Energy storage battery temperature control early warning system

Abstract

The application relates to an energy storage battery temperature control early warning system, belonging to the technical field of energy storage power station control, which comprises a power grid load dynamic prediction unit, which is used for processing based on historical energy storage power data and a real-time correction factor of a power grid, adopting a dynamic weighted power prediction model to obtain a predicted energy storage power curve, a battery thermal characteristic digital twin unit, which is used for receiving the predicted energy storage power curve, combining current battery states obtained from a battery management system and operation power of a temperature control system fed back by a temperature control system fine execution unit, driving an electro-thermal coupling simulation model to perform simulation processing to obtain a predicted maximum temperature difference, a temperature control early warning and decision unit, which is used for receiving the predicted maximum temperature difference, combining a preset temperature control system response time, adopting a thermal risk index model to perform calculation processing to obtain a thermal risk index, and the application intervenes in adjustment in advance, eliminates response delay, effectively avoids local overheating of the battery, and prevents thermal runaway from occurring.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy storage power station control, in particular to a kind of energy storage battery temperature control early warning system. BACKGROUND

[0002] When large energy storage power station participates in high frequency dynamic frequency modulation of power grid, battery system needs to withstand high rate, bidirectional power impact, this working condition aggravates battery heat production, and it appears that significant temperature difference between each monomer battery in module, and it is well known to those skilled in the art that the maximum monomer temperature difference (ΔT max ) in battery module is the core index that influences the overall life and safety of battery cluster, if it frequently exceeds 8 ℃, etc. Key threshold, will trigger the barrel effect inside battery cluster, cause irreversible capacity premature attenuation and life loss.

[0003] Existing temperature control technology is mostly passive response mode, i.e. high-power cooling is started after monitoring that temperature or temperature difference exceeds limit, this kind of scheme exists response lag, temperature difference out of control risk due to inherent delay of heat transfer, and the extensive control strategy of on-off causes huge parasitic energy consumption of temperature control system, influence the overall operation efficiency of energy storage power station, therefore, there is a profound technical contradiction between guaranteeing temperature control performance and reducing parasitic energy consumption.

[0004] The above information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, therefore it can include information that does not constitute prior art known to those skilled in the art. SUMMARY

[0005] The purpose of the present application is to provide an energy storage battery temperature control early warning system to solve the problems raised in the above background.

[0006] The technical scheme of the present application is, including: power grid load dynamic prediction unit, for based on historical energy storage power data and power grid real-time correction factor, using dynamic weighted power prediction model to process, obtains predicted energy storage power curve;

[0007] Battery thermal characteristics digital twin unit, for receiving predicted energy storage power curve, combine current battery state obtained from battery management system and temperature control system operation power fed back by temperature control system refinement execution unit, drive electro-thermal coupling simulation model to simulate processing, obtain predicted maximum temperature difference;

[0008] Temperature control early warning and decision unit, for receiving predicted maximum temperature difference, combine preset temperature control system response time, using thermal risk index model to calculate processing, obtain thermal risk index;

[0009] The temperature control system fine execution unit is configured to receive the thermal risk index, adjust and process the thermal risk index using a proportional-integral control law, obtain a temperature control system control power, and feed back the temperature control system control power as a temperature control system operation power to the battery thermal characteristic digital twin unit.

[0010] Preferably, the processing procedure of the dynamic weighted power prediction model is as follows:

[0011] The historical energy storage power data is weighted and summed with the corresponding dynamic weight, and the weighted and summed result is superimposed with a real-time correction factor of the power grid to obtain a predicted energy storage power curve.

[0012] Preferably, the dynamic weight is periodically self-adaptively adjusted according to a prediction error of a previous period by an online learning algorithm.

[0013] Preferably, the processing procedure of the battery thermal characteristic digital twin unit comprises:

[0014] S1. Based on the predicted energy storage power curve, a coulomb integration method is used for processing to obtain a predicted state of charge.

[0015] S2. In combination with the predicted state of charge, a battery equivalent circuit model is used for processing to obtain a predicted terminal voltage and a predicted current.

[0016] S3. Based on the predicted current and the predicted state of charge, a Bernardi battery heat generation model is used for processing to obtain a heat generation rate of each single battery.

