Method, device and platform for detecting state of electrically-driven cooling system of new energy automobile

Abstract

The embodiment of the invention provides a new energy automobile electric drive cooling system state detection method, device and platform, and belongs to the technical field of automobile state detection, and the method comprises the steps: obtaining the real-time working data of a vehicle in a plurality of time periods; inputting the real-time working data into a pre-constructed state scoring model to obtain a plurality of state scores of the vehicle; a state of the vehicle is determined based on the distribution of the plurality of state scores. According to the new energy automobile electric drive cooling system state detection method provided by the embodiment of the invention, the state score calculation and statistics are performed on the real-time working data through the state scoring model, the driving data fragment of the faulty automobile can be effectively identified and detected, further deterioration of the fault is prevented, the fault detection accuracy is improved, and the fault detection efficiency is improved. And effective response to complex working conditions is ensured, and a solid data basis is provided for health monitoring of a vehicle electric drive system.

Description

Technical Field

[0001] This application relates to the field of vehicle condition detection technology, and in particular to a method, device and platform for condition detection of electric drive cooling system of new energy vehicle. Background Technology

[0002] Current vehicle monitoring and maintenance methods primarily rely on periodic manual inspections and simple sensor alarms. These methods cannot accurately assess whether the electric drive cooling system is malfunctioning, and are difficult to quickly identify potential problems. Furthermore, the alarm information provided by sensors is usually limited to a single parameter, failing to comprehensively depict the complex condition of the vehicle's electric drive system, resulting in relatively poor accuracy and reliability in vehicle fault detection. Summary of the Invention

[0003] This application provides a method, apparatus, and platform for detecting the status of an electric drive cooling system in a new energy vehicle, in order to solve the aforementioned technical problems.

[0004] In a first aspect, an embodiment of this application provides a method for detecting the state of an electric drive cooling system for a new energy vehicle, comprising:

[0005] Acquire real-time operational data of the vehicle across multiple time periods;

[0006] The real-time working data is input into a pre-built state scoring model to obtain multiple state scores for the vehicle;

[0007] The state of the vehicle is determined based on the distribution of the multiple state scores.

[0008] In conjunction with the first aspect, the method for determining the state of the vehicle based on the distribution of the plurality of state scores includes:

[0009] Determine the proportion of the state scores of the first type among the plurality of state scores;

[0010] If the percentage value is greater than or equal to a first preset threshold, the vehicle is determined to be in a first state;

[0011] If the percentage value is less than or equal to a second preset threshold, the vehicle is determined to be in a second state.

[0012] If the percentage value is less than the first preset threshold and greater than the second preset threshold, the vehicle is determined to be in a third state.

[0013] Wherein, the first state is used to characterize the normal state of the vehicle, the second state is used to characterize the fault state of the vehicle, and the third state is used to characterize the questionable state of the vehicle.

[0014] In conjunction with the first aspect, the method for determining the proportion of the state scores of the first type among the plurality of state scores includes:

[0015] Determine the quantized value of the state score for each time period;

[0016] If the quantized value is greater than or equal to a preset base value, the state score is determined to be of the first type;

[0017] If the quantized value is less than the preset base value, the state score is determined to be of the second type;

[0018] Determine the percentage of the first type of state score among all state scores under the multiple time periods.

[0019] In conjunction with the first aspect, the quantified value includes any one of the mean, variance, standard deviation, and extreme values.

[0020] In conjunction with the first aspect, the method further includes:

[0021] If the vehicle is determined to be in the second state, an alarm is issued.

[0022] In conjunction with the first aspect, the method for constructing the state scoring model includes:

[0023] An initial model is built based on a basic network architecture, which includes any one of the following: isolated forest, radial basis function kernel, autoencoder and its variants, deep embedding clustering, and GMM neural network.

[0024] The acquired historical working data of the vehicle is input into the initial model for training;

[0025] The initial model that has been trained and validated is determined as the state scoring model.

[0026] In conjunction with the first aspect, the method for acquiring real-time operating data of a vehicle across multiple time periods includes:

[0027] Obtain the operating parameters of the vehicle during its operation;

[0028] The operating parameters are preprocessed;

[0029] The motor temperature during vehicle operation is determined based on the preprocessed operating parameters.

[0030] If the motor temperature is greater than or equal to a preset temperature threshold, the operating parameters corresponding to the motor temperature are retained.

[0031] The retained motor temperature and corresponding operating parameters are used as the real-time operating data.

[0032] In conjunction with the first aspect, the method for determining the motor temperature during vehicle operation based on the preprocessed operating parameters includes:

[0033] The operating parameters are input into a pre-trained temperature prediction model to obtain the motor temperature;

[0034] The temperature prediction model is trained based on the vehicle's historical data.

[0035] Secondly, a condition detection device for the electric drive cooling system of a new energy vehicle is provided, comprising:

[0036] The acquisition module is used to acquire real-time working data of the vehicle under multiple time periods;

[0037] An input module is used to input the real-time working data into a pre-built state scoring model to obtain multiple state scores for the vehicle.

[0038] The calculation module is used to determine the state of the vehicle based on the distribution of the plurality of state scores.

[0039] Secondly, a condition monitoring platform for the electric drive cooling system of a new energy vehicle is provided, including:

[0040] A data acquisition module is used to collect real-time operating data of the vehicle.

[0041] A wireless module, which is used to send the real-time working data to the cloud;

[0042] The detection module is used to detect the state of a vehicle using the new energy vehicle electric drive cooling system state detection method as described in any one of the first aspects;

[0043] A sending module is used to send the detection results of the vehicle to the vehicle.

