Temperature prediction method and device, electronic equipment and computer readable storage medium

Abstract

The application provides a temperature prediction method and device, electronic equipment and a computer readable storage medium. The method comprises: obtaining at least one set of historical operation parameters of a target inverter; obtaining at least one set of historical heat dissipation parameters of the target inverter, one set of historical heat dissipation parameters corresponding to the same historical time information as one set of historical operation parameters; inputting each set of historical operation parameters, each set of historical heat dissipation parameters and the historical time information into a trained temperature prediction model, and predicting the operation temperature of the target inverter through the temperature prediction model. Thus, the temperature prediction method provided by the application not only considers the historical operation parameters of the target inverter, but also considers the historical heat dissipation parameters corresponding to the historical operation parameters at the same historical time information, so that the operation temperature of the inverter at a future time can be more effectively and accurately predicted. If the predicted temperature of the inverter in the future is high, appropriate warning measures can be taken to avoid damage to the inverter due to the high temperature in the future.

Description

Technical Field

[0001] This application relates to the field of electrical technology, specifically to a temperature prediction method, apparatus, electronic device, and computer-readable storage medium. Background Technology

[0002] An inverter is a device that converts direct current (DC) into alternating current (AC) for use by loads or for connection to the power grid. It is one of the core components of a power generation system. Inverters directly affect power generation efficiency and operational stability, and therefore occupy an important position in the power system.

[0003] The core component of an inverter, the IGBT (Insulated Gate Bipolar Transistor), generates a significant amount of heat during operation. Approximately 1-1.5% of the rated power is consumed within the IGBT and converted into heat. This heat causes the power device's die to heat up and the junction temperature to rise. If this heat is not dissipated promptly and effectively, it will affect the device's performance, thereby reducing system reliability and even damaging the device. Therefore, it is essential to know the inverter's operating temperature to prevent damage caused by overheating. Summary of the Invention

[0004] This application provides a temperature prediction method that can effectively predict the operating temperature of an inverter.

[0005] In a first aspect, this application provides a temperature prediction method, the method comprising:

[0006] Obtain at least one set of historical operating parameters for the target inverter;

[0007] Obtain at least one set of historical heat dissipation parameters of the target inverter, wherein one set of historical heat dissipation parameters corresponds to the same historical time information as one set of historical operating parameters;

[0008] The historical operating parameters, historical heat dissipation parameters, and historical time information of each group are input into the temperature prediction model to complete the training, and the operating temperature of the target inverter is predicted by the temperature prediction model.

[0009] In some embodiments of this application, the step of inputting the historical operating parameters of each group, the historical heat dissipation parameters of each group, and the historical time information into a trained temperature prediction model, and predicting the operating temperature of the target inverter through the temperature prediction model, includes:

[0010] Obtain the cycle operation parameters and cycle heat dissipation parameters corresponding to the information at each cycle time in the historical cycle;

[0011] The historical operating parameters and heat dissipation parameters of each group are compared with the periodic operating parameters and heat dissipation parameters of each periodic time to determine the target periodic operating parameters and the target periodic heat dissipation parameters, and the target periodic time of the target periodic operating parameters and the target periodic heat dissipation parameters is obtained.

[0012] Determine the time difference between the time to be predicted and the most recent historical time in the historical time information;

[0013] Based on the time difference, determine the associated time in the historical period corresponding to the target periodic time;

[0014] Determine the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time.

[0015] Based on the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time, the operating temperature of the target inverter at the time to be predicted is predicted.

[0016] In some embodiments of this application, before obtaining at least one set of historical operating parameters of the target inverter, the method further includes:

[0017] Obtain multiple sets of inverter operating parameters and multiple sets of inverter heat dissipation parameters for training the temperature prediction model;

[0018] Time encoding is performed on inverter operating parameters and inverter heat dissipation parameters that have the same time information to obtain time encoding information of the inverter operating parameters and time encoding information of the inverter heat dissipation parameters.

[0019] Based on the time-coded information of the inverter operating parameters and the time-coded information of the inverter heat dissipation parameters, the inverter operating parameters and the inverter heat dissipation parameters of each group are sorted to obtain training samples including time series information.

[0020] The temperature prediction model is trained based on the training samples, which include time series information.

[0021] In some embodiments of this application, obtaining multiple sets of inverter operating parameters for training the temperature prediction model includes:

[0022] Obtain multiple sets of initial inverter operating parameters;

[0023] Determine the parameter types included in the initial inverter operating parameters for each group;

[0024] Using each set of initial inverter operating parameters as the target initial inverter operating parameter set, determine whether the target initial inverter operating parameter set has any missing parameter types;

[0025] If the target initial inverter operating parameter group does not have any missing parameter types, the target initial inverter operating parameter group is directly used as a set of inverter operating parameters;

[0026] If there are missing target parameter types in the target initial inverter operating parameter group, fill in the missing parameters of the target parameter types to obtain a set of inverter operating parameters.

[0027] In some embodiments of this application, the method of filling in the missing parameters of the target parameter type includes:

[0028] Determine the target timestamp for the target initial inverter operating parameter group;

[0029] Determine the adjacent initial inverter operating parameter group corresponding to the adjacent timestamps adjacent to the target timestamp;

[0030] Determine the parameter similarity between the adjacent initial inverter operating parameter group including the target parameter type and the target initial inverter operating parameter group, and obtain the target adjacent parameter group whose parameter similarity meets the target setting;

[0031] Based on the parameters corresponding to the target parameter type in the target adjacent parameter group, determine the parameters of the target parameter type in the target initial inverter operating parameter group.

[0032] In some embodiments of this application, obtaining multiple sets of inverter operating parameters for training the temperature prediction model includes:

[0033] Obtain multiple sets of initial inverter operating parameters;

[0034] Each parameter in each set of initial inverter operating parameters is standardized and processed to obtain multiple sets of inverter operating parameters.

[0035] In some embodiments of this application, obtaining multiple sets of inverter heat dissipation parameters for training the temperature prediction model includes:

[0036] Using the heat dissipation parameters of each inverter as the target inverter heat dissipation parameters, determine the current temperature of the target inverter under the corresponding target time information;

[0037] Determine the apparent power of the target inverter at the target time information;

[0038] The power ratio is determined based on the apparent power and the rated power of the target inverter;

[0039] If the current temperature is less than the minimum endpoint of the preset temperature range, the first heat dissipation parameter is determined based on the power ratio;

[0040] If the current temperature is within the temperature range, determine the second heat dissipation parameter based on the current temperature;

[0041] The maximum heat dissipation parameter among the first heat dissipation parameter and the second heat dissipation parameter is determined as a set of inverter heat dissipation parameters, thereby obtaining multiple sets of inverter heat dissipation parameters.

[0042] In some embodiments of this application, training the temperature prediction model based on the training samples including time-series information includes:

[0043] The training samples, including time series information, are divided into training set samples, validation set samples, and test set samples according to a set ratio.

[0044] Each training sample in the training set is sequentially used as the current training sample and input into the temperature prediction model to be trained to obtain the predicted temperature at the second historical moment after the first historical moment corresponding to the current training sample. The second historical moment corresponds to at least one target training sample in the training set other than the current training sample.

[0045] The loss function value is calculated based on the predicted temperature at the second historical moment and the operating temperature parameters of the inverter corresponding to the second historical moment in the target training sample.

