High-precision intelligent lithium battery management system for energy storage

Abstract

The invention discloses a high-precision intelligent lithium battery management system for energy storage, and relates to the technical field of lithium batteries, the high-precision intelligent lithium battery management system comprises a system architecture module and a core function module, the system architecture module comprises a local control module, a data analysis cloud platform and a user side module, the core function module comprises a data acquisition and preprocessing module, a high-precision SOC, an SOH estimation module and a bidirectional active equalization module. According to the high-precision intelligent lithium battery management system for energy storage, a three-layer architecture of the local control module, the data analysis cloud platform and the user side module is adopted, and the computing power distribution and data interaction efficiency is greatly optimized; the local master and slave controllers are clear in division of labor, the slave controllers accurately collect cell core parameters and environment working condition data, the master controller quickly organizes and summarizes the data, and the communication module is matched to transfer complex calculation tasks to the cloud, so that the calculation force pressure of a local chip is effectively reduced, and the smoothness of real-time data processing is guaranteed; the data analysis cloud platform carries an evaluation algorithm, and intelligent matching can be carried out according to different scenes.

Description

Technical Field

[0001] This invention relates to the field of lithium battery technology, specifically a high-precision intelligent lithium battery management system for energy storage. Background Technology

[0002] Lithium batteries are a type of battery that uses lithium metal or lithium alloy as the positive / negative electrode material and a non-aqueous electrolyte solution. They are used in many fields such as power and factories. In my country, lithium batteries for energy storage in railway locomotives are replacing lead-acid batteries. Whether it is lithium batteries for locomotives or lithium batteries for energy storage in other fields, they all rely on high-precision intelligent lithium battery management systems.

[0003] Currently, the lithium battery management system for locomotives suffers from significant errors in estimating the state of charge (SOC) and state of health (SOH) due to the complex nonlinear chemical reaction characteristics, process noise, and measurement noise of lithium batteries. Furthermore, the energy conversion efficiency is low due to the large energy dissipation of passive balancing technology and the complex circuit topology and control of active balancing technology. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a high-precision intelligent lithium battery management system for energy storage, which solves the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a high-precision intelligent lithium battery management system for energy storage, comprising a system architecture module and a core functional module. The system architecture module includes a local control module, a data analysis cloud platform, and a user terminal module. The core functional module includes a data acquisition and preprocessing module, a high-precision SOC and SOH estimation module, and a bidirectional active balancing module. The local control module consists of a master controller, a slave controller, and a communication module; The data analysis cloud platform is used to store historical data of the entire battery life cycle, and is equipped with evaluation models and algorithms. It adapts to different scenarios through multi-threaded calculations, supports evaluation of multiple working conditions and multiple types of lithium batteries, and can obtain prediction results through multi-threaded calculations to achieve early warning 100 seconds in advance. The user terminal module is used to display key data such as battery status, warning information, and remaining cycle life to users, and supports visual operation and query to meet the needs of different users. The data acquisition and preprocessing module is used to collect key battery data in real time, providing basic data support for the estimation module and cloud platform, including core cell parameters, environmental and operating condition data; The high-precision SOC and SOH estimation module adopts the EKF-MIAUKF joint estimation method, which combines the extended Kalman filter (EKF) and the multi-novel adaptive unscented Kalman filter (MIAUKF). The EKF algorithm updates the SOH of the battery in real time, and the MIAUKF algorithm estimates the SOC of the lithium battery in real time. The bidirectional active balancing module adopts a master-slave architecture of a centralized controller and a module controller, supports the management of 32 modules and 256 batteries, and is compatible with 1000V high-voltage scenarios. Based on the bidirectional flyback converter, a bidirectional active balancing circuit and control method between modules are designed to achieve efficient bidirectional energy transfer. That is, energy channels are established between modules through the bidirectional flyback converter, supporting bidirectional energy flow between any modules, rather than unidirectional transfer.

[0006] Furthermore, the slave controller module is used to collect cell data, the master controller module is used to receive cell and environmental data from the slave controller, and organize and summarize them, and the communication module is used to hand over computing tasks to the cloud platform to reduce the computing power pressure on the local chip.

