging profile based on the determined predicted output and charging current associated with training data of the charging profile.

[0015] In another aspect, the present disclosure describes the processing unit for predicting the optimal charging profile. The processing unit is further configured to compare the determined real-time battery characteristic parameters with the trained data. Subsequently, the processing unit is further configured to identify the optimal profile based on the comparison.

[0016] In another aspect, the present disclosure describes the processing unit for determining additional information related to charging the electric vehicle. The processing unit is further configured to perform a mechanistic analysis based on the identified optimal profile. Subsequently, the processing unit is further configured to determine the charging efficiency and the optimal battery temperature based on the performed physics-based analysis.

[0017] In one aspect, the present disclosure may describe a method for predicting the optimal charging profile for electric vehicles. The method includes determining real-time battery characteristic parameters of an electric vehicle and real-time charging source characteristic parameters to estimate the time required for charging. The method further includes comparing the departure time with the estimated time. Finally, if the departure time is greater than the estimated charging time, the method includes predicting an optimal charging profile based on the determined real-time battery characteristic parameters and the departure time using a pre-trained machine learning (ML) model. The pre-trained model corresponds to a model trained based on training data of the charging profile to generate an optimal charging profile.

[0018] In another aspect, the present disclosure describes that the real-time battery characteristic parameters correspond to the temperature and state of charge (SOC) of a battery of the electric vehicle.

[0019] In another aspect, the present disclosure describes a method for training the ML model, including obtaining the charging profile training data from at least one of a server and the battery management unit over a period of time, wherein the charging profile training data includes a relationship between battery characteristic parameters, departure time, and charging current. The method further includes determining an output based on an activation function, battery characteristic parameters, and departure time associated with charging profile training data. Furthermore, the method further includes training the ML model to predict the optimal charging profile based on the determined predicted output and charging current associated with charging profile training data.

[0020] In another aspect, the present disclosure describes a method for predicting the optimal charging profile, including comparing the determined real-time battery characteristic parameters with the trained data. The method further includes identifying the optimal profile based on the comparison.

[0021] In another aspect, the present disclosure describes a method for determining additional information related to charging the electric vehicle, including performing a mechanistic analysis based on the identified optimal profile. The method includes determining the charging efficiency and optimal battery temperature based on the performed physics-based analysis.

[0022] The foregoing summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, other aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description.

[0023] The embodiments of the disclosure itself, as well as a preferred mode of use, further objects, and advantages thereof, are best understood by reference to the following detailed description of an illustrative embodiment when read in conjunction with the accompanying drawings. One or more embodiments will now be described, by way of example only, with reference to the accompanying drawings, in which: Fig.1 illustrates an environment architecture for a system for predicting optimal charging profiles for electric vehicles according to an embodiment of the present disclosure. Fig. 2 illustrates a system for predicting optimal charging profiles for electric vehicles according to an embodiment of the present disclosure. Fig. 3 illustrates training the ML model to provide a trained dataset to generate an optimal charging profile, according to an embodiment of the present disclosure. Fig. 4 is a flowchart illustrating a method for predicting an optimal charging profile for electric vehicles according to an embodiment of the present disclosure.

[0024] The figures depict embodiments of the disclosure for illustrative purposes only. Those skilled in the art will readily appreciate from the following description that alternative embodiments of the structures and methods illustrated herein may be employed without departing from the principles of the disclosure described herein.

[0025] The foregoing has broadly outlined the features and technical advantages of the present disclosure in order to better understand the following detailed description of the disclosure. It should be apparent to those skilled in the art that the disclosed conception and specific embodiment may be readily utilized as a basis for modifying or designing other structures to carry out the same purposes of the present disclosure.

[0026] Various embodiments of the present invention will now be described more fully with reference to the accompanying drawings, in which some, but not all, embodiments of the invention are shown. Thus, the invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. The term "or" is used herein in both the alternative and conjunctive senses unless otherwise indicated. The terms "illustrative," "example," and "exemplary" are used as examples without indication of degree of quality. Like numbers refer to like elements throughout.