[0044] One of the above technical solutions has the following advantages or beneficial effects:

[0045] This application provides a method for detecting the state of an electric drive cooling system in a new energy vehicle, including: acquiring real-time operating data of the vehicle at multiple time periods; inputting the real-time operating data into a pre-built state scoring model to obtain multiple state scores for the vehicle; and determining the vehicle's state based on the distribution of the multiple state scores. The method for detecting the state of an electric drive cooling system in a new energy vehicle provided by this application uses a state scoring model to calculate and statistically analyze state scores from real-time operating data. This effectively identifies and detects driving data segments of faulty vehicles, preventing further deterioration of the fault. It not only improves the accuracy of fault detection but also ensures effective handling of complex operating conditions, providing a solid data foundation for the health monitoring of the vehicle's electric drive system. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] To gain a more complete understanding of this application and its beneficial effects, the following description will be provided in conjunction with the accompanying drawings, wherein the same reference numerals in the following description denote the same parts.

[0048] Figure 1 A flowchart of a method for detecting the state of a new energy vehicle electric drive cooling system provided in this application embodiment;

[0049] Figure 2 This is a schematic diagram of the motor temperature determination process in the new energy vehicle electric drive cooling system status detection method provided in the embodiments of this application;

[0050] Figure 3 This is a schematic diagram of data transmission in the new energy vehicle electric drive cooling system status detection method provided in the embodiments of this application;

[0051] Figure 4 A schematic diagram illustrating the percentage of the first type of state score provided in this application embodiment;

[0052] Figure 5 A schematic diagram illustrating the comparison results of the proportion values ​​provided in this application embodiment;

[0053] Figure 6 A schematic diagram illustrating three determination structures for the state of an electric drive system provided in an embodiment of this application;

[0054] Figure 7 A schematic diagram of the overall process of the new energy vehicle electric drive cooling system status detection method provided in the embodiments of this application;

[0055] Figure 8 This is a schematic diagram of a module for a new energy vehicle electric drive cooling system status detection device provided in an embodiment of this application;

[0056] Figure 9 A schematic diagram of the data processing flow of the new energy vehicle electric drive cooling system status detection device provided in this application embodiment;

[0057] Figure 10 A schematic diagram of the module of the new energy vehicle electric drive cooling system status detection platform provided in the embodiments of this application;

[0058] Figure 11 A schematic diagram of the data processing flow of the new energy vehicle electric drive cooling system status monitoring platform provided in this application embodiment;

[0059] Figure 12 A line graph illustrating the quantitative value changes of the first testing vehicle provided in this embodiment of the application;

[0060] Figure 13 This is a line graph illustrating the quantitative value changes of the second testing vehicle provided in this embodiment of the application;

[0061] Figure 14 This is a line graph illustrating the quantitative value changes of the third testing vehicle provided in this embodiment of the application.

[0062] Figure 15 This is a line graph illustrating the quantitative value change of the third testing vehicle after maintenance, as provided in an embodiment of this application. Detailed Implementation

[0063] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the protection scope of this application.

[0064] In the description of this application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0065] In the embodiments of this application, "at least one" refers to one or more; "multiple" refers to two or more. In the description of this application, the terms "first," "second," "third," etc., are used only for the purpose of distinguishing descriptions and should not be construed as indicating or implying relative importance, nor should they be construed as indicating or implying order.

[0066] References such as “one embodiment” or “some embodiments” as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the terms “comprising,” “including,” “having,” and variations thereof, as used in this specification, mean “including, but not limited to,” unless otherwise specifically emphasized.

[0067] It should be noted that in the embodiments of this application, "and / or" describes the relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. In addition, the character " / ", unless otherwise specified, generally indicates that the associated objects before and after it are in an "or" relationship.

[0068] It should be noted that in the embodiments of this application, "connection" can be understood as electrical connection. The connection between two electrical components can be a direct or indirect connection between the two electrical components. For example, the connection between A and B can be a direct connection between A and B, or an indirect connection between A and B through one or more other electrical components.

[0069] Currently, the new energy vehicle industry is experiencing rapid development. As a core component of new energy vehicles, the performance of the electric drive system directly affects the vehicle's operating efficiency, stability, and safety. This system typically consists of key components such as a motor, controller, and reducer, which generate a significant amount of heat during operation. Therefore, an efficient cooling mechanism is crucial for the normal operation of the electric drive system.

[0070] Most current new energy vehicles use oil cooling technology to maintain a suitable temperature for their electric drive systems. Specifically, cooling oil flows through the motor and other heat sources via a circulation path, absorbing the generated heat, and then exchanges heat with the coolant through a heat exchanger to lower the oil temperature. This cooling method effectively controls the temperature of the electric drive system, ensuring its stable operation under various complex conditions.

[0071] In practice, electric drive systems may encounter various types of malfunctions, the most common being cooling system failures such as water or oil leaks. These problems not only weaken the performance of the electric drive system but can also pose significant safety hazards. For example, oil leaks can drastically reduce cooling efficiency, leading to motor overheating or even damage; while water leaks not only affect cooling performance but can also cause electrical short circuits, and in extreme cases, even fires. Therefore, timely detection and resolution of these malfunctions are crucial for ensuring the safe operation of vehicles.

[0072] Current vehicle monitoring and fault detection methods mainly include:

[0073] (1) Regular manual inspection: Professional technicians conduct comprehensive inspections of the vehicle according to predetermined cycles, identifying potential problems through visual inspection and manual testing. This method relies on the experience and expertise of the technicians, cannot provide continuous real-time data, and is difficult to quickly identify potential problems;

[0074] (2) Sensor-triggered alarm: Various sensors installed on the vehicle (such as temperature sensors, pressure sensors, etc.) are used to monitor the operating status of the electric drive system and issue an alarm when an abnormality is detected. Although this method can provide immediate feedback, simple sensor alarms usually only reflect one aspect of information and cannot fully capture the complex state of the electric drive system;

[0075] (3) Rule-based diagnostic systems: These systems use pre-defined rules and logic to determine whether a system has a fault. This type of method is suitable for simple and well-defined fault modes, but when dealing with complex multivariate time series data, it lacks an effective data analysis model and struggles to accurately analyze abnormal patterns in multivariate time series, resulting in low accuracy and reliability of fault detection.

