AI-based mobile energy storage vehicle operation status monitoring and analysis system

Abstract

The present invention relates to the technical field of mobile energy storage vehicles, and discloses an AI-based operating status monitoring and analysis system for mobile energy storage vehicles. The system includes a main controller, an AI monitoring coprocessor, and a status analysis coprocessor. The main controller retrieves the monitoring and analysis instructions and sends them to the status analysis coprocessor; the AI ​​monitoring coprocessor schedules the status analysis instructions and processes multimodal data; the status analysis coprocessor analyzes the parameters to generate a benchmark monitoring signal, and monitors the battery temperature, charge and discharge efficiency, and load fluctuations in real time, and generates abnormal correction signals and status control instructions in combination with the analysis optimization model. The system realizes accurate analysis of multi-source data and dynamic control of energy storage units through data preprocessing, feature dimensionality reduction, and deep neural network prediction, thereby improving monitoring accuracy, dynamic adaptability, and intelligence level, and ensuring the safe and efficient operation of mobile energy storage vehicles.

Description

Technical Field

[0001] The present invention relates to the technical field of mobile energy storage vehicles, and specifically to an AI-based mobile energy storage vehicle operating status monitoring and analysis system. Background Art

[0002] With the rapid development of new energy technologies, mobile energy storage vehicles, as a flexible energy storage and supply device, have been widely used in emergency power supply, electric vehicle charging, renewable energy consumption and other fields. However, mobile energy storage vehicles face complex working environments and variable load demands during operation. The operating status of their core energy storage units (such as battery packs) directly affects the safety, stability and service life of the entire system. Currently, traditional mobile energy storage vehicle operating status monitoring and analysis systems have the following significant problems:

[0003] During operation, mobile energy storage vehicles need to collect multiple sensor signals, including voltage, current, and temperature. Traditional systems often use a single data processing architecture, making it difficult to effectively integrate and analyze multimodal data. For example, different sensors have different sampling frequencies, data formats, and noise characteristics, resulting in data timing offsets and information distortion, making it impossible to accurately reflect the actual operating status of the energy storage vehicle. Furthermore, traditional methods lack efficient feature extraction and analysis methods for unstructured data, such as thermal imaging signals, making it difficult to achieve precise control of battery thermal management.

[0004] During the charging and discharging process of mobile energy storage vehicles, parameters such as battery temperature, charge and discharge efficiency, and load fluctuations exhibit dynamic changes. Traditional systems typically use a monitoring mode with fixed thresholds, which cannot dynamically adjust monitoring strategies and optimize parameters based on real-time operating conditions. For example, when energy storage vehicles face extreme operating conditions such as high load shocks or transient charge and discharge, traditional systems have difficulty responding quickly and iterating parameters, which can easily lead to battery overcharging, over-discharging, or excessive temperatures, thereby causing safety hazards or shortening battery life. In addition, for different operating scenarios (such as emergency power supply for urban power grids and power supply for field operations), traditional systems lack flexible mode adaptation capabilities and cannot achieve optimal monitoring and analysis results.

[0005] When traditional systems detect abnormal operating conditions in energy storage vehicles, they often only issue simple warning signals, lacking in-depth analysis of the cause of the anomaly and effective correction strategies. For example, when battery temperatures exceed a safe range, traditional systems may simply activate the cooling fan, failing to dynamically compensate for variations in temperature distribution and load fluctuations and fine-tune parameters. Furthermore, traditional systems struggle to perform coordinated analysis and comprehensive processing for complex anomalies driven by multiple parameters (such as simultaneous voltage fluctuations and temperature increases), resulting in ineffective correction and impacting the continued stable operation of the energy storage vehicle.

[0006] Existing mobile energy storage vehicle monitoring systems mostly rely on manually set parameters and empirical judgment, lacking AI-based autonomous learning and predictive capabilities. For example, it's impossible to train models using historical data to predict the state evolution of energy storage units, making it difficult to proactively detect potential faults and perform preventative maintenance. Furthermore, when it comes to multi-module collaboration, the control logic of traditional systems is relatively rigid, preventing intelligent scheduling and efficient collaboration between the main controller, monitoring coprocessor, and analysis coprocessor, limiting overall system performance. Summary of the Invention

[0007] The purpose of the present invention is to provide an AI-based mobile energy storage vehicle operation status monitoring and analysis system to solve the problems raised in the above background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solution: an AI-based mobile energy storage vehicle operating status monitoring and analysis system, the system comprising:

[0009] Main controller, AI monitoring coprocessor, and status analysis coprocessor;

[0010] The main controller is used to retrieve corresponding monitoring and analysis instructions according to the operating status data of the energy storage vehicle, and send the monitoring and analysis instructions to the status analysis coprocessor;

[0011] The AI ​​monitoring coprocessor is used to schedule the state analysis instructions of the state analysis coprocessor when executing the multimodal data parsing instructions;

[0012] The state analysis coprocessor is used to parse the dynamic monitoring parameters in the monitoring and analysis instructions to generate a baseline monitoring signal; perform real-time collaborative monitoring of battery temperature, charge and discharge efficiency, and load fluctuations, and select the corresponding analysis optimization mode for dynamic adaptation based on the monitoring results to generate an abnormal correction signal, and generate a state control instruction for the energy storage unit in combination with the baseline monitoring signal.

[0013] Preferably, the AI ​​monitoring coprocessor includes:

[0014] A data preprocessing module, configured to filter noise from multi-source sensor signals and convert the processed sensor signals into standard state parameters as the reference monitoring signals; wherein the sensor signals include voltage signals, current signals, and thermal imaging signals;

[0015] a feature extraction module, the feature extraction modules being respectively deployed in the dynamic adaptation module and the anomaly correction module, for performing feature dimensionality reduction on the dynamic adaptation module and the anomaly correction module, respectively, to extract a first feature set corresponding to the dynamic adaptation module and a second feature set corresponding to the anomaly correction module;

[0016] A prediction model module is used to generate a time series prediction signal through a deep neural network based on the first feature set and the second feature set, and generate a state control instruction for the energy storage unit in combination with the reference monitoring signal.

[0017] Preferably, the main controller is further configured to retrieve the corresponding preset monitoring threshold as a first target analysis instruction through a parameter storage unit according to the first target monitoring task, and send the first target analysis instruction to the state analysis coprocessor.

[0018] Preferably, the main controller is further configured to use the dynamic optimization algorithm based on timing correlation analysis as a second target analysis instruction according to the second target monitoring task, and send the second target analysis instruction to the state analysis coprocessor.

[0019] Preferably, the analysis and optimization mode includes: bypassing the data verification function, performing global parameter verification, and performing local parameter verification.

[0020] Preferably, the prediction model module includes:

[0021] a dynamic adjustment module, configured to select, based on the first feature set and the second feature set, optimization correction strategies corresponding to the dynamic adaptation module and the abnormality correction module, respectively, to perform parameter iteration and generate the abnormality correction signal; wherein the optimization correction strategies include a load balancing strategy, a temperature suppression strategy, and a charge-discharge smoothing strategy;

[0022] A signal fusion module, configured to perform a spatiotemporal alignment operation on the reference monitoring signal and the abnormal correction signal to generate a final control instruction;

[0023] The instruction issuing module is used to transmit the final control instruction to the control terminal of the energy storage unit.

[0024] Preferably, the dynamic adjustment module includes:

[0025] a first adjustment module, configured to, when detecting that the battery temperature corresponding to the reference monitoring signal exceeds a safe range, superimpose a temperature compensation signal within a current monitoring period to generate a first target adjustment signal;

[0026] The second adjustment module is used to perform trend prediction correction according to the charge and discharge efficiency attenuation curve to generate a second target adjustment signal.

