A highland energy storage battery monitoring method and system based on physical data double driving

Abstract

The application relates to a highland energy storage battery monitoring method and system based on physical data double driving, wherein the method comprises the following steps: firstly, a battery internal resistance physical model considering highland environment correction is constructed, internal resistance is predicted in real time, and failure prior probability is generated. Meanwhile, a support vector machine (SVM) and a random forest (RF) are used to construct a data-driven model, battery operation characteristics are extracted, a fault probability and uncertainty are output. On this basis, a Bayesian fusion method is used to dynamically adjust the weights of the physical model and the data-driven model, calculate the battery failure posterior probability, and quantify the total uncertainty. Finally, a decision threshold is set, when the posterior probability exceeds the threshold, a warning is triggered and protection measures are taken. The method effectively combines the theoretical advantages of the physical model and the actual data mining ability of the data-driven model, adapts to the highland environment, and provides protection for the safe and stable operation of the highland photovoltaic power station energy storage battery.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of reliability monitoring of energy storage batteries in plateau regions, and in particular to a plateau energy storage battery monitoring method and system based on physical data double driving. BACKGROUND

[0002] With the growing global demand for clean energy, photovoltaic power stations are increasingly deployed in plateau regions. Plateau regions have abundant solar energy resources, providing a unique condition for photovoltaic power generation. However, the particularity of the plateau environment poses a severe challenge to the reliability and performance of energy storage batteries. The low air pressure, thin air, large diurnal temperature difference, high ultraviolet intensity, and possible low temperature in plateau regions are all adverse environmental factors that seriously affect the chemical reaction process, thermal management, and mechanical structural stability of energy storage batteries. For example, low air pressure can cause the expansion of gases inside the battery, affecting the battery's sealing and internal structure; large diurnal temperature difference can cause thermal expansion and contraction of battery materials, accelerating battery aging. These environmental factors together make the failure rate of energy storage batteries in plateau regions significantly higher than in plain regions, thereby affecting the stability and power generation efficiency of the entire photovoltaic power station. Traditional battery monitoring methods are mainly based on a single physical model or data statistical method, which is difficult to accurately reflect the true state of the battery in the complex plateau environment. The physical model often deviates from the parameters when describing the internal electrochemical reactions of the battery due to the particularity of the plateau environment; while the data-driven model can capture some of the rules in the actual operation data, but it lacks understanding of the deep physical properties of the battery, and may not be accurate in predicting extreme working conditions in the plateau environment. Therefore, there is an urgent need for a reliability monitoring method for energy storage batteries that can consider both physical mechanisms and actual operation data, and adapt to the plateau environment, to ensure the safe and stable operation of plateau photovoltaic power stations. SUMMARY

[0003] The present application aims to provide a plateau energy storage battery monitoring method based on physical data double driving, which realizes accurate and real-time monitoring of the reliability of plateau energy storage batteries by deeply integrating data-driven methods with physical mechanism models and introducing a Bayesian fusion framework. The present application fully considers the influence of plateau environmental factors on battery performance, uses physical models to provide prior knowledge of battery reliability status, and at the same time uses data-driven models to mine potential fault features in actual operation data, and then dynamically fuses the advantages of the two models through Bayesian methods, quantifies uncertainty and adaptively adjusts the weight. At the decision level, potential faults are timely warned according to the fused posterior probability, triggering appropriate protective measures, thereby significantly improving the operational safety and stability of plateau energy storage batteries, reducing the risk of failure, and ensuring efficient power generation and stable operation of plateau photovoltaic power stations.

[0004] The specific scheme is as follows:

[0005] A monitoring method for high-altitude energy storage batteries based on dual physical data driving force includes the following steps:

[0006] S1. Construct a physical model of battery internal resistance that takes into account the high-altitude environment, predict internal resistance in real time and generate prior probability of failure.

[0007] S2. Use the Support Vector Machine (SVM) model and the Random Forest (RF) model to build a data-driven model, extract battery operating features, and output failure probability and uncertainty.

[0008] S3. Using the Bayesian fusion method, dynamically adjust the weights of the battery internal resistance physical model and the data-driven model, calculate the posterior probability of battery failure, and quantify the total uncertainty.