[0027] The terms "in one embodiment," "according to an embodiment," or the like generally mean that the particular feature, structure, or characteristic following the term may be included in at least one embodiment of the present disclosure and may be included in more than one embodiment of the present disclosure (importantly, such terms do not necessarily refer to the same embodiment).

[0028] The word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any implementation described herein as "exemplary" should not necessarily be construed as preferred or advantageous over other implementations.

[0029] If the description states that a component or feature "preferably," "may," "usually," "optionally," "for example," "often" (or similar language) may, "could," "should," or "would" be included or have a characteristic, it is not required that that particular component or feature be included or have that characteristic. That component or feature may be optionally included in some embodiments, or it may be excluded.

[0030] The terms "vehicle" and "electric vehicle" are used interchangeably throughout this disclosure. The electric vehicle refers to battery electric vehicles (BEVs): These vehicles are powered entirely by batteries that store electricity used to power the electric motor that drives the wheels. The electric vehicle may not have a gasoline engine and may not use fossil fuels. The electric vehicle may be charged by plugging the electric vehicle into a wall outlet or charging station. The electric vehicle may be a car, a truck, a semi-trailer truck, a motorcycle, an airplane, a train, a moped, a scooter, or another type of transportation. Further, the electric vehicle may use many types of powertrains. For example, the electric vehicle may be a plug-in electric vehicle, a plug-in hybrid electric vehicle, a hybrid electric vehicle, or a fuel cell vehicle.

[0031] Throughout this disclosure, the term "machine learning" is used. Machine learning, in its broadest sense, describes a function of systems that learn from data. A machine learning system, engine, or module may include a machine learning algorithm that can be trained to learn currently unknown functional relationships between inputs and outputs. In one or more embodiments, the machine learning functionality may be implemented using an artificial neural network (ANN) that can be trained to perform a currently unknown function. In machine learning and cognitive science, ANNs are a family of statistical learning models inspired by the biological neural networks of animals, and in particular, the brain.ANNs can be used to estimate or approximate systems and functions that depend on a large number of inputs.

[0032] The "battery" may be a rechargeable battery / battery module that includes (or is formed from) a plurality of unit battery cells coupled in series and / or parallel to provide a high energy density. The battery module may be formed by connecting electrode terminals of the unit battery cells together, with the number of unit battery cells determined based on a desired (or required) amount of power, to provide a high-power rechargeable battery. The unit cells may be connected in series, parallel, or a mixture of both to provide the desired voltage, capacity, and / or power density.

[0033] Disclosed herein is a system for predicting optimal charging profiles for electric vehicles. The system may include an interface unit that enables the selection of a departure time for an electric vehicle. The system may further include a battery management system and a processing unit operatively connected to the interface unit and the battery management unit. The processing may be configured to determine real-time battery characteristic parameters of the electric vehicle from the battery management unit and real-time charging source characteristic parameters to estimate the time required for charging. Subsequently, the processing unit may be configured to compare the selected departure time with the estimated time.Finally, if the selected departure time is greater than the estimated charging time, the processing unit may be configured to predict an optimal charging profile based on the determined real-time battery characteristic parameters and the selected departure time using a pre-trained machine learning (ML) model, wherein the pre-trained model corresponds to a model trained based on training data of the charging profile to generate an optimal charging profile. An embodiment of the present disclosure may provide a method and system for predicting optimal charging profiles that extend the lifespan and long-term performance of the battery by avoiding overstressing the battery. Thus, the system may provide a simple and efficient process that eliminates the need for a complex process for predicting the optimal charging profile.In addition, the system can improve battery life, increasing the capacity and performance of the electric vehicle.