[0027] Preferably, the first adjustment module includes:

[0028] The steady-state monitoring module is used to determine whether the deviation between the current load fluctuation and the expected load exceeds the tolerance threshold when the energy storage unit is monitored to be in steady-state operation. If so, the load balancing strategy is fine-tuned; if not, the original strategy is maintained;

[0029] The transient response module is used to determine whether the difference between the thermal imaging signal and the preset temperature distribution exceeds the critical value before the charge and discharge operation begins when the energy storage unit is monitored to be in transient charge and discharge. If so, dynamic compensation is performed on the temperature suppression strategy; otherwise, no compensation is performed.

[0030] The exception handling module is used to dynamically adapt the charge and discharge smoothing strategy based on voltage fluctuation data when it detects that the energy storage unit is under high load impact, and adjust the smoothing parameters once the voltage fluctuation exceeds the threshold;

[0031] The single parameter adjustment module is used to locally adapt the correction strategy corresponding to a single sensor signal within the monitoring period when a single sensor signal is detected to be abnormal, while maintaining the original correction strategy for the remaining signals.

[0032] Preferably, the second adjustment module includes:

[0033] A prediction configuration module is used to activate the trend prediction function for specific operating conditions; wherein the specific operating conditions include high load cycles, abnormal temperature gradients, and current mutation intervals, each of which is configured with a prediction parameter register and a prediction action period register;

[0034] A log recording module is used to continuously record the parameter evolution data of each prediction node under the specific working condition after the prediction function is activated;

[0035] The nonlinear prediction module is used to perform a nonlinear compensation based on the evolution data when the prediction target is achieved within a specified time period of the specific working condition.

[0036] Preferably, the data preprocessing module further includes a timestamp synchronization unit for performing clock calibration on the sampling frequencies of different sensors to eliminate timing offsets of multi-source data.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The system uses the AI ​​monitoring coprocessor's data preprocessing module to filter noise and calibrate clocks on multi-source sensor signals, eliminating timing offsets in multi-source data. The processed sensor signals are converted into standard state parameters as baseline monitoring signals, ensuring data accuracy and reliability and providing a solid foundation for subsequent monitoring and analysis. The feature extraction module, deployed in the dynamic adaptation module and the anomaly correction module, performs feature dimensionality reduction and extracts the corresponding feature set. Combined with the deep neural network of the prediction model module, it generates time-series prediction signals, enabling forward-looking predictions of energy storage unit status, identifying potential issues and enabling intervention, effectively improving system safety and stability.

[0039] Based on different monitoring tasks, the main controller flexibly retrieves preset monitoring thresholds or dynamic optimization algorithms as analysis instructions and sends them to the state analysis coprocessor. This enables the system to dynamically adjust monitoring strategies based on different operating conditions and monitoring requirements, improving the system's adaptability and flexibility. The state analysis coprocessor analyzes dynamic monitoring parameters to generate a baseline monitoring signal, performing real-time collaborative monitoring of battery temperature, charge and discharge efficiency, and load fluctuations. Based on the monitoring results, it dynamically adapts the corresponding analysis and optimization mode (including data verification function bypass, global parameter verification, and local parameter verification), generates anomaly correction signals, and generates state control instructions based on the baseline monitoring signal. This achieves precise control and dynamic optimization of the energy storage unit's operating status, effectively improving the unit's charge and discharge efficiency and service life.

[0040] The dynamic adjustment module of the predictive model module selects optimized correction strategies (load balancing, temperature suppression, and charge-discharge smoothing) based on the feature set, iterates parameters, and generates anomaly correction signals. The first adjustment module applies corresponding adjustment measures for different operating conditions, including battery temperature exceeding the safe range, steady-state operation of the energy storage unit, transient charge and discharge, high-load shock, and single sensor signal anomalies. This enables refined processing and precise control of various complex operating conditions. The second adjustment module activates trend prediction for specific operating conditions (high load cycles, abnormal temperature gradients, and current mutations), records parameter evolution data, and performs nonlinear compensation, further improving the system's predictive and adaptable capabilities for dynamic conditions. The signal fusion module performs spatiotemporal alignment of the baseline monitoring signal and the anomaly correction signal to generate the final control command. The command issuance module transmits this command to the energy storage unit's control terminal, ensuring the accuracy and timeliness of the control command and achieving efficient control of the energy storage unit.

[0041] Through the application of artificial intelligence technology, the entire system realizes the intelligent analysis, feature extraction and predictive analysis of multimodal data, as well as the collaborative work and intelligent scheduling between modules, which significantly improves the intelligence level of mobile energy storage vehicle operation status monitoring and analysis, and provides a strong guarantee for the safe, stable and efficient operation of mobile energy storage vehicles, with significant economic and social benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a working principle diagram of the AI-based mobile energy storage vehicle operation status monitoring and analysis system described in the present invention;

[0043] Figure 2 This is the design diagram of the AI ​​monitoring coprocessor module;

[0044] Figure 3 The workflow diagram of the state analysis coprocessor;

[0045] Figure 4 Design diagram for dynamically adjusting module logic;

[0046] Figure 5 A design diagram for tuning strategies for predictive models. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 efforts are within the scope of protection of the present invention.

[0048] See also Figure 1-Figure 5 The present invention relates to an AI-based mobile energy storage vehicle operating status monitoring and analysis system, which includes a main controller, an AI monitoring coprocessor, and a status analysis coprocessor. Specifically, it includes the following steps:

[0049] The main controller is configured to retrieve corresponding monitoring and analysis instructions based on the operating status data of the energy storage vehicle and send the monitoring and analysis instructions to the status analysis coprocessor. In this process, the main controller, as the core control hub of the system, obtains various status data of the energy storage vehicle in real time during operation, such as basic information such as battery voltage, current, temperature, charge and discharge efficiency, load changes, etc., and judges the type of monitoring and analysis task that needs to be performed based on preset logical rules or algorithms, and then retrieves the corresponding monitoring and analysis instructions from the internal storage unit or external interaction module. These instructions may include monitoring parameter settings, analysis algorithm selection, control strategy calls, etc. for different operating scenarios. The main controller accurately transmits the instructions to the status analysis coprocessor through standardized communication interfaces and protocols to trigger subsequent analysis and processing processes.

[0050] The AI ​​monitoring coprocessor is used to schedule the state analysis instructions of the state analysis coprocessor in the execution of multimodal data parsing instructions. The coprocessor has the ability to parse and process multi-source heterogeneous data, and can receive data of different modalities such as voltage signals, current signals, thermal imaging signals from multiple sensors, and perform preliminary preprocessing and parsing of these data. During the parsing process, the AI ​​monitoring coprocessor identifies the valid information in the data according to the requirements of the multimodal data parsing instructions, and determines whether it is necessary to call the function of the state analysis coprocessor. For example, when abnormal features are detected in the data or trends that require in-depth analysis, the AI ​​monitoring coprocessor will actively schedule the state analysis instructions of the state analysis coprocessor, and pass the preprocessed key data or feature parameters to the state analysis coprocessor to achieve more in-depth analysis and processing.

[0051] The state analysis coprocessor parses the dynamic monitoring parameters in the monitoring and analysis instructions to generate a baseline monitoring signal. It also performs real-time collaborative monitoring of battery temperature, charge and discharge efficiency, and load fluctuations. Based on the monitoring results, it selects the corresponding analysis and optimization mode for dynamic adaptation, generates anomaly correction signals, and, combined with the baseline monitoring signals, generates state control instructions for the energy storage unit. Specifically, the state analysis coprocessor first parses the monitoring and analysis instructions sent by the main controller, extracting the dynamic monitoring parameters contained therein, such as temperature thresholds, charge and discharge efficiency standards, and load fluctuation tolerances. Based on these parameters, it generates a baseline monitoring signal for subsequent monitoring and control. Simultaneously, it collects real-time data from devices such as battery temperature sensors, charge and discharge efficiency monitoring modules, and load sensors to simultaneously monitor the three key indicators. During the monitoring process, the system continuously compares the actual monitoring data with the baseline monitoring signal. When an anomaly is detected (such as a temperature exceeding a safety threshold, a significant decrease in charge and discharge efficiency, or severe load fluctuations), the system selects an appropriate mode from a set of analysis and optimization modes (such as bypassing the data verification function, global parameter verification, or local parameter verification) based on pre-set rules or AI algorithms. It then dynamically adjusts the monitoring data or control strategy to generate an anomaly correction signal. Finally, the baseline monitoring signal and the abnormal correction signal are fused and processed to form state control instructions for energy storage units (such as battery packs, charging and discharging modules, etc.), and real-time regulation of the operating status of the energy storage vehicle is achieved through the actuator.

