High and low temperature explosion-proof test box for battery

Abstract

The invention relates to the technical field of battery explosion prevention, in particular to a battery high-low temperature explosion-proof test box, which comprises a data acquisition and detection module used for monitoring the change of parameters inside and outside the test box in real time through a sensor to obtain monitoring data; the data preprocessing module is used for cleaning, denoising and formatting the collected monitoring data; the prediction model building module is used for building a lasso regression model optimized by an alternating direction multiplier algorithm on the basis of historical data, and analyzing a relationship between external environment parameters and own operation parameters to obtain a lasso training model; and the decision execution module takes the monitoring data as the input of the lasso training model, outputs lasso data, and adjusts the operation mode and parameters of the test box according to the lasso data, so that the test box reaches the optimal working state. The lasso regression model can automatically select important features, and the ADMM algorithm can efficiently process the features, so that the model keeps concise, and meanwhile, the prediction accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery explosion-proof, and in particular to a battery high and low temperature explosion-proof test box. Background Art

[0002] With the widespread application of batteries in various fields, especially in electric vehicles, energy storage systems and various portable electronic devices, battery safety has become a crucial issue.

[0003] In actual use, batteries face various complex environmental conditions, among which high and low temperature environments have a significant impact on battery performance and safety. High temperature may intensify the chemical reaction inside the battery, causing the battery to overheat, swell, or even catch fire and explode; low temperature will reduce the performance of the battery, such as capacity decay, reduced charging and discharging efficiency, etc. It may also cause changes in the internal structure of the battery and deterioration of material properties, increase the risk of short circuits and other faults, and cause serious consequences such as explosions.

[0004] In order to ensure the safety and reliability of batteries under different temperature environments, the battery high and low temperature explosion-proof test chamber came into being. It can simulate various extreme high and low temperature environments, accurately test and evaluate the performance changes, stability and explosion-proof performance of batteries under these conditions, provide key data support and guarantee for battery research and development, production and quality control, help companies improve battery design and optimize production processes, thereby effectively reducing the possibility of safety accidents such as battery explosions in actual use, protecting the lives and property of users and the stable operation of various equipment, meeting the growing stringent requirements for battery safety performance, and promoting the healthy and sustainable development of battery-related industries.

[0005] In the prior art, the battery high and low temperature explosion-proof test chamber cannot adjust its own parameters and modes according to changes in the external environment, and most of the existing control systems work based on preset fixed parameters. Before the test begins, the operator needs to manually set parameters such as temperature, humidity, and test time. During the test, the control system controls according to these preset values ​​and will not automatically adjust according to dynamic changes in the external environment. The present invention provides a battery high and low temperature explosion-proof test chamber that can automatically adjust the operating mode and parameters according to actual needs, thereby improving the operating efficiency of the equipment. Summary of the invention

[0006] The present invention provides a battery high and low temperature explosion-proof test box, which is used to solve the defect in the prior art that the parameters and modes thereof cannot be adjusted according to the changes in the external environment.

[0007] The present invention provides a battery high and low temperature explosion-proof test box, comprising: a data acquisition and detection module, which is used for real-time monitoring of changes in parameters inside and outside the test box through sensors to obtain monitoring data.

[0008] The data preprocessing module cleans, denoises and formats the collected monitoring data.

[0009] The prediction model building module builds a lasso regression model optimized by the alternating direction multiplier algorithm based on historical data, analyzes the relationship between external environmental parameters and its own operating parameters, and obtains the lasso training model.

[0010] The decision-making execution module takes the monitoring data as the input of the lasso training model, outputs the lasso data, and adjusts the operating mode and parameters of the test chamber according to the lasso data to make the test chamber reach the optimal working state.

[0011] The present invention provides a battery high and low temperature explosion-proof test box, and the data preprocessing module includes:

[0012] The deduplication unit is used to identify and delete duplicate data in the monitoring data using the deduplication function in the database query statement.

[0013] The missing value processing unit is used to fill the missing values ​​of monitoring data using linear interpolation.

