A voltage detection device and detection method for a series battery pack for dynamic balancing

Abstract

The present invention discloses a voltage detection device and a detection method for a series battery pack for dynamic equalization, which relate to the field of battery detection. The voltage detection device for a series battery pack comprises: a balancing control module, which is used to access each battery cell of a battery pack to be detected, receive a preset balancing strategy, and adjust the charge and discharge state of each battery cell in real time according to battery voltage and temperature data; a voltage detection module, which is used to sample voltage data of each battery cell in real time through a voltage sensor, and convert the analog signal into a digital signal through an analog-to-digital converter; a model building unit, which is used to build a machine learning model for fault analysis, identify and mark abnormal sets, and output fault causes and adjustment strategy references; real-time collection of voltage and temperature data of each battery cell, monitoring the abnormal state of the battery cell, combining the machine learning model with the analysis of historical data, to generate a dynamic adjustment strategy based on an influence coefficient, which has stronger adaptability and effectively improves the charging and discharging efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of battery detection, and in particular to a voltage detection device and a detection method for a series battery pack for dynamic equalization. Background Art

[0002] The battery pack is composed of single cells with small power and low voltage connected in series. When charging and discharging, the single cells connected in series often have the problem of voltage and energy imbalance of the single cells due to the difference in battery characteristics and manufacturing process. When charging, if one single cell in the series battery pack is fully charged, the whole battery pack can no longer be charged. When discharging, if one single cell reaches the discharge limit, the whole battery pack can no longer be discharged. After long-term use, some batteries in the battery pack cannot be fully utilized, while some batteries will be seriously consumed. Therefore, one of the important parts of the battery management system is the balanced management of battery energy.

[0003] One of the available methods for detecting the balancing management status of a series battery pack is to detect the voltage. However, traditional battery monitoring systems are often unable to monitor the voltage and temperature of each battery cell in real time, resulting in the failure to detect abnormal conditions of some battery cells in a timely manner. The dynamic adjustment of the balancing strategy is relatively simple, resulting in low charging and discharging efficiency of the battery pack, and may even cause further damage to the battery. Some adjustment strategies may work in the current cycle for emergency considerations, but are not conducive to the healthy operation of multiple cycles. Moreover, these problems are difficult to discover through voltage detection data, lack effective fault analysis capabilities, and are unable to deeply explore the causes of faults. They lack the ability to learn and optimize from data, making it difficult to provide intelligent feedback and adjustment strategies for historical data, resulting in data islands. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides a voltage detection device and a detection method for a series battery pack for dynamic balancing, which can effectively solve the problems of the prior art.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] The present invention discloses a voltage detection device for a series battery pack for dynamic equalization, comprising:

[0009] The balancing control module is used to access each battery cell of the battery pack to be tested, receive a preset balancing strategy, and adjust the charge and discharge state of each battery cell in real time according to the battery voltage and temperature data;

[0010] The voltage detection module is used to sample the voltage data of each battery cell in real time through a voltage sensor and convert the analog signal into a digital signal through an analog-to-digital converter;

[0011] The temperature monitoring module is used to sample the temperature data of each battery cell in real time through a temperature sensor, execute synchronously with the voltage detection module, and convert the analog signal into a digital signal through an analog-to-digital converter;

[0012] The acquisition storage module is used to record the voltage, temperature, charge and discharge status and historical data of the battery unit, establish a time series data set, and update it in real time;

[0013] The anomaly detection module is used to set the upper and lower limits of the battery cell voltage and temperature, and compare the time series data set in real time. When abnormal data is detected, the alarm program is triggered, and the abnormal type is recorded. According to the abnormal type, it is classified and marked, and the marked abnormal set is output;

[0014] A model building unit, which is used to build a machine learning model for fault analysis, identify and mark abnormal sets, and output fault causes and adjustment strategy references;