[0046] The model metrics of the temperature prediction model are iteratively updated based on the loss function value until the loss function value meets the iteration termination condition, thus obtaining the trained temperature prediction model.

[0047] The trained temperature prediction model is validated and tested using training samples from the validation set and the test set, respectively, thus completing the training of the temperature prediction model.

[0048] In some embodiments of this application, training the temperature prediction model based on the training samples including time-series information includes:

[0049] Based on the time series information of the training samples in the training set, the trend characteristics, seasonal characteristics, and error characteristics of temperature information under different operating parameters and different inverter heat dissipation parameters are determined.

[0050] Based on the trend characteristics, the seasonal characteristics, and the error characteristics, a prediction sample including time series information is obtained;

[0051] The first feature loss is determined based on the trend characteristics, seasonal characteristics, and error characteristics of the predicted sample.

[0052] The second feature loss is determined based on the time series information of the predicted sample;

[0053] The temperature prediction model is adjusted based on the first feature loss and the second feature loss to obtain a temperature prediction model that has completed initial training.

[0054] Secondly, this application also provides a temperature prediction device, the device comprising:

[0055] The first acquisition module is used to acquire at least one set of historical operating parameters of the target inverter;

[0056] The second acquisition module is used to acquire at least one set of historical heat dissipation parameters of the target inverter, wherein the set of historical heat dissipation parameters corresponds to the same historical time information as a set of historical operating parameters.

[0057] The prediction module is used to input the historical operating parameters of each group, the historical heat dissipation parameters of each group, and the historical time information into the temperature prediction model to complete the training, and to predict the operating temperature of the target inverter through the temperature prediction model.

[0058] Thirdly, this application also provides an electronic device, the electronic device including a processor, a memory, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the steps in any of the temperature prediction methods described above.

[0059] Fourthly, this application also provides a computer-readable storage medium storing a computer program that is executed by a processor to implement the steps in any of the temperature prediction methods described above.

[0060] The temperature prediction method provided in this application not only considers the historical operating parameters of the target inverter, but also the historical heat dissipation parameters corresponding to the historical operating parameters at the same historical time. This enables a more effective and accurate prediction of the inverter's operating temperature in the future. If the predicted temperature of the inverter is high in the future, corresponding early warning measures can be taken to avoid damage to the inverter due to excessively high temperatures. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of a temperature prediction system provided in the embodiments of this application;

[0063] Figure 2 This is a schematic flowchart of one embodiment of the temperature prediction method in this application;

[0064] Figure 3 This is a table diagram illustrating the historical operating parameters of the target inverter in one embodiment of the temperature prediction method in this application.

[0065] Figure 4 This is a schematic diagram of the architecture of the temperature prediction model used in one embodiment of the temperature prediction method in this application.

[0066] Figure 5 This is a schematic diagram of a functional module of the temperature prediction device in the embodiments of this application;

[0067] Figure 6 This is a schematic diagram of the structure of the electronic device in the embodiments of this application. Detailed Implementation

[0068] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0069] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0070] In this application, the term "exemplary" is used to mean "used as an example, illustration, or description." Any embodiment described as "exemplary" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. Furthermore, it is understood that in the specific embodiments of this application, user information, user data, and other related data are involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions.

[0071] To enable any person skilled in the art to implement and use this application, the following description is provided. In this description, details are set forth for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be implemented without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of this application with unnecessary detail. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0072] This application provides a temperature prediction method, apparatus, device, and storage medium, which are described in detail below.

[0073] Please see Figure 1 , Figure 1 This is a schematic diagram of a temperature prediction system provided in an embodiment of this application. The temperature prediction system may include an electronic device 100 and a storage device 200, and the storage device 200 may transmit data to the electronic device 100. Figure 1 The electronic device 100 in the storage device 200 can obtain the historical operating parameters and historical heat dissipation parameters of the inverter stored in the storage device 200, so as to execute the temperature prediction method in this application and realize the prediction of the operating temperature of the inverter at future times.

[0074] In this embodiment, the electronic device 100 may be, but is not limited to, a desktop computer, a portable computer, a network server, a PDA (Personal Digital Assistant), a tablet computer, a wireless electronic device, an embedded device, a standalone server, or a server network or server cluster composed of servers.

[0075] In the embodiments of this application, the electronic device 100 and the storage device 200 can communicate through any communication method, including but not limited to mobile communication based on the 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), and Worldwide Interoperability for Microwave Access (WiMAX), or computer network communication based on the TCP / IP Protocol Suite (TCP / IP) and User Datagram Protocol (UDP).

[0076] It should be noted that, Figure 1The schematic diagram of the temperature prediction system shown is merely an example. The temperature prediction system and scenario described in this application are intended to more clearly illustrate the technical solutions of this application and do not constitute a limitation on the technical solutions provided in this application. As those skilled in the art will know, with the evolution of temperature prediction systems and the emergence of new business scenarios, the technical solutions provided in this application are also applicable to similar technical problems.

[0077] like Figure 2 As shown, Figure 2 This is a flowchart illustrating a temperature prediction method in one embodiment of the present application. The temperature prediction method may include the following steps 201 to 203.

[0078] 201. Obtain at least one set of historical operating parameters of the target inverter.

[0079] In this embodiment, the target inverter can be the inverter whose temperature is to be predicted. A set of historical operating parameters can include various parameters, such as temperature, inverter current, and inverter voltage. Specific parameters are not limited in this embodiment. As an example, obtaining the historical operating parameters of the target inverter can be achieved by recording the operating parameters at a certain historical moment when the target inverter is in operation. Simultaneously, operating parameters with the same historical moment can be stored in the same storage location, such as in the same folder. When historical operating parameters need to be obtained, the historical operating parameters with the same historical moment are grouped together to obtain a set of historical operating parameters.

[0080] 202. Obtain at least one set of historical heat dissipation parameters of the target inverter, and the set of historical heat dissipation parameters corresponds to the same historical time information as the set of historical operating parameters.

[0081] In this embodiment, the heat dissipation parameters of the target inverter can be the speed of the cooling fan, the temperature of the cooling water, etc. A set of historical heat dissipation parameters can include multiple parameters, such as the speed of the cooling fan and the temperature of the cooling water simultaneously. The specific parameters depend on the actual heat dissipation devices installed in the target inverter. For example, if both an air-cooled fan and a water-cooled pump are installed, the heat dissipation parameters can include both the fan speed and the cooling water temperature. Alternatively, if the target inverter only has one type of heat dissipation device, the heat dissipation parameters of that device can be obtained. In other words, if the target inverter only has one type of heat dissipation device, a set of historical heat dissipation parameters can include only one type of heat dissipation parameter; this embodiment does not limit the specifics. In one implementation of this embodiment, when the target inverter is in operation, the heat dissipation parameters at a certain historical moment can be recorded. Simultaneously, heat dissipation parameters with the same historical moment are stored in the same storage location, such as in the same folder. When historical heat dissipation parameters need to be obtained, a set of historical heat dissipation parameters can be obtained by grouping the historical heat dissipation parameters with the same historical moment into a single data set.

[0082] 203. Input the historical operating parameters, historical heat dissipation parameters and historical time information of each group into the trained temperature prediction model, and predict the operating temperature of the target inverter through the temperature prediction model.

