A method and system for controlling temperature variations across a battery pack of a vehicle

Abstract

A method of controlling temperature variations across a battery pack of a vehicle comprises receiving cooling inlet parameters associated with a cell cooling plate 104, at least one of coolant inlet f

Description

PREAMBLE TO THE DESCRIPTION The following specification particularly describes the invention and the manner in which it is to be performed: TECHNICAL FIELD

[0001] The present disclosure relates to the field of automobiles, and more particularly relates to controlling temperature variations across a battery pack of a vehicle. BACKGROUND OF THE DISCLOSURE

[0002] Recently, the demand for Electric Vehicles (EVs) in the automotive sector has significantly increased. Batteries are identified as the most crucial and costly components in any EV. As a result, it is imperative for the battery to demonstrate extended longevity and maintain optimal efficiency throughout its operational lifespan. Accordingly, a robust mechanism must be implemented to attenuate battery aging and degradation. Battery aging can be attributed to various factors, including, but not limited to, inadequate thermal management, which accelerates the aging process. Elevated operating temperatures result in a gradual decline in battery efficiency and a corresponding reduction in overall battery life span.

[0003] To enhance battery performance and longevity, precise measurement of individual cell temperatures is essential during both the design and operational phases. However, due to constraints related to cost and space requirements, the deployment of temperature sensors for monitoring the temperature of each cell is often limited. Typically, a battery pack consists of multiple cells organized into modules, with each module equipped with a minimal number of temperature sensors for monitoring cell temperatures. This restricted utilization of temperature sensors proves inadequate for detecting variations in individual cell temperatures. Consequently, this limits the scope of varying coolant flow rates based on the maximum temperature deviation across the plurality of cells in the battery pack. Current mechanisms depend on the limited sensor data obtained from these temperature sensors to implement control strategies. However, no existing mechanism effectively monitors the temperature at interface point also known as the cell contact surface between the cell cooling plate and the plurality of cells.

[0004] Accordingly, there is a need to provide a robust and reliable thermal management system for vehicle batteries. SUMMARY OF THE DISCLOSURE

[0005] In one non-limiting embodiment of the present disclosure, a method of controlling temperature variations across a battery pack of a vehicle is disclosed. The method comprises receiving cooling inlet parameters associated with a cell cooling plate and cell parameters of a plurality of cells of the battery. The cooling inlet parameters comprise at least one of a coolant inlet flow rate and a coolant inlet temperature. The cell parameters comprise at least one of: a heat dissipation rate and location information of each cell. Further, the method comprises determining the temperature at the cell cooling plate interface by applying a Physical Informed Neural Network (PINN) model. Based on the determined temperature, the method calculates a maximum temperature deviation across each cell. Thereafter, the method comprises determining at least one of an optimal coolant inlet flow rate and an optimal coolant inlet temperature for the cell cooling plate interface when the maximum temperature deviation exceeds a temperature threshold value. Finally, the method comprises controlling the temperature variations across the battery pack based on the determined optimal coolant inlet flow rate and the optimal coolant inlet temperature.

[0006] In another non-limiting embodiment of the present disclosure, a system for controlling temperature variations across a battery pack of a vehicle is disclosed. The system comprises one or more sensors configured to sense cooling inlet parameters associated with the cell cooling plate and cell parameters of a plurality of cells of the battery. The cooling inlet parameters comprise at least one of a coolant inlet flow rate and a coolant inlet temperature. The cell parameters comprise at least one of: a heat dissipation rate and location information of each cell. The system comprises at least one processor communicatively coupled to the one or more sensors. The at least one processor is configured to receive the cooling inlet parameters associated with the cell cooling plate and the cell parameters of a plurality of cells of the battery. Thereafter, the processor is configured to determine the temperature at the cell cooling plate interface by applying a Physical Informed Neural Network (PINN) model. Further, the at least one processor is configured to calculate a maximum temperature deviation across each cell based on the determined temperature and determine at least one of an optimal coolant inlet flow rate and an optimal coolant inlet temperature for the cell cooling plate interface when the maximum temperature deviation exceeds a temperature threshold value. Finally, the at least one processor is configured to control the temperature variations across the battery pack based on the determined optimal coolant inlet flow rate and the optimal coolant inlet temperature.

[0007] The foregoing summary is illustrative only and is not intended to be in any way limiting. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features will become apparent by reference to the drawings and the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0008] The embodiments of the disclosure itself, as well as a preferred mode of use, further objectives and advantages thereof, will best be understood by reference to the following detailed description of an illustrative embodiment when read in conjunction with the accompanying drawings. One or more embodiments are now described, by way of example only, with reference to the accompanying drawings in which:

[0009] Fig. 1 illustrates an exemplary environment of a battery pack of a vehicle interfacing with a cooling plate, in accordance with an embodiment of the present disclosure;

[0010] Fig. 2A shows an exemplary architecture of the system for controlling temperature variations across a battery pack of a vehicle, in accordance with an embodiment of the present disclosure;

[0011] Fig. 2B shows a detailed block diagram of the system for controlling temperature variations across a battery pack of a vehicle, in accordance with an embodiment of the present disclosure;

[0012] Fig. 3 exemplary schematic diagram of a PINN model for determining temperature at the cell cooling plate interface, in accordance with an embodiment of the present disclosure;

[0013] Fig. 4 illustrates another exemplary environment of cell cooling plate indicating cell contact surface area, in accordance with an embodiment of the present disclosure;

[0014] Fig. 5 shows a flowchart depicting a method of controlling temperature variations across a battery pack of a vehicle, in accordance with an embodiment of the present disclosure; and

[0015] Fig. 6 represents a flowchart depicting a method of determining temperature at the cell cooling plate interface by implementing an optimization model, in accordance with an embodiment of the present disclosure; DETAILED DESCRIPTION

[0016] The novel features which are believed to be characteristic of the disclosure, both as to its organization and method of operation, together with further objects and advantages will be better understood from the following description when considered in connection with the accompanying Figures. It is to be expressly understood, however, that each of the Figures is provided for the purpose of illustration and description only and is not intended as a definition of the limits of the present disclosure.

[0017] As described earlier, to optimize battery performance and longevity, accurately determining individual cell temperature is crucial during design and operation of the battery. Battery cell temperatures are frequently affected by their interaction with the cooling system, which exhibits thermal inhomogeneities. To address these issues, the present disclosure provides a method and system for controlling temperature variations across a battery pack of a vehicle. In particular, the disclosed method and system utilize a Physics Informed Neural network (PINN) model to determine the temperature at the cell cooling plate interface (the process of determining temperature is disclosed in upcoming paragraphs). Based on the determined temperature, one or more cooling parameters such as coolant inlet flow rate and coolant inlet temperature are regulated. For instance, in cases of overheating the determined temperature may increase and the coolant inlet flow may also be increased significantly such that the temperature across the battery pack decreases.

