Temperature prediction method and device for power conversion equipment

Abstract

The invention discloses a temperature prediction method and device for power conversion equipment, and belongs to the field of power conversion. The temperature prediction method for the power conversion equipment comprises the following steps: acquiring time sequence electrical parameters of a to-be-detected position in the power conversion equipment at a plurality of acquisition moments; converting the time sequence electrical parameter into a feature image according to a preset time sequence-to-image processing method; and based on the time sequence electrical parameters, the feature image and the environment temperature of the power conversion equipment, performing prediction through a temperature prediction model to obtain the predicted temperature of the to-be-measured position. According to the temperature prediction method of the power conversion equipment, the temperature prediction of the to-be-detected position is carried out based on the feature image obtained through conversion, the time sequence electrical parameters and the environment temperature, the incidence relation of the time sequence electrical parameters in time can be fully considered, the accuracy and authenticity of the predicted temperature obtained through prediction are improved, and the accuracy of the temperature prediction is improved. Therefore, normal operation of equipment such as power conversion equipment is effectively maintained.

Description

Temperature prediction method and device for power conversion equipment Technical Field

[0001] This application belongs to the field of power conversion, and in particular relates to a method and apparatus for predicting the temperature of power conversion equipment. Background Technology

[0002] Temperature assessment within power conversion equipment is fundamental for anomaly analysis, fault and lifespan assessment of individual units within the inverter, as well as for the inverter itself. Related technologies primarily assess the internal temperature of the inverter by establishing the physical relationship between current, voltage, and temperature, based on current and voltage operating parameters. However, the accuracy of the temperature assessment obtained using these methods is relatively poor, thus affecting the normal operation of the inverter. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a method and apparatus for predicting the temperature of power conversion equipment, which can fully consider the temporal correlation of time-series electrical parameters, improve the accuracy and authenticity of the predicted temperature, and thus effectively maintain the normal operation of power conversion equipment and other equipment.

[0004] In a first aspect, this application provides a method for predicting the temperature of a power conversion device, the method comprising:

[0005] Acquire the time-series electrical parameters of the location to be measured in the power conversion equipment at multiple acquisition times;

[0006] The time-series electrical parameters are converted into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images.

[0007] Based on the time-series electrical parameters, the feature image, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the temperature of the location to be measured; the temperature prediction model is constructed based on a multimodal fusion algorithm.

[0008] According to the temperature prediction method for power conversion equipment of this application, by performing image conversion on the collected time-series electrical parameters, the temperature prediction of the target location is made based on the converted feature image, time-series electrical parameters and ambient temperature. This method can fully consider the temporal correlation of time-series electrical parameters, improve the accuracy and authenticity of the predicted temperature, and thus effectively maintain the normal operation of power conversion equipment and other equipment.

[0009] According to one embodiment of this application, the step of converting the time-series electrical parameters into a feature image according to a preset time-to-image processing method includes:

[0010] Sliding sampling is performed on the time-series electrical parameters based on the target time window to obtain at least two sets of time-series data;

[0011] The at least two sets of time-series data are converted into images to obtain at least one frame of feature image.

[0012] According to one embodiment of this application, the step of performing image transformation on the at least two sets of time-series data to obtain at least one frame of feature image includes:

[0013] The time-series data of each group are standardized to obtain standard electrical parameters;

[0014] Convert the standard electrical parameters into polar coordinates;

[0015] Based on the polar coordinates, the at least one frame of feature image is obtained.

[0016] According to one embodiment of this application, the target time window is updated based on the following steps:

[0017] Under the condition of satisfying the iterative update, the target time window is adjusted based on the target step size;

[0018] Based on the adjusted target time window, the following steps are performed: "Converting the time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images; based on the time-series electrical parameters, the feature images, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the predicted temperature of the location to be measured."

[0019] If the accuracy metric is satisfied based on the predicted temperature and the actual temperature at the location to be measured, the iterative update stops.

[0020] If the accuracy target is not met based on the predicted temperature and the actual temperature at the location to be measured, the target time window is adjusted based on the target step size.

[0021] According to one embodiment of this application, the step of predicting the predicted temperature of the location to be measured based on the time-series electrical parameters, the feature image, and the ambient temperature of the power conversion equipment using a temperature prediction model includes:

[0022] Obtain the simulated temperature corresponding to the location to be tested, output by the physical model; the physical model is a pre-constructed temperature simulation model of the component at the location to be tested of the power conversion equipment;

[0023] Based on the time-series electrical parameters, the feature image, the ambient temperature, and the simulated temperature, the predicted temperature is obtained by using the temperature prediction model.