[0083] After obtaining each set of historical operating parameters and each set of historical heat dissipation parameters, the historical operating parameter groups and historical heat dissipation parameter groups with the same historical time information can be input into the trained temperature prediction model, either sequentially or simultaneously. This allows the trained temperature prediction model to determine, based on the same historical time information, which specific historical heat dissipation parameters correspond to which specific historical operating parameters of the target inverter at a given moment. It should be noted that the more arrays of historical operating parameters and historical heat dissipation parameters input, the better the temperature prediction model can identify the actual temperature of the inverter under specific operating and heat dissipation parameters, and consequently, the better the temperature prediction model can predict the temperature of the target inverter at future moments.

[0084] For example: Assuming the current time is 9:00 AM on [Date], we obtain the historical operating parameters and historical heat dissipation parameters of the target inverter for each minute between 8:30 AM and 9:00 AM on [Date]. If the model analysis shows that the current, voltage, and temperature within the target inverter are all increasing between 8:30 AM and 9:00 AM on [Date], and heat dissipation parameters, such as fan speed, are also increasing, then the target inverter is in a phase of increasing load pressure. To predict the temperature of the target inverter at 9:01 AM on [Date], we can obtain a predicted temperature value based on the temperature change curve, operating parameter change curve, and heat dissipation parameter change curve obtained from the model analysis. For example, if the operating parameter change curve indicates that the inverter is in the load increase phase and the inverter load subsequently reaches a stable level, and the temperature change curve also shows a pattern of first rising and then stabilizing, and the heat dissipation parameter change curve also shows a pattern of first rising and then stabilizing, then it can be determined that the target inverter's load will not change in the future, and the heat dissipation parameters will not change. In this case, it can be determined that the predicted temperature value at 9:01 AM on [Date] is the same as the stable value in the temperature change curve, thus completing the temperature prediction. Of course, there are other situations, and the specific temperature can be predicted based on the specific actual situation, which will not be elaborated here.

[0085] It should be noted that the model architecture of the temperature prediction model trained in this application can be any model architecture, and the training method can also be any training method, such as supervised training or unsupervised training. Specific embodiments of this application are not limited.

[0086] The temperature prediction method provided in this application can effectively predict the future operating temperature of the inverter by acquiring the historical operating parameters of the target inverter and the historical heat dissipation parameters corresponding to the historical operating parameters at the same historical time. If the predicted future temperature of the inverter is high, corresponding early warning measures can be taken to avoid damage to the inverter due to excessively high temperatures.

[0087] To better implement the embodiments of this application, in one embodiment, inputting each set of historical operating parameters, each set of historical heat dissipation parameters, and historical time information into a trained temperature prediction model, and predicting the operating temperature of the target inverter through the temperature prediction model, may include the following steps:

[0088] First, obtain the cycle operation parameters and cycle heat dissipation parameters corresponding to the information of each cycle moment in the historical cycle;

[0089] Then, the historical operating parameters and the historical heat dissipation parameters of each group are compared with the periodic operating parameters and periodic heat dissipation parameters at each periodic time to determine the target periodic operating parameters and the target periodic heat dissipation parameters, and obtain the target periodic time of the target periodic operating parameters and the target periodic heat dissipation parameters.

[0090] Next, determine the time difference between the time to be predicted and the most recent historical time in the historical time information;

[0091] Secondly, based on the time difference, determine the corresponding time in the historical cycle for the target period.

[0092] In addition, determine the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time points;

[0093] Finally, based on the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time, the operating temperature of the target inverter at the time to be predicted is predicted.

[0094] In the above embodiments, a model is provided that extracts features from input data to obtain corresponding curves, thereby obtaining the temperature change pattern of the target inverter and predicting its temperature. However, at certain special times, the inverter may reduce its load from a high-load state, leading to a temperature drop. If the historical operating parameters and historical heat dissipation parameters are obtained at the peak of the high load, the inverter temperature may subsequently decrease. Therefore, the prediction result may not match the actual situation, with the predicted temperature being higher than the actual temperature. Therefore, the model needs to learn the operating pattern of the target inverter during the training phase. For example, the historical operating parameters and historical heat dissipation parameters of the target inverter at different times each day can be provided to the model, allowing the model to grasp the historical periodicity of the target inverter. Therefore, the historical period in this embodiment can be a periodic feature that the temperature prediction model learns on its own after training. For example, after training, the model determines that the target inverter operates according to the same pattern every day, including that the inverter inside the solar panel starts working at 6:00 AM when the sun rises and ends at 6:00 PM when the sun sets. Furthermore, longer historical periods, including seasonal historical periods, are also considered. For example, the duration of daylight hours differs between winter and summer. If the current time is winter, a winter-related historical period is used; if the current time is summer, a summer-related historical period is used. In summary, when a temperature prediction model is needed, it can determine which specific historical period applies. Assuming seasonal factors are not considered, the historical period can be determined in days.

[0095] Simultaneously, after determining the historical period, it is necessary to obtain the daily operating patterns and temperature variation patterns of the target inverter at different times. Specifically, with a daily historical period, each historical period corresponds to specific periodic operating parameters and periodic heat dissipation parameters. After obtaining each set of input historical operating parameters and each set of historical heat dissipation parameters, the model can match the time information corresponding to each set of historical operating parameters and each set of historical heat dissipation parameters with the time information in the historical period, calculating the similarity between the historical heat dissipation parameters, historical operating parameters, and the corresponding periodic operating parameters and periodic heat dissipation parameters at the matched time information. When the parameter similarity is greater than a certain similarity threshold, it is determined that the currently input historical parameters match the parameters in the period, further confirming that the operating pattern of the current target inverter still satisfies periodicity.

[0096] Specifically, this can be represented as follows: Assuming we obtain and compare the historical operating parameters and heat dissipation parameters of each set with the periodic features in the model, if the historical operating parameters and heat dissipation parameters match the periodic parameters within the model—for example, assuming the periodic features in the model are on a daily basis, and the historical operating parameters match the periodic operating parameters from 8:30 AM to 9:00 AM within a single period—and simultaneously, the historical heat dissipation parameters also match the periodic heat dissipation parameters from 8:30 AM to 9:00 AM, then we can determine that the current historical operating parameters and heat dissipation parameters essentially map to the data from 8:30 AM to 9:00 AM each day. Therefore, we can confirm that the obtained historical operating parameters and heat dissipation parameters satisfy periodicity.

[0097] Once it's confirmed that the model still satisfies periodicity, the time to be predicted is obtained according to the instructions, and then the prediction is performed. It should be noted that the time to be predicted may not be adjacent to the most recent historical time in the input historical parameters. For example, the input historical parameters correspond to the historical operating parameters and historical heat dissipation parameters per minute between 8:30 AM on [Month] XX and 9:00 AM on [Year] XX. As in the above embodiment, the predicted time could be the temperature at 9:01 AM on [Year] XX. In reality, temperatures at other times may be needed. For example, 9:05 AM or 9:10 AM on [Year] XX, which are not adjacent to 9:00. Therefore, assuming the obtained historical operating parameters and historical heat dissipation parameters are between 8:30 AM on [Month] XX and 9:00 AM on [Year] XX, and the time to be predicted is 9:05 AM on [Year] XX, then 9:05 AM on [Year] XX is 5 minutes different from the most recent historical time between 8:30 AM on [Month] XX and 9:00 AM on [Year] XX. This 5-minute difference is the time difference. Therefore, if the historical cycle is based on time, 9:05 AM in the historical cycle is the associated time of 9:05 AM on XX / XX / XX / XX, which can also be considered the corresponding time. Furthermore, the temperature information corresponding to 9:05 AM in the historical cycle can be regarded as the temperature information corresponding to 9:05 AM on XX / XX / XX / XX. At this point, the temperature prediction for the time to be predicted is complete. Similarly, assuming that 2 PM is the turning point of light intensity every day, if the obtained historical parameters are between 1:30 PM on XX / XX / XX and 2 PM on XX / XX / XX, according to the above embodiment, it may be impossible to predict that the inverter temperature may subsequently decrease. However, through this embodiment, it is possible to effectively determine that the temperature will reverse, thus enabling more accurate temperature prediction.