[0018] Thus, the present disclosure provides a robust and reliable thermal management system for controlling temperature variations at each point of the cell cooling plate interface. Further, by accurately determining individual cell temperatures with the help of the PINN model, the present disclosure ensures optimized temperature management, thereby enhancing battery performance and lifespan. Moreover, the PINN model utilized in determining the temperature provides precise temperature predictions at any point on the cell cooling plate interface and may be extended seamlessly to entire battery modules. Additionally, the utilization of PINN model is effective in modeling multi-physics system, capturing both the interaction between solid and liquid / fluid components and the liquid flow inside the battery cooling system. An overview of the PINN model is described hereafter.

[0019] A Physics-Informed Neural Network (PINN) is a type of Neural Network (ANN) that embeds the knowledge of any physical laws into its architecture to model complex systems. Unlike conventional neural networks, which rely purely on data, PINN leverages both observed data and known physics principles or physics laws to improve accuracy, generalization and efficiency in simulations. Most of the physical laws that govern the dynamics of a system may be described by Partial Differential Equations (PDEs). For example, the Navier-Stokes equations are a set of PDEs derived from the conservation laws (i.e., conservation of mass, momentum, and energy) that govern fluid mechanics. The solution of the Navier-stokes equations with appropriate initial and boundary conditions allows the quantification of flow dynamics in a precisely defined geometry. The initial and boundary conditions are physical constraints which ensure that the neural network learns solutions that are physically valid. The initial and boundary conditions are integrated into the PINN model as additional terms in the loss function. Thus, the PINN holds the ability to solve differential equations with a given set of initial conditions and boundary conditions using the dataset. The prior knowledge of the equations governing the physical laws acts as regularizing agents that limit the space of admissible solutions, which increases the correctness of the predictions. The PINN achieves the accuracy by defining a unique loss function and collocation points. The collocation points may be defined as a set of points in an input domain (typically space and / or time) where the governing physical equations, such as PDEs, are enforced during training. The unique loss function may be calculated by combining MSE loss and the physics loss (PDE). The description of the MSE loss and the PDE loss is discussed in the following description.

[0020] In general, the PINN architecture comprises an input layer, one or more hidden layers, and output layer. The input data at the input layer may include variables such as time, space, or other relevant system parameters. For example, if a heat distribution in a material is being modeled, the input may include spatial coordinates of different points within the material and the time at which the prediction is needed. In addition to the input data, initial conditions (i.e., initial temperature distribution) and boundary conditions (e.g., fixed temperatures at boundaries) may be incorporated to guide the solution and ensure that the network adheres to known physical constraints. More the number of collocation points and initial-boundary conditions are present, the accuracy of the PINN model is better. Once all the inputs (input data, initial-boundary conditions) are provided to the network, the inputs are passed through the one or more hidden layers. The one or more hidden layers processes the inputs through a series of non-linear transformations using activation functions. Thereafter, the output layer produces output in the form of predicted physical quantities such as temperature, pressure, velocity, etc.,

[0021] As described earlier, the PINN differs from the conventional neural network in manner that in addition to the traditional data-based loss also knows as Mean Squared Error (MSE) loss or simply data loss (Ldata), the physics loss (Lpde) is also included. The physics loss or PDE loss (Lpde) may be expressed as: — A,-! m—y + u—(1) Nc^-iX. dtt2 dtt lJ v 7

[0022] Where, — represent the average squared error over Nc, which represents the number of collocation points (specific points in space or time where the PDE is evaluated), d2x m is a physical constant, possibly the mass in a system, is the second derivative of the predicted output with respect to time, represents acceleration, / r is a physical constant, potentially representing damping or viscosity in the system, is the first derivative of the dt[ predicted output with respect to time, representing velocity, k is a constant (possibly the spring constant or stiffness in a system) and xt is the predicted output which represents displacement. The expression inside the parentheses is the residual of the PDE at each collocation point. Squaring the equation may ensure that positive and negative errors do not cancel out. At last, helps in normalizing the loss by the number of collocation points, averaging the squared residuals. Thus, the PDE loss ensures that the xt predicted output satisfies the underlying physics described by the PDE. The closer the residual is to zero at each collocation point, the better the solution matches the true physical behavior.

[0023] Based on the above, it may be concluded that the PINN calculates derivatives of the output with respect to inputs (e.g., time, space) using Automatic Differentiation (AD). Thereafter, the derivatives are used to determine if the physical laws (such as Navier-Stokes Equations, conservation of energy, etc.,) are being satisfied. Further, the PINN is trained by minimizing the combined loss function (i.e., the MSE loss, physics loss and boundary conditions) using any suitable optimization algorithm. The combined loss function may be expressed as: L XdataLdata~t ^PDeLpDE-*-^-boundaryLboundary (2) Where Xdata, Xpde, and ^boundary are boundary coefficients to balance the importance of the data, physics, and the boundary / initial condition losses. During the training, the network adjusts or finetunes the number of neurons, layers, collocation points and weights of the losses at each iteration to reduce the total loss. Consequently, the PINN model helps in minimizing errors related to physical laws instead of just minimizing errors between the predicted output and training data. A detailed explanation of the proposed technique(s) is disclosed in the forthcoming paragraphs.

[0024] Fig. 1 illustrates an exemplary environment 100 of a battery pack 102 of a vehicle interfacing with a cooling plate 104, in accordance with an embodiment of the present disclosure. The environment 100 is exemplified in a scenario where a plurality of cells (106a, 106b, 106c, ... 106n) of the battery pack 102 may be comprised within a plurality of battery modules and may be in direct contact of the cooling plate 104. The battery pack 102 may be configured in electrical vehicles (EVs) and is a collection of the plurality of battery cells 106 that stores and provides power to the vehicle’s electric motor. The plurality of cells 106a, 106n may be arranged into modules connected in series and enclosed in a battery casing. Although Fig. 1 illustrates a single cell (106a) within a battery module, it should be noted that the battery module may comprise one or more cells. The battery pack 102 generally comprises a cooling and heating system which helps in maintaining optimal temperatures of the battery. The cooling and heating system may comprise the cooling plate 104 with a coolant or fluid or circulating fluid (as shown in dashed lines in Fig. 1) that removes heat from the battery cells to keep the cells at the optimal temperature. In other words, the coolant or fluid flowing inside the cooling plate 104 may comprise water or refrigerant having a relatively high heat capacity to absorb the heat generated by the plurality of cells 106. The heat generated by the plurality of cells 106 is carried away with the help of the coolant via an outlet port.