[0024] According to one embodiment of this application, the physical model is constructed based on the following steps:

[0025] The physical model is constructed based on the performance parameters of the location to be tested within a preset time period, the operating parameters of the location to be tested under normal working conditions, and the ambient temperature.

[0026] According to one embodiment of this application, the temperature prediction model is updated based on the following steps:

[0027] Under the condition of satisfying the iterative update, the network parameters of the temperature prediction model are updated based on the predicted temperature and the actual temperature of the location to be measured.

[0028] Secondly, this application provides a temperature prediction device for power conversion equipment, the device comprising:

[0029] The first processing module is used to acquire the time-series electrical parameters of the location to be measured in the power conversion equipment at multiple acquisition times.

[0030] The second processing module is used to convert the time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images.

[0031] The third processing module is used to predict the temperature of the location to be measured by using a temperature prediction model based on the time-series electrical parameters, the feature image, and the ambient temperature of the power conversion equipment; the temperature prediction model is constructed based on a multimodal fusion algorithm.

[0032] According to the temperature prediction device for power conversion equipment of this application, by performing image conversion on the collected time-series electrical parameters, and based on the converted feature image, time-series electrical parameters and ambient temperature, the temperature of the location to be measured is predicted. This device can fully consider the temporal correlation of the time-series electrical parameters, improve the accuracy and authenticity of the predicted temperature, and thus effectively maintain the normal operation of power conversion equipment and other equipment.

[0033] Thirdly, this application provides a power conversion device that uses a temperature prediction method based on the power conversion device described in the first aspect to predict temperature.

[0034] Fourthly, this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the temperature prediction method for power conversion equipment as described in the first aspect above.

[0035] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the temperature prediction method for power conversion equipment as described in the first aspect above.

[0036] The above-described one or more technical solutions in the embodiments of this application have at least one of the following technical effects:

[0037] By performing image conversion on the collected time-series electrical parameters, and then predicting the temperature at the target location based on the converted feature image, time-series electrical parameters, and ambient temperature, the correlation between time-series electrical parameters over time can be fully considered, improving the accuracy and realism of the predicted temperature, thereby effectively maintaining the normal operation of power conversion equipment and other devices.

[0038] Furthermore, by performing sliding sampling and image transformation on the collected time-series electrical parameters through the target time window, more time-series correlation information can be retained, the information contained in the transformed feature image can be improved, and the accuracy and authenticity of the predicted temperature can be further improved, thereby effectively maintaining the normal operation of power conversion equipment and other equipment.

[0039] Furthermore, by combining multimodal data such as time-series electrical parameters, feature images, ambient temperature, and simulated temperature output from the physical model for temperature prediction, and using simulated temperature as a basic reference value, the prediction bias of the temperature prediction model can be reduced, and the accuracy and realism of the prediction results can be improved.

[0040] Furthermore, by setting iterative update conditions, the temperature prediction model can be updated based on real-time data from the power conversion equipment during operation, provided these conditions are met. This enables online updates and dynamic adjustments to the temperature prediction model, allowing it to adapt to different environmental changes and achieve higher temperature estimation results. This improves the prediction accuracy and precision of the temperature prediction model and enhances its environmental adaptability.

[0041] 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

[0042] 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:

[0043] Figure 1 is one of the flowcharts illustrating the temperature prediction method for power conversion equipment provided in this application embodiment;

[0044] Figure 2 is a second schematic flowchart of the temperature prediction method for power conversion equipment provided in the embodiments of this application;

[0045] Figure 3 is one of the schematic diagrams showing the results of the temperature prediction method for power conversion equipment provided in the embodiments of this application;

[0046] Figure 4 is a second schematic diagram of the results of the temperature prediction method for power conversion equipment provided in the embodiments of this application;

[0047] Figure 5 is a third flowchart illustrating the temperature prediction method for power conversion equipment provided in this application embodiment;

[0048] Figure 6 is a schematic diagram of the temperature prediction device for power conversion equipment provided in an embodiment of this application;

[0049] Figure 7 is a schematic diagram of the structure of the electronic device provided in an embodiment of this application. Detailed Implementation

[0050] 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.

[0051] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0052] The following description, in conjunction with the accompanying drawings, details the temperature prediction method, temperature prediction device, power conversion equipment, and readable storage medium provided in this application through specific embodiments and application scenarios.

[0053] The temperature prediction method for power conversion equipment can be applied to the terminal, specifically executed by the hardware or software in the terminal.