[0014] The error value correction unit is used to correct the error values ​​of monitoring data in batches using the replacement function.

[0015] The standardization processing unit converts the monitoring data into a distribution with a mean of 0 and a standard deviation of 1, and scales the monitoring data to a specific range.

[0016] The present invention provides a battery high and low temperature explosion-proof test box, comprising: a lasso regression model is:

[0017]

[0018] Among them, y i is the dependent variable, β 0 , β 1 , …, β p is the regression coefficient, x i1 , x i2 , …x ip is the external environment parameter, n is the number of samples, p is the number of independent variables, λ is the regularization parameter of lasso regression, and the regularization term is the L1 regularization term.

[0019] The present invention provides a battery high and low temperature explosion-proof test box, comprising: in a prediction model building module, the relationship between external environmental parameters and its own operating parameters is expressed as follows:

[0020]

[0021] Where, △T = |Tint -T ext | is the difference between the internal and external temperatures, △H = |H int -H ext ∣ is the difference between internal and external humidity, △P = ∣P int -P ext ∣ is the difference between internal and external pressure, P heat is the heating or cooling power, F speed is the fan speed, α, β, γ, δ, and ε are weight coefficients determined according to experimental data, reflecting the relative importance of each parameter to OESI.

[0022] The present invention provides a battery high and low temperature explosion-proof test box, including: a prediction model building module, using historical data to train a lasso regression model including:

[0023] Set the regularization parameter unit to control the strength of the L1 regularization term.

[0024] Initialize the model parameter unit and initialize the historical data.

[0025] Construct a loss function unit to construct the loss function of the lasso regression model, measure the difference between the model prediction results and the actual observation values, and guide the direction and speed of model training.

[0026] Select the Algorithm unit and choose the ADMM algorithm as the algorithm for optimizing the lasso regression model.

[0027] The iterative optimization unit is used to substitute historical parameters into the lasso regression model and perform multiple optimizations according to the iterative rules of the ADMM algorithm.

[0028] The model evaluation unit uses regression model evaluation indicators to evaluate the model prediction performance and compare the differences between the model prediction results and the actual observations.

[0029] The present invention provides a battery high and low temperature explosion-proof test box, comprising: in a loss function construction unit, the formula of the loss function is expressed as:

[0030]

[0031] In the formula, w is the coefficient vector of the characteristic variable, b is the intercept term, and x i is the feature vector of the i-th sample, y i is the true value of the i-th sample, N is the number of samples, and M is the number of features.

[0032] The present invention provides a battery high and low temperature explosion-proof test box, comprising: in a selection algorithm unit, the mathematical model of the ADMM algorithm is:

[0033] minf 1(x 1 )+f 2 (x 2 )

[0034] A 1 x 2 +A 1 x 2 +b

[0035] In the formula, x 1 ∈R m and x 2 ∈R n is the decision variable, A 1 ∈R p×m , A 2 ∈R p×n and b∈R p The coefficient matrix representing the equality constraints, f 1 and f 2 is a suitable closed convex function.

[0036] The present invention provides a battery high and low temperature explosion-proof test box, comprising: in a selection algorithm unit, an augmented Lagrangian function formula constructed for an ADMM algorithm mathematical model is expressed as:

[0037] L p (x 1 , x 2 ,λ)=f 1 (x 1 )+f 2 (x 2 )+λ(A 1 x 1 +A 2 x 2 ―b)+φ

[0038]

[0039] Where λ is the Lagrange multiplier and ρ is the penalty factor.

[0040] The present invention provides a battery high and low temperature explosion-proof test box, comprising: in a selection algorithm unit, the iterative process of ADMM is:

[0041]

[0042] In the formula, argmin is the decision variable value of the optimal solution of the objective function, k is the number of iterations, is the Lagrange multiplier.

[0043] The present invention provides a battery high and low temperature explosion-proof test box, comprising: in a model evaluation unit, using a regression model evaluation index to evaluate the model prediction performance, including:

[0044] The mean square error is expressed as:

[0045]

[0046] In the formula, y i is the true value of the i-th sample, which is the target value that the model attempts to predict or estimate. is the predicted value of the i-th sample, which is the output value calculated by the model based on the input features, and n is the sample size.