[0025] In some embodiments, a rate at which the coolant flows inside the cooling plate 104 via an inlet port is known as a cooling inlet flow rate 108 and a temperature measured at the inlet port is known as coolant inlet temperature 110. In some embodiments, a rate at which the coolant flows outside the cooling plate 104 via an outlet port is known as a cooling outlet flow rate 116 and a temperature measured at the outlet port is known as coolant outlet temperature 118. In some embodiments, the cooling plate 104 may be constructed from a heat-conducting material, such as aluminum or ceramic. In some embodiments, the cooling plate 104 may be configured as either a direct cooling system or an indirect cooling system. In some embodiments, a rate at which the heat generated or dissipated from each of the plurality of cells 106 of the battery may be known as a heat dissipation rate 112. In some embodiments, location information 114 pertaining to each cell and the cell cooling interface may be required to determine the temperature at particular location or point. Thereafter, the coolant inlet flow rate, the coolant inlet temperature, the heat dissipation rate 112, and the location information 114 may be received by a battery management system (referred to as system) for controlling temperature variations across the battery pack 102. A detailed description of the system is provided in forthcoming paragraphs.

[0026] Fig. 2A shows an exemplary architecture of the system 202 for controlling temperature variations across a battery pack 102 of a vehicle, in accordance with an embodiment of the present disclosure. The system 202 may be a computing system that may comprise a processor 204, a PINN model 206, and a memory 208, etc. In some implementation, the system 202 may include other components (not shown in this fig.) to implement desired functions of the present disclosure. In one embodiment, the system 202 may be configured internally within the vehicle. In an alternative embodiment, the system 202 may be configured externally to the vehicle. The processor 204 may be configured to implement the functionality of the system 202. In some embodiments, the processor 204 may be communicatively coupled to the PINN model 206 and the memory 208 for implementing the functions of the system 202. The system 202 may be configured for controlling the temperature variations across the battery pack by receiving cooling inlet parameters 226 and cell parameters from one or more sensors 210.

[0027] In some embodiments, the one or more sensors 210 may be placed in the battery pack 102 in several locations for determining flow rate, temperature, pressure, and other factors. In some embodiments, one or more sensors 210 may comprise, but not limited to, temperature sensors, internal sensors, a ring terminal temperature sensor, a flow sensor or flow meter, thermal flow sensor, thermal sensor, thermocouples, etc., In some embodiments, the one or more sensors 210 may be configured to sense cooling inlet parameters 226 at the inlet port and cooling outlet parameters at the outlet port. The cooling inlet parameters 226 may comprise, but not limited to, a coolant inlet flow rate 108 a coolant inlet temperature 110, a coolant outlet flow rate 116, and coolant outlet temperature 118.

[0028] In some embodiments, the one or more sensors 210 may be configured to sense cell parameters of the plurality of cells 106 of the battery pack 102. The cell parameters may comprise, but are not limited to, a heat dissipation rate 112 from each of the plurality of cells 106 and location information 114 of the cell cooling plate interface. The cell cooling plate interface may be an interface between the cooling plate 104 and the plurality of cells 106. In some embodiments, the one or more sensors 210 may continuously measure the cooling inlet parameters 226 and the cell parameters in real-time. In some embodiments, the one or more sensors 210 may provide the coolant inlet parameters and the cell parameters to the system 202. The system 202 may utilize the cooling inlet parameters 226 and the cell parameters for controlling temperature variations across the battery pack 102.

[0029] In accordance with the present disclosure, the processor 204 may be configured to determine the temperature at the cell cooling plate interface based on the coolant inlet parameters and the cell parameters. In some embodiments, the cooling inlet parameters 226 and the cell parameters may be stored in the memory 208. The processor 204 may retrieve the cooling inlet parameters 226 and the cell parameters as per the requirements from the memory 208. To determine the temperature, the processor 204 may be configured to apply the PINN model 206. Further, the processor 204 may be configured to calculate a maximum temperature deviation across each cell based on the determined temperature. Thereafter, the processor 204 may be configured to determine an optimal coolant inlet flow rate 108 and an optimal coolant inlet temperature 110 for the cell cooling plate interface when the maximum temperature deviation exceeds a temperature threshold value. Based on the determined optimal inlet flow rate and the optimal inlet temperature, the processor 204 may be configured to control the temperature variations across the battery pack 102. A detailed description of the system 202 is provided in forthcoming paragraphs in conjunction with Figs. 1 and 2A.

[0030] Fig. 2B shows a detailed block diagram of the system 202 for controlling temperature variations across a battery pack 102 of a vehicle, in accordance with an embodiment of the present disclosure. The system 202 may comprise the processor 204, the PINN model 206, the memory 208, an Input / Output (I / O) interface 234, and one or more modules 212, but not limited thereto. In one implementation, the processor 204, the PINN model 206, and the term “module” may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and / or any devices that manipulate signals based on operational instructions. In one embodiment, the processor 204 may be configured as a standalone processing unit or combination of one or more processors for implementing the subject matter of the present disclosure. In some embodiments, the processor 204 may be implemented through software or hardware or a suitable combination of software and hardware as per the implementation requirements of the present disclosure. Among other capabilities, the processor 204 may be configured to fetch and execute computer-readable instructions and other information stored in the memory 208. In some embodiments, the PINN model 206 may be implemented for determining the temperature at the cell cooling plate interface. In a non-limiting embodiment, the PINN model 206 may use the cooling inlet parameters 226 and the cell parameters for determining the temperature at the cell cooling plate interface. In one implementation, the PINN model 206 may be internally configured within the processor 204. Alternatively, the PINN model 206 may be externally configured or separately built outside the system 202. In an exemplary embodiment, the PINN model 206 may be stored in the memory 208 and the processor 204 may fetch the instructions from the memory 208 to perform the desired functions. Further details related to the PINN model 206 are explained in reference to Fig. 3.