[0034] Referring to the drawings, the following detailed description, taken in conjunction with the accompanying drawings, is intended as a description of various configurations and is not intended to represent the only configurations in which the concepts described herein may be practiced. The detailed description includes specific details for the purpose of providing a thorough understanding of various concepts, with like reference numerals indicating like components throughout the several views. However, it will be apparent to one skilled in the art that these concepts may be practiced without these specific details.

[0035] Fig.Figure 1 illustrates an environmental architecture 100 for a system for predicting optimal charging profiles for electric vehicles according to an embodiment of the present disclosure. The environmental architecture 100 may consist of an external power source 102, a charging station 104, an electric vehicle 106, a server 108, and a system 110. All Fig. 1 are essential associated elements, but the environment architecture 100 may be implemented by more than those illustrated in Fig.1, which are not explained for brevity. System 110 can communicate with all associated elements of environmental architecture 100. The components present in system 110 can communicate with charging station 104, electric vehicle 106, and server 108 via a wireless communication network, a wired communication network, or a combination of both.

[0036] In one embodiment, external power source 102 may be a utility grid, a generator, a renewable energy source (e.g., solar, wind, hydropower, geothermal, etc.), or another power source. External power source 102 may be in electrical communication with charging station 104. Charging station 104 may receive power from external power source 102. Further, charging station 104 may provide power to electric vehicles by establishing a physical and electronic connection between charging station 104 and electric vehicle 106 via a charging cable and connector.

[0037] In one embodiment, the electric vehicle 106 may include a battery management unit 112. The battery management unit 112 may be any electronic system that manages a rechargeable battery, such as by protecting the battery from operation outside its safe operating range, monitoring its condition, calculating secondary data, reporting that data, and controlling its environment.

[0038] In one embodiment, server 108 may be configured to communicate with system 110 to provide various information to system 110. Specifically, the server may include an application programming interface (API) to provide various information to system 110. The API may be used to enable software applications to interact with other software applications or services. APIs use standardized protocols and data formats to enable communication between applications, such as HTTP or REST. Through the functionality of the API, the API may provide various information to system 110.

[0039] In one embodiment, the system 110 may include a processing unit 114 and an interface unit 116. The associated elements of the system 110 may communicate with all other associated elements of the environmental architecture 100 to perform a prediction of the optimized charging profile for electric vehicles. In one embodiment, the system 110 may be present in the electric vehicle 106. In an alternative embodiment, the system 110 may be present in the charging station 104. The detailed operation of the system 110 in conjunction with other Fig. 1 disclosed elements can be Fig. 2 will be further explained in subsequent sections of the present disclosure.

[0040] Referring now to Fig. 2, which uses a system for predicting optimal charging profiles (such as the system 110 of Fig.2) for electric vehicles according to an embodiment of the present disclosure. It will be apparent to those skilled in the art that the architecture may also be used in various environments other than Fig. 2 can be implemented.

[0041] According to one embodiment of the present disclosure, the system 200 may include an interface unit 202 (same as the interface unit 116 of Fig. 1) and a processing unit 204 (equal to the processing unit 114 of Fig. 1) exist. All in Fig. 2 are essential associated elements, and the system 200 may be configured by more than those illustrated in Fig.2 may be implemented, but these are not explained for the sake of brevity. All associated elements of system 200 may communicate with each other via a wireless connection, an electrical connection, or a combination of both.

[0042] In one embodiment, the interface unit 202 may be a user interface that allows a person to communicate with either a charging station (such as the charging station 104 of Fig. 1) or an electric vehicle (such as the vehicle 106 of Fig.1). In one embodiment, the interface unit 202 may include a display unit, a speaker, a microphone, and so on. In one embodiment, the display unit may be used to display information regarding optimal charging profiles. In one embodiment, the speaker may be used to provide information regarding optimal charging profiles. In one embodiment, the microphones could be used by drivers to provide voice commands regarding optimal charging profiles.

