Defrosting methods for vehicle heat exchangers and training methods for frosting detection models

Abstract

This application discloses a defrosting method for vehicle heat exchangers and a training method for a frosting detection model, belonging to the field of vehicle technology. The defrosting method for vehicle heat exchangers includes: acquiring frosting feature data related to frosting scenarios on the surface of the vehicle's heat exchanger; processing the frosting feature data using a frosting detection model to obtain the degree of frosting on the heat exchanger; the frosting detection model is a pre-trained machine learning model; determining whether defrosting conditions are met based on the degree of frosting and the vehicle's location information, and defrosting the heat exchanger if the defrosting conditions are met. This application can accurately detect the degree of frosting on the surface of the heat exchanger.

Description

Technical Field

[0001] This application belongs to the field of vehicle technology, and in particular relates to a defrosting method for vehicle heat exchangers and a training method for a frosting detection model. Background Technology

[0002] Vehicles face several challenges when driving in winter. While range is a major concern, if the ambient temperature is close to zero degrees Celsius and there is a certain level of humidity, frost will form on the surface of the front heat exchanger, which will hinder airflow, increase thermal resistance, and consequently increase compressor energy consumption and reduce the overall vehicle range.

[0003] Existing methods for detecting the degree of frost require vehicles to be equipped with dedicated cameras or sensors for frost detection, which are not applicable to vehicles lacking such equipment. Alternatively, methods can be based on the parameters of components in the air conditioning system, but this method has lower accuracy.

[0004] Therefore, how to accurately defrost the heat exchanger of a vehicle without increasing vehicle production costs is a problem that urgently needs to be solved. Summary of the Invention

[0005] This application aims to address at least one of the technical problems existing in the related art. To this end, this application proposes a defrosting method for vehicle heat exchangers and a training method for a frosting detection model to accurately predict the degree of frosting on the heat exchanger surface, thereby enabling accurate defrosting.

[0006] In a first aspect, this application provides a defrosting method for a vehicle heat exchanger, the method comprising:

[0007] Acquire frosting feature data related to frosting scenarios on vehicle heat exchanger surfaces;

[0008] The frost detection model is called to process the frost feature data to obtain the degree of frost on the heat exchanger; the frost detection model is a pre-trained machine learning model.

[0009] Based on the degree of frost and the vehicle's location information, it is determined whether the defrosting conditions are met, and if the defrosting conditions are met, the heat exchanger is defrosted.

[0010] The defrosting method for vehicle heat exchangers according to this application acquires frosting feature data related to the frosting scenario on the vehicle's heat exchanger surface, and processes this data using a frosting detection model. This allows for real-time prediction of the frosting degree of the heat exchanger. Through multi-dimensional data analysis, the method accurately determines the frosting degree, significantly improving the accuracy and real-time performance of frosting detection. Furthermore, based on the frosting degree and the collected vehicle location information, the method intelligently determines whether immediate defrosting is necessary, optimizing the defrosting strategy logic and avoiding excessive energy consumption caused by frequent defrosting initiation, thus improving the vehicle's energy-saving performance. Moreover, by using a pre-trained frosting detection model to detect the frosting degree, there is no need to install additional cameras or sensors specifically for detecting frosting on the vehicle. This method is applicable to various vehicle configurations and driving environments, demonstrating good adaptability.

[0011] According to one embodiment of this application, determining whether the defrosting conditions are met based on the degree of frost and the vehicle's location information includes: determining whether the degree of frost exceeds an upper limit threshold; determining the remaining travel time for the vehicle to move from its current location to its destination; and determining that the degree of frost and the vehicle's location information meet the defrosting conditions if the degree of frost exceeds the upper limit threshold and the remaining travel time exceeds a time threshold.

[0012] According to one embodiment of this application, the step of determining whether the defrosting conditions are met based on the degree of frost and the vehicle's location information further includes: determining whether there is a heat source in the vehicle that is in a usable state; and determining that the degree of frost and the vehicle's location information meet the defrosting conditions if there is at least one heat source in a usable state, the degree of frost exceeds an upper limit threshold, and the remaining driving time exceeds a duration threshold.

[0013] According to one embodiment of this application, defrosting the heat exchanger when the defrosting conditions are met includes: performing a preset defrosting operation when the defrosting conditions are met, so that the degree of frost is lower than a lower threshold; wherein the preset defrosting operation includes at least one of heating defrosting and compression defrosting.

[0014] Secondly, this application provides a training method for a frosting detection model, the method comprising:

[0015] Multiple sets of frosting feature samples related to the frosting scenario on the heat exchanger surface of a vehicle are obtained; each set of frosting feature samples is labeled with a sample tag, and the sample tag indicates the actual degree of frosting corresponding to the frosting feature sample;

[0016] The frost detection model to be trained is used to process each group of frost feature samples and output the predicted degree of frost corresponding to each group of frost feature samples.

[0017] Determine the difference between the predicted degree of frost and the actual degree of frost for each group of frost feature samples;

[0018] The frost detection model is trained with the goal of minimizing the difference until the difference is less than the error threshold. The training ends and a trained frost detection model is obtained. The trained frost detection model is used to predict the degree of frost on the heat exchanger.

[0019] According to the training method of the frost detection model in this application, multiple sets of frost feature samples related to the frost scene on the surface of the vehicle's heat exchanger are acquired. The frost detection model to be trained processes each set of frost feature samples, outputting the predicted frost degree corresponding to each set of frost feature samples. This predicted frost degree is then compared with the actual frost degree represented by the sample label. The difference obtained from the comparison represents the accuracy of the model's prediction of the frost degree. Furthermore, the frost detection model is trained with the goal of minimizing the difference until the difference is less than the error threshold. The training ends, and a well-trained frost detection model is obtained, capable of accurately judging the frost degree of the vehicle's heat exchanger. Through the above supervised learning method, the model can learn the correlation between various features and the actual frost degree. The trained frost detection model has high prediction accuracy, strong generalization ability, and robustness, making it applicable to various driving environments.

[0020] According to one embodiment of this application, the method further includes: acquiring a first sample image and a second sample image corresponding to each group of frost feature samples, wherein the first sample image is used to represent the surface frost area of ​​the heat exchanger and the second sample image is used to represent the side frost thickness of the heat exchanger; and determining a sample label corresponding to each group of frost feature samples based on the surface frost area and the side frost thickness.

[0021] Thirdly, this application provides a defrosting device for a vehicle heat exchanger, the device comprising:

[0022] The data acquisition module is used to acquire frosting feature data related to the frosting scenario on the heat exchanger surface of the vehicle;

[0023] The prediction module is used to call the frost detection model to process the frost feature data and obtain the degree of frost on the heat exchanger; the frost detection model is a pre-trained machine learning model.

[0024] An execution module is used to determine whether defrosting conditions are met based on the degree of frost and the vehicle's location information, and to defrost the heat exchanger if the defrosting conditions are met.

[0025] The defrosting device for vehicle heat exchangers according to this application acquires frost feature data related to frost formation on the heat exchanger surface and processes this data using a frost detection model. This allows for real-time prediction of the frost level on the heat exchanger. Through multi-dimensional data analysis, the device accurately determines the frost level, significantly improving the accuracy and real-time performance of frost detection. Furthermore, based on the frost level and the collected vehicle location information, it intelligently determines whether immediate defrosting is necessary, optimizing the defrosting strategy logic and avoiding excessive energy consumption caused by frequent defrosting activation, thus improving the vehicle's energy efficiency. Moreover, by using a pre-trained frost detection model to detect the frost level, there is no need to install additional cameras or sensors specifically for frost detection on the vehicle. This makes it suitable for various vehicle configurations and driving environments, demonstrating good adaptability.

[0026] Fourthly, this application provides a training device for a frosting detection model, the device comprising:

[0027] The acquisition module is used to acquire multiple sets of frost feature samples related to the frost scene on the heat exchanger surface of the vehicle; wherein, each set of frost feature samples is labeled with a sample label, and the sample label indicates the actual degree of frost corresponding to the frost feature sample;

[0028] The training module is used to process each group of frost feature samples using the frost detection model to be trained, and output the predicted degree of frost corresponding to each group of frost feature samples.

[0029] The training module is also used to determine the difference between the predicted degree of frost and the actual degree of frost corresponding to each group of frost feature samples;

[0030] The training module is further configured to train the frosting detection model with the goal of minimizing the difference, until the difference is less than the error threshold, at which point the training ends and a trained frosting detection model is obtained; the trained frosting detection model is used to predict the degree of frosting on the heat exchanger.

[0031] According to the training device of the frosting detection model of this application, multiple sets of frosting feature samples related to the frosting scenario on the surface of a vehicle's heat exchanger are acquired. The frosting detection model to be trained processes each set of frosting feature samples, outputting the predicted frosting degree corresponding to each set of frosting feature samples. This predicted degree is then compared with the actual frosting degree represented by the sample label. The difference obtained from the comparison represents the accuracy of the model's prediction of the frosting degree. Furthermore, the frosting detection model is trained with the goal of minimizing the difference until the difference is less than the error threshold. The training ends, and a well-trained frosting detection model is obtained, capable of accurately judging the frosting degree of the vehicle's heat exchanger. Through the above-mentioned supervised learning method, the model can learn the correlation between various features and the actual frosting degree. The trained frosting detection model has high prediction accuracy, strong generalization ability, and robustness, making it applicable to various driving environments.

[0032] Fifthly, this application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the defrosting method for a vehicle heat exchanger as described in the first aspect above, or the training method for a frosting detection model as described in the second aspect.

[0033] Sixthly, this application provides a vehicle including the computer equipment described in the fifth aspect.

[0034] In a seventh aspect, this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the defrosting method for a vehicle heat exchanger as described in the first aspect above, or the training method for a frosting detection model as described in the second aspect.