[0098] To better implement the embodiments of this application, in an alternative example, before obtaining at least one set of historical operating parameters of the target inverter, the temperature prediction method of this application further includes a step of training the temperature prediction model, as detailed below.

[0099] First, multiple sets of inverter operating parameters and multiple sets of inverter heat dissipation parameters are obtained for training the temperature prediction model.

[0100] Secondly, time encoding is performed on the inverter operating parameters and inverter heat dissipation parameters that have the same time information to obtain the time encoding information of the inverter operating parameters and the time encoding information of the inverter heat dissipation parameters.

[0101] Then, based on the time-coded information of the inverter operating parameters and the time-coded information of the inverter heat dissipation parameters, the inverter operating parameters and the inverter heat dissipation parameters of each group are sorted to obtain training samples including time series information.

[0102] Finally, a temperature prediction model is trained based on training samples that include time-series information.

[0103] The above embodiments provide a scheme for temperature prediction based on historical cycles. Therefore, this application provides a training scheme for a model to grasp the periodic patterns of an inverter. For example, it obtains various types of operating parameters and heat dissipation parameters of the target inverter within 5 years. Of course, the heat dissipation parameters and operating parameters used for training also need specific historical time information. To optimize the input operating parameters and heat dissipation parameters, the time information can be encoded, for example, the time information of XX month XX day XX hour can be converted into an information code. The specific encoding formula is shown in Table 1 below:

[0104] minute current_minute / 59.0-0.5 Hour current_hour / 23.0-0.5 sky current_day / 30.0-0.5 moon current_month / 11.0-0.5

[0105] Table 1

[0106] After time-coding the inverter operating parameters and heat dissipation parameters for each input group, the parameters can be sorted according to the size of the encoding, i.e., the chronological order, to obtain training samples containing time-series information. It should be noted that when sorting the parameters by time, different periods such as year, month, and day can be used to create various training samples. After obtaining multiple training samples, these samples are then fed into the model for training, enabling the model to capture the periodic characteristics of the inverter.

[0107] To better implement the embodiments of this application, in one embodiment of this application, obtaining multiple sets of inverter operating parameters for temperature prediction model training may include the following steps:

[0108] First, obtain multiple sets of initial inverter operating parameters;

[0109] Then, determine the parameter types included in the initial inverter operating parameters for each group;

[0110] Finally, using each set of initial inverter operating parameters as the target initial inverter operating parameter set, determine whether there are any missing parameter types in the target initial inverter operating parameter set. If there are no missing parameter types in the target initial inverter operating parameter set, directly use the target initial inverter operating parameter set as a set of inverter operating parameters. If there are missing target parameter types in the target initial inverter operating parameter set, fill in the missing parameters of the target parameter types to obtain a set of inverter operating parameters.

[0111] The above embodiments provide an implementation method for feeding training samples into a model for training. However, when obtaining the historical operating parameters of the inverter, certain preprocessing operations can be performed on the historical operating parameters to improve the quality of the subsequently generated training samples.

[0112] Specifically, the above embodiments illustrate that a set of historical operating parameters can include various types of data, such as, but not limited to, daily power generation, ambient temperature, phase A voltage, phase A current, input voltage, input current, active power, inverter temperature, etc. However, when collecting data from the inverter during inverter operation, incomplete data collection may occur, resulting in the inability to obtain some types of data. For example, assuming there are multiple sets of historical operating data, theoretically, each set of historical operating data should include various types of data such as daily power generation, ambient temperature, phase A voltage, phase A current, input voltage, input current, active power, and inverter temperature. However, due to collection issues, some historical operating parameter sets may lack inverter temperature, active power, or other types of parameters. Therefore, this application embodiment provides a scheme for filling in missing parameters.

[0113] Specifically, when acquiring each historical parameter group, the total number of data types involved in all historical parameter groups can be detected to determine the total number of parameter types. Then, it is determined whether the number of parameter types in each historical parameter group matches the total number of parameter types. If they do not match, it can be determined that a parameter type is missing in the historical parameter group. If a parameter type is missing in the current historical parameter group, the specific missing target parameter type is further detected. Then, based on the time information of the current historical parameter group, the two parameters of the same type corresponding to the two most recent adjacent time information of the current historical parameter group are identified, and the weighted average of these two parameters is used to fill the missing target parameter type in the current historical parameter group.

[0114] For example: Suppose the current historical operating parameter group is missing inverter temperature information, and the corresponding time information for the current historical operating parameter group is XX year XX month XX day XX hour. In this case, determine the two temperature information from the two most recent and adjacent time information before and after XX year XX month XX day XX hour in the two historical operating parameter groups, perform a weighted average calculation, and fill the missing temperature information in the current historical operating parameter group with the calculated result. It should be noted that adjacent historical operating parameter groups may also have missing parameters, such as missing parameter types that are the same as those in the current historical operating parameter group. In this case, search for the parameter with the smallest time difference between the current historical operating parameter group and the previous and next historical operating parameter groups, and fill it with the parameter that does not have a missing target parameter type. Of course, if there is no missing parameter type in the current historical operating parameter group, skip the parameter filling step.

[0115] Furthermore, the initial inverter operating parameters obtained for training can be as follows: Figure 3 As shown. Figure 3 As shown, the air temperature type parameter in the fourth initial inverter operating parameter group corresponding to serial number 4 is missing. In this case, the air temperature in the initial inverter operating parameter group with serial number 3 and the initial inverter operating parameter group with serial number 5 can be averaged, and the average value can be filled into the temperature type of the fourth initial inverter operating parameter group as the temperature parameter. This will give us a set of inverter operating parameters for training. By analogy, we can get all the inverter operating parameter groups.

[0116] To better implement the embodiments of this application, in one embodiment of this application, filling in parameters with missing target parameter types includes:

[0117] First, determine the target timestamp for the target initial inverter operating parameter group;

[0118] Then, determine the adjacent initial inverter operating parameter group corresponding to the adjacent timestamps adjacent to the target timestamp;

[0119] Secondly, determine the parameter similarity between the adjacent initial inverter operating parameter group, including the target parameter type, and the target initial inverter operating parameter group, to obtain the target adjacent parameter group whose parameter similarity meets the target setting;

[0120] Finally, based on the parameters corresponding to the target parameter type in the target adjacent parameter group, determine the parameters of the target parameter type in the target initial inverter operating parameter group.

