System and method for determining the rechargeable or dischargeable energy of a battery energy storage system

Abstract

A system and method for determining the total rechargeable / dischargeable energy of a subsystem in a battery energy storage system (BESS) are disclosed. The iterative process is performed over a dynamic time period divided into multiple iterations using a neural network model that includes an energy prediction submodel and a state prediction submodel. The energy prediction submodel outputs the rechargeable / dischargeable energy of the subsystem for the current iteration. The state prediction submodel outputs the subsystem voltage for the next iteration, the subsystem charge rate for the next iteration, the subsystem's maximum temperature for the next iteration, and the difference in charge rate for the next iteration. The total rechargeable / dischargeable energy of the BESS subsystem is determined by summing the dischargeable energies determined for each iteration.

Description

TECHNICAL FIELD

[0001] This application claims priority to and the benefit of U.S. Provisional Application No. 63 / 601,641 filed on November 21, 2023, the disclosure of which is incorporated herein by reference in its entirety.

[0002] The present disclosure relates to managing a battery energy storage system (BESS), and more specifically, to accurately determining the chargeable / dischargeable energy of the BESS. BACKGROUND ART

[0003] BESS has become a critical component in modern energy management systems. Essentially, with the increasing integration of intermittent renewable power sources such as wind and solar power, energy storage solutions are required to ensure power grid stability and efficient power distribution. BESS technology enables storing excess power during low-demand periods and discharging insufficient power during high-demand periods, thereby optimizing energy usage (by reducing curtailment of solar and wind power output) and reducing dependence on fossil fuel-based power generation such as gas turbines. Such capabilities become particularly important as the global transition to cleaner energy sources accelerates and intermittent power sources gain a larger share in the power supply mix.

[0004] Because battery cells degrade over time (also known as "fading"), it is crucial to accurately predict the available charge capacity at any point in the BESS lifecycle. In some cases, and even when battery cells are predicted to pass the test, they may fail the capacity test. A conventional calculation for estimating the available charge capacity of a BESS is as follows: (Site average SOC (State of Charge)%) x (Rated capacity) x (Capacity fade%) x (Number of BESS racks). This calculation has inherent inaccuracies and can affect the ability to fully discharge a BESS, thus reducing market revenue. [Overview of the project] [Problems that the invention aims to solve]

[0005] Accordingly, this disclosure describes a system and method for determining the rechargeable / dischargeable energy of a BESS subsystem at rated (specified) power. A neural network model considers the voltage, charge rate, and maximum temperature of the BESS subsystem to predict the total rechargeable / dischargeable energy of the BESS over a dynamic time period divided into multiple iterations (i.e., time intervals). The neural network model includes an energy prediction submodel and a state prediction submodel. The energy prediction submodel predicts the rechargeable / dischargeable energy for the current iteration, and the state prediction submodel predicts the input features (e.g., operating state including voltage, charge rate, and maximum temperature) for the next iteration. The neural network model repeats the multiple iterations over the time period, summing the rechargeable / dischargeable energy for each iteration until it reaches the total rechargeable / dischargeable energy of the BESS subsystem over the time period. This system and method offers several advantages over conventional BESS energy calculations.

[0006] Firstly, BESS users using conventional energy calculations may incur penalties and consequently revenue losses because they may submit bids to electricity markets with incorrect rechargeable / dischargeable energy values ​​(e.g., overestimated or underestimated). In contrast, this system and method can consistently provide relatively accurate dischargeable (i.e., marketable) energy values, thus significantly reducing the likelihood of inducing penalties and revenue losses.

[0007] Secondly, conventional energy calculations can misestimate the rechargeable / dischargeable energy of a BESS for the entire SOC range (e.g., 95%), allowing users to bid conservative estimates in the market, which can lead to revenue losses. In contrast, this system and method can automatically calculate the rechargeable / dischargeable energy at time intervals (e.g., every 2 minutes) based on the current operating state of the system. This real-time calculation means that there is no need for any estimation regarding how much energy to discharge (and thus improve revenue generation opportunities in the electricity market).

[0008] Thirdly, users may not be aware of the (significant) energy loss due to cell voltage imbalance, temperature changes, and fade (i.e., battery degradation), which may be the dominant factors, on a daily basis. In contrast, the rechargeable / dischargeable energy values ​​generated by this system and method automatically indicate the gap (relating to the installed or identified energy), thereby highlighting the negative impact of any cell voltage imbalance and other factors on the rechargeable / dischargeable energy.

[0009] In one aspect, the disclosure relates to a system for determining the total rechargeable / dischargeable energy of a subsystem of a battery energy storage system (BESS), wherein the subsystem includes battery cells, and the system includes a controller which includes one or more processing modules and one or more non-temporary memory storage modules for storing computing instructions, wherein the computing instructions are configured to perform an iterative process over a dynamic time period when executed by one or more processing modules, and the dynamic time period is divided into multiple iterations by using a neural network model which includes an energy prediction submodel and a state prediction submodel, and for each iteration of the multiple iterations, the controller includes: (1) the energy prediction submodel which includes the subsystem's voltage for the current iteration of the multiple iterations, and the subsystem's voltage for the current iteration. (2) The energy prediction submodel is configured to output the charge rate and the maximum temperature of the subsystem for the current iteration, where the energy prediction submodel is configured to output the chargeable / dischargeable energy of the subsystem for the current iteration; and (2) the state prediction submodel is configured to input the voltage of the subsystem for the current iteration, the charge rate of the subsystem for the current iteration, the maximum temperature of the subsystem for the current iteration, and the difference in the charge rate for the current iteration; the difference in the charge rate for the current iteration is obtained by subtracting the charge rate of the subsystem for previous iterations of the multiple iterations from the charge rate of the subsystem for the current iteration; and the state prediction submodel is configured to output the voltage of the subsystem for the next iteration of the multiple iterations, the charge rate of the subsystem for the next iteration, the maximum temperature of the subsystem for the next iteration, and the difference in the charge rate for the next iteration.

[0010] In some cases, the subsystem voltage for the first iteration of multiple iterations can be determined using the subsystem voltage imbalance for the first iteration by the following equation:

[0011]

number

[0012] Here, CV t ε is the minimum cell voltage of the subsystem's battery cells for the first iteration, ε is the voltage imbalance of the subsystem for the first iteration, and α is the scaling factor. The voltage imbalance is a function of the minimum cell voltage of the subsystem for the first iteration and the maximum cell voltage of the subsystem for the first iteration.

[0013] In some cases, the controller may be configured to repeat steps (1) and (2) until the last iteration of the iterative process is performed.

[0014] In some cases, the controller may be configured to determine the total rechargeable / dischargeable energy of the BESS subsystem over a dynamic time period by summing the rechargeable / dischargeable energy output by the energy prediction submodel for each iteration.

[0015] In some cases, the controller may be configured to repeat steps (1) and (2) until the subsystem voltage for the next iteration of the iterative process output by the state prediction submodel reaches a voltage limit.

[0016] In some cases, the dynamic time period may be determined by the time required to charge or discharge the BESS subsystem from the first iteration of the iterative process to the last iteration of the iterative process, the last iteration of the iterative process being the iteration in which the subsystem voltage output by the state prediction submodel reaches a voltage limit.