[0007] Furthermore, the data analysis cloud platform is equipped with eight evaluation algorithms, including but not limited to EKF-MIAUKF, LSTM+PSO, and LS-SVM, which balance computing speed and cloud platform load according to the scenario matching algorithm.

[0008] Furthermore, the core parameters of the battery cell collected by the data acquisition and preprocessing module include voltage, capacity, and raw SOC / SOH data.

[0009] Furthermore, the environmental and operating condition data collected by the data acquisition and preprocessing module includes temperature and vibration data. It collects temperature and vibration data through temperature and vibration sensors to adapt to complex operating conditions such as locomotives.

[0010] Furthermore, the high-precision SOC and SOH estimation module introduces the Adaptive Unscented Kalman Filter (AUKF) algorithm to achieve adaptive correction of system noise; it also introduces the multiple innovation theory to construct a historical multiple innovation matrix for state estimation, thereby improving the estimation stability under nonlinear conditions and solving the problem of low estimation accuracy of battery state of charge (SOC) and state of health (SOH), with a target estimation error ≤2%.

[0011] Furthermore, the bidirectional active equalization module includes core hardware units such as a sampling circuit, a driving circuit, a control circuit, and a PWM signal-controlled switch.

[0012] Furthermore, the bidirectional active balancing module can provide a constant balancing current of 5A, with an energy conversion efficiency of ≥93%, and a fast balancing speed, which can quickly reduce the voltage difference between modules and avoid battery life degradation caused by untimely balancing.

[0013] Furthermore, the inter-module bidirectional active equalization circuit control method of the bidirectional active equalization module includes the following process: S1. Check battery voltage after start. After the process is started, the voltage at both ends of the battery is checked first; S2. Determine if the voltage is between 30V and 50V. If the voltage is not within the range, return to re-detect the voltage; if the voltage is within the range, proceed to the next step. S3, Estimate Battery SOC Estimate the SOC of the battery pack based on the collected voltage values; S4. Determine if the SOC of the two battery packs are equal. If the SOC is equal, return to re-detect the voltage; if the SOC is not equal, proceed to the next step. S5. Calculate the equalization time and PWM duty cycle. The equalization time is obtained by the ampere-hour integration method, and then the PWM duty cycle is obtained based on the open-circuit voltage; S6, Output PWM and start timing. Output a PWM signal to light up the LED, and simultaneously set the timer threshold and start the timer. The initial timer t=0. S7. Determine if the timer has reached the set value 1. If the target is not reached, continue timing; if the target is reached, proceed to the next step. S8, Stop PWM output and reset timer. Stop outputting PWM, turn off the LED, and reset t to 0; S8. Determine if the timer has reached the set value 2. If the voltage is not reached, continue timing; if it is reached, return to re-detect the voltage and repeat the process.

[0014] This invention provides a high-precision intelligent lithium battery management system for energy storage, which has the following advantages: 1. This high-precision intelligent lithium battery management system for energy storage adopts a three-layer architecture consisting of a local control module, a data analysis cloud platform, and a user-end module, significantly optimizing computing power allocation and data interaction efficiency. The local master and slave controllers have clearly defined roles: the slave controller accurately collects core cell parameters and environmental condition data, while the master controller quickly organizes and summarizes this data. In conjunction with the communication module, complex computational tasks are offloaded to the cloud, effectively reducing the computing power pressure on the local chip and ensuring smooth real-time data processing. The data analysis cloud platform is equipped with eight evaluation algorithms that can intelligently match different scenarios. Through multi-threaded calculations, it provides 100-second early fault warnings and stores full-lifecycle battery data, providing ample data support for battery status assessment and lifespan prediction. The user-end module's visual operation and query functions allow maintenance personnel to intuitively grasp key data such as battery status and warning information, significantly reducing maintenance difficulty and adapting to the operational needs of different users.