[0015] The loss analysis module is used to receive the model building unit data, analyze the impact of the current faulty battery cell on the battery pack, calculate the loss coefficient caused by the faulty battery cell in the current and future cycles, and use the loss coefficient as a reference to analyze the impact coefficient of the faulty battery cell on the overall performance of the battery pack in the current and future cycles;

[0016] The strategy generation module is used to analyze the impact coefficients, rank the current and future impact coefficients with reference to the historical adjustment strategies provided by the model building unit, and formulate adjustment plans based on the ranking results;

[0017] The dynamic feedback module is used to feed back the generated adjustment plan to the balance control module, apply the adjustment strategy in real time, and identify the completion degree of the adjustment plan.

[0018] Furthermore, the model building unit includes a sample training module, a data output module and a correlation analysis module, the sample training module and the data output module are interactively connected via a wireless network, and the data output module and the loss analysis module are interactively connected via a wireless network, wherein:

[0019] A sample training module is used to pre-process the historical available adjustment data and the corresponding marked anomaly set as training samples;

[0020] The data output module selects a deep learning algorithm, builds a fault recognition model, inputs a labeled anomaly set, and outputs the cause of the fault.

[0021] Furthermore, the sample training module divides the training samples into a training set and a test set. The fault identification model constructed by the data output module uses the training set as input, executes back propagation and optimization algorithms to perform model training, continuously adjusts the weights of the fault identification model during the training process to minimize the loss function, uses the validation set to evaluate the performance of the fault identification model, and adjusts the hyperparameters of the fault identification model based on the evaluation results.

[0022] Furthermore, the data output module is interactively connected to a correlation analysis module via a wireless network, and the correlation analysis module is interactively connected to a strategy generation module via a wireless network. The correlation analysis module is used to generate an adjustment strategy based on a current marked anomaly set and its output fault cause, matching historical available adjustment data, and submitting the strategy to the strategy generation module.

[0023] Furthermore, the calculation formula of the influence coefficient of the faulty battery unit on the overall performance of the battery pack in the current cycle and the future cycle in the loss analysis module is:

[0024] ;

[0025] In the formula, P represents the influence coefficient of the faulty battery cell on the overall performance of the battery pack in the current and future cycles, F represents the power output of the faulty battery cell in the current working state, and W represents the total power output of the battery pack. represents the loss coefficient of the faulty battery cell in the current c cycle, represents the decay function, where e is a natural constant, k is a decay rate constant, t represents a time variable, and N represents the number of faulty battery cells that affect the performance of the battery pack in the current cycle. represents the loss caused by the nth faulty battery cell in the current cycle, Q represents the overall performance index of the battery pack, Represents the time point when the nth battery cell is identified.

[0026] Furthermore, in the process of the dynamic feedback module identifying the completion of the solution, the voltage and temperature performance data of the next cycle is obtained through the acquisition and storage module, the performance parameters are analyzed, and compared with the preset standards. If it is higher than the standard, it means that the completion degree has met the standard. Otherwise, it has not met the standard. If it has not met the standard, the working cache of the abnormality detection module is reset, real-time data collection is performed again, and alarm feedback is performed.

[0027] Furthermore, the balancing control module is interactively connected with the voltage detection module, the temperature monitoring module and the dynamic feedback module through an electrical medium, the acquisition and storage module is interactively connected with the voltage detection module, the temperature monitoring module and the abnormality detection module through a wireless network, the abnormality detection module and the model building unit are interactively connected through a wireless network, and the strategy generation module is interactively connected with the loss analysis module and the dynamic feedback module through a wireless network.