[0017] In some cases, the subsystem is selected from at least one of the following: multiple battery packs containing battery cells, or multiple battery racks containing battery packs.

[0018] In some cases, the battery type of the subsystem's battery cells is selected from a group that includes: lithium nickel manganese cobalt (NMC), lithium iron phosphate (LFP), lithium sulfur (Li-S), solid state, nickel cadmium, nickel-metal hydride (NiMH), zinc-air, iron-air, vanadium redox flow, sodium ions, potassium ions, aluminum ions, lead-acid, silicon anode, or combinations thereof.

[0019] In some cases, the architecture of a neural network model is selected from a group consisting of: multilayer perceptrons (MLPs), recurrent neural networks (RNNs), long-term / short-term memory networks, gated recurrent units (GRUs), or combinations thereof.

[0020] In another embodiment, the disclosure relates to a method for determining the rechargeable / dischargeable energy of a subsystem of a battery energy storage system (BESS), wherein the subsystem includes battery cells, and the method includes: a step of performing an iterative process over a dynamic time period, using a neural network model including an energy prediction submodel and a state prediction submodel, wherein the dynamic time period is divided into multiple iterations, and each of the multiple iterations is the following step: (1) inputting into the energy prediction submodel: the voltage of the subsystem for the current iteration of the multiple iterations, the charge rate of the subsystem for the current iteration, and the maximum temperature of the subsystem for the current iteration, where the energy prediction submodel (1) The system outputs the rechargeable / dischargeable energy of the subsystem for the current iteration; and (2) the state prediction submodel inputs: the subsystem voltage for the current iteration, the subsystem charge rate for the current iteration, the subsystem maximum temperature for the current iteration, and the difference in the charge rate for the current iteration; the difference in the charge rate for the current iteration is obtained by subtracting the subsystem charge rate for previous iterations of multiple iterations from the subsystem charge rate for the current iteration; and the state prediction submodel outputs the subsystem voltage for the next iteration of multiple iterations, the subsystem charge rate for the next iteration, the subsystem maximum temperature for the next iteration, and the difference in the charge rate for the next iteration.

[0021] It should be noted that the technical effects obtained through this disclosure are not limited to those described herein, and other effects not mentioned herein can be clearly understood by those skilled in the art from the following description. [Brief explanation of the drawing]

[0022] The accompanying drawings illustrate exemplary embodiments of this disclosure and, together with the following detailed description, serve to further illustrate the technical ideas of this disclosure. However, this disclosure shall not be construed as being limited to the drawings.

[0023] [Figure 1]FIG. 1 is a perspective view schematically showing a configuration of a battery container according to an aspect of the present disclosure.

[0024] [Figure 2] FIG. 2 is a perspective view schematically showing a form in which some constituent elements of the battery container according to an aspect of the present disclosure are separated or moved.

[0025] [Figure 3] FIG. 3 is a view showing an internal configuration of the battery container according to an aspect of the present disclosure, as viewed from above.

[0026] [Figure 4] FIG. 4 is a graph showing a discrepancy between the estimated chargeable / dischargeable energy of a BESS subsystem and the actually discharged energy.

[0027] [Figure 5] FIG. 5 is a conceptual diagram showing the effect of unbalanced voltage levels between battery cells on the chargeability and dischargeability of a BESS subsystem.

[0028] [Figure 6a] FIG. 6 is a conceptual diagram illustrating a BESS subsystem including a power block according to an aspect of the present disclosure.

[0029] [Figure 6b] FIG. 7 is a conceptual diagram illustrating a BESS site including a plurality of power blocks according to an aspect of the present disclosure.

[0030] [Figure 7] FIG. 8 is a graph showing the effect of unbalanced voltage levels between racks on the operating power of a power block.

[0031] [Figure 8a] FIG. 9 is a conceptual diagram illustrating a model for determining the chargeable / dischargeable energy of a BESS subsystem according to an aspect of the present disclosure.

[0032] [Figure 8b] This flowchart illustrates an example of a model for determining the rechargeable / dischargeable energy of a subsystem of a BESS according to the aspects of this disclosure.

[0033] [Figure 9] This is a schematic diagram illustrating a subsystem of a BESS including a power block with a controller that embodies the models of Figures 6a and 6b according to an aspect of the present disclosure. [Modes for carrying out the invention]

[0034] This disclosure may be modified in various ways and may have various embodiments, and the specific embodiments disclosed in detail herein are used to help those skilled in the art to understand this disclosure.

[0035] Therefore, it should be understood that there is no intention to limit this disclosure to any particular form disclosed, but rather, that this disclosure includes all modifications, equivalents, and substitutes that fall within the concept and technical scope of this disclosure.

[0036] In this application, terms such as “includes” or “having” are intended to indicate the presence of features, figures, stages, actions, components, parts, or combinations thereof described in the specification, and it should be understood that such terms do not preclude the presence or possibility of the addition of one or more other features or figures, stages, actions, components, parts, or combinations thereof.

[0037] Figure 1 is a schematic perspective view showing the configuration of the battery container 1000 according to an embodiment of the present disclosure. Figure 2 is a schematic perspective view showing a configuration in which some components of the battery container 1000 according to an embodiment of the present disclosure are separated or movable. Figure 3 is a diagram showing the internal configuration of the battery container 1000 as seen from above, according to an embodiment of the present disclosure.

[0038] Referring to Figures 1 to 3, the battery container 1000 according to this disclosure includes a battery rack 100, a container housing 200, a main connector 300, and a main busbar 400.

[0039] The battery rack 100 may include a plurality of battery modules 110, where each battery module 110 may be configured such that a plurality of battery cells (secondary batteries) are housed in a module case. The battery modules 110 can also be stacked in one direction, for example, vertically, to form the battery rack 100. In particular, the battery rack 100 may include a rack case that enables the stacking of the battery modules 110. In this case, the plurality of battery modules 110 can be housed in individual storage spaces provided in the rack case to form a module stack. In some embodiments, the battery modules 110 may be arranged in other configurations, for example, in a parallel or matrix pattern. The rack case may include features such as cooling channels or structural reinforcement to support the weight of the stacked modules. In some cases, the battery rack 100 may include sensors for monitoring the temperature, voltage, or other parameters of the battery modules 110.

[0040] The battery modules 110 contained within the battery rack 100 may further include control units such as a battery management system (BMS) for each group or a specific group. For example, a separate pack BMS may be provided for each battery module 110. In this case, each battery module 110 may be referred to as a battery pack. That is, the battery rack 100 may be considered to contain multiple battery packs. In the various descriptions below, the battery module 110 may be replaced by a battery pack. In some cases, the battery rack 100 may include sensors for monitoring parameters such as temperature, voltage, or current of the battery modules 110. The BMS for each battery module or pack may communicate with a higher-level rack BMS to coordinate the overall rack performance and safety.

[0041] One or more battery racks 100 may be included in the battery container 1000. In particular, multiple battery racks 100 may be included in the battery container 1000. Furthermore, multiple battery racks 100 may be arranged in at least one direction, for example, horizontally. For example, eight battery racks 100 may be included in the battery container 1000, and the multiple battery racks 100 may be arranged in the left-right direction (X-axis direction) inside the battery container 1000. When multiple battery racks 100 are included, a separate control unit such as a rack BMS may be provided for each battery rack 100. In this case, the rack BMS can be connected to multiple pack BMSs to exchange data and control multiple pack BMSs. On the other hand, when the battery container 1000 includes at least one rack BMS, the rack BMS may be connected to a separate control device provided outside the battery container 1000, for example, a control container. Furthermore, the control container can be connected to the rack BMS or pack BMS of the battery container 1000 to control them or to exchange data with them.

