Thermal safety coordination management method and system for battery energy storage

Abstract

The invention discloses a thermal safety coordination management method for battery energy storage. The thermal safety coordination management method comprises the following steps: acquiring real-time operation data and historical operation data of a battery; constructing a battery performance attenuation prediction model based on a machine learning algorithm, and obtaining a predicted battery performance attenuation value in combination with the historical operation data and the real-time operation data; acoustic emission and optical fiber sensors are adopted to monitor the battery to obtain a battery health index, and an early warning instruction is sent out according to a preset multi-stage thermal management strategy in combination with a predicted battery performance attenuation value; a gas sensor is adopted to monitor a battery to obtain real-time gas data, the real-time gas data is analyzed, an out-of-control early warning instruction is sent out according to a preset multi-stage thermal out-of-control management strategy, and thermal safety coordination management of battery energy storage is completed. The invention also provides a system for realizing the thermal safety coordination management method for battery energy storage. According to the method, it is ensured that the battery can be maintained in the optimal working temperature range at different environment temperatures, thermal runaway can be warned in advance, and the safety and reliability of the battery are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery management, and in particular relates to a thermal safety coordination management method and system for battery energy storage. Background Art

[0002] With the rapid development of new energy vehicles and large-scale energy storage systems, battery safety issues are becoming increasingly prominent. During operation, batteries may suffer from thermal runaway due to electrical abuse, thermal abuse, or mechanical abuse, leading to safety issues such as fire, combustion, or even explosion. On the other hand, the temperature of the battery is very important for the safe operation and performance of the battery. The temperature has a direct impact on the energy storage performance of the battery. Too high or too low a temperature will reduce the discharge capacity and charging efficiency of the battery. By controlling the temperature of the battery, the battery is kept in a suitable temperature range to ensure efficient operation and good performance of the battery. At the same time, it effectively avoids safety risks caused by overheating of the battery under high-temperature working conditions, promptly discovers and resolves potential safety hazards, ensures safe operation of the system, extends battery life, and ensures system performance, and ensures stable and reliable operation of the energy storage system. Therefore, effective thermal management and thermal runaway warning of the battery are crucial.

[0003] Existing batteries often use a single management method for thermal management. This management method is mainly based on unified management, without considering the need for different management in different situations. When a situation occurs, the power supply is often directly turned off, and the coordination and management capabilities are poor. It cannot ensure that the battery can maintain the optimal operating temperature range under different ambient temperatures. Summary of the invention

[0004] One of the purposes of the present invention is to provide a method for thermal safety coordinated management of battery energy storage, which uses machine learning algorithms to make predictions based on historical data and real-time data, and implements thermal safety coordinated management of battery energy storage by performing multi-level thermal management of the battery.

[0005] A second objective of the present invention is to provide a system for implementing the thermal safety coordinated management method of battery energy storage.

[0006] The present invention provides a thermal safety coordination management method for battery energy storage, comprising the following steps: S1. Obtain real-time and historical operating data of the battery; S2. Based on the machine learning algorithm, a battery performance degradation prediction model is constructed, and the historical operation data and real-time operation data obtained in step S1 are combined to obtain the predicted battery performance degradation value; S3. Use acoustic emission and optical fiber sensors to monitor the battery, obtain monitoring data, analyze the monitoring data to obtain a battery health index, and then combine the predicted battery performance attenuation value obtained in step S2 to determine whether to issue a warning instruction according to a preset multi-level thermal management strategy. If no warning instruction is issued, proceed to step S4; S4. Monitor the battery in different areas, use gas sensors to obtain real-time gas data, and then perform gas chromatography analysis on the real-time gas data. According to the preset multi-level thermal runaway management strategy, issue a runaway warning instruction to complete the thermal safety coordination management of battery energy storage.

[0007] Step S2 specifically includes the following steps: Construct an initial battery performance degradation prediction model based on a decision tree algorithm; Normalizing the historical operation data and the real-time operation data obtained in step S1; The initial battery performance degradation prediction model is trained using historical operation data to obtain a trained battery performance degradation prediction model; The mean square error and determination coefficient are used as performance indicators of the battery performance degradation prediction model after training to judge the stability of the model, and finally the battery performance degradation prediction model is obtained, and the average performance indicator after training is calculated; Based on the battery performance degradation prediction model, the normalized real-time operating data is input to obtain the predicted battery performance degradation value PAV.