[0028] A voltage detection method for a series battery pack for dynamic balancing comprises the following steps:

[0029] Step 1: Monitor the voltage and temperature data of each battery cell in real time, formulate corresponding balancing strategies according to the voltage conditions of each battery cell, set balancing thresholds, and record the voltage and temperature indicators during the balancing process;

[0030] Step 2: Collect and organize the historical voltage, temperature, and charge and discharge status data of each battery cell, generate a data set for training the machine learning model, and record the performance data of each battery cell under different states;

[0031] Step 3: Apply machine learning models to analyze the real-time collected data, identify abnormal battery cell occurrences, and analyze the causes of failed battery cells;

[0032] Step 4: Calculate the loss coefficients of several battery cells associated with the current and future respectively, analyze the influence coefficient of the faulty battery cells in the current cycle, and evaluate the influence coefficient of the current battery pack performance in the next cycle;

[0033] Step 5: Use historical available adjustment data to train the machine learning model, generate the associated adjustment strategy for the faulty battery cells through the output of the influence coefficient set, sort them according to the influence coefficients of the current and future cycles, give priority to the battery cells with the greatest influence, and output the optimized adjustment strategy;

[0034] Step 6: Establish a real-time feedback mechanism to monitor the operating status of the battery pack in real time through the optimized adjustment strategy, and re-optimize the strategy based on real-time data feedback.

[0035] Furthermore, the calculation logic of the influence coefficient of Step 4 is as follows: by calculating the proportion of the power of the faulty battery to the total power, the influence of the faulty battery on the overall battery pack performance is reflected, and the influence of the fault influence changing with the time period is simulated, and the influence of all faulty batteries in the current cycle is uniformly analyzed by summing up, and the comprehensive influence coefficient on the overall performance of the battery pack is calculated.

[0036] Furthermore, in the Step 6, if the feedback result does not reach the preset threshold, the alarm data and optimization suggestions are sent to the management end. If no feedback is received within the preset time, the influence coefficient is automatically adjusted according to the preset adjustment standard to re-formulate the adjustment strategy and re-apply it.

[0037] (III) Beneficial effects

[0038] Compared with the known prior art, the technical solution provided by the present invention has the following beneficial effects:

[0039] By collecting the voltage and temperature data of each battery cell in real time, the abnormal status of the battery cell can be monitored to avoid potential damage caused by untimely monitoring. Combined with the analysis of historical data by the machine learning model, a dynamic adjustment strategy based on the influence coefficient is generated. Compared with the traditional simple balancing strategy, it is more adaptable, can effectively improve the charging and discharging efficiency, and extend the battery life.

[0040] By using machine learning models to identify and analyze the causes of failures, automatically identifying the root causes of failures, and providing accurate feedback based on historical and real-time data, future failures can be predicted and prevented. By deeply exploring the causes of failures, data support can be provided for subsequent battery management, avoiding simple emergency measures, and promoting healthy operation of multiple cycles. By evaluating the loss coefficient and combining historical data analysis, not only does it focus on the response of the current cycle, but it also considers future impacts, ensuring the long-term effectiveness of the adjustment strategy, making forward-looking management of the health status of the battery pack, and avoiding short-term performance improvements that damage long-term use.

[0041] By analyzing historical available adjustment data, intelligent feedback and optimization strategies are generated, data is integrated to break data silos, and adjustment suggestions based on intelligent analysis are provided. The charging and discharging strategies are dynamically adjusted to efficiently balance the voltage differences within the battery pack, improve the overall charging and discharging efficiency, continuously optimize the battery pack management strategy, and enhance the overall intelligence level of operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0043] Figure 1 It is a schematic diagram of the framework of the present invention.

[0044] The numbers in the figure represent, respectively, 1. Balance control module; 2. Voltage detection module; 3. Temperature monitoring module; 4. Acquisition and storage module; 5. Anomaly detection module; 6. Model building unit; 61. Sample training module; 62. Data output module; 63. Correlation analysis module; 7. Loss analysis module; 8. Strategy generation module; 9. Dynamic feedback module. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments 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 making creative work are within the scope of protection of the present invention.

[0046] The present invention will be further described below in conjunction with the embodiments.