[0035] Eighthly, this application provides a chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the defrosting method for a vehicle heat exchanger as described in the first aspect, or the training method for a frosting detection model as described in the second aspect.

[0036] Ninthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the defrosting method for a vehicle heat exchanger as described in the first aspect above, or the training method for a frosting detection model as described in the second aspect.

[0037] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0038] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0039] Figure 1 This is a schematic diagram illustrating an application scenario of the defrosting method for a vehicle heat exchanger provided in some embodiments of this application;

[0040] Figure 2 This is a schematic flowchart of a defrosting method for a vehicle heat exchanger provided in some embodiments of this application;

[0041] Figure 3 This is a circuit diagram of a direct heat pump system provided in some embodiments of this application;

[0042] Figure 4 This is a schematic diagram of the circuit structure of a secondary heat exchange heat pump system provided in some embodiments of this application;

[0043] Figure 5 This is a schematic diagram of the structure of the decision tree model provided in some embodiments of this application;

[0044] Figure 6 This is a flowchart illustrating the training method of the frost detection model provided in some embodiments of this application;

[0045] Figure 7 This is a schematic diagram illustrating the training process of the frost detection model provided in some embodiments of this application;

[0046] Figure 8 This is a schematic diagram of the overall process in a real-world application scenario provided in some embodiments of this application;

[0047] Figure 9 This is a flowchart illustrating the defrosting condition judgment logic provided in some embodiments of this application;

[0048] Figure 10 This is a schematic diagram of the defrosting device for a vehicle heat exchanger provided in some embodiments of this application;

[0049] Figure 11 This is a schematic diagram of the structure of the training device for the frost detection model provided in some embodiments of this application;

[0050] Figure 12 This is a schematic diagram of the structure of a computer device provided in some embodiments of this application. Detailed Implementation

[0051] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0052] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used in the description of this application is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms "comprising" and "having," and any variations thereof, in the description, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the description, claims, or accompanying drawings of this application are used to distinguish different objects, not to describe a specific order or hierarchy.

[0053] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.

[0054] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," "linking," and "attachment" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0055] In this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, in this application, the character " / " generally indicates that the preceding and following related objects have an "or" relationship.

[0056] In this application, "multiple" refers to two or more (including two), and similarly, "multiple groups" refers to two or more (including two), and "multiple pieces" refers to two or more (including two).

[0057] Traditional methods of vehicle frost detection mostly focus on checking for frost on the windshield, neglecting the more critical issue of defrosting the heat exchanger, which significantly impacts vehicle performance. Furthermore, traditional frost detection requires additional onboard cameras or sensors for assessment. For example, detecting frost on the windshield necessitates using a camera or sensor near the window to measure the thickness of the frost layer, thus determining the degree of frost buildup. This increases production costs and makes the system unsuitable for vehicles without such devices. Conversely, relying solely on parameters of components within the air conditioning system results in highly inaccurate readings. This is particularly problematic for new energy vehicles with limited battery capacity. Inaccurate frost detection leads to either neglecting defrosting even when heat exchanger frost is severe, or frequently performing defrosting when frost is mild, resulting in significant waste of remaining battery power.

[0058] In view of this, embodiments of this application provide a defrosting method for a vehicle heat exchanger and a training method for a frost detection model, thereby achieving accurate prediction of the degree of frost on the vehicle heat exchanger through a machine learning model without increasing vehicle development and production costs, and determining a specific defrosting strategy based on the degree of frost, thereby further reducing the energy consumption of the entire vehicle.

[0059] It should be noted that the vehicles mentioned in the embodiments of this application include, but are not limited to, gasoline vehicles, plug-in hybrid electric vehicles or new energy vehicles, etc., and this application does not make specific limitations in this regard.

[0060] The defrosting method for vehicle heat exchangers and the training method for the frost detection model provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios.

[0061] The defrosting method for vehicle heat exchangers provided in this application embodiment can be applied to, for example... Figure 1 The application environment shown is such that vehicle 110 is communicatively connected to computer equipment 120, and a heat exchanger is installed at the front end of vehicle 110.

[0062] Computer equipment can be installed inside or outside the vehicle. It can be a terminal device or a server. For example, terminal devices include, but are not limited to, one or more of various desktop computers, laptops, smartphones, tablets, in-vehicle terminals, IoT devices, or portable wearable devices. IoT devices can include one or more of smart speakers, smart TVs, smart air conditioners, or smart in-vehicle devices. Portable wearable devices can include one or more of smartwatches, smart bracelets, or head-mounted devices. Servers can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, and big data and artificial intelligence platforms.

[0063] The defrosting method for vehicle heat exchangers can be applied to computer equipment, specifically executed by the hardware or software within the computer equipment.

[0064] The defrosting method for vehicle heat exchangers provided in this application embodiment can be executed by a computer device or a functional module or entity within a computer device capable of implementing the defrosting method for the vehicle heat exchanger. The following description uses a computer device as the executing entity to illustrate the defrosting method for vehicle heat exchangers provided in this application embodiment.

[0065] like Figure 2 As shown, the defrosting method for the vehicle heat exchanger includes steps 210, 220, and 230.

[0066] Step 210: Obtain frost feature data related to the frost scene on the heat exchanger surface of the vehicle.

[0067] Vehicles face challenges in winter, including range anxiety and frost buildup on heat exchangers. When air temperature is low and humidity is high, or when the vehicle's air conditioning is in cooling mode, moisture in the air condenses and frosts on the heat exchanger surface. If this frost isn't defrosted promptly, the frost layer reduces heat exchange efficiency, impacting vehicle performance and increasing energy consumption.

[0068] Therefore, the frosting characteristic data related to the frosting scenario on the surface of a vehicle's heat exchanger usually refers to the data or characteristics of components closely related to the function, structure, or operating status of the heat exchanger, such as the temperature of the heat exchanger itself and the ambient humidity. These data or characteristics are related to whether or not frost will form on the surface of the heat exchanger.

[0069] Specifically, equipment related to heat exchangers includes, but is not limited to: equipment that directly works with the heat exchanger, i.e., equipment that directly participates in or assists the heat exchanger in completing the heat exchange process, such as one or more of condensers, evaporators, radiators, compressors, and fans; control and monitoring equipment, i.e., sensors and control components used to detect and regulate the state of the heat exchanger, such as one or more of temperature sensors, humidity sensors, coolant pumps, heat pumps, and expansion valves; and equipment that affects the operating conditions of the heat exchanger, i.e., various environmental and operating condition data that affect the operation of the heat exchanger, including one or more of ambient temperature, ambient humidity, airflow, coolant state, vehicle speed, and intake air volume.

[0070] For example, the frosting characteristic data includes, but is not limited to, one or more of the following data types: compressor inlet pressure, compressor inlet temperature, external heat exchanger inlet refrigerant pressure, external heat exchanger inlet refrigerant temperature, cooling fan duty cycle, ambient heat pump running time, system mode, ambient temperature, ambient humidity, vehicle speed, variable intake grille opening, and external heat exchanger inlet coolant temperature.

[0071] The computer device acquires frosting characteristic data of the vehicle related to a frosting scenario on the heat exchanger surface, and can directly acquire frosting characteristic data transmitted by various devices or sensors. In some embodiments, or when the computer device acts as a server, it can acquire frosting characteristic data sent by the in-vehicle terminal.

[0072] Step 220: Call the frost detection model to process the frost feature data and obtain the degree of frost on the heat exchanger; the frost detection model is a pre-trained machine learning model.

[0073] The computer equipment analyzes the collected frosting feature data by calling a trained machine learning model to output the current degree of frosting on the heat exchanger.

[0074] The degree of frost can be expressed as one or more of the following: frost volume, frost area, frost thickness, frost severity level, or frost severity score. For example, the frost severity level can be preset as: no frost, light frost, heavy frost, etc. The frost severity score can characterize the degree of frost by the magnitude of the score; the higher the score, the more severe the frost, and so on.

[0075] The frost detection model, trained on extensive test data under various operating conditions, can accurately identify and determine the frost condition of heat exchangers under different conditions. The frost detection model is a supervised machine learning model; for example, it can be, but is not limited to, one of the following: XGBoost (Extreme Gradient Boosting), Random Forest, Support Vector Machine, or Naive Bayes.

[0076] Step 230: Determine whether the defrosting conditions are met based on the degree of frost and the vehicle's location information, and defrost the heat exchanger if the defrosting conditions are met.

[0077] In addition to frost characteristic data, the computer equipment also acquires the vehicle's location information for use in determining the defrosting strategy. For example, the computer equipment can obtain the vehicle's location at various times through the vehicle's GPS module. The computer equipment combines the degree of frost and the vehicle's location information to determine whether defrosting conditions are met. Defrosting conditions specify that defrosting should be performed under certain conditions to meet requirements such as vehicle performance or driving comfort.

[0078] For example, when the computer device determines that the degree of frost has reached a state of slight or severe frost, and determines that the vehicle is located in a low-temperature environment area based on the location information, the computer device determines that the defrosting conditions are met and starts the defrosting mode to defrost the heat exchanger.

[0079] For example, when the computer system determines that the frost layer volume is greater than a threshold and the vehicle is located in a high-humidity environment based on the location information, the computer system judges that the defrosting conditions are met and starts the defrosting mode to defrost the heat exchanger.

[0080] The defrosting method for vehicle heat exchangers disclosed in this application acquires frosting feature data related to frosting scenarios on the heat exchanger surface and processes this data using a frosting detection model. This allows for real-time prediction of the frosting degree. Through multi-dimensional data analysis, the method accurately determines the frosting degree, significantly improving the accuracy and real-time performance of frosting detection. Furthermore, based on the frosting degree and the collected vehicle location information, the method intelligently determines whether immediate defrosting is necessary, optimizing the defrosting strategy logic and avoiding excessive energy consumption caused by frequent defrosting initiation, thus improving the vehicle's energy efficiency. Moreover, by using a pre-trained frosting detection model to detect the frosting degree, there is no need to install additional cameras or sensors specifically for frosting detection on the vehicle. This method is applicable to various vehicle configurations and driving environments, demonstrating good adaptability.