[0121] The above embodiments provide a scheme for data imputation by calculating the average value. However, in order to make the imputed data as close to reality as possible, this application embodiment also provides a method for calculating data similarity. Specifically, using... Figure 3 To illustrate, Figure 3 In the initial inverter operating parameter group with serial number 4, the air temperature parameter is missing. In this case, the similarity can be calculated by taking the Euclidean distance between the three sets of data adjacent to the timestamp of serial number 4. Specifically, refer to the following formula (1):

[0122]

[0123] In formula (1), i is the parameter type number, x can be a parameter in the current target initial inverter operating parameter group, and y is a parameter in the adjacent target initial inverter operating parameter group. Through the above formula (1) and... Figure 3 The data can be obtained from Table 2:

[0124] European distance 38.3 28.5 25.0 24.1 45.9

[0125] Table 2

[0126] From Table 2, we can see that Figure 3 The third and fifth groups in the sequence have the lowest similarity values ​​with the fourth group, which are 25.0 and 24.1 respectively. At this point, the average value of the temperature values ​​of the third and fifth groups is taken and filled into the temperature type of the fourth group.

[0127] To better implement the embodiments of this application, in one embodiment, multiple sets of inverter operating parameters for training a temperature prediction model are obtained, including:

[0128] Obtain multiple sets of initial inverter operating parameters; perform standardized data processing on each parameter in each set of initial inverter operating parameters to obtain multiple sets of inverter operating parameters.

[0129] When obtaining the operating parameters of each inverter group, this application can perform data standardization processing on each inverter operating parameter group. For example, the z-score standardization method is used to standardize the data based on the mean and standard deviation of the original data. Generally, the mean is normalized to 0 and the variance is normalized to 1. The specific formula is shown in the following formula (2):

[0130]

[0131] X' is the standardized value, and x is a parameter value of a specific parameter type in a set of inverter operating parameters. σ represents the average value of parameters of the same parameter type, and σ represents the standard deviation of parameters of the same parameter type. It should be noted that the average value and standard deviation can be determined according to the actual situation, and this application does not limit them.

[0132] To better implement the embodiments of this application, in one embodiment, multiple sets of inverter heat dissipation parameters are obtained for training the temperature prediction model, including:

[0133] Using the heat dissipation parameters of each inverter as the target inverter heat dissipation parameters, determine the current temperature of the target inverter at the corresponding target time information; determine the apparent power of the target inverter at the target time information; determine the power ratio based on the apparent power and the rated power of the target inverter; if the current temperature is less than the minimum endpoint of the preset temperature range, determine the first heat dissipation parameter based on the power ratio; if the current temperature is within the temperature range, determine the second heat dissipation parameter based on the current temperature; determine the maximum heat dissipation parameter between the first and second heat dissipation parameters as a set of inverter heat dissipation parameters, thereby obtaining multiple sets of inverter heat dissipation parameters.

[0134] The above embodiments provide a scheme for obtaining inverter operating parameters for model training. This application also provides a scheme for obtaining inverter heat dissipation parameters. In this application, the inverter heat dissipation parameters can refer to the rotational speed of the inverter cooling fan.

[0135] Since the heat dissipation parameters of each target inverter group include the inverter's operating temperature at the corresponding moment, the operating temperature at the corresponding moment can be determined directly based on the heat dissipation parameters of each target inverter group. Simultaneously, the apparent power of the inverter at the corresponding moment can be determined using formula (3):

[0136] S = P a / cosθ……(3)

[0137] Where: S is the apparent power of the inverter at the corresponding moment, P a Let θ be the active power of the inverter at the corresponding moment, and cos θ be the reactive power of the inverter at the corresponding moment. Then, the power ratio can be calculated using the following formula (4):

[0138] ratio p =S / P r ……(4)

[0139] Where: ratio p P represents the power percentage of the inverter at the corresponding time, S represents the apparent power of the inverter at the corresponding time, and P represents the apparent power of the inverter at the corresponding time. r This represents the rated power of the inverter at the corresponding moment.

[0140] It should be noted that, in this embodiment of the application, the preset temperature range for setting the module temperature is [T min T max The power ratio constraint range is [r]. min rmax The temperature in the current target inverter's heat dissipation parameter set is less than T. min When: If the power percentage is less than r min Set the fan speed to 0; if the power percentage is within [r] min r max Within, rotational speed r cur The formula for the normalized speed value calculated based on the power ratio when adjusting from 0 to 100% is as follows (5):

[0141] n F-P =(r cur -r min ) / (r max -r min )……(5)

[0142] Final normalized fan speed: n F =n F-P This yields the first heat dissipation parameter. If the power percentage is greater than r... max Set the fan speed normalization value to 100%. If the current module temperature is within [T... min T max The formula for calculating the rotational speed based on the module temperature is as follows (6):

[0143] n F-T =(T cur -T min ) / (T max -T min )……(6)

[0144] Where, n F-T This is the second cooling parameter. The final normalized fan speed value: n F =max(n F-T, n F-P This means selecting the maximum value between the first and second heat dissipation parameters as a set of inverter heat dissipation parameters. Additionally, if the current temperature is greater than T... max At this time: the fan speed normalization value is set to 100%. In summary, based on the above scheme for determining the inverter heat dissipation parameters, the heat dissipation parameters for each group of inverters can be calculated separately.

[0145] To better implement the embodiments of this application, in one embodiment of this application, training a temperature prediction model based on training samples including time series information may include the following steps:

[0146] First, multiple training samples containing time series information are divided into training set samples, validation set samples, and test set samples according to a set ratio.

[0147] Then, each training sample in the training set is used as the current training sample and input into the temperature prediction model to be trained to obtain the predicted temperature of the second historical moment after the first historical moment corresponding to the current training sample. The second historical moment corresponds to at least one target training sample in the training set other than the current training sample.

[0148] Secondly, the loss function value is calculated based on the predicted temperature at the second historical moment and the operating temperature parameters of the inverter corresponding to the second historical moment in the target training sample.

[0149] Next, the model metrics of the temperature prediction model are iteratively updated based on the loss function value until the loss function value meets the iteration termination condition, thus obtaining the trained temperature prediction model.

[0150] Finally, the trained temperature prediction model is validated and tested using training samples from both the validation set and the test set, thus completing the training of the temperature prediction model.

[0151] In this embodiment of the application, the scheme for dividing the training sample set, the validation sample set, and the test sample set can be divided in a 7:1:2 ratio. Of course, other methods can also be used. The specific implementation of this application is not limited.

[0152] Since the training samples include time-series information, and the inverter operating parameters and heat dissipation parameters involved in the training samples are collected based on the historical operation of the inverter, the temperature prediction model can predict the training samples based on the learned features to obtain the prediction results. It should be noted that, assuming the total time-series range of the training samples covers inverter operating parameters and heat dissipation parameters within one year, the temperature prediction model can predict the data for the first six months of that year. That is, data before the first historical moment can be considered as data for the first six months of the year, while data after the first historical moment can be considered as data for the second half of the year. Since the inverter operating parameters and heat dissipation parameters are essentially real operating parameters, the sample data for the second half of the year can be considered as real data and used to compare the temperature prediction model's prediction results for the first six months' data, thereby obtaining the model's training loss. Then, by adjusting the model's parameters, the loss value can be made less than a preset requirement. It should be noted that the first and second historical moments listed in this application are not necessarily six-month time points; they can be other time points, and the specific embodiments of this application are not limited to these.