[0043] In another embodiment, the interface unit 202 may receive user input information based on the interaction between a user and the interface unit 202. In particular, the user may provide the departure time to the system 200 based on the operation of the interface unit 202. The departure time may refer to the time at which the electric vehicle 106 is scheduled to depart the charging station 104. To receive the departure time, the interface unit 202 may include a keypad, a keyboard, a touchscreen, a touchpad, a mouse, a microphone, and / or a speaker.

[0044] In one embodiment, the processing unit 204 may consist of a receiving unit 206, a determining unit 208, a comparing unit 210, and a predicting unit 212. All of the Fig.2 are essential associated elements, but the processing unit 204 may be implemented by more than those illustrated in Fig. 2 may be implemented. However, for the sake of brevity, these will not be explained. All associated elements of the processing unit 204 can communicate with each other via a wireless / wired communication network. The functionality of the individual associated elements of the processing unit 204 is explained in subsequent paragraphs.

[0045] In a non-limiting embodiment of the present disclosure, the receiving unit 206 may be configured to receive real-time battery characteristic parameters of the electric vehicle 106 from a battery management unit (such as the battery management unit 112 of Fig.1) and receive real-time charging source characteristic parameters. Specifically, the receiving unit 206 may receive the real-time battery characteristic from the battery management unit 112.

[0046] In one embodiment, the real-time battery characteristic parameters may include a maximum power, which indicates the highest rated power at which a battery of the electric vehicle 106 can accept electrical energy without causing damage or significantly impacting performance and lifetime. In an additional embodiment, the real-time battery characteristic parameters may include the current state of charge (SOC) of the battery, which may represent the current state of charge of the battery relative to the maximum capacity of the battery. In an additional embodiment, the real-time battery characteristic parameters may include the battery temperature (T_init), which may indicate the initial or starting temperature of the battery when the battery is commissioned or begins a charge or discharge cycle.In an additional embodiment, the battery thermal limit may indicate the maximum safe operating temperature of the battery. In an additional embodiment, the real-time battery characteristic parameters may include information regarding the battery's lithium plating. In an additional embodiment, the depth of charge / discharge (DCB) limit may correspond to the maximum depth to which the battery can be charged or discharged without causing damage or reducing the battery's service life.

[0047] Furthermore, the receiving unit 206 may receive parameters related to an environmental condition. The parameters related to the environmental condition may include the ambient temperature (Tamb), which may represent the temperature of the environment.

[0048] Furthermore, the receiving unit 206 may receive real-time charging source characteristic parameters. The charging source characteristic parameters may include the infrastructure boundary of the charging station 104, which may indicate the maximum capacity of the charging station 104 in terms of power output, the number of charging points, and other related features. In an additional embodiment, the charging source characteristic parameters may include a system limit, which may indicate the maximum operating capacity of the battery.

[0049] In one embodiment, the receiving unit 206 may receive all of the information mentioned in the preceding paragraphs sequentially or simultaneously.

[0050] In one embodiment, the receiving unit 206 may be, for example, a receiver, which may include an antenna, an antenna array, an input interface, a pin, a circuit, or the like.

[0051] In one non-limiting embodiment of the present disclosure, the determination unit 208 may be configured to determine the real-time battery characteristic parameters of the electric vehicle 106 from the battery management unit 112 and the real-time charging source characteristic parameters to estimate the time required for charging.

[0052] In particular, the determination unit 208 may be configured to receive the real-time battery characteristic parameters of the electric vehicle 106 from the battery management unit 112 and the real-time charging source characteristic parameters from the receiving unit 206. After receiving the information, the determination unit 208 may generate a conventional charging profile that may illustrate how the battery SoC changes over time during the charging process. The determination unit 208 may generate the charging profile taking into account the optimal charging time (minimum charging time) to charge the battery from the initial SoC to the desired final SoC.