[0009] S4. Set a decision threshold. When the posterior probability exceeds the threshold, trigger an early warning and take protective measures.

[0010] Furthermore, step S1 specifically includes: processing and Bayesian fusion of the physical model output results.

[0011] (1) Output results and acquisition method of energy storage battery physical model

[0012] ① Thermal characteristic model of lithium-ion batteries based on internal resistance (high-altitude environment correction model)

[0013] Predicted output battery internal resistance R b :

[0014] R b =R b0 +K·(T s -T s0 )+ΔR b,alt

[0015] in:

[0016] R b It is the predicted internal resistance of the battery under the current operating conditions, reflecting the comprehensive resistance characteristics of the battery's internal electrodes and electrolyte, etc., and the unit is ohms (Ω).

[0017] R b0 The reference surface temperature T s0 The reference value of the battery internal resistance under standard atmospheric pressure was obtained through experimental calibration, and the unit is ohms (Ω).

[0018] K is the internal resistance temperature coefficient, which represents the rate at which the battery's internal resistance changes with surface temperature. It is obtained through experimental fitting and is measured in ohms per degree Celsius (Ω / ℃).

[0019] T s It is the actual surface temperature of the battery, which is collected in real time by a temperature sensor and is expressed in degrees Celsius (°C).

[0020] T s0 It is the reference surface temperature, which is usually selected as the surface temperature of the battery under standard test conditions, and the unit is degrees Celsius (°C).

[0021] ΔR b,alt This is a high-altitude environment correction term, used to quantify the impact of low air pressure at high altitudes on battery internal resistance. The calculation formula is:

[0022]

[0023] in:

[0024] β alt It is the plateau environment sensitivity coefficient, reflecting the sensitivity of the battery's internal resistance to changes in air pressure. It is obtained through experimental fitting and is dimensionless.

[0025] P a,sea It is the standard atmospheric pressure at sea level, usually taken as 101.325 kPa;

[0026] P a,alt This is the actual air pressure at the plateau, collected in real time by a pressure sensor, and the unit is kilopascal (kPa).

[0027] ② Optimization of digital twin parameters

[0028] Output optimized physical model parameters (e.g.) K * , Minimize the objective function J using the Particle Swarm Optimization (PSO) algorithm.

[0029]

[0030] in:

[0031] R b,measured (t i ) is at time t i Real-time battery internal resistance measurement, in ohms (Ω);

[0032] R b,model (t i ) is at time t i The predicted internal resistance of the battery is calculated based on the physical model, and the unit is ohms (Ω).

[0033] N is the number of data points collected, used to evaluate the deviation between the model's predicted values ​​and the actual measured values.

[0034] (2) Probabilistic processing of physical model output

[0035] Transform the physical model output into a probability distribution:

[0036] Battery internal resistance R b Assume that the distribution follows a normal distribution. Where μ phy =R b That is, the mean of the distribution is the predicted internal resistance value calculated by the physical model; the variance Determined by model error analysis, and taking into account the additional uncertainties brought about by the high-altitude environment:

[0037]

[0038] in:

[0039] This is the basic model error variance, obtained through statistical analysis of historical data. It reflects the degree of deviation between the predicted and actual values ​​of the physical model, and is expressed in ohm squares (Ω). 2 );

[0040] γ alt It is the uncertainty coefficient of the plateau environment, used to quantify the enhancing effect of the plateau environment on the uncertainty of model predictions. It is determined through experiments or historical data analysis and is dimensionless.

[0041] H represents the actual altitude, which is collected in real time by an altitude sensor or obtained through data from a Geographic Information System (GIS), and is measured in meters (m).

[0042] H0 is the reference altitude, which is usually a known altitude used for model calibration, and is expressed in meters (m).

[0043] (3) Prior probability construction in Bayesian fusion

[0044] The physical model output is used as a prior distribution, taking into account the impact of the high-altitude environment on the failure threshold:

[0045]

[0046] in:

[0047] P(fail) is the prior probability of battery failure, representing the likelihood of battery failure after considering the prediction results of the physical model and the correction for the high-altitude environment. Its value ranges from [0,1].