[0031] In one implementation, the I / O interface 234 may be any hardware component which may allow the system 202 to communicate with external devices such as peripherals and other computers. In some embodiments, the I / O interface 234 may be configured to receive the cooling inlet parameters 226 and the cell parameters, etc. from the one or more sensors. In some implementations, the memory 208 may be configured to store the received the cooling inlet parameters 226 and the cell parameters which may be used by the system 202 for controlling temperature variations across the battery pack of the vehicle.

[0032] In some embodiments, the one or more modules 212 may comprise a receiver module 214, a determining module 216, a calculation module 218, a control module 220, and other modules 222. It may be well appreciated by a person skilled in the art that the one or more modules 212 may include additional modules to implement desired functions of the present disclosure. In some embodiments, the one or more modules 212 may perform functions carried by the processor 204. It shall be appreciated that such the modules 214-222 may be represented as a single module or a combination of different modules 212.

[0033] In some implementations, the memory 208 may be an external memory chip or an inbuilt Electrically Erasable Programmable Read-Only Memory (EEPROM) memory of the vehicle. The memory 208 may be configured to store data 224 which may comprise, but not limited to, cooling inlet parameters 226, cell parameters 228, temperature threshold value 230, or any other data 232 in the form of various data structures. Additionally, the data may be organized using data models, such as relational or hierarchical data models. In one embodiment, the cooling inlet parameters 226 may comprise, but not limited to, the coolant inlet flow rate 108, the coolant inlet temperature 110, the coolant outlet flow rate 116, the coolant outlet temperature 118 etc., In some embodiments, the cell parameters 228 may comprise, but not limited to, the dissipation rate, location information 114 of each cell, etc., In another embodiment, the temperature threshold 230 may include, but not limited to, a predefined temperature threshold value. In some embodiments, the other data 232 may include various temporary data and files generated by the processor 204 or one or more modules 212 while performing various functions of the present disclosure. In some embodiments, the other data 232 may comprise but not limited to percentage of variation, a modified coolant inlet flow rate 108, a modified coolant inlet temperature 110, an optimal coolant inlet flow rate 108, and an optimal coolant inlet temperature 110, etc., In some embodiments, the memory 208 may be communicatively coupled to the processor 204 and the one or more modules 212 and may be configured to store various data processed by the processor 204 and the one or more modules 212.

[0034] As discussed in paragraph

[0039] , the processor 204 may be configured to receive the cooling inlet parameters 226 associated with the cell cooling plate 104 and the cell parameters 228 of the plurality of cells 106 of the battery. The processor 204 may receive the cooling inlet parameters 226 and the cell parameters 228 from the one or more sensors 210 placed on several locations of the battery pack and the cooling plate 104. Upon receiving the cooling inlet parameters 226 and the cell parameters 228, the processor 204 may be configured to determine the temperature at the cell cooling plate interface by applying the PINN model 206. In some embodiments, the PINN model 206 may be pretrained PINN model used to determine the temperature at any desired location of the plurality of cells 106 or the cell cooling plate interface. The training of the PINN model 206 is described in reference to Fig. 3. Thereafter, the processor 204 may be configured to calculate a maximum temperature deviation across each cell based on the determined temperature. In a non-limiting embodiment, the processor 204 may calculate the maximum temperature deviation by computing a difference between a maximum temperature and a minimum temperature determined at each cell. In some embodiments, the processor 204 may be configured to determine at least one of: the optimal coolant inlet flow rate 108 and the optimal coolant inlet temperature 110 for the cell cooling plate interface when the maximum temperature deviation exceeds the temperature threshold value 230. In an exemplary embodiment, the optimal coolant inlet temperature 110 may be a desired coolant inlet temperature 110 which optimizes the battery performance and the battery lifespan. In another exemplary embodiment, the optimal coolant inlet flow rate 108 may be a desired inlet flow rate which is required to flow inside the cooling plate 104 to cool the cells to maintain the optimal temperature of the cells. Further, the processor 204 may be configured to control the temperature variations across the battery pack based on the optimal coolant inlet flow rate 108 and the optimal coolant inlet temperature 110. In some embodiments, the processor 204 may control the temperature variations by controlling the coolant inlet flow rate 108 and the coolant inlet temperature 110 based on the optimal coolant inlet flow rate 108 and the optimal coolant inlet temperature 110. In one implementation, to control the coolant inlet flow rate 108 and the coolant inlet temperature 110, the processor 204 may be configured to increase or decrease the coolant inlet flow rate 108 based on the determined optimal coolant inlet flow rate 108. In another implementation, the processor 204 may be configured to increase or decrease the coolant inlet temperature 110 based on the determined optimal coolant inlet temperature 110.

[0035] In some embodiments, the processor 204 may be configured to control internal cell temperatures based on the optimal coolant inlet flow rate 108 and the optimal coolant inlet temperature 110. The internal cell temperatures may be determined in a manner like the temperature measurement at the cooling plate interface. It may be noted to a person skilled in the art that the internal cell temperature may be determined at different locations on the plurality of cells 106 apparat from the cell cooling plate interface. Post determining the internal cell temperature, the processor 204 may be configured to identify higher temperature values and lower temperature values within the plurality of cells 106. In an exemplary embodiment, identifying a region on the cell with higher temperature values may be known as hotspots where the temperature may be significantly higher than the surrounding areas. In another exemplary embodiment, identifying a region on the cell with lower temperature values may be known as cold spots where the temperature may be significantly lower than the surrounding areas.

[0036] In some embodiments, to determine the temperature at the cell cooling plate interface, the processor 204 may be configured to implement an optimization model. It may be worth noting that the present disclosure may implement any suitable optimization model for determining the temperature at the cell cooling plate interface. In one implementation, the processor 204 may be configured to implement the optimization model by evaluating a modified coolant inlet flow rate 108 and a modified coolant inlet temperature 110. To evaluate the modified coolant inlet flow rate 108 and the modified coolant inlet temperature 110, the processor 204 may be configured to calculate a difference between the maximum temperature variation and the temperature threshold variation. Thereafter, the processor 204 may be configured to calculate a percentage of variation (percentage_varl) which indicates a degree to which the coolant inlet temperature 110 exceeds the temperature threshold value 230. In some embodiments, the processor 204 may be configured to determine the first percentage of variation by dividing the percentage of variation (percentage_varl) by a factor of 2.

[0037] In an exemplary embodiment, the processor 204 may be configured to determine the modified coolant inlet flow rate 108 and the modified coolant inlet temperature 110 using below equations: modified coolant inlet flow rate = coolant inlet flow rate*(l + percentage_varl) (3) modified coolant inlet temperature = coolant inlet temperature*(l - percentage_varl) (4) It may be worth noting that the coolant inlet flow rate 108 and the coolant inlet temperature 110 refers to the received coolant inlet flow rate 108 the received coolant inlet temperature 110.