[0081] In some scenarios, whether a vehicle performs a defrosting operation is also related to its remaining battery power. For example, for new energy vehicles, if multiple defrosting operations are performed at a distance from the destination, the battery power will be consumed more quickly than expected (i.e., only considering the power consumption of the travel route), which may result in insufficient battery power to support the vehicle to reach its destination. In view of this, in some embodiments, the determination of whether the defrosting conditions are met is based on the degree of frost and the vehicle's location information, including: determining whether the degree of frost exceeds an upper limit threshold; determining the remaining travel time for the vehicle to move from its current location to its destination; and determining that the degree of frost and location information meet the defrosting conditions if the degree of frost exceeds the upper limit threshold and the remaining travel time exceeds a time threshold.

[0082] The destination location could be, for example, the end point of the current trip, or the location of a charging station / charging pile.

[0083] The computer device obtains a pre-set upper limit threshold, which indicates the degree of frost buildup on the heat exchanger and the need for defrosting. For example, assuming the model outputs the degree of frost buildup as represented by frost volume, when the frost volume exceeds a preset volume, the computer device considers that there is a large amount of frost accumulation on the heat exchanger surface and that defrosting is required.

[0084] Furthermore, the computer device also obtains the vehicle's current location and destination location. Based on these locations, the computer device can determine the remaining travel time from the current location to the destination. For example, the computer device can obtain this remaining travel time by calling navigation services or map services. Correspondingly, a threshold is set for the remaining travel time. For instance, if the remaining travel time is 20 minutes, exceeding a preset threshold (e.g., 15 minutes), the computer device determines that the vehicle is far from its destination and requires a longer driving time. During this process, the frost layer on the heat exchanger surface may further worsen. Therefore, it determines that the degree and location of the frost meet the defrosting conditions and a defrosting operation is necessary.

[0085] Conversely, if either of the above two conditions is not met, the computer equipment determines that the defrosting conditions are not met. For example, if the degree of frost does not exceed the upper limit threshold, it means that the frost layer on the heat exchanger surface does not need to be defrosted, and the computer equipment determines that the defrosting conditions are not met. As another example, if the degree of frost exceeds the upper limit threshold, but the remaining driving time is 5 minutes, which is less than the preset time threshold (e.g., 15 minutes), it means the vehicle is about to reach its destination. In this case, the defrosting operation can be performed after the vehicle arrives at its destination, either after parking or while charging, thereby avoiding unnecessary defrosting operations and further reducing overall vehicle energy consumption.

[0086] In the above embodiments, by combining the degree of frost and the remaining driving time from the current location to the destination location, defrosting is only triggered when the degree of frost exceeds the upper limit threshold and the driving time is relatively long, thus avoiding the energy waste of frequently starting the defrosting device. Furthermore, by automatically detecting the degree of frost and the remaining driving time, defrosting can be automatically performed when the defrosting conditions are met, without requiring manual operation by the driver. This reduces operational complexity and improves the intelligence and convenience of vehicle use.

[0087] In some embodiments, determining the remaining travel time for a vehicle to move from its current location to its destination includes: calling a navigation service to obtain the remaining travel time for the vehicle to move from its current location to its destination; or, obtaining traffic information and determining the remaining travel time for the vehicle to move from its current location to its destination based on the traffic information.

[0088] Computer devices can obtain the remaining travel time from the current location to the destination location by calculating it themselves or by calling other existing services, such as navigation services and map services.

[0089] For example, once a computer determines that a vehicle meets the defrosting requirements, it obtains the remaining travel time from the current location to the destination via a navigation service interface. The navigation service provides an accurate remaining travel time based on factors such as the distance between the current and destination locations, the route, and the current speed. Alternatively, the computer can calculate the remaining time itself. For instance, if the computer is an in-vehicle terminal integrated with local map services, it can obtain traffic information and estimate the remaining travel time from the current location to the destination based on that information. Traffic information includes, but is not limited to, one or more of the following: congestion conditions along the current route, speed limits, and traffic light information.

[0090] In the above embodiments, by flexibly selecting the method of obtaining the remaining driving time through navigation services or traffic information, the driving time estimation is more accurate, and thus the determination of whether the defrosting conditions are met is more accurate.

[0091] In gasoline-powered vehicles, defrosting is typically achieved using engine heat. In new energy vehicles, since the electric motor replaces the engine, defrosting can be performed using heat generated by the motor or by the vehicle's positive temperature coefficient (PTC) or compressor. Of course, other defrosting methods are also possible, and this application does not limit this approach. Based on the above embodiments, determining whether defrosting conditions are met based on the degree of frost and the vehicle's location information further includes: determining whether there is a usable heat source in the vehicle; and determining that the degree of frost and location information meet the defrosting conditions if at least one usable heat source exists, the degree of frost exceeds an upper limit threshold, and the remaining driving time exceeds a duration threshold.

[0092] In other words, in addition to determining whether the degree of frost exceeds the upper limit threshold and whether the remaining driving time exceeds the time threshold, the computer equipment also needs to determine whether there is a usable heat source in the vehicle. The heat source can be one or more of the following: engine waste heat, engine exhaust heat, motor waste heat, and electric heating devices. If at least one heat source is usable, it indicates that the hardware conditions for defrosting are met. The computer equipment then determines that the defrosting conditions are met and initiates the defrosting mode to defrost the heat exchanger.

[0093] In the above embodiments, by comprehensively considering the degree of frost, driving time, and availability of heat sources, the system intelligently determines when defrosting is needed. This allows for timely activation of defrosting operations when frost accumulation is severe, the journey is long, and no additional heat source is required, thereby improving the intelligence level and driving safety of the vehicle. Furthermore, by only activating defrosting operations when an available heat source is detected, the computer equipment avoids the energy waste caused by frequent heating, effectively saving electricity or fuel resources and making the vehicle more energy-efficient and environmentally friendly overall.

[0094] In any of the above embodiments, defrosting the heat exchanger when the defrosting conditions are met includes: performing a preset defrosting operation when the defrosting conditions are met, so that the degree of frost is lower than a lower threshold; wherein the preset defrosting operation includes at least one of heating defrosting and compression defrosting.

[0095] When the computer equipment determines that the defrosting conditions are met based on various conditions, namely, the degree of frost exceeds the upper limit threshold and the remaining driving time is greater than the set duration threshold; or, the degree of frost exceeds the upper limit threshold, the remaining driving time is greater than the set duration threshold, and there is a usable heat source, then the preset defrosting operation is executed.

[0096] Preset defrosting operations can include preset defrosting methods, defrosting durations, etc. The preset defrosting method indicates the heat source used for defrosting, and at least includes one of heating defrosting and compression defrosting. Heating defrosting includes, for example, engine heating or PTC heating for defrosting, while compression defrosting includes, for example, running a compressor for defrosting. Preset defrosting operations can also specify the degree to which frost is removed, such as reducing the frost level below a lower threshold.

[0097] For example, the computer equipment controls an electric heating device or waste heat from the engine to heat the heat exchanger, raising its temperature and accelerating the melting of frost until the degree of frost falls below a set lower threshold. Alternatively, the computer equipment adjusts the compressor operating mode of the air conditioning system to reverse the refrigerant flow, raising the surface temperature of the heat exchanger and rapidly melting the frost. When the frost level drops below the lower threshold, the computer equipment controls the compressor to return to normal operating mode.

[0098] During this process, the computer equipment can continuously monitor the degree of frost formation. Once the set lower threshold is reached (such as below 10% of the frost volume), the compression defrosting operation is stopped, thereby reducing vehicle energy consumption.

[0099] In the above embodiments, when the defrosting conditions are met, the computer equipment can quickly reduce the accumulation of frost by performing heating or compression defrosting operations, ensuring the normal operation of the heat exchanger; and the defrosting process is automatically terminated when the frost layer is below the lower threshold, avoiding unnecessary heating or compression operations, effectively saving electricity or fuel resources, and achieving energy conservation and environmental protection.

[0100] Furthermore, vehicles can have different types of heat pump systems. For example, a vehicle's heat pump system can be a direct heat exchange type or a secondary heat exchange heat pump type. Different heat pump system types will affect the efficiency, energy consumption, and selection of defrosting methods for the heat exchanger. For instance, when a vehicle uses a direct heat pump system, its circuit structure related to the heat exchanger can be as follows: Figure 3 As shown, it includes an external heat exchanger, fan, compressor, condenser, and electronic expansion valve. When the vehicle is a secondary heat exchange heat pump system, its circuit structure related to the heat exchanger can be as follows: Figure 4 As shown, it includes an external heat exchanger, a fan, a battery cooler, a compressor, a condenser, and an electronic expansion valve.

[0101] In a direct heat pump system, the refrigerant circulates directly through the heat exchanger, and the heat is directly applied to the heat exchanger surface. Therefore, the heat exchanger can quickly absorb heat to achieve defrosting. Due to the shorter heat transfer path and less heat loss, direct heat pump systems have high defrosting efficiency. In contrast, secondary heat pump systems require heat transfer to the heat exchanger surface via a heat exchange fluid (such as refrigerant or water). Because heat transfer involves two steps (from the heat pump to the heat exchange fluid, and then from the heat exchange fluid to the heat exchanger), there is some heat loss, resulting in a relatively longer defrosting time. For secondary heat exchange systems, defrosting can be achieved using methods such as compression heating and circulating heating.