[0047] The accuracy is calculated as follows:

[0048]

[0049] In the formula, TP is a true positive example and FP is a false positive example.

[0050] The F1 score is calculated as:

[0051]

[0052] In the formula, P is the precision and R is the recall.

[0053] The battery high and low temperature explosion-proof test box provided by the present invention solves the problem that the test box cannot adjust its own parameters according to the external environment by constructing a lasso regression model optimized by an alternating direction multiplier algorithm based on historical data, and the beneficial effects achieved are: lasso regression can effectively perform feature selection by shrinking the coefficient to zero, that is, automatically selecting the set of independent variables that have the most explanatory power for the dependent variable. When there is a high correlation between the predicted variables, lasso regression will select one of the factors and shrink the other factors to zero, thereby solving the problem of multicollinearity. Through feature selection and reducing model complexity, lasso regression can improve the prediction accuracy and generalization performance of the model. Due to the existence of L1 regularization, lasso regression tends to compress the coefficients of unimportant independent variables to zero, thereby automatically performing feature selection. This makes lasso regression very useful in high-dimensional data sets, which can reduce the risk of overfitting and improve the interpretability of the model.

[0054] The lasso regression model can automatically select important features, while the ADMM algorithm can efficiently process these features, so that the model can maintain simplicity while improving the accuracy of prediction. In addition, in a distributed computing environment, the ADMM algorithm can effectively distribute computing tasks to multiple processors or nodes, thereby improving the overall computing efficiency, which is of great significance for processing large-scale data sets. The ADMM algorithm is highly flexible and scalable and can adapt to different types of constraints and objective functions. This enables the lasso regression model to form a more powerful hybrid optimization strategy when combined with other optimization algorithms. At the same time, as the data scale increases, the ADMM algorithm can maintain stable performance, which makes the lasso regression model more scalable in a big data environment.

[0055] Through intelligent scheduling and optimization, the equipment can automatically adjust the operation mode and parameters according to actual needs, thereby improving the operation efficiency of the equipment. This helps to shorten the test cycle and improve the test efficiency. Intelligent scheduling and optimization technology can predict and adjust according to the operation status and performance trend of the equipment. This helps to timely discover and deal with potential performance problems and improve the overall performance of the equipment. Through intelligent scheduling and optimization, the equipment can operate at the lowest energy consumption while ensuring performance. This helps to reduce the energy consumption cost of the equipment and reduce the maintenance cost caused by equipment failure. Intelligent scheduling and optimization technology can reduce the operation cost of the equipment and improve the economy and sustainability of the equipment. This helps enterprises achieve energy conservation and emission reduction goals and enhance their social responsibility and brand image. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 It is a module diagram of a battery high and low temperature explosion-proof test box provided by an embodiment of the present invention;

[0058] Figure 2 It is a module diagram of a battery high and low temperature explosion-proof test box prediction model construction module provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Combine the following Figure 1-Figure 2 The invention discloses a battery high and low temperature explosion-proof test box.

[0061] Figure 1 It is a module diagram of a battery high and low temperature explosion-proof test box provided in an embodiment of the present invention.

[0062] like Figure 1 As shown, the battery high and low temperature explosion-proof test box provided by the embodiment of the present invention mainly includes the following steps:

[0063] The data acquisition and detection module is used to monitor the changes in the temperature, humidity and pressure inside the test chamber in real time through temperature sensors, humidity sensors and pressure sensors to obtain monitoring data. The sensors transmit the monitored data to the data acquisition system via wired or wireless means.

[0064] The data preprocessing module cleans, denoises and formats the collected data to ensure the accuracy and consistency of the data.

[0065] Remove duplicate data units, use database query statements or deduplication functions in programming languages ​​to identify and delete duplicate data records.