[0038] In a non-limiting embodiment, the processor 204 may be configured to provide the modified coolant inlet flow rate 108 and the modified coolant inlet temperature 110 as inputs to the PINN model 206 by maintaining the cell parameters 228 for determining whether the maximum temperature variation exceeds the temperature threshold value 230. In one embodiment, upon determining that the maximum temperature variation does not exceed the temperature threshold value 230, the processor 204 may be configured to control the coolant inlet flow rate 108 and the coolant inlet temperature 110 in accordance with the modified coolant inlet flow rate 108 and modified coolant inlet temperature 110. In another embodiment, upon determining that the maximum temperature variation exceeds the temperature threshold value 230, the processor 204 may be configured to apply a Bisection method or Secant method to determine a percentage of variation (percentage_var2). It may be well appreciated by a person skilled the art that the percentage of variation (percentage_varl) and percentage of variation (percentage_var2) differ from each other.

[0039] Based on the determined percentage of variation (percentage_var2), the processor 204 may be configured to subsequently modify or reupdate the coolant inlet flow rate 108 and the subsequently modify coolant inlet temperature 110. Thereafter, the processor 204 may be configured to provide the subsequently modified coolant inlet flow rate 108 and the subsequently modified coolant inlet temperature 110 to the PINN model 206 as new inputs. The new inputs may be provided to determine whether the subsequently modified coolant inlet flow rate 108 and the subsequently modified coolant inlet temperature 110 meets the temperature threshold value 230. Finally, the processor 204 may be configured to control the temperature variations based on the subsequently modified coolant inlet flow rate 108 and the subsequently modified coolant inlet temperature 110 for which the temperature threshold value 230 is satisfied. It is further noted that any other suitable method may be utilized for determining the percentage variation apart from the Bisection method and the Secant method.

[0040] Thus, the present disclosure provides a robust and reliable thermal management system for controlling temperature variations at each point of the cell cooling plate interface. Further, by accurately determining individual cell temperatures with the help of the PINN model 206, the present disclosure ensures optimized temperature management, thereby enhancing battery performance and lifespan. Moreover, the PINN model 206 utilized in determining the temperature provides precise temperature predictions at any point on the cell cooling plate interface and can be extended seamlessly to entire battery modules. Additionally, the utilization of PINN model 206 is effective in modeling multiphysics system, capturing both the interaction between solid and liquid / fluid components and the liquid flow inside the battery pack 102.

[0041] Fig. 3 exemplary schematic diagram of a PINN model 206 for determining temperature at the cell cooling plate interface, in accordance with an embodiment of the present disclosure. The PINN architecture comprises an input layer, one or more hidden layers, and an output layer. In an exemplary embodiment, the PINN model 206 may be trained using a dataset at the input layer. The dataset may comprise the cooling inlet parameters 226 (also referred to as liquid flow parameters) associated with the cell cooling plate and the cell parameters 228 (also referred to as solid-state parameters) associated with the plurality of cells 106. The cooling inlet parameters 226 may include the coolant inlet flow rate 108, the coolant inlet temperature 110 and the cell parameters 228 may include the heat dissipation rate (Q) 112 and the location information (also referred to as 3-D location coordinates in X, Y, and Z directions) 114 of each cell. In addition to the dataset, initial conditions (i.e., initial temperature distribution) and boundary conditions (e.g., fixed temperatures at boundaries) are incorporated to guide the output and to ensure that the network adheres to known physical constraints. All the inputs (dataset, and initial-boundary conditions) are fed into the neural network model 302. The inputs are passed through the one or more hidden layers of the neural network model 302 to train the PINN model 206. The one or more hidden layers processes the inputs through a series of non-linear transformations using activation functions. In a non-limiting embodiment, the non-linear transformations allow the network to learn and model more complex relationships within the inputs. Further, the activation function is a mathematical function applied to the output of each neuron in the hidden layer.

[0042] The PINN model 206 may be trained to determine outputs 304 at the output layer. The outputs may comprise at least one of: a velocity of the coolant inlet flow rate (u), a coolant pressure (p), a coolant temperature (Tf), and a temperature associated with the cell cooling plate interface (Ts). In other words, the PINN may be trained by determining the PDE loss and the MSE loss. In some embodiments, the PDE loss may be determined by using at least one of: a Navier-Stokes Equation, a heat diffusion equation, and a coupling loss (i.e., where the heat transfer takes place between the solid and the fluid). In a nonlimiting embodiment, the Navier-Stokes Equation may be used for calculating coolant / fluid physics (PDE) loss using the following equations: Continuity Equation: 7 ■ V = 0 (5) Momentum Equation: = g — Up + V2V (6) Energy Equation: = ^qx,y,z + V2T (7) Where 7 ( nabla operator) is divergence of the velocity vector field u, p is dynamic viscosity of the coolant / fluid, p is density of coolant / fluid, cp means thermal capacity of the coolant / fluid, kftuid is thermal conductivity of the fluid / coolant, DV is total derivative of the velocity, Dt is the derivative of time, Tf is the temperature of the coolant / fluid, 7 is the velocity, qXiyiZ is an array of heat dissipation rate 112 in 3-D location coordinates 114 from the plurality of cells 106, Vp is the pressure through Laplace.

[0043] In a non-limiting embodiment, the heat diffusion equation may apply Laplace equation for calculating the solid (i.e., the interaction between the cooling plate 104 and the plurality of cells 106) physics loss using the following equation: + g2U + g2U = n / 9A ax2 + dy2 + az2 0 (8) Where Ts is the temperature of the solid and x, y, z are 3D location coordinates 114 of each of the plurality of cells 106.