[0102] Therefore, in some embodiments, before calling the frost detection model to process the frost feature data, the above method further includes: determining the type of heat pump system in the vehicle; the heat pump system type includes a direct heat pump system and a secondary heat exchange heat pump system; if the vehicle is a direct heat pump system, the frost detection model is determined to be the frost detection model corresponding to the direct heat pump system; if the vehicle is a secondary heat exchange heat pump system, the frost detection model is determined to be the frost detection model corresponding to the secondary heat exchange heat pump system; wherein, the model structure of the frost detection models corresponding to different heat pump system types is different, and the data types of the frost feature data required for input are different for each type.

[0103] The computer equipment identifies the type of heat pump system in the vehicle based on the vehicle's configuration information, such as whether the current vehicle is equipped with a direct heat pump system or a secondary heat exchange heat pump system.

[0104] When the vehicle's heat pump system is confirmed to be a direct heat pump system, the computer automatically loads the corresponding frosting detection model. This frosting detection model is trained using training data that matches the direct heat pump system, making it more suitable for it. Frosting characteristic data corresponding to the direct heat pump system includes, but is not limited to: compressor inlet pressure, compressor inlet temperature, external heat exchanger inlet refrigerant pressure, external heat exchanger inlet refrigerant temperature, cooling fan duty cycle, ambient heat pump operating time, system mode, ambient temperature, ambient humidity, vehicle speed, and variable grille opening.

[0105] When the vehicle's heat pump system is confirmed to be a secondary heat exchange heat pump system, the computer automatically loads the frosting detection model corresponding to this system. This frosting detection model is trained using training data specific to secondary heat exchange heat pump systems, making it more suitable for them. Frosting characteristic data corresponding to secondary heat exchange heat pump systems includes, but is not limited to, the inlet coolant temperature of the external heat exchanger.

[0106] In the above embodiments, by automatically identifying direct heat pump systems and secondary heat exchange heat pump systems, selecting the corresponding frost detection model according to the vehicle heat pump system type, and loading the adapted frost detection model, the adaptability of the frost detection model to different vehicle configurations is realized, meeting the frost detection requirements of different heat pump systems. Furthermore, by setting different frost detection models specifically for different heat pump system types, the frost detection results are made more accurate.

[0107] In some scenarios, such as when a vehicle arrives at its destination and enters a charging state, energy consumption can be disregarded, and defrosting can be performed solely to completely remove the frost layer. Therefore, in some embodiments, the method further includes: when the vehicle is charging, determining a preset defrosting operation duration based on the degree of frost and / or a preset charging time; and performing the preset defrosting operation according to the operation duration.

[0108] Specifically, when the computer equipment detects that the vehicle is charging, it determines the preset defrosting operation duration based on the determined degree of frost. For example, if the frost volume output by the frost detection model is x, and the preset operation duration corresponding to this frost volume x is 10 minutes, then the computer equipment will perform a preset defrosting operation for 10 minutes.

[0109] For example, the computer equipment obtains the vehicle's preset charging time from the charging system. For instance, if the preset charging time is 1 hour, the computer equipment allocates a certain amount of time for defrosting based on that charging time; for example, if the operation time is 20 minutes, the computer equipment performs a 20-minute preset defrosting operation. Of course, both the degree of frost and the preset charging time can be used as the basis for determining the operation time simultaneously. For example, it could be the longer of the separately determined operation times, or a weighted average of the two, etc.

[0110] The above embodiments improve time utilization by completing the defrosting operation while the vehicle is charging, ensuring that the vehicle is in good working condition of the heat exchanger after charging is completed, thus enhancing the user experience. Furthermore, by dynamically adjusting the defrosting time based on the degree of frost and the charging time, unnecessary energy consumption during the defrosting process is avoided, achieving intelligent energy-saving defrosting.

[0111] For example, the frost detection model may include multiple decision tree models. Each decision tree model is a decision tree. The frost detection model is invoked to process the frost feature data to obtain the degree of frost on the heat exchanger, including: processing the frost feature data separately using multiple decision tree models to obtain model prediction values ​​corresponding to each decision tree model; and determining the degree of frost on the heat exchanger based on the model prediction values ​​corresponding to all decision tree models.

[0112] The computer system inputs the frosting feature data into each decision tree model, resulting in multiple independent model predictions. For example, given the input frosting feature data, different decision tree models might output predicted frosting levels of 0.5, 0.7, 0.6, and so on. Then, the computer system combines the model predictions from each decision tree model to obtain the final overall model prediction, which represents the degree of frosting on the heat exchanger.

[0113] For example, a computer device can integrate multiple model predictions using an averaging or weighted averaging method to determine the degree of frost on a heat exchanger. For instance, if the computer device calculates an average of multiple model predictions of 0.6, then 0.6 is the current degree of frost on the heat exchanger.

[0114] In the above embodiments, by using multiple decision tree models to process the frost feature data, the overfitting or underfitting problems that may exist in a single model are avoided, enabling a more comprehensive assessment of the frost condition of the heat exchanger, thereby improving prediction accuracy. Furthermore, multiple decision tree models can focus on different frost feature data respectively, making the frost detection model more robust. Moreover, the degree of frost is determined based on the prediction results of multiple decision tree models, and the final prediction result can be determined by flexible weighted averaging or other combinations. Even if individual decision tree models deviate, it will not have a significant impact on the overall prediction result, thus improving the stability and reliability of frost detection.

[0115] As mentioned above, frosting characteristic data includes one or more data types. For example, as shown in Table 1 below, in the frosting detection model corresponding to the direct heat exchanger heat pump type, the required input frosting characteristic data includes 11 types: compressor inlet pressure, compressor inlet temperature, external heat exchanger inlet refrigerant pressure, external heat exchanger inlet refrigerant temperature, cooling fan duty cycle, ambient heat pump running time, system mode, ambient temperature, ambient humidity, vehicle speed, and variable intake grille opening. Similarly, in the frosting detection model corresponding to the secondary heat exchanger heat pump type, the required input frosting characteristic data may include 10 types: compressor inlet pressure, compressor inlet temperature, external heat exchanger inlet coolant temperature, cooling fan duty cycle, ambient heat pump running time, system mode, ambient temperature, ambient humidity, vehicle speed, and variable intake grille opening, etc.

[0116] Table 1

[0117]

[0118]

[0119] Therefore, in some embodiments, the frost feature data is processed by multiple decision tree models to obtain model prediction values ​​corresponding to each decision tree model. This includes: for any decision tree model, determining one or more types of frost feature data corresponding to the target decision tree model from the collected frost feature data; judging whether the corresponding one or more types of frost feature data satisfy the judgment conditions corresponding to each node through multiple nodes in the target decision tree model; the judgment conditions corresponding to each node are used to judge whether the frost feature data of one type of data satisfies a threshold condition; when the corresponding one or more types of frost feature data satisfy the judgment conditions corresponding to any node, the judgment result is output; and based on the judgment result, the model prediction value corresponding to the target decision tree model is obtained.

[0120] Each decision tree model consists of multiple nodes, including a root node, internal nodes, and leaf nodes. The model calculation result corresponding to the leaf node is the model prediction value output by that decision tree model. A decision tree model outputs one model prediction value. During the training phase, the structure of the decision tree model is trained; that is, for a given decision tree model, after training, the data type of the frost feature data input to the root node and the judgment conditions corresponding to each node are determined.

[0121] The judgment condition corresponding to each node is used to determine whether the frosting feature data of a certain data type meets the threshold condition. For example, such as... Figure 5 As shown, for node A, the corresponding judgment condition is: determine the ambient temperature T. amb The system checks if the temperature is less than 2℃, where 2℃ is a preset threshold. If yes, it outputs the corresponding result; otherwise, it proceeds to the next node. The next node's judgment can also be based on the ambient temperature, or it can be based on other data types related to frosting characteristics. For example, the next node, B, might have the condition of determining the compressor inlet temperature T. in If the temperature is less than -4℃, the corresponding judgment result is output; otherwise, the judgment continues to the next node. Furthermore, when it is determined at any node that the frost feature data meets the corresponding judgment condition, the judgment result is output, and this judgment result is used as the model prediction value corresponding to the decision tree model.

[0122] In the above embodiments, by having each decision tree model focus on one or more specific data types, the computer device can more effectively select feature data directly related to the frosting situation for processing, thereby improving the accuracy of judgment. Furthermore, by judging the frosting feature data layer by layer, the decision tree model can adapt to different frosting conditions in terms of data type, realize flexible conditional branch judgment, and thus improve the model's adaptability and generalization ability to complex data.

[0123] After obtaining the model prediction values ​​corresponding to each decision tree model, the computer equipment can integrate the model prediction values ​​corresponding to each decision tree model to determine the final total model prediction value, and thus determine the degree of frost on the heat exchanger.

[0124] Therefore, in some embodiments, the degree of frost on the heat exchanger is determined based on the model prediction values ​​corresponding to all decision tree models, including: determining a preset initial prediction value; obtaining a total model prediction value based on the model prediction values ​​corresponding to all decision tree models; and obtaining the degree of frost on the heat exchanger by combining the initial prediction value with the total model prediction value.

[0125] For decision tree models, an initial prediction value is usually preset. For example, the initial prediction value is usually set to 0.5 to represent the degree of frost on the heat exchanger in the initial or default state.

[0126] The computer device, based on the model predictions corresponding to each decision tree model, performs one or more operations such as summation, weighted summation, or sum of squares on the predictions of each decision tree model to obtain a total model prediction value. For example, if the frost detection model includes three decision tree models, and the computer device obtains frost severity prediction values ​​of 0.4, 0.6, and 0.5 from the three decision tree models respectively, then the computer device adds up 0.4, 0.6, and 0.5 and averages them to obtain a total model prediction value of 0.5.