[0076] (4) Statistical analysis methods: These methods extract features from the data using statistical tools and then perform fault detection based on these features. These methods are suitable for handling high-dimensional datasets, but their ability to capture non-linear relationships is relatively limited.

[0077] (5) Machine learning model: A fault diagnosis model is established using machine learning algorithms. This type of method can effectively handle complex data relationships, especially nonlinear cases, but requires carefully designed feature engineering to ensure the effectiveness of the model and thus accurately distinguish between normal and fault states;

[0078] (6) Deep learning technology: using deep neural networks for fault diagnosis. Deep learning methods are known for their powerful automatic feature extraction capabilities and nonlinear modeling advantages, but they may suffer from overfitting when the sample size is small.

[0079] In summary, traditional fault detection methods have limitations in terms of real-time performance and comprehensiveness, while emerging data-driven methods offer more advanced solutions, especially demonstrating significant advantages when handling complex, multivariate data. However, each method has its applicable scenarios and technical challenges, and selecting an appropriate fault diagnosis strategy requires comprehensive consideration of the specific application scenario and the feasibility of technical implementation.

[0080] In conclusion, in order to improve the reliability and safety of electric drive systems in new energy vehicles, it is urgent to develop more advanced monitoring and diagnostic technologies to achieve early detection and prevention of faults and ensure the safety of users' travel.

[0081] In response, the relevant technical personnel applied the isolation forest concept to score the vehicle's condition. The characteristics and advantages of this method are reflected in the following aspects:

[0082] (1) Optimized data processing: By deeply mining a large number of vehicle operation records and performing meticulous data cleaning, a high-quality dataset can be obtained, which provides a solid foundation for subsequent analysis and modeling work;

[0083] (2) Unsupervised learning characteristics: Isolation Forest is an unsupervised learning algorithm, particularly suitable for situations where a large number of labeled anomalous samples are lacking. In the operation of electric drive cooling systems, anomalous data generated by faulty vehicles usually accounts for a minority, while normal data from normal vehicles accounts for the majority. Isolation Forest does not rely on a large amount of labeled data for training; it can directly learn "normal" patterns from normal data and effectively identify anomalous points that deviate from these patterns.

[0084] (3) High efficiency in anomaly detection: Isolation forests quickly isolate samples by randomly selecting features and split values. Anomalies are more easily isolated due to their uniqueness, and can therefore be identified with fewer splits. This characteristic makes isolation forests highly sensitive and accurate when dealing with a small number of anomaly samples, making them very suitable for the sparse anomaly data in electric drive cooling systems.

[0085] (4) Ability to process high-dimensional data: Monitoring of electric drive cooling systems involves multiple strongly correlated indicators such as current, voltage, torque, speed, inlet water temperature, and controller temperature, forming a high-dimensional and complex dataset. Isolation forests excel at processing high-dimensional data and can effectively capture the potential relationships between various dimensions without losing information, thereby enabling more accurate anomaly detection;

[0086] (5) Adaptive and Transfer Learning: An adaptive mechanism and fine-tuning strategy are introduced on the basis of isolated forest to further improve the detection effect. By fine-tuning the pre-trained model with a small number of target domain samples, it can be ensured that the model can quickly adapt to new working conditions and maintain a high response to new anomaly patterns. This flexibility and adaptability are particularly important for dynamically changing electric drive cooling systems;

[0087] (6) Reduced computational costs: Compared to other complex deep learning models, Isolation Forest trains faster and requires fewer computational resources. This not only improves detection efficiency but also reduces deployment costs, making it easier to apply on a large scale in real-world industrial environments;

[0088] (7) Real-time early warning mechanism: Using the anomaly detection model established above, the data uploaded from vehicles every day can be analyzed in real time, and judgments can be made accordingly. If a vehicle is identified as having a fault in the data statistics for the day, the system will trigger an alarm and send a notification to the user or maintenance team so that timely action can be taken to prevent the fault from escalating;

[0089] By implementing this technical solution, not only has the speed and accuracy of fault detection in electric drive systems been improved, but the safety and reliability of vehicles have also been enhanced, thereby improving the user's driving experience.

[0090] The specific implementation methods of this application are illustrated below through specific embodiments:

[0091] like Figure 1 As shown in the figure, this application provides a method for detecting the status of an electric drive cooling system in a new energy vehicle, including:

[0092] S1: Obtain real-time operating data of the vehicle at multiple time periods.

[0093] The specific methods include: collecting operating parameters of the vehicle's electric drive system during operation using various sensors installed on the vehicle. These parameters include inlet water temperature, controller temperature, motor speed, motor torque, current, and voltage. Specifically, inlet water temperature and controller temperature can be collected using multiple temperature sensors; motor speed can be collected using any one of the following: photoelectric encoder, Hall effect sensor, magnetoelectric speed sensor, rotary transformer, and laser speed sensor; motor torque can be collected using any one of the following: strain gauge torque sensor, inductive / magnetoelastic torque sensor, piezoelectric torque sensor, fiber optic torque sensor, and multifunctional torque-speed sensor; motor current can be collected using any one of the following: Hall effect current sensor, current transformer, shunt resistor sensor, and magnetically modulated current sensor; and motor voltage can be collected using any one of the following: voltage divider resistor sensor, voltage transformer, Hall voltage sensor, and fiber optic voltage sensor. It should be noted that the frequency range for sensor parameter acquisition is from 1 second to 60 seconds, specifically 1 second, 2 seconds, 3 seconds, 5 seconds, 8 seconds, or 10 seconds, etc. The sampling frequency of the working parameters can be adjusted according to the specific application scenario, and this application embodiment does not impose too many restrictions here.

[0094] It is worth noting that the embodiments of this application only exemplify some common sensors used to collect operating parameters. The specific sensor used to collect operating parameters can be adjusted according to the actual situation. The embodiments of this application do not exhaustively list the types of sensors.