[0053] In one non-limiting embodiment of the present disclosure, the comparison unit 210 may be configured to compare the selected departure time with the estimated time. Specifically, the comparison unit 210 may receive the departure time from the interface unit 202 and may receive the estimated time from the determination unit 208. After receiving the departure time and the estimated time, the comparison unit 210 may compare the departure time with the estimated time.

[0054] If the departure time is less than or equal to the estimated time, in one embodiment, the comparison unit 210 may provide an indication to the prediction unit 212, which may instruct the prediction unit 212 to extract the conventional charging profile from the determination unit 208 and provide the conventional charging profile to the charging station 104 to charge the electric vehicle 106.

[0055] In an alternative embodiment, if the departure time is greater than the estimated time, the comparison unit 210 may provide an indication to the prediction unit 212, which may instruct the prediction unit 212 to predict an optimal charging profile.

[0056] In one non-limiting embodiment of the present disclosure, after receiving an instruction to predict the optimal charging profile, the prediction unit 212 may predict the optimal charging profile based on the determined real-time battery characteristic parameters and the selected departure time. To predict the optimal charging profile, the prediction unit 212 may use a pre-trained machine learning (ML) model 214.

[0057] The pre-trained ML model 214 may correspond to a model trained based on charging profile training data to generate an optimal charging profile. After training using charging profile data, the ML model 214 may analyze real-time battery characteristics, including state of charge, voltage, temperature, and the selected departure time. By comparing this data with the patterns captured during training, the ML model 214 may generate a charging current profile. This profile may outline recommended charging rates over time, thereby optimizing the charging process to meet specific requirements, whether to maintain battery integrity, reduce costs, or maximize energy efficiency. The training of the pre-trained ML model 214 is described in subsequent paragraphs of Fig. 3 is further explained.

[0058] Referring now to Fig.3, the training of the ML model 300 (equal to the pre-trained ML model 214 of Fig. 2) to provide a trained data set to generate an optimal charging profile, according to an embodiment of the present disclosure. The ML model 300 may consist of a maintenance unit 302 and a training unit 304. All in Fig. 3 are essential related elements, and the ML model 300 may be supplemented by more related elements than those illustrated in Fig. 3, which, however, are not explained for the sake of brevity. All associated elements of the ML model 300 can communicate with each other via a wireless connection, an electrical connection, or a combination of both.

[0059] In one non-limiting embodiment of the present disclosure, the obtaining unit 302 may be configured to obtain charging profile training data including a relationship between battery characteristic parameters, departure time, and charging current, as shown below. CURRENT_SOC T_INIT DEPARTURE TIME CHARGING CURRENT 0,75 25 30 30 0,60 27 45 25 0,85 28 35 35

[0060] In one embodiment, data corresponding to the current SOC, T_init, and the departure time may be input data. The data corresponding to the charging current may be output data. In one embodiment, the obtaining unit 302 may obtain the charging profile training data from the server (such as the server 108 of Fig. 1), from the battery management unit (such as the battery management unit 112 of Fig.1) or from both the server 108 and the battery management unit 112. The maintenance unit 302 may include a process for continuously collecting the charging profile training data over a specified period to capture the dynamic nature of charging profiles and enable a more comprehensive data set.

[0061] After obtaining the training data of the loading profile, the obtaining unit 302 may provide the training data to the training unit 304. In one non-limiting embodiment of the present disclosure, the training unit 304 may be configured to scale the input data and the output data of the training data between 0 and 1. Further, the training unit 304 may initially assign weights and biases as random values.

[0062] After initialization, the training unit 304 may determine the weighted sum of inputs by taking the dot product of the input values ​​and the weights. After the weighted sum, the training unit 304 may add the biases in the weighted sum of inputs to generate a preliminary output value. The training unit 304 may perform this process using Equation 1 below, as shown below: A=xi*wi+bi

[0063] After determining the preliminary output value, the training unit 304 may be configured to generate a predicted output value using an activation function for the preliminary output value. The training unit 304 may perform this process using Equation 2 below, as shown below: Y_pred=tanh(A)

[0064] The training unit 304 can calculate the error between the predicted output (Y_pred) and the actual output and adjust the weights and biases using optimization techniques to minimize the mean square error (MSE) between the predicted output and the actual output. The MSE (i.e., an R2 score) can measure how well the model predicts the charging current compared to the actual values. By performing the steps described in the description of Fig. 3 mentioned steps, the training of the ML model can be carried out for the different training data.