[0127] In some embodiments, the frost detection model has a learning rate as a model parameter. The computer device then determines the degree of frost on the heat exchanger based on the initial predicted value, the learning rate, and the total predicted value of the model. For example, the computer device can calculate the degree of frost as follows: Frost degree = Initial predicted value + Learning rate * Total predicted value of the model.

[0128] In the above embodiments, by setting an initial prediction value, the initial prediction value of the degree of frost can be dynamically adjusted according to real-time monitoring conditions such as season and climate, so as to make a flexible initial estimate of the degree of frost, and combine it with the total prediction value of the model to more accurately reflect the true situation of the degree of frost.

[0129] Corresponding to the above embodiments, this application also provides a training method for a frosting detection model. The defrosting method for a vehicle heat exchanger can be applied to computer equipment, specifically executed by hardware or software within the computer equipment.

[0130] Computer equipment can be a terminal device or a server. For example, terminal devices include, but are not limited to, one or more of various desktop computers, laptops, smartphones, tablets, in-vehicle terminals, IoT devices, or portable wearable devices. IoT devices can include one or more of smart speakers, smart TVs, smart air conditioners, or smart in-vehicle devices. Portable wearable devices can include one or more of smartwatches, smart bracelets, or head-mounted devices. Servers can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, and big data and artificial intelligence platforms.

[0131] The frost detection model training method provided in this application embodiment can be executed by a computer device or a functional module or entity within a computer device capable of implementing the frost detection model training method. The following description uses a computer device as the execution entity to illustrate the frost detection model training method provided in this application embodiment.

[0132] like Figure 6 As shown, the training method for the frost detection model includes steps 610, 620, 630, and 640.

[0133] Step 610: Obtain multiple sets of frost feature samples related to the frost scene on the heat exchanger surface of the vehicle; each set of frost feature samples is labeled with a sample label, which indicates the actual degree of frost corresponding to the frost feature sample.

[0134] Corresponding to the frost feature data, the frost feature data used in the training process constitutes the frost feature samples. For example, a computer device can divide multiple sets of frost feature samples related to a vehicle's heat exchanger into a training set and a validation set, for example, in an 8:2 ratio or other suitable proportion. The frost detection model is iteratively trained using the training set data, and the trained model is validated using the validation set to avoid overfitting.

[0135] For example, the frosting feature samples include, but are not limited to, one or more of the following data types: compressor inlet pressure, compressor inlet temperature, external heat exchanger inlet refrigerant pressure, external heat exchanger inlet refrigerant temperature, cooling fan duty cycle, ambient heat pump running time, system mode, ambient temperature, ambient humidity, vehicle speed, variable intake grille opening, and external heat exchanger inlet coolant temperature.

[0136] Frost feature samples can be collected during calibration road tests, bench tests, environmental chamber tests, and other processes. Correspondingly, the computer equipment also acquires sample labels corresponding to each group of frost feature samples to indicate the actual degree of frost for that group of frost feature samples.

[0137] The frost detection model is a supervised machine learning model. For example, the frost detection model can be, but is not limited to, one of the following: XGBoost (Extreme Gradient Boosting), Random Forest, Support Vector Machine, or Naive Bayes.

[0138] Step 620: Using the frost detection model to be trained, process each group of frost feature samples and output the predicted degree of frost corresponding to each group of frost feature samples.

[0139] The computer equipment uses a frost detection model to predict the degree of frost for each set of feature samples and outputs a predicted frost level value. For example, for a set of samples, the model predicts an output of 0.55, which is 0.05 different from the actual frost level label of 0.6.

[0140] Step 630: Determine the difference between the predicted degree of frost and the actual degree of frost for each group of frost feature samples.

[0141] The computer equipment processes each set of samples one by one, calculates the difference between the predicted and actual values ​​for each set, and adjusts the model using an optimization algorithm that minimizes the difference. For example, the model parameters are adjusted using a gradient descent algorithm to make the predicted degree of frost closer to the actual value. Exemplarily, the difference can be calculated in ways including but not limited to one or more of variance, square root, standard deviation, and mean variance.

[0142] Step 640: With minimizing the difference as the optimization objective, train the frosting detection model until the difference is less than the error threshold, then end the training and obtain the trained frosting detection model; the trained frosting detection model is used to predict the degree of frosting on the heat exchanger.

[0143] With minimizing the difference as the optimization objective, the computer device can continuously adjust model parameters, such as the learning rate and model gradient, until the difference between the predicted value and the actual value is less than a preset error threshold (e.g., 0.01), thus satisfying the training termination condition. At this point, the computer device saves the trained frost detection model as the final model for predicting frost on the heat exchanger.

[0144] For example, when the frost detection model is a decision tree model, the computer device can adjust the model parameters, which may include adjusting the number of nodes in the decision tree model, the judgment conditions corresponding to the nodes, or increasing or decreasing the number of decision trees, etc.

[0145] For example, such as Figure 7 As shown, the computer equipment will process frosting feature samples of various data types (e.g., N compressor inlet pressure data X). 11 ,X 12 ,……,X N-1,1 ,X N,1 N compressor inlet temperature data X 12 ,X 22 ,……,X N-1,2 ,X N,2 ..., N cooling fan duty cycle data X 1,N-1 ,X 2,N-1 ,……,X N-1,N-1 ,X N,N-1 ..., N vehicle speed data X 1,N ,X 2,N ,……,X N-1,N ,X N,N Using data such as (etc.) as input, N decision tree models are trained through supervised machine learning, and errors are reduced through continuous iterative optimization. Each of the N decision tree models will output its corresponding degree of frost, for example, Y1, Y2, ... Y N-1 ,Y N .

[0146] The training method for the frost detection model provided in this application involves acquiring multiple sets of frost feature samples and processing each set of frost feature samples using the frost detection model to be trained. The predicted frost level corresponding to each set of frost feature samples is then output and compared with the actual frost level represented by the sample label. The difference obtained from the comparison represents the accuracy of the model's prediction of the frost level. Furthermore, the frost detection model is trained with the goal of minimizing the difference until the difference is less than the error threshold. The training then ends, and a trained frost detection model is obtained, capable of accurately judging the frost level of vehicle heat exchangers. Through the above supervised learning method, the model learns the correlation between various features and the actual frost level. The trained frost detection model has high prediction accuracy, strong generalization ability, and robustness, making it applicable to various driving environments.

[0147] In some embodiments, the frost detection model to be trained includes multiple decision tree models to be trained. Accordingly, each set of frost feature samples is processed using the frost detection model to be trained, and a predicted frost degree corresponding to each set of frost feature samples is output. This includes: determining a set of frost feature samples corresponding to each decision tree model to be trained; for any decision tree model to be trained, inputting the corresponding set of frost feature samples into the target decision tree model to obtain the model training value output by the target decision tree model; and using the model training value as the predicted frost degree corresponding to the corresponding set of frost feature samples of the target decision tree model.

[0148] The computer acquires multiple decision tree models and assigns the collected frost feature data to each model according to their requirements. For example, the computer extracts data such as temperature, humidity, and refrigerant flow rate from the frost feature samples, groups this data according to specific rules, and assigns each group of samples to a decision tree model to be trained. The data types of the frost feature samples input to each decision tree model can be the same or different. For example, decision tree model A processes inlet and outlet refrigerant temperatures, while decision tree model B processes airflow rate and humidity.

[0149] For any decision tree model, the computer device inputs a set of assigned frost feature samples into the decision tree. For example, for a certain decision tree model, the input sample data includes a temperature of 0°C and humidity of 85%. The decision tree processes the input data layer by layer according to the node judgment conditions to obtain the model training values. After completing the node judgment, the computer device uses the obtained model training values ​​as the output values ​​of the decision tree model to predict the degree of frost on the input samples. For example, for a set of frost feature samples with a temperature of 0°C and humidity of 85%, the training value output by the decision tree model is 0.7, which is the predicted degree of frost for that set of frost feature samples.

[0150] In the above embodiments, by assigning multiple frost feature samples to different decision tree models for processing, each decision tree model can focus on processing specific types of feature data, which improves the recognition of feature data and the overall prediction accuracy of the model, and avoids the computational burden caused by a single model processing all features, effectively improving the prediction performance of the model.

[0151] In some embodiments, for any decision tree model to be trained, a corresponding set of frost feature samples is input into the target decision tree model to obtain the model training value output by the target decision tree model. This includes: for any decision tree model to be trained, judging one by one whether the input frost feature sample meets the judgment condition corresponding to each node through multiple nodes in the target decision tree model; the judgment condition corresponding to each node is used to judge whether the frost feature data of a certain data type meets the threshold condition; when the input frost feature sample meets the judgment condition corresponding to any node, the judgment result is output; based on the judgment result, the model training value corresponding to the target decision tree model is obtained.

[0152] For any decision tree model to be trained, the computer device inputs a set of frost feature samples into the model. For example, the input set of frost feature samples might be a temperature of -3℃, humidity of 90%, and refrigerant flow rate of 0.3 kg / s. The computer device then checks each node of the decision tree model to see if the input sample meets the corresponding condition. For example, if the condition for the first node is temperature > -5℃, and the input sample's temperature is -3℃, then the condition is met, and the result is output. If not, the computer device passes the data to the next node. The decision tree model traverses each node progressively downwards until a leaf node is reached or the node's condition is no longer met.

[0153] In the above embodiments, by making conditional judgments at each node, the decision tree model can deeply analyze each data type of the frost feature sample, realize a refined prediction of the degree of frost on the heat exchanger, and help improve the prediction accuracy. Furthermore, by utilizing the hierarchical judgment mechanism of the decision tree, the computer device can make rapid judgments at different nodes, making the calculation more hierarchical, avoiding redundant judgments, and improving the efficiency of frost prediction.