[0095] In this embodiment, after acquiring the operating parameters, they are uploaded to a cloud server using IoT and wireless communication technologies, where the data is stored. This ensures timely and accurate capture of the vehicle's operating status, providing reliable support for subsequent data analysis.

[0096] In this embodiment, after obtaining the working parameters, data cleaning is required to remove any missing parameters. Simultaneously, the working parameters are standardized or normalized to eliminate dimensional differences between different parameters.

[0097] In this embodiment of the application, after preprocessing the operating parameters, the motor temperature during vehicle operation is determined based on the preprocessed operating parameters.

[0098] The specific method includes: inputting operating parameters into a pre-trained temperature prediction model to obtain the motor temperature; wherein, the temperature prediction model is trained based on historical vehicle data. The temperature prediction model can predict the motor temperature at the end of each time period based on the acquired operating parameters. The motor temperature is used to filter out effective data for vehicle condition detection, especially high-temperature data, thereby enhancing the condition detection model's sensitivity to identifying abnormal data.

[0099] like Figure 2 As shown, it's worth noting that the method for filtering operating parameters based on motor temperature includes: retaining the operating parameters corresponding to the motor temperature when the motor temperature is greater than or equal to a preset temperature threshold; and using the retained motor temperature and corresponding operating parameters as real-time operating data. It should be noted that in the early stages of a vehicle's motor system malfunction, abnormal temperature increases in certain areas or components are common. Therefore, when the motor temperature is greater than or equal to the preset temperature threshold, this data is considered suitable for further processing and analysis. Conversely, if the motor temperature is less than the preset temperature threshold, this data is insufficient to effectively distinguish between the vehicle's normal and faulty states and is therefore discarded.

[0100] Understandably, by predicting the corresponding motor temperature based on the collected operating parameters, comparing the motor temperature with a preset temperature threshold, retaining the operating parameters exceeding the temperature threshold, and removing the operating parameters below the temperature threshold, data segments that cannot effectively distinguish between fault and normal states are eliminated, thereby reducing unnecessary consumption of computing resources and improving the model's judgment accuracy.

[0101] S2: Input real-time working data into a pre-built state scoring model to obtain multiple state scores for the vehicle.

[0102] Specifically, the method for constructing the state scoring model includes: building an initial model based on a basic network architecture, which can be any one of isolated forest, radial basis function kernel, autoencoder and variants, deep embedding clustering, and GMM neural network; inputting the acquired historical working data of the vehicle into the initial model for training; and determining the initial model that has been trained and validated as the state scoring model.

[0103] In this embodiment, the Isolation Forest is constructed by building multiple isolation trees to form a forest. Each tree is created by recursively partitioning the data until all data points are completely isolated. In an isolation tree, outliers are usually closer to the root because they are easily isolated. Therefore, measuring the distance from a data point to the root can assess whether the point is an outlier. The main application scenarios in this embodiment include: real-time monitoring of parameters such as motor current, voltage, speed, torque, and temperature, and using the Isolation Forest algorithm to identify patterns of change in these parameters. The appearance of outliers may indicate an internal fault or impending fault in the motor.

[0104] Correspondingly, the Radial Basis Function Kernel (RBF) uses radial basis functions (such as Gaussian kernels) to map the input to a high-dimensional space, achieving classification / regression through linear combination. The core formula is: Among them, ||xy|| 2 σ is the square of the Euclidean distance between the two points, and σ is the kernel width parameter. The main application scenarios in the embodiments of this application include: in electric drive system fault diagnosis, the RBF kernel is used for classification tasks such as support vector machines to improve the nonlinear classification ability of the model.

[0105] Correspondingly, an autoencoder (AE) is an unsupervised learning method used to learn efficient encoding of data. An autoencoder consists of an encoder and a decoder, and its goal is to minimize reconstruction error. Variations of autoencoders achieve similar functions. The main application scenarios in this application include: in electric drive system fault diagnosis, training an autoencoder on normal and fault data can distinguish faults in the vehicle's electric drive system.

[0106] Correspondingly, Deep Embedded Clustering (DEC) is a deep learning-based clustering method that clusters data by learning its embedded representation. The goal of DEC is to minimize the distance between the embedded representation and the cluster centers. In the embodiments of this application, the main application scenarios include: in the fault diagnosis of vehicle electric drive systems, DEC is used to perform cluster analysis on high-dimensional data to identify different fault modes.

[0107] Accordingly, the GMM neural network combines a Gaussian Mixture Model (GM) with the neural network, including: a compression network: using an autoencoder to reduce dimensionality and calculate the reconstruction error; and an estimation network: using the neural network to predict the posterior probability of the GMM and jointly optimizing the reconstruction loss and the GMM likelihood. The optimization objective is: L = reconstruction error + λGMM likelihood term. The main application scenarios in this application include: acquiring data from a vehicle's electric drive system and performing cluster analysis on the high-dimensional data to identify different fault modes.

[0108] Based on the above, this application's embodiments employ an isolated forest to construct a state scoring model. For example... Figure 3 As shown, on-board sensors collect time-window data that meets temperature thresholds generated during daily vehicle operation. This data is transmitted via a wireless communication module and received by a server. The data is then forwarded to a cloud-based data preprocessor via a forwarding module. After data cleaning, the preprocessor uses a temperature prediction model to predict the motor temperature. The motor temperature and the collected data are then input into an isolated forest model to obtain the vehicle's state score. The daily average state score is then calculated. By using a pre-fitted and trained isolated forest model, state scores are calculated on real-time window data. The more windows accumulated, the more stable the average state score becomes, and the better it reflects the current state of the vehicle's electric drive system. This is because the outlier score output by the isolated forest model provides a quantitative indicator for each sample, assessing the likelihood of a fault. A positive value indicates a normal vehicle tendency, while a negative value indicates a fault tendency. The specific value reflects the degree to which the sample deviates from the normal distribution.

