A battery energy storage plant thermal management method, system, computing device, and medium

Abstract

The application discloses a battery energy storage power station thermal management method and system, a computing device and a medium, relates to the technical field of battery cooling control, and solves the problems of untimely control and large energy consumption in the traditional mode. The method is applied to a battery energy storage power station, and specifically comprises the following steps: obtaining temperature data of the battery; based on the temperature data, a first control strategy or a second control strategy is selected to control a cooling unit to cool the battery; the first control strategy is based on an MPC algorithm for control, and the second control strategy is based on the MPC algorithm and an optimal expected cost of the MPC algorithm for control, wherein the optimal expected cost is calculated based on an SDP algorithm. The application calculates the optimal expected cost of the MPC algorithm by the SDP algorithm as a constraint, so that when the controller cools the battery by using the second control strategy, the energy consumption required for temperature control can be reduced on the basis of ensuring the temperature control capability, and the economy of the battery energy storage power station is greatly improved.

Description

Technical Field

[0001] This invention relates to the field of thermal management technology for battery energy storage power stations, and in particular to a thermal management method, system, computing device, and computer-readable storage medium for battery energy storage power stations. Background Technology

[0002] With the development of energy storage power stations in recent years, battery performance and safety have become increasingly important. Batteries are widely used in energy storage power stations, but they are susceptible to damage at extremely high and low temperatures. If the temperature is not within the appropriate range, battery performance and efficiency will decrease, and batteries may even explode at high temperatures. Therefore, regulating the battery temperature within the required range is crucial. While ensuring safety, the battery cooling unit consumes a significant amount of energy. Current research mainly focuses on optimizing the power management of energy storage power stations, but research on the optimal control of thermal management systems is limited. Therefore, the cooling unit should first ensure the safe and stable operation of the battery. Furthermore, with the development of energy storage power stations, the economics of energy storage power stations should also be considered in the thermal management system, achieving cost reduction and efficiency improvement while ensuring safe and stable operation.

[0003] Therefore, the controller of the thermal management system needs to be optimized. This system has multiple objectives, such as regulating battery temperature and minimizing energy consumption, which involves a trade-off. Current technology mainly relies on intuition and empirical observation to solve this problem. Some methods currently use simple rule-based or fuzzy strategy-based approaches. A typical strategy is thermostat control, which activates cooling when the temperature is too high and deactivates it when the temperature returns to normal. This design is simple but cannot guarantee timely temperature control and results in significant energy loss.

[0004] Therefore, there is an urgent need for a thermal management method, system, computing device, and computer-readable storage medium for battery energy storage power stations. Summary of the Invention

[0005] To address the problems of insufficient control timeliness and high energy loss in existing technologies, this invention provides a thermal management method, system, computing device, and computer-readable storage medium for battery energy storage power stations. This method reduces the energy consumption required for temperature control while maintaining temperature control capabilities, significantly improving the economic efficiency of battery energy storage power stations. The specific technical solution is as follows:

[0006] In a first aspect, embodiments of this application provide a thermal management method for a battery energy storage power station, applied to a thermal management system for a battery energy storage power station. The system includes a controller and a cooling unit, wherein the controller controls the cooling unit to cool the batteries in the battery energy storage power station; the method includes:

[0007] The controller acquires the temperature data of the battery; if the temperature data is greater than or equal to a preset temperature threshold, the controller controls the cooling unit to cool the battery based on a first control strategy; if the temperature data is less than the temperature threshold, the controller controls the cooling unit to cool the battery based on a second control strategy.

[0008] The first control strategy is a control strategy based on the model predictive control (MPC) algorithm; the second control strategy is a control strategy based on the MPC algorithm and the optimal expected cost of the MPC algorithm. The optimal expected cost is calculated based on the stochastic dynamic programming (SDP) algorithm and serves as a constraint of the MPC algorithm.

[0009] Preferably, the controller controls the cooling unit to cool the battery based on a first control strategy, including: the controller acquiring the dynamic expression equation and cost function of the model of the cooling unit; the controller calculating the approximate operating cost of the cooling unit based on the dynamic expression equation and cost function; the controller acquiring the average heat generation of the battery; the controller calculating the predicted heat generation of the cooling unit within a predicted time range based on the average heat generation and a moving average filter algorithm; the controller determining a first optimal control input set based on the approximate operating cost and the predicted heat generation; and the controller controlling the cooling unit to cool the battery based on the first optimal control input set.

[0010] Preferably, the controller controls the cooling unit to cool the battery based on a second control strategy, including: the controller obtaining the cost function of the model of the cooling unit and obtaining the optimal expected cost; the controller establishing a random variable based on a Markov process, the random variable being used to indicate the heat generation of the cooling unit; the controller determining a second optimal control input set based on the random variable, the optimal expected cost, and the cost function; and the controller controlling the cooling unit to cool the battery based on the second optimal control input set.

[0011] Preferably, the optimal expected cost is preset in the controller.

[0012] Preferably, obtaining the optimal expected cost includes: constructing an optimal expected cost function; obtaining the corresponding infinite-term stochastic problem equation based on the optimal expected cost function; obtaining the random variable; and solving the infinite-term stochastic problem equation based on a preset first recursive equation, a second recursive equation, and the random variable to obtain the optimal expected cost, wherein the first recursive equation is used to iterate the optimal expected cost under a given strategy, and the second recursive equation is used to iterate the optimal strategy.

[0013] Preferably, after the controller acquires the temperature data of the battery, the method further includes: the controller establishing a model based on the cooling unit; and the controller constructing a dynamic expression equation and cost function for the model of the cooling unit.