[0008] Step S2 is specifically as follows: A battery performance degradation prediction model is constructed based on the decision tree algorithm. The battery performance degradation prediction model includes B decision trees. Each decision tree is run separately to obtain the result. The final output of the battery performance degradation prediction model is the average value of the output results of all decision trees, expressed using the following formula: in, is the final prediction result of the model; B is the number of decision trees in the model; is the prediction result of the bth decision tree for the input feature X; b is the serial number of a single decision tree; The historical operation data and the real-time operation data obtained in step S1 are normalized and expressed using the following formula: Among them, X is the historical operation data after normalization; For the original historical running data; It is the minimum value in the historical running data; It is the maximum value in the historical operation data; It is the real-time operation data after normalization; It is the original real-time operation data; is the minimum value in the real-time running data; is the maximum value in the real-time running data; The initial battery performance degradation prediction model is trained using historical operation data to obtain a trained battery performance degradation prediction model; The mean square error and determination coefficient are used as performance indicators of the battery performance degradation prediction model after training to judge the stability of the model, and finally the battery performance degradation prediction model is obtained, and the average performance indicator after training is calculated; The mean square error MSE is expressed using the following formula: Where n is the total number of data; is the i-th real-time running data; Coefficient of determination Use the following formula to express it: in, is the average value of real-time running data; Set the mean square error threshold ; like , it means the model is stable; like , it means that the model is unstable. Add L1 regularization term to MSE and retrain to prevent overfitting of the model; New mean square error Use the following formula to express it: Among them, h is the regularization parameter; is the i-th element in the model parameters; Setting the coefficient of determination threshold ; like , it means the model is stable; like , it means that the model is unstable, and the coefficient of determination Add L2 regularization term and retrain to prevent the model from overfitting; New coefficient of determination Use the following formula to express it: ; Based on the battery performance degradation prediction model, the normalized real-time operating data is input to obtain the predicted battery performance degradation value PAV.

[0009] Step S3 specifically includes the following steps: Acoustic emission and optical fiber sensors are used to monitor the microscopic activities inside the battery to obtain monitoring data; the monitoring data includes acoustic emission data and optical fiber sensor data; the acoustic emission data is the maximum amplitude of the AE signal and the total energy of the AE signal; the optical fiber sensor data is temperature, strain and pressure; The acoustic emission data and optical fiber sensor data in the monitoring data are analyzed separately, and then the battery health index BHI is obtained through weighted calculation; According to the preset multi-level thermal management strategy, by comparing the health index BHI, the battery performance decay value PAV obtained in step S2 with the corresponding preset threshold, it is determined whether to issue a warning instruction. If no warning instruction is issued, step S4 is entered.

[0010] Step S3 is specifically as follows: Divide the battery into n equal areas along its length, marked as arrive ; Acoustic emission sensors and optical fiber sensors are used to arrive n areas for collection; Before the battery starts charging and discharging cycles, record the initial readings of all acoustic emission sensors and optical fiber sensors, and set a detection cycle for periodic sampling; The data collected by the acoustic emission sensor are the maximum amplitude of the AE signal in each area and the total energy of the AE signal in each area; The data collected by the fiber optic sensor are the temperature, strain and pressure of each area; The acoustic emission data collected by the acoustic emission sensor are analyzed to obtain the acoustic emission processing data AEI, which is expressed using the following formula: in, is the AE signal amplitude of the jth region at time t; is the AE signal energy of the jth region at time t; is the maximum value operation for time t; The temperature data collected by the optical fiber sensor is analyzed to obtain the temperature processing data TeI, which is expressed by the following formula: in, is the temperature of the jth region at time t; The strain data collected by the optical fiber sensor is analyzed to obtain the strain processing data StI, which is expressed by the following formula: in, is the strain of the jth region at time t; The pressure data collected by the optical fiber sensor is analyzed to obtain the pressure processing data PrI, which is expressed by the following formula: in, is the pressure of the jth region at time t; The acoustic emission processing data AEI, temperature processing data TeI, strain processing data StI and pressure processing data PrI are weighted to obtain the battery health index BHI, which is expressed by the following formula: in, The weight of acoustic emission processing data AEI; is the weight of temperature processing data TeI; is the weight of strain processing data StI; is the weight of the pressure treatment data PrI; According to the preset multi-level thermal management strategy, by comparing the health index BHI, the battery performance attenuation value PAV obtained in step S2 with the corresponding preset threshold, it is determined whether to issue a warning instruction. The multi-level thermal management strategy is specifically as follows: Set battery health threshold , battery performance attenuation threshold ; when or When the heating mode is started, a first-level instruction is issued; The heating power Hpr of the heating mode is calculated using the following formula: in, is the proportionality coefficient of heating power; is the battery target temperature; is the current temperature of the battery; Battery target temperature Use the following formula to calculate: in, is the basic target temperature of the battery; Adjust the battery target temperature scaling factor for BHI; A scaling factor for adjusting the battery target temperature for PAV; when and When Further judgment: when When the cooling mode is started, a secondary instruction is issued, in which is the maximum threshold of battery temperature; The heat dissipation power Hdp of the heat dissipation mode is calculated using the following formula: in, is the proportionality coefficient of heat dissipation power; when , it is determined to be in normal mode, no instruction is issued, and the process goes to step S4.