[0017] S4. The heat generation rate is taken as an internal heat source, and the temperature control system operation power is taken as an active heat dissipation item, which are substituted into a single battery temperature evolution model for simulation processing to obtain a predicted temperature of each single battery, and based on the predicted temperature of each single battery, a predicted maximum temperature difference is obtained.

[0018] Preferably, the processing procedure of the temperature control warning and decision unit comprises:

[0019] According to the temperature evolution information output by the battery thermal characteristic digital twin unit, a temperature control buffer time required from the current time to the first time when the predicted maximum temperature difference reaches a preset maximum temperature difference critical value is calculated.

[0020] Preferably, the processing procedure of the temperature control warning and decision unit further comprises:

[0021] A temperature difference risk degree is determined, which is a ratio of the predicted maximum temperature difference to the maximum temperature difference critical value.

[0022] A time risk degree is determined, which is a ratio of a temperature control system response time to the temperature control buffer time.

[0023] The temperature difference risk degree and the time risk degree are weighted and summed to obtain a thermal risk index.

[0024] Preferably, the adjustment process of the proportional-integral control law is as follows:

[0025] The deviation of the thermal risk index from a preset target risk value is obtained, and the deviation is subjected to proportional operation and integral operation, and the operation result is superimposed with a preset basic operation power to obtain a control power of the temperature control system.

[0026] The present application improves the energy storage battery temperature control early warning system, compared with the prior art, has the following improvements and advantages:

[0027] Firstly, the present application realizes the fundamental change of the temperature control strategy from passive response to forward-looking management, significantly improves the operation safety of the energy storage power station, predicts the future short-time domain energy storage power impact in advance through the setting of the power grid load dynamic prediction unit and the use of the dynamic weighted power prediction model; the prediction result drives the electro-thermal coupling simulation model in the battery thermal characteristic digital twin unit, and then deduces the future temperature evolution trend; this design enables the system to foresee the risk before the temperature difference exceeds the limit, so as to intervene in advance, eliminate the response delay, effectively avoid the local overheating of the battery, and prevent the occurrence of thermal runaway;

[0028] Secondly, the present application introduces a quantitative risk decision mechanism, making the temperature control adjustment more accurate and reasonable, and creatively constructs a temperature control early warning and decision unit, so that the system can respond differently and more reasonably to different degrees of potential thermal risk according to the size of the thermal risk index, avoiding unnecessary excessive cooling or insufficient intervention;

[0029] Thirdly, the present application realizes the refinement and high efficiency of temperature control execution, greatly reduces the parasitic energy consumption of the energy storage power station, and sets a temperature control system refinement execution unit to solve the energy loss caused by frequent start-stop and excessive refrigeration in the traditional control mode, ensures that the battery module maximum monomer temperature difference is reliably maintained within the safety threshold, and significantly reduces the parasitic energy consumption of the temperature control system, thereby improving the overall operation efficiency and economy of the energy storage power station;

[0030] Fourthly, the application effectively suppresses the temperature inconsistency inside the battery cluster, thereby prolonging the overall service life of the energy storage asset; the significant temperature difference between each single battery in the battery module is the core factor leading to the barrel effect of the battery cluster, thereby triggering irreversible capacity attenuation; the application effectively suppresses the temperature inconsistency inside the battery cluster through the closed-loop control of prediction-simulation-decision-execution-feedback; by combining the prospective prediction of the maximum temperature difference, the core index, with fine control, the system slows down the premature attenuation of the battery caused by temperature difference, ensures the consistency of the overall performance of the battery cluster, and ultimately significantly prolongs the service life and value of the energy storage asset. BRIEF DESCRIPTION OF DRAWINGS

[0031] The application will be further explained in conjunction with the accompanying drawings and embodiments:

[0032] Figure 1 is a flowchart of a temperature control and early warning system for an energy storage battery according to the application. DETAILED DESCRIPTION

[0033] To make the purpose, technical solutions and advantages of the application clearer, the application will be further described in detail below in conjunction with specific embodiments.