[0042] An empty space may be formed inside the container housing 200. Furthermore, the container housing 200 can accommodate the battery rack 100 within its internal space. More specifically, the container housing 200 may be formed in a substantially rectangular parallelepiped shape, as shown in Figure 1. In this case, the container housing 200 may include an upper housing 201, a lower housing, a front housing 203, a rear housing, a left-side housing 205, and a right-side housing around the internal space. The container housing 200 can also accommodate the battery rack 100 within the internal space defined by these six unit housings.

[0043] The container housing 200 may consist of materials that ensure a certain level of rigidity and stably protect the internal components from external physical and chemical factors. For example, the container housing 200 may consist of or have such metallic materials as steel, aluminum, or titanium. In some embodiments, the container housing 200 may consist of composite materials such as carbon fiber reinforced polymer or fiberglass that provide a high strength-to-weight ratio. The housing may also incorporate corrosion-resistant alloys such as stainless steel or galvanized steel in areas exposed to harsh environmental conditions. In some cases, the container housing 200 may use a combination of materials such as a steel frame with aluminum panels to balance strength, weight, and cost considerations. Furthermore, the housing may include special coatings or treatments such as powder coating or anodizing to improve durability and weather resistance.

[0044] Container housings may be the same size as or similar to the size of shipping containers. They may also conform to shipping container standards predetermined by ISO standards, etc. For example, a container housing may be designed to be the same size as or similar to a 20-foot or 40-foot container. However, the size of the container housing may be appropriately designed depending on the circumstances. In particular, the size or shape of the container housing can be varied depending on the construction scale, shape, topography, etc., of the system to which the battery container is applied, such as an energy storage system. This disclosure may not be limited by the size or shape of the container housing. In some embodiments, for example, the container housing may have different shapes, such as cylindrical, spherical, or customized polygonal shapes. The housing may also be modular, allowing for expansion or contraction based on capacity requirements. In some cases, the container housing may include features such as a sloped roof or reinforced walls for water drainage to increase durability in harsh environments.

[0045] The main connector 300 may be configured to be electrically connected to the outside. That is, in relation to the battery container 1000, the main connector 300 may be configured to connect to other external components of the battery container 1000, such as another battery container 1000 or a control container equipped with a control unit such as a BSC (Battery System Controller).

[0046] The main connector 300 may be located on at least one side of the container housing 200. For example, the main connector 300 may be located on the left or right side of the container housing 200. Furthermore, multiple main connectors 300 may be included in the battery container 1000. For example, as shown in Figures 2 and 3, the main connector 300 may include two main connectors 300, namely a first connector 301 and a second connector 302.

[0047] Multiple main connectors 300 may be located on different sides of the container housing 200. Furthermore, multiple main connectors 300 may be located on both sides of the container housing 200. For example, as shown in Figures 1 to 3, the first connector 301 and the second connector 302 may be provided on the left and right sides of the container housing 200, respectively. In some embodiments, the main connectors 300 may be located on the roof or floor of the container housing 200. In some cases, the main connectors 300 may be located at the corners or edges of the container housing 200. The main connectors 300 may also be arranged in a variety of configurations, for example, in an alternating pattern or aligned vertically along the sides of the container housing 200. In some embodiments, additional main connectors may be included on the front or rear side of the container housing 200 to provide additional connection options.

[0048] The main busbar 400 can be configured to transmit power. In particular, the main busbar 400 can serve as a path through which charging power and discharging power are transmitted to the battery rack 100 contained in the corresponding battery container 1000. For this purpose, the main busbar 400 can be electrically connected to each terminal of the battery module 110 provided to the battery rack 100. The main busbar 400 can also be connected to the main connector 300. Thus, the main busbar 400 can serve as a path through which charging power is transmitted from the main connector 300 to the battery module 110. The main busbar 400 can also serve as a path through which discharging power is transmitted from the battery module 110 to the main connector 300.

[0049] Furthermore, the main busbar 400 can function as a power transmission line between multiple main connectors 300. For this purpose, the differing ends of the main busbar 400 can be connected to different main connectors 300. For example, the main busbar 400 may be a power line that extends long in one direction, for example, in the left-right direction. In this case, both ends of the main busbar 400 can be connected to different main connectors 300, for example, a first connector 301 and a second connector 302. The main busbar 400 can also serve as a path for transmitting power between different main connectors 300, for example, between a first connector 301 and a second connector 302.

[0050] The main busbar 400 may include two unit busbars, namely a positive busbar 410 and a negative busbar 420, to function as a power transmission path. The positive busbar 410 may be connected to the positive terminal of the battery rack 100 or the positive terminal of the battery module 110 contained therein. The negative busbar 420 may also be connected to the negative terminal of the battery rack 100 or the negative terminal of the battery module 110 contained therein.

[0051] Furthermore, the main connector 300 may be provided separately at the respective ends of the positive busbar 410 and the negative busbar 420. For example, the first connector 301 and the second connector 302 may be provided at the left and right ends of the positive busbar 410, respectively. The first connector 301 and the second connector 302 provided at the ends of both ends of the positive busbar 410 may be the positive connector 310. Also, the first connector 301 and the second connector 302 may be provided at the left and right ends of the negative busbar 420, respectively. The two connectors provided at the ends of both ends of the negative busbar 420, namely the first connector 301 and the second connector 302, may both be the negative connector 320.

[0052] Furthermore, the battery container 1000 according to this disclosure may include a cable cover (CC). The cable cover (CC) may be configured to cover cables connected to the battery container 1000. For example, multiple power cables may be connected to a terminal busbar (TB) to transmit power. In this case, the cable cover (CC) may be located at one end of the terminal cover (TC), for example, the lower end, to protect the multiple power cables connected to the terminal busbar (TB). Alternatively, the battery container 1000 may be connected to data cables to exchange various data with other external components, such as a control container 2000. In this case, the cable cover (CC) may be configured to protect the data cables and the like connected to the battery container 1000 from external sources.

[0053] In particular, the cable cover (CC) may include a cable tray (CC1) and a tray cover (CC2). The cable tray (CC1) may include a main body that adheres to the outer wall of the container housing 200 and side walls that protrude outward from the edges of the main body. For example, the side walls may be formed to protrude to the left from the front and rear edges of the main body. The tray cover (CC2) may be combined with the ends of the side walls protruding from the main body of the cable tray (CC1) to form an internal space together with the main body and side walls. In particular, such an internal space may be formed in a hollow shape. This allows cables to be extended outward from the battery container 1000 through the internal space of the cable cover (CC). The cables extended outward may also be connected to other external components such as the control container 2000 or other battery containers 1000.

[0054] In this configuration, the cables extending from the battery container 1000 are protected from external exposure, thereby preventing damage or breakage of the cables. Furthermore, the cable cover (CC) is configured to have a hollow formed at the bottom of the side of the container housing, allowing the cables housed inside to be exposed to the outside below. This can be advantageous for cable installation, management, and undergrounding.