[0052] Example 1:

[0053] In this embodiment, the AI ​​monitoring coprocessor includes a data preprocessing module, a feature extraction module, and a prediction model module. The data preprocessing module performs noise filtering on multi-source sensor signals (including voltage, current, and thermal imaging signals). Specifically, it uses digital filtering algorithms (such as Kalman filtering and Butterworth filtering) to remove high-frequency noise and low-frequency drift from the signals, ensuring the accuracy and stability of the sensor signals. The processed sensor signals are then converted into standard state parameters (such as standardized voltage, current, and temperature values) as reference monitoring signals.

[0054] The data preprocessing module first receives raw signals from various sensor types. Voltage signals are typically acquired by high-precision voltage sensors, reflecting changes in the potential difference within the energy storage vehicle's battery pack or other electrical components. Current signals are acquired via Hall-effect sensors or shunt resistors to monitor the current intensity in the circuit. Thermal imaging signals are acquired by infrared thermal imagers, providing images of the temperature distribution of key energy storage vehicle components. Due to factors such as the sensor's inherent electrical characteristics, environmental interference, and electromagnetic interference during signal transmission, the raw signals often contain various noise components.

[0055] The data preprocessing module uses the Kalman filter algorithm for voltage and current signals. The Kalman filter is a recursive optimal estimation algorithm that uses prior knowledge of the system state and current measurements to optimally estimate the system state through two steps: prediction and update. When processing voltage and current signals, the Kalman filter treats the signal as the output of a dynamic system and estimates and predicts the true value of the signal by establishing a state-space model. Specifically, the filter first predicts the current state value based on the previous state estimate and the system model. It then compares the current measured value with the predicted value and calculates the error. Finally, based on the magnitude of the error, it adjusts the predicted value to obtain the optimal estimate for the current moment. By repeating this process, the Kalman filter effectively suppresses random noise in the signal and improves signal quality.

[0056] For thermal imaging signals, due to their spatial distribution characteristics, the data preprocessing module uses the Butterworth filter algorithm. A Butterworth filter is a low-pass filter with a maximally flat amplitude response. It provides the flattest frequency response within the passband while achieving rapid attenuation within the stopband. When processing thermal imaging signals, the Butterworth filter first performs a Fourier transform on the image, converting it from the spatial domain to the frequency domain. It then designs an appropriate low-pass filter in the frequency domain to remove high-frequency noise components. Finally, an inverse Fourier transform is performed to convert the filtered image back to the spatial domain. In this way, the Butterworth filter can effectively remove high-frequency noise from thermal images while preserving image detail.

[0057] After noise filtering, the data preprocessing module converts the processed sensor signals into standard state parameters. Voltage and current signals are converted to standardized values, mapping the actual measured values ​​to a standard numerical range for ease of subsequent analysis and processing. Thermal imaging signals are converted to standardized temperature values. This involves analyzing and processing the thermal image to extract temperature information at each pixel and convert it to actual temperature values. These standardized state parameters serve as baseline monitoring signals for subsequent feature extraction and state analysis.

[0058] The feature extraction modules are deployed in the dynamic adaptation module and the anomaly correction module respectively. Through dimensionality reduction algorithms such as principal component analysis (PCA) and linear discriminant analysis (LDA), the data input into the dynamic adaptation module and the anomaly correction module are processed by feature dimensionality reduction to extract the first feature set corresponding to the dynamic adaptation module and the second feature set corresponding to the anomaly correction module.

[0059] In the dynamic adaptation module, the feature extraction module receives preprocessed baseline monitoring signals, including standardized voltage, current, and temperature values. To extract key features from these data, the feature extraction module uses the principal component analysis (PCA) algorithm to reduce the dimensionality of these features. PCA is an unsupervised learning algorithm that projects the original data into a new, lower-dimensional space through linear transformation, maximizing the variance of the projected data. When processing the baseline monitoring signals, PCA first calculates the data's covariance matrix, then solves for the eigenvalues ​​and eigenvectors of the covariance matrix. It then selects several eigenvectors with the largest eigenvalues ​​to form a projection matrix. Finally, the original data is projected onto this projection matrix to obtain the reduced-dimensional data. Using the PCA algorithm, the feature extraction module extracts the principal components from the baseline monitoring signals that best represent the data's changing trends. These principal components form the first feature set corresponding to the dynamic adaptation module. For example, the first feature set may include principal components reflecting load trends or characterizing temperature distribution. These features help the system better understand the operating status of the energy storage vehicle, enabling dynamic adaptation.

[0060] In the anomaly correction module, the feature extraction module uses the linear discriminant analysis (LDA) algorithm to perform dimensionality reduction on the input data. LDA is a supervised learning algorithm whose goal is to find a projection direction that maximizes the separation of data of different categories in the projected space and the proximity of data of the same category. When processing the baseline monitoring signal, LDA first calculates the inter-class scatter matrix and the intra-class scatter matrix based on known anomaly and normal samples. It then solves the generalized eigenvalue problem to determine the optimal projection direction. Finally, the original data is projected onto this projection direction to obtain the reduced-dimensional data. Using the LDA algorithm, the feature extraction module extracts the features that best distinguish abnormal and normal states from the baseline monitoring signal. These features constitute the second feature set corresponding to the anomaly correction module. For example, the second feature set may include features that characterize the charge and discharge efficiency decay pattern or those that reflect abnormal voltage fluctuations. These features help the system promptly detect anomalies during the operation of the energy storage vehicle and make appropriate corrections.

[0061] The dynamic adaptation module implements dynamic adaptation by receiving the first feature set output by the feature extraction module. This module contains a built-in strategy selection matrix, which pre-stores adaptation strategy templates for different operating conditions, including steady-state operation strategies, transient response strategies, and high-load response strategies. When the first feature set is input, a real-time matching algorithm calculates the similarity between the feature vector and the operating condition label in the strategy matrix and selects the strategy template with the highest match as the base strategy. The parameter fine-tuning unit dynamically adjusts the base strategy parameters based on the feature values. For steady-state operation, if the load fluctuation deviation exceeds a threshold, the load balancing strategy response coefficient is increased by 0.1-0.3. For transient charging and discharging temperature anomalies, the temperature suppression compensation strength is increased by a preset ratio. The adjusted strategy parameters are ultimately packaged as an adaptation strategy instruction output. After receiving the second feature set, the anomaly correction module's anomaly pattern recognizer identifies the anomaly type using a decision tree model constructed from a historical fault database. The strategy generator invokes a library of correction strategies based on the anomaly type. For temperature abrupt changes, a temperature compensation strategy is extracted and the compensation amount is calculated. For efficiency drops, a charge-discharge smoothing strategy is added to dynamically adjust the charge-discharge curve slope. All correction strategies are packaged as instruction outputs. The two modules use a real-time pipeline architecture to update processing every 100 milliseconds and exchange intermediate data through shared memory to avoid repeated calculations.

[0062] Based on the first and second feature sets, the prediction model module uses deep neural networks (such as LSTM and Transformer models) to learn and predict time series data, generating a time series prediction signal that reflects the future trend of the energy storage vehicle's operating status. The prediction model module further combines the baseline monitoring signal with a signal fusion algorithm (such as weighted average fusion and Kalman filter fusion) to generate state control instructions for the energy storage unit, enabling proactive regulation of the energy storage vehicle's operating status.