[0048] Φ(·) is the cumulative distribution function of the standard normal distribution, used to convert the standardized internal resistance deviation into a probability value, and is dimensionless;

[0049] R b,crit,alt This is the battery internal resistance failure threshold under high-altitude conditions. When the battery internal resistance exceeds this value, the battery is considered to be in a potentially failed state. The calculation formula is as follows:

[0050]

[0051] in:

[0052] R b,crit It is the failure threshold of the battery's internal resistance under standard conditions, determined through experiments or data provided by the manufacturer, and the unit is ohms (Ω).

[0053] δ alt It is the plateau environment failure threshold adjustment coefficient, used to quantify the impact of the plateau environment on the battery internal resistance failure threshold. It is determined through experiments or data analysis and is dimensionless.

[0054] Furthermore, step S2 specifically includes: processing and Bayesian fusion of the data-driven model output results.

[0055] (1) Output results of the data-driven model

[0056] ①Support Vector Machine (SVM)

[0057] Output failure probability p svm : Convert classification results into probabilities using Platt Scaling:

[0058]

[0059] in:

[0060] p svm The predicted battery failure probability based on the SVM model has a value range of [0,1].

[0061] f(x) is the decision function output of the SVM model, and its calculation formula is:

[0062]

[0063] in:

[0064] α i It is a Lagrange multiplier, determined through the SVM training process, and is dimensionless;

[0065] y i It is the class label (normal or faulty) of the training sample, with a value of +1 or -1;

[0066] K(x i ,x) is a kernel function used to map input features to a high-dimensional space and calculate the similarity between samples. Commonly used kernel functions include radial basis functions (RBF), which are dimensionless.

[0067] x i It is the feature vector of the i-th support vector;

[0068] x is the current input feature vector, which contains parameters such as battery voltage, current, and temperature after plateau correction;

[0069] b is the bias term, which is determined through the SVM training process and is dimensionless.

[0070] The SVM model converts the classification results into failure probabilities p using Platt Scaling. svm The formula is A and B are parameters of Platt Scaling, calibrated through cross-validation, used to convert the decision function output of SVM into dimensionless probability values.

[0071] ② Random Forest (RF)

[0072] Output failure probability P rf Class probabilities based on majority voting:

[0073]

[0074] in:

[0075] P rf It is a battery failure probability prediction value based on the RF model, and the value range is [0,1].

[0076] T is the number of decision trees in the random forest. A larger value is usually chosen to improve the stability and accuracy of the model. It is a positive integer.

[0077] I(·) is an indicator function that takes the value of 1 when the prediction result of decision tree t is a fault, and 0 otherwise;

[0078] f t (x) is the prediction result of the t-th decision tree, which determines whether the battery state corresponding to the current input feature vector x is faulty.

[0079] (2) Quantification of uncertainty in probability output

[0080] SVM variance estimation: based on confidence scores of support vector distances, considering the impact of the high-altitude environment.

[0081]

[0082] in:

[0083] It is the variance of the SVM output probability after considering the plateau environment correction, used to quantify the uncertainty of the model prediction, and the unit is dimensionless.

[0084] N sv It represents the number of support vectors, i.e., the number of samples selected during training to construct the decision boundary, and is a positive integer.

[0085] RF variance estimation: Based on the variance of the Bootstrap sample, considering the influence of the high-altitude environment:

[0086]

[0087] in:

[0088] It is the variance of the RF output probability after considering the plateau environment correction, used to quantify the uncertainty of the model prediction, and the unit is dimensionless.

[0089] It is the predicted failure probability output by the t-th decision tree, and its value ranges from [0,1].

[0090] (3) Likelihood function construction in Bayesian fusion

[0091] The data-driven model output is used as a likelihood function, with adjustments made for the high-altitude environment.

[0092]

[0093] in:

[0094] P(D|fail) is the likelihood probability of observing the current data D under the condition of battery failure. It is used to quantify the relationship between the output of the data-driven model and the battery failure state. The value range is [0,1].

[0095] It is a normal distribution, meaning that the probability value output by the data-driven model follows a mean of p. data,alt variance is The normal distribution;

[0096] p data,alt This is the predicted fault probability value of the fused data-driven model, calculated using the following formula:

[0097]

[0098] in:

[0099] It takes into account the plateau environment correction, and the weights of the SVM model reflect the reliability of the SVM model's prediction results.

[0100] The weights of the RF model after considering the plateau environment correction reflect the reliability of the RF model's prediction results.