[0015] 2. This high-precision intelligent lithium battery management system for energy storage significantly improves the management accuracy of energy storage devices through high-precision SOC and SOH estimation technology. It employs an EKF-MIAUKF joint estimation method, using the EKF algorithm to update the battery SOH in real time and the MIAUKF algorithm to accurately estimate the SOC. Combined with adaptive unscented Kalman filtering and multiple innovation theory, it can achieve adaptive correction of system noise and construction of historical multiple innovation matrices, effectively solving the problem of low SOC and SOH estimation accuracy under nonlinear environments, with a target estimation error ≤2%. Accurate state estimation provides a scientific basis for optimizing charging and discharging strategies, avoiding battery life degradation caused by overcharging and over-discharging, and improving battery safety and reliability. Attached Figure Description

[0016] Figure 1 This is a flowchart of the high-precision SOC and SOH estimation module of a high-precision intelligent lithium battery management system for energy storage according to the present invention; Figure 2 This is a hardware structure diagram of a bidirectional active balancing module controller for a high-precision intelligent lithium battery management system for energy storage according to the present invention. Figure 3 This is a schematic diagram of the overall circuit structure of the bidirectional active balancing module of a high-precision intelligent lithium battery management system for energy storage according to the present invention. Figure 4 This is a flowchart of the bidirectional active balancing module of a high-precision intelligent lithium battery management system for energy storage according to the present invention. Detailed Implementation

[0017] The embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and should not be construed as limiting the scope of the invention.

[0018] like Figures 1-4As shown, the present invention provides a technical solution: a high-precision intelligent lithium battery management system for energy storage, comprising a system architecture module and a core functional module. The system architecture module includes a local control module, a data analysis cloud platform, and a user terminal module. The core functional module includes a data acquisition and preprocessing module, a high-precision SOC and SOH estimation module, and a bidirectional active balancing module. The local control module consists of a master controller, a slave controller, and a communication module. The slave controller module is used to collect battery cell data, the master controller module is used to receive and organize the battery cell and environmental data from the slave controller, and the communication module is used to offload computing tasks to the cloud platform to reduce the computing power pressure on the local chip. The data analysis cloud platform is used to store historical data of the entire battery life cycle, and is equipped with evaluation models and algorithms. It adapts to different scenarios through multi-threaded computing, supports evaluation of multiple working conditions and multiple types of lithium batteries, and can obtain prediction results through multi-threaded computing to achieve early warning 100 seconds in advance. The data analysis cloud platform is equipped with 8 evaluation algorithms, including but not limited to EKF-MIAUKF, LSTM+PSO, and LS-SVM. The algorithm is matched according to the scenario to balance the computing speed and the cloud platform load. The user terminal module is used to display key data such as battery status, warning information, and remaining cycle life to users, and supports visual operation and query to meet the needs of different users. The data acquisition and preprocessing module is used to collect key battery data in real time, providing basic data support for the estimation module and cloud platform, including core cell parameters, environmental and operating condition data; the core cell parameters collected by the data acquisition and preprocessing module include voltage, capacity, and raw SOC / SOH data; the environmental and operating condition data collected by the data acquisition and preprocessing module include temperature and vibration data, which are collected by temperature and vibration sensors to adapt to complex operating conditions such as locomotives; The high-precision SOC and SOH estimation module employs the EKF-MIAUKF joint estimation method, combining Extended Kalman Filter (EKF) and Multi-Innovation Adaptive Unscented Kalman Filter (MIAUKF). The EKF algorithm updates the battery's SOH in real time, while the MIAUKF algorithm estimates the lithium battery's SOC in real time. The high-precision SOC and SOH estimation module introduces the Adaptive Unscented Kalman Filter (AUKF) algorithm to achieve adaptive correction of system noise. It also incorporates multi-innovation theory to construct a historical multi-innovation matrix for state estimation, improving estimation stability under nonlinear environments. This addresses the problem of low estimation accuracy for battery state of charge (SOC) and state of health (SOH), with a target estimation error ≤2%. The bidirectional active balancing module adopts a master-slave architecture of a centralized controller and a module controller, supporting the management of 32 modules and 256 batteries, and is compatible with 1000V high-voltage scenarios. Based on a bidirectional flyback converter, it designs a bidirectional active balancing circuit and control method between modules to achieve efficient bidirectional energy transfer. That is, energy channels are established between modules through the bidirectional flyback converter, supporting bidirectional energy flow between any modules, rather than unidirectional transfer. The bidirectional active balancing module includes core hardware units such as sampling circuit, drive circuit, control circuit, and PWM signal control switch. The bidirectional active balancing module can provide a constant balancing current of 5A, with an energy conversion efficiency of ≥93%, fast balancing speed, and can quickly reduce the voltage difference between modules, avoiding battery life degradation caused by untimely balancing. The inter-module bidirectional active equalization circuit control method of the bidirectional active equalization module includes the following process: S1. Check battery voltage after start. After the process is started, the voltage at both ends of the battery is checked first; S2. Determine if the voltage is between 30V and 50V. If the voltage is not within the range, return to re-detect the voltage; if the voltage is within the range, proceed to the next step. S3, Estimate Battery SOC Estimate the SOC of the battery pack based on the collected voltage values; S4. Determine if the SOC of the two battery packs are equal. If the SOC is equal, return to re-detect the voltage; if the SOC is not equal, proceed to the next step. S5. Calculate the equalization time and PWM duty cycle. The equalization time is obtained by the ampere-hour integration method, and then the PWM duty cycle is obtained based on the open-circuit voltage; S6, Output PWM and start timing. Output a PWM signal to light up the LED, and simultaneously set the timer threshold and start the timer. The initial timer t=0. S7. Determine if the timer has reached the set value 1. If the target is not reached, continue timing; if the target is reached, proceed to the next step. S8, Stop PWM output and reset timer. Stop outputting PWM, turn off the LED, and reset t to 0; S8. Determine if the timer has reached the set value 2. If the voltage is not reached, continue timing; if it is reached, return to re-detect the voltage and repeat the process.

