Aviation power supply vehicle lithium battery thermal runaway monitoring method based on digital twinborn technology

Abstract

The invention provides an aviation power supply vehicle lithium battery thermal runaway monitoring method based on a digital twinborn technology, and belongs to the technical field of aviation power supply vehicle lithium battery thermal runaway monitoring. The technical problem to be solved is to provide the aviation power supply vehicle lithium battery thermal runaway monitoring method based on the digital twinborn technology. According to the technical scheme, a sensor used for collecting voltage, current, frequency and temperature data when a lithium battery works is installed on a to-be-monitored power van, and basic information of the to-be-monitored power van is preset in advance through the installed sensor; a PLC programmable controller is arranged in the data processing area, a three-dimensional model building module, a twinning monitoring module and a lithium battery temperature prediction module are preset in the controller, a data storage module and a data communication module are connected with the controller, and the controller performs data interaction with a remote PC terminal through the data communication module; the method is applied to aviation power supply vehicle lithium battery thermal runaway monitoring.

Description

Technical Field

[0001] The present invention provides a method for monitoring thermal runaway of lithium batteries of aviation power supply vehicles based on digital twin technology, belonging to the technical field of thermal runaway monitoring of lithium batteries of aviation power supply vehicles. Background Art

[0002] At present, when maintaining aircraft at airports, the necessary energy support for maintenance operations is mainly provided by aviation power supply vehicles, which can provide starting power, auxiliary power and emergency power for aircraft. Due to the harsh working environment of the power supply vehicle, higher requirements are also placed on the stability and reliability of the internal lithium battery. When the temperature of the lithium battery is abnormal, the lithium battery may experience thermal runaway, causing the power supply vehicle to catch fire, explode and other accidents, which may cause huge losses to the aircraft. Therefore, the power supply vehicle needs to monitor the working parameters of the lithium battery in real time and predict its temperature change trend. When the predicted data is abnormal, the power supply vehicle also needs to take corresponding fault handling measures to prevent thermal runaway of the battery.

[0003] At present, the lithium battery monitoring system used in power supply vehicles is a simple on-board monitoring panel, that is, the battery temperature is displayed on a real-time digital screen. The operator needs to stay at the monitoring panel for a long time to observe the lithium battery temperature, voltage, current and other working parameters, and cannot remotely monitor in real time; at the same time, the operator cannot access the historical temperature data of the lithium battery, and thus cannot give a prediction of the thermal runaway failure that may occur in the lithium battery in the future after comprehensive analysis. The predictive analysis method used for the working temperature of the lithium battery is relatively simple, and the prediction is made only based on the operator's experience. However, because the thermal runaway speed of the lithium battery is fast and can easily cause serious consequences, it is necessary to realize fault diagnosis and prediction of the lithium battery. Therefore, the prediction of thermal runaway of the lithium battery requires the analysis and processing of a large amount of data to quickly and accurately obtain the prediction conclusion, which is impossible to achieve with current monitoring methods. Summary of the invention

[0004] In order to overcome the deficiencies in the prior art, the present invention aims to solve the technical problem of providing a thermal runaway monitoring method for lithium batteries of aviation power supply vehicles based on digital twin technology.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for monitoring thermal runaway of lithium batteries of aviation power supply vehicles based on digital twin technology, comprising the following monitoring steps:

[0006] Step 1: Install sensors on the power supply vehicle to be monitored to collect voltage, current, frequency, and temperature data of the lithium battery during operation. The installed sensors are pre-set with basic information of the power supply vehicle to be monitored;

[0007] Step 2: Set up a PLC programmable controller in the data processing area, pre-install a 3D model building module, a twin monitoring module, and a lithium battery temperature prediction module in the controller, and connect the data storage module and the data communication module to the controller. The controller exchanges data with the remote PC through the data communication module;

[0008] Step 3: Use the 3D model building module to perform 3D modeling on the lithium battery on the power supply vehicle to be monitored, and display the modeling results on the screen in the data processing area through the twin monitoring module;

[0009] Step 4: Collect the working data of the lithium battery through the sensor and send the working data to the controller in real time. The controller generates a line graph with the historical working data through the twin monitoring module, sends the collected data and the line graph to the data storage module for storage, and sends the collected data and the line graph to the remote PC through the data communication module and displays them on the screen;

[0010] Step 5: Analyze and process the line graph through the lithium battery temperature prediction module, predict and evaluate the change of the current lithium battery operating temperature through the EMD-Informer temperature prediction model, send the evaluation results to the data storage module for storage, and send the evaluation results to the remote PC through the data communication module and display them on the screen;

[0011] Step 6: The controller performs emergency stop and alarm operations through the twin monitoring module based on the evaluation results.