[0044] In a non-limiting embodiment, the coupling loss may be determined using the following equation: Tf = Ts &= ^Y(x,y,z) e Interface Points (9)

[0045] It will be appreciated by those skilled in the art that the Mean Squared Error (MSE) loss is a well-known concept and may be computed using any suitable method. In an exemplary embodiment, the MSE loss may be determined to measure the difference between predicted values and actual values. Also, it may be important to note here that the PDE loss and the MSE loss may be determined at each iteration during the training. Now the combination of PDE loss and the MSE loss is known as the loss function 306. The loss function is also determined iteratively during the training. Further, it is determined whether the loss function is less than or greater than the interface points. Upon determining that the loss function is greater than the interface points, a backpropagation algorithm is utilized by adjusting the weights of the PINN model 206 based on the MSE loss and the physics loss such that the PINN model 206 is trained to provide minimum physics loss at each interface point. Alternatively, upon determining that the loss function is less than the interface points, the computation ends 312. Accordingly, to ensure optimal efficiency and performance of the battery, the loss function should be as minimal as possible. Once the PINN model 206 is trained, the PINN model 206 is deployed or integrated within the system 202 for determining the temperature at the interface of the cell cooling plate 104. In accordance with the present disclosure, the PINN model 206 is configured to effectively model multiphysics systems, capturing both the interactions between solid and fluid components, as well as the fluid flow within the battery cooling systems. The PINN model 206 can be seamlessly extended across the entire battery modules, facilitating precise temperature predictions at any point along the battery coolant interface (also referred to as cell-cooling interface). 5

[0046] Fig. 4 illustrates an exemplary environment of a cell cooling plate 104, showing the cell contact surface area for solving a multi-physics problem using the PINN model 206, in accordance with an embodiment of the present disclosure. The exemplary environment indicates a cell contact surface area of the cell cooling plate interface 402 at which the temperature is determined. In an exemplary scenario, considering the coolant 10 inlet flow rate is 1LPM, the coolant inlet temperature is -20°C, and the heat dissipation rate 112 for all the plurality of cells is 5.2W. In an exemplary embodiment, the PINN model 206 may be trained using the exemplary dataset shown below: 1. Solid Wall Points: 51,150 2. Inlet point 504 3. Outlet Points: 790 4. Interface Points: 27,688 5. Fluid Internal Points 167,047 6. Solid Internal points: 156,617 7. Internal points used fortraining: 0 8. Interface Points used for training 0 9. Collocation points Fluid internal points + solid internal points + interface points 15 Table A: Exemplary Training dataset Thus, based on the above, the PINN model 206 may be trained based on the training dataset and the coolant inlet flow rate 108, the coolant inlet temperature 110, and the heat dissipation rate 112 and the neural network model may be employed for each output. 20 Further, at least one of the outputs is applied in at least one of: the Navier-Stokes Equation for calculating fluid physics loss, the heat diffusion equation for calculating solid physics loss, and the coupling loss. It may be noted that the boundary points from the above training data include the inlet-outlet points used for calculating the MSE loss of the cooling plate 104 and solid wall points used for calculating the MSE loss of the plurality of cells 106. 25 Further, the fluid volume points are used as collocation points to solve the Navier-Stokes Equation for calculating fluid physics loss. Furthermore, the solid volume points are used as collocation points to solve the Laplace Equation for heat diffusion. Finally, the interface points are used to ensure that the determined temperature of both the solid and fluid model are equal. The following illustrates the exemplary fluid and solid properties applied in the aforementioned equations for determining the physics (PDE) loss: kg J u = 0.029229325 Pa-s -, p = 1093.5 = 3037-^--^ m3 p kg W kfiutd = 0-307- ■ K; ksoUd = 237 W / m ■ K; Where p is dynamic viscosity of the coolant / fluid, p is density of coolant / fluid, cp means thermal capacity of the coolant / fluid, kftuid is thermal conductivity of the fluid / coolant, and ksoiid is thermal conductivity of the solid or cell. It will be readily understood by those skilled in the art that the foregoing is provided solely for illustrative purposes to explain the concept of applying PINN to solve solid and fluid problems in a single cell cooling plate 104. Hence, it should not be construed as limiting the scope of the invention.

[0047] Fig. 5 depicts a method 500 of controlling temperature variations across a battery pack of a vehicle, in accordance with an embodiment of the present disclosure. The description of Fig. 5 is provided in conjunction with Figs. 1-4 for clarity purposes. The method 500 may be described in the general context of computer executable instructions. Generally, computer executable instructions may include routines, programs, objects, components, data structures, procedures, modules, and functions, which perform specific functions or implement specific abstract data types. The order in which the method 500 is described is not intended to be construed as a limitation, and any number of the described method blocks may be combined in any order to implement the method. Additionally, individual blocks may be deleted from the methods without departing from the spirit and scope of the subject matter described. Further, it may be noted that the method 500 may be performed by a processor 204 or any of the modules 212 of the system 202, as shown in Figs. 2B.

[0048] At block 502, the method 500 may comprise receiving cooling inlet parameters 226 associated with the cell cooling plate 104 and cell parameters 228 of a plurality of cells 106 of the battery. In a non-limiting embodiment, the cooling inlet parameters 226 may comprise at least one of a coolant inlet flow rate 108 and a coolant inlet temperature 110. Further, the cell parameters 228 may comprise at least one of: a heat dissipation rate 112 and location information 114 of each cell. The method at block 502 may be performed by the processor 204 or the receiver module 214. At block 504, the method 500 may comprise determining the temperature at the cell cooling plate interface by applying the PINN model 206. The method at block 504 may be performed by the processor 204 or the determining module 216. In some embodiments, the PINN model 206 may comprise building the PINN model 206. To build the PINN model 206, the method may comprise providing a dataset related to the cell cooling plate 104 and the plurality of cells 106 of the battery to train the PINN model 206. The PINN model 206 may comprise an input layer, one or more hidden layers, and an output layer. In some embodiments, the dataset may include at least one of: 3-dimensional (3-D) location coordinates 114 of each cell, liquid flow parameters of the cell cooling plate 104, and solid-state parameters of the plurality of cells 106. In a nonlimiting embodiment, the liquid flow parameters may relate to a mass flow rate and an inlet temperature of the coolant flowing inside the cooling plate 104. In some embodiments, the solid-state parameters may comprise at least one of 3D location coordinates and a heat dissipation rate 112. In a non-limiting embodiment, the 3D location coordinates 114 may be spatial coordinates of each cell in X-direction, Y-direction, and Z-direction where the temperature is to be determined. In an embodiment, the heat dissipation rate 112 may relate to heat rejection from the plurality of cells 106.