[0034] Embodiment 1:

[0035] Please refer to Figure 1 The application provides a temperature control and early warning system for an energy storage battery, comprising: a power grid load dynamic prediction unit for processing based on historical energy storage power data and real-time correction factors of the power grid using a dynamic weighted power prediction model to obtain a predicted energy storage power curve;

[0036] A battery thermal characteristic digital twin unit for receiving the predicted energy storage power curve, combining the current battery state obtained from the battery management system and the operating power of the temperature control system fed back by the temperature control system fine execution unit, driving an electro-thermal coupling simulation model for simulation processing to obtain a predicted maximum temperature difference;

[0037] A temperature control and early warning and decision unit for receiving the predicted maximum temperature difference, combining a preset temperature control system response time, and calculating and processing using a thermal risk index model to obtain a thermal risk index;

[0038] A temperature control system fine execution unit for receiving the thermal risk index, adjusting and processing using a proportional-integral control law to obtain temperature control system control power, and feeding back the temperature control system control power as the operating power of the temperature control system to the battery thermal characteristic digital twin unit.

[0039] The core of the energy storage battery temperature control early warning system disclosed in the embodiment is a closed-loop control architecture composed of a power grid load dynamic prediction unit, a battery thermal characteristic digital twin unit, a temperature control early warning and decision unit, and a temperature control system fine execution unit. In the architecture, the power grid load dynamic prediction unit constitutes the logical starting point of the entire forward-looking control, and the output predicted energy storage power curve provides a key data basis for the subsequent units. Based on the predicted data, the battery thermal characteristic digital twin unit drives an electrical-thermal coupling simulation model that runs synchronously with the physical battery system to accurately deduce the future battery temperature difference evolution trend. The temperature control early warning and decision unit calculates a thermal risk index based on the deduced results through a quantitative risk model to guide the final control action. The temperature control system fine execution unit generates smooth and accurate power instructions based on the risk index and applies them to the temperature control hardware, while feeding the power instructions back to the digital twin unit as feedback signals, thereby forming a complete prediction-simulation-decision-execution-feedback closed loop. This architecture aims to fundamentally solve the problems of temperature control response lag and high parasitic energy consumption caused by inherent thermal transfer delay when large-scale energy storage stations participate in high-frequency dynamic frequency modulation of the power grid. The technical effect is that the maximum single cell temperature difference within the battery module can be reliably controlled within a preset safety threshold, such as 8℃, under severe dynamic frequency modulation conditions, while significantly reducing the overall operating energy consumption of the temperature control system.

[0040] Embodiment 2

[0041] The processing process of the dynamic weighted power prediction model is as follows:

[0042] The historical energy storage power data and the corresponding dynamic weights are weighted and summed, and the weighted sum result is superimposed with the real-time correction factor of the power grid to obtain the predicted energy storage power curve.

[0043] The dynamic weight is periodically self-adapted according to the prediction error of the last period through an online learning algorithm.

[0044] In the embodiment, the function of the power grid load dynamic prediction unit is realized through a dynamic weighted power prediction model, and the mathematical expression of the model is The design of the model aims to provide a lightweight and fast-responding second-level power prediction mechanism for large-scale energy storage stations that need to participate in high-frequency dynamic frequency modulation of the power grid. The internal logic is to combine the periodicity of historical data with the instant randomness of power grid frequency modulation instructions. P pred (t+Δt p ) represents the predicted energy storage power at future Δt p time from the current time t, which is the direct output of the model; P hist (t-(i-1)Δt hist) represents the actual power value of the energy storage system at the i-th historical sampling point before the current time t; w i (t) is the dynamic weight of the i-th historical data point, which is a dimensionless core adjustable parameter. This weight is periodically self-adaptively adjusted by an online learning algorithm, such as the least mean square algorithm, according to the prediction error of the last period, so that the historical data segment with better prediction effect obtains a higher weight, and the sum of all weights is normalized to 1; t: current time; i: index of historical data point, which is a loop variable;

[0045] Taking the normalized least mean square algorithm as an example, at the end of each prediction step Δt p , the actual power value P actual (t) is obtained, and the prediction error e(t-Δt p ) of the last step can be calculated; the update rule of the weight vector W(t) can be represented as:

[0046]

[0047] In the formula, W(t) = [w1(t), w2(t),..., w N (t)] T is the weight vector, P hist (t-Δt p ) is the historical power data vector corresponding to the last time, and μ is the step factor that determines the convergence speed and stability; the updated weight vector needs to be normalized to ensure Through this method, the weight can be quickly self-adaptively adjusted according to the latest prediction performance; e(t-Δt p ): prediction error of the last step; ||P hist (t-Δt p )| 2 : L2 norm of the historical power data vector, Euclidean norm, which is a standard mathematical symbol; T: standard mathematical symbol for vector transposition, used to convert a row vector into a column vector;