[0012] The three-dimensional model building module is specifically the 3DMax software installed in the controller.

[0013] The UDP user datagram communication protocol is specifically used to transmit data between the remote PC and the controller.

[0014] When sending the collected data to the data storage module for storage, the controller specifically stores the received data in a specified Excel table, stores the time point and value of each data in each row of the table, and stores the data in the order of time points.

[0015] The specific method for predicting and evaluating the change of the current lithium battery operating temperature through the EMD-Informer temperature prediction model is as follows:

[0016] Step 5.1: The controller collects and obtains the lithium battery temperature time series data set;

[0017] Step 5.2: Divide the acquired data set into training set and test set;

[0018] Step 5.3: Decompose the original temperature time series data using the EMD empirical mode to obtain multiple IMF intrinsic mode function components, expressed as:

[0019]

[0020] Where: x(t) is the original temperature time series data, imf f (t) is the fth mutually orthogonal and periodic IMFs, r(t) is the residual term left after decomposition of the original temperature time series data, f is the current decomposition round number, and m is the IMF obtained by decomposition f (t) quantity;

[0021] Step 5.4: Reconstruct the low, medium and high correlation components based on the correlation between IMFs and the original data, and input the components into the sub-model in combination with the characteristic data;

[0022] Step 5.5: Design the architecture of the Informer model, including the input, encoder, decoder, and output parts, and implement a multi-head probabilistic sparse self-attention mechanism by running multiple independent probabilistic sparse self-attention mechanisms in parallel.

[0023] The expression for running the probabilistic sparse self-attention mechanism is:

[0024]

[0025] Where: is the set of the most positive vectors, K is the feature vector in the input sequence, V is the value vector of the weighted output, d is the dimension of K, and softmax(·) is the activation function that converts the output value into a probability distribution;

[0026] Step 5.6: Set the number of layers, number of heads, and learning rate parameters of the Informer model;

[0027] Step 5.7: Select MSE mean square error, MAE mean absolute error, and MAXE maximum absolute error as indicators for evaluating model performance;

[0028] Step 5.8: Train the model on the training set of the Informer model and fine-tune it on the validation set.

[0029] Step 5.9: Evaluate the performance of the model on the test set of the Informer model.

[0030] Step 5.10: Adjust the Informer model architecture or parameters based on the analysis results to further optimize the performance and achieve lithium battery temperature prediction.

[0031] The twin monitoring module also provides a visual monitoring panel, on which a model display interface, a graphic monitoring interface, and an early warning interface for monitoring and operation are provided, wherein:

[0032] The model display interface is used to display the three-dimensional model of lithium battery at any angle;

[0033] The graphic monitoring interface is used to display the real-time temperature of lithium batteries, vehicle operating parameters, data thresholds, historical data and the line graphs generated by them;

[0034] The early warning interface is used to set the thresholds of various parameters and issue an alarm when the real-time monitored data exceeds the thresholds.

[0035] The specific method for the twin monitoring module to perform emergency stop and alarm operations is:

[0036] First, the operator checks the basic information of the power supply vehicle to be monitored on the early warning interface and sets the threshold value based on the basic information;

[0037] When the real-time operating parameter data of the lithium battery of the power supply vehicle or the operating data predicted by the temperature prediction model exceeds the threshold set on the early warning interface, the twin monitoring module sends an emergency stop command to the controller, causing the power supply vehicle to stop working urgently and sound an alarm in the early warning interface.

[0038] The beneficial effects of the present invention compared with the prior art are as follows: the present invention provides an intelligent real-time monitoring solution for thermal runaway of an aviation power supply vehicle, which realizes remote real-time monitoring of parameters such as temperature, voltage, current, etc. of a lithium battery through a PLC communication module of a monitoring system, and provides a line graph to display historical data. It uses low-cost digital twin prediction and visualization technology, and adopts an EMD-Informer prediction model to quickly and accurately predict the working status of a lithium battery, and reflects the status of the lithium battery of the power supply vehicle in real time on a large visualization screen through a wireless communication module. When sensor data or predicted data exceeds a threshold, an emergency stop operation is performed on the system and the power supply vehicle, and an alarm is sounded, effectively avoiding property losses caused by thermal runaway of the lithium battery of the power supply vehicle. The digital twin monitoring system adopted by the present invention can simultaneously perform remote real-time monitoring of the lithium batteries of multiple power supply vehicles, so that one person can monitor the lithium batteries of multiple power supply vehicles, thereby reducing the labor cost during the operation of the power supply vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The present invention will be further described below in conjunction with the accompanying drawings:

[0040] Figure 1 It is a flow diagram of the monitoring method of the present invention;

[0041] Figure 2 A flow chart of the steps of thermal runaway prediction in the present invention;

[0042] Figure 3 It is a working principle diagram of the visual monitoring system of the present invention;

[0043] Figure 4 This is a flow chart of the steps of temperature prediction using the EMD-Informer prediction model in the present invention. DETAILED DESCRIPTION

[0044] like Figures 1 to 4 As shown, the present invention provides a real-time monitoring and prediction solution for thermal runaway of lithium batteries used in aviation power supply vehicles. The solution is implemented based on digital twin technology and mainly includes:

[0045] Model display: The complex internal structure of lithium batteries leads to higher performance of the equipment required for loading the three-dimensional model. Before the lithium battery model is imported into the digital twin monitoring system, the present invention optimizes the surface of the model without losing the display effect, and simplifies the small parts and fasteners that cannot be observed inside. The three-dimensional model building module can be observed at any angle in the model interface, which is convenient for operators to understand the internal structure and working principle of the battery. Due to the complex internal structure of lithium batteries, it is difficult for operators to understand the internal structure of the battery. The present invention optimizes the surface of the three-dimensional model in the 3DMax software and then imports it into the remote PC.

[0046] Data communication: In order to realize remote real-time monitoring, the present invention needs to use a wireless communication protocol to send data to the remote end, and due to the high frequency of sensor acquisition, the amount of data sent and received is huge, and there are few types of communication protocols common to the PLC controller and the remote PC. After comprehensive consideration, the present invention uses the User Datagram Protocol (UDP) to realize the sending of PLC data and the reception of remote PC data. The UDP communication protocol adopted can realize connectionless communication, and has the advantages of large data transmission volume and high transmission efficiency.

[0047] Data visualization: The monitoring panel provided by the twin monitoring module of the present invention can display the lithium battery temperature, historical data, data threshold settings, vehicle operating parameters, etc. in real time. By building multiple data visualization interfaces, it is convenient for staff to operate the digital twin monitoring panel proficiently. In the graphic monitoring interface, a line graph is generated according to the historical temperature data to clearly display the temperature changes of the lithium battery of the power supply vehicle. The thresholds of various parameters can be set in the early warning interface, and the lithium battery model can be displayed at any angle in the model display interface.

[0048] Intelligent data prediction: Since lithium batteries can experience thermal runaway very quickly, the twin monitoring module provided by the present invention needs to predict the operating temperature of the lithium battery to prevent thermal runaway. However, since the battery temperature is strongly correlated with the voltage and current, the present invention uses the EMD-Informer temperature prediction model to predict battery temperature changes. When the data collected by the lithium battery temperature sensor or the battery temperature data predicted by the model is abnormal, the monitoring system controls the power supply vehicle to perform an emergency stop.

[0049] Simultaneous monitoring of multiple power supply vehicles: When the present invention is in use, the sensors provided on multiple power supply vehicles can be uniformly connected to the PLC controller, and real-time monitoring of multiple power supply vehicles can be achieved by switching the monitored power supply vehicles in the monitoring interface.

[0050] When the present invention is used, it is necessary to transmit the collected real-time data such as temperature, voltage, current, etc. to the PLC storage area through the sensor, use the UDP communication protocol to send the data in the storage area to the remote PC end, and the remote PC end stores the data in the monitoring system after receiving the data; when the PLC detects that there is data stored in the storage area, the data will be sent through the UDP communication protocol in the order of position, and the client IP address and port number of the data receiving need to be set inside the PLC, and each data sending time interval is set. When the power supply car restarts, a new database will be created to store the data in the newly created database; the remote PC end receives the data stored inside the PLC through the UDP communication protocol, and the IP address and port number of the sending end are set in the monitoring system to facilitate the reception of data. The receiving frequency is required to be the same as the sending frequency set by the PLC, ensuring that the monitoring system can receive the data completely and prevent the data from accumulating too much at the data sending end; the monitoring system stores the received data in a specified Excel table, stores the time point and value of each data in each row of the table, and stores the data in the order of the time point to facilitate the later reading of the data. When the monitoring system restarts to receive data again, the data in the table is deleted to facilitate the storage of a new round of data.