[0153] For example, the current training samples include Sample 1 and Sample 2, where Sample 2 is the training sample whose time series follows Sample 1. Sample 1 corresponds to the first historical time T1, and Sample 2 corresponds to the second historical time T2, with the first historical time T1 being earlier than the second historical time T2. During training, the inverter operating parameters in Sample 1 are used to predict the inverter operating temperature Pre-TempT2 at the second historical time T2. Then, Pre-TempT2 is compared and analyzed with the inverter operating temperature Sam-TempT2 recorded in Sample 2 corresponding to the second historical time T2. The calculated loss function value is used to iteratively update the weight parameters of the temperature prediction model according to the set gradient direction. This training method eliminates the need for manual or machine calibration of each training sample, saving sample processing time.

[0154] Furthermore, the training samples in the validation set and the training samples in the test set are the same as the training samples in the training set, which will not be elaborated here.

[0155] To better implement the embodiments of this application, in one embodiment, a temperature prediction model is trained based on training samples including time series information, including:

[0156] Based on the time series information of the training samples in the training set, the trend characteristics, seasonal characteristics, and error characteristics of temperature information under different operating parameters and different inverter heat dissipation parameters are determined. Based on the trend characteristics, seasonal characteristics, and error characteristics, prediction samples including time series information are obtained. Based on the trend characteristics, seasonal characteristics, and error characteristics of the prediction samples, the first feature loss is determined. Based on the time series information of the prediction samples, the second feature loss is determined. Based on the first feature loss and the second feature loss, the temperature prediction model is adjusted to obtain the temperature prediction model that has completed the initial training.

[0157] The above embodiments provide a scheme for learning the periodic characteristics of an inverter through a temperature prediction model. However, in reality, the inverter temperature is easily affected by the ambient temperature, which typically follows a certain pattern. For example, from winter to spring and then to summer, the ambient temperature gradually rises, and the inverter's operating temperature shows an increasing trend, while from summer to autumn and then to winter, the inverter's operating temperature shows a decreasing trend. Therefore, when training the model, it can be trained to learn the trend characteristics of temperature changes, seasonal characteristics, and some error characteristics.

[0158] For example, after obtaining training samples containing time-series information, trend features and seasonal features can be extracted based on the time-series information of the training samples, as shown in the following formula (7):

[0159] X(t)=T(t)+S(t)+R(t)……(7)

[0160] Here, X(t) represents the training samples including time-series information, and T(t), S(t), and R(t) correspond to trend features, seasonal features, and error features, respectively. When the temperature prediction model determines that the inverter is showing a downward trend over a certain period, it can be represented by 1; conversely, when it determines that the temperature is showing an upward trend over a certain period, it can be represented by 0. Seasonal features can be represented by 00 for spring, 01 for summer, 10 for autumn, and 11 for winter. Error features can be represented by 1 indicating the presence of error and 0 indicating the absence of error. For example, during the transition from summer to winter, temperatures may fluctuate during certain autumn periods, eventually returning to summer temperatures; in this case, the high-temperature moments within the autumn period can be identified as error terms. Subsequently, the trend features, seasonal features, and error features are integrated into the periodic features learned by the model for further learning.

[0161] After extracting the features, this application performs loss calculation, and the training loss of the model can also be calculated using formula (8), as shown in the following formula (8):

[0162]

[0163] in, The first feature loss is used to compare the predicted sequence with the true sequence; The time loss between the predicted sequence and the true sequence is called the second feature loss. The reason for this time loss is that at certain moments, the feature similarity is high, but the moments don't correspond. α is a hyperparameter with a value between [0,1], used to weight the shape and time. The method for obtaining the predicted and true sequences is the same as in the previous embodiment, and will not be repeated here.

[0164] Furthermore, it should be noted that the model architecture of the temperature prediction model used in the embodiments of this application can be as follows: Figure 4 As shown, each layer consists of three modules: Auto-Correlation, Series Decomp, and Feed Forward, while also using a residual structure. The Encoder eliminates long-term trend-period terms through the sequence decomposition module and focuses on modeling seasonal features. The Decoder progressively accumulates the trend-period terms extracted from the latent variables. The encoder-decoder-Auto-Correlation (Decoder) utilizes past seasonal term information from the Encoder.

[0165] Simultaneously, during training, when determining periodic features, the model can truncate and shift the acquired feature vectors, with the truncation length adaptively adjusted by the model. Each time a vector is truncated, it is shifted sequentially, calculating which part of the subsequent vector is similar to the truncated vector, thus determining the displacement. This displacement is essentially a temporal shift; therefore, as the model progressively completes vector truncation, shifting, and feature similarity calculation, it can obtain the periodic features.

[0166] To better implement the temperature prediction method in the embodiments of this application, a temperature prediction device is also provided in the embodiments of this application, such as... Figure 5 As shown, the temperature prediction device 300 includes:

[0167] The first acquisition module 301 is used to acquire at least one set of historical operating parameters of the target inverter;

[0168] The second acquisition module 302 is used to acquire at least one set of historical heat dissipation parameters of the target inverter, wherein the set of historical heat dissipation parameters corresponds to the same historical time information as a set of historical operating parameters.

[0169] The prediction module 303 is used to input the historical operating parameters, historical heat dissipation parameters and historical time information of each group into the temperature prediction model after training, and predict the operating temperature of the target inverter through the temperature prediction model.

[0170] The temperature prediction device provided in this application acquires the historical operating parameters of the target inverter and the historical heat dissipation parameters corresponding to the historical operating parameters at the same historical time through the first acquisition module 301 and the second acquisition module 302. Then, the prediction module 303 can effectively predict the operating temperature of the inverter in the future. If the predicted temperature of the inverter in the future is high, corresponding early warning measures can be taken to avoid damage to the inverter due to excessively high temperatures.

[0171] In some embodiments of this application, the prediction module 303 is specifically used for:

[0172] Obtain the cycle operation parameters and cycle heat dissipation parameters corresponding to the information at each cycle time in the historical cycle;

[0173] The historical operating parameters and heat dissipation parameters of each group are compared with the periodic operating parameters and heat dissipation parameters of each periodic time to determine the target periodic operating parameters and target periodic heat dissipation parameters, and thus obtain the target periodic time of the target periodic operating parameters and target periodic heat dissipation parameters.

[0174] Determine the time difference between the time to be predicted and the most recent historical time in the historical time information;

[0175] Based on the time difference, determine the corresponding time in the historical cycle for the target period.

[0176] Determine the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time points;

[0177] Based on the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time, the operating temperature of the target inverter at the time to be predicted is predicted.

[0178] In some embodiments of this application, the temperature prediction device further includes a training module, which is specifically used for:

[0179] Obtain multiple sets of inverter operating parameters and multiple sets of inverter heat dissipation parameters for training the temperature prediction model;

[0180] Time-encode the inverter operating parameters and inverter heat dissipation parameters that have the same time information to obtain the time-encoded information of the inverter operating parameters and the time-encoded information of the inverter heat dissipation parameters.

[0181] Based on the time-coded information of inverter operating parameters and inverter heat dissipation parameters, the inverter operating parameters and inverter heat dissipation parameters of each group are sorted to obtain training samples including time series information.