[0055] Furthermore, the battery container 1000 according to this disclosure may further include an air conditioning module 600 as shown in Figures 1 and 2. The air conditioning module 600 may be configured to regulate the air inside the container housing 200. In particular, the air conditioning module 600 can control the temperature state of the air inside. The air conditioning module 600 can also circulate the air inside the container housing 200 and control the temperature of various electronic equipment, such as the battery rack 100 or rack BMS, contained in the battery container 1000 within a certain range. In particular, the air conditioning module 600 can cool the air inside the container housing 200. For example, the air conditioning module 600 may be configured to absorb heat from the air inside the container housing 200 and release the heat to the outside. The air conditioning module 600 may also be configured to remove dust or foreign matter contained in the air inside the container housing 200.

[0056] Typically, the air conditioning module 600 may include at least one HVAC (Heating, Ventilation, and Air Conditioning). For example, the battery container 1000 according to this disclosure may include four HVACs. The HVACs can allow air to circulate inside the container housing 200. In this case, the temperature of the battery racks 100 may be lowered, and the temperature difference between the battery racks 100 or the battery modules 110 contained within the container housing 200 may be reduced.

[0057] In particular, the container housing 200 may include at least one door, such as the one represented by E in Figures 1 and 2, to allow for the installation, maintenance, or repair of the battery rack 100. For example, the container housing 200 may have eight doors E on the front. Also, two doors E may be openable and closable as a pair in the form of a case. In addition, such doors E may be provided on other parts of the container housing 200, for example, on the rear.

[0058] In this manner, when a door E is provided to the container housing 200, an HVAC can be installed in the door E of the container housing 200. For example, when two doors E form a pair, the HVAC can be provided in one of the two doors E. Alternatively, the HVAC, i.e., the air conditioning module 600, can be configured to penetrate the container housing 200, particularly the door E. In this case, one surface of the air conditioning module 600 may be exposed to the outside of the container housing 200, and the other surface of the air conditioning module 600 may be exposed to the inside of the container housing 200. This allows the inner surface of the air conditioning module 600 to come into contact with the air inside the container housing 200 and absorb heat, while the outer surface of the air conditioning module 600 can come into contact with the air outside the container housing 200 and release heat.

[0059] The air conditioning module 600 may be configured to prevent direct contact between internal and external air. That is, the air conditioning module 600 may be configured to prevent internal air from being released to the outside and to prevent external air from flowing into the inside. Therefore, even if the temperature inside the container housing 200 rises, the air conditioning module 600 can absorb only the heat from the internal air and release the heat to the outside without directly releasing the internal air to the outside. In this embodiment, even if a fire or toxic gas occurs inside the battery container 1000, it is possible to prevent the fire or toxic gas from being released to the outside and causing damage to other devices such as other nearby battery containers 1000 or to workers outside.

[0060] Furthermore, the battery container 1000 according to this disclosure may further include a venting module 700 as shown in Figures 1 and 2. The venting module 700 may be configured to release gas from inside the container housing 200 to the outside. The venting module 700 may also allow outside air to flow into the container housing 200. Therefore, the ventilating module 700 can function as a ventilation device. That is, the ventilating module 700 can exchange or circulate gas between the inside and outside of the container housing 200.

[0061] In particular, the ventilation module 700 may be configured to operate in abnormal situations, for example, when ventilation gas or fire occurs in a specific battery module 110. The ventilation module 700 may also be configured to release gas to the outside when gas is generated inside the container housing 200 due to phenomena such as thermal runaway of the battery rack 100. Furthermore, the ventilation module 700 may be configured to be closed under normal conditions and switch to an open state in abnormal conditions such as thermal runaway. In this case, since the ventilation module 700 performs active ventilation, the ventilation module may be referred to as an AVS (Active Ventilation System) or may include such a system.

[0062] In this case, it is possible to prevent more serious problems from occurring, such as an explosion due to increased internal pressure in the battery container 1000. Also, in this case, by rapidly releasing the flammable gas inside the container housing 200 to the outside, it is possible to reduce the likelihood of a fire inside the battery container 1000 or delay the onset of a fire, potentially reducing the scale of the fire.

[0063] On the other hand, in an embodiment that includes both a ventilation module 700 and an air conditioning module 600, under normal circumstances, the ventilation module 700 may not operate, but the air conditioning module 600 may operate. In this case, it is possible to prevent foreign matter or moisture from entering the container housing 200 through the ventilation module 700 during the cooling process. In such an embodiment, since the battery container 1000 includes the air conditioning module 600, the ventilation module 700, etc., the air conditioning module 600 or the ventilation module 700 can be transported and installed together with the battery container 1000 by simply transporting and installing the battery container 1000. Therefore, on-site installation work for the installation of the energy storage system can be minimized, and the connection structure can be simplified.

[0064] In this embodiment, the air conditioning module 600 and / or the ventilation module 700 can operate under the control of the control container 2000. Alternatively, the air conditioning module 600 and / or the ventilation module 700 may be controlled by a control unit contained within the battery container 1000, for example, a rack BMS or other separate control unit that controls the charge / discharge operation of each battery rack 100.

[0065] Furthermore, the battery container 1000 according to this disclosure includes at least one sensor that can provide sensing information to a rack BMS, another separate control unit, or a control container 2000 included in the battery container 1000. For example, a temperature sensor, a smoke sensor, an H2 sensor, and / or a CO sensor may be included in the battery container 1000. In this case, the operation of the air conditioning module 600 and / or the ventilation module 700 may be controlled based on information sensed by such sensors. The battery container 1000 may further include a fire connector 810 for a fire protection module (not shown).

[0066] Figure 4 shows graph 1100 as an example illustrating the discharge of a BESS site identified from 10 MWhr (energy) and 2.5 MW (power) with a discharge time of 4 hours at the beginning of life (BOL). Based on conventional calculations (represented by curve 1102), BESS was estimated to discharge approximately 10 MWhr. However, when the BESS site was completely discharged from a specific power, the actual energy discharged before reaching the power limit was approximately 9 MWhr over 4 hours (represented by curve 1104). Therefore, although BESS should have passed the capacity test, the BESS site failed the capacity test.

[0067] The discrepancy between the estimated discharge energy and the actual discharge energy can be explained by cell voltage imbalance, which is the difference between the maximum and minimum cell voltages across a given BESS subsystem (e.g., pack, module, rack, string, or group of racks). The BESS takes these minimum and maximum cell voltages into account to ensure that batteries are not discharged below the minimum cell voltage or charged above the maximum cell voltage. When discharging, cell voltage imbalance traps energy that would otherwise be discharged (from cells with voltages above the minimum cell voltage), and when charging, cell voltage imbalance traps energy that would otherwise be charged (from cells with voltages below the maximum cell voltage). Cell voltage imbalance can cause problems in balancing supply and demand in the electric grid and may necessitate the use of polluting and inefficient backup power sources such as gas turbines. To resolve cell voltage imbalance, a battery management system (BMS) can passively balance the cells, which can occur when the SOC of a BESS subsystem is below a threshold current drawn above a threshold. For these reasons, continuous use of BESS without rest periods exacerbates cell voltage imbalances.