[0065] In one non-limiting example of the present disclosure, the training data 304 may receive the training data as shown below: CURRENT_SOC T_INIT DEPARTURE TIME CHARGING CURRENT 0 -15 260,9 50 0 -7 260,9 50 0 0 260,9 50 0 5 260,9 50 0 10 260,9 50

[0066] After receiving the data, the training unit 304 can predict the output using Equations 1 and 2 discussed in the previous paragraphs. The relationship between the predicted output and the actual output can be used to generate the optimized weights and bias values ​​as shown below: Weightings

[0067] W(i): [-0.43930491 0.55847266 -0.35635229 0.58199735 -0.0873219 0.05342527 0.29500304 -0.00133969] Biases:

[0068] [0.09124053 0.42535492 -0.16889854 0.67697059 0.31785347 0.68196372 -0.76449376 0.52555509 0.38929549)

[0069] The updated weights and biases can be calculated during the training process to minimize the MSE between the predicted output and the actual output. After training, the updated weights and biases reflect the relationship between the input and output to predict the optimal loading profile.

[0070] With further reference to Fig.2 illustrates the prediction unit 212 predicting the optimal charging profile after training the ML model 214. The prediction unit 212 may receive the determined real-time battery characteristic parameters and the selected departure time. The ML model of the prediction unit 212 may analyze the received real-time battery characteristic parameters and the selected departure time by comparing the data received in real time with the trained data. After the comparison, the ML model 214 may provide the charging current (i.e., the charging current profile), which may be an optimal charging profile for charging the vehicle.

[0071] In one non-limiting embodiment of the present disclosure, after determining the charging current, the ML model 214 may perform a mechanistic analysis based on the identified optimal profile and determine the charging efficiency and the optimal battery temperature. In particular, the ML model 214 may perform a mechanistic analysis, such as a physics-based analysis. During the physics-based analysis, the charging efficiency may be calculated by comparing the energy input to the battery (calculated from the predicted charging current profiles) with the change in the battery's energy storage capacity. This comparison may evaluate how efficiently the battery can be charged. Further, the physics-based analysis may include tracking the battery temperature throughout the charging process. In this way, the optimal temperature range for charging may be determined.

[0072] In one embodiment, the prediction unit 212 may provide the predicted charge along with the predicted maximum temperature and charging efficiency to the interface unit 202 before starting the charging process.

[0073] After predicting the charging profile, the charging process between charging station 104 and electric vehicle 106 is initiated. Once the charging process is complete, prediction unit 212 can provide the charging profile, along with the maximum temperature and charging efficiency, to interface unit 202. Interface unit 202 can display the information received from prediction unit 212.

[0074] Referring to Fig.4, the steps of a method for predicting the optimal charging profile 400 for electric vehicles are shown. The method begins at step 402. At step 402, the method 400 may include determining real-time battery characteristic parameters of an electric vehicle (such as the electric vehicle 106 of Fig. 1) and real-time charging source characteristic parameters to estimate the time required for charging, wherein the real-time battery characteristic parameters correspond to the temperature and state of charge (SOC) of a battery of the electric vehicle 106. In an exemplary aspect, a determination unit 208 of Fig. 2 of the system 200 may be configured to perform the process steps disclosed in step 402.

[0075] At step 404, the method 400 may include comparing the departure time with the estimated time. In an exemplary aspect, a comparison unit 210 of Fig.2 of the system 200 may be configured to perform the process steps disclosed in step 404.