[0048] δ grid (t) is a real-time correction factor of the power grid, which is calculated by the energy management system of the power station according to the real-time deviation of the power grid frequency, and is used to compensate for the prediction lag caused by relying solely on historical data;

[0049] This correction factor can be generated by a proportional controller, and the calculation formula is:

[0050] δ grid (t) = K f · (f ref -f actual (t))

[0051] wherein f actual (t) is the real-time frequency of the power grid, f ref is the reference frequency of the power grid, for example, 50 Hz, K f is the frequency deviation correction gain, which can be pre-calibrated according to the specific response characteristics of the power grid and the regulation capacity of the energy storage power station, wherein K f has the dimension of power / frequency, unit: W / Hz, and this correction term enables the predicted power to respond quickly to the real-time demand of the power grid; f ref : reference frequency of the power grid; f actual (t): real-time frequency of the power grid; K f : frequency deviation correction gain;

[0052] N is the size of the sliding window of the number of historical data points; Δt p and Δt hist are the prediction step and the historical data sampling interval, respectively; in system application, the power grid load dynamic prediction unit collects the historical power sequence and obtains the real-time correction factor from the EMS, and through the above model calculation, outputs the accurate prediction of the energy storage power curve P pred (t) in the future short time domain; the dynamic weight adaptive capability of this prediction mechanism ensures that the model can be continuously optimized, and the predicted power curve output by the prediction mechanism provides a data basis for the forward-looking control of the entire system, and ensures the predictability of the subsequent temperature control decision from the source.

[0053] Embodiment 3

[0054] The processing process of the battery thermal characteristic digital twin unit includes:

[0055] S1, based on the predicted energy storage power curve, the coulomb integral method is used for processing to obtain the predicted state of charge;

[0056] S2, combined with the predicted state of charge, the battery equivalent circuit model is used for processing to obtain the predicted terminal voltage and the predicted current;

[0057] S3, based on the predicted current and the predicted state of charge, the Bernardi battery heat generation model is used for processing to obtain the heat generation rate of each single battery;

[0058] S4, taking the heat generation rate as an internal heat source and taking the temperature control system operating power as an active heat dissipation term, the single battery temperature evolution model is simulated to obtain the predicted temperature of each single battery; and based on the predicted temperature of each single battery, the predicted maximum temperature difference is obtained.

[0059] In this embodiment, the function of the battery thermal characteristic digital twin unit is to establish an accurate mapping from predicted power to predicted temperature difference. This process is achieved through a multi-step coupled model that includes electric state of charge prediction, heat generation rate calculation and temperature evolution simulation. There are close logical and data transfer relationships between each step.

[0060] To perform electric-charge state prediction, based on the predicted current I pred (t), predicting the state of charge (SoC) using the Coulomb integral method. pred Evolution of the state of charge (SoC); the state of charge (SoC) is a dimensionless parameter representing the ratio of current charge to rated capacity; the calculation formula is: Where Δt: time step, which, depending on the context, is the simulation time interval; Q rated,As The rated capacity of the battery, measured in ampere-seconds (A·s); SoC(t) is the current SoC value obtained from the battery management system, and is dimensionless; Q rated,As The rated capacity of the battery is expressed in ampere-seconds (A·s). If the rated capacity Q... rated If the unit is ampere-hour (Ah), then the formula is expressed as follows: This formula calculates the change in SoC using the Coulomb integral method, and this change is also dimensionless; based on this, the prediction result is Q. rated The rated capacity of a battery, measured in ampere-hours (Ah); using the battery's equivalent circuit model U... pred (t)=U ocv (SoC pred (t))-I pred (t)·R int,j (T j (t),SoC pred (t) is used to predict the battery's terminal voltage U. pred (t), where U ocv The open-circuit voltage of the battery is the SoC. pred The function, R int,j The internal impedance of a single cell is related to temperature T. j and SoC pred Related, open-circuit voltage function U ocv (SoC) and internal impedance function R int,j (T j The SoC (System-on-a-Chip) is fixed for a specific cell model and can be calibrated using standard experimental testing methods. For example, using hybrid pulse power characteristic testing, a series of charge-discharge pulse tests are performed on the battery at different ambient temperatures and state of charge (SOC). By analyzing the voltage response data, the U can be fitted. ocv and R int,j Lookup tables or empirical functions related to temperature and SoC; j: index of a single cell;