[0047] Example 1

[0048] A voltage detection device for a series battery pack connected in a dynamic and balanced manner according to the present embodiment is as follows: Figure 1 As shown, including:

[0049] The balancing control module 1 is used to access each battery cell of the battery pack to be tested, receive the preset balancing strategy, and adjust the charge and discharge state of each battery cell in real time according to the battery voltage and temperature data; monitor the voltage and temperature of the battery cell in real time to ensure that the battery operates within a safe range, implement the preset balancing strategy, and improve the overall performance and safety of the battery pack;

[0050] The voltage detection module 2 is used to sample the voltage data of each battery cell in real time through a voltage sensor and convert the analog signal into a digital signal through an analog-to-digital converter;

[0051] The temperature monitoring module 3 is used to sample the temperature data of each battery cell in real time through a temperature sensor, execute synchronously with the voltage detection module 2, and convert the analog signal into a digital signal through an analog-to-digital converter;

[0052] The acquisition storage module 4 is used to record the voltage, temperature, charge and discharge status and historical data of the battery unit, establish a time series data set, and update it in real time; establish a detailed historical data record of the battery unit to provide a solid data foundation for subsequent fault analysis and strategy optimization, and ensure the scientific nature of decision-making;

[0053] The abnormality detection module 5 is used to set the upper and lower limits of the battery cell voltage and temperature, and compare the time series data set in real time. When abnormal data is detected, the alarm program is triggered, and the abnormal type is recorded. According to the abnormal type, it is classified and marked, and the marked abnormal set is output; the abnormal situation is quickly identified and the alarm is processed to protect the battery cell from damage and improve the safety and stability of the system;

[0054] The model building unit 6 is used to build a machine learning model for fault analysis, identify and mark abnormal sets, and output fault causes and adjustment strategy references; the model building unit 6 includes a sample training module 61, a data output module 62 and an association analysis module 63, the sample training module 61 and the data output module 62 are interactively connected via a wireless network, and the data output module 62 and the loss analysis module 7 are interactively connected via a wireless network, wherein:

[0055] The sample training module 61 is used to pre-process the historical available adjustment data and the corresponding marked abnormal set as training samples; the sample training module 61 divides the training samples into a training set and a test set, and the fault identification model constructed by the data output module 62 uses the training set as input, executes the back propagation and optimization algorithm to train the model, continuously adjusts the weight of the fault identification model during the training process to minimize the loss function, uses the validation set to evaluate the performance of the fault identification model, and adjusts the hyperparameters of the fault identification model according to the evaluation results; optimizes the training samples to improve the effectiveness and efficiency of the model training;

[0056] The data output module 62 selects a deep learning algorithm, builds a fault identification model, inputs a marked abnormal set, and outputs the cause of the fault; the data output module 62 is interactively connected to the association analysis module 63 through a wireless network, and the association analysis module 63 is interactively connected to the strategy generation module 8 through a wireless network. The association analysis module 63 is used to generate an adjustment strategy based on the current marked abnormal set and its output fault cause, matching the historical available adjustment data, and submitting it to the strategy generation module 8; the deep learning algorithm is selected to achieve efficient fault identification, support rapid decision-making, generate appropriate adjustment strategies based on historical data and current fault conditions, and realize intelligent adaptive control

[0057] The loss analysis module 7 is used to receive the data of the model building unit 6, analyze the impact of the current faulty battery cell on the battery pack, calculate the loss coefficient caused by the faulty battery cell in the current and future cycles respectively, and use the loss coefficient as a reference to analyze the impact coefficient of the faulty battery cell on the overall performance of the battery pack in the current and future cycles; by calculating the loss coefficient, comprehensively evaluate the impact of the faulty cell on the overall performance of the battery pack, and support accurate fault impact prediction and management decision-making;