[0011] Step S4 specifically includes the following steps: Use gas sensors to monitor the battery in real time to obtain real-time gas data; Use gas chromatograph to analyze the real-time gas data and establish standard curves for different gas components; Calculate gas concentration based on real-time gas data and standard curve; According to the preset multi-level thermal runaway management strategy, by comparing the gas concentration with the preset threshold, a thermal runaway command is issued to complete the thermal safety coordinated management of battery energy storage.

[0012] Step S4 is specifically as follows: The area around the battery is divided into u different monitoring areas, and gas sensors are deployed in each monitoring area. For a certain monitoring area h, the real-time gas data is calculated using the following formula: in, is the output voltage of the gas sensor in the hth area; is the supply voltage of the hth region; is the resistance value of the gas sensor in the hth area; is the load resistance of the hth region; For a certain monitoring area h, a gas chromatograph is used for analysis, and combined with the existing standard gas of known concentration, a standard curve of different gas components is established, which is expressed by the following formula: in, is the chromatographic peak area of ​​the pth gas in the hth region; is the correction factor for the pth gas in the hth region; is the mass of the pth gas in the hth region; Calculate the gas concentration in the hth area based on real-time gas data and the standard curve , use the following formula to calculate: in, is the gas sensitive resistance value of the gas sensor in the hth area in clean air; is the constant of the gas sensor in the hth area; is another constant of the gas sensor in the hth region; is the constant of the hth region; Gas sensor resistance value in clean air in the hth area Use the following formula to express it: According to the multi-level thermal runaway management strategy, by comparing the gas concentration with the preset threshold, a thermal runaway command is issued to complete the thermal safety coordination management of battery energy storage, specifically: Set the minimum threshold of gas concentration to , the maximum threshold is ; like , issued a mild thermal runaway command; like , issued a moderate thermal runaway command; like , issuing a severe thermal runaway command.

[0013] Specifically, the thermal safety coordination management of battery energy storage also includes: When a level 1 command is received, a level 1 warning is issued and a reminder is broadcast; When a level 2 command is received, a level 2 warning is issued and the alarm is sounded; When a mild thermal runaway command is received, a first-level runaway warning is issued and the alarm light starts to light up; When a moderate thermal runaway command is received, a level 2 runaway warning is issued and the warning light starts flashing; When a severe thermal runaway command is received, a level 3 runaway warning is issued, the warning light starts flashing and an alarm sounds.

[0014] The present invention also provides a system for implementing the thermal safety coordinated management method of battery energy storage, including an acquisition module, a prediction module, a monitoring thermal management module, and a runaway thermal management module; The acquisition module acquires the real-time operation data and historical operation data of the battery and uploads the data to the prediction module; The prediction module builds a battery performance degradation prediction model based on the received data and machine learning algorithm, and then combines the historical operation data with the real-time operation data to obtain the predicted battery performance degradation value, and uploads the data to the monitoring thermal management module; The monitoring thermal management module uses acoustic emission and optical fiber sensors to monitor the battery based on the received data, obtains monitoring data, and analyzes the monitoring data to obtain the battery health index. Combined with the predicted battery performance attenuation value, it determines whether to issue a warning instruction based on the preset multi-level thermal management strategy. If no warning instruction is issued, the data is uploaded to the multi-level thermal management module. The runaway thermal management module monitors the battery in different areas based on the received data, uses gas sensors to obtain real-time gas data, and then performs gas chromatography analysis on the real-time gas data. Based on the preset multi-level thermal runaway management strategy, it issues a runaway warning instruction to complete the thermal safety coordinated management of battery energy storage.