[0014] Secondly, embodiments of this application provide a thermal management system for a battery energy storage power station. This system includes a controller and a cooling unit. The controller controls the cooling unit to cool the batteries in the battery energy storage power station. Specifically,

[0015] The controller is used to acquire the temperature data of the battery; when the temperature data is greater than or equal to a preset temperature threshold, the controller is also used to control the cooling unit to cool the battery based on a first control strategy; when the temperature data is less than the temperature threshold, the controller is also used to control the cooling unit to cool the battery based on a second control strategy.

[0016] The first control strategy is a control strategy based on the Model Predictive Control (MPC) algorithm; the second control strategy is a control strategy based on the MPC algorithm and the optimal expected cost of the MPC algorithm, wherein the optimal expected cost is calculated based on the Stochastic Dynamic Programming (SDP) algorithm and serves as a constraint of the MPC algorithm.

[0017] Preferably, the optimal expected cost is preset in the controller.

[0018] Thirdly, embodiments of this application provide a computing device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in the first aspect above.

[0019] Fourthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the method described in the first aspect above.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows: The technology of this application, through the first control strategy based on the MPC algorithm, has learning and predictive capabilities compared with the traditional passive temperature control strategy. Its temperature control response speed is faster and its stability is better, which can effectively avoid thermal runaway events of the battery. By using the optimal expected cost of the MPC algorithm calculated by the SDP algorithm as a constraint, the controller can reduce the energy consumption required for temperature control while ensuring temperature control capability when using the second control strategy to cool the battery, which greatly improves the economics of the battery energy storage power station. By adopting the first control strategy with control priority or the second control strategy with cost priority under different temperature conditions, the safety and economy of the battery energy storage power station can be balanced. Attached Figure Description

[0021] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0022] Figure 1 A schematic flowchart illustrating a thermal management method for a battery energy storage power station provided in an embodiment of this application;

[0023] Figure 2 This is a schematic diagram of the structure of a cooling unit provided in an embodiment of this application;

[0024] Figure 3 A schematic diagram of the MPC control process under the first control strategy provided in the embodiments of this application;

[0025] Figure 4 A flowchart of the iterative algorithm for the optimal expected cost and optimal strategy provided in the embodiments of this application;

[0026] Figure 5 This is a schematic diagram of the MPC control process under the second control strategy provided in the embodiments of this application;

[0027] Figure 6 A set of simulation data comparison charts under different control strategies are provided for embodiments of this application;

[0028] Figure 7 This is a schematic diagram of the structure of a battery energy storage power station provided in an embodiment of this application;

[0029] Figure 8 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0032] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0033] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0034] To address the issues of poor timeliness and high energy consumption in traditional thermal management methods for battery energy storage power stations, this application provides a thermal management method, system, computing device, and computer-readable storage medium for battery energy storage power stations. This method can reduce the energy consumption required for temperature control while ensuring temperature control capabilities, thereby significantly improving the economic efficiency of battery energy storage power stations.

[0035] Please see Figure 1 , Figure 1 This application provides a flowchart illustrating a thermal management method for a battery energy storage power station. The method is applied to a thermal management system for the battery energy storage power station, which includes a controller and a cooling unit. The controller controls the cooling unit to cool the batteries in the battery energy storage power station. The method specifically includes the following steps:

[0036] Step 101: The controller acquires the temperature data of the battery.

[0037] The system also includes a temperature sensor located near or in direct contact with the battery, and the controller is connected to the temperature sensor; the controller can obtain the battery temperature data based on the temperature sensor.

[0038] Preferably, the temperature sensor can also be installed in the coolant channel of the cooling unit to collect the coolant temperature. Based on the battery temperature data and the coolant temperature data, the controller controls the speed of the air conditioning compressor and the speed of the cooling water pump in the cooling unit to cool the battery.

[0039] Specifically, in the formulas of this document, battery temperature data is represented by T. batt The coolant temperature is indicated by T. clnt The compressor speed is represented by ω. comp The rotational speed of the cooling water pump is expressed in ω. pump This indicates that the unit of rotational speed is revolutions per minute (rpm).

[0040] Preferably, after the controller acquires the temperature data of the battery, the method further includes: the controller establishing a model based on the cooling unit; and the controller constructing a dynamic expression equation and cost function for the model of the cooling unit.

[0041] It should be noted that the model for this cooling unit was constructed to achieve temperature control in battery energy storage power stations. By simulating and predicting the thermal behavior of the battery under different operating conditions, the model can dynamically adjust the operating parameters of the cooling unit based on the real-time operating status of the battery and the ambient temperature to achieve optimal cooling performance. Furthermore, the model can also predict temperature change trends over a future period based on the battery's charge / discharge status and historical temperature data, thereby making corresponding adjustments to the cooling strategy in advance to prevent battery overheating or thermal runaway. In this way, the thermal management system of the battery energy storage power station can achieve more intelligent and efficient temperature control, ensuring the safe and stable operation of the battery.

[0042] It should be noted that the cooling unit model (hereinafter referred to as the cooling unit model) is designed to control the temperature of battery modules in a battery energy storage power station. This model focuses on the thermal characteristics of the battery modules under different charge-discharge cycles. By accurately simulating the thermal behavior of the battery modules, the cooling unit model can monitor the temperature changes of the battery modules in real time and dynamically adjust the operating parameters of the cooling unit based on the temperature data. The establishment of this model enables the battery energy storage power station to more accurately control the temperature of the battery modules, thereby effectively avoiding battery performance degradation or damage caused by excessive temperature, and improving the overall operating efficiency and safety of the power station.