[0058] The strategy generation module 8 is used to analyze the influence coefficient, sort the current and future influence coefficients with the historical adjustment strategy provided by the model building unit 6 as the reference of the associated adjustment strategy, and formulate an adjustment plan according to the sorting result; formulate a priority ranking based on the historical adjustment strategy and the influence coefficient, and provide an optimized adjustment plan for the battery pack to improve the battery use efficiency;

[0059] The dynamic feedback module 9 is used to feed back the generated adjustment plan to the balancing control module 1, apply the adjustment strategy in real time, and identify the completion of the adjustment plan. The generated adjustment plan is applied in real time, and the completion of the plan implementation is monitored to ensure that the dynamic balancing strategy is timely and effective to achieve the optimal state of battery pack performance.

[0060] like Figure 1 As shown, the balancing control module 1 is interactively connected with the voltage detection module 2, the temperature monitoring module 3 and the dynamic feedback module 9 through an electrical medium, the acquisition storage module 4 is interactively connected with the voltage detection module 2, the temperature monitoring module 3 and the abnormality detection module 5 through a wireless network, the abnormality detection module 5 and the model building unit 6 are interactively connected through a wireless network, and the strategy generation module 8 is interactively connected with the loss analysis module 7 and the dynamic feedback module 9 through a wireless network.

[0061] In the specific implementation of this embodiment, the balancing control module 1 controls the global modules and units to formulate the initial balancing strategy, the voltage detection module 2 and the temperature monitoring module 3 respectively detect the voltage and temperature data of each battery cell in the battery pack, and the acquisition and storage module 4 acquires and stores the data, the abnormality detection module 5 determines whether the obtained data has abnormalities, the sample training module 61 in the model building unit 6 obtains historical abnormal data for sample training, builds a fault identification model, runs the constructed fault identification model through the data output module 62, outputs the fault data, analyzes the loss data of the abnormal battery cell through the loss analysis module 7, obtains the associated impact data of the faulty battery cell based on the real-time changes of the historical data and the associated battery cells through the association analysis module 63, and outputs the adjustment strategy after comprehensive analysis by the strategy generation module 8, performs real-time feedback through the dynamic feedback module 9, and analyzes the completion degree.

[0062] Example 2

[0063] In other aspects, this embodiment further provides a voltage detection method for a series battery pack for dynamic balancing, comprising the following steps:

[0064] Step 1: Monitor the voltage and temperature data of each battery cell in real time, formulate corresponding balancing strategies according to the voltage conditions of each battery cell, set balancing thresholds, and record the voltage and temperature indicators during the balancing process;

[0065] Step 2: Collect and organize the historical voltage, temperature, and charge and discharge status data of each battery cell, generate a data set for training the machine learning model, and record the performance data of each battery cell under different states;

[0066] Step 3: Apply machine learning models to analyze the real-time collected data, identify abnormal battery cell occurrences, and analyze the causes of failed battery cells;

[0067] Step 4: Calculate the loss coefficients of several battery cells associated with the current and future respectively, analyze the influence coefficient of the faulty battery cells in the current cycle, and evaluate its influence coefficient on the battery pack performance in the next cycle; The influence coefficient calculation logic of Step 4 is: by calculating the proportion of the faulty battery power to the total power, reflect the influence of the faulty battery on the overall battery pack performance, and simulate the influence of the fault influence changing with the time cycle, and analyze the influence of all faulty batteries in the current cycle in a unified manner by summing, and calculate the comprehensive influence coefficient on the overall performance of the battery pack;

[0068] Step 5: Use historical available adjustment data to train the machine learning model, generate the associated adjustment strategy for the faulty battery cells through the output of the influence coefficient set, sort them according to the influence coefficients of the current and future cycles, give priority to the battery cells with the greatest influence, and output the optimized adjustment strategy;

[0069] Step 6: Establish a real-time feedback mechanism to monitor the operating status of the battery pack in real time through the optimized adjustment strategy, and re-optimize the strategy based on the real-time data feedback. If the feedback result does not reach the preset threshold, send alarm data and optimization suggestions to the management end. If no feedback is received within the preset time, the influence coefficient is automatically adjusted according to the preset adjustment standard to re-formulate the adjustment strategy and re-apply it.