[0049] Further, the method may comprise training the PINN model 206 by determining a Partial Differential Equation (PDE) loss and a Mean Squared Error (MSE) loss. The PINN model 206 is trained to determine at least one of outputs comprising a velocity of the coolant inlet flow rate (u), a coolant pressure (p), a coolant temperature (Tf), and a temperature associated with the cell cooling plate interface (Ts). It may be worth noting that method may comprise determining the PDE loss and the MSE loss at each iteration during the training. In some embodiments, for determining the PDE loss, the method may further comprise evaluating a coolant PDE loss function by applying Navier-Stokes Equation using at least one of the outputs. Thereafter, the method may comprise evaluating a cell cooling plate interface PDE loss function by applying a heat diffusion equation using at least one of the outputs. Finally, the method may comprise evaluating a coupling PDE loss function using the at least one of the outputs. Next, for the training the PINN model 206, the method may comprise determining a loss function (previously referred to as combined loss function) based on the determined PDE loss and the MSE loss. It may be worth noting that method may comprise determining the PDE loss and the MSE loss at each iteration during the training. Thereafter, at block 506, the method may comprise calculating a maximum temperature deviation across each cell based on the determined temperature. The maximum temperature deviation may be calculated by computing difference between the maximum temperature and the minimum temperature of each cell. The method at block 506 may be performed by the processor 204 or the calculation module 218. At block 508, the method may comprise determining at least one of an optimal coolant inlet flow rate 108 and an optimal coolant inlet temperature 110 for the cell cooling plate interface when the maximum temperature deviation exceeds a temperature threshold value 230. The method at block 508 may be performed by the processor 204 or the determining module 216. At block 510, the method may comprise controlling temperature variations across the battery 102 pack based on the determined optimal coolant inlet flow rate 108 and the optimal coolant inlet temperature 110. For controlling the temperature variations, the method may further comprise controlling the coolant inlet flow rate 108 and the coolant inlet temperature 110 based on the determined optimal coolant inlet flow rate 108 and the optimal coolant inlet temperature 110. In one embodiment, for controlling the coolant inlet flow rate 108, the method may comprise increasing or decreasing the coolant inlet flow rate 108 based on the determined optimal coolant inlet flow rate 108. In another embodiment, for controlling the coolant inlet temperature 110, the method may comprise increasing or decreasing the coolant inlet temperature 110 based on the determined optimal coolant inlet temperature 110. The method at block 510 may be performed by the processor 204 or the control module 220.

[0050] Fig. 6 represents a flowchart depicting a method of determining temperature at the cell cooling plate interface by implementing an optimization model, in accordance with an embodiment of the present disclosure. The description of Fig. 6 is provided in conjunction with Figs. 1-5 for clarity purposes. The method 600 may be described in the general context of computer executable instructions.

[0051] The order in which the method 600 is described is not intended to be construed as a limitation, and any number of the described method blocks may be combined in any order to implement the method. Additionally, individual blocks may be deleted from the methods without departing from the spirit and scope of the subject matter described. Further, it may be noted that the method 500 may be performed by a processor 204 or other modules 222 of the system 202, as shown in Figs. 2B.

[0052] At block 602, the method may comprise calculating a difference between the maximum temperature variation and the temperature threshold variation. Further, the method at block 604 may comprise calculating a percentage of variation (percentage_varl) based on the calculated difference. The percentage of variation indicates a degree to which the coolant inlet temperature 110 exceeds the temperature threshold value 230. In some embodiments, the method may comprise determining the first percentage of variation by dividing the percentage of variation (percentage_varl) by a factor of 2. Thereafter, at block 606, the method may comprise evaluating a modified coolant inlet flow rate 108 and a modified coolant inlet temperature 110 based on the percentage of variation. In some embodiments, the method may comprise providing the modified coolant inlet flow rate 108 and the modified coolant inlet temperature 110 as inputs to the PINN model 206 by maintaining the cell parameters 228 for determining whether the maximum temperature variation exceeds the temperature threshold value 230 as shown in block 608. In one embodiment, upon determining that the maximum temperature variation does not exceed the temperature threshold value 230, the method may comprise controlling the coolant inlet flow rate 108 and the coolant inlet temperature 110 in accordance with the modified coolant inlet flow rate 108 and modified coolant inlet temperature 110 as shown in block 612. In another embodiment, upon determining that the maximum temperature variation exceeds the temperature threshold value 230, the method may comprise applying a Bisection method or Secant method for determining a percentage of variation (percentage_var2). It may be well appreciated by a person skilled the art that the percentage of variation (percentage_varl) and percentage of variation (percentage_var2) differ from each other. Based on the determined percentage of variation (percentage_var2), the method may comprise subsequently modifying or reupdating the coolant inlet flow rate 108 and subsequently modifying the coolant inlet temperature 110. Thereafter, the method may comprise providing the subsequently modified coolant inlet flow rate 108 and the subsequently modified coolant inlet temperature 110 to the PINN model 206 as new inputs. Furthermore, the method may comprise determining whether the subsequently modified coolant inlet flow rate 108 and the subsequently modified coolant inlet temperature 110 meets / satisfies the temperature threshold value 230. Finally, at block 610, the method may comprise controlling the temperature variations based on the subsequently modified coolant inlet flow rate 108 and the subsequently modified coolant inlet temperature 110 for which the temperature threshold value 230 is satisfied. It is further noted that any other suitable method may be utilized for determining the percentage variation apart from the Bisection method and the Secant method.

[0053] The Physics-Informed Neural Network (PINN) method disclosed in the present disclosure will optimize the cooling efficiency of battery systems, directly contributing to prolonged battery life while significantly reducing the risk of thermal runaway events. As batteries operate within their designed temperature thresholds, predictive accuracy will improve, leading to an extended electric vehicle (EV) range. This increase in range will enhance customer experience, reducing any associated discomfort. Furthermore, mitigating the risks of thermal runaway provides customers with peace of mind, preventing potential accidents that could result in serious injuries or fatalities and avoiding negative impacts on the company's reputation. The extended longevity of the batteries further adds to the system’s efficiency and reliability.

[0054] _The proposed disclosure is particularly advantageous for direct cooling systems. In such systems, where the cells are submerged in the coolant, the coupling loss will include an additional term to account for the physics loss at the coolant-cell interface. However, since the PINN formulation incorporates the geometry of the battery pack by defining boundary conditions as boundary loss, the careful segmentation of each individual cell wall's coordinates allows these to be treated as boundaries. This enables the deployment of additional PINN to model the thermal behavior of the battery cooling system in direct cooling scenarios. The model may be extended to account for the behavior of the thermal interface material (TIM) with minimal modifications to the existing conduction equations governing the solid region. The proposed Navier-Stokes Equation Battery Cooling System, as described in paragraph

[0064] , addresses incompressible flows. However, it can be adapted to account for coolant properties that vary with temperature, facilitating the 5 modeling of compressible flows. The proposed Navier-Stokes Equation Battery Cooling System as described in paragraph

[0064] is adaptable to account for both steady and unsteady coolant flow behavior for temperature predictions.