[0043] Optionally, the cooling unit model can also be designed to include an ambient temperature model, which is used to simulate and predict temperature changes in the environment where the battery energy storage power station is located. By integrating the ambient temperature model, the thermal management system can more comprehensively consider the impact of external factors on battery temperature, thereby making the cooling strategy more accurate and efficient. For example, the ambient temperature model can predict temperature changes at different times of the day, helping the cooling unit to adjust its operating status in advance to adapt to fluctuations in ambient temperature and ensure that the battery operates within its optimal temperature range.

[0044] Optionally, the cooling unit model can also integrate a battery health status monitoring module. This module is responsible for monitoring the battery's health status in real time, including key parameters such as battery aging level and internal resistance changes. By analyzing these parameters, the thermal management system can adjust the cooling strategy in a timely manner to prevent thermal management problems caused by battery aging and extend the battery's lifespan.

[0045] Optionally, the cooling unit model can also integrate a fault diagnosis module, which can detect potential faults and anomalies in the battery energy storage power station in real time. Through advanced diagnostic algorithms, the fault diagnosis module can quickly identify the source of the problem and provide corresponding handling suggestions or automatically execute fault handling procedures, thereby reducing downtime and improving system reliability and safety. Furthermore, the fault diagnosis module can record historical fault data, providing a reference for future maintenance and upgrades.

[0046] It should be noted that this cooling unit model includes various models used in battery cooling units, which can be designed using different design methods.

[0047] Optionally, the controller can establish a twin model, a data-driven modeling approach that constructs a virtual model highly similar to the actual system behavior by collecting operational data from the actual battery cooling unit. This model can simulate the dynamic response of the battery cooling unit in real time, providing accurate prediction and control basis for the thermal management system. Machine learning algorithms, such as neural networks, can be used to train the twin model, enabling it to learn and mimic the complex behavior of the battery cooling unit. In this way, the twin model can accurately predict battery temperature changes and provide effective cooling strategies in actual operation to cope with various operating conditions and environmental changes. Furthermore, the twin model can also be used for fault prediction and performance optimization. By analyzing the differences between the model output and the actual system, potential problems can be identified and adjusted in a timely manner, ensuring the efficient and safe operation of the battery energy storage power station.

[0048] Optionally, the controller can also represent the refrigeration and battery cooling processes using relevant mathematical expressions. These mathematical expressions can be equations based on thermodynamics and fluid mechanics principles, describing the flow and heat transfer processes of the cooling medium in the refrigeration and battery cooling circuits. Through precise mathematical models, the performance of the cooling unit can be quantitatively analyzed, thereby optimizing the cooling strategy and ensuring optimal temperature control of the battery under different operating conditions. Furthermore, mathematical models can be used to simulate and predict the thermal behavior of the battery under extreme operating conditions, providing theoretical support for the safe operation of the battery energy storage power station. In practical applications, these mathematical models can be combined with actual sensor data to adjust the operating parameters of the cooling unit in real time to adapt to changes in battery operating status and ambient temperature. In this way, the thermal management system of the battery energy storage power station can achieve more precise and efficient temperature control, further improving the system's stability and reliability.

[0049] Optionally, the controller can also establish a physics-based simulation model that can simulate the thermodynamic behavior of the battery under various operating conditions. This simulation model allows for in-depth analysis of the battery's thermal characteristics at different charge / discharge rates, ambient temperatures, and aging stages. The establishment of the simulation model helps predict potential thermal problems during the battery design phase, enabling corresponding preventative measures to be taken during battery manufacturing and energy storage power station design. Furthermore, the simulation model can be used to train operators, improving their understanding of the thermal management system and their emergency response capabilities by simulating different fault scenarios. Through these comprehensive measures, the thermal management system of the battery energy storage power station can more comprehensively ensure battery safety and performance, providing solid technical support for the long-term stable operation of the power station.

[0050] The following section will use the example of the controller establishing relevant mathematical expressions to represent the cooling process and the battery cooling process to illustrate how the controller establishes a cooling unit model.

[0051] Optionally, the cooling unit model includes a refrigeration circuit, which may include one or more refrigerant circulation pumps to drive the refrigerant flow within the circuit; one or more heat exchangers to transfer heat between the battery module and the refrigerant; and one or more control units to dynamically adjust the operating state of the refrigeration circuit based on the battery module's temperature data and the prediction results of an ambient temperature model. The control units may include temperature sensors, pressure sensors, and flow sensors to monitor the operating parameters of the refrigeration circuit in real time and ensure that the battery module's temperature remains within a safe and efficient range through a feedback control mechanism. Furthermore, the control units can be connected to the power station's central control system to achieve centralized control and optimization of the entire power station's thermal management. Through this integrated control strategy, the cooling unit can respond more intelligently and adaptively to the thermal management needs of the battery energy storage power station, ensuring the power station's efficient, stable, and safe operation.

[0052] Preferably, the cooling unit model includes at least a model portion representing the compressor, a model portion representing the condenser, a model portion representing the thermal expansion valve, a model portion representing the evaporator, and a model portion representing the cooler.

[0053] Preferably, the cooling unit model includes at least a mathematical expression for representing the battery cooling rate, a mathematical expression for representing the coolant temperature, and a mathematical expression for representing the pump.

[0054] Please see Figure 2 , Figure 2 This is a schematic diagram of a cooling unit provided in an embodiment of this application. Figure 2As shown, the cooling unit consists of a battery cooling circuit and a refrigeration circuit. In practical applications, the cooling unit may also include power electronics and motor cooling circuits.