[0154] During training, to further optimize prediction accuracy, the computer device can also filter the data types of the frost feature samples. In some embodiments, the above method further includes: determining the number of times each data type is used; the number of times a data type is used as a judgment condition corresponding to a node in the entire decision tree model to be trained; determining the weight of each data type based on the number of times it is used; the higher the weight, the more correlated the frost feature sample of the data type corresponding to the weight is with the predicted degree of frost on the heat exchanger; and removing frost feature samples of data types with weights lower than a preset threshold from multiple sets of frost feature samples.

[0155] The computer system counts the number of times each type of frost feature is used as a node decision condition in all the decision tree models being trained. For example, temperature data is used 10 times across all decision tree models, humidity data is used 8 times, and air velocity data is used 3 times. Based on this usage frequency, the computer system assigns a corresponding weight to each data type. For instance, temperature data, which is used more frequently, is assigned a higher weight, humidity data has a moderate weight, and air velocity data, which is used less frequently, has a lower weight. Assume the weight of temperature data is 0.5, humidity data is 0.3, and air velocity data is 0.1.

[0156] Therefore, the computer equipment can remove data types with weights below a preset threshold from the feature sample set. If the preset weight threshold is 0.2, then the air velocity data (weight 0.1) is below the threshold, indicating that it is basically irrelevant to whether frost forms on the heat exchanger surface. The computer equipment will then remove this data and retain the temperature and humidity data. Thus, the computer equipment can identify one or more frost feature samples of data types most relevant to the frost formation scenario on the heat exchanger surface.

[0157] For example, after training, the computer equipment, through screening, finally obtains the weights of frosting feature samples of various data types corresponding to direct heat pump type and secondary heat exchange heat pump type, as shown in Table 2:

[0158] Table 2

[0159]

[0160]

[0161] In the above embodiments, by calculating the number of times data types are used and assigning weights accordingly, the model can identify data types that are more critical to predicting the degree of frost. Eliminating low-weight data types can reduce the participation of irrelevant or less influential data in model training, thereby further improving prediction accuracy.

[0162] The sample labeling can be done manually or in the following way: obtain a first sample image and a second sample image corresponding to each group of frost feature samples. The first sample image is used to represent the surface frost area of ​​the heat exchanger, and the second sample image is used to represent the side frost thickness of the heat exchanger. Based on the surface frost area and the side frost thickness, determine the sample label corresponding to each group of frost feature samples.

[0163] During calibration road tests, bench tests, and environmental chamber tests, a camera can be temporarily installed on the front and side of the heat exchanger of the vehicle under test to detect the actual degree of frost on the vehicle under test, which will serve as the sample label corresponding to a set of frost feature data collected.

[0164] Computer equipment can acquire images representing the area of ​​frost on the surface of a heat exchanger using a camera or similar device, and record these images as first sample images. For example, the computer equipment can capture an image of a frost-covered surface on the heat exchanger using a camera; the proportion of white pixels in this image can represent the area of ​​the frost layer. The computer equipment can also acquire an image representing the thickness of the frost layer on the side of the heat exchanger from another angle, and record this image as a second sample image. For example, a side-view camera can capture the thickness distribution of the frost layer on the side of the heat exchanger.

[0165] Therefore, the computer equipment analyzes the first sample image to extract the coverage area of ​​the frost layer on the heat exchanger surface; simultaneously, it analyzes the second sample image to obtain the average thickness of the frost layer on the side of the heat exchanger. Assuming the surface frost area is 20 square centimeters and the side frost thickness is 2 millimeters, the computer equipment then determines a sample label for each group of frost characteristic samples based on the aforementioned frost area and thickness values. Assuming this sample label represents the severity of the frost, the area and thickness results are combined to determine the value of this sample label as "Level 2".

[0166] For example, when the degree of frosting is expressed in terms of frosting volume, frosting volume = frosting area * frosting thickness.

[0167] In the above embodiments, by acquiring sample images of surface area and side thickness, the area and thickness of the frost layer can be comprehensively considered, thereby improving the accuracy of detecting the degree of frost formation on the heat exchanger. Furthermore, the automatic generation of sample labels based on image analysis avoids the tedious process of manual annotation, simplifies data preprocessing, and improves model training efficiency.

[0168] As previously mentioned, vehicle heat pump systems include direct heat pump systems and secondary heat exchange heat pump systems. Accordingly, during training, the computer equipment will also train appropriate frosting detection models based on both direct and secondary heat exchange heat pump systems. Specifically, in some embodiments, the method further includes: determining frosting feature samples of a first data type matching the direct heat pump system, and training based on these samples to obtain a frosting detection model corresponding to the direct heat pump system; and determining frosting feature samples of a second data type matching the secondary heat exchange heat pump system, and training based on these samples to obtain a frosting detection model corresponding to the secondary heat exchange heat pump system.

[0169] When the computer identifies a vehicle's heat pump system as a direct heat pump system, it filters out the first data type that matches the direct heat pump system from the frosting feature dataset. For example, the frosting features of a direct heat pump system are mainly related to refrigerant temperature, refrigerant pressure, and outside air temperature; the computer selects these data types as the first data type.

[0170] When the computer identifies a vehicle's heat pump system as a secondary heat exchange system, it filters out a second data type that matches the secondary heat exchange system. For example, the frosting characteristics of a secondary heat exchange system are more affected by coolant temperature, coolant flow rate, and heat exchange efficiency; the computer selects these data types as the second data type.

[0171] Therefore, the computer equipment establishes and trains frost detection models based on the determined first and second data types, respectively, and finally obtains frost detection models specifically for direct heat pump systems and frost detection models specifically for secondary heat exchange heat pump systems.

[0172] In the above embodiments, by matching the characteristic data types of different heat pump systems, the resulting frost detection model can better fit the actual operating characteristics of the heat pump system, improving the accuracy of frost degree prediction. Furthermore, by selecting the most relevant data type for training based on the heat pump system type, the model becomes more adaptable to the operating conditions of different heat pump systems, thereby reducing misjudgments and false alarms and improving the model reliability in practical applications. Moreover, by matching the different characteristic data types of direct heat pump systems and secondary heat exchange heat pump systems to the corresponding models, interference from irrelevant data is reduced, training samples are optimized, and computational efficiency is improved.

[0173] In a specific example, the control logic for the vehicle to perform defrosting operations based on the degree of frost and location information can be as follows: Figure 8 The example illustrates the data collection, model training, and practical application processes. In this example, the data collection process gathers frost feature samples of various data types and determines the corresponding sample labels. During model training, multiple decision tree models are trained, and the weights for each data type are obtained. After training, the model can be deployed to a cloud server for in-vehicle use. For example, a computer device can act as the in-vehicle terminal, calling the frost detection model deployed on the cloud server. Alternatively, the model can be deployed locally on the vehicle to better protect driving privacy and security. During vehicle operation, various data can be collected to cover more driving conditions. This allows for subsequent updates and upgrades to improve detection accuracy. The defrosting strategy is as follows: when the frost level exceeds the upper threshold, a usable heat source exists, and the remaining driving time to the destination exceeds 30 minutes, the defrosting mode is activated, and defrosting operations are performed until the detected frost level drops below the lower threshold. Otherwise, if any condition is not met, the defrosting mode is not activated.

[0174] The specific judgment process can be as follows: Figure 9As shown, the computer equipment predicts the degree of frost using a frost detection model and makes the following judgments sequentially: First, it checks if the frost level has reached the upper threshold Fmax. If so, it checks if there is a usable heat source. If so, it further checks if the current location is more than 30 minutes away from the destination. If so, it activates the PTC or compressor to start defrosting until the frost level falls below the lower threshold Fmin. If any condition is not met, the computer equipment does not perform defrosting until the next defrosting condition is determined. For example, if the computer equipment predicts that the frost level exceeds the upper threshold Fmax and there is a usable heat source, but the distance to the destination is less than 30 minutes, the computer equipment will wait until the vehicle arrives at the destination before starting defrosting.

[0175] The defrosting method for vehicle heat exchangers provided in this application can be executed by a defrosting device for the vehicle heat exchanger. This application uses an example of a defrosting device for the vehicle heat exchanger executing the defrosting method to illustrate the defrosting device for the vehicle heat exchanger provided in this application.

[0176] This application also provides a defrosting device for a vehicle heat exchanger, such as... Figure 10 As shown, the defrosting device for the vehicle heat exchanger includes a data acquisition module 1010, a prediction module 1020, and an execution module 1030. Wherein:

[0177] The data acquisition module 1010 is used to acquire frost feature data related to frost formation on the surface of a vehicle's heat exchanger.

[0178] The prediction module 1020 is used to call the frost detection model to process the frost feature data and obtain the degree of frost on the heat exchanger; the frost detection model is a pre-trained machine learning model.

[0179] The execution module 1030 is used to determine whether the defrosting conditions are met based on the degree of frost and the vehicle's location information, and to defrost the heat exchanger if the defrosting conditions are met.

[0180] The defrosting device for vehicle heat exchangers provided in this application acquires frost feature data related to frost formation on the heat exchanger surface and processes this data using a frost detection model. This allows for real-time prediction of the frost level on the heat exchanger. Through multi-dimensional data analysis, the device accurately determines the frost level, significantly improving the accuracy and real-time performance of frost detection. Furthermore, based on the frost level and the collected vehicle location information, the device intelligently determines whether immediate defrosting is necessary, optimizing the defrosting strategy logic and avoiding excessive energy consumption caused by frequent defrosting cycles, thus improving the vehicle's energy efficiency. Moreover, by using a pre-trained frost detection model to detect the frost level, there is no need to install additional cameras or sensors specifically designed for frost detection on the vehicle. This makes it suitable for various vehicle configurations and driving environments, demonstrating good adaptability.