[0063] The prediction model module first receives the first and second feature sets from the feature extraction module. Since these feature sets typically contain time series data—data sequences arranged in chronological order—the prediction model module uses a long short-term memory (LSTM) network to learn and predict this time series data. LSTM is a special type of recurrent neural network (RNN) that effectively handles long-term dependencies within long sequences of data. By introducing a gating mechanism, including input, forget, and output gates, LSTM selectively memorizes and forgets information, thereby better capturing dynamic changes in time series data. When processing the first and second feature sets, the LSTM first inputs the feature sequence into the network. Through hidden layer processing, it extracts characteristic patterns from the sequence. Based on these patterns, it then predicts future states and generates a time series prediction signal. For example, based on historical load trends and temperature distribution characteristics, the LSTM can predict the load demand and temperature changes of energy storage vehicles over a period of time, providing a basis for proactive system control.

[0064] In some cases, the prediction model module can also use the Transformer model to process time series data. The Transformer model is a deep learning model based on the attention mechanism. It can process sequence data in parallel and capture long-distance dependencies in the sequence. Compared with LSTM, the Transformer model has higher efficiency and better performance when processing long sequence data. When processing the first feature set and the second feature set, the Transformer model first embeds the feature sequence and maps each feature into a low-dimensional vector space. Then, through the multi-head attention mechanism, it calculates the degree of correlation between each position in the sequence. Finally, through the feedforward neural network, it processes the attention-weighted features to generate a time series prediction signal.

[0065] After generating the time series prediction signal, the prediction model module fuses it with the reference monitoring signal. Specifically, the prediction model module uses the Kalman filter fusion algorithm to fuse the time series prediction signal and the reference monitoring signal. The Kalman filter fusion algorithm is an optimal estimation fusion algorithm that can perform weighted fusion of multiple signals based on the statistical characteristics of the signal to obtain the optimal estimation result. During the fusion process, the Kalman filter first calculates the respective weights of the time series prediction signal and the reference monitoring signal based on the variance; then, based on these weights, the two signals are weighted averaged to obtain the fused signal; finally, through the two steps of prediction and update, the fusion result is continuously adjusted to make it more accurately reflect the operating status of the energy storage vehicle. In this way, the prediction model module can combine the advantages of the time series prediction signal and the reference monitoring signal to generate more accurate and reliable state control instructions.

[0066] The generated state control instructions are sent to the energy storage unit's actuators, such as the battery management system (BMS) and charge / discharge controller, to proactively control the energy storage vehicle's operating status. For example, if the predictive model module predicts that the energy storage vehicle's load demand will increase in the future, it can proactively adjust the battery's charge and discharge strategy to increase the battery's output power to meet the load demand. If it predicts that the battery temperature will rise, it can proactively activate the cooling system to reduce the battery temperature and ensure safe operation. Through this proactive control, the system can more efficiently manage the energy storage vehicle's operating status, improving its reliability and performance.

[0067] The data preprocessing module also includes a timestamp synchronization unit. This unit uses hardware clock synchronization (such as the PTP precision clock protocol or GPS clock synchronization) or a software timestamp calibration algorithm to align the sampling frequencies of different sensors. This ensures timeline consistency for multi-source data, eliminates timing offsets caused by sampling time asynchrony, and provides a reliable data foundation for subsequent data analysis and condition monitoring. The timestamp synchronization unit first obtains the sampling timestamps of each sensor and then uses hardware clock synchronization or a software algorithm to align these timestamps to a common time base. For example, when using the PTP precision clock protocol, the timestamp synchronization unit synchronizes the clocks of each sensor to a high-precision master clock, ensuring that all sensor sampling timestamps share the same time base. When using a software timestamp calibration algorithm, the timestamp synchronization unit corrects the sampling timestamps based on communication and processing delays between sensors, eliminating any time offsets caused by these delays. This process ensures that data collected by different sensors is accurately aligned on the timeline, providing a reliable data foundation for subsequent feature extraction and condition analysis.

[0068] Example 2:

[0069] In this embodiment, the main controller also has the function of retrieving preset monitoring thresholds based on a first target monitoring task. Specifically, when the system receives a first target monitoring task (e.g., safety threshold monitoring during normal operation), the main controller retrieves the corresponding preset monitoring thresholds from a parameter storage unit (e.g., an EEPROM, flash memory, or other storage device) as a first target analysis instruction. These thresholds include the upper safety limit for battery temperature, the minimum charge and discharge efficiency standard, and the allowable load fluctuation range. The thresholds include, for example, the upper safety limit for battery temperature, the minimum charge and discharge efficiency standard, and the allowable load fluctuation range. The main controller sends the first target analysis instruction to the state analysis coprocessor via a serial communication interface (e.g., a UART or CAN bus). The state analysis coprocessor compares and analyzes the preset monitoring thresholds with real-time monitored battery temperature, charge and discharge efficiency, and load fluctuation data. If the actual data exceeds the preset thresholds, the coprocessor triggers corresponding alarms or control mechanisms, such as activating the temperature cooling system, adjusting the charge and discharge power, or limiting load access, to ensure that the energy storage vehicle operates within a safe and reliable range.

[0070] As the core control unit of the entire system, the main controller bears the important responsibility of receiving, processing and assigning various tasks. When faced with the first target monitoring task, the main controller first needs to accurately identify the task type. This identification process relies on the system's preset task classification mechanism, which may be designed based on multiple factors such as task priority, time characteristics, functional requirements, etc. For example, a safety threshold monitoring task under normal operating conditions may be marked as a high-priority task, and the system will allocate special processing procedures and resources to it. When the main controller receives the first target monitoring task from external input or internal triggering, it will perform a detailed analysis of the task through the task parsing module, extract the key features and parameters of the task, and determine the specific monitoring requirements.

[0071] After determining the task type, the main controller accesses the parameter storage unit and retrieves the preset monitoring thresholds corresponding to the first target monitoring task. The parameter storage unit is a module within the system specifically used to store various preset parameters and configuration information. It can be implemented using a variety of storage technologies, such as EEPROM (Electrically Erasable Programmable Read-Only Memory) and flash memory. These storage devices are non-volatile and maintain data integrity even after system power is lost. The preset monitoring thresholds are a series of reference values ​​pre-set during the system design and commissioning phase based on the energy storage vehicle's performance parameters, safety standards, and operational requirements. These include the upper limit of battery temperature safety, minimum charge and discharge efficiency standards, and the allowable range of load fluctuations. These thresholds are crucial for ensuring the safe and stable operation of the energy storage vehicle.

[0072] The main controller communicates with the parameter storage unit via an internal data bus, accurately locating and reading the corresponding preset monitoring thresholds based on the task type and requirements. To ensure data accuracy and reliability, the main controller may implement data verification mechanisms such as CRC and parity checks during the reading process to validate the read data. If errors or inconsistencies are detected, the main controller will take appropriate measures, such as re-reading the data or triggering an error alarm.

[0073] Once the preset monitoring thresholds are successfully acquired, the main controller converts them into a unified data format and protocol within the system, forming the first target analysis instruction. This conversion process must take into account the interface requirements and data format specifications of subsequent processing modules to ensure that the instruction can be correctly understood and executed by the state analysis coprocessor. In addition to the preset monitoring thresholds, the first target analysis instruction may also include auxiliary data such as a task identifier, timestamp, and priority information to ensure that the state analysis coprocessor can accurately identify and process the instruction.

[0074] The main controller sends the first target analysis command to the state analysis coprocessor via a serial communication interface. Serial communication interfaces are commonly used for data transmission, offering advantages such as long transmission distances, strong anti-interference capabilities, and low cost. In this system, serial communication technologies such as UART (Universal Asynchronous Receiver / Transmitter) or CAN bus (Controller Area Network) may be used. UART is a full-duplex asynchronous communication interface that enables bidirectional data transmission via transmit (TX) and receive (RX) lines, making it suitable for point-to-point communication scenarios. CAN bus, on the other hand, is a multi-master-slave serial communication bus that uses differential signaling for high reliability and real-time performance, making it suitable for communication between multiple nodes. When sending commands, the main controller packages and encodes them according to the protocol specifications of the selected communication interface, adding control information such as start bits, stop bits, and parity bits to ensure correct data transmission.