[0109] It is worth noting that, for a single vehicle, the average state score calculated by the model over multiple consecutive days is statistically integrated with the state score index for reliable analysis. This effectively avoids misjudgments caused by accidental factors such as sudden severe operating conditions or adverse driving environments on a single day, which may prevent the state score from accurately reflecting the vehicle's cooling system status. Statistical analysis of state scores over multiple consecutive days can effectively filter out noise information.

[0110] It should be noted that after calculating the vehicle's status score over multiple days, the distribution of the status score will fall into three categories, corresponding to the three vehicle statuses:

[0111] Normal status confirmed:

[0112] The electric drive system is considered to be in normal working condition when the number of days with a positive average status score significantly exceeds the number of days with a negative average. That is, by comparing the number of days with a positive average status score and the number of days with a negative average, if the number of positive days significantly exceeds the number of negative days, the system is considered to be operating well. Continue operation only when the system is indeed functioning normally to avoid unnecessary maintenance interventions.

[0113] Fault status confirmation:

[0114] A fault is identified in the electric drive system when the number of days with a positive daily average status score is significantly less than the number of days with a negative daily average. This is achieved by comparing the number of days with a positive and negative daily average status score; if the number of positive days is significantly less than the number of negative days, a system problem is considered to have occurred. Rapid fault identification and response prevent potential problems from escalating.

[0115] Status in doubt:

[0116] When the number of positive and negative daily average status scores is close and difficult to distinguish, no conclusion is drawn temporarily. That is, if the number of positive and negative daily average status scores is roughly equal and insufficient to form a clear judgment, the current data is considered insufficient to support a final conclusion. Specifically, when the ratio of the number of days with positive to the number of days with negative daily average status scores differs by ±5%, it is determined that the two are roughly equal. Setting up a questionable status reduces the risk of misjudgment of vehicles, and the true status of the system can be more accurately confirmed through accumulating more data or manual verification.

[0117] Understandably, by using the isolated forest model to output state scores under multiple time windows and combining them with statistical methods, a more effective quantitative indicator of the current state of the vehicle's electric drive system is obtained. This leads to a more accurate judgment of the vehicle's electric drive system state, enabling timely detection and handling of faults while avoiding premature conclusions under uncertain circumstances, thereby improving the reliability and safety of diagnosis.

[0118] S3: Determine the vehicle's state based on the distribution of multiple state scores. Specific methods include:

[0119] S31: Determine the percentage of the first type of state scores among multiple state scores.

[0120] like Figure 4As shown, the method includes: determining the quantized value of the state score for each time period, where the quantized value includes any one of the mean, variance, standard deviation, and extreme values; determining the state score as the first type when the quantized value is greater than or equal to a preset base value; determining the state score as the second type when the quantized value is less than the preset base value; and determining the proportion of the first type of state score to all state scores across multiple time periods. Specifically, the quantized values ​​corresponding to the state scores for all time periods are obtained based on the mean calculation formula, variance calculation formula, standard deviation calculation formula, or extreme value determination method; each quantized value is compared with the preset base value, i.e., the quantized values ​​are classified by the preset base value, and the state scores corresponding to quantized values ​​greater than or equal to the preset base value are determined as the first type, and the state scores corresponding to quantized values ​​less than the preset base value are determined as the second type; then, the proportion of the first type in the total data is calculated according to the proportion calculation formula. It is worth noting that the specific value of the preset base value needs to be determined based on the calculation method of the quantized value and in combination with the vehicle status in historical data. For example, if the quantized value is calculated using the mean calculation formula, and the preset base value is set to 0 in combination with historical data. When other calculation formulas are used for the quantification value, the preset base value will also be adjusted accordingly. The specific adjustment method needs to be determined based on the actual situation. This application will not list them in detail here.

[0121] S311: If the percentage value is greater than or equal to a first preset threshold, the vehicle is determined to be in a first state, which represents the normal state of the vehicle. Specifically, after obtaining the percentage value of the first type of state score, the first preset threshold is determined in conjunction with historical data. First, historical state scores are determined based on historical data from multiple vehicles; then, the type of historical state score is determined; next, the percentage value of the first type of historical state score is determined; finally, the value range of the percentage values ​​of all normal vehicles is determined to determine the size of the first preset threshold. Figure 5 As shown, the status scores of a vehicle over multiple days are determined, and the status scores are summarized to determine the type of the status score for each day. The percentage of the status scores over multiple days is then calculated. When the percentage falls within the defined range, the vehicle is considered to be in a normal state.

[0122] S312: If the percentage value is less than or equal to a second preset threshold, the vehicle is determined to be in a second state, which characterizes the vehicle's fault condition. Furthermore, if the vehicle is determined to be in the second state, an alarm is issued. Specifically, after obtaining the percentage value of the first type of state score, the second preset threshold is determined using the same method combined with the vehicle's historical data. Figure 5As shown, the system determines the vehicle's status score over multiple days, summarizes these scores, identifies the type of each day's status score, and calculates the percentage of the multi-day status score. This percentage is then compared to a second preset threshold to determine if the vehicle has malfunctioned. When a malfunction is detected, a warning is sent to the vehicle via wireless communication technology to remind the driver and passengers to perform timely vehicle maintenance and prevent accidents.

[0123] S313: If the percentage value is less than a first preset threshold but greater than a second preset threshold, the vehicle is determined to be in a third state. This third state characterizes the vehicle's questionable state. Specifically, the first preset threshold is greater than the second preset threshold. If the percentage value is less than the first preset threshold but greater than the second preset threshold, it indicates that the vehicle's current state cannot be determined. Figure 5 As shown, the vehicle's status score over multiple days is determined and summarized. The type of each day's status score is identified, and the percentage of the multi-day status score is calculated. This percentage is then compared to a first preset threshold and a second preset threshold to determine if the vehicle's status is questionable. There are many reasons why a vehicle might be considered to have a questionable status. It could be due to insufficient real-time vehicle operating data, resulting in an inaccurate percentage of vehicles in normal condition; environmental factors could lead to incomplete data for vehicles in normal condition; or the vehicle might be in the early stages of a fault, not yet exhibiting a faulty state. Therefore, based on these reasons, vehicles exhibiting this condition can be further observed.