[0055] A description of an embodiment with several components in communication with each other does not imply that all such components are required. On the contrary, a variety 10 of optional components are described to illustrate the wide variety of possible embodiments of the disclosure.

[0056] While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be 15 limiting, with the true scope and spirit being indicated by the following claims.

[0057] Claims:

Claims

1. A method of controlling temperature variations across a battery pack of a vehicle, the method comprising:receiving cooling inlet parameters associated with a cell cooling plate and cell parameters of a plurality of cells of the battery, wherein the cooling inlet parameters comprise at least one of a coolant inlet flow rate and a coolant inlet temperature, and wherein the cell parameters comprise at least one of: a heat dissipation rate and location information of each cell;determining the temperature at the cell cooling plate interface by applying a Physical Informed Neural Network (PINN) model;calculating a maximum temperature deviation across each cell based on the determined temperature;determining at least one of an optimal coolant inlet flow rate and an optimal coolant inlet temperature for the cell cooling plate interface when the maximum temperature deviation exceeds a temperature threshold value; andcontrolling the temperature variations across the battery pack based on the determined optimal coolant inlet flow rate and the optimal coolant inlet temperature.

2. The method of claim 1, wherein controlling the temperature variations further comprising:controlling the coolant inlet flow rate and the coolant inlet temperature based on the determined optimal coolant inlet flow rate and the optimal coolant inlet temperature.

3. The method of claim 1, wherein the PINN model is built using steps comprising: providing a dataset related to the cell cooling plate and the plurality of cells of the battery to train the PINN model, wherein the PINN model comprises an input layer, one or more hidden layers and an output layer, and wherein the dataset includes at least one of 3-dimensional (3-D) location coordinates of each cell, liquid flow parameters of the cell cooling plate, and solid-state parameters of the plurality of cells, wherein the liquid flow parameters relate to the cell cooling plate and the solid-state parameters relate to the plurality of cells, liquid flow within the battery pack, and the interaction between cell cooling plate and the plurality of cells; andtraining the PINN Model by determining a Partial Differential Equation (PDE) loss and a Mean Squared Error (MSE) loss,wherein the PINN model is trained to determine at least one of outputs comprising a velocity of the coolant inlet flow rate (u), a coolant pressure (p), a coolant temperature (Tf), and a temperature associated with the cell cooling plate interface (Ts), andwherein the PDE loss and the MSE loss is determined iteratively during the training.

4. The method of claim 3, wherein determining the PDE loss comprising:evaluating a coolant PDE loss function by applying Navier-Stokes Equation using at least one of the outputs;evaluating a cell cooling plate interface PDE loss function by applying a heat diffusion equation using at least one of the outputs; andevaluating a coupling PDE loss function using the at least one of the outputs.

5. The method of claim 1, wherein controlling the coolant inlet flow rate and the coolant inlet temperature, further comprising:increasing or decreasing the coolant inlet flow rate based on the determined optimal coolant inlet flow rate; andincreasing or decreasing the coolant inlet temperature based on the determined optimal coolant inlet temperature.

6. The method of claim 1, further comprising:controlling internal cell temperatures based on the determined optimal coolant inlet flow rate and the optimal coolant inlet temperature; andidentifying higher temperature values and lower temperature values within the plurality of cells based on the internal cell temperatures.

7. The method of claim 1, wherein the temperature at the cell cooling plate interface is determined by implementing an optimization model using steps comprising:evaluating a modified coolant inlet flow rate and a modified coolant inlet temperature iteratively based on a percentage of variation, wherein the percentage of variation indicates a degree to which the coolant inlet temperature exceeds the temperature threshold value,providing the modified coolant inlet flow rate and the modified coolant inlet temperature to the PINN model iteratively to determine whether the maximum temperature variation exceeds the temperature threshold value; andcontrolling the coolant inlet flow rate and the coolant inlet temperature according to the modified coolant inlet flow rate and the modified coolant inlet temperature for which the temperature threshold value is satisfied.

8. A system for controlling temperature variations across a battery pack of a vehicle, the system comprising:one or more sensors configured to sense cooling inlet parameters associated with the cell cooling plate and cell parameters of a plurality of cells of the battery, wherein the cooling inlet parameters comprise at least one of a coolant inlet flow rate and a coolant inlet temperature, and wherein the cell parameters comprise at least one of: a heat dissipation rate and location information of each cell; andat least one processor coupled to the one or more sensors, wherein the at least one processor is configured to:receive the cooling inlet parameters associated with the cell cooling plate and the cell parameters of a plurality of cells of the battery;determine the temperature at the cell cooling plate interface by applying a Physical Informed Neural Network (PINN) model;calculate a maximum temperature deviation across each cell based on the determined temperature;determine at least one of an optimal coolant inlet flow rate and an optimal coolant inlet temperature for the cell cooling plate interface when the maximum temperature deviation exceeds a temperature threshold value; andcontrol the temperature variations across the battery pack based on the determined optimal coolant inlet flow rate and the optimal coolant inlet temperature, wherein to control the temperature variations, the at least one processor is configured to control at least the coolant inlet flow rate and the coolant inlet temperature.

9. The system of claim 8 , wherein the at least one processor is further configured to built the PINN model by:providing a dataset related to the cell cooling plate and the plurality of cells of the battery to train the PINN model, wherein the PINN model comprises an input layer, one or more hidden layers and an output layer, and wherein the dataset includes at least one of 3-dimensional (3-D) location coordinates of each cell, liquid flow parameters of the cell cooling plate, and solid-state parameters of the plurality of cells, wherein the liquid flow parameters and the solid-state parameters relate to the cell cooling plate, liquid flow within the battery pack, and the interaction between the cell cooling plate and the plurality of cells; andtraining the PINN Model by determining a Partial Differential Equation (PDE) loss and a Mean Squared Error (MSE) loss, wherein the PINN model is trained to determine at least one of outputs comprising a velocity of the coolant inlet flow rate (u), a coolant pressure (p), a coolant temperature (Tf), and a temperature associated with the cell coolingplate interface (Ts), and wherein the PDE loss and the MSE loss is determined iteratively during the training.

10. The system of claim 8, wherein to control the coolant inlet flow rate and the coolant 5 inlet temperature, the at least one processor is further configured to:increase or decrease the coolant inlet flow rate based on the determined optimal coolant inlet flow rate; andincrease or decrease the coolant inlet temperature based on the determined optimal coolant inlet temperature.10s

Images