[0076] At step 406, if the departure time is greater than the estimated charging time, the method 600 may include predicting an optimal charging profile based on the determined real-time battery characteristic parameters and the departure time using a pre-trained machine learning (ML) model. Predicting the optimal charging profile may include comparing the determined real-time battery characteristic parameters to the trained data and identifying the optimal profile based on the comparison. After predicting the charging profile, a mechanistic analysis is further performed based on the identified optimal profile, and the charging efficiency and optimal battery temperature are determined based on the performed physics-based analysis. Training of the ML may be performed prior to predicting the charging profile.In particular, the pre-trained model corresponds to a model that is trained based on charging profile training data to generate an optimal charging profile. The training steps of the ML model may include obtaining the charging profile training data from at least one of a server or the battery management unit 112 over a period of time. The charging profile training data includes a relationship between battery characteristic parameters, departure time, and charging current. Subsequently, the training step of the ML model may include determining an output based on an activation function, battery characteristic parameters, and departure time associated with charging profile training data, and training the ML model to predict the optimal charging profile based on the determined predicted output and charging current associated with charging profile training data.In one example aspect, a prediction unit 212 of . Fig. 2 of the system 200 may be configured to perform the process steps disclosed in step 406.

[0077] The technical effect associated with the present disclosure is shown below: An embodiment of the present disclosure may provide a smart home inverter with a DC-AC conversion process that eliminates the need for additional converter hardware on the vehicle. This elimination improves the overall performance of the electric vehicle and minimizes the impact on the vehicle's structural integrity.

[0078] An embodiment of the present disclosure may provide a method and system for predicting optimal charging profiles that extend battery life and long-term performance by avoiding battery overstress.

[0079] An embodiment of the present disclosure may provide a method and system for predicting optimal charging profiles to optimize charging efficiency by adjusting charging speed and other parameters. This optimization reduces energy losses during the charging process.

[0080] An embodiment of the present disclosure may provide a method and system for predicting optimal charging profiles that improve grid integration by reducing grid peak loads.

[0081] The foregoing method descriptions and process flow diagrams have been provided merely as illustrative examples and are not intended to prescribe or imply that the steps of the various embodiments must be performed in the order shown. As will be appreciated by one skilled in the art, the order of steps in the foregoing embodiments may be performed in any order. Words such as "thereafter," "then," "next," etc., are not intended to limit the order of steps; these words serve merely to guide the reader through the description of the methods. Furthermore, any reference to claim elements in the singular, for example, using the articles "a," "an," or "the," should not be construed as limiting the element to the singular.

[0082] As used herein, the term "unit" may be implemented in hardware and / or software. When implemented in hardware, the unit may be configured as a device, e.g., a computer or a processor, or as part of a system, e.g., a computer system. When implemented in software, the unit may be configured as a computer program product, a function, a routine, or program code.

[0083] The hardware used to implement the various illustrative logic, logic blocks, modules, and circuits described in connection with the aspects disclosed herein may include a general-purpose processor, a digital signal processor (DSP), a special-purpose processor such as an application-specific integrated circuit (ASIC) or a field-programmable gate array (FPGA), a programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but alternatively, the processor may be a conventional processor, a conventional controller, a conventional microcontroller, or a conventional state machine. A processor may also be embodied as a combination of computing devices, e.g.,a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or another such configuration. Alternatively or additionally, some steps or methods may be performed by logic specific to a given function.

[0084] Many modifications and other embodiments of the inventions set forth herein will occur to one skilled in the art to which these inventions belong when considering the teachings presented in the foregoing descriptions and the accompanying drawings. Although the figures illustrate only certain components of the devices and systems described herein, it should be understood that various other components may be used in conjunction with the utility management system. It is therefore to be understood that the inventions are not limited to the specific embodiments disclosed, and that modifications and other embodiments are intended to be included within the scope of the appended claims.Furthermore, the steps in the method described above need not necessarily occur in the order depicted in the accompanying diagrams, and in some cases, one or more of the depicted steps may occur substantially concurrently, or additional steps may be involved. Although specific terms are used herein, they are used only in a generic and descriptive sense and not for purposes of limitation.