[0066] To handle missing value units, for time series data, you can use linear interpolation, polynomial interpolation and other methods to fill in missing values. Calculate the mean of the column (or row) where the missing value is located and use the mean to replace the missing value. However, it should be noted that this method may introduce bias, especially when the missing values ​​are not randomly distributed. If there are many missing values ​​or the missing pattern has an important impact on the analysis, you can choose to delete the records with missing values.

[0067] Correct the erroneous cells in the data and use the Replace function to batch correct these erroneous values ​​to ensure that the data format is consistent with the expected one.

[0068] Normalizes the data to a distribution with a mean of 0 and a standard deviation of 1. This helps with data that has different dimensions and distributions. Scales the data to a specific range. This helps with data that has different dimensions but you want to keep the relative size relationship.

[0069] like Figure 2As shown, the prediction model building module uses historical data to train the lasso regression model optimized by the alternating direction multiplier algorithm to analyze the relationship between external environmental parameters and its own operating parameters.

[0070] The external environment includes external temperature, external humidity, and external pressure, and the operating parameters of the test chamber itself include internal temperature, internal humidity, internal pressure, heating / cooling power, and fan speed.

[0071] The relationship between external environment parameters and their own operating parameters is expressed as:

[0072]

[0073] Where, △T = |T int -T ext | is the difference between the internal and external temperatures, △H = |H int -H ext ∣ is the difference between internal and external humidity, △P = ∣P int -P ext ∣ is the difference between internal and external pressure, P heat is the heating or cooling power, F speed is the fan speed, α, β, γ, δ, and ε are weight coefficients determined according to experimental data, reflecting the relative importance of each parameter to OESI.

[0074] Lasso regression is a regularization method for linear regression, similar to ridge regression. Lasso regression solves the problem of multicollinearity by introducing the L1 regularization term and has the ability of feature selection. Unlike ridge regression, lasso regression tends to make some regression coefficients become exactly zero, thus achieving automatic feature selection. When looking for a linear model that minimizes the sum of squared errors, lasso regression adds an L1 penalty term to restrict the coefficients of the model. This L1 penalty term is the sum of the absolute values ​​of the regression coefficients, which is different from the L2 penalty term (sum of squared regression coefficients) used in ridge regression. Due to the characteristics of the L1 penalty term, when the penalty value is large enough, the estimated values ​​of some regression coefficients will be exactly shrunk to 0, thereby achieving the function of feature selection.

[0075] The lasso regression model can be expressed as:

[0076]

[0077] Where: y i is the dependent variable, β 0 , β 1 , …, β p is the regression coefficient, x i1 , x i2 , …x ipis the independent variable, n is the number of samples, p is the number of independent variables, and λ is the regularization parameter of lasso regression. is the L1 regularization term, which encourages the regression coefficients to take smaller values ​​and compresses some coefficients to zero.

[0078] Training a lasso regression model using historical data involves:

[0079] Set the regularization parameter. In lasso regression, the regularization parameter λ is a very important hyperparameter. It is used to control the strength of the L1 regularization term, which is the sum of the absolute values ​​of the feature variable coefficients multiplied by λ. The choice of λ directly affects the sparsity and generalization ability of the model. If λ is too large, the model may be too simple, resulting in underfitting. If λ is too small, the model may be too complex and contain many unnecessary features, resulting in overfitting.

[0080] In order to select a suitable λ, a cross-validation method (such as K-fold cross-validation) is usually used. Cross-validation divides the data set into a training set and a validation set (or test set), trains the model on the training set, and evaluates the model performance on the validation set, thereby selecting the optimal λ value.

[0081] Initialize model parameters. Before model training begins, the parameters of the lasso regression model need to be initialized. These parameters include the intercept term (also called the bias term or constant term) and the coefficients of each feature variable. These parameters will be continuously updated in the subsequent iterative optimization process to minimize the loss function.

[0082] Construct the loss function. The loss function of lasso regression consists of two parts: prediction error and L1 regularization term. The prediction error usually uses mean square error (MSE) or other similar metrics to measure the difference between the model prediction value and the true value. The L1 regularization term is the sum of the absolute values ​​of the feature variable coefficients multiplied by λ, which is used to introduce sparsity and make the coefficients of some features zero, thereby achieving feature selection.