[0075] After receiving the first target analysis instruction, the state analysis coprocessor first parses and decodes the instruction, extracting the preset monitoring thresholds and other relevant information. It then acquires real-time monitoring data from various sensors, including battery temperature, charge and discharge efficiency, and load fluctuation data. These sensors are located in key areas of the energy storage vehicle, such as the battery pack, charge and discharge circuits, and load interfaces. They can sense and collect the vehicle's operating status parameters in real time.

[0076] The state analysis coprocessor compares and analyzes real-time monitoring data with preset monitoring thresholds. This analysis process is a multi-dimensional, multi-level comparison process. For battery temperature, the state analysis coprocessor compares the real-time measured temperature value with the preset temperature safety upper limit. If the real-time temperature exceeds the safety upper limit, it means that the battery may be at risk of overheating, and timely measures must be taken to cool it down. For charge and discharge efficiency, the state analysis coprocessor calculates the efficiency value of the current charge and discharge process and compares it with the preset minimum standard. If the efficiency value is lower than the minimum standard, it may mean that there is a fault in the charge and discharge circuit or excessive energy loss, which requires inspection and adjustment. For load fluctuations, the state analysis coprocessor analyzes changes in load current or power to determine whether they are within the preset allowable range. If the load fluctuation exceeds the allowable range, it may cause an impact on the energy storage vehicle's power system, affecting the stability and reliability of the system.

[0077] During the comparative analysis process, the state analysis coprocessor may employ a variety of analytical methods and techniques. For example, it can perform statistical analysis on real-time monitoring data, calculating statistical characteristics such as the mean, standard deviation, maximum, and minimum values, to more comprehensively understand the data's distribution and changing trends. It can also employ time series analysis to model and predict historical data changes, determining whether current anomalies are temporary fluctuations or potential signs of failure. Furthermore, the state analysis coprocessor can incorporate machine learning algorithms to perform pattern recognition and classification on monitoring data, enabling more accurate identification of abnormal conditions and fault types.

[0078] When the state analysis coprocessor detects that real-time monitoring data exceeds a preset monitoring threshold, it triggers the corresponding alarm or control mechanism. Alarm mechanisms can be implemented in a variety of ways, such as audible and visual alarms, SMS alerts, and email alerts. The audible and visual alarm system emits distinct audible and visual signals on the energy storage vehicle's control panel or monitoring center to alert operators to abnormal conditions. SMS and email alerts promptly send abnormality information to the relevant personnel's mobile phones or email addresses, ensuring they are aware of the situation and can take appropriate action.

[0079] The control mechanism will take appropriate control measures based on the type and severity of the abnormal situation. If the battery temperature exceeds the safety upper limit, the state analysis coprocessor will send a control instruction to the temperature cooling system to start the cooling fan or liquid cooling system to speed up the heat dissipation of the battery and reduce the battery temperature. If the charge and discharge efficiency is lower than the minimum standard, the state analysis coprocessor will adjust the charge and discharge power, optimize the charge and discharge strategy, and improve the charge and discharge efficiency. If the load fluctuation exceeds the allowable range, the state analysis coprocessor will limit the load access to avoid the impact of excessive load fluctuations on the system. In some cases, in order to ensure the safety of the system, the state analysis coprocessor may take emergency shutdown measures to cut off the power supply of the energy storage vehicle to prevent the fault from further expanding.

[0080] To ensure the reliability and stability of the entire monitoring and control process, the state analysis coprocessor provides real-time monitoring and feedback on the execution of control instructions. It uses sensors to obtain status information from the actuators and determine whether the control instructions are being correctly executed. If any anomalies are detected during execution, the state analysis coprocessor promptly adjusts the control strategy and implements remedial measures to ensure that the system can resume normal operation as soon as possible.

[0081] In actual applications, the preset monitoring thresholds are not static but need to be dynamically adjusted based on factors such as the energy storage vehicle's operating environment, operating status, and performance changes. The main controller can receive threshold update instructions from the cloud server or operations personnel via a remote communication interface to update and maintain the preset monitoring thresholds in the parameter storage unit. Furthermore, the system can automatically optimize the preset monitoring thresholds through a self-learning mechanism based on long-term operating data and experience, improving the system's adaptability and accuracy.

[0082] Example 3:

[0083] In this embodiment, the main controller can invoke a dynamic optimization algorithm based on a second-target monitoring task. When the system needs to execute a second-target monitoring task (such as optimizing operating conditions under complex operating conditions), the main controller uses a dynamic optimization algorithm based on time-series correlation analysis (such as a particle swarm optimization algorithm or a genetic algorithm) as a second-target analysis instruction and transmits it to the state analysis coprocessor via the data bus. Upon receiving the instruction, the state analysis coprocessor performs time-series correlation analysis on the energy storage vehicle's operating data, identifying time-series dependencies between different parameters (such as the lag correlation between battery temperature changes and charge and discharge currents, and the causal relationship between load fluctuations and voltage response). It then uses the dynamic optimization algorithm to globally optimize the monitoring parameters and control strategies.

[0084] The main controller, as the task scheduling core of the system, first distinguishes between the first target monitoring task and the second target monitoring task through the task identification module. The second target monitoring task usually corresponds to more complex operating conditions, such as high load cycles, abnormal temperature gradients, or sudden current changes. Such tasks need to break through the fixed limitations of preset thresholds and instead use dynamic algorithms to achieve adaptive optimization of parameters. After receiving such task trigger signals (such as abnormal operating condition marks from sensors or manually set optimization instructions), the main controller calls the dynamic optimization algorithm based on time series correlation analysis from the algorithm storage unit. The algorithm storage unit can store a variety of optimization algorithms in the form of a firmware library, including particle swarm optimization (PSO), genetic algorithm (GA), simulated annealing algorithm, etc. Each algorithm corresponds to different parameter optimization scenarios.

[0085] The main controller encapsulates the selected dynamic optimization algorithm as a second-target analysis instruction. The instruction structure includes the algorithm type identifier, optimization target parameters (such as maximizing charge and discharge efficiency and minimizing temperature fluctuations), constraints (such as voltage safety range and load power limit), and timing analysis window parameters (such as the number of historical data sampling cycles). The instruction is transmitted to the state analysis coprocessor via the system data bus (such as the AMBA AHB bus). The data bus supports high-speed parallel transmission, ensuring that the algorithm instruction and the accompanying historical data block reach the target module in real time.

[0086] After receiving the instruction, the state analysis coprocessor first starts the timing correlation analysis module. This module uses Granger Causality Test or Vector Autoregression Model (VAR) to analyze the dependency relationship between parameters based on the timing characteristics of the energy storage vehicle operation data. ) and the charge and discharge current ( ) as an example, by constructing the following lagged regression model:

[0087]

[0088] in, Indicates the current time The battery temperature, Indicates hysteresis The charge and discharge current per time unit, Indicates hysteresis The battery temperature in time units, is a constant term, and is the regression coefficient, is the random error term, Indicates the hysteresis order of the charge and discharge current, Represents the hysteresis order of battery temperature. The hysteresis order of current on temperature can be determined by estimating the coefficient using the least squares method and performing a significance test. , thereby establishing a temporal correlation model between parameters.

[0089] The output of the time series correlation analysis is a correlation matrix containing the causal relationship of parameters and the lag time. This matrix is ​​used as one of the input conditions of the dynamic optimization algorithm. Taking the particle swarm optimization algorithm as an example, the state analysis coprocessor defines the optimization variable as the adjustment factor of the monitoring parameter (such as the charge and discharge power correction coefficient). , temperature threshold offset ) and control policy parameters (such as load balancing weights , cooling system startup threshold ), construct a multi-dimensional search space. Each particle in the particle swarm represents a set of parameter combinations, and its position vector Corresponding to different control strategy configurations.