[0070] Compared with the existing technology, this method can monitor the voltage and temperature of the battery cell in real time and obtain status information in time, so as to dynamically adjust the balancing strategy and improve the overall performance and safety of the battery pack. By using the machine learning model to analyze the real-time data, the abnormal conditions of the battery cells can be identified earlier and more accurately. Compared with the traditional threshold-based monitoring method, the sensitivity and accuracy of fault detection can be significantly improved. By using the historical battery cell voltage, charge and discharge and temperature data for model training, the performance of the battery cell can be better understood and predicted. This big data-driven method increases the reliability of the model.

[0071] Calculating the impact coefficient of the faulty battery and comprehensively analyzing the impact of the fault on the overall battery pack performance can more scientifically evaluate which battery cell has the greatest impact on the performance degradation of the battery pack, thereby formulating a priority processing strategy and establishing a real-time feedback mechanism. Automatically optimizing and adjusting the strategy based on the operating status and feedback data can effectively improve the adaptability and flexibility of the system and ensure that the battery pack remains in the best condition under various working conditions. When a fault occurs, it can send alarm information and optimization suggestions in a timely manner to improve safety. If no feedback is received, the battery system can adjust itself autonomously, which improves the level of intelligence and reduces the need for manual intervention. Battery cells are prioritized according to the impact coefficient, effectively concentrating resources to process the batteries that have the greatest impact on performance, optimizing the processing process and improving management efficiency.

[0072] Example 3

[0073] In this embodiment, a calculation formula for the influence coefficient of a faulty battery cell on the overall performance of the battery pack in the current cycle and the future cycle is provided, which is specifically:

[0074] ;

[0075] In the formula, P represents the influence coefficient of the faulty battery cell on the overall performance of the battery pack in the current and future cycles, which indicates the comprehensive influence of the faulty battery cell on the overall performance of the battery pack. F represents the power output of the faulty battery cell in the current working state, which directly affects the performance of the battery pack. W represents the total power output of the battery pack, which is used as a benchmark to calculate the influence of the faulty battery cell on the overall performance. Represents the loss coefficient of the faulty battery cell in the current c cycle, indicating the degree of performance loss caused by the faulty battery cell. The larger its value, the more significant the loss. represents the decay function, where e is a natural constant, k is a decay rate constant, t represents a time variable, and simulates the decrease of the fault impact over time. As time goes by, the impact of the fault will gradually decrease. N represents the number of faulty battery cells that affect the performance of the battery pack in the current cycle, and is an iteration count when calculating the impact of all faulty battery cells on the overall performance. Represents the amount of loss caused by the nth faulty battery cell in the current cycle, indicating the actual loss caused by the specific faulty battery to the battery pack performance in the current cycle. Q represents the overall performance index of the battery pack, and as a benchmark comparison, it shows the proportion of the loss caused by the battery cell in the overall performance. Represents the point in time when the nth battery cell is identified, recording the exact time when the fault occurred, which serves as a key reference for analyzing the impact attenuation and the extent of the impact in future cycles.

[0076] In summary: The present invention ensures that the status of each battery cell is monitored in real time through a complete voltage and temperature detection module, and timely discovers potential faults. By collecting and analyzing the real-time data of each battery cell, dynamic balancing adjustment can be achieved according to the real-time status of each battery cell. The machine learning model is integrated, and through historical data training, the cause of the fault can be more accurately identified and its impact on the entire battery pack can be analyzed. According to the historical available data and the current status, an optimized adjustment strategy is dynamically generated, thereby improving management efficiency;

[0077] The present invention designs a multi-level abnormality detection and early warning mechanism, which can promptly issue an alarm when an abnormality occurs in a battery cell, and provide targeted adjustment suggestions, integrate the monitoring data of all battery cells to form a data set, and conduct comprehensive analysis and decision-making based on this data, thereby improving the overall system performance.