[0085] The various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the particular application and the design constraints imposed on the overall system. One skilled in the art may implement the described functionality in different ways for each particular application, but such implementation choices should not be construed as a departure from the scope of the present invention. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] US 9440547B2

[0009]

Claims

[1] System for predicting an optimal charging profile (200) for electric vehicles, the system (200) comprising: an interface unit (202) enabling the selection of a departure time for an electric vehicle; a processing unit (204) operatively connected to the interface unit (202) and a battery management unit of an electric vehicle, the processing unit (204) being configured to: Determining real-time battery characteristic parameters of the electric vehicle from the battery management unit and real-time charging source characteristic parameters to estimate the time required for charging; Comparing the selected departure time with the estimated time; and if the selected departure time is greater than the estimated time required for loading, the processing unit is further configured to: Predicting, using a pre-trained machine learning (ML) model (214), an optimal charging profile based on the determined real-time battery characteristic parameters and the selected departure time, wherein the pre-trained model corresponds to a model trained based on training data of the charging profile to generate an optimal charging profile. [2] The system (200) of claim 1, wherein the real-time battery characteristic parameters correspond to the temperature and state of charge (SOC) of a battery of the electric vehicle. [3] The system (200) of claim 2, wherein the processing unit (204) for training the ML model (214) is further configured to: Obtaining the charging profile training data from at least one of a server or the battery management unit over a period of time, the server being communicatively connected to the processing unit, the charging profile training data including a relationship between battery characteristic parameters, departure time, and charging current; Determining an output based on an activation function, battery characteristic parameters, and departure time associated with training data of the charging profile; and Training the ML model (214) to predict the optimal charging profile based on the determined predicted output and charging current associated with training data of the charging profile. [4] The system (200) of claim 3, wherein the processing unit (204) for predicting the optimal charging profile is further configured to: Comparing the determined real-time battery characteristic parameters with the trained data and Identify the optimal profile based on the comparison. [5] The system (200) of claim 4, wherein the processing unit (204) is further configured to: Conducting a mechanistic analysis based on the identified optimal profile; Determine the charging efficiency and optimal battery temperature based on the physics-based analysis performed. [6] A method for predicting the optimal charging profile (400) for electric vehicles, the method (400) comprising: Determining (402) real-time battery characteristic parameters of an electric vehicle and real-time charging source characteristic parameters to estimate the time required for charging; Comparing (404) the departure time with the estimated time; and if the departure time is greater than the estimated charging time, predicting (406), using a pre-trained machine learning (ML) model, an optimal charging profile based on the determined real-time battery characteristic parameters and the departure time, wherein the pre-trained model corresponds to a model trained based on training data of the charging profile to generate an optimal charging profile. [7] The method (400) of claim 6, wherein the real-time battery characteristic parameters correspond to the temperature and state of charge (SOC) of a battery of the electric vehicle. [8] The method (400) of claim 7 for training the ML model, further comprising: Obtaining the charging profile training data from at least one of a server or the battery management unit over a period of time, wherein the charging profile training data includes a relationship between battery characteristic parameters, departure time, and charging current; Determining (406) an output based on an activation function, battery characteristic parameters, and departure time associated with training data of the charging profile; and Training (406) the ML model to predict the optimal charging profile based on the determined predicted output and charging current associated with training data of the charging profile. [9] The method (400) of claim 8 for predicting the optimal charging profile, further comprising: Comparing (406) the determined real-time battery characteristic parameters with the trained data and identifying (406) the optimal profile based on the comparison. [10] The method (400) of claim 9, further comprising: Conducting (406) a mechanistic analysis based on the identified optimal profile and Determining (406) the charging efficiency and the optimal battery temperature based on the physics-based analysis performed.

Images