[0182] A temperature prediction model is trained based on training samples that include time-series information.

[0183] In some embodiments of this application, the temperature prediction device further includes a parameter acquisition module, which is specifically used for:

[0184] Obtain multiple sets of initial inverter operating parameters;

[0185] Determine the parameter types included in the initial inverter operating parameters for each group;

[0186] Using each set of initial inverter operating parameters as the target initial inverter operating parameter set, determine whether there are any missing parameter types in the target initial inverter operating parameter set;

[0187] If the target initial inverter operating parameter group does not have any missing parameter types, the target initial inverter operating parameter group is directly used as a set of inverter operating parameters;

[0188] If there are missing target parameter types in the target initial inverter operating parameter group, fill in the missing parameters of the target parameter types to obtain a set of inverter operating parameters.

[0189] In some embodiments of this application, the parameter acquisition module is further used for:

[0190] Determine the target timestamp for the target initial inverter operating parameter group;

[0191] Determine the adjacent initial inverter operating parameter group corresponding to the adjacent timestamps adjacent to the target timestamp;

[0192] Determine the parameter similarity between the adjacent initial inverter operating parameter group (including the target parameter type) and the target initial inverter operating parameter group to obtain the target adjacent parameter group whose parameter similarity meets the target setting;

[0193] Based on the parameters corresponding to the target parameter type in the adjacent parameter group of the target, determine the parameters of the target parameter type in the target initial inverter operating parameter group.

[0194] In some embodiments of this application, the parameter acquisition module is further used for:

[0195] Obtain multiple sets of initial inverter operating parameters;

[0196] Each parameter in each set of initial inverter operating parameters is standardized and processed to obtain multiple sets of inverter operating parameters.

[0197] In some embodiments of this application, the parameter acquisition module is further used for:

[0198] Using the heat dissipation parameters of each inverter as the target inverter heat dissipation parameters, determine the current temperature of the target inverter under the corresponding target time information;

[0199] Determine the apparent power of the target inverter under the target time information;

[0200] The power ratio is determined based on the apparent power and the rated power of the target inverter;

[0201] If the current temperature is less than the minimum endpoint of the preset temperature range, the first heat dissipation parameter is determined based on the power ratio.

[0202] If the current temperature is within the temperature range, determine the second heat dissipation parameter based on the current temperature;

[0203] The maximum heat dissipation parameter among the first and second heat dissipation parameters is determined as a set of inverter heat dissipation parameters, thereby obtaining multiple sets of inverter heat dissipation parameters.

[0204] In some embodiments of this application, the training module is further used for:

[0205] Multiple training samples containing time series information are divided into training set samples, validation set samples, and test set samples according to a set ratio;

[0206] Each training sample in the training set is used as the current training sample and input into the temperature prediction model to be trained. The predicted temperature of the second historical moment after the first historical moment corresponding to the current training sample is obtained. The second historical moment corresponds to at least one target training sample in the training set other than the current training sample.

[0207] The loss function value is calculated based on the predicted temperature at the second historical moment and the operating temperature parameters of the inverter corresponding to the second historical moment in the target training sample.

[0208] The model metrics of the temperature prediction model are iteratively updated based on the loss function value until the loss function value meets the iteration termination condition, thus obtaining the trained temperature prediction model.

[0209] The trained temperature prediction model was validated and tested using training samples from both the validation set and the test set, thus completing the training of the temperature prediction model.

[0210] In some embodiments of this application, the training module is further used for:

[0211] Based on the time series information of the training samples in the training set, the trend characteristics, seasonal characteristics and error characteristics of temperature information under different operating parameters and different inverter heat dissipation parameters are determined.

[0212] Based on trend characteristics, seasonal characteristics, and error characteristics, a prediction sample including time series information is obtained;

[0213] The first feature loss is determined based on the trend characteristics, seasonal characteristics, and error characteristics of the predicted samples.

[0214] The second feature loss is determined based on the time series information of the predicted samples;

[0215] The temperature prediction model is adjusted based on the first feature loss and the second feature loss to obtain the temperature prediction model that has completed the initial training.

[0216] This application also provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the steps of any of the temperature prediction methods in this application. This electronic device integrates any of the temperature prediction methods provided in this application, such as... Figure 6 As shown, it illustrates a structural schematic diagram of the electronic device involved in the embodiments of this application, specifically:

[0217] The electronic device may include components such as a processor 401 with one or more processing cores, a memory 402 with one or more computer-readable storage media, a power supply 403, and an input unit 404. Those skilled in the art will understand that... Figure 6 The electronic device structure shown does not constitute a limitation on the electronic device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0218] The processor 401 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 402, and by calling data stored in the memory 402, it performs various functions and processes data, thereby providing overall monitoring of the electronic device. Optionally, the processor 401 may include one or more processing cores. The processor 401 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor. Preferably, the processor 401 may integrate an application processor and a modem processor. The application processor mainly handles the operating system, user interface, and application programs, while the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 401.

[0219] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and data processing by running the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 402 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 with access to the memory 402.

[0220] The electronic device also includes a power supply 403 that supplies power to the various components. Preferably, the power supply 403 can be logically connected to the processor 401 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 403 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0221] The electronic device may also include an input unit 404, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0222] Although not shown, the electronic device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 401 in the electronic device loads the executable files corresponding to the processes of one or more applications into the memory 402 according to the following instructions, and the processor 401 runs the applications stored in the memory 402 to realize various functions, such as:

[0223] Obtain at least one set of historical operating parameters for the target inverter;

[0224] Obtain at least one set of historical heat dissipation parameters of the target inverter, where one set of historical heat dissipation parameters corresponds to the same historical moment information as one set of historical operating parameters.

[0225] The historical operating parameters, historical heat dissipation parameters, and historical time information of each group are input into the trained temperature prediction model, and the operating temperature of the target inverter is predicted by the temperature prediction model.

[0226] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0227] Therefore, embodiments of this application provide a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), a magnetic disk, or an optical disk, etc. A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in any of the temperature prediction methods provided in embodiments of this application. For example, the computer program loaded by the processor can execute the following steps:

[0228] Obtain at least one set of historical operating parameters for the target inverter;

[0229] Obtain at least one set of historical heat dissipation parameters of the target inverter, where one set of historical heat dissipation parameters corresponds to the same historical moment information as one set of historical operating parameters.

[0230] The historical operating parameters, historical heat dissipation parameters, and historical time information of each group are input into the trained temperature prediction model, and the operating temperature of the target inverter is predicted by the temperature prediction model.

[0231] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the detailed descriptions of other embodiments above, which will not be repeated here.

[0232] In practice, each of the above units or structures can be implemented as an independent entity or can be arbitrarily combined to be implemented as the same or several entities. For the specific implementation of each of the above units or structures, please refer to the previous method embodiments, which will not be repeated here.