[0068] Figure 5 illustrates the effect of cell voltage imbalance. An unbalanced group of battery cells 1202 may include cell 1204 with the highest voltage level 1208, cell 1206 with the lowest voltage level 1210, and cells 1203 and 1205 with intermediate voltage levels 1207 and 1209. As shown on the left, when discharging, energy cannot be fully discharged from cells 1203, 1204, and 1205 (represented as hatched areas) due to being limited by cell 1206 with the lowest voltage level 1210. As shown on the right, energy cannot be fully charged from cells 1203, 1205, and 1206 (represented as hatched areas) due to being limited by cell 1204 with the highest voltage level 1208.

[0069] Figure 6a shows a subsystem 1300a of BESS according to an embodiment of the present disclosure. The subsystem 1300a of BESS may include a power block 1301 which includes a plurality of battery racks 1302a, 1302b, and 1302c, which in turn include a plurality of battery packs 1303aa-ar, 1303ba-br, and 1303ca-cr and battery protection units (BPUs) 1310a, 1310b, and 1310c. The racks 1302a-c may include a physical structure having a standardized form (e.g., a steel or aluminum frame) to enable easy installation, management, and expansion. A suitable exemplary battery rack may be, for example, the TR1300 (model ERT5422CN201) manufactured by LG Energy Solutions. In some embodiments, the racks may be constructed of other materials such as carbon fiber composite, fiberglass, or reinforced plastic to reduce weight while maintaining strength. Each of the battery packs 1303aa-ar, 1303ba-br, and 1303ca-cr may contain one or more battery modules and may be connected to monitoring and management electronic devices such as a battery management system (BMS). Each battery module may contain multiple battery cells connected to each other, and the battery cells may be covered and managed as a single unit. A battery cell is the smallest unit of a BESS in which an electrochemical reaction occurs, storing and releasing energy. Cells may have different form factors, such as cylindrical, pouch-type, or prismatic. In each of the battery racks 1302a-c, the battery packs 1303aa-ar, 1303ba-br, and 1303ca-cr may be electrically connected in series with respect to each other, but this disclosure is not limited thereto. The battery racks 1302a-c may be electrically connected in parallel with respect to each other, but this disclosure is not limited thereto. The battery racks 1302a-c and battery packs 1303aa-ar, 1303ba-br, and 1303ca-cr can be connected in any series or parallel arrangement to achieve the target power output.While battery packs and / or modules are described in such specific embodiments, other racks that exclude packs and / or modules are considered within the scope of this disclosure. For example, a rack may contain multiple battery cells without any modules.

[0070] The BPU1310a-c, which may be referred to as a rack BMS, includes electrical and communication interfaces connected to packs 1303aa-ar, 1303ba-br, and 1303ca-cr within each individual rack 1302a, 1302b, and 1302c. The BPU1310a-c may be electrically and communicatively connected to voltage line 1304 and one or more power block controllers 1308.

[0071] Battery racks 1302a-c may be electrically connected to the power grid 1307 via voltage lines 1304. DC switches 1305 may be used, in particular, to disconnect or isolate batteries from other components from the power conversion system (PCS) 1306 or grid-coupled inverter for maintenance, safety, or in case of failure. In case of overvoltage, overcurrent, or other failure of the system, DC switches 1305 can quickly interrupt the current to prevent damage to the power block 1301 or other components. The PCS 1306 manages the conversion between DC power from the power block 1301 and AC power for use by the power grid 1307 (i.e., loads). The PCS 1306 may include both an inverter (DC to AC) and a rectifier (AC to DC), thus enabling bidirectional energy flow between the power block 1301 and the power grid 1307. The PCS1306 synchronizes the output from the power block 1301 with the voltage, frequency, and phase of the grid 1307, enabling the power block 1301 to smoothly inject electricity into the grid 1307 or absorb power from the grid 1307.

[0072] The energy management system (EMS) 1309 can coordinate and optimize the overall energy flow in the BESS subsystem 1300a. The EMS 1309 can handle strategic decisions regarding when and how energy must be stored or discharged, and can integrate multiple energy resources (e.g., solar and wind power jointly located and connected to the microgrid and / or grid 1307). Based on load demand, market signals (e.g., electricity prices such as Location Marginal Price (LMP)), and the availability of renewable power, the EMS 1309 can determine when the BESS subsystem 1300a must store or discharge electricity, and can manage the interaction between the power block 1301 and the grid 1307 to provide services such as frequency regulation, voltage support, and demand response.

[0073] The Power Block Controller (PBC) 1308 can control and operate individual components within the BESS subsystem 1300a, such as the battery packs 1303aa-ar, 1303ba-br, and 1303ca-cr, as well as the battery modules and cells within them, ensuring the safe and efficient operation of the BESS subsystem 1300a at the hardware level. The PBC 1308 works with one or more BMSs to ensure the safe operation of the battery cells and can prevent overcharging, deep discharging, or temperature issues. Working in cooperation with the EMS 1309, the PBC 1308 can manage the conversion of DC power from the power block 1301 to AC power for the grid 307 and vice versa (working with the EMS to adhere to power setpoints) and can adjust the power output from the PBC 1301 to ensure in real time that it meets the voltage and frequency requirements of the grid 1307. The PBC1308 monitors the power block 1301 for defects and, together with the DC switch 1305, can execute protection mechanisms in response to problems such as overvoltage, overcurrent, or overheating. The PBC1308 can also be responsible for executing commands from the EMS1309 at the hardware level.

[0074] Figure 6b shows a BESS site 1300b containing multiple power blocks 1301a-h (e.g., similar to power block 1301) each connected to individual PCS 1307a-h. A single cell in a given power block 1301a-h can reduce the overall capacity of the BESS site 1300b. For example, if a single cell in power block 1301a reaches minimum voltage more quickly than other cells in power block 1301, the overall capacity of the BESS site 1300b could potentially be reduced by about 1 / 8 of its identified capacity, illustrating the importance of managing cell voltage imbalances.

[0075] Figure 7 shows graphs 1400a and 1400b illustrating the effect of unbalanced voltage levels between racks on the operating power of power blocks. As shown in graph 1400a, racks A1, A2, and A3 have nearly identical minimum voltage levels among them. Over time, the operating power of power blocks including racks A1 through A3 is nearly identical at the same point in time. In contrast, as shown in graph 1400b, racks B1, B2, and B3 have differing minimum voltage levels among them. As time progresses, rack B2 forces the power block to reach a specific operating power more quickly, thus limiting the ability of racks B1 and B3 to discharge at a higher operating power at the same point in time.

[0076] Figure 8a is a conceptual diagram illustrating a model for determining the rechargeable / dischargeable energy 1507 of a BESS subsystem according to an aspect of the present disclosure (for example, the BESS subsystem 1300a shown in Figure 6a). The model determines the rechargeable / dischargeable energy at the rated power of the BESS subsystem and is a function of the operating state of the BESS subsystem (voltage, maximum cell temperature, and charge rate).