[0061] To calculate the heat generation rate, the predicted current I pred (t) and the updated state parameters, the Bernardi cell heat generation model is adopted The model can accurately describe the Joule heat and reaction heat of the battery and is the basis for building a high-fidelity thermal model; where Q gen,j (t) is the instantaneous heat generation rate of monomer j at time t; I pred (t) is the predicted current obtained by iterative solution in the previous step; R int,j (T j , SoC pred (t)) is the internal impedance of monomer j, which is a function of the current temperature T j and the predicted state of charge SoC pred ; is the entropy heat coefficient of monomer j; T j (t) is the real-time temperature of monomer j at time t obtained from the BMS, which is used to calculate the entropy heat term in the Bernardi cell heat generation model. The temperature T j (t) in the Bernardi cell heat generation model must use the absolute temperature scale, with the unit of Kelvin (K); T: in the entropy heat coefficient , T represents the absolute temperature;

[0062] The entropy heat coefficient in the model can also be determined by experiment, that is, the open-circuit voltage U ocv of the battery is measured at different temperatures, and the rate of change of voltage with temperature is calculated. The mass m j of the monomer battery can be obtained by direct weighing, and the specific heat capacity c p can be measured by an accelerating rate calorimeter or other equipment;

[0063] To perform temperature evolution simulation, the calculated heat generation rate Q gen,j (t) is taken as an internal heat source, and the heat dissipation power Q tcs (t) corresponding to the operating power P cool,j of the temperature control system fed back by the execution unit is taken as the active heat dissipation term, which is substituted into the monomer battery temperature evolution model This model directly links the control behavior and the thermodynamic state, solving the decoupling problem of heat dissipation capacity and control output in traditional models; where T j,pred (t+Δt) is the predicted temperature of monomer j at future time Δt; m j and c p are the mass and specific heat capacity of monomer j, respectively; Q cool,j (t) is the total heat dissipation power of the temperature control system, which is the sum of the internal heat source and the active heat dissipation term; P tcs (t current) decision, which is the key feedback input to realize control loop; m j : mass of monomer battery j; c p : specific heat capacity of monomer battery j; Q cool,j (t): active heat dissipation power corresponding to the operating power of the temperature control system;

[0064] The conversion relationship from the system operating power P tcs to the monomer heat dissipation power Q cool,j can be modeled as:

[0065] Q cool,j (t) = η cop · P tcs (t) · β j

[0066] In the formula, η cop is the overall energy efficiency ratio of the temperature control system, such as an air conditioner or a cooling unit, which converts the input electric power into the total refrigeration power. This parameter can be calibrated by performance testing of the temperature control hardware. β j is the heat dissipation allocation coefficient of monomer battery j, representing the proportion allocated from the total cooling capacity. This coefficient is related to the flow field distribution inside the battery module. For all monomers, ∑β j = 1; The value of β j can be obtained by computational fluid dynamics simulation or experimental calibration by sticking multiple temperature sensors on the battery module. This model accurately links the macroscopic control power with the microscopic monomer heat dissipation effect; η cop : overall energy efficiency ratio of the temperature control system; P tcs (t): actual operating power of the temperature control system; β j : heat dissipation allocation coefficient of monomer battery j;

[0067] Through the above steps, the digital twin unit receives the predicted power, combines the real-time state of the BMS and the actual operating power of the temperature control system, simulates the temperature evolution trajectory of each monomer battery in the future short time domain, and finally outputs the predicted maximum temperature difference ΔT max,pred ; This process provides high-precision, future-oriented temperature difference data for subsequent decision-making, enabling the system to make decisions based on scientific prediction of the future.