[0055] In this circuit, the power electronics and motors are cooled by a coolant in the power electronics / motor circuit. The heated coolant is cooled by a fan as it passes through a radiator. Generally, the temperature control requirements for power electronics and motors are relatively undemanding because they operate over a wide temperature range. Therefore, the components and coolant can be adequately cooled by ambient air.

[0056] Figure 2 The external circuit is the battery cooling circuit, consisting of a pump, a radiator (not shown in the diagram), and the battery. Coolant is circulated by the pump and cools the battery. Heated coolant from the battery is cooled by ambient air at the front radiator. However, if the ambient air is insufficient to cool the battery and cabin, the refrigeration system will operate to provide additional cooling.

[0057] A refrigeration system is essential for providing additional cooling to the overall thermal management system of a power plant. Fluids in the refrigeration and battery cooling circuits exchange heat at the cooler. The coolant heated in the battery cooling circuit can be cooled by the refrigerant. However, this additional cooling increases the energy consumption of the cooling unit.

[0058] In the thermal management system of a battery energy storage power station, this cooling unit is operated by a controller, aiming to minimize cooling energy consumption while adequately cooling the components. Therefore, when the cooling demand of heat-generating components is low, ambient air is typically used for cooling, without the need for compressor operation. On the other hand, if components require additional cooling, compressors and pumps will operate at high speeds to regulate the component temperature.

[0059] It should be noted that when the refrigeration circuit is running, most of the energy in the thermal management system is consumed by the compressor and pump; this is known as active cooling mode. The way the speeds of the two actuators are controlled determines most of the energy consumption in the battery cooling unit. Therefore, the thermal management strategy, i.e., the control strategy of the cooling unit, is essential for the overall system efficiency.

[0060] The controller modeled the cooling unit using physical simulation. The following paragraphs describe the battery temperature dynamics and their relationship with the components in the cooling unit.

[0061] Specifically, the battery heating mechanism: During the operation of an energy storage power station, the battery current (I... batt The following can be calculated using an equivalent circuit model:

[0062]

[0063] Among them, V oc and Rbatt These represent the open-circuit voltage and internal resistance of the battery, respectively, P. batt It is the sum of the compressor power and pump power in the thermal management system.

[0064] Battery internal resistance causes the battery to heat up, generating heat. The formula is as follows:

[0065]

[0066] The cooling unit includes a compressor, condenser, thermal expansion valve, evaporator, and cooler. In the compressor, the refrigerant is compressed, and its temperature rises during the compression process. The compressor's mass flow rate... The model can be modeled as follows:

[0067]

[0068] Among them, V comp and ρ rfg These are the compressor displacement and refrigerant density, respectively. Volumetric efficiency (η) comp PR is a function of compressor speed and compression ratio. PR can be calculated using the following formula:

[0069] PR = p comp,out / p comp,in (4)

[0070] Where, p comp,out and p comp,in These represent the pressures at the compressor inlet and outlet, respectively. Compressor mechanical power (P) comp,mech The formula for calculating ) is:

[0071]

[0072] Among them, h comp,out h is the enthalpy value at the compressor outlet. eva,out η is the enthalpy value at the evaporator outlet. isen This refers to the isentropic efficiency of the compressor.

[0073] The power of the compressor (P) comp,elec ) can be based on motor efficiency (η) c The calculation is as follows:

[0074] P comp,elec =P comp,mech / η c (6)

[0075] It should be noted that in the condenser, the refrigerant is cooled and liquefied by radiating heat to the surrounding air. Upon passing through the thermal expansion valve, the liquefied refrigerant evaporates due to the pressure drop. In the evaporator, the refrigerant absorbs energy from the surrounding air and coolant through a heat exchange process. As a result, it changes from a liquid-vapor state to a gaseous state. The heat exchange between the refrigerant and coolant in the evaporator is modeled using a convection model with a parallel flow approach. The output fluid temperature of the evaporator can be described as follows:

[0076]

[0077] Among them, T clnt,eva,in and T clnt,eva,out These are the refrigerant temperatures at the evaporator inlet and outlet, respectively. T rfg,eva,out and T rfg,eva,in These are the refrigerant temperatures at the compressor outlet and inlet, respectively. eva and A eva These represent the heat transfer coefficient and cross-sectional area between the evaporator and the fluid, respectively. C is the coolant mass flow rate. clnt This refers to the specific heat capacity of the coolant.

[0078] In this system, the coolant in the battery cooling circuit circulates through coolant channels within the battery to absorb heat. Battery cooling rate. This can be represented by a uniform wall temperature model:

[0079]

[0080] Among them, T clnt,in and T clnt,out These represent the coolant temperatures at the inlet and outlet of the coolant passage, respectively. The coolant temperature at the outlet can be calculated using the following formula:

[0081]

[0082] Among them, h batt and A batt This represents the heat transfer coefficient and cross-sectional area between the battery and the coolant. The dynamic temperature of the battery is expressed as:

[0083]

[0084] Where, m batt and C batt These are the battery's thermal mass and specific heat capacity, respectively.

[0085] The mass flow rate of a pump can be expressed similarly to that of a compressor, as follows:

[0086]

[0087] Where, ωpump V pump and ρ pump These represent pump speed (in rpm), displacement, and coolant density, respectively. η pump This refers to the pump's volumetric efficiency. The pump's mechanical power (P) pump,mech It can be calculated in the following ways:

[0088]

[0089] Where, Δp pump Pump pressure drop. Pump power (P) pump,elec Available motor efficiency (η) p The following is an example:

[0090] P pump,elec =P pump,mech / η p (13)

[0091] It should be noted that the overall system described above is modeled as a nonlinear dynamic system with two states and two control inputs. The system equations can be expressed as a state-space model. If the model is discretized with a sampling time of Δt, the dynamic equations of the battery cooling unit can be:

[0092]

[0093] It should be noted that the benefits of establishing a cooling unit model based on a battery energy storage power station for step 102 are obvious. By establishing such a model, the controller can perform corresponding calculations based on model predictive control (MPC) and stochastic dynamic programming (SDP) algorithms, thus enabling better precise control and optimization of the cooling unit.