[0090] Optimize the objective function The design needs to comprehensively consider multi-dimensional indicators. For example, under high load conditions, the objective function can be defined as:

[0091]

[0092] in, is the function of charge and discharge efficiency with respect to the power correction factor, is the standard deviation of battery temperature, is the maximum value of the voltage deviating from the rated value, is the weight coefficient of each indicator, reflecting the optimization focus (for example, when voltage stability is given priority, higher value).

[0093] The particle swarm optimization algorithm searches for the optimal solution that minimizes (or maximizes) the objective function by iteratively updating the position and velocity of the particles. In each iteration, the particles are and the global optimal position Update speed:

[0094]

[0095]

[0096] in, is the inertia weight, which is used to balance the global search and local search capabilities; is the acceleration constant, which controls the particle's tendency to move toward its own optimal and global optimal positions; is a random number between [0,1], introducing randomness to avoid local optimality.

[0097] In each iteration, the state analysis coprocessor substitutes the parameter combination corresponding to the particle into the time series correlation model, predicts the optimized operating state of the energy storage vehicle, and calculates the objective function value. After a preset number of iterations or when the convergence condition is met (such as no significant change in the optimal solution for several consecutive generations), the algorithm terminates and outputs the optimal parameter combination. For example, in the optimization of the charge and discharge smoothing strategy, the power correction coefficient determined by particle swarm optimization is The voltage fluctuation amplitude can be reduced to a certain proportion of the initial value, while ensuring that the charging and discharging efficiency remains in a reasonable range.

[0098] The output of the dynamic optimization algorithm is transmitted via a standardized interface to the control strategy generation module. This module generates specific control instructions based on the optimal parameter combination, such as adjusting the charge and discharge current limits of the battery management system (BMS), updating the fan speed control curve of the thermal management system, and modifying the power scheduling logic of the load distributor. These instructions are then sent to the actuators of the energy storage units via a fieldbus (such as CANopen), achieving global optimization of the energy storage vehicle's operating status.

[0099] To ensure real-time dynamic optimization, the state analysis coprocessor uses hardware acceleration units (such as parallel computing modules implemented in FPGAs) to accelerate the algorithm iteration process, shortening the time required for a single iteration. Furthermore, the system uses an optimization cycle timer to dynamically adjust the optimization frequency based on the complexity of the operating conditions. For example, during steady-state operation, the optimization cycle can be extended to reduce the computational load, while during transient conditions, the cycle can be shortened to quickly respond to parameter changes.

[0100] Example 4:

[0101] In this embodiment, the analysis optimization mode includes three modes: data verification function bypass, global parameter verification, and local parameter verification. The three modes are suitable for different working scenarios, and the automatic switching and execution of the modes are realized through the logic judgment module of the state analysis coprocessor.

[0102] The data verification bypass mode is typically triggered in emergency situations or data transmission anomalies. For example, a mobile energy storage vehicle operating outdoors might encounter a sudden downpour, temporarily interrupting data transmission from some sensors due to moisture, or a sudden load short-circuit failure during charging or discharging, requiring a rapid system response to prevent a safety incident. In this situation, the fault detection unit of the state analysis coprocessor identifies the data anomaly by monitoring the communication link status (such as CAN bus error frame counts) or by missing sensor heartbeat signals. The logic judgment module immediately triggers data verification bypass mode, temporarily skipping the verification of the abnormal sensor data and instead generating control instructions based on the most recently stored valid data (such as historical data from 100ms before the fault). For example, in battery pack charge and discharge control, if a current sensor fails at the moment of a short circuit, the system directly sends a command to the charge and discharge controller to shut down the output based on the pre-fault current value and the preset safety current threshold, avoiding response delays caused by waiting for data verification. In this mode, the system maintains only basic safety control functions while sending data anomaly warnings to maintenance personnel via flashing indicators or remote communication interfaces, prompting them to perform sensor maintenance.

[0103] The global parameter verification mode is suitable for scenarios where the system is operating normally and a comprehensive assessment of the status of the energy storage vehicle is required, such as daily routine inspections, pre-inspections before starting long-term charging and discharging tasks, or status confirmation when switching between operating modes (such as switching from static energy storage mode to mobile driving mode). After receiving a global verification trigger signal (such as scheduled task scheduling, manual verification instructions initiated through the touch screen), the state analysis coprocessor starts a multi-threaded parallel processing mechanism to perform a systematic verification of all monitoring parameters. The verification scope covers core indicators such as battery temperature (temperature sensor data of all battery modules in the vehicle), charge and discharge efficiency (calculation of input and output power of each charge and discharge circuit), load fluctuation (comparison of real-time load current with historical load curves), and also includes consistency verification of multi-modal data such as voltage, current, and thermal imaging.

[0104] Taking battery temperature verification as an example, the system first compares each module's real-time temperature data with the temperature distribution under the same operating conditions in historical operating data. If a module's temperature deviates significantly from the average (e.g., a temperature difference exceeding 5°C), a local thermal runaway warning is triggered for that module. Thermal imaging data is then used to determine whether there is poor contact or battery cell anomaly. During charge and discharge efficiency verification, the state analysis coprocessor calculates the difference between input and output energy based on the principle of energy conservation. If this difference exceeds a preset loss threshold (e.g., greater than 5%), an impedance test of the charge and discharge circuit is initiated to identify potential issues such as circuit aging or increased contact resistance. During the global verification process, the system generates a detailed verification report, documenting the deviation values ​​for each parameter, verification timestamps, and the locations of abnormalities for O&M personnel to review and analyze. Upon completion, if all parameters meet preset standards, the system automatically returns to normal operation. If multiple anomalies are detected, a deeper diagnostic process is triggered, invoking more complex analysis algorithms (such as fault tree analysis) to locate the root cause of the problem.

[0105] The local parameter verification mode precisely responds to abnormal alarms for specific monitoring parameters or subsystems, preventing global verification from interfering with system operations. For example, when an energy storage vehicle is operating under low-load conditions, the temperature sensor of a battery module suddenly sends an over-threshold signal. The single-parameter anomaly detection unit of the state analysis coprocessor first determines whether the signal is a valid anomaly using a signal transition recognition algorithm (e.g., exceeding the threshold for three consecutive sampling periods and a temperature difference greater than 3°C with adjacent modules). If an anomaly is confirmed, the logic judgment module immediately activates the local parameter verification mode, performing a deep verification of only that module and its associated subsystems (e.g., the corresponding BMS sub-controller and cooling fan branch).

[0106] During the specific implementation process, the system first self-calibrates the temperature sensor of the module, and verifies the accuracy of the original data by switching to the redundant backup sensor or calling the historical temperature-voltage correlation model (a mapping relationship between temperature and open-circuit voltage established based on the historical data of the same module). If the temperature is still abnormal after calibration, the system will further check whether the charging and discharging current of the module exceeds the rated value, whether the cooling fan is operating normally, and whether the thermal conductive silicone between the modules is aging or falling off. In a local verification scenario with abnormal load fluctuations, if the current of a load interface suddenly changes beyond the allowable range, the system will isolate the load branch, test its impedance characteristics separately, and compare it with the nominal parameters of the load device to determine whether it is caused by a load device failure or a surge current at the moment of access.

[0107] The execution of the local parameter verification mode is highly targeted and timely. At the hardware level, the state analysis coprocessor uses an independent sub-module controller (such as a parallel processing unit implemented by FPGA) to separately address and collect data for the target subsystem to avoid affecting other normally operating subsystems; at the software level, an incremental verification algorithm is used to conduct in-depth analysis only on the data stream related to the abnormal parameters, rather than repeatedly processing the full amount of data. For example, when processing the anomaly of a single temperature sensor, the system only retrieves the historical data of the module where the sensor is located, the temperature field distribution data of the adjacent modules, and the corresponding heat dissipation control logic, without having to reload the configuration information of all sensors in the vehicle, thereby controlling the verification time to milliseconds.