[0015] The present invention discloses a thermal safety coordinated management method and system for battery energy storage. By combining a machine learning algorithm with historical data and real-time operation data, the performance degradation and potential failures of the battery can be predicted more accurately, and the problems that may occur in the battery can be predicted in advance. In addition, a multi-level thermal management module can be used to provide early warning. At the same time, the use of acoustic emission and optical fiber sensors can be combined to enable real-time monitoring of microscopic activities inside the battery. The battery can be better monitored under multi-dimensional data, so that the status of the battery in use can be understood in real time. Moreover, under a flexible thermal management strategy, different levels of management can be guaranteed when different problems occur. In addition, a gas sensor can be used to detect the gas emitted by the battery in real time. Real-time monitoring can respond quickly when battery thermal runaway occurs. The ability of real-time monitoring and synchronous adjustment of multi-level thermal management modes can effectively control and alleviate thermal runaway events and protect the safety of personnel and equipment. Under thermal management measures, different levels of early warning measures are allocated and corresponding response measures are triggered. This hierarchical early warning mechanism can ensure that appropriate response measures are taken under different levels of risks, improve the safety and reliability of the entire system, achieve the effect of rapid response and timely control in the early stage, ensure the safe use of the battery, ensure that the battery can be maintained within the optimal operating temperature range under different ambient temperatures, and can give early warning of thermal runaway, thereby improving the safety and reliability of the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 It is a structural schematic diagram of the system of the present invention; Figure 3 A battery temperature variation diagram under single thermal management in an embodiment of the present invention; Figure 4 This is a diagram of battery temperature changes under management using the method of the present invention in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The present invention provides a thermal safety coordination management method for battery energy storage, the flow diagram of which is as follows: Figure 1 As shown, the following steps are included: S1. Obtain real-time and historical operating data of the battery; S2. Based on the machine learning algorithm, a battery performance degradation prediction model is constructed, and then the historical operation data and real-time operation data obtained in step S1 are combined to obtain the predicted battery performance degradation value, specifically: A battery performance degradation prediction model is constructed based on the decision tree algorithm. The battery performance degradation prediction model includes B decision trees. Each decision tree is run separately to obtain the result. The final output of the battery performance degradation prediction model is the average value of the output results of all decision trees, expressed using the following formula: in, is the final prediction result of the model; B is the number of decision trees in the model; is the prediction result of the bth decision tree for the input feature X; b is the serial number of a single decision tree; The historical operation data and the real-time operation data obtained in step S1 are normalized and expressed using the following formula: Among them, X is the historical operation data after normalization; For the original historical running data; It is the minimum value in the historical running data; It is the maximum value in the historical operation data; It is the real-time operation data after normalization; It is the original real-time operation data; is the minimum value in the real-time running data; is the maximum value in the real-time running data; The correlation analysis between historical operation data and real-time operation data is performed using the following formula: Where r is the correlation coefficient and ; is the i-th historical running data; is the i-th real-time running data; n is the total number of data; is the average value of historical running data; is the average value of real-time running data; Identify the features most relevant to battery performance degradation through correlation analysis, reduce model complexity and improve model generalization capabilities; The initial battery performance degradation prediction model is trained using historical operation data to obtain a trained battery performance degradation prediction model; The mean square error and determination coefficient are used as performance indicators of the battery performance degradation prediction model after training to judge the stability of the model, and finally the battery performance degradation prediction model is obtained, and the average performance indicator after training is calculated; The mean square error MSE is expressed using the following formula: Where n is the total number of data; is the i-th real-time running data; Coefficient of determination Use the following formula to express it: in, is the average value of real-time running data; Set the mean square error threshold ; like , it means the model is stable; like , it means that the model is unstable. Add L1 regularization term to MSE and retrain to prevent overfitting of the model; New mean square error Use the following formula to express it: Among them, h is the regularization parameter; is the i-th element in the model parameters; Setting the coefficient of determination threshold ; like , it means the model is stable; like , it means that the model is unstable, and the coefficient of determination Add L2 regularization term and retrain to prevent the model from overfitting; New coefficient of determination Use the following formula to express it: Based on the battery performance degradation prediction model, the normalized real-time operating data is input to obtain the predicted battery performance degradation value PAV.

[0018] S3. Use acoustic emission and optical fiber sensors to monitor the battery, obtain monitoring data, analyze the monitoring data to obtain a battery health index, and then combine the predicted battery performance attenuation value obtained in step S2 to issue a warning instruction according to the preset multi-level thermal management strategy, specifically: Divide the battery into n equal areas along its length, marked as arrive ; Acoustic emission sensors and optical fiber sensors are used to arrive n areas for collection; Before the battery starts charging and discharging cycles, record the initial readings of all acoustic emission sensors and optical fiber sensors, and set a detection cycle for periodic sampling; The data collected by the acoustic emission sensor are the maximum amplitude of the AE signal in each area and the total energy of the AE signal in each area; The data collected by the fiber optic sensor are the temperature, strain and pressure of each area; The acoustic emission data collected by the acoustic emission sensor are analyzed to obtain the acoustic emission processing data AEI, which is expressed using the following formula: in, is the AE signal amplitude of the jth region at time t; is the AE signal energy of the jth region at time t; is the maximum value operation for time t; The temperature data collected by the optical fiber sensor is analyzed to obtain the temperature processing data TeI, which is expressed by the following formula: in, is the temperature of the jth region at time t; The strain data collected by the optical fiber sensor is analyzed to obtain the strain processing data StI, which is expressed by the following formula: in, is the strain of the jth region at time t; The pressure data collected by the optical fiber sensor is analyzed to obtain the pressure processing data PrI, which is expressed by the following formula: in, is the pressure of the jth region at time t; The acoustic emission processing data AEI, temperature processing data TeI, strain processing data StI and pressure processing data PrI are weighted to obtain the battery health index BHI, which is expressed by the following formula: in, The weight of acoustic emission processing data AEI; is the weight of temperature processing data TeI; is the weight of strain processing data StI; is the weight of the pressure treatment data PrI; According to the preset multi-level thermal management strategy, by comparing the health index BHI, the battery performance attenuation value PAV obtained in step S2 with the corresponding preset threshold, it is determined whether to issue a warning instruction. The multi-level thermal management strategy is specifically as follows: Set battery health threshold , battery performance attenuation threshold ; when or When the heating mode is started, a first-level instruction is issued; The heating power Hpr of the heating mode is calculated using the following formula: in, is the proportionality coefficient of heating power; is the battery target temperature; is the current temperature of the battery; Battery target temperature Use the following formula to calculate: in, is the basic target temperature of the battery; Adjust the battery target temperature scaling factor for BHI; A scaling factor for adjusting the battery target temperature for PAV; when and When Further judgment: when When the cooling mode is started, a secondary instruction is issued, in which is the maximum threshold of battery temperature; The heat dissipation power Hdp of the heat dissipation mode is calculated using the following formula: in, is the proportionality coefficient of heat dissipation power; when , it is determined to be in normal mode, no instruction is issued, and the process goes to step S4.