[0019] Example: Taking an urban rail transit energy storage system as an example; Hypothetical scenario: This urban rail transit energy storage system uses 6 groups of 32 modules and 256 ternary lithium batteries. The voltage of a single module is 35V±5V, and the total system voltage is 1000V. It needs to achieve high-precision estimation of battery SOC / SOH, rapid equalization between modules, and early warning of faults. System Architecture Deployment: Local Control Module: Each set of slave controllers corresponds to 256 battery cells, collecting raw data on cell voltage, capacity, and SOC through voltage / current sensors, and simultaneously collecting ambient temperature and vibration acceleration through temperature and vibration sensors; The main controller receives data from 6 sets of slave controllers, performs preliminary processing, and then uploads the full lifecycle data and real-time operating condition data to the data analysis cloud platform via a 4G communication module; Data Analysis Cloud Platform: Stores historical data of the energy storage system for the past 3 years, including charging and discharging, temperature changes, and faults, and is equipped with 8 algorithms such as EKF-MIAUKF, LSTM+PSO, and LS-SVM; User-end Module: Metro maintenance personnel can view real-time SOC, SOH, and balance status through a web interface, and receive overvoltage / undervoltage warnings 100 seconds in advance; The core functional module operation flow is as follows: (I) Data Acquisition and Preprocessing Data is collected from the controller at a frequency of 10ms / time: Key parameters of the battery cell: single cell voltage (3.2-3.6V), total module capacity (200Ah); Environmental and operating condition data: Temperature inside the energy storage cabinet (30℃), vibration data during train braking (2g); After data collection, outliers are automatically filtered out, packaged, and uploaded to the main controller.

[0020] (II) High-precision SOC and SOH estimation A joint EKF-MIAUKF estimation method is adopted, combining Extended Kalman Filter (EKF) and Multi-Innovation Adaptive Unscented Kalman Filter (MIAUKF). The EKF algorithm updates the battery's State of Charge (SOH) in real time, while the MIAUKF algorithm estimates the lithium battery's State of Charge (SOC) in real time. The high-precision SOC and SOH estimation module introduces the Adaptive Unscented Kalman Filter (AUKF) algorithm to achieve adaptive correction of system noise. Furthermore, multi-innovation theory is introduced to construct a historical multi-innovation matrix for state estimation, improving the estimation stability under nonlinear environments. This solves the problem of low estimation accuracy for battery State of Charge (SOC) and State of Health (SOH), with a target estimation error ≤2%.