[0051] When the present invention performs data analysis, the twin monitoring module uses the stored data to display it in real time on a twin monitoring interface similar to the vehicle-mounted monitoring panel, and generates a line chart to display historical data. The temperature data change is predicted by deep learning of the EMD-Informer temperature prediction model. When the predicted temperature exceeds the threshold set by the system, the system performs emergency processing. The temperature data change prediction steps used are specifically as follows:

[0052] (1) Data collection: Obtain lithium battery temperature time series data set.

[0053] (2) Data segmentation: Divide the dataset into training set and test set.

[0054] (3) Empirical mode decomposition (EMD): The original temperature time series data is decomposed by EMD to obtain multiple IMF (intrinsic mode function) components, which are expressed as follows:

[0055]

[0056] Where: x(t) is the original temperature time series data, imf f (t) is the fth mutually orthogonal and periodic IMFs, r(t) is the residual term left after decomposition of the original temperature time series data, f is the current decomposition round number, and m is the IMF obtained by decomposition f (t) Quantity.

[0057] (4) Correlation reconstruction: Based on the correlation between IMFs and the original data, the data are reconstructed into low, medium, and high correlation components, and the components are combined with the feature data and input into the sub-model.

[0058] (5) Informer model architecture: Design the architecture of the Informer model, including the input, encoder, decoder, and output parts, and implement a multi-head probabilistic sparse self-attention mechanism by running multiple independent probabilistic sparse self-attention mechanisms in parallel. The mechanism of probabilistic sparse self-attention is:

[0059]

[0060] Where: is the set of the most positive vectors, K is the feature vector in the input sequence, V is the value vector of the weighted output, d is the dimension of K, and softmax(·) is the activation function that converts the output value into a probability distribution.

[0061] (6) Informer model parameter setting: set the number of model layers, number of heads, and learning rate.

[0062] (7) Evaluation indicators: Mean square error (MSE), mean absolute error (MAE), and maximum absolute error (MAXE) are selected as indicators for evaluating model performance.

[0063] (8) Model training and tuning: The model is trained on the training set and tuned on the validation set.

[0064] (9) Model evaluation: Evaluate the performance of the model on the test set.

[0065] (10) Result analysis and optimization: Adjust the model architecture or parameters based on the analysis results to further optimize the performance and realize lithium battery temperature prediction.

[0066] In order to reduce the learning cost of the operators for the digital twin monitoring system, the present invention builds a twin monitoring interface that is the same as the real power supply vehicle in the twin monitoring module. The monitoring panel can display the voltage, current, frequency, temperature, basic vehicle information, etc. of the power supply vehicle when it is working; in order to display historical data, a graphic monitoring interface is added to the monitoring system, and the historical data is inserted into the line chart according to the chart plug-in XCharts in the graphic monitoring interface to realize dynamic data monitoring. The operator can also judge the working condition of the lithium battery according to the line chart, and the basic information of the monitored vehicle, such as vehicle number, factory information, vehicle status, etc., can be viewed in the early warning interface. The operator can set the threshold according to the basic information of the power supply vehicle; when the value of the working parameter of the power supply vehicle exceeds the threshold set in the early warning interface, the digital twin monitoring system will send an emergency stop command to the PLC controller to control the corresponding power supply vehicle to stop working urgently, and at the same time, an alarm will be sounded in the early warning interface to attract the attention of the operator.

[0067] The EMD-Informer prediction model provided by the present invention has good performance in predicting long-sequence temperatures. The single prediction time is less than 0.01s, which meets the requirement of fast prediction speed. The multi-head sparse self-attention mechanism is used to occupy less resources on the remote PC side during prediction. When the prediction time is within 60s and the absolute error is within ±0.5℃, the prediction accuracy reaches 100%, which meets the requirement of high prediction accuracy.