[0068] Embodiment 4

[0069] The processing process of the temperature control warning and decision unit includes:

[0070] According to the temperature evolution information output by the battery thermal characteristic digital twin unit, the temperature control buffer time required from the current time to the first time when the predicted maximum temperature difference reaches the preset maximum temperature difference threshold is calculated;

[0071] The processing process of the temperature control warning and decision unit also includes:

[0072] determining a temperature difference risk degree, the temperature difference risk degree being a ratio of the predicted maximum temperature difference and a maximum temperature difference threshold value;

[0073] determining a time risk degree, the time risk degree being a ratio of a temperature control system response time and a temperature control buffer time;

[0074] performing weighted summation processing on the temperature difference risk degree and the time risk degree to obtain a thermal risk index.

[0075] In the embodiment, the temperature control early warning and decision unit serves as a decision core module of the system, receives the predicted maximum temperature difference ΔT max,pred from the digital twin unit, and performs processing based on a self-defined thermal risk index model; the core decision model is represented as The design logic of this model is that the temperature control decision should not only depend on the peak value of the predicted temperature difference, but also consider the time urgency required by the system to respond to the risk, so as to upgrade the decision from a single-dimensional threshold judgment to a comprehensive quantitative assessment of the risk urgency. th (t): dimensionless comprehensive thermal risk index; ω T : dimensionless weight factor; ΔT max,pred (t+τ response ): predicted maximum temperature difference; ΔT crit : maximum temperature difference threshold value; ω τ : dimensionless weight factor; τ response : temperature control system response time; τ buffer (t): temperature control buffer time;

[0076] wherein R th (t) is the output dimensionless comprehensive thermal risk index, which is the only input of the subsequent control module; the first term is the temperature difference risk degree, which is determined by the ratio of the predicted maximum temperature difference ΔT max,pred (t+τ response ) and the maximum temperature difference threshold value ΔT crit , and the ΔT crit is a threshold value set according to the battery safety and service life requirements through experimental data or industry standards, for example, 8℃; the second term is the time risk degree, which is determined by the ratio of the temperature control system response time τ response and the temperature control buffer time τ buffer (t), wherein τ response is the inherent response time obtained by step response test calibration on the hardware system, and τ buffer (t) is calculated by the temperature evolution curve output by the digital twin unit from the current time to the first time when ΔT max,pred reaches the threshold value ΔTcrit Time required; ω T With ω τ These are dimensionless weighting factors, and their sum is 1. They are used to adjust the conservatism or economy of the decision-making strategy, and their ratio can be optimized in the simulation according to the operating objectives of the power plant.

[0077] In system applications, this unit calculates the temperature control buffer time τ. buffer (t) is used to determine the temperature difference risk level and the time risk level, and the thermal risk index R is obtained by weighted summation. th (t); This index provides a quantitative scientific basis for subsequent refined adjustments, enabling control decisions to simultaneously consider both the magnitude of the risk and the urgency of the time.

[0078] Example 5

[0079] The adjustment process of the proportional-integral control law is as follows:

[0080] The deviation between the thermal risk index and the preset target risk value is obtained, and proportional and integral calculations are performed on the deviation. The calculation results are then superimposed with the preset base operating power to obtain the control power of the temperature control system.

[0081] In this embodiment, the function of the temperature control system's refined execution unit is to process the risk index R output by the decision unit. th (t) is transformed into precise and continuous power control commands for the temperature control hardware. The adjustment process is achieved through a proportional-integral (PI) control law; the mathematical expression of this control law is: The technical reason for adopting this control law is that it replaces the traditional on / off control with a regulator that can output continuous and smooth power, thereby achieving a precise match between heat dissipation energy output and heat dissipation demand, maximizing energy savings while ensuring temperature control effect.

[0082] Where P tcs (t) represents the final output control power applied to the temperature control system hardware at time t, measured in watts (W); P base The base operating power (W) necessary to maintain the system's minimum cycle length; R th (t) represents the real-time thermal risk index input by the decision-making unit, which is dimensionless; R th,target K represents an ideal steady-state target value below the risk trigger threshold, for example, 0.5, dimensionless; p The proportional gain, measured in watts (W), determines the strength of the controller's response to the current risk deviation; K i The integral gain, measured in watts per second (W / s), is responsible for eliminating steady-state errors; K p and K iThese two gain parameters can be calibrated by Ziegler-Nichols and other engineering tuning methods during system commissioning. base : basic operating power th,target : target risk value of ideal steady state p : proportional gain i : integral gain : integral variable, representing the time in the integral interval, is the standard symbol in calculus