[0101] Furthermore, step S3 specifically includes: the working mechanism of the Bayesian fusion model.

[0102] (1) Calculation of posterior probability

[0103] Based on Bayes' theorem:

[0104]

[0105] Substituting the prior P(fail) of the physical model after the plateau environment correction and the data-driven likelihood P(D|fail), we get:

[0106]

[0107] in:

[0108] P(fail|D) refers to the posterior probability of battery failure given observed data D, with a value range of [0,1]. It is the final reliability assessment result after integrating the physical model and the data-driven model.

[0109] P(D) is the marginal probability of data D, used for normalization to ensure that the sum of the posterior probabilities is 1;

[0110] Fusion Mean

[0111] Fusion variance

[0112] (2) Dynamic weight adjustment

[0113] The weighting coefficients are adaptively adjusted according to the environmental complexity Cenv, taking into account the high-altitude environmental factors:

[0114]

[0115] in:

[0116] ω phy,alt (C env The dynamic weights of the physical model are calculated after considering environmental complexity and adjustments for the plateau environment.

[0117] ω data,alt (C env The dynamic weights of the data-driven model are calculated after considering environmental complexity and adjustments for the plateau environment.

[0118] These are the initial weight values, usually determined based on prior knowledge or initial model performance evaluation, and are dimensionless.

[0119] λ is the weight adjustment coefficient, which controls the degree of influence of environmental complexity on the weight. It is determined through experiments or data analysis and is dimensionless.

[0120] C env To account for environmental complexity, multiple environmental factors such as temperature change rate, charge / discharge rate fluctuation, and air pressure change are comprehensively considered. The result is obtained through real-time data calculation and is dimensionless.

[0121] ζ alt It is the plateau environment weight adjustment coefficient, used to quantify the impact of the plateau environment on the dynamic adjustment of weights. It is determined through experiments or data analysis and is dimensionless.

[0122] (3) Propagation of uncertainty

[0123] Total uncertainty after fusion:

[0124]

[0125] in:

[0126] U fuse,alt It represents the total uncertainty after fusion, used to quantify the confidence interval of the final posterior probability, and is in dimensionless form.

[0127] This refers to the uncertainty of the physical model, specifically the width of the confidence interval for the predicted values ​​after considering the plateau environment correction.

[0128] This refers to the uncertainty of the data-driven model, specifically the confidence interval width of the data-driven model's predicted values ​​after considering the plateau environment correction.

[0129] σ phy,alt σ data,alt These are the standard deviations of the physical model and the data-driven model after considering the plateau environment correction, respectively, in dimensionless units.

[0130] These are the critical values ​​of the t-distribution for the corresponding physical model and data-driven model, respectively, used to construct confidence intervals, and compared with a given significance level α and degrees of freedom df. phy,alt df data,alt Related, dimensionless.

[0131] Furthermore, step S4 specifically includes:

[0132] Set a reasonable decision threshold α(C) env The calculated posterior probability P(fail|D) is compared with the decision threshold α(C). env If P(fail|D) exceeds the decision threshold, it is determined that the energy storage battery may have a failure risk, triggering an early warning signal and taking corresponding protective measures to ensure the safe operation of the energy storage battery.

[0133] A high-altitude energy storage battery monitoring system based on dual physical data driving mechanism, comprising sensors, a data acquisition module, a data processing module, a model calculation module, and a decision-making module, including:

[0134] ① Input layer: The data acquisition module collects environmental parameters (temperature, air pressure, altitude), electrical signals (battery voltage, current), and thermal parameters (battery surface temperature) monitored by the sensors in real time and sends them to the data processing module.

[0135] ②Physical Model Layer: The battery internal resistance Rb after high-altitude environment correction is calculated through the model calculation module, and the prior probability P(fail) is generated.

[0136] ③ Data-driven layer: Extracts features (voltage ripple rate γ) after correction for the plateau environment through the model calculation module. ripple Charging and discharging current fluctuation rate γ current Temperature gradient ▽T, output fault probability p data,alt .