[0083] The formula of the loss function is expressed as:

[0084]

[0085] In the formula, w is the coefficient vector of the characteristic variable, b is the intercept term, and x i is the feature vector of the i-th sample, y i is the true value of the i-th sample, N is the number of samples, and M is the number of features.

[0086] The ADMM algorithm is selected as the optimization algorithm. The alternating direction multiplier (ADMM) algorithm has the advantages of strong flexibility, good convergence and excellent robustness. Its basic idea is to split the original problem into multiple sub-problems, and then coordinate each sub-problem to obtain a global solution after solving it separately. The ADMM algorithm has been widely recognized and applied in distributed optimization scheduling.

[0087] The mathematical model of the ADMM algorithm is:

[0088] minf 1 (x 1 )+f 2 (x 2 )

[0089] A 1 x 2 +A 1 x 2 +b

[0090] In the formula, x 1 ∈R m and x 2 ∈R n is the decision variable, A 1 ∈R p×m , A 2 ∈R p×n and b∈R p The coefficient matrix representing the equality constraints, f 1 and f 2 is a suitable closed convex function.

[0091] The augmented Lagrangian function formula for the ADMM algorithm mathematical model is expressed as:

[0092] L p (x 1 , x 2 ,λ)=f 1 (x 1 )+f 2 (x 2 )+λ(A 1 x 1 +A 2 x 2 ―b)+φ

[0093]

[0094] Where λ is the Lagrange multiplier and ρ is the penalty factor.

[0095] The iterative process of ADMM is:

[0096]

[0097] In the formula, argmin is the decision variable value of the optimal solution of the objective function, k is the number of iterations, is the Lagrange multiplier.

[0098] The subproblem solution is updated and the final iteration is performed until convergence. Whether the algorithm converges is determined based on whether the original residual r and the dual residual s of each iteration meet the convergence accuracy value.

[0099]

[0100] In the formula, ε prim is the convergence accuracy value of the original residual, ε dual is the convergence accuracy value of the dual residual.

[0101] Iterative optimization is an iterative optimization of model parameters. In each iteration, the algorithm calculates the gradient or directional derivative of the loss function under the current parameters and updates the parameter value based on this information. The iterative process will continue until the convergence condition is met or the maximum number of iterations is reached. The convergence condition is determined by the change in the loss function. If the change in the loss function is less than a preset threshold after multiple consecutive iterations, the model is considered to have converged, and the iterative process can be stopped to obtain a trained optimized model.

[0102] The model evaluation unit uses regression model evaluation indicators to evaluate the model prediction performance:

[0103] The mean square error is used to reflect the deviation between the predicted value and the true value. It is sensitive to outliers. The calculation formula of the mean square error is:

[0104]

[0105] In the formula, yi is the true value of the i-th sample, which is the target value that the model attempts to predict or estimate. is the predicted value of the i-th sample, which is the output value calculated by the model based on the input features, and n is the sample size.

[0106] Coefficient of determination R 2 , which is used to evaluate the value range between 0 and 1. The closer it is to 1, the better the model performance. The calculation formula of the determination coefficient is:

[0107]

[0108] In the formula, yi is the true value of the i-th sample, which is the target value that the model attempts to predict or estimate. is the predicted value of the i-th sample, which is the output value calculated by the model based on the input features.

[0109] Adjusted R 2, which is used to penalize the model complexity for the number of independent variables p in the model. The calculation formula of the adjusted determination coefficient is:

[0110]

[0111] In the formula, p is the number of independent variables, R 2 is the coefficient of determination, and n is the sample size.

[0112] The precision reflects the reliability of the model's prediction of positive samples, and the calculation formula is:

[0113]

[0114] In the formula, TP is a true positive example and FP is a false positive example.

[0115] The recall rate reflects the model's ability to identify positive samples, and the calculation formula is:

[0116]

[0117] Where TP is a true positive example and FN is a false negative example.