[0124] like Figure 6 As shown, after determining the vehicle's status, it is necessary to mark the vehicle in the system. Based on the above results, it can be determined whether the vehicle's electric drive system is in a normal, faulty, or questionable state. When the vehicle's electric drive system is determined to be in a normal state, the vehicle is marked as normal and no further action is required; when the vehicle's electric drive system is determined to be in a faulty state, the vehicle is marked as faulty and a warning message is sent to the vehicle or the owner; when the vehicle's electric drive system is determined to be in a questionable state, the vehicle is marked as questionable and continued observation is required.

[0125] Based on the above solutions, such as Figure 7 As shown, this application provides a detailed flowchart of the solution, which specifically includes:

[0126] Data stream transmission: The system collects and transmits the operating parameters of the vehicle's electric drive system through onboard sensors.

[0127] Large databases: including storage devices such as servers, used to store the acquired working parameters;

[0128] Data cleaning: Preprocessing data to remove incomplete or invalid data;

[0129] Real-time motor temperature prediction: Input real-time operating data into the temperature prediction model to obtain the real-time predicted motor temperature;

[0130] Generate motor temperature: Generate the motor temperature corresponding to the operating parameters for each time period;

[0131] Motor temperature determination: Compare the motor temperature with a preset temperature threshold;

[0132] Determine if the temperature is greater than a preset temperature threshold. If the temperature is less than or equal to the preset temperature threshold, continue to determine the relationship between the motor temperature of the next vehicle and the preset temperature threshold. If the temperature is greater than the preset temperature threshold, retain the data for subsequent processing and analysis. Extract high temperature data: that is, retain the data where the motor temperature is greater than the preset temperature threshold.

[0133] Extract N time points: Before inputting the data into the model, extract the corresponding time points from the retained high-temperature data to ensure that it meets the input of the model. N is a positive integer greater than 0, which is the total amount of high-temperature data.

[0134] Isolation forest model scoring: A scoring model is built based on the isolation forest model. Processed high temperature data is input into the isolation forest model for scoring to obtain the state score of each vehicle.

[0135] State score statistics: Statistical analysis of the obtained state scores;

[0136] Store and record the results: Store the status scores and record the results;

[0137] Results are summarized daily: The stored state scores are categorized and summarized on a daily basis;

[0138] Statistical quantification indicators: quantify the state score and statistically analyze the resulting quantification indicators;

[0139] Calculate daily indicators: Calculate the quantitative indicators for each day;

[0140] Result determination: The quantitative indicators are classified on a daily basis to determine the number of normal states, the number of fault states, and the number of questionable states of the quantitative indicators.

[0141] Statistical analysis of each indicator: Count the number of quantitative indicators under each state. If it is a "normal state", count the normal number. If the number of normal states of the determined quantitative indicators is much greater than the number of fault states, then it is recorded as a normal state.

[0142] Statistical results: Quantitative indicators of the recorded normal state were statistically analyzed;

[0143] Output: If the vehicle's electric drive system is confirmed to be normal, its status is recorded as "normal". For vehicles in normal status, the database will not display additional specific information by default in order to save resources and focus attention on faulty vehicles that need attention, thereby ensuring that vehicles can continue to operate in normal status and avoiding unnecessary maintenance activities.

[0144] If it is a "fault state", that is, the number of faults is counted: if the number of fault states of the determined quantitative indicators is much greater than the number of normal states, then it is recorded as a fault state.

[0145] Statistical results: Quantitative indicators of the recorded fault states were statistically analyzed;

[0146] Output: If the electric drive system is identified as faulty, its status is marked as "faulty";

[0147] Report generation: Generate a vehicle fault report based on the fault status data;

[0148] Send reports via email: Send fault reports to customers via email recorded by the user, or inform relevant personnel which vehicles have faults and require immediate action through the user interface or message notifications;

[0149] Recommended inspection and maintenance: Remind users to inspect and maintain their vehicles in a timely manner to identify and resolve faults and prevent them from deteriorating further.

[0150] If a vehicle is in a "questionable" state, it is marked as such when the condition of the electric drive system cannot be clearly determined. A special list of vehicles suspected of malfunction is established to store information about these vehicles in a "questionable" state. Similarly, relevant personnel are alerted to the status of these vehicles via the user interface or push notifications, and suggestions are made to conduct a more detailed inspection, thereby avoiding misjudgments and ensuring that the true condition of the vehicle can be accurately verified and addressed.

[0151] Understandably, the design of the above process not only improves the accuracy of diagnostic results for the vehicle's electric drive system, but also optimizes resource allocation, ensures that critical issues are addressed first, and reduces the possibility of false alarms.

[0152] In summary, this application fully utilizes big data and IoT technologies to achieve real-time collection and preprocessing of vehicle operation data. This not only ensures the timeliness and accuracy of the data but also comprehensively captures the multi-dimensional operating status of the electric drive system. By introducing a temperature prediction model, potentially abnormal data segments can be pre-screened based on motor temperature, providing a more efficient and accurate foundation for subsequent fault diagnosis. By introducing the concept of comprehensive state score calculation, the multi-day state scores output by the isolated forest model are cleverly combined with statistical indicators. The magnitude of the state score effectively represents whether the electric drive cooling system is in normal mode; while different statistical indicators help to accurately identify abnormal data. The combination of these two aspects further improves the accuracy of fault diagnosis, making anomaly detection under complex operating conditions more reliable. Therefore, this application not only solves some key problems in existing technologies but also provides a solid guarantee for the safe operation of new energy vehicles, optimizes the fault detection process, ensures vehicle safety and reliability, and significantly improves the user's travel experience.