[0108] The three analysis and optimization modes work together through the state analysis coprocessor's mode management module. The mode management module maintains a priority queue, with data verification bypass mode taking the highest priority (for emergency situations), global parameter verification mode taking second place (for scheduled maintenance needs), and local parameter verification mode serving as the standard response mode. When the system receives multiple mode trigger signals simultaneously, the mode management module arbitrates based on pre-set conflict resolution rules (e.g., "emergency first" and "local over global") to ensure stable system operation under all conditions.

[0109] In practical applications, the switching logic between the three modes can be hardware-accelerated using a field-programmable gate array (FPGA), ensuring that mode switching latency is less than 10ms. For example, when the system detects an anomaly in a subsystem during global parameter verification mode, it can immediately interrupt the global process and switch to local parameter verification mode. Once the local issue is resolved, global verification can be resumed from the breakpoint. This flexible mode switching mechanism enables the system to complete comprehensive health checks under normal operating conditions while responding quickly under abnormal conditions, enabling refined management of the energy storage vehicle's operating status.

[0110] By combining these three analysis and optimization modes, the system can dynamically adjust verification strategies in different scenarios, balancing the requirements of safety, reliability, and operational efficiency. The data verification bypass mode ensures system survivability in emergency situations, the global parameter verification mode supports the establishment of a preventive maintenance system, and the local parameter verification mode enables the precise location and rapid repair of abnormal events. Together, these modes form the intelligent decision-making core of the mobile energy storage vehicle operating status monitoring and analysis system.

[0111] Example 5:

[0112] In this embodiment, the prediction model module includes a dynamic adjustment module, a signal fusion module, and an instruction issuance module. Each module achieves precise control of the energy storage unit through data interaction and logical coordination. The following describes its implementation in detail with reference to specific scenarios:

[0113] 1. Implementation of the Dynamic Adjustment Module

[0114] The dynamic adjustment module selects the corresponding strategy from the set of optimized correction strategies for parameter iteration based on the first feature set (such as load distribution features) and the second feature set (such as temperature anomaly features) output by the feature extraction module. For example, when an energy storage vehicle is connected to a temporary power load at a construction site, the simultaneous activation of multiple power tools may cause load imbalance. The first feature set identifies through principal component analysis that the load power deviation of each energy storage unit exceeds 20%, and the dynamic adjustment module triggers the load balancing strategy. At this time, the module calculates the optimal power distribution coefficient by traversing parameters such as the remaining capacity and charge and discharge efficiency of each battery pack. For example, it transfers part of the power of the high-load unit to the low-load unit, and achieves load balancing by adjusting the current limit of each charge and discharge controller.

[0115] If the second feature set indicates that a battery module's temperature rises by 15°C within five minutes, exceeding the preset temperature rise rate threshold, the dynamic adjustment module invokes the temperature suppression strategy. The specific process is as follows: First, confirm whether the module's corresponding cooling fan is running at full speed. If not, a command is sent to increase the fan speed to 100%. If the fan is already at its maximum speed but the temperature continues to rise, the solenoid valve of the liquid cooling system is further triggered to open, increasing the coolant flow. During this process, the module monitors the temperature change rate in real time and updates the cooling parameters every 10 seconds, forming a closed-loop control.

[0116] If voltage fluctuations exceeding ±5% of the nominal value (e.g., a sudden drop from 48V to 45.6V) are detected during charging or discharging, the dynamic adjustment module activates the charge and discharge smoothing strategy. Taking charging conditions as an example, the module calculates the charging current adjustment step size (e.g., 5A per step) based on the frequency and amplitude of the voltage fluctuations. By gradually reducing the charging current, voltage overshoots or dips are suppressed. If the fluctuations are caused by a sudden increase in load, the module prioritizes powering non-critical loads (e.g., lighting systems) to ensure voltage stability in the energy storage unit.

[0117] 2. Implementation of submodules of dynamic adjustment module

[0118] First adjustment module (temperature and load related adjustment)

[0119] When the energy storage unit is in steady-state operation (e.g., continuously powering a communication base station), the steady-state monitoring module compares the current load fluctuations with the expected load curve (a 24-hour load template generated based on historical data) in real time. If the actual load during a period exceeds the expected value by 15% (e.g., the expected load is 5kW, and the actual load reaches 5.75kW), the power allocation coefficient of the load balancing strategy is fine-tuned (e.g., increasing the power transfer by 10%). If the deviation is within the tolerance range (±10%), the original strategy is maintained.

[0120] During transient charging and discharging scenarios (such as the start-stop phase of fast charging for electric vehicles), the transient response module captures thermal imaging signals the moment the charging plug is plugged in and compares them to a preset temperature distribution model (battery pack temperature uniformity deviation is ≤2°C during normal charging). If the temperature difference in a certain area exceeds 3°C, the temperature suppression strategy is immediately dynamically compensated, for example by increasing the coolant flow rate in the corresponding liquid cooling line by 5%. If the difference does not exceed the critical value, no additional action is taken.

[0121] When an energy storage vehicle encounters a high-load shock (such as when connected to a large electric motor for startup), the abnormal handling module dynamically adapts the charge and discharge smoothing strategy based on voltage fluctuation data (e.g., a voltage drop to 85% of the rated value) and current peaks (exceeding 120% of the rated current). Specific measures include temporarily raising the current limit threshold to 110% of the rated value (for no more than 5 seconds) to allow for a short period of high current flow, and simultaneously activating the battery pack's pre-discharge mechanism to mitigate voltage drops by releasing some stored energy.

[0122] If a single sensor (such as a battery module's voltage sensor) experiences a signal jump (e.g., from 3.6V to 4.2V within 1 second), the single parameter adjustment module locally adapts the correction strategy corresponding to that signal within the monitoring period (e.g., 1 minute). For example, it temporarily switches to redundant sensor data for control and applies a sliding average filter to the historical data of the faulty sensor. If the signal remains abnormal for five consecutive cycles, the sensor is marked as faulty, triggering an operation and maintenance alarm.

[0123] Second adjustment module (trend prediction and compensation)

[0124] The prediction configuration module activates trend prediction for specific operating conditions, such as high load cycles, abnormal temperature gradients, and current mutation intervals. For example, when an energy storage vehicle enters an industrial park to power multiple welding machines, the system identifies periodic fluctuations in load current between 50 and 150A (a high load cycle condition) and automatically activates the prediction parameter register (which stores the number of training cycles for the prediction model, the sliding window size, and so on) and the prediction action period register (which sets predictions for the first 10 seconds of each load cycle).

[0125] After the prediction function is activated, the logging module continuously records parameter evolution data under specific operating conditions. For example, in the case of an abnormal temperature gradient, when a battery cluster experiences a 10°C temperature difference from inlet to outlet (exceeding the 5°C standard under normal operating conditions), the module records data such as the temperature of each battery cell, the corresponding charge and discharge currents, and the cooling fan speed in real time, forming a time series log. This data is used to train a nonlinear prediction model and serves as a historical reference for subsequent similar operating conditions.

[0126] When the nonlinear prediction module reaches its prediction target (e.g., battery temperature is expected to exceed a safety threshold) within a specified time period under specific operating conditions (e.g., 30 minutes before the end of charging), it applies compensation based on evolving data. For example, if the voltage rise rate slows at the end of charging (indicating near-full charge), the module gradually reduces the charging current five minutes in advance to prevent a sudden temperature rise caused by overcharging. This compensation process utilizes an exponential smoothing algorithm, dynamically adjusting the prediction step size based on historical errors to ensure that the lead time of control commands matches actual operating conditions.

[0127] 3. Implementation of the Signal Fusion Module and the Command Issuance Module

[0128] The signal fusion module performs spatiotemporal alignment operations on baseline monitoring signals (such as standardized temperature and voltage data) and abnormal correction signals (such as power adjustment values ​​after load balancing). Taking the scenario of multiple battery packs connected in parallel as an example, the voltage signals of each battery pack may have slight time series offsets (such as a maximum deviation of 5ms) due to differences in sampling clocks. The fusion module uses a timestamp interpolation algorithm (such as linear interpolation) to unify the signals to the global clock reference of the main controller, ensuring that parameters such as voltage, current, and temperature are aligned at the same time. Spatial alignment of thermal imaging signals and temperature sensor data uses a coordinate mapping algorithm (such as a one-to-one correspondence between thermal imaging pixels and the physical location of the sensor) to eliminate spatial errors caused by differences in installation positions.