[0021] (III) Two-way active equilibrium Taking "Module A (SOC=72%) and Module B (SOC=80%)" as an example, the load balancing process is as follows: Detection voltage: Detection module A voltage is 35V, within the 30-50V range; Estimated SOC: Module ASOC = 72%, Module BSOC = 80%; Determine SOC difference: If the SOCs of the two modules are not equal, enter the equalization process; Calculation parameters: The equalization time is 120s according to the ampere-hour integration method, and the PWM duty cycle is calculated to be 40% based on the open-circuit voltage; Output PWM and timing: Output a PWM signal with a 40% duty cycle, light up the equalization indicator, and start the timer (t=0). Timing judgment 1: When t reaches 120s (set value 1), stop PWM output, turn off the indicator light, and reset t to 0; Timing judgment 2: When t reaches 30s (set value 2, wait for voltage to stabilize), return to re-detect the voltage; After equalization, the SOC difference between modules A and B is reduced to ≤1%, and the energy conversion efficiency is 94%. After implementation of this embodiment, the SOC estimation error is ≤1.5% and the SOH estimation error is ≤2%; the inter-module balancing current is stable at 5A, and the voltage difference after balancing is ≤50mV; the fault warning is 100 seconds in advance, effectively avoiding the lifespan degradation caused by battery overcharging / over-discharging, and fully verifying the effectiveness of the present invention.

[0022] Based on the above, this invention adopts a three-layer architecture consisting of a local control module, a data analysis cloud platform, and a user terminal module, significantly optimizing computing power allocation and data interaction efficiency. The local master and slave controllers have clearly defined roles: the slave controller accurately collects core cell parameters and environmental operating condition data, while the master controller quickly organizes and summarizes this data. In conjunction with the communication module, complex computational tasks are transferred to the cloud, effectively reducing the computing power pressure on the local chip and ensuring smooth real-time data processing. The data analysis cloud platform is equipped with eight evaluation algorithms that can intelligently match different scenarios. Through multi-threaded computation, it provides 100-second early fault warnings and stores battery lifecycle data, providing ample data support for battery status assessment and lifespan prediction. The user terminal module's visual operation and query functions allow maintenance personnel to intuitively grasp key data such as battery status and warning information, significantly reducing maintenance difficulty and adapting to the operational needs of different users.

[0023] This invention significantly improves the management accuracy of energy storage devices through high-precision SOC and SOH estimation technology. It employs an EKF-MIAUKF joint estimation method, using the EKF algorithm to update battery SOH in real time and the MIAUKF algorithm to accurately estimate SOC. Combined with adaptive unscented Kalman filtering and multiple innovation theory, it achieves adaptive correction of system noise and construction of historical multiple innovation matrices, effectively solving the problem of low SOC and SOH estimation accuracy under nonlinear environments, with a target estimation error ≤2%. Accurate state estimation provides a scientific basis for optimizing charging and discharging strategies, avoiding battery life degradation caused by overcharging and over-discharging, and improving battery safety and reliability.

[0024] The embodiments of the present invention are given for illustrative and descriptive purposes only, and are not intended to be exhaustive or to limit the invention to the forms disclosed. Many modifications and variations will be apparent to those skilled in the art. The embodiments were chosen and described in order to better illustrate the principles and practical application of the invention, and to enable those skilled in the art to understand the invention and to design various embodiments with various modifications suitable for a particular purpose.