[0153] like Figure 8 As shown, this application embodiment provides a state detection device for an electric drive cooling system of a new energy vehicle, including: an acquisition module for acquiring real-time operating data of the vehicle at multiple time periods; an input module for inputting the real-time operating data into a pre-built state scoring model to obtain multiple state scores of the vehicle; and a calculation module for determining the state of the vehicle based on the distribution of the multiple state scores. Specifically, as shown... Figure 9 As shown, the acquisition module can acquire real-time operating parameters obtained by the vehicle-mounted sensors at regular intervals, and preprocess the real-time operating parameters to obtain time-series data, i.e., real-time operating data; the input module inputs the real-time operating data into the state scoring model constructed by the isolated forest for scoring, and the state scoring model outputs state scores; the average value of the vehicle on a single day is calculated based on the acquired state scores; the average values ​​of multiple days are summarized to obtain the state scores for multiple days; finally, the state of the vehicle is determined according to the distribution of the state scores.

[0154] Understandably, a state scoring model built using isolated forests is employed to calculate and statistically analyze state scores on time-series window data. This model, based on the isolated forest concept, effectively identifies and detects driving data segments from faulty vehicles. After training and validation on a large dataset, this model demonstrates good generalization ability and reduces the risk of overfitting. Furthermore, leveraging this model, the system can analyze uploaded vehicle operating data daily in real time and promptly notify users or maintenance teams to take action when anomalies are detected, preventing further deterioration of the fault. This not only improves the accuracy of fault detection but also ensures effective handling of complex operating conditions, providing a solid data foundation for the health monitoring of vehicle electric drive systems.

[0155] like Figure 10 As shown in the embodiments of this application, a status detection platform for a new energy vehicle electric drive cooling system is provided, including: a data acquisition module for acquiring real-time operating data of the vehicle; a wireless module for transmitting the real-time operating data to the cloud; a detection module for acquiring the real-time operating data and detecting the vehicle's status using the new energy vehicle electric drive cooling system status detection method provided in any of the above embodiments; and a transmission module for transmitting the vehicle's detection results to the vehicle. Specifically, the real-time operating parameters of the vehicle are acquired in real time by the data acquisition module installed on the vehicle, and the real-time operating parameters are preprocessed to obtain real-time operating data; the real-time operating data is transmitted to the cloud via the wireless module, which includes 4G or 5G networks, etc.; the cloud stores the acquired real-time operating data and sends it to a status scoring model for scoring to obtain a status score; after statistical analysis of the status score, the final diagnostic result of the vehicle is obtained; if the vehicle has a fault, a fault report is sent to the vehicle to remind the user to have it repaired in time.

[0156] Figure 11 The specific operation flow of the new energy vehicle electric drive cooling system status monitoring platform is shown, including: "Acquiring vehicle data": collecting real-time operating parameters of the electric drive system through onboard sensors; "Data acquisition and preprocessing": after acquiring the real-time operating parameters, preprocessing the real-time operating parameters to obtain real-time operating data; "Real-time prediction of motor temperature": inputting the acquired real-time operating data into the temperature prediction model to obtain the motor temperature; "Temperature > preset temperature threshold?": used to judge the predicted motor temperature, thereby filtering the real-time operating data; if the motor temperature is less than the preset temperature threshold, the current segment ends, that is, the real-time operating data of this part is removed; if the motor temperature is greater than the preset temperature threshold, "status scoring model scoring" is performed: Real-time working data is input into a state scoring model built from an isolated forest to obtain corresponding state scores; "Summarize Scoring Results": Score results from multiple time periods are summarized, and the state scores are uniformly quantified to obtain quantified indicators; "Output Module": Based on the quantified indicators, the state of the vehicle's electric drive system is judged, and different results are output; When the number of "normal" indicators is high, the judgment result is "high probability of normal", and the "output result: normal" is output; When the number of "fault" indicators is high, the judgment result is "high probability of fault", and the "output result: fault" is output; When the number of both indicators is "equal", the judgment result is "high probability of doubt", and the "output result: doubt" is output.

[0157] It is understood that the new energy vehicle electric drive cooling system status detection platform provided in this application embodiment significantly improves the real-time performance, accuracy and reliability of electric drive system fault detection by integrating real-time data acquisition and preprocessing, typical feature extraction, efficient anomaly detection algorithms and powerful fault diagnosis models.

[0158] Based on the above-described methods, apparatus, or platforms for detecting the state of electric drive cooling systems in new energy vehicles according to embodiments of this application, corresponding embodiments are provided, specifically including:

[0159] Suppose a new energy vehicle company conducts a comprehensive analysis of the operating data of all vehicles at the end of each month to assess the overall health status and guide maintenance work. The specific process includes the following steps:

[0160] 1. Data Acquisition and Preprocessing

[0161] Data extraction: The vehicle status detection system collects operational data of all vehicles from the cloud server over the past month and performs data cleaning and preprocessing steps.

[0162] Temperature prediction: Based on the vehicle's operating data, a temperature prediction model is used to predict the motor temperature, generating corresponding temperature prediction values ​​for each time window.

[0163] 2. Temperature assessment

[0164] Threshold determination: Check the motor temperature prediction data of each vehicle one by one, and filter out the time window when the motor temperature exceeds the preset temperature threshold.

[0165] 3. State score output and statistics

[0166] State score output: The state score is calculated and output using a pre-trained isolated forest model on the time window data. The daily state score is calculated, and the daily average score is recorded and the average score over multiple consecutive days is recorded.

[0167] 4. Result Determination

[0168] Status Assessment: Based on multi-day statistical results of status scores, determine whether the vehicle is in a faulty or normal condition. Statistical analysis of vehicle status changes across different categories categorized by day within the current month (e.g., from normal to questionable to faulty).