[0094] The controller in this embodiment can employ two control strategies to control the cooling unit: a first control strategy and a second control strategy. The first control strategy is a typical MPC control strategy, whose cost function consists of energy cost and penalty cost. This first control strategy exhibits optimal temperature control performance. The second control strategy is an MPC control strategy that considers a cost function composed of energy cost and a stochastically approximated residual cost. This strategy makes trade-offs between temperature control and energy consumption cost, ensuring optimal cost-effectiveness of the control strategy.

[0095] In optimal control theory, the optimal control problem of a dynamic system is defined by the cost function J as follows:

[0096]

[0097] Where x(k) is the state vector at time step k, u(k) is the control input, g(*) is the transition cost from time step k to N, and h(*) is the terminal cost used to constrain the final state. The goal of the optimal control problem (OCP) is to compute a set of optimal control inputs that minimize J:

[0098]

[0099] Among them, J * (x(k),k) represents the minimum cost, U * (k) is the optimal control input set.

[0100] By dividing the conversion cost into two levels, the above equation (16) can be expressed in the following recursive form:

[0101]

[0102] The optimal cost in the above formula is expressed as:

[0103]

[0104] Among them, J * (x(k+n),k+n) is the optimal cost from step k+n to N. Equation (18) above becomes the principle of optimization when n=1:

[0105]

[0106] Equation (19) above provides the foundation for dynamic programming (DP), a powerful OCP solver. DP can use equation (19) above and J*(x(N),N)=h(x(N)) to calculate the optimal control input from k=N to k=1 in a backward recursive manner.

[0107] For the DP formula of the battery cooling unit, OCP is formulated as energy consumption minimization with equality and inequality constraints.

[0108] The cooling unit has battery temperature data T. batt and coolant temperature using T clnt Two states, and compressor speed ω comp and cooling water pump speed ω pump Two control variables. The functions g(*) and h(*) in the above formula can be defined as:

[0109]

[0110] Where N is the number of time steps in a given work cycle; P cooling(k) represents the power consumption of the cooling unit in the k-th time step; ω is the weighting factor for the battery temperature deviation at the end of the operating cycle; T des It is the expected battery temperature at the final time step.

[0111] It should be noted that the future information required to calculate accurate operating costs is the battery heat generation for all future time instances in Equation (14) above. Once a duty cycle is given, the required battery power and the corresponding battery heat generation can be determined in advance.

[0112] After obtaining the dynamic expression equation and cost function describing the cooling unit model, as well as related functions, the controller can execute steps 102 and 103.

[0113] Optionally, in addition to the MPC algorithm mentioned in steps 102 and 103, the aforementioned cost function can also be solved using optimization algorithms, such as genetic algorithms, particle swarm optimization algorithms, or simulated annealing algorithms. These algorithms can handle complex nonlinear problems and find optimal or near-optimal control strategies. Through these algorithms, it can be ensured that the controller can quickly respond and adjust the control strategy under different operating conditions to adapt to the real-time thermal state of the battery module and environmental changes. Furthermore, the controller can automatically adjust control parameters based on historical data and predictive models to achieve long-term optimization and energy-saving effects. Through this intelligent control, the cooling unit of the battery energy storage power station can not only improve efficiency but also extend the battery's lifespan and reduce operating costs.

[0114] Optionally, the cost function can also be solved using machine learning algorithms, such as support vector machines, random forests, or neural networks. These algorithms can learn from large amounts of data and identify the optimal control strategy, thereby further improving the intelligence level of the controller. Through training with machine learning algorithms, the controller can automatically adjust its control strategy based on real-time data and historical operating conditions of the battery modules to adapt to different working environments and load changes. This adaptive control method not only improves the operating efficiency of the battery energy storage power station but also effectively prevents potential risks such as thermal runaway, ensuring the safe and stable operation of the power station.

[0115] Optionally, the cost function can also be solved using a hybrid algorithm, combining the advantages of optimization and machine learning algorithms to achieve more precise control. For example, an optimization algorithm can be used to first determine a general control strategy framework, and then machine learning algorithms can be used to fine-tune and optimize the strategy to adapt to the complexities and dynamic changes in power plant operation. This hybrid approach can balance the computational efficiency and control accuracy of the algorithm, providing a more reliable and efficient solution for the thermal management of battery energy storage power plants. Through this comprehensive strategy, the thermal management system of battery energy storage power plants can more flexibly cope with various operating conditions, ensuring that battery modules operate within their optimal temperature range, thereby improving the overall performance and safety of the power plant.

[0116] In steps 102 and 103 below, this paper will take solving the cost function based on the MPC algorithm and performing corresponding control based on the solution results as an example to illustrate the thermal management method provided in the embodiments of this application.

[0117] Step 102: When the temperature data is greater than or equal to the preset temperature threshold, the controller controls the cooling unit to cool the battery based on the first control strategy.

[0118] The temperature threshold can be set based on experience, or it can be set based on the controller's control speed and the safe handling time for thermal runaway.

[0119] The first control strategy is a control strategy based on the model predictive control (MPC) algorithm.