[0181] In some embodiments, the frost detection model includes multiple decision tree models; the prediction module is further configured to process the frost feature data through the multiple decision tree models respectively to obtain the model prediction value corresponding to each decision tree model; and determine the degree of frost on the heat exchanger based on the model prediction values ​​corresponding to all decision tree models respectively.

[0182] In some embodiments, the frost feature data includes one or more data types; the prediction module is further configured to, for any decision tree model, determine one or more data types of frost feature data corresponding to the target decision tree model from the collected frost feature data; through multiple nodes in the target decision tree model, determine whether the corresponding one or more data types of frost feature data satisfy the judgment conditions corresponding to each node; the judgment conditions corresponding to each node are used to determine whether the frost feature data of one data type satisfies the threshold condition; when the corresponding one or more data types of frost feature data satisfy the judgment conditions corresponding to any node, output the judgment result; based on the judgment result, obtain the model prediction value corresponding to the target decision tree model.

[0183] In some embodiments, the prediction module is further configured to determine a preset initial prediction value; obtain a total model prediction value based on the model prediction values ​​corresponding to all decision tree models; and obtain the degree of frost formation on the heat exchanger based on the initial prediction value and the total model prediction value.

[0184] In some embodiments, the execution module is further configured to determine whether the degree of frost exceeds an upper limit threshold; determine the remaining travel time for the vehicle to move from its current location to its destination location; and, if the degree of frost exceeds the upper limit threshold and the remaining travel time exceeds a duration threshold, determine that the degree of frost and the vehicle's location information meet the defrosting conditions.

[0185] In some embodiments, the execution module is further configured to invoke a navigation service to obtain the remaining travel time for the vehicle to move from its current location to its destination; or, to obtain traffic information and determine the remaining travel time for the vehicle to move from its current location to its destination based on the traffic information.

[0186] In some embodiments, the execution module is further configured to determine whether there is a heat source in the vehicle that is in a usable state; and if there is at least one heat source in a usable state, the degree of frost exceeds the upper limit threshold, and the remaining driving time exceeds the duration threshold, determine that the degree of frost and the vehicle's location information meet the defrosting conditions.

[0187] In some embodiments, the execution module is further configured to perform a preset defrosting operation when the defrosting conditions are met, so that the degree of frost is lower than a lower threshold; wherein the preset defrosting operation includes at least one of heating defrosting and compression defrosting.

[0188] In some embodiments, the above-mentioned apparatus further includes a verification model for determining the type of heat pump system in the vehicle; the heat pump system type includes a direct heat pump system and a secondary heat exchange heat pump system; if the vehicle is a direct heat pump system, the frosting detection model is determined to be the frosting detection model corresponding to the direct heat pump system; if the vehicle is a secondary heat exchange heat pump system, the frosting detection model is determined to be the frosting detection model corresponding to the secondary heat exchange heat pump system; wherein, the model structure of the frosting detection models corresponding to different heat pump system types is different, and the data types of the frosting feature data required for input are different for each type.

[0189] In some embodiments, the execution module is further configured to determine the operation duration of a preset defrosting operation based on the degree of frost and / or a preset charging time when the vehicle is in a charging state; and to perform the preset defrosting operation according to the operation duration.

[0190] The method for training a frost detection model for a vehicle heat exchanger provided in this application embodiment can be executed by a training device for the frost detection model of a vehicle heat exchanger. This application embodiment uses the example of a training device for the frost detection model of a vehicle heat exchanger executing the training method for the frost detection model of a vehicle heat exchanger to illustrate the training device for the frost detection model of a vehicle heat exchanger provided in this application embodiment.

[0191] This application also provides a training device for a frosting detection model of a vehicle heat exchanger, such as... Figure 11 As shown, the training device for the frost detection model of the vehicle heat exchanger includes an acquisition module 1111 and a training module 1120, wherein:

[0192] The acquisition module 1111 is used to acquire multiple sets of frost feature samples related to the frost scene on the heat exchanger surface of the vehicle; each set of frost feature samples is labeled with a sample label, which indicates the actual degree of frost corresponding to the frost feature sample.

[0193] The training module 1120 is used to process each group of frost feature samples through the frost detection model to be trained, and output the predicted degree of frost corresponding to each group of frost feature samples.

[0194] The training module 1120 is also used to determine the difference between the predicted degree of frost and the actual degree of frost corresponding to each group of frost feature samples.

[0195] The training module 1120 is also used to train the frosting detection model with the goal of minimizing the difference until the difference is less than the error threshold, at which point the training ends and the trained frosting detection model is obtained; the trained frosting detection model is used to predict the degree of frosting on the heat exchanger.

[0196] The training device for the frost detection model of a vehicle heat exchanger provided in this application acquires multiple sets of frost feature samples and processes each set of frost feature samples using the frost detection model to be trained. It outputs a predicted degree of frost corresponding to each set of frost feature samples, and then compares this predicted degree with the actual degree of frost represented by the sample label. The difference obtained from the comparison represents the accuracy of the model's prediction of the degree of frost. Furthermore, by minimizing the difference as the optimization objective, the frost detection model is trained until the difference is less than the error threshold. The training then ends, and a trained frost detection model is obtained, capable of accurately judging the degree of frost on the vehicle heat exchanger. Through the above-mentioned supervised learning method, the model can learn the correlation between various features and the actual degree of frost. The trained frost detection model has high prediction accuracy, strong generalization ability, and robustness, and can be applied to various driving environments.

[0197] In some embodiments, the frost detection model to be trained includes multiple decision tree models to be trained; the training module is further configured to determine a set of frost feature samples corresponding to each decision tree model to be trained; for any decision tree model to be trained, the corresponding set of frost feature samples is input into the target decision tree model to be trained to obtain the model training value output by the target decision tree model to be trained; the model training value is used as the predicted degree of frost corresponding to the corresponding set of frost feature samples of the target decision tree model to be trained.

[0198] In some embodiments, the frost feature samples include one or more data types; the training module is further configured to, for any decision tree model to be trained, determine, through multiple nodes in the decision tree model to be trained, whether the input frost feature sample satisfies the judgment condition corresponding to each node; the judgment condition corresponding to each node is used to determine whether the frost feature data of a certain data type satisfies the threshold condition; when the input frost feature sample satisfies the judgment condition corresponding to any node, the judgment result is output; based on the judgment result, the model training value corresponding to the decision tree model to be trained is obtained.

[0199] In some embodiments, the training module is further configured to determine the number of times a data type is used for one or more data types; the number of times a data type is used as a judgment condition for a node in the entire decision tree model to be trained; based on the number of times it is used, the weight of each data type is determined; the higher the weight, the more correlated the frosting feature sample of the data type corresponding to the weight is with the predicted frosting degree of the heat exchanger; in multiple sets of frosting feature samples, frosting feature samples of data types with weights lower than a preset threshold are removed.

[0200] In some embodiments, the training module is further configured to acquire a first sample image and a second sample image corresponding to each group of frost feature samples, wherein the first sample image represents the surface frost area of ​​the heat exchanger and the second sample image represents the side frost thickness of the heat exchanger; and based on the surface frost area and the side frost thickness, determine the sample label corresponding to each group of frost feature samples.

[0201] In some embodiments, the vehicle's heat pump system type includes a direct heat pump system and a secondary heat exchange heat pump system; the training module is further configured to determine frost feature samples of a first data type that match the direct heat pump system, and train based on the frost feature samples of the first data type to obtain a frost detection model corresponding to the direct heat pump system; determine frost feature samples of a second data type that match the secondary heat exchange heat pump system, and train based on the frost feature samples of the second data type to obtain a frost detection model corresponding to the secondary heat exchange heat pump system.

[0202] The defrosting device for the vehicle heat exchanger or the training device for the frost detection model of the vehicle heat exchanger in the embodiments of this application can be a computer device or a component of the computer device, such as an integrated circuit or a chip. The computer device can be a terminal or other devices besides a terminal. For example, the computer device can be a mobile phone, tablet computer, laptop computer, handheld computer, in-vehicle computer device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM or self-service machine, etc. The embodiments of this application do not specifically limit it.

[0203] The defrosting device for the vehicle heat exchanger or the training device for the frost detection model of the vehicle heat exchanger in the embodiments of this application can be a device with an operating system. The operating system can be a Microsoft (Windows) operating system, an Android operating system, an iOS operating system, or other possible operating systems. This application embodiment does not specifically limit the specific operating system.

[0204] The defrosting device for vehicle heat exchangers provided in this application embodiment can achieve... Figure 2 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0205] The training device for the frost detection model of the vehicle heat exchanger provided in this application embodiment can achieve... Figure 6 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0206] In some embodiments, such as Figure 12 As shown, this application embodiment also provides a computer device 1200, including a processor 1201, a memory 1202, and a computer program stored in the memory 1202 and executable on the processor 1201. When the program is executed by the processor 1201, it implements the various processes of the above-described method embodiments and can achieve the same technical effects. To avoid repetition, it will not be described again here.

[0207] It should be noted that the computer devices in this application embodiment include the mobile computer devices and non-mobile computer devices described above.

[0208] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described vehicle heat exchanger defrosting method embodiment and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0209] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described training method embodiment for the frost detection model of a vehicle heat exchanger and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0210] The processor is the processor in the computer device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0211] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described defrosting method for a vehicle heat exchanger.

[0212] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the training method for the above-described frost detection model of a vehicle heat exchanger.

[0213] The processor is the processor in the computer device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0214] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described vehicle heat exchanger defrosting method embodiment and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0215] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described training method embodiment for the frost detection model of the vehicle heat exchanger, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0216] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0217] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0218] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0219] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

[0220] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0221] Unless otherwise specified, all embodiments and optional embodiments of this application can be combined to form new technical solutions.

[0222] Unless otherwise specified, all technical features and optional technical features of this application may be combined to form new technical solutions.