[0019] S4. Monitor the battery in different areas, use gas sensors to obtain real-time gas data, and then perform gas chromatography analysis on the real-time gas data. According to the preset multi-level thermal runaway management strategy, issue a runaway warning instruction to complete the thermal safety coordination management of battery energy storage, specifically: The area around the battery is divided into u different monitoring areas, and gas sensors are deployed in each monitoring area. For a certain area h, the real-time gas data is calculated using the following formula: in, is the output voltage of the gas sensor in the hth area; is the supply voltage of the hth region; is the resistance value of the gas sensor in the hth area; is the load resistance of the hth region; For a certain monitoring area h, a gas chromatograph is used for analysis, and combined with the existing standard gas of known concentration, a standard curve of different gas components is established, which is expressed by the following formula: in, is the chromatographic peak area of ​​the pth gas in the hth region; is the correction factor for the pth gas in the hth region; is the mass of the pth gas in the hth region; Calculate the gas concentration in the hth area based on real-time gas data and the standard curve , use the following formula to calculate: in, is the gas sensitive resistance value of the gas sensor in the hth area in clean air; is the constant of the gas sensor in the hth area; is another constant of the gas sensor in the hth region; is the constant of the hth region; Gas sensor resistance value in clean air in the hth area Use the following formula to express it: According to the multi-level thermal runaway management strategy, by comparing the gas concentration with the preset threshold, a thermal runaway command is issued to complete the thermal safety coordination management of battery energy storage, specifically: Set the minimum threshold of gas concentration to , the maximum threshold is ; like , issued a mild thermal runaway command; like , issued a moderate thermal runaway command; like , issuing a severe thermal runaway command.

[0020] Specifically, the thermal safety coordination management of battery energy storage also includes: When a level 1 command is received, a level 1 warning is issued and a reminder is broadcast; When a level 2 command is received, a level 2 warning is issued and the alarm is sounded; When a mild thermal runaway command is received, a first-level runaway warning is issued and the alarm light starts to light up; When a moderate thermal runaway command is received, a level 2 runaway warning is issued and the warning light starts flashing; When a severe thermal runaway command is received, a level 3 runaway warning is issued, the warning light starts flashing and an alarm sounds.

[0021] The present invention also provides a system for implementing the thermal safety coordinated management method of battery energy storage, and its structural schematic diagram is as follows: Figure 2 As shown, it includes an acquisition module, a prediction module, a monitoring thermal management module, and a runaway thermal management module; The acquisition module acquires the real-time operation data and historical operation data of the battery and uploads the data to the prediction module; The prediction module builds a battery performance degradation prediction model based on the received data and machine learning algorithm, and then combines the historical operation data with the real-time operation data to obtain the predicted battery performance degradation value, and uploads the data to the monitoring thermal management module; The monitoring thermal management module uses acoustic emission and optical fiber sensors to monitor the battery based on the received data, obtains monitoring data, and analyzes the monitoring data to obtain the battery health index. Combined with the predicted battery performance attenuation value, it determines whether to issue a warning instruction based on the preset multi-level thermal management strategy. If no warning instruction is issued, the data is uploaded to the multi-level thermal management module. The runaway thermal management module monitors the battery in different areas based on the received data, uses gas sensors to obtain real-time gas data, and then performs gas chromatography analysis on the real-time gas data. Based on the multi-level thermal runaway management strategy, it issues a runaway warning instruction to complete the thermal safety coordinated management of battery energy storage.