[0078] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A voltage detection device for a series battery pack for dynamic balancing, characterized in that: include: A balancing control module (1) is used to access each battery cell of the battery pack to be tested, receive a preset balancing strategy, and adjust the charge and discharge state of each battery cell in real time according to battery voltage and temperature data; A voltage detection module (2) is used to sample voltage data of each battery cell in real time through a voltage sensor and convert the analog signal into a digital signal through an analog-to-digital converter; A temperature monitoring module (3) is used to sample temperature data of each battery cell in real time through a temperature sensor, and is executed synchronously with the voltage detection module (2) to convert the analog signal into a digital signal through an analog-to-digital converter; The acquisition storage module (4) is used to record the voltage, temperature, charge and discharge status and historical data of the battery unit, establish a time series data set, and update it in real time; The abnormality detection module (5) is used to set the upper and lower limits of the battery cell voltage and temperature, and compare the time series data set in real time. When abnormal data is detected, an alarm program is triggered, and the abnormality type is recorded. According to the abnormality type, it is classified and marked, and a marked abnormality set is output; A model building unit (6), used to build a machine learning model for fault analysis, identify and mark abnormal sets, and output fault causes and adjustment strategy references; A loss analysis module (7) is used to receive data from the model building unit (6), analyze the impact of the current faulty battery cell on the battery pack, calculate the loss coefficient caused by the faulty battery cell in the current and future cycles, and use the loss coefficient as a reference to analyze the impact coefficient of the faulty battery cell on the overall performance of the battery pack in the current and future cycles; A strategy generation module (8) is used to analyze the impact coefficients, rank the current and future impact coefficients with reference to the historical adjustment strategies provided by the model building unit (6), and formulate an adjustment plan based on the ranking results; A dynamic feedback module (9) is used to feed back the generated adjustment plan to the balancing control module (1), apply the adjustment strategy in real time, and identify the completion degree of the adjustment plan; The calculation formula of the influence coefficient of the faulty battery unit on the overall performance of the battery pack in the current cycle and the future cycle in the loss analysis module (7) is: ； In the formula, P represents the influence coefficient of the faulty battery cell on the overall performance of the battery pack in the current and future cycles, F represents the power output of the faulty battery cell in the current working state, and W represents the total power output of the battery pack. represents the loss coefficient of the faulty battery cell in the current c cycle, represents the decay function, where e is a natural constant, k is a decay rate constant, t represents a time variable, and N represents the number of faulty battery cells that affect the performance of the battery pack in the current cycle. represents the loss caused by the nth faulty battery cell in the current cycle, Q represents the overall performance index of the battery pack, Represents the time point when the nth battery cell is identified.

2. The voltage detection device for a series battery pack for dynamic equalization according to claim 1, characterized in that: The model building unit (6) comprises a sample training module (61), a data output module (62) and a correlation analysis module (63), wherein the sample training module (61) and the data output module (62) are interactively connected via a wireless network, and the data output module (62) and the loss analysis module (7) are interactively connected via a wireless network, wherein: A sample training module (61) is used to pre-process the historical available adjustment data and the corresponding marked anomaly set and use them as training samples; The data output module (62) selects a deep learning algorithm, builds a fault recognition model, inputs a labeled anomaly set, and outputs the cause of the fault.

3. The voltage detection device for a series battery pack for dynamic equalization according to claim 2, characterized in that: The sample training module (61) divides the training samples into a training set and a test set. The fault identification model constructed by the data output module (62) uses the training set as input, executes back propagation and optimization algorithms to perform model training, continuously adjusts the weights of the fault identification model during the training process to minimize the loss function, uses the validation set to evaluate the performance of the fault identification model, and adjusts the hyperparameters of the fault identification model according to the evaluation results.