[0120] Preferably, the controller can obtain the dynamic expression equation and cost function of the model of the cooling unit; then calculate the approximate operating cost of the cooling unit based on the dynamic expression equation and cost function; then obtain the average heat generation of the battery, and calculate the predicted heat generation of the cooling unit within the prediction time range based on the average heat generation and the moving average filter algorithm; then determine the first optimal control input set based on the approximate operating cost and the predicted heat generation; finally, control the cooling unit to cool the battery based on the first optimal control input set.

[0121] The process of the controller acquiring the dynamic expression equation and cost function of the cooling unit model has been explained in step 101 and will not be repeated here.

[0122] Given the optimal operating cost for all possible states k+n, the controller can calculate the optimal control input set over the time range from k to k+n-1. However, this real-time implementation faces two obstacles: 1) the optimal operating cost for all possible states must be known; 2) future information or future disturbances from k to k+n-1 are required.

[0123] The first obstacle can be mitigated by approximating the running cost as a function of the current system state. One of the most typical forms of approximating the running cost is a quadratic function of the state:

[0124] J * (x(k+n),k+n)≈α(x(k+n)-x d ) 2 (twenty one)

[0125] Where α and x d These are design variables. This secondary cost is called the penalty cost, and it is the cost between the penalty state and the desired state x. d The deviation. This form seems heuristic, but if α and x... d Once properly determined, it is strongly correlated with the true optimal residual cost value.

[0126] The optimal cost-benefit function x(k+n) with a time step of k+n can be expressed as the optimal value x around x(k+n). * The Taylor expansion of (k+n) approximates:

[0127]

[0128] Where α0, α1, α2, x d and β are only affected by x * The effect of (k+n) is independent of x(k+n) or any control input.

[0129] The optimal input can be represented as:

[0130]

[0131] Wherein, the transition cost g(*) is defined as the energy consumption of the battery cooling unit, x d Let α be the desired temperature regulation, and let α be the weighting coefficient between temperature regulation and energy minimization. The disturbance information needed within the prediction timeframe is battery heat generation. Assuming it remains constant within the prediction range and its value is the same as the recent average heat generation, it can be obtained using a moving average filter:

[0132]

[0133] Where m is the length of the filter. It is the average heat output of the battery (i.e., the predicted heat output).

[0134] like Figure 3 As shown, Figure 3The schematic diagram of the MPC control process under the first control strategy provided in the embodiment of this application shows the controller structure obtained by using approximate operating cost and assumed disturbance information. After calculating the approximate operating cost based on the above formula (22) and the predicted heat generation of the battery based on the above formula (24), the controller can use the approximate operating cost and the predicted heat generation as inputs to the MPC algorithm and calculate the first optimal control input set through the MPC algorithm.

[0135] Then, the controller can send corresponding control commands to the cooling unit based on the first optimal control input set to perform corresponding control.

[0136] Step 103: If the temperature data is less than the temperature threshold, the controller controls the cooling unit to cool the battery based on the second control strategy.

[0137] The second control strategy is a control strategy based on the MPC algorithm and the optimal expected cost of the MPC algorithm. The optimal expected cost is calculated based on the stochastic dynamic programming (SDP) algorithm and serves as a constraint of the MPC algorithm.

[0138] It is understandable that the optimal expected cost of this MPC algorithm refers to the operating cost of the controller controlling the cooling unit based on the MPC algorithm.

[0139] In order to reduce energy consumption, when the controller controls the cooling unit based on the MPC algorithm, the controller's operating cost should be as close as possible to the optimal operating cost function.

[0140] Optionally, the controller can calculate the optimal expected cost itself, specifically including: constructing an optimal expected cost function; obtaining the corresponding infinite-term stochastic problem equation based on the optimal expected cost function; obtaining the random variable; solving the infinite-term stochastic problem equation based on the preset first recursive equation, second recursive equation, and the random variable to obtain the optimal expected cost, wherein the first recursive equation is used to iterate the optimal expected cost under a given strategy, and the second recursive equation is used to iterate the optimal strategy.

[0141] Optionally, the optimal expected cost can also be calculated by a computing device outside the battery energy storage power station and preset in the controller. For example, the controller can receive or update the latest optimal expected cost through a wireless network module.

[0142] Taking the calculation of the optimal expected cost using computing equipment as an example, the specific calculation method can be described as follows:

[0143] When future information is uncertain in terms of determinism but statistically available, computing devices can use statistical information to approximate the optimal operating cost function. An optimal expected cost is proposed to approximate the optimal operating cost:

[0144]

[0145] Optimal expected cost Given the current state, this represents the optimal expected value of the operating cost.

[0146]

[0147] Where d(k) is the disturbance input that affects the dynamics of the system.

[0148] Calculating equation (26) requires a relative time step with respect to N, where N is the length of a given work cycle. This implies that the controller should know the work cycle information, which is still non-causal. To avoid this non-causal problem, the search for the optimal expected cost can be treated as an infinite stochastic problem:

[0149]

[0150] Where γ is the discount factor and π is the control strategy. It is the penalty cost function for the state change in the next time step. The discount factor is a value between 0 and 1, which guarantees the convergence of the infinite optimization problem. It is the best expected cost for an indefinite problem given perturbation statistics. This is obtained through solving, and the formula is the standard formula for SDP. To solve using SDP, a probability distribution of the perturbation or future information is needed. The perturbation of the battery cooling unit is battery heat generation, which is modeled as a random variable. Typically, the probability distribution function is modeled as a Markov process, meaning that future states depend only on the current state. The Markov process for battery heat generation is represented by the following probability distribution:

[0151]

[0152] in, It is the heat generated in the next time step. v(k) is the battery heat generation at the current time step, and v(k) is the vehicle longitudinal speed at the current time step.