[0118] The harmonic average of the F1 score precision and recall can comprehensively reflect the performance of the model. The calculation formula is:

[0119]

[0120] In the formula, P is the precision and R is the recall.

[0121] The decision-making execution module is responsible for using the real-time monitored data as input information for the lasso training model. These monitoring data may cover a variety of key environmental parameters such as temperature, humidity, and pressure in the test chamber. Once the data is input into the lasso training model, the model will use its powerful data analysis and prediction capabilities to identify potential optimization space for the current test chamber operation status. The intelligent scheduling system will automatically and intelligently adjust the operation mode and various parameters of the test chamber based on the output results of the lasso training model. This process may involve fine-tuning the heating system, cooling system, humidification / dehumidification system, and airflow control system. By analyzing the difference between the monitoring data and the model prediction results in real time, the system can dynamically optimize the working state of the test chamber to ensure that it always operates in the optimal or near-optimal range. The intelligent scheduling system also has the ability to self-learn and optimize. With the continuous accumulation of monitoring data and the continuous iterative update of the lasso training model, the system can gradually improve the accuracy and efficiency of its decision execution, thereby more accurately meeting the needs of various complex test scenarios.

[0122] The lasso regression model can automatically select important features, while the ADMM algorithm can efficiently process these features, so that the model can maintain simplicity while improving the accuracy of prediction. In a distributed computing environment, the ADMM algorithm can effectively distribute computing tasks to multiple processors or nodes, thereby improving the overall computing efficiency, which is of great significance for processing large-scale data sets. The ADMM algorithm is highly flexible and scalable and can adapt to different types of constraints and objective functions. This enables the lasso regression model to form a more powerful hybrid optimization strategy when combined with other optimization algorithms. At the same time, as the data scale increases, the ADMM algorithm can maintain stable performance, which makes the lasso regression model more scalable in a big data environment. The ADMM algorithm has parallel computing capabilities and can accelerate the optimization process. In the lasso regression model, the ADMM algorithm can take advantage of this feature to accelerate the model training process by processing multiple sub-problems in parallel. The lasso regression model adds an L1 penalty term to limit the sum of the absolute values ​​of the model coefficients, thereby avoiding the situation where the model performs well on the training set but performs poorly on the test set, that is, overfitting. When solving the lasso regression model, the ADMM algorithm can maintain the generalization performance of the model and avoid overfitting.

[0123] Through intelligent scheduling and optimization, the equipment can automatically adjust the operation mode and parameters according to actual needs, thereby improving the operation efficiency of the equipment. This helps to shorten the test cycle and improve the test efficiency. Intelligent scheduling and optimization technology can predict and adjust according to the operation status and performance trend of the equipment. This helps to timely discover and deal with potential performance problems and improve the overall performance of the equipment. Through intelligent scheduling and optimization, the equipment can operate at the lowest energy consumption while ensuring performance. This helps to reduce the energy consumption cost of the equipment and reduce the maintenance cost caused by equipment failure. Intelligent scheduling and optimization technology can reduce the operation cost of the equipment and improve the economy and sustainability of the equipment. This helps enterprises achieve energy conservation and emission reduction goals and enhance their social responsibility and brand image.

[0124] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative effort.

[0125] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0126] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A battery high and low temperature explosion-proof test box, characterized in that: include: The data acquisition and detection module is used to monitor the changes of parameters inside and outside the test chamber in real time through sensors to obtain monitoring data; A data preprocessing module cleans, removes noise and formats the collected monitoring data to obtain preprocessed data; The prediction model building module builds a lasso regression model optimized by the alternating direction multiplier algorithm based on historical data, analyzes the relationship between external environmental parameters and its own operating parameters, and obtains a lasso training model; The decision execution module uses the monitoring data as the input of the lasso training model, outputs the lasso data, and adjusts the operation mode and parameters of the test box according to the lasso data to make the test box reach the optimal working state.