[0077] The rechargeable / dischargeable energy may be determined individually for each subsystem of a BESS site. The rechargeable / dischargeable energy calculated for each subsystem may be aggregated to determine the total rechargeable / dischargeable energy of the BESS site (or the total rechargeable / dischargeable energy of a group of subsystems). In some cases, the rechargeable / dischargeable energy is calculated for each battery pack in a group of battery packs (battery racks), and the rechargeable / dischargeable energy calculated for each battery pack is aggregated to determine the rechargeable / dischargeable energy of the battery rack. In some cases, the rechargeable / dischargeable energy is calculated for each battery rack in a group of battery racks (power blocks), and the dischargeable energy calculated for each battery rack is aggregated to determine the rechargeable / dischargeable energy of the power block. In some cases, the rechargeable / dischargeable energy is calculated for each power block in a group of power blocks (BESS site), and the rechargeable / dischargeable energy calculated for each power block is aggregated to determine the rechargeable / dischargeable energy of the BESS site.

[0078] The model may include an energy prediction submodel 1501 and a state prediction submodel 1502. One or both of submodels 1501 or 1502 may be implemented as a neural network model performing machine learning. The energy prediction submodel 1501 predicts the chargeable / dischargeable energy for the current iteration, and the state prediction submodel 1502 predicts the operating state (e.g., voltage, charge rate, and maximum cell temperature) for the next iteration. The model may run as an iterative process over a dynamic time period (i.e., a cycle) divided into multiple iterations (i.e., time intervals). Time can be minutes, hours, days, months, etc. Each iteration may be defined or composed of seconds, minutes, hours, etc. (fewer iterations may produce more accurate results). In some cases, each iteration may be defined or composed of any value in the range of 1 minute to 15 minutes. In one example, the dynamic time period is determined to be 24 hours, and each iteration is defined or composed of 2 minutes. Thus, the rechargeable / dischargeable energy of the BESS subsystem can be calculated every two minutes, and the total rechargeable / dischargeable energy of the BESS subsystem can be calculated by summing the respective dischargeable energies calculated in each two-minute iteration.

[0079] As shown in flowchart 1500 in Figure 8b, the model can be realized by inputting a feature unit 1503 (i.e., the operating state of the BESS subsystem) into the energy prediction submodel 1501 and the state prediction submodel 1502. The feature unit 1503 may include the subsystem voltage for the current iteration (i.e., operating voltage), the subsystem charge rate for the current iteration, and the subsystem's maximum temperature for the current iteration (e.g., the highest temperature of the battery cells). In the case of the state prediction submodel 1502, the difference in charge rates (the subsystem's charge rate for the current iteration minus the subsystem's charge rate for the previous iteration) is additionally input.

[0080] The energy prediction submodel 1501 outputs the rechargeable / dischargeable energy 1504 of the subsystem for the current iteration. The state prediction submodel 1502 outputs the operating state 1506 for the next iteration (which becomes the input feature 1503 for the next iteration), which may include the subsystem voltage for the next iteration, the subsystem charge rate for the next iteration, the subsystem's maximum temperature for the next iteration, and the difference in charge rate for the next iteration. After inputting the initial voltage, initial charge rate, and initial maximum temperature for the first iteration of the time period, the model can predict the rechargeable / dischargeable energy 1504 and the next operating state 1506 for the remainder of multiple iterations of the time period.

[0081] The initial voltage of the subsystem for the first iteration of a time period can be determined using the subsystem's voltage imbalance for the first iteration by the following equation:

[0082]

number

[0083] Here, CV t ε is the minimum cell voltage of the subsystem's battery cells for the first iteration, ε is the voltage imbalance of the subsystem for the first iteration, and α is the scaling factor. The voltage imbalance is a function (i.e., difference) of the minimum cell voltage of the subsystem's battery cells for the first iteration and the maximum cell voltage of the subsystem's battery cells for the first iteration. The minimum and maximum cell voltages can be determined (i.e., measured) using one or more voltage sensors (which can be integrated with the BMS). The initial charge rate of the subsystem for the first iteration of a period of time can be determined by the following equation: [Formula 2] Charging rate = (Rated power / Rated energy capacity) The initial maximum temperature for the first iteration of a period of time can be determined (i.e., measured) using one or more temperature sensors (which may be integrated with the BMS).

[0084] The energy prediction submodel 1501 and the state prediction submodel 1502 may run until the final iteration of the iterative process (in other words, the last iteration of the dynamic time period), where the decision may be made at decision point 1505. In some cases, the energy prediction submodel 1501 and the state prediction submodel 1502 may run until the state prediction submodel 1502 outputs the maximum operating voltage (i.e., voltage limit), which may be made at decision point 1505. The maximum operating voltage may depend on the cell type and may be, for example, about 1V to about 10V. In some cases, the maximum operating voltage may be about 2.5V to 4V (e.g., about 3.2V). In some cases, the dynamic time period may be determined by the time required to charge or discharge the subsystem of the BESS from the first iteration of the iterative process to the last iteration of the iterative process, where the last iteration of the iterative process is the iteration in which the voltage of the subsystem output by the state prediction submodel 1502 reaches maximum operation.

[0085] The total rechargeable / dischargeable energy 1507 of the BESS subsystem over a dynamic time period can be determined by summing the rechargeable / dischargeable energy 1504 output by the energy prediction submodel 1501 for each iteration. For a BESS site, the overall rechargeable / dischargeable energy of the BESS site over a dynamic time period can be determined by summing the total dischargeable energy 1507 determined for each power block (group of battery racks).

[0086] In some cases, the battery type (i.e., chemical) of the battery cells in a BESS subsystem may be one or more of the following: lithium-ion, lithium nickel-manganese-cobalt (NMC), lithium iron phosphate (LFP), lithium sulfur (Li-S), solid state, nickel-cadmium, nickel-metal hydride (NiMH), zinc-air, iron-air, vanadium redox flow, sodium ion (e.g., sodium sulfur (NaS), sodium nickel chloride), potassium ion, aluminum ion, lead-acid, silicon anode, or a combination thereof, but this disclosure is not limited to these. Other battery types may include lithium titanate (LTO), lithium-air, magnesium-ion, calcium-ion, organic radical batteries, zinc-brom, or flow batteries such as iron-chromium, sodium-air, lithium-organic, and solid polymer electrolyte batteries. In some embodiments, emerging technologies such as graphene-based batteries, glass batteries, or sand batteries may also be utilized in a BESS subsystem.

[0087] The architecture of neural network models may be one or more of the following: multilayer perceptron (MLP), recurrent neural network (RNN), long-term short-term memory network, gated recurrent unit (GRU), convolutional neural network (CNN), transformer network, deep belief network (DBN), radial basis function network (RBFN), echo state network (ESN), Hopfield network, Boltzmann machine, restricted Boltzmann machine (RBM), deep autoencoder, variational autoencoder (VAE), generative adversarial network (GAN), or a combination thereof, but this disclosure is not limited to these. In some cases, the model architecture may be a tree-based model such as a decision tree, random forest, gradient boosting machine (GBM), or a combination thereof, rather than a neural network, but this disclosure is not limited to these.

[0088] An MLP model can consist of multiple layers, including an input layer, one or more hidden layers, and an output layer. Each layer can be fully connected, meaning that each neuron in one layer is connected to all neurons in subsequent layers through weighted connections. The network can operate by processing input data through these layers, applying activation functions at each neuron to introduce nonlinearity, and generating an output at the final layer. The weights of the connections can be adjusted during training through a process known as backpropagation, which minimizes the difference between the predicted output and the actual target. By learning to map inputs to outputs based on the provided training data, an MLP model can perform complex tasks such as classification or regression. In some embodiments, an MLP model can integrate additional techniques such as dropout or batch normalization to improve generalization and training stability. The architecture of an MLP can be customized with a diverse number and size of hidden layers to suit the complexity of different problems. In some cases, an MLP model can be combined with other neural network types in an ensemble manner to further improve predictive performance.