[0083] In system application, the execution unit continuously receives the thermal risk index R th (t) and calculates its deviation from the target risk value R th,target ; the PI controller generates an accurate adjustment power value through proportional and integral operation according to the deviation, and the final temperature control system control power P tcs (t) is formed by superimposing the basic operating power; this power instruction directly drives the temperature control equipment to realize fine adjustment of cooling intensity; as the temperature control system operating power, it is fed back to the battery thermal characteristic digital twin unit for simulation correction in the next round; this closed-loop feedback ensures that each prediction of the system takes into account the actual impact of the last control action, thereby realizing dynamic and adaptive predictive control, effectively avoiding temperature difference overshoot, and maintaining the temperature consistency of the battery cluster within a safe range at a lower energy cost.

[0084] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A temperature control and early warning system for an energy storage battery, characterized in that, include: The grid load dynamic prediction unit is used to process historical energy storage power data and real-time grid correction factors using a dynamic weighted power prediction model to obtain the predicted energy storage power curve. A digital twin unit for battery thermal characteristics is used to receive the predicted energy storage power curve, combine the current battery state obtained from the battery management system with the operating power of the temperature control system fed back by the temperature control system fine execution unit, drive the electro-thermal coupling simulation model to perform simulation processing, and obtain the predicted maximum temperature difference of the battery. The temperature control early warning and decision-making unit is used to receive the predicted maximum temperature difference, combine it with the preset temperature control system response time, and use the thermal risk index model to calculate and process the thermal risk index. This index provides a quantitative scientific basis for subsequent fine-tuning, enabling control decisions to take into account both the magnitude of the risk and the urgency of the time. The temperature control system's refined execution unit is used to receive the thermal risk index, perform adjustment processing using a proportional-integral control law, obtain the temperature control system's control power, and feed the temperature control system's control power as the temperature control system's operating power back to the battery's thermal characteristic digital twin unit. The real-time grid correction factor is a parameter calculated by the power plant energy management system based on the real-time deviation of the grid frequency, in order to compensate for the prediction lag caused by relying solely on historical data. The processing procedure of the dynamic weighted power prediction model is as follows: The historical energy storage power data is weighted and summed with the corresponding dynamic weights, and the weighted summation result is superimposed with the real-time correction factor of the power grid to obtain the predicted energy storage power curve. The processing steps of the temperature control early warning and decision-making unit include: Based on the temperature evolution information output by the battery thermal characteristic digital twin unit, the temperature control buffer time required from the current moment until the predicted maximum temperature difference first reaches the preset maximum temperature difference critical value is calculated; The processing steps of the temperature control early warning and decision-making unit also include: Determine the temperature difference risk level, which is the ratio of the predicted maximum temperature difference to the critical value of the maximum temperature difference; Determine the time risk level, which is the ratio of the temperature control system response time to the temperature control buffer time; The thermal risk index is obtained by weighting and summing the temperature difference risk and the time risk.

2. The energy storage battery temperature control and early warning system according to claim 1, characterized in that, The dynamic weights are periodically and adaptively adjusted based on the prediction error of the previous period through an online learning algorithm.

3. The energy storage battery temperature control and early warning system according to claim 1, characterized in that, The processing steps of the digital twin unit for battery thermal characteristics include: S1. Based on the predicted energy storage power curve, the Coulomb integral method is used to process the data to obtain the predicted state of charge. S2. Combining the predicted state of charge, the battery equivalent circuit model is used to process the data to obtain the predicted terminal voltage and predicted current. S3. Based on the predicted current and predicted state of charge, the Bernardi battery heat generation model is used to process the data and obtain the heat generation rate of each individual cell. S4. Using the heat generation rate as the internal heat source and the operating power of the temperature control system as the active heat dissipation term, we input them into the single cell temperature evolution model for simulation processing to obtain the predicted temperature of each single cell; and based on the predicted temperature of each single cell, we obtain the predicted maximum temperature difference.

4. The energy storage battery temperature control and early warning system according to claim 1, characterized in that, The adjustment process of the proportional-integral control law is as follows: The deviation between the thermal risk index and the preset target risk value is obtained, and proportional and integral calculations are performed on the deviation. The calculation results are then superimposed with the preset base operating power to obtain the control power of the temperature control system.

Images