[0137] ④ Bayesian Fusion Layer: Dynamically weights and fuses prior and likelihood data through the model calculation module to calculate the posterior probability P(fail|D). ⑤ Decision Layer: Makes a judgment through the decision module; if P(fail|D) > α(C)... env ,H), triggering early warning or protection measures, such as limiting charging and discharging power, activating the heating / heat dissipation system, etc., where α(C env H) represents the decision threshold considering the plateau environment.

[0138] The beneficial effects of this invention are as follows:

[0139] (1) The physical model of battery internal resistance is improved according to the characteristics of plateau environment. It can accurately reflect the change of battery internal resistance under plateau environment, reduce prediction deviation, improve the accuracy of failure prediction, and provide early warning to ensure safety.

[0140] (2) Improve the data-driven model by using support vector machine (SVM) and random forest (RF), mine the fault characteristics in the actual operation data, fully capture complex fault modes, improve the accuracy of fault probability prediction, and enhance the reliability of reliability assessment.

[0141] (3) The Bayesian fusion method is adopted to improve the model fusion method, which combines the advantages of physical model and data-driven model, dynamically adjusts weights, quantifies total uncertainty, accurately judges battery failure risk, and improves the accuracy and robustness of assessment.

[0142] (4) Set decision thresholds based on posterior probability and take early warning protection measures to effectively avoid faults, ensure the safe operation of energy storage batteries, and reduce the impact on the stability of photovoltaic power plants.

[0143] (5) Construct a reliability monitoring system for energy storage batteries in plateau areas. All modules work together to achieve comprehensive and real-time reliability monitoring, improve monitoring efficiency and accuracy, and provide strong support for the operation and management of plateau photovoltaic power stations. Attached Figure Description

[0144] Figure 1 A system block diagram of a method for monitoring the reliability of energy storage batteries in plateau regions is presented, showing the connection relationships between modules and the data flow.

[0145] Figure 2 This is a schematic diagram of the physical model of an energy storage battery, which details the physical relationship between the battery's internal resistance and the effects of temperature and high-altitude environment.

[0146] Figure 3 The flowchart shows the training process for data-driven models, illustrating how to train support vector machines and random forest models using historical data.

[0147] Figure 4 The flowchart illustrates the computation process of the Bayesian fusion model, showing the fusion process of the physical model and the data-driven model outputs, as well as the method for calculating the posterior probability. Detailed Implementation

[0148] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the present invention.

[0149] As shown in the figure, this invention provides a monitoring method for high-altitude energy storage batteries based on dual physical data driving, comprising the following steps:

[0150] 1. Monitoring System Setup

[0151] Construct an energy storage battery monitoring system that includes sensors, data acquisition modules, data processing modules, model calculation modules, and decision-making modules. These modules work collaboratively to achieve comprehensive monitoring of the reliability of energy storage batteries in high-altitude photovoltaic power stations.

[0152] 2. Application of physical models

[0153] (1) Model parameter acquisition

[0154] ① Experimental calibration: The energy storage battery was subjected to charge-discharge cycle tests in a laboratory environment. The internal resistance of the battery was measured under different temperature (e.g., 0℃, 20℃, 40℃) and air pressure (simulating different altitude air pressures on a plateau, such as 80kPa, 70kPa, 60kPa).

[0155] ② Data Fitting: The measured internal resistance data is correlated with conditions such as temperature and air pressure. Through data fitting, the internal resistance Rb0, internal resistance temperature coefficient K, and high-altitude environment sensitivity coefficient β of the energy storage battery under different temperature and air pressure conditions are obtained. alt Physical model parameters, etc.

[0156] (2) Model Calculation and Correction

[0157] ① Real-time calculation: During actual operation, the data acquisition module collects the battery surface temperature Ts and the actual air pressure at high altitude P in real time. a,alt Parameters such as these are transmitted to the data processing module.

[0158] ② Internal resistance prediction: The data processing module predicts the internal resistance based on the physical model formula R. b =R b0 +K·(T s -T s0 )+ΔR b,alt The collected data is substituted into the model to calculate the predicted value R of the battery internal resistance. b Among them, the plateau environment correction term ΔR b,alt According to the formula The calculations yielded a result used to correct for the impact of the high-altitude environment on the battery's internal resistance.

[0159] ③ Model function: The physical model is based on the physical characteristics of the battery and can accurately describe the relationship between the battery's internal resistance and temperature and high-altitude environment, providing a theoretical basis and prior knowledge for subsequent reliability assessment.