[0223] Unless otherwise specified, all steps of this application may be performed sequentially or randomly, preferably sequentially. For example, the method includes steps (a) and (b), indicating that the method may include steps (a) and (b) performed sequentially, or it may include steps (b) and (a) performed sequentially. For example, the mention that the method may also include step (c) indicates that step (c) may be added to the method in any order; for example, the method may include steps (a), (b), and (c), or it may include steps (a), (c), and (b), or it may include steps (c), (a), and (b), etc.

[0224] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A defrosting method for a vehicle heat exchanger, characterized by, The method comprises: obtaining frosting feature data related to a frosting scenario of a heat exchanger surface of a vehicle; calling a frosting detection model to process the frosting feature data to obtain a frosting degree of the heat exchanger; the frosting detection model is a pre-trained machine learning model; determining whether a defrosting condition is met based on the frosting degree and location information of the vehicle, and defrosting the heat exchanger if the defrosting condition is met.

2. The method of claim 1, wherein, The frosting detection model comprises a plurality of decision tree models; the calling of the frosting detection model to process the frosting feature data to obtain the frosting degree of the heat exchanger comprises: processing the frosting feature data through the plurality of decision tree models respectively to obtain model prediction values respectively corresponding to each decision tree model; determining the frosting degree of the heat exchanger based on the model prediction values respectively corresponding to all the decision tree models.

3. The method of claim 2, wherein, The frosting feature data comprises one or more data types; the processing of the frosting feature data through the plurality of decision tree models respectively to obtain the model prediction values respectively corresponding to each decision tree model comprises: for any decision tree model, determining, from the collected frosting feature data, frosting feature data of one or more data types corresponding to the decision tree model; judging, through a plurality of nodes in the decision tree model, whether the frosting feature data of the one or more data types corresponds to a judgment condition respectively corresponding to each node; the judgment condition corresponding to each node is used to judge whether the frosting feature data of one data type meets a threshold condition; outputting a judgment result when the frosting feature data of the one or more data types meets the judgment condition corresponding to any node; obtaining the model prediction value corresponding to the decision tree model based on the judgment result.

4. The method according to claim 2 or 3, characterized in that, The determination of the frosting degree of the heat exchanger based on the model prediction values respectively corresponding to all the decision tree models comprises: determining a preset initial prediction value; obtaining a total model prediction value based on the model prediction values respectively corresponding to all the decision tree models; obtaining the frosting degree of the heat exchanger based on the initial prediction value and the total model prediction value.

5. The method according to any one of claims 1 to 4, characterized in that, The determination of whether the defrosting condition is met based on the frosting degree and the location information of the vehicle comprises: determining whether the frosting degree exceeds an upper threshold; determining a remaining driving duration of the vehicle from a current location to a destination location; determining that the frosting degree and the location information of the vehicle meet the defrosting condition if the frosting degree exceeds the upper threshold and the remaining driving duration exceeds a duration threshold.

6. The method according to any one of claims 1 to 5, characterized in that, The determination of whether the defrosting condition is met based on the frosting degree and the location information of the vehicle further comprises: determining whether there is a heat source in an available state in the vehicle; determining that the frosting degree and the location information of the vehicle meet the defrosting condition if there is at least one heat source in an available state, the frosting degree exceeds the upper threshold, and the remaining driving duration exceeds the duration threshold.

7. The method according to claim 5 or 6, characterized in that, The defrosting of the heat exchanger if the defrosting condition is met comprises: When the defrosting condition is met, a preset defrosting operation is performed to make the frosting degree lower than a lower threshold; wherein the preset defrosting operation at least includes one of heating defrosting and compression defrosting.

8. The method according to any one of claims 1 to 7, characterized in that, Before the calling the frosting detection model to process the frosting feature data, the method further comprises: determining a heat pump system type in the vehicle; the heat pump system type includes a direct heat pump system and a secondary heat exchange heat pump system; if the vehicle is a direct heat pump system, determining the frosting detection model as a frosting detection model corresponding to the direct heat pump system; if the vehicle is a secondary heat exchange heat pump system, determining the frosting detection model as a frosting detection model corresponding to the secondary heat exchange heat pump system; wherein the model structures of the frosting detection models corresponding to different heat pump system types are different, and the data types of the frosting feature data required by each frosting detection model are different.

9. The method according to any one of claims 1 to 8, characterized in that, The method further comprises: when the vehicle is in a charging state, determining an operation time length of the preset defrosting operation based on the frosting degree and / or a preset charging time length; performing the preset defrosting operation according to the operation time length.

10. A training method for a frosting detection model, characterized in that, comprises: obtaining a plurality of sets of frosting feature samples related to a frosting scene of a heat exchanger surface of a vehicle; wherein each set of frosting feature samples is respectively labeled with a sample label, and the sample label represents an actual frosting degree corresponding to the frosting feature sample; processing each set of frosting feature samples respectively through a to-be-trained frosting detection model to output a predicted frosting degree corresponding to each set of frosting feature samples respectively; determining a difference between the predicted frosting degree and the actual frosting degree corresponding to each set of frosting feature samples; training the frosting detection model with the minimization of the difference as an optimization objective, until the difference is less than an error threshold, ending the training and obtaining a trained frosting detection model; the trained frosting detection model is used to predict the frosting degree of the heat exchanger.

11. The method of claim 10, wherein, The to-be-trained frosting detection model comprises a plurality of to-be-trained decision tree models; the processing each set of frosting feature samples respectively through the to-be-trained frosting detection model to output a predicted frosting degree corresponding to each set of frosting feature samples respectively comprises: determining a set of frosting feature samples corresponding to each to-be-trained decision tree model respectively; for any to-be-trained decision tree model, inputting the corresponding set of frosting feature samples into the to-be-trained decision tree model to obtain a model training value output by the to-be-trained decision tree model; taking the model training value as a predicted frosting degree corresponding to the corresponding set of frosting feature samples of the to-be-trained decision tree model.

12. The method according to claim 10 or 11, characterized in that, The frosting feature sample comprises one or more data types; the inputting the corresponding set of frosting feature samples into the to-be-trained decision tree model to obtain the model training value output by the to-be-trained decision tree model for any to-be-trained decision tree model comprises: For any to-be-trained decision tree model, whether the input frost feature sample meets the judgment condition corresponding to each node in the to-be-trained decision tree model is judged one by one through the nodes in the to-be-trained decision tree model; When the input frost feature sample meets the judgment condition corresponding to any node, a judgment result is outputted; Based on the judgment result, a model training value corresponding to the to-be-trained decision tree model is obtained.

13. The method according to any one of claims 10 to 12, characterized in that, The method further comprises: determining the number of uses corresponding to the one or more data types; the number of uses represents the number of times that a data type is used as a data type in the judgment condition corresponding to a node in all to-be-trained decision tree models; based on the number of uses, the weight of each data type is determined respectively; the higher the weight, the more relevant the frost feature sample corresponding to the weight to the predicted frost degree of the heat exchanger; in the plurality of groups of frost feature samples, the frost feature samples under the data type with a weight lower than a preset threshold are removed.

14. The method according to any one of claims 10 to 13, characterized in that, The method further comprises: obtaining a first sample image and a second sample image corresponding to each group of frost feature samples respectively; the first sample image is used to represent the surface frost layer area of the heat exchanger, and the second sample image is used to represent the side frost layer thickness of the heat exchanger; based on the surface frost layer area and the side frost layer thickness, the sample label corresponding to each group of frost feature samples is determined respectively.

15. The method according to any one of claims 10 to 14, characterized in that, The heat pump system type of the vehicle includes a direct heat pump system and a secondary heat exchange heat pump system; the method further comprises: determining the frost feature sample of the first data type matched with the direct heat pump system, and training based on the frost feature sample of the first data type to obtain the frost detection model corresponding to the direct heat pump system; determining the frost feature sample of the second data type matched with the secondary heat exchange heat pump system, and training based on the frost feature sample of the second data type to obtain the frost detection model corresponding to the secondary heat exchange heat pump system.

16. A defrosting device for a vehicle heat exchanger, characterized by The device comprises: a collection module for obtaining frost feature data related to the frost formation scene on the surface of the heat exchanger of the vehicle; a prediction module for calling the frost detection model to process the frost feature data to obtain the frost degree of the heat exchanger; the frost detection model is a pre-trained machine learning model; an execution module for determining whether the defrosting condition is met based on the frost degree and the location information of the vehicle, and defrosting the heat exchanger if the defrosting condition is met.

17. A training device for a frosting detection model of a vehicle heat exchanger, characterized in that, The device comprises: an acquisition module for acquiring a plurality of groups of frost feature samples related to the frost formation scene on the surface of the heat exchanger of the vehicle; wherein each group of frost feature samples is labeled with a sample label, and the sample label represents the actual frost degree corresponding to the frost feature sample; a training module for processing each group of frost feature samples through a to-be-trained frost detection model to output a predicted frost degree corresponding to each group of frost feature samples respectively; The training module is further configured to determine a difference between a predicted frost degree corresponding to each set of frost feature samples and an actual frost degree. The training module is further configured to train the frost detection model with minimization of the difference as an optimization objective, and end the training and obtain a trained frost detection model when the difference is less than an error threshold.

18. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, The processor implements the defrosting method of the vehicle heat exchanger according to any one of claims 1-9, or the training method of the frost detection model according to claims 10-15 when executing the program.

19. A vehicle characterized by comprising: The computer device according to claim 18 is provided. 20.A non-transitory computer-readable storage medium having stored thereon a computer program. The computer program is executed by the processor to implement the defrosting method of the vehicle heat exchanger according to any one of claims 1-9, or the training method of the frost detection model according to any one of claims 10-15.

21. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the defrosting method of the vehicle heat exchanger according to any one of claims 1-9, or the training method of the frost detection model according to any one of claims 10-15.

Images