Claims

1. A high-precision intelligent lithium battery management system for energy storage, characterized in that: It includes a system architecture module and a core functional module. The system architecture module includes a local control module, a data analysis cloud platform, and a user terminal module. The core functional module includes a data acquisition and preprocessing module, a high-precision SOC and SOH estimation module, and a bidirectional active balancing module. The local control module consists of a master controller, a slave controller, and a communication module; The data analysis cloud platform is used to store historical data of the entire battery life cycle, and is equipped with evaluation models and algorithms. It adapts to different scenarios through multi-threaded calculations, supports evaluation of multiple working conditions and multiple types of lithium batteries, and can obtain prediction results through multi-threaded calculations to achieve early warning 100 seconds in advance. The user terminal module is used to display key data such as battery status, warning information, and remaining cycle life to users, and supports visual operation and query to meet the needs of different users. The data acquisition and preprocessing module is used to collect key battery data in real time, providing basic data support for the estimation module and cloud platform, including core cell parameters, environmental and operating condition data; The high-precision SOC and SOH estimation module adopts the EKF-MIAUKF joint estimation method, which combines the extended Kalman filter (EKF) and the multi-novel adaptive unscented Kalman filter (MIAUKF). The EKF algorithm updates the SOH of the battery in real time, and the MIAUKF algorithm estimates the SOC of the lithium battery in real time. The bidirectional active balancing module adopts a master-slave architecture of a centralized controller and a module controller, supports the management of 32 modules and 256 batteries, and is compatible with 1000V high-voltage scenarios. Based on the bidirectional flyback converter, a bidirectional active balancing circuit and control method between modules are designed to achieve efficient bidirectional energy transfer. That is, energy channels are established between modules through the bidirectional flyback converter, supporting bidirectional energy flow between any modules, rather than unidirectional transfer.

2. The high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The slave controller module is used to collect battery cell data, the master controller module is used to receive battery cell and environmental data from the slave controller, and organize and summarize them, and the communication module is used to hand over computing tasks to the cloud platform to reduce the computing power pressure on the local chip.

3. The high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The data analysis cloud platform is equipped with eight evaluation algorithms, including but not limited to EKF-MIAUKF, LSTM+PSO, and LS-SVM, which are matched with the scenario to balance computing speed and cloud platform load.

4. The high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The core parameters of the battery cell collected by the data acquisition and preprocessing module include voltage, capacity, and raw SOC / SOH data.

5. A high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The data acquisition and preprocessing module collects environmental and operating condition data, including temperature and vibration data. It collects temperature and vibration data through temperature and vibration sensors to adapt to complex operating conditions such as locomotives.

6. The high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The high-precision SOC and SOH estimation module introduces the adaptive unscented Kalman filter algorithm AUKF to achieve adaptive correction of system noise; Furthermore, the theory of multiple innovations is introduced to construct a historical multiple innovation matrix for state estimation, thereby improving the estimation stability under nonlinear conditions and solving the problem of low estimation accuracy of battery state of charge (SOC) and state of health (SOH), with a target estimation error ≤2%.

7. A high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The bidirectional active equalization module includes core hardware units such as a sampling circuit, a driving circuit, a control circuit, and a PWM signal-controlled switch.

8. A high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that: The bidirectional active balancing module can provide a constant balancing current of 5A, with an energy conversion efficiency of ≥93%. It has a fast balancing speed and can quickly reduce the voltage difference between modules, avoiding battery life degradation caused by untimely balancing.

9. A high-precision intelligent lithium battery management system for energy storage according to claim 1, characterized in that, The inter-module bidirectional active equalization circuit control method of the bidirectional active equalization module includes the following process: S1. Check battery voltage after start. After the process is started, the voltage at both ends of the battery is checked first; S2. Determine if the voltage is between 30V and 50V. If the voltage is not within the range, return to re-detect the voltage; if the voltage is within the range, proceed to the next step. S3, Estimate Battery SOC Estimate the SOC of the battery pack based on the collected voltage values; S4. Determine if the SOC of the two battery packs are equal. If the SOC is equal, return to re-detect the voltage; if the SOC is not equal, proceed to the next step. S5. Calculate the equalization time and PWM duty cycle. The equalization time is obtained by the ampere-hour integration method, and then the PWM duty cycle is obtained based on the open-circuit voltage; S6, Output PWM and start timing. Output a PWM signal to light up the LED, and simultaneously set the timer threshold and start the timer. The initial timer t=0. S7. Determine if the timer has reached the set value. If the target is not reached, continue timing; if the target is reached, proceed to the next step. S8, Stop PWM output and reset timer. Stop outputting PWM, turn off the LED, and reset t to 0; S8. Determine if the timer has reached the set value 2. If the voltage is not reached, continue timing; if it is reached, return to re-detect the voltage and repeat the process.

Images