[0169] In the embodiments of this application, such as Figure 12 , Figure 13 , Figure 14 and Figure 15 The diagram illustrates the changes in the quantified status scores of different vehicles over a month. Figure 12The chart shown is a line graph illustrating the quantitative value changes of the first vehicle inspected over a month. The graph shows that the quantitative value of the first vehicle was positive for 25 days, negative for 2 days, and zero for 3 days (a value of 0 indicates the vehicle was not used that day). Therefore, it can be concluded that the first vehicle inspected was in a normal state. Figure 13 The chart shows the line graph of the quantitative value changes of the second vehicle under inspection over a month. As can be seen from the graph, the quantitative value of the second vehicle was positive for 14 days, negative for 15 days, and zero for 1 day. Therefore, it can be concluded that the second vehicle under inspection is in a questionable state. Figure 14 The chart shows the line graph of the quantitative value changes of the third inspection vehicle over a month. As can be seen from the graph, the quantitative value of the third inspection vehicle was positive for 0 days, negative for 24 days, and zero for 6 days. Therefore, it can be concluded that the third inspection vehicle was in a faulty state. Figure 15 The graph shows the change in the quantitative value of the third vehicle after it passed the maintenance in one month. As can be seen from the graph, the quantitative value of the third vehicle was positive for 23 days, negative for 0 days, and zero for 7 days. Therefore, it can be concluded that the third vehicle has recovered from the faulty state to the normal state after maintenance.

[0170] 5. Report Generation and Output

[0171] Report preparation: The system generates detailed monthly reports that summarize the status evolution of each vehicle. In particular, for vehicles that have experienced malfunctions, the report usually shows the trend of change from normal to questionable and then to malfunction, and is sent to the relevant departments via email.

[0172] Maintenance Recommendations: For vehicles identified as faulty, the system provides recommendations for further inspection and maintenance to ensure that potential problems are addressed promptly.

[0173] The above process not only helps to gain a comprehensive understanding of the fleet's overall health, but also allows for the early identification and response to potential problems, thereby optimizing maintenance plans and improving operational efficiency.

[0174] The above are merely preferred embodiments of this application and are not intended to limit this application in any way. Although this application has disclosed preferred embodiments as above, it is not intended to limit this application. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the technical solution of this application. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of this application without departing from the scope of the technical solution of this application shall still fall within the scope of the technical solution of this application.

Claims

1. A method for detecting the status of an electric drive cooling system in a new energy vehicle, characterized in that, include: Acquire real-time operational data of the vehicle across multiple time periods; The real-time working data is input into a pre-built state scoring model to obtain multiple state scores for the vehicle; The state of the vehicle is determined based on the distribution of the multiple state scores.

2. The method for detecting the state of a new energy vehicle electric drive cooling system according to claim 1, characterized in that, The method for determining the state of the vehicle based on the distribution of the plurality of state scores includes: Determine the proportion of the state scores of the first type among the plurality of state scores; If the percentage value is greater than or equal to a first preset threshold, the vehicle is determined to be in a first state; If the percentage value is less than or equal to a second preset threshold, the vehicle is determined to be in a second state. If the percentage value is less than the first preset threshold and greater than the second preset threshold, the vehicle is determined to be in a third state. Wherein, the first state is used to characterize the normal state of the vehicle, the second state is used to characterize the fault state of the vehicle, and the third state is used to characterize the questionable state of the vehicle.

3. The method for detecting the state of a new energy vehicle electric drive cooling system according to claim 2, characterized in that, The method for determining the proportion of the first type of state score among the plurality of state scores includes: Determine the quantized value of the state score for each time period; If the quantized value is greater than or equal to a preset base value, the state score is determined to be of the first type; If the quantized value is less than the preset base value, the state score is determined to be of the second type; Determine the percentage of the first type of state score among all state scores under the multiple time periods.

4. The method for detecting the status of a new energy vehicle electric drive cooling system according to claim 3, characterized in that, The quantified values ​​include any one of the mean, variance, standard deviation, and extreme values.

5. The method for detecting the state of a new energy vehicle electric drive cooling system according to claim 2, characterized in that, The method further includes: If the vehicle is determined to be in the second state, an alarm is issued.

6. The method for detecting the state of a new energy vehicle electric drive cooling system according to claim 1, characterized in that, The method for constructing the state scoring model includes: An initial model is built based on a basic network architecture, which includes any one of the following: isolated forest, radial basis function kernel, autoencoder and its variants, deep embedding clustering, and GMM neural network. The acquired historical working data of the vehicle is input into the initial model for training; The initial model that has been trained and validated is determined as the state scoring model.

7. The method for detecting the state of a new energy vehicle electric drive cooling system according to claim 1, characterized in that, The method for acquiring real-time operating data of a vehicle across multiple time periods includes: Obtain the operating parameters of the vehicle during its operation; The operating parameters are preprocessed; The motor temperature during vehicle operation is determined based on the preprocessed operating parameters. If the motor temperature is greater than or equal to a preset temperature threshold, the operating parameters corresponding to the motor temperature are retained. The retained motor temperature and corresponding operating parameters are used as the real-time operating data.

8. The method for detecting the state of an electric drive cooling system for new energy vehicles according to claim 7, characterized in that, The method for determining the motor temperature during vehicle operation based on the preprocessed operating parameters includes: The operating parameters are input into a pre-trained temperature prediction model to obtain the motor temperature; The temperature prediction model is trained based on the vehicle's historical data.

9. A status detection device for an electric drive cooling system of a new energy vehicle, characterized in that, include: The acquisition module is used to acquire real-time working data of the vehicle under multiple time periods; An input module is used to input the real-time working data into a pre-built state scoring model to obtain multiple state scores for the vehicle. The calculation module is used to determine the state of the vehicle based on the distribution of the plurality of state scores.

10. A condition monitoring platform for an electric drive cooling system of a new energy vehicle, characterized in that, include: A data acquisition module is used to collect real-time operating data of the vehicle. A wireless module, which is used to send the real-time working data to the cloud; The detection module is used to detect the vehicle's status using the new energy vehicle electric drive cooling system status detection method as described in any one of claims 1-8; A sending module is used to send the detection results of the vehicle to the vehicle.

Images