[0022] The present invention is further described below with reference to an embodiment: Select lithium electronic batteries of the same model, keep the initial state consistent under the same working conditions, set the initial temperature of all batteries to 11 degrees, set the experimental time to 10 minutes, and use single thermal management and the method of the present invention to manage the batteries; Among them, single thermal management: Observe and record the peak battery temperature within 10 minutes, such as Figure 3 As shown; The method of the present invention manages: Observe and record the peak battery temperature within 10 minutes, such as Figure 4 As shown; The predicted battery performance attenuation value PAV is calculated based on the formula of the model in step S2. It is assumed that each temperature fluctuation causes 1% performance attenuation. The PAV is 8% under single thermal management within 10 minutes, and the PAV is 7% under the management of the method of the present invention within 10 minutes. According to a study on the lithium-ion battery attenuation model published by the School of Mechanical Engineering of the University of Shanghai for Science and Technology, the experimental research concluded that under the same operating conditions, the battery capacity decayed from 2.6515 Ah to 2.4390 Ah, a decay of 8%.

[0023] Among them, it can be seen that under the management of the method of the present invention, the battery performance attenuation value PAV of the battery is significantly improved compared with the single thermal management. At the same time, compared with the existing literature, the attenuation rate of ten minutes is reduced by 1%, which shows that the battery can be better managed under multi-level thermal management; The battery health index BHI is calculated based on the formula under S3. The BHI is 96% under single thermal management within 10 minutes, and the BHI is 94% under the management of the method of the present invention within 10 minutes. According to the joint estimation method of SOH and RUL of lithium-ion batteries based on health characteristic parameters published by the School of Electrical Automation and Information Engineering of Tianjin University, according to the experimental study, under the same working conditions, the initial capacity of the battery decays from 2010mAh to 1935mAh within 10 minutes, and the health index is 95%; In summary, it can be seen that under the management of the method of the present invention, the battery health index BHI is higher than the battery health index under single thermal management. Compared with the existing literature, the health index is increased by 1%, which shows that under the management of the method of the present invention, the battery can be effectively maintained within the optimal operating temperature range under different ambient temperatures, thereby ensuring the battery health index.

[0024] At the same time, from Figure 3 and Figure 4 It can be seen intuitively that Figure 3 and Figure 4The initial temperature of the batteries is 11 degrees. Both single thermal management and management by the method of the present invention can adjust the battery temperature. Under single thermal management, the battery temperature peaks within 10 minutes are 11 degrees, 18 degrees, 12 degrees and 19 degrees, and the overall peak fluctuation rate is 8 degrees. Under the management of the method of the present invention, the battery temperature peaks within 10 minutes are 11 degrees, 18 degrees, 12 degrees, 17 degrees, 12 degrees, 17 degrees, 13 degrees, 17 degrees, 13 degrees, 17 degrees and 11 degrees, and the overall peak fluctuation rate is 7 degrees. Compared with single thermal management, the management method of the present invention can adjust the battery temperature more timely and can keep the battery within the optimal operating temperature range under different ambient temperatures, thereby ensuring the reliability and safety of the battery.

Claims

1. A thermal safety coordination management method for battery energy storage, characterized in that: The following steps are involved: S1. Obtain real-time and historical operating data of the battery; S2. Based on the machine learning algorithm, a battery performance degradation prediction model is constructed, and the historical operation data and real-time operation data obtained in step S1 are combined to obtain the predicted battery performance degradation value; S3. The battery is monitored using acoustic emission and optical fiber sensors to obtain monitoring data, and the monitoring data is analyzed to obtain a battery health index. Combined with the predicted battery performance attenuation value obtained in step S2, a determination is made based on a preset multi-level thermal management strategy whether to issue a warning instruction. If no warning instruction is issued, the process proceeds to step S4; S4. Monitor the battery in different areas, use gas sensors to obtain real-time gas data, and then perform gas chromatography analysis on the real-time gas data. According to the preset multi-level thermal runaway management strategy, issue a runaway warning instruction to complete the thermal safety coordination management of battery energy storage.

2. The thermal safety coordination management method for battery energy storage according to claim 1 is characterized in that: Step S2 specifically includes the following steps: Construct an initial battery performance degradation prediction model based on a decision tree algorithm; Normalizing the historical operation data and the real-time operation data obtained in step S1; The initial battery performance degradation prediction model is trained using historical operation data to obtain a trained battery performance degradation prediction model; The mean square error and determination coefficient are used as performance indicators of the battery performance degradation prediction model after training to judge the stability of the model, and finally the battery performance degradation prediction model is obtained, and the average performance indicator after training is calculated; Based on the battery performance degradation prediction model, the normalized real-time operating data is input to obtain the predicted battery performance degradation value PAV.