[0089] RNN models can process sequences of data by maintaining connections between neurons across different time stages, allowing the network to retain information from previous inputs. This processing can be achieved through the use of circular connections, where the output from one time stage is fed back to the network as input for the next time stage, enabling the network to capture the time dependency of the data. The network can be trained using backpropagation through time (BPTT), which adjusts weights to minimize the error between the predicted output and the actual output over the entire sequence. In some embodiments, RNN models can integrate gating mechanisms to better handle long-term dependencies. These gating mechanisms can allow the network to selectively update, forget, or output information at each time stage. Furthermore, RNN models can be stacked with numerous layers to generate deep RNN architectures, which can potentially increase their capacity to model complex time patterns. In some cases, attention mechanisms can be integrated into RNN models to allow the network to focus on differing parts of the input sequence when making predictions, which can be particularly useful for tasks involving long sequences or when certain parts of the input are more relevant than others.

[0090] LSTM network models are a type of RNN model that solves the limitations of traditional RNN models in capturing long-range dependencies from sequential data. LSTM network models can integrate unique cell structures, including memory cells and gating mechanisms, i.e., input, output, and forget gates. These gates regulate the flow of information, allowing the network to selectively retain or forget data over time, thus preserving important information over longer sequences and avoiding problems such as gradient vanishing or exploding. LSTM networks can be particularly effective for working with time-series data or natural language processing. In some cases, LSTM models can be stacked or combined with other neural network architectures to produce more complex and powerful models for specific applications.

[0091] GRU models are similar to LSTM models but can have a simpler architecture and may have non-computational requirements while still being effective for sequential data prediction. GRU models can process sequential data by integrating a gating mechanism that regulates the flow of information through the network. The GRU structure features two main gates: an update gate that determines the amount of historical information to retain, and a reset gate that controls the impact of new input data. These gates allow the GRU to dynamically maintain or modify its internal state, enabling the network to effectively capture time dependencies without the complexity of separate input and forget gates found in other architectures. In some embodiments, GRU models may be particularly well-suited for tasks involving shorter sequences or high-priority tasks where computational efficiency is a priority. Additionally, GRU models may be easier to train and converge more quickly than LSTM models for certain applications, making them a widely used choice for modeling diverse sequences in natural language processing and time series analysis.

[0092] Figure 9 is a schematic diagram illustrating a subsystem 1300a of a BESS that includes a power block 1301, which includes a controller 1600 that embodies the neural network model described in relation to Figures 8a and 8b, according to an aspect of the present disclosure.

[0093] The controller 1600 may include one or more processors 1602 (i.e., processing modules) configured to execute program instructions maintained on memory 1604 (i.e., memory modules(s)). In this regard, one or more processors 1602 of the controller 1600 may perform any of the various methods, processes, steps, or algorithms described throughout this disclosure, including, for example, the embodiment of a neural network model including inputs of feature units 1503 to energy prediction submodels 1501 and state prediction submodels 1502, the determination of rechargeable / dischargeable energy 1504 for each iteration of a dynamic time period, the output of an operating state 1506 for the next iteration by the state prediction submodel 1502, the determination of whether to terminate the execution of submodels 1501 and 1502 at decision point 1505, and the determination of total rechargeable / dischargeable energy 1507.

[0094] Additionally, controller 1600 may be configured to receive data including, but not limited to, historical data on the operating state of the BESS subsystem (from power block controllers 1308 and / or EMS 1309). Submodels 1501 and 1502 may be trained using historical data of a BESS subsystem having a specific cell type (i.e., a chemical such as Li NMC or LFP). The historical data may include voltage, charge rate, and maximum cell temperature for the BESS subsystem over a period of time (e.g., several months). The historical data may be measured using sensors (e.g., voltage and temperature sensors). Submodels 1501 and 1502 may be trained using historical data 1605 derived from the same BESS site where the rechargeable / dischargeable energy is determined. In some cases, submodels 1501 and 1502 may be trained using historical data 1606 derived from different BESS sites where the rechargeable / dischargeable energy is determined. In some cases, submodels 1501 and 1502 may be trained using a combination of historical data 1605 derived from the same BESS site and historical data 1606 derived from different BESS sites. In some cases, controller 1600 may receive initial operating conditions 1607 of the BESS subsystem 1300a, such as initial minimum cell voltage, initial maximum cell voltage, initial cell voltage imbalance, initial operating voltage, initial charge rate, and initial maximum temperature, but the disclosure is not limited to these. In some cases, controller 1600 may provide data of total rechargeable / dischargeable energy 1507 to the EMS 1309 and / or power block controller 1308, which can then be used to determine when and how much energy to discharge.

[0095] The controller 1600 (i.e., computing device) may include a desktop computer, a mainframe computer system, a workstation, an image computer, a parallel processor, or any other computer system (e.g., a networked computer). One or more processors 1602 of the controller 1600 may include any processing elements known to the art. In this sense, one or more processors 1602 may include any microprocessor-type device configured to execute algorithms and / or instructions, such as an ASIC (application-specific integrated circuit), an FPGA (field-programmable gate array), a parallel processor, a GPU (graphics processing unit), a CPU (central processing unit), other chipsets, logic circuits, and / or electronic processors. It is further recognized that the term “processor” may be broadly defined to include any device having one or more processing elements that execute program instructions from non-temporary memory 1604. Also, the stages described throughout this disclosure may be performed by a single controller 1600 or, alternatively, by a number of controllers. For example, the power block controller 1308, EMS 1309, PCS 1306, and controller 1600 may be the same controller or multiple controllers. Additionally, controller 1600 may include one or more controllers housed in a common housing or multiple housings. In this manner, any controller or combination of controllers can be individually packaged as modules suitable for integration into the BESS subsystem 1300a.

[0096] Memory 1604 may include any storage medium known in the industry that is suitable for storing program instructions executable by one or more associated processors 1602. For example, memory 1604 may include non-temporary memory media. Other examples include, but are not limited to, read-only memory, random-access memory, magnetic or optical memory devices (e.g., disks), magnetic tapes, solid-state drives, etc. It should be added that memory 1604 may be housed in a common control housing together with the processors 1602. In some cases, memory 1604 may be located remotely in relation to the physical locations of the processors 1602 and the control 1600. For example, one or more processors 1602 of the control 1600 may have access to remote memory (e.g., a server or cloud) accessible via a network (e.g., the internet, an intranet, etc.).

[0097] As stated above, the present disclosure has been described in more detail through the drawings and embodiments. However, the embodiments of the specification or the configurations described in the drawings are merely embodiments of the present disclosure and do not represent the entire technical concept of the present disclosure. Therefore, when filing this application, it should be understood that there may be a variety of equivalents and variations included in the claims instead.