[0153] For example, the computing device can use historical data to extract a probability distribution. To calculate the probability distribution, the computing device can use Autonomous software to simulate a battery energy storage power station with a cooling unit over approximately 80 different duty cycles.

[0154] The optimal expected cost and control strategy of the SDP can be obtained using iterative algorithms. Examples include value iteration and policy iteration. Computational devices can use policy iteration, which has a faster convergence speed. The policy iteration algorithm uses the Bellman equation form, as shown below:

[0155]

[0156] The above Bellman equation can be derived as follows:

[0157]

[0158] In this embodiment, the policy iteration algorithm consists of two iterations: a policy evaluation iteration and a policy improvement iteration, such as... Figure 4 As shown, Figure 4 A flowchart of the iterative algorithm for the optimal expected cost and optimal policy provided in embodiments of this application. In the policy evaluation iteration, given a policy π... j Update the expected cost using the following first recursive equation:

[0159]

[0160] Where i represents the number of iterations in policy evaluation, and j represents the number of iterations in policy improvement. π is the expected cost calculated in the i-th iteration. j Let Xcur be the control strategy updated in the j-th iteration, and Xnext be the variable at the next time step. This can be obtained from the above formula.

[0161] In this embodiment of the application, the expected cost obtained by the first recursive equation (31) in two iterations and The absolute value of the difference is less than or equal to the preset first iteration threshold ε J Convergence occurs at the given policy π; the convergence value is π. j The optimal expected cost under these conditions is expressed as:

[0162]

[0163] Understandably, this cost only applies to a given policy π. j The optimal cost should be calculated using the optimal policy, which can be achieved through iterative policy improvement using the following second recursive equation:

[0164]

[0165] In this embodiment, the policy improvement is iterative until the optimal policy π is obtained after two iterations. j and π j+1The absolute value of the difference is less than or equal to the preset second iteration threshold ε π The second recursive equation (33) converges when the number of iterations reaches the maximum. After convergence, π and J represent the optimal expected cost and optimal strategy for the infinite problem, respectively.

[0166] like Figure 5 As shown, Figure 5 The schematic diagram of the MPC control process under the second control strategy provided in the embodiments of this application shows the process of using random variables corresponding to the optimal expected cost and battery heat generation as inputs to the MPC algorithm, and the controller obtaining the second optimal control input set based on these inputs and the MPC algorithm.

[0167] Specifically, the controller obtains the cost function of the model of the cooling unit and obtains the optimal expected cost; the controller establishes a random variable based on a Markov process, which is used to indicate the heat generation of the cooling unit; the controller determines a second optimal control input set based on the random variable, the optimal expected cost and the cost function; the controller controls the cooling unit to cool the battery based on the second optimal control input set.

[0168] In this embodiment, the artificial intelligence-based optimization model, compared to traditional passive temperature control strategies, possesses learning and predictive capabilities, resulting in faster temperature control response and better stability, effectively preventing battery thermal runaway events. Compared to other artificial intelligence models, this technology can pre-determine the optimal expected cost through offline optimization, thus avoiding increased real-time computing load, lowering hardware computing power requirements, and significantly reducing control system costs. The two control strategies proposed in this application can balance both speed-priority and cost-priority considerations, thereby reducing the energy consumption required for temperature control while ensuring temperature control capabilities, and significantly improving the economics of battery energy storage power stations.

[0169] Please see Figure 6 , Figure 6 A set of simulation data comparison charts under different control strategies provided for embodiments of this application.

[0170] This application verifies the control strategy through simulation in three aspects: performance evaluation, robustness evaluation, and computational efficiency evaluation.

[0171] Specifically, this application applies multiple control strategies to the same system. Figure 6(a) Data for the cooling unit under the traditional MPC algorithm control strategy; (b) Data for the cooling unit under the first control strategy in this embodiment; (c) Data for the cooling unit under the second control strategy in this embodiment; (d) Data for the cooling unit under thermostat control. The upper and lower limits of the thermostat are set to 34℃ and 32.5℃, respectively. The operating rules are adjusted so that the pump and compressor operate at 2000rpm and 3000rpm at the upper limit, and 0rpm at the lower limit.

[0172] As can be seen, compared with the traditional MPC algorithm, the battery temperature fluctuation is smaller and more stable under the first control strategy and the second control strategy provided in this application embodiment; and the energy consumption of the cooling unit is lower under the second control strategy.

[0173] The method provided in the embodiments of this application has been described above. The system provided in the embodiments of this application is described below.

[0174] Please see Figure 7 , Figure 7 This is a schematic diagram of the structure of a battery energy storage power station 700 provided in an embodiment of this application. Figure 7 As shown, the battery energy storage power station 700 includes a battery energy storage power station thermal management system 701 and a battery 702. The system 701 includes a controller 7011 and a cooling unit 7012. The controller 7011 is used to control the cooling unit 7012 to cool the battery 702; specifically,

[0175] The controller 7011 is used to acquire the temperature data of the battery; if the temperature data is greater than or equal to a preset temperature threshold, the controller 7011 is also used to control the cooling unit 7012 to cool the battery 702 based on a first control strategy; if the temperature data is less than the temperature threshold, the controller 7011 is also used to control the cooling unit 7012 to cool the battery 702 based on a second control strategy.

[0176] The first control strategy is a control strategy based on the Model Predictive Control (MPC) algorithm; the second control strategy is a control strategy based on the MPC algorithm and the optimal expected cost of the MPC algorithm, wherein the optimal expected cost is calculated based on the Stochastic Dynamic Programming (SDP) algorithm and serves as a constraint of the MPC algorithm.