[0160] 3. Application of data-driven models

[0161] (1) Model Training

[0162] ① Data Collection: Collect a large amount of operational data from energy storage batteries in high-altitude photovoltaic power stations, including sample data under normal operation and fault conditions. This data covers electrical and thermal parameters such as battery voltage, current, temperature, and internal resistance, as well as corresponding environmental parameters (such as temperature, air pressure, and altitude).

[0163] ② Feature extraction: Feature extraction is performed on the collected data. The extracted features include voltage ripple rate, charge and discharge current fluctuation rate, temperature gradient, etc. These features can reflect the battery's operating status and potential fault information.

[0164] ③ Model Training: The extracted feature data is divided into training and test sets. Support Vector Machine (SVM) and Random Forest (RF) algorithms are used to train the model. During training, the algorithm parameters (such as the kernel function parameters of SVM and the number of decision trees in RF) are adjusted to enable the model to accurately predict battery failure probabilities. The trained model is validated and evaluated using the test set to ensure its accuracy and reliability.

[0165] (2) Model output and uncertainty quantification

[0166] ① Fault Probability Prediction: During actual monitoring, the data processing module inputs the real-time collected and processed data into the trained SVM and RF models. The SVM model converts the classification results into fault probabilities p using Platt Scaling. svm The formula is in Parameters A and B are calibrated using cross-validation. The RF model outputs the failure probability P based on the majority vote's class probability. rf The formula is Where I(·) is the indicator function.

[0167] ② Uncertainty Quantification: Based on the uncertainty quantification method of the model, calculate the variance of the output probabilities of SVM and RF. and SVM variance estimation is based on the confidence score of the support vector distance, taking into account the influence of the high-altitude environment. The formula is as follows: RF variance estimation is based on the variance of the Bootstrap sample, taking into account the influence of the high-altitude environment. The formula is as follows:

[0168] ③ Model function: Data-driven models can mine potential fault characteristics and patterns of batteries based on actual operating data, identify and predict complex fault modes that cannot be accurately described by physical models, and provide data support for reliability assessment.

[0169] 4. Bayesian Fusion and Decision Making

[0170] (1) Bayesian fusion computation

[0171] ① Construction of Prior Probability and Likelihood Function: The predicted internal resistance value and its uncertainty output by the physical model are transformed into a prior probability P(fail). The influence of the plateau environment on the failure threshold is considered. The formula is as follows: Where Φ(·) is the cumulative function of the standard normal distribution. Let be the battery internal resistance failure threshold under high-altitude conditions. The failure probability and its uncertainty output by the data-driven model are constructed as a likelihood function P(D|fail), and a high-altitude environment correction is considered; the formula is... in Weight

[0172] ② Posterior probability calculation: Based on Bayes' theorem, the prior probability P(fail) and the likelihood function P(D|fail) are fused to calculate the posterior probability P(fail|D). The formula is as follows: Substituting the prior a priori and data-driven likelihood of the physical model after plateau environment correction, the expanded result is:

[0173]

[0174] Among them, the fusion mean variance

[0175] ③ Dynamic weight adjustment: The weight coefficients are dynamically adjusted based on the environmental complexity Cenv and altitude H, using the following formula:

[0176]

[0177] Where ζ alt This is an adjustment factor for the weights of the plateau environment. As environmental complexity and altitude increase, the weights of the physical model are reduced to enhance the robustness of the data-driven approach.

[0178] ④ Uncertainty propagation: Calculate the total uncertainty after fusion, using the following formula:

[0179]

[0180] in

[0181] ⑤ Model Function: Bayesian fusion models can combine the advantages of physical models and data-driven models. By dynamically adjusting weights and quantifying uncertainties, they can achieve accurate assessment of the reliability status of energy storage batteries, thereby improving the accuracy and reliability of monitoring results.

[0182] (2) Decision-making and early warning

[0183] ① Set a decision threshold: Based on the historical operating data and reliability requirements of the energy storage batteries in the plateau photovoltaic power station, set a reasonable decision threshold α(C). env H). This threshold takes into account the impact of factors such as environmental complexity and altitude on battery reliability.