Claims

1. A system for determining the total rechargeable / dischargeable energy of a subsystem of a battery energy storage system (BESS), wherein the subsystem includes battery cells, and the system: A controller including one or more processing modules and one or more non-temporary memory storage modules for storing computing instructions, wherein the computing instructions are executed by the one or more processing modules, Configured to perform an iterative process over a dynamic time period, and using a neural network model including an energy prediction submodel and a state prediction submodel, the dynamic time period is divided into multiple iterations, and for each of the multiple iterations, the controller: (1) Input to the energy prediction submodel: the voltage of the subsystem for the current iteration of the plurality of iterations, the charge rate of the subsystem for the current iteration, and the maximum temperature of the subsystem for the current iteration, wherein the energy prediction submodel is configured to output the rechargeable / dischargeable energy of the subsystem for the current iteration. (2) The state prediction submodel is configured to receive the following inputs: the voltage of the subsystem for the current iteration, the charge rate of the subsystem for the current iteration, the maximum temperature of the subsystem for the current iteration, and the difference in charge rates for the current iteration. The difference in the charging rate for the current iteration is obtained by subtracting the charging rate of the subsystem for previous iterations of the plurality of iterations from the charging rate of the subsystem for the current iteration. The state prediction submodel is configured to output the voltage of the subsystem for the next iteration of the plurality of iterations, the charge rate of the subsystem for the next iteration, the maximum temperature of the subsystem for the next iteration, and the difference in the charge rate for the next iteration of the system.

2. The voltage of the subsystem for the first iteration of the plurality of iterations is determined using the voltage imbalance of the subsystem for the first iteration by the following equation: [Math 1] CVmin ｔ ε is the minimum cell voltage of the battery cell of the subsystem for the first iteration, ε is the voltage imbalance of the subsystem for the first iteration, and α is the scaling factor. The system according to claim 1, wherein the voltage imbalance is a function of the minimum cell voltage of the subsystem for the first iteration and the maximum cell voltage of the subsystem for the first iteration.

3. The aforementioned controller is: The system according to claim 1, configured to repeat steps (1) and (2) until the last iteration of the iterative process is performed.

4. The aforementioned controller is: The system according to claim 3, configured to determine the total rechargeable / dischargeable energy of the subsystem of BESS over the dynamic time period by summing the rechargeable / dischargeable energy output by the energy prediction submodel for each iteration.

5. The aforementioned controller is: The system according to claim 1, configured to repeat steps (1) and (2) until the voltage of the subsystem for the next iteration of the iterative process output by the state prediction submodel reaches a voltage limit.

6. The aforementioned controller is: The system according to claim 5, configured to determine the total rechargeable / dischargeable energy of the subsystem of BESS over the dynamic time period by summing the rechargeable / dischargeable energy output by the energy prediction submodel for each iteration.

7. The system according to claim 5, wherein the dynamic time period is determined by the time required to charge or discharge the subsystem of BESS from the first iteration of the iterative process to the last iteration of the iterative process, the last iteration of the iterative process being the iteration in which the voltage of the subsystem output by the state prediction submodel reaches the voltage limit.

8. The subsystem is selected from at least one of a plurality of battery packs including the battery cells, or a plurality of battery racks including the battery packs, according to claim 1.

9. The system according to claim 1, wherein the battery type of the battery cell of the subsystem is selected from the group including: lithium nickel manganese cobalt (NMC), lithium iron phosphate (LFP), lithium sulfur (Li-S), solid state, nickel cadmium, nickel-metal hydride (NiMH), zinc-air, iron-air, vanadium redox flow, sodium ions, potassium ions, aluminum ions, lead-acid, silicon anode, or combinations thereof.

10. The system according to claim 1, wherein the architecture of the neural network model is selected from a group consisting of: a multilayer perceptron (MLP), a recurrent neural network (RNN), a long-term / short-term memory network, a gated recurrent unit (GRU), or a combination thereof.

11. A method for determining the total rechargeable / dischargeable energy of a subsystem of a battery energy storage system (BESS), wherein the subsystem includes battery cells, and the method is: The process includes a step of performing an iterative process over a dynamic time period, and by using a neural network model that includes an energy prediction submodel and a state prediction submodel, the dynamic time period is divided into multiple iterations, each of which iterations is the following step: (1) The energy prediction submodel is input with the following: the voltage of the subsystem for the current iteration of the plurality of iterations, the charge rate of the subsystem for the current iteration, and the maximum temperature of the subsystem for the current iteration, wherein the energy prediction submodel outputs the rechargeable / dischargeable energy of the subsystem for the current iteration. (2) The state prediction submodel includes the step of inputting: the voltage of the subsystem for the current iteration, the charge rate of the subsystem for the current iteration, the maximum temperature of the subsystem for the current iteration, and the difference in charge rates for the current iteration, The difference in the charging rate for the current iteration is obtained by subtracting the charging rate of the subsystem for previous iterations of the plurality of iterations from the charging rate of the subsystem for the current iteration. A method for which the state prediction submodel outputs the voltage of the subsystem for the next iteration of the plurality of iterations, the charge rate of the subsystem for the next iteration, the maximum temperature of the subsystem for the next iteration, and the difference in the charge rate for the next iteration.

12. The voltage of the subsystem for the first iteration of the plurality of iterations is determined using the voltage imbalance of the subsystem for the first iteration by the following equation: [Math 1] CVmin ｔ ε is the minimum cell voltage of the battery cell of the subsystem for the first iteration, ε is the voltage imbalance of the subsystem for the first iteration, and α is the scaling factor. The method according to claim 11, wherein the voltage imbalance is a function of the minimum cell voltage of the subsystem for the first iteration and the maximum cell voltage of the subsystem for the first iteration.

13. The method according to claim 11, further comprising the step of repeating steps (1) and (2) until the last iteration of the iterative process is performed.

14. The method according to claim 13, further comprising the step of determining the total rechargeable / dischargeable energy of the subsystem of BESS over the dynamic time period by summing the rechargeable / dischargeable energy output by the energy prediction submodel for each iteration.

15. The method according to claim 11, further comprising the step of repeating steps (1) and (2) until the voltage of the subsystem for the next iteration of the iterative process output by the state prediction submodel reaches a voltage limit.

16. The method according to claim 15, further comprising the step of determining the total rechargeable / dischargeable energy of the subsystem of BESS over the dynamic time period by summing the rechargeable / dischargeable energy output by the energy prediction submodel for each iteration.

17. The method according to claim 15, wherein the dynamic time period is determined by the time required to charge or discharge the subsystem of BESS from the first iteration of the iterative process to the last iteration of the iterative process, the last iteration of the iterative process being the iteration in which the voltage of the subsystem output by the state prediction submodel reaches the voltage limit.

18. The method according to claim 11, wherein the subsystem is selected from at least one of a plurality of battery packs including the battery cells, or a plurality of battery racks including the battery packs.

19. The method according to claim 11, wherein the battery type of the battery cell of the subsystem is selected from the group consisting of: lithium nickel manganese cobalt (NMC), lithium iron phosphate (LFP), lithium sulfur (Li-S), solid state, nickel cadmium, nickel-metal hydride (NiMH), zinc-air, iron-air, vanadium redox flow, sodium ions, potassium ions, aluminum ions, lead-acid, silicon anode, or combinations thereof.

20. The method according to claim 11, wherein the architecture of the neural network model is selected from the group including: multilayer perceptrons (MLPs), recurrent neural networks (RNNs), long-term and short-term memory networks, gated recurrent units (GRUs), or combinations thereof.

Images