2. The battery high and low temperature explosion-proof test box according to claim 1, characterized in that: The data preprocessing module comprises: Removing duplicate data units, for identifying and deleting duplicate data in the monitoring data using a deduplication function in a database query statement; A missing value processing unit, used for filling the missing values ​​of the monitoring data by using linear interpolation; an error value correction unit, used to correct the error values ​​of the monitoring data in batches using a replacement function; The standardization processing unit converts the monitoring data into a distribution with a mean of 0 and a standard deviation of 1, and scales the monitoring data to a specific range.

3. The battery high and low temperature explosion-proof test box according to claim 1, characterized in that: The lasso regression model is: Among them, y i is the dependent variable, β0, β1, …, β p is the regression coefficient, x i1 , x i2 , …x ip is the external environment parameter, n is the number of samples, p is the number of independent variables, λ is the regularization parameter of lasso regression, and the regularization term is the L1 regularization term.

4. The battery high and low temperature explosion-proof test box according to claim 1, characterized in that: In the prediction model building module, the relationship between the external environment parameters and the self-operation parameters is expressed as follows: Where, △T = |T int -T ext | is the difference between the internal and external temperatures, △H = |H int -H ext ∣ is the difference between internal and external humidity, △P = ∣P int -P ext ∣ is the difference between internal and external pressure, P heat is the heating or cooling power, F speed is the fan speed, α, β, γ, δ, and ε are weight coefficients determined according to experimental data, reflecting the relative importance of each parameter to OESI.

5. The battery high and low temperature explosion-proof test box according to claim 1, characterized in that: The prediction model building module uses historical data to train the lasso regression model, including: Set the regularization parameter unit to control the strength of the L1 regularization term; Initializing a model parameter unit to initialize the historical data; Construct a loss function unit to construct the loss function of the lasso regression model, measure the difference between the model prediction results and the actual observation values, and guide the model training direction and speed; Select the Algorithm unit and select the ADMM algorithm as the algorithm for optimizing the lasso regression model; Iterative optimization unit, used to substitute historical parameters into the lasso regression model and perform multiple optimizations according to the iterative rules of the ADMM algorithm; The model evaluation unit uses regression model evaluation indicators to evaluate the model prediction performance and compare the differences between the model prediction results and the actual observations.

6. The battery high and low temperature explosion-proof test box according to claim 5, characterized in that: In the loss function construction unit, the loss function is expressed as: In the formula, w is the coefficient vector of the characteristic variable, b is the intercept term, and x i is the feature vector of the i-th sample, y i is the true value of the i-th sample, N is the number of samples, and M is the number of features.

7. The battery high and low temperature explosion-proof test box according to claim 4, characterized in that: In the algorithm selection unit, the mathematical model of the ADMM algorithm is: minf1(x1)+f2(x2) A1x2+A1x2+b Where x1∈R m and x2∈R n is the decision variable, A1∈R p×m , A2∈R p×n and b∈R p represents the coefficient matrix of the equality constraint, and f1 and f2 are appropriate closed convex functions.

8. The battery high and low temperature explosion-proof test box according to claim 4, characterized in that: In the algorithm selection unit, the augmented Lagrangian function formula for the ADMM algorithm mathematical model is expressed as: L p (x1, x2, λ)=f1(x1)+f2(x2)+λ(A1x1+A2x2―b)+φ Where λ is the Lagrange multiplier and ρ is the penalty factor.

9. The battery high and low temperature explosion-proof test box according to claim 4, characterized in that: In the selection algorithm unit, the iterative process of ADMM is: In the formula, argmin is the decision variable value of the optimal solution of the objective function, k is the number of iterations, is the Lagrange multiplier.

10. The battery high and low temperature explosion-proof test box according to claim 4, characterized in that: In the model evaluation unit, the model prediction performance is evaluated using regression model evaluation indicators, including: The mean square error is expressed as: In the formula, y i is the true value of the i-th sample, which is the target value that the model attempts to predict or estimate. is the predicted value of the i-th sample, which is the output value calculated by the model based on the input features, and n is the sample size; The accuracy is calculated as follows: In the formula, TP is a true positive example and FP is a false positive example; The F1 score is calculated as: In the formula, P is the precision and R is the recall.

Images