[0233] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0234] The above provides a detailed description of a temperature prediction method and apparatus provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A temperature prediction method, characterized in that, The method includes: Obtain at least one set of historical operating parameters for the target inverter; Obtain at least one set of historical heat dissipation parameters of the target inverter, wherein one set of historical heat dissipation parameters corresponds to the same historical time information as one set of historical operating parameters; The historical operating parameters, historical heat dissipation parameters and historical time information of each group are input into the temperature prediction model to complete the training, and the operating temperature of the target inverter is predicted by the temperature prediction model. The step of inputting the historical operating parameters, historical heat dissipation parameters, and historical time information into a trained temperature prediction model, and then predicting the operating temperature of the target inverter using the temperature prediction model, includes: Obtain the periodic operation parameters and periodic heat dissipation parameters corresponding to each periodic moment in the historical period; the historical period refers to the periodicity characteristics that the temperature prediction model has mastered after completing training. The historical operating parameters and heat dissipation parameters of each group are compared with the periodic operating parameters and heat dissipation parameters of each periodic time to determine the target periodic operating parameters and the target periodic heat dissipation parameters, and the target periodic time of the target periodic operating parameters and the target periodic heat dissipation parameters is obtained. Determine the time difference between the time to be predicted and the most recent historical time in the historical time information; Based on the time difference, determine the associated time in the historical period corresponding to the target periodic time; Determine the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time. Based on the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated time, the operating temperature of the target inverter at the time to be predicted is predicted.

2. The temperature prediction method according to claim 1, characterized in that, Before obtaining at least one set of historical operating parameters of the target inverter, the method further includes: Obtain multiple sets of inverter operating parameters and multiple sets of inverter heat dissipation parameters for training the temperature prediction model; Time encoding is performed on inverter operating parameters and inverter heat dissipation parameters that have the same time information to obtain time encoding information of the inverter operating parameters and time encoding information of the inverter heat dissipation parameters. Based on the time-coded information of the inverter operating parameters and the time-coded information of the inverter heat dissipation parameters, the inverter operating parameters and the inverter heat dissipation parameters of each group are sorted to obtain training samples including time series information. The temperature prediction model is trained based on the training samples, which include time series information.

3. The temperature prediction method according to claim 2, characterized in that, The acquisition of multiple sets of inverter operating parameters for training the temperature prediction model includes: Obtain multiple sets of initial inverter operating parameters; Determine the parameter types included in the initial inverter operating parameters for each group; Using each set of initial inverter operating parameters as the target initial inverter operating parameter set, determine whether the target initial inverter operating parameter set has any missing parameter types; If the target initial inverter operating parameter group does not have any missing parameter types, the target initial inverter operating parameter group is directly used as a set of inverter operating parameters; If there are missing target parameter types in the target initial inverter operating parameter group, fill in the missing parameters of the target parameter types to obtain a set of inverter operating parameters.

4. The temperature prediction method according to claim 3, characterized in that, The parameters used to fill in the missing target parameter type include: Determine the target timestamp for the target initial inverter operating parameter group; Determine the adjacent initial inverter operating parameter group corresponding to the adjacent timestamps adjacent to the target timestamp; Determine the parameter similarity between the adjacent initial inverter operating parameter group including the target parameter type and the target initial inverter operating parameter group, and obtain the target adjacent parameter group whose parameter similarity meets the target setting; Based on the parameters corresponding to the target parameter type in the target adjacent parameter group, determine the parameters of the target parameter type in the target initial inverter operating parameter group.

5. The temperature prediction method according to claim 2, characterized in that, The acquisition of multiple sets of inverter operating parameters for training the temperature prediction model includes: Obtain multiple sets of initial inverter operating parameters; Each parameter in each set of initial inverter operating parameters is standardized and processed to obtain multiple sets of inverter operating parameters.

6. The temperature prediction method according to claim 2, characterized in that, The acquisition of multiple sets of inverter heat dissipation parameters for training the temperature prediction model includes: Using the heat dissipation parameters of each inverter as the target inverter heat dissipation parameters, determine the current temperature of the target inverter under the corresponding target time information; Determine the apparent power of the target inverter at the target time information; The power ratio is determined based on the apparent power and the rated power of the target inverter; If the current temperature is less than the minimum endpoint of the preset temperature range, the first heat dissipation parameter is determined based on the power ratio; If the current temperature is within the temperature range, determine the second heat dissipation parameter based on the current temperature; The maximum heat dissipation parameter among the first heat dissipation parameter and the second heat dissipation parameter is determined as a set of inverter heat dissipation parameters, thereby obtaining multiple sets of inverter heat dissipation parameters.

7. The temperature prediction method according to claim 2, characterized in that, The step of training the temperature prediction model based on the training samples including time-series information includes: The training samples, including time series information, are divided into training set samples, validation set samples, and test set samples according to a set ratio. Each training sample in the training set is sequentially used as the current training sample and input into the temperature prediction model to be trained to obtain the predicted temperature at the second historical moment after the first historical moment corresponding to the current training sample. The second historical moment corresponds to at least one target training sample in the training set other than the current training sample. The loss function value is calculated based on the predicted temperature at the second historical moment and the operating temperature parameters of the inverter corresponding to the second historical moment in the target training sample. The model metrics of the temperature prediction model are iteratively updated based on the loss function value until the loss function value meets the iteration termination condition, thus obtaining the trained temperature prediction model. The trained temperature prediction model is validated and tested using training samples from the validation set and the test set, respectively, thus completing the training of the temperature prediction model.

8. The temperature prediction method according to claim 7, characterized in that, The step of training the temperature prediction model based on the training samples including time-series information includes: Based on the time series information of the training samples in the training set, the trend characteristics, seasonal characteristics, and error characteristics of temperature information under different operating parameters and different inverter heat dissipation parameters are determined. Based on the trend characteristics, the seasonal characteristics, and the error characteristics, a prediction sample including time series information is obtained; The first feature loss is determined based on the trend characteristics, seasonal characteristics, and error characteristics of the predicted sample. The second feature loss is determined based on the time series information of the predicted sample; The temperature prediction model is adjusted based on the first feature loss and the second feature loss to obtain a temperature prediction model that has completed initial training.

9. A temperature prediction device, characterized in that, The device includes: The first acquisition module is used to acquire at least one set of historical operating parameters of the target inverter; The second acquisition module is used to acquire at least one set of historical heat dissipation parameters of the target inverter, wherein the set of historical heat dissipation parameters corresponds to the same historical time information as a set of historical operating parameters. The prediction module is used to input the historical operating parameters of each group, the historical heat dissipation parameters of each group, and the historical time information into the temperature prediction model after training, and to predict the operating temperature of the target inverter through the temperature prediction model. The prediction module is specifically used to acquire the periodic operating parameters and periodic heat dissipation parameters corresponding to each periodic moment in the historical cycle. The historical cycle refers to the periodic characteristics that the temperature prediction model has acquired after training. The module compares each set of historical operating parameters and each set of historical heat dissipation parameters with the periodic operating parameters and periodic heat dissipation parameters at each periodic moment to determine the target periodic operating parameters and target periodic heat dissipation parameters, thus obtaining the target periodic moment for the target periodic operating parameters and target periodic heat dissipation parameters. It then determines the time difference between the moment to be predicted and the most recent historical moment in the historical time information. Based on the time difference, it determines the associated moment corresponding to the target periodic moment in the historical cycle. Finally, it determines the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated moment. Based on the periodic operating parameters and periodic heat dissipation parameters corresponding to the associated moment, it predicts the operating temperature of the target inverter at the moment to be predicted.

10. An electronic device, characterized in that, The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, the processor executing the computer program to implement the steps of the temperature prediction method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that is executed by a processor to implement the steps of the temperature prediction method according to any one of claims 1 to 8.

Images