3. The thermal safety coordination management method for battery energy storage according to claim 2 is characterized in that: Step S2 is specifically as follows: A battery performance degradation prediction model is constructed based on the decision tree algorithm. The battery performance degradation prediction model includes B decision trees. Each decision tree is run separately to obtain the result. The final output of the battery performance degradation prediction model is the average value of the output results of all decision trees, expressed using the following formula: in, is the final prediction result of the model; B is the number of decision trees in the model; is the prediction result of the bth decision tree for the input feature X; b is the serial number of a single decision tree; The historical operation data and the real-time operation data obtained in step S1 are normalized and expressed using the following formula: Among them, X is the historical operation data after normalization; For the original historical running data; It is the minimum value in the historical running data; It is the maximum value in the historical operation data; It is the real-time operation data after normalization; It is the original real-time operation data; is the minimum value in the real-time running data; is the maximum value in the real-time running data; The initial battery performance degradation prediction model is trained using historical operation data to obtain a trained battery performance degradation prediction model; The mean square error and determination coefficient are used as performance indicators of the battery performance degradation prediction model after training to judge the stability of the model, and finally the battery performance degradation prediction model is obtained, and the average performance indicator after training is calculated; The mean square error MSE is expressed using the following formula: Where n is the total number of data; is the i-th real-time running data; Coefficient of determination Use the following formula to express it: in, is the average value of real-time running data; Set the mean square error threshold ; like , it means the model is stable; like , it means that the model is unstable. Add L1 regularization term to MSE and retrain to prevent overfitting of the model; New mean square error Use the following formula to express it: Among them, h is the regularization parameter; is the i-th element in the model parameters; Setting the coefficient of determination threshold ; like , it means the model is stable; like , it means that the model is unstable, and the coefficient of determination Add L2 regularization term and retrain to prevent the model from overfitting; New coefficient of determination Use the following formula to express it: ; Based on the battery performance degradation prediction model, the normalized real-time operating data is input to obtain the predicted battery performance degradation value PAV.

4. The thermal safety coordination management method for battery energy storage according to claim 1, characterized in that: Step S3 specifically includes the following steps: Acoustic emission and optical fiber sensors are used to monitor the microscopic activities inside the battery to obtain monitoring data; the monitoring data includes acoustic emission data and optical fiber sensor data; the acoustic emission data is the maximum amplitude of the AE signal and the total energy of the AE signal; the optical fiber sensor data is temperature, strain and pressure; The acoustic emission data and optical fiber sensor data in the monitoring data are analyzed separately, and then the battery health index BHI is obtained through weighted calculation; According to the preset multi-level thermal management strategy, by comparing the health index BHI, the battery performance attenuation value PAV obtained in step S2 with the corresponding preset threshold, it is determined whether to issue a warning instruction. If no warning instruction is issued, step S4 is entered.

5. The thermal safety coordination management method for battery energy storage according to claim 4 is characterized in that: Step S3 is specifically as follows: Divide the battery into n equal areas along its length, marked as arrive ; Acoustic emission sensors and optical fiber sensors are used to arrive n areas for collection; Before the battery starts charging and discharging cycles, record the initial readings of all acoustic emission sensors and optical fiber sensors, and set a detection cycle for periodic sampling; The data collected by the acoustic emission sensor are the maximum amplitude of the AE signal in each area and the total energy of the AE signal in each area; The data collected by the fiber optic sensor are the temperature, strain and pressure of each area; The acoustic emission data collected by the acoustic emission sensor are analyzed to obtain the acoustic emission processing data AEI, which is expressed using the following formula: in, is the AE signal amplitude of the jth region at time t; is the AE signal energy of the jth region at time t; is the maximum value operation for time t; The temperature data collected by the optical fiber sensor is analyzed to obtain the temperature processing data TeI, which is expressed by the following formula: in, is the temperature of the jth region at time t; The strain data collected by the optical fiber sensor is analyzed to obtain the strain processing data StI, which is expressed by the following formula: in, is the strain of the jth region at time t; The pressure data collected by the optical fiber sensor is analyzed to obtain the pressure processing data PrI, which is expressed by the following formula: in, is the pressure of the jth region at time t; The acoustic emission processing data AEI, temperature processing data TeI, strain processing data StI and pressure processing data PrI are weighted to obtain the battery health index BHI, which is expressed by the following formula: in, The weight of acoustic emission processing data AEI; is the weight of temperature processing data TeI; is the weight of strain processing data StI; is the weight of the pressure treatment data PrI; According to the preset multi-level thermal management strategy, by comparing the health index BHI, the battery performance attenuation value PAV obtained in step S2 with the corresponding preset threshold, it is determined whether to issue a warning instruction. The multi-level thermal management strategy is specifically as follows: Set battery health threshold , battery performance attenuation threshold ; when and When the heating mode is started, a first-level instruction is issued; The heating power Hpr of the heating mode is calculated using the following formula: in, is the proportionality coefficient of heating power; is the battery target temperature; is the current temperature of the battery; Battery target temperature Use the following formula to calculate: in, is the basic target temperature of the battery; Adjust the battery target temperature scaling factor for BHI; A scaling factor for adjusting the battery target temperature for PAV; when or When Further judgment: when When the cooling mode is started, a secondary instruction is issued, in which is the maximum threshold of battery temperature; The heat dissipation power Hdp of the heat dissipation mode is calculated using the following formula: in, is the proportionality coefficient of heat dissipation power; when , it is judged as normal mode and no command is issued.