[0184] ② Decision making: The calculated posterior probability P(fail|D) is compared with the decision threshold α(C). env The system compares P(fail|D) with H. If P(fail|D) exceeds the decision threshold, it is determined that the energy storage battery may have a failure risk, triggering an early warning signal and taking corresponding protective measures, such as limiting charging and discharging power and starting the heating / cooling system, to ensure the safe operation of the energy storage battery.

[0185] ③ Model function: The decision module makes timely and accurate decisions based on the output of the Bayesian fusion model, ensuring the reliable operation of the energy storage battery in the high-altitude environment, reducing the risk of failure, and improving the safety and stability of the system.

[0186] The above description is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several adjustments and improvements without departing from the core concept of the present invention, and these adjustments and improvements should also be considered within the scope of protection of the present invention.

Claims

1. A monitoring method for high-altitude energy storage batteries based on dual-drive physical data, characterized in that, Includes the following steps: S1. Construct a physical model of battery internal resistance that takes into account the high-altitude environment, predict internal resistance in real time and generate prior probability of failure. S2. Use the Support Vector Machine (SVM) model and the Random Forest (RF) model to build a data-driven model, extract battery operating features, and output failure probability and uncertainty. S3. Using the Bayesian fusion method, dynamically adjust the weights of the battery internal resistance physical model and the data-driven model, calculate the posterior probability of battery failure, and quantify the total uncertainty. S4. Set a decision threshold. When the posterior probability exceeds the threshold, trigger an early warning and take protective measures. In step S1, the formula for calculating the real-time predicted internal resistance is: Among them, R b This is the predicted internal resistance of the battery under current operating conditions. Reference surface temperature The reference value of battery internal resistance under standard atmospheric pressure; K is the temperature coefficient of internal resistance; This is the actual surface temperature of the battery; It is the reference surface temperature; This is the plateau environment correction term, and its calculation formula is: in, It is the sensitivity coefficient to the plateau environment; That is the standard atmospheric pressure at sea level; This is the actual air pressure at the plateau; Output optimized physical model parameters: Minimize the objective function J using the particle swarm optimization algorithm: in, At any moment Real-time acquisition of battery internal resistance measurement values; At any moment The predicted internal resistance of the battery is calculated based on the physical model; N is the number of data points collected, used to evaluate the deviation between the model's predicted value and the actual measured value. In step S3, the method for calculating the total uncertainty is as follows: The weighting coefficients are adaptively adjusted according to the environmental complexity Cenv, taking into account the high-altitude environmental factors: in, The dynamic weights of the physical model are adjusted to account for environmental complexity and high-altitude environment. The dynamic weights of the data-driven model are adjusted to account for environmental complexity and high-altitude environment. , These are the initial weight values; This is the weighting adjustment factor; For environmental complexity; It is the plateau environment weighting adjustment coefficient; The formula for calculating the total uncertainty after fusion is: in, It is the total uncertainty after fusion; It is the uncertainty of the physical model; It is the uncertainty of the data-driven model; , These are the standard deviations of the physical model and the data-driven model, respectively, after considering the plateau environment correction. , These are the critical values ​​of the t-distribution for the corresponding physical model and data-driven model, respectively, used to construct the confidence region.

2. The plateau energy storage battery monitoring method based on dual physical data driving according to claim 1, characterized in that, In step S1, the method for calculating the prior probability of failure is as follows: Transform the physical model output into a probability distribution: Battery internal resistance R b Assume that the distribution follows a normal distribution. ,in That is, the mean of the distribution is the predicted internal resistance value calculated by the physical model; the variance Determined by model error analysis, and taking into account the additional uncertainties brought about by the high-altitude environment: in: It is the variance of the basic model error; H is the uncertainty coefficient of the plateau environment; H is the actual altitude. H0 is the reference altitude; The physical model output is used as a prior distribution, taking into account the impact of the high-altitude environment on the failure threshold: in, It is the prior probability of battery failure, with a value range of [0,1]. It is the cumulative distribution function of the standard normal distribution; This is the battery internal resistance failure threshold under high-altitude conditions. When the battery internal resistance exceeds this value, the battery is considered to be in a potentially failed state. The calculation formula is as follows: in, It is the failure threshold of the battery's internal resistance under standard conditions; It is the adjustment coefficient for the failure threshold in high-altitude environments.