[0129] The fused signal undergoes logical judgment (such as checking whether the voltage is within the safe range and whether the temperature correction value exceeds the upper limit of the cooling system capacity) to generate the final control instruction. The instruction issuing module transmits the instruction to the control terminal of the energy storage unit via the CAN bus, for example:

[0130] Send a charge and discharge current limit update command to the battery management system (BMS) (for example, adjust from 100A to 80A);

[0131] Send a coolant flow adjustment instruction to the liquid cooling system controller (for example, increase from 5L / min to 8L / min);

[0132] Send power switching instructions to the load distributor (such as switching non-critical loads from the main battery pack to the backup battery pack).

[0133] The command transmission process adopts a priority mechanism. Emergency commands (such as shutdown commands when the temperature exceeds the safety threshold) have the highest transmission priority, ensuring that they are delivered to the control terminal within 1ms; routine adjustment commands (such as load balancing parameter updates) are sent periodically (for example, every 100ms) to reduce bus load.

[0134] 4. The role of the timestamp synchronization unit

[0135] The data preprocessing module's timestamp synchronization unit calibrates the sampling frequencies of sensors such as voltage, current, and thermal imaging sensors using hardware clock synchronization circuits (such as the PTP precision clock protocol). For example, if a voltage sensor uses a 100Hz sampling rate and a thermal imager uses a 25Hz sampling rate, the synchronization unit uses the PTP protocol to ensure that the sampling error between the two is less than 10μs. For sensors that cannot support hardware synchronization (such as older current sensors), a software algorithm performs post-calibration of timestamps. This algorithm compensates for signal transmission delay (average delay measured by ping-pong testing is 2.3ms) and processor processing time (fixed at 0.5ms) to ensure time consistency across multiple data sources.

[0136] Through the coordinated operation of these modules, the predictive model module achieves dynamic prediction and precise control of the energy storage unit's operating status. From real-time load balancing to trend compensation for temperature anomalies, and then to the integration and alignment of multimodal signals, the entire process combines logical judgment with data-driven approaches to ensure the mobile energy storage vehicle maintains stable operation under complex operating conditions, while also providing a solid data foundation for fault warning and optimization strategy iteration.

[0137] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0138] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An AI-based mobile energy storage vehicle operating status monitoring and analysis system, characterized in that: include: Main controller, AI monitoring coprocessor, and status analysis coprocessor; The main controller is used to retrieve corresponding monitoring and analysis instructions according to the operating status data of the energy storage vehicle, and send the monitoring and analysis instructions to the status analysis coprocessor; The AI ​​monitoring coprocessor is used to schedule the state analysis instructions of the state analysis coprocessor when executing the multimodal data parsing instructions; The state analysis coprocessor is used to parse the dynamic monitoring parameters in the monitoring and analysis instructions to generate a reference monitoring signal; perform real-time collaborative monitoring of battery temperature, charge and discharge efficiency, and load fluctuations, and select a corresponding analysis optimization mode for dynamic adaptation based on the monitoring results to generate an abnormality correction signal, and generate a state control instruction for the energy storage unit in combination with the reference monitoring signal; The AI ​​monitoring coprocessor includes: A data preprocessing module, configured to filter noise from multi-source sensor signals and convert the processed sensor signals into standard state parameters as the reference monitoring signals; wherein the sensor signals include voltage signals, current signals, and thermal imaging signals; a feature extraction module, the feature extraction modules being respectively deployed in the dynamic adaptation module and the anomaly correction module, for performing feature dimensionality reduction on the dynamic adaptation module and the anomaly correction module, respectively, to extract a first feature set corresponding to the dynamic adaptation module and a second feature set corresponding to the anomaly correction module; a dynamic adaptation module, configured to dynamically adapt the battery temperature, charge and discharge efficiency, and load fluctuations according to the first feature set, and generate an adaptation strategy instruction; an abnormality correction module, configured to correct abnormal conditions of battery temperature, charge and discharge efficiency, and load fluctuation according to the second feature set, and generate a correction strategy instruction; The prediction model module is used to generate a timing prediction signal through a deep neural network based on the adaptation strategy instruction and the correction strategy instruction, and generate a state control instruction for the energy storage unit in combination with the reference monitoring signal.

2. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 1 is characterized in that: The main controller is further configured to retrieve a corresponding preset monitoring threshold as a first target analysis instruction through a parameter storage unit according to the first target monitoring task, and send the first target analysis instruction to the state analysis coprocessor.

3. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 1 is characterized in that: The main controller is further configured to use the dynamic optimization algorithm based on timing correlation analysis as a second target analysis instruction according to the second target monitoring task, and send the second target analysis instruction to the state analysis coprocessor.

4. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 1 is characterized in that: The analysis and optimization mode includes: bypassing the data verification function, performing global parameter verification, and performing local parameter verification.

5. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 1 is characterized in that: The prediction model module includes: a dynamic adjustment module, configured to select, based on the adaptation strategy instruction and the correction strategy instruction, the optimization correction strategies corresponding to the dynamic adaptation module and the abnormality correction module, respectively, for parameter iteration to generate the abnormality correction signal; wherein the optimization correction strategies include a load balancing strategy, a temperature suppression strategy, and a charge-discharge smoothing strategy; A signal fusion module, configured to perform a spatiotemporal alignment operation on the reference monitoring signal and the abnormal correction signal to generate a final control instruction; The instruction issuing module is used to transmit the final control instruction to the control terminal of the energy storage unit.

6. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 5 is characterized in that: The dynamic adjustment module includes: a first adjustment module, configured to, when detecting that the battery temperature corresponding to the reference monitoring signal exceeds a safe range, superimpose a temperature compensation signal within a current monitoring period to generate a first target adjustment signal; The second adjustment module is used to perform trend prediction correction according to the charge and discharge efficiency attenuation curve to generate a second target adjustment signal.

7. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 6 is characterized in that: The first adjustment module includes: The steady-state monitoring module is used to determine whether the deviation between the current load fluctuation and the expected load exceeds the tolerance threshold when the energy storage unit is monitored to be in steady-state operation. If so, the load balancing strategy is fine-tuned; if not, the original strategy is maintained; The transient response module is used to determine whether the difference between the thermal imaging signal and the preset temperature distribution exceeds the critical value before the charge and discharge operation begins when the energy storage unit is monitored to be in transient charge and discharge. If so, dynamic compensation is performed on the temperature suppression strategy; otherwise, no compensation is performed. The exception handling module is used to dynamically adapt the charge and discharge smoothing strategy based on voltage fluctuation data when it detects that the energy storage unit is under high load impact, and adjust the smoothing parameters once the voltage fluctuation exceeds the threshold; The single parameter adjustment module is used to locally adapt the correction strategy corresponding to a single sensor signal within the monitoring period when a single sensor signal is detected to be abnormal, while maintaining the original correction strategy for the remaining signals.

8. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 6 is characterized in that: The second adjustment module includes: A prediction configuration module is used to activate the trend prediction function for specific operating conditions; wherein the specific operating conditions include high load cycles, abnormal temperature gradients, and current mutation intervals, each of which is configured with a prediction parameter register and a prediction action period register; A log recording module is used to continuously record the parameter evolution data of each prediction node under the specific working condition after the prediction function is activated; The nonlinear prediction module is used to perform a nonlinear compensation based on the evolution data when the prediction target is achieved within a specified time period of the specific working condition.

9. The AI-based mobile energy storage vehicle operation status monitoring and analysis system according to claim 1 is characterized in that: The data preprocessing module also includes a timestamp synchronization unit for performing clock calibration on the sampling frequencies of different sensors to eliminate the timing offset of multi-source data.

Images