[0177] Preferably, the optimal expected cost is preset in the controller 7011.

[0178] Understandably, the controller 7011 can also achieve the above. Figures 3 to 5 The methods described in any of the embodiments will not be repeated here.

[0179] like Figure 8As shown, Figure 8 This is a schematic diagram of a possible logical structure of a computing device provided in an embodiment of this application. The computing device 800 includes a processor 801, a communication interface 802, a memory 803, and a bus 804, which are interconnected via the bus 804. In an embodiment of this application, the processor 801 is used to control and manage the operation of the computing device 800; for example, the processor 801 is used to execute... Figures 1 to 5 The steps in any embodiment and / or other processes used in the techniques described herein. Communication interface 802 is used to support communication by computing device 800. Memory 803 is used to store program code and data of computing device 800.

[0180] In another embodiment of this application, a computer-readable storage medium is also provided, which stores computer-executable instructions. When at least one processor of the device executes the computer-executable instructions, the device performs the aforementioned... Figure 1 , Figures 3 to 5 The method described in any of the embodiments.

[0181] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0182] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0183] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0184] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0185] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A battery energy storage plant thermal management method, characterized by, The application is applied to a battery energy storage power station thermal management system, the system comprises a controller and a cooling unit, the controller is used for controlling the cooling unit to cool the battery of the battery energy storage power station; the method comprises: The controller obtains temperature data of the battery; In the case that the temperature data is greater than or equal to a preset temperature threshold, the controller controls the cooling unit to cool the battery based on a first control strategy; wherein the first control strategy is a control strategy based on a model predictive control (MPC) algorithm; In the case that the temperature data is less than the temperature threshold, the controller controls the cooling unit to cool the battery based on a second control strategy; the second control strategy is a control strategy based on the MPC algorithm and an optimal expected cost of the MPC algorithm, the optimal expected cost is calculated based on a stochastic dynamic programming (SDP) algorithm, and the optimal expected cost is used as a constraint of the MPC algorithm; The controller controls the cooling unit to cool the battery based on the first control strategy, comprising: The controller obtains a dynamic expression equation and a cost function of a model of the cooling unit; The controller calculates an approximate running cost of the cooling unit based on the dynamic expression equation and the cost function; The controller obtains an average heat generation of the battery; The controller calculates a predicted heat generation of the cooling unit in a predicted time range based on the average heat generation and a moving average filter algorithm; The controller determines a first optimal control input set based on the approximate running cost and the predicted heat generation; The controller controls the cooling unit to cool the battery based on the first optimal control input set; The controller controls the cooling unit to cool the battery based on the second control strategy, comprising: The controller obtains a cost function of a model of the cooling unit and an optimal expected cost; The controller establishes a random variable based on a Markov process, the random variable is used for indicating a heat generation of the cooling unit; The controller determines a second optimal control input set based on the random variable, the optimal expected cost and the cost function; The controller controls the cooling unit to cool the battery based on the second optimal control input set.

2. The method of claim 1, wherein, The optimal expected cost is preset in the controller.

3. The method of claim 1, wherein, The optimal expected cost is obtained, comprising: An optimal expected cost function is constructed; A corresponding infinite-period random problem equation is obtained based on the optimal expected cost function; The random variable is obtained; The infinite-period random problem equation is solved based on a preset first recursive equation, a second recursive equation and the random variable to obtain the optimal expected cost, wherein the first recursive equation is used for iterating an optimal expected cost under a given strategy, and the second recursive equation is used for iterating an optimal strategy.

4. The method of claim 1, wherein, After the controller obtains the temperature data of the battery, the method further comprises: The controller establishes a model based on the cooling unit; The controller constructs a dynamic expression equation and a cost function of the model of the cooling unit.

5. A battery energy storage plant thermal management system, characterized by, The system comprises a controller and a cooling unit, the controller being configured to control the cooling unit to cool the battery of the battery energy storage power station; The controller is configured to obtain temperature data of the battery; When the temperature data is greater than or equal to a preset temperature threshold, the controller is further configured to control the cooling unit to cool the battery based on a first control strategy; wherein the first control strategy is a model predictive control (MPC) algorithm-based control strategy; When the temperature data is less than the temperature threshold, the controller is further configured to control the cooling unit to cool the battery based on a second control strategy; the second control strategy is a control strategy based on the MPC algorithm and an optimal expected cost of the MPC algorithm, the optimal expected cost being calculated based on a stochastic dynamic programming (SDP) algorithm, and the optimal expected cost being used as a constraint of the MPC algorithm; The controller is specifically configured to: When the temperature data is greater than or equal to a preset temperature threshold, obtain a dynamic expression equation and a cost function of a model of the cooling unit; calculate an approximate running cost of the cooling unit based on the dynamic expression equation and the cost function; obtain an average heat generation of the battery; calculate a predicted heat generation of the cooling unit within a prediction time range based on the average heat generation and a moving average filter algorithm; determine a first optimal control input set based on the approximate running cost and the predicted heat generation; and control the cooling unit to cool the battery based on the first optimal control input set. When the temperature data is less than the temperature threshold, obtain a cost function of a model of the cooling unit and an optimal expected cost; establish a random variable based on a Markov process, the random variable being configured to indicate the heat generation of the cooling unit; determine a second optimal control input set based on the random variable, the optimal expected cost and the cost function; and control the cooling unit to cool the battery based on the second optimal control input set.

6. The battery energy storage plant thermal management system of claim 5, wherein, The optimal expected cost is preset in the controller.

7. A computing device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method of any one of claims 1 to 4.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method of any one of claims 1 to 4.

Images