3. The plateau energy storage battery monitoring method based on dual physical data driving according to claim 2, characterized in that, In step S2, the fault probability output by the SVM model is: in, The predicted battery failure probability based on the SVM model has a value range of [0,1]; f(x) is the decision function output of the SVM model, and its calculation formula is as follows: in, They are Lagrange multipliers; It is the class label of the training sample, with a value of +1 or -1; It is a kernel function; x i is the feature vector of the i-th support vector; x is the feature vector of the current input; b is the bias term.

4. The plateau energy storage battery monitoring method based on dual physical data driving according to claim 3, characterized in that, In step S2, the fault probability output by the RF model is: Among them, P rf is the predicted battery failure probability based on the RF model, with a value range of [0,1]; T is the number of decision trees in the random forest, taking a positive integer; It is an indicator function, which takes the value of 1 when the prediction result of decision tree t is a fault, and 0 otherwise; It is the prediction result of the t-th decision tree, which determines whether the battery state corresponding to the current input feature vector x is faulty.

5. The plateau energy storage battery monitoring method based on dual physical data driving according to claim 4, characterized in that, In step S2, the uncertainty of the probability output is quantified, specifically as follows: Calculate the SVM variance estimate based on the confidence score of the support vector distance, taking into account the impact of the high-altitude environment: in, It is the variance of the SVM output probability after considering the high-altitude environment correction; It is the number of support vectors; Calculate the variance estimate of the RF sample, based on the variance of the Bootstrap sample, taking into account the influence of the high-altitude environment: in, It is the variance of the RF output probability after considering the high-altitude environment correction; It is the predicted failure probability output by the t-th decision tree, and its value ranges from [0,1].

6. The plateau energy storage battery monitoring method based on dual physical data drive according to claim 5, characterized in that, In step S3, the method for calculating the posterior probability of battery failure is as follows: First, the predicted internal resistance value and its uncertainty output by the physical model are converted into a prior probability P(fail), as shown in the formula: in, It is the cumulative function of the standard normal distribution; Then, the failure probability and its uncertainty output by the data-driven model are constructed into a likelihood function P(D|fail), and considering the high-altitude environment correction, the formula is: in, It is the likelihood probability of observing the current data D under the condition of battery failure, and the value range is [0,1]. It is a normal distribution, meaning that the probability values ​​output by the data-driven model follow a mean of . variance is The normal distribution; This is the predicted fault probability value of the fused data-driven model, calculated using the following formula: in, This refers to the weights of the SVM model after considering the plateau environment correction; This refers to the weights of the RF model after considering the plateau environment correction; Next, based on Bayes' theorem, the prior probability P(fail) and the likelihood function P(D|fail) are fused to calculate the posterior probability P(fail|D), as shown in the formula: Substituting the prior P(fail) of the physical model after plateau environment correction and the data-driven likelihood P(D|fail), we get: Where: P(fail|D) refers to the posterior probability of battery failure given observed data D, with a value range of [0,1]; P(D) is the marginal probability of data D, used for normalization to ensure that the sum of the posterior probabilities is 1; Fusion Mean ; Fusion variance .

7. The plateau energy storage battery monitoring method based on dual physical data drive according to claim 6, characterized in that, In step S4, a reasonable decision threshold is set. The calculated posterior probability P(fail|D) is compared with the decision threshold. If P(fail|D) exceeds the decision threshold, it is determined that the energy storage battery may have a failure risk, triggering an early warning signal and taking corresponding protective measures to ensure the safe operation of the energy storage battery.

8. A high-altitude energy storage battery monitoring system based on dual physical data driving, characterized in that, To implement the method of any one of claims 1-7, comprising: Input layer: Real-time acquisition of environmental parameters, electrical signals, and thermal parameters; Physical model layer: Calculate the battery internal resistance R after correction for high-altitude environment. b Generate the prior probability P(fail); Data-driven layer: Extracts features corrected for the high-altitude environment, including voltage ripple rate. Charging and discharging current fluctuation rate Temperature gradient Output failure probability ; Bayesian fusion layer: dynamically weights and fuses priors and likelihoods, and calculates the posterior probability P(fail|D); Decision-making level: If This triggers early warning or protective measures, among which To take into account the decision threshold of the plateau environment.

Images