[0068] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology, characterized in that: The monitoring steps include the following: Step 1: Install sensors on the power supply vehicle to be monitored to collect voltage, current, frequency, and temperature data of the lithium battery during operation. The installed sensors are pre-set with basic information of the power supply vehicle to be monitored; Step 2: Set up a PLC programmable controller in the data processing area, pre-install a 3D model building module, a twin monitoring module, and a lithium battery temperature prediction module in the controller, and connect the data storage module and the data communication module to the controller. The controller exchanges data with the remote PC through the data communication module; Step 3: Use the 3D model building module to perform 3D modeling on the lithium battery on the power supply vehicle to be monitored, and display the modeling results on the screen in the data processing area through the twin monitoring module; Step 4: Collect the working data of the lithium battery through the sensor and send the working data to the controller in real time. The controller generates a line graph with the historical working data through the twin monitoring module, sends the collected data and the line graph to the data storage module for storage, and sends the collected data and the line graph to the remote PC through the data communication module and displays them on the screen; Step 5: Analyze and process the line graph through the lithium battery temperature prediction module, predict and evaluate the change of the current lithium battery operating temperature through the EMD-Informer temperature prediction model, send the evaluation results to the data storage module for storage, and send the evaluation results to the remote PC through the data communication module and display them on the screen; Step 6: The controller performs emergency stop and alarm operations through the twin monitoring module based on the evaluation results.

2. The method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology according to claim 1 is characterized in that: The three-dimensional model building module is specifically the 3DMax software installed in the controller.

3. The method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology according to claim 1 is characterized in that: The UDP user datagram communication protocol is specifically used to transmit data between the remote PC and the controller.

4. The method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology according to claim 1 is characterized in that: When sending the collected data to the data storage module for storage, the controller specifically stores the received data in a specified Excel table, stores the time point and value of each data in each row of the table, and stores the data in the order of time points.

5. The method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology according to claim 1 is characterized in that: The specific method for predicting and evaluating the change of the current lithium battery operating temperature through the EMD-Informer temperature prediction model is as follows: Step 5.1: The controller collects and obtains the lithium battery temperature time series data set; Step 5.2: Divide the acquired data set into training set and test set; Step 5.3: Decompose the original temperature time series data using the EMD empirical mode to obtain multiple IMF intrinsic mode function components, expressed as: Where: x(t) is the original temperature time series data, imf f (t) is the fth mutually orthogonal and periodic IMFs, r(t) is the residual term left after decomposition of the original temperature time series data, f is the current decomposition round number, and m is the IMF obtained by decomposition f (t) quantity; Step 5.4: Reconstruct the low, medium and high correlation components based on the correlation between IMFs and the original data, and input the components into the sub-model in combination with the characteristic data; Step 5.5: Design the architecture of the Informer model, including the input, encoder, decoder, and output parts, and implement a multi-head probabilistic sparse self-attention mechanism by running multiple independent probabilistic sparse self-attention mechanisms in parallel. The expression for running the probabilistic sparse self-attention mechanism is: Where: Q is the set of the most positive vectors, K is the feature vector in the input sequence, V is the value vector of the weighted output, d is the dimension of K, and softmax(·) is the activation function that converts the output value into a probability distribution; Step 5.6: Set the number of layers, number of heads, and learning rate parameters of the Informer model; Step 5.7: Select MSE mean square error, MAE mean absolute error, and MAXE maximum absolute error as indicators for evaluating model performance; Step 5.8: Train the model on the training set of the Informer model and fine-tune it on the validation set. Step 5.9: Evaluate the performance of the model on the test set of the Informer model. Step 5.10: Adjust the Informer model architecture or parameters based on the analysis results to further optimize the performance and achieve lithium battery temperature prediction.

6. The method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology according to claim 1 is characterized in that: The twin monitoring module also provides a visual monitoring panel, on which a model display interface, a graphic monitoring interface, and an early warning interface for monitoring and operation are provided, wherein: The model display interface is used to display the three-dimensional model of lithium battery at any angle; The graphic monitoring interface is used to display the real-time temperature of lithium batteries, vehicle operating parameters, data thresholds, historical data and the line graphs generated by them; The early warning interface is used to set the thresholds of various parameters and issue an alarm when the real-time monitored data exceeds the thresholds.

7. The method for monitoring thermal runaway of lithium batteries in aviation power supply vehicles based on digital twin technology according to claim 6 is characterized in that: The specific method for the twin monitoring module to perform emergency stop and alarm operations is: First, the operator checks the basic information of the power supply vehicle to be monitored on the early warning interface and sets the threshold value based on the basic information; When the real-time operating parameter data of the lithium battery of the power supply vehicle or the operating data predicted by the temperature prediction model exceeds the threshold set on the early warning interface, the twin monitoring module sends an emergency stop command to the controller, causing the power supply vehicle to stop working urgently and sound an alarm in the early warning interface.

Images