6. The thermal safety coordination management method for battery energy storage according to claim 5, characterized in that: Step S4 specifically includes the following steps: Use gas sensors to monitor the battery in real time to obtain real-time gas data; Use gas chromatograph to analyze the real-time gas data and establish standard curves for different gas components; Calculate gas concentration based on real-time gas data and standard curve; According to the preset multi-level thermal runaway management strategy, by comparing the gas concentration with the preset threshold, a thermal runaway command is issued to complete the thermal safety coordinated management of battery energy storage.

7. The thermal safety coordination management method for battery energy storage according to claim 6, characterized in that: Step S4 is specifically as follows: The area around the battery is divided into u different monitoring areas, and gas sensors are deployed in each monitoring area. For a certain monitoring area h, the real-time gas data is calculated using the following formula: in, is the output voltage of the gas sensor in the hth area; is the supply voltage of the hth region; is the resistance value of the gas sensor in the hth area; is the load resistance of the hth region; For a certain monitoring area h, a gas chromatograph is used for analysis, and combined with the existing standard gas of known concentration, a standard curve of different gas components is established, which is expressed by the following formula: in, is the chromatographic peak area of ​​the pth gas in the hth region; is the correction factor for the pth gas in the hth region; is the mass of the pth gas in the hth region; Calculate the gas concentration in the hth area based on real-time gas data and the standard curve , use the following formula to calculate: in, is the gas sensitive resistance value of the gas sensor in the hth area in clean air; is the constant of the gas sensor in the hth area; is another constant of the gas sensor in the hth region; is the constant of the hth region; Gas sensor resistance value in clean air in the hth area Use the following formula to express it: According to the preset multi-level thermal runaway management strategy, by comparing the gas concentration with the preset threshold, a thermal runaway command is issued to complete the thermal safety coordination management of battery energy storage, specifically: Set the minimum threshold of gas concentration to , the maximum threshold is ; like , issued a mild thermal runaway command; like , issued a moderate thermal runaway command; like , issuing a severe thermal runaway command.

8. The thermal safety coordination management method for battery energy storage according to claim 7, characterized in that: Thermal safety coordination management of battery energy storage also includes: When a level 1 command is received, a level 1 warning is issued and a reminder is broadcast; When a level 2 command is received, a level 2 warning is issued and the alarm is sounded; When a mild thermal runaway command is received, a first-level runaway warning is issued and the alarm light starts to light up; When a moderate thermal runaway command is received, a level 2 runaway warning is issued and the warning light starts flashing; When a severe thermal runaway command is received, a level 3 runaway warning is issued, the warning light starts flashing and an alarm sounds.

9. A system for implementing the thermal safety coordination management method of battery energy storage according to any one of claims 1 to 8, characterized in that: Including acquisition module, prediction module, monitoring thermal management module, and out-of-control thermal management module; The acquisition module acquires the real-time operation data and historical operation data of the battery and uploads the data to the prediction module; The prediction module builds a battery performance degradation prediction model based on the received data and machine learning algorithm, and then combines the historical operation data with the real-time operation data to obtain the predicted battery performance degradation value, and uploads the data to the monitoring thermal management module; The monitoring thermal management module monitors the battery using acoustic emission and optical fiber sensors according to the received data to obtain monitoring data, and analyzes the monitoring data to obtain the battery health index, and then combines the predicted battery performance attenuation value obtained in step S2 to determine whether to issue a warning instruction according to the preset multi-level thermal management strategy. If no warning instruction is issued, the data is uploaded to the multi-level thermal management module; The runaway thermal management module monitors the battery in different areas based on the received data, uses gas sensors to obtain real-time gas data, and then performs gas chromatography analysis on the real-time gas data. Based on the multi-level thermal runaway management strategy, it issues a runaway warning instruction to complete the thermal safety coordinated management of battery energy storage.

Images