4. The voltage detection device for a series battery pack for dynamic balancing according to claim 2, characterized in that: The data output module (62) is interactively connected to the association analysis module (63) via a wireless network. The association analysis module (63) is interactively connected to the strategy generation module (8) via a wireless network. The association analysis module (63) is used to match the historical available adjustment data based on the current marked abnormal set and its output fault cause to generate an adjustment strategy, and submit it to the strategy generation module (8).

5. The voltage detection device for a series battery pack for dynamic balancing according to claim 1, characterized in that: In the process of identifying the degree of completion of the scheme, the dynamic feedback module (9) obtains the voltage and temperature performance data of the next cycle through the acquisition storage module (4), analyzes the performance parameters, and compares them with the preset standards. If it is higher than the standard, it means that the degree of completion has reached the standard. Otherwise, it has not reached the standard. If it has not reached the standard, the working cache of the abnormality detection module (5) is reset, real-time data acquisition is performed again, and alarm feedback is performed.

6. The voltage detection device for a series battery pack for dynamic equalization according to claim 1, characterized in that: The equalization control module (1) is interactively connected with the voltage detection module (2), the temperature monitoring module (3) and the dynamic feedback module (9) via an electrical medium; the acquisition storage module (4) is interactively connected with the voltage detection module (2), the temperature monitoring module (3) and the abnormality detection module (5) via a wireless network; the abnormality detection module (5) is interactively connected with the model construction unit (6) via a wireless network; and the strategy generation module (8) is interactively connected with the loss analysis module (7) and the dynamic feedback module (9) via a wireless network.

7. A method for detecting voltage of a series battery pack for dynamic balancing, the method being an implementation method of a device for detecting voltage of a series battery pack for dynamic balancing according to any one of claims 1 to 6, characterized in that: The following steps are involved: Step 1: Monitor the voltage and temperature data of each battery cell in real time, formulate corresponding balancing strategies according to the voltage conditions of each battery cell, set balancing thresholds, and record the voltage and temperature indicators during the balancing process; Step 2: Collect and organize the historical voltage, temperature, and charge and discharge status data of each battery cell, generate a data set for training the machine learning model, and record the performance data of each battery cell under different states; Step 3: Apply machine learning models to analyze the real-time collected data, identify abnormal battery cell occurrences, and analyze the causes of failed battery cells; Step 4: Calculate the loss coefficients of several battery cells associated with the current and future respectively, analyze the influence coefficient of the faulty battery cells in the current cycle, and evaluate the influence coefficient of the current battery pack performance in the next cycle; Step 5: Use historical available adjustment data to train the machine learning model, generate the associated adjustment strategy for the faulty battery cells through the output of the influence coefficient set, sort them according to the influence coefficients of the current and future cycles, give priority to the battery cells with the greatest influence, and output the optimized adjustment strategy; Step 6: Establish a real-time feedback mechanism to monitor the operating status of the battery pack in real time through the optimized adjustment strategy, and re-optimize the strategy based on real-time data feedback.

8. The method for detecting voltage of a series battery pack for dynamic balancing according to claim 7, characterized in that: The calculation logic of the influence coefficient of Step 4 is as follows: by calculating the ratio of the power of the faulty battery to the total power, the influence of the faulty battery on the overall performance of the battery pack is reflected, and the influence of the fault influence changing with the time period is simulated. The influence of all faulty batteries in the current cycle is analyzed uniformly by summing up, and the comprehensive influence coefficient on the overall performance of the battery pack is calculated.

9. The method for detecting voltage of a series battery pack for dynamic balancing according to claim 7, characterized in that: In the Step 6, if the feedback result does not reach the preset threshold, the alarm data and optimization suggestions are sent to the management end. If no feedback is received within the preset time, the influence coefficient is automatically adjusted according to the preset adjustment standard to re-formulate the adjustment strategy for re-application.

Images