[0054] The temperature prediction method for power conversion equipment provided in this application embodiment can be implemented by an energy storage system, an electronic device electrically connected to the energy storage system, or a functional module or entity in the energy storage system / electronic device that can implement the temperature prediction method for the power conversion equipment. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablets, computers, cameras, and wearable devices. The following uses an energy storage system as the implementing entity to illustrate the temperature prediction method for power conversion equipment provided in this application embodiment.

[0055] The temperature prediction method for this power conversion equipment can be used in fields such as energy storage or photovoltaic power generation.

[0056] As shown in Figure 1, the temperature prediction method for the power conversion equipment includes steps 110, 120 and 130.

[0057] Among them, power conversion equipment may include: DC-AC converters and DC-DC converters, etc.

[0058] Step 110: Obtain the time-series electrical parameters of the location to be measured in the power conversion equipment at multiple acquisition times;

[0059] In this step, the power conversion equipment may include multiple units.

[0060] Taking an inverter as an example, it can include multiple units such as an insulated-gate bipolar transistor (IGBT), diodes, DC modules, output modules, and modules composed of capacitors and inductors.

[0061] The location to be measured can be any location in the power conversion equipment. In some embodiments, the location to be measured can also be the unit corresponding to the location to be measured.

[0062] Timing electrical parameters are electrical parameters that have timing characteristics.

[0063] Electrical parameters may include, but are not limited to: voltage data, current data, and power data.

[0064] In actual execution, time-series electrical parameters can be collected based on certain time intervals, such as time intervals at the second or minute level.

[0065] Taking a time interval of 1 minute as an example, 60 time-series electrical parameters can be collected within one hour.

[0066] As can be understood, as shown in Figure 3, the collected time-series electrical parameters are time-series data, where the horizontal axis represents the acquisition time and the vertical axis represents the specific values. The time-series electrical parameters change dynamically with time, and the time-series electrical parameters at different acquisition times have a certain correlation in time sequence.

[0067] Step 120: Convert the time series electrical parameters into feature images according to the preset time series to image processing method;

[0068] In this step, the time-to-image processing method is a method for converting time-series data into feature images.

[0069] In some embodiments, the time-to-image processing method may include: correlation method, Gram angle field method or other arbitrary time-to-image processing algorithm.

[0070] In actual implementation, the corresponding time-to-image processing method can be selected according to the requirements, and this application does not limit it.

[0071] Feature images are used to characterize the temporal correlation of time-series electrical parameters.

[0072] Figure 4 illustrates a feature image used to characterize the correlation between time-series electrical parameters at any two acquisition times. The feature image includes multiple image pixel blocks to represent the degree of correlation between any two acquisition times. The darker the color of the image pixel block, the higher the degree of correlation.

[0073] In some embodiments, step 120 may further include:

[0074] High-frequency filtering is applied to the timing electrical parameters;

[0075] Based on a pre-defined time-to-image processing method, the time-series electrical parameters after high-frequency filtering are converted into feature images.

[0076] In this embodiment, by performing signal decomposition on the time-series electrical parameters to filter the high-frequency components, abnormal data can be effectively removed, and data with a small impact on temperature can also be removed, while retaining data with a relatively high impact on temperature. This ensures that the main energy in the retained time-series electrical parameters is the data that causes temperature changes.

[0077] In some embodiments, the signal can be decomposed using its own time-frequency distribution information, such as using the Ensemble Empirical Mode Decomposition (EEMD) method to decompose the signal, which is applicable to any environment and the influence of factors such as the DC side current and voltage being greatly affected by radiation, thereby improving the accuracy and applicability of the signal decomposition.

[0078] After filtering, a time-to-image processing method is used to sample the retained data to perform image transformation based on the sampled time-series data, thereby obtaining a feature image.

[0079] According to the temperature prediction method for power conversion equipment provided in the embodiments of this application, by performing signal decomposition and high-frequency filtering on the collected time-series electrical parameters before image conversion, errors caused by data loss, interference, and acquisition accuracy can be reduced, thereby improving data quality. In addition, data with a relatively high degree of influence on temperature can be retained, thereby improving the accuracy and precision of the converted feature image and increasing computational efficiency.

[0080] Step 130: Based on time-series electrical parameters, feature images, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the temperature of the location to be measured. The temperature prediction model is constructed based on a multi-modal fusion algorithm.

[0081] In this step, the time-series electrical parameters and characteristic images used for temperature prediction should be located in the same or similar time periods.

[0082] The ambient temperature can be the average of the actual ambient temperatures within the current time period, or the actual ambient temperature at a specific moment within the current time period, such as the ambient temperature at the start of the current time period. In some embodiments, the collected time-series electrical parameters and ambient temperature data can also be time-aligned to ensure data consistency over time.

[0083] The temperature prediction model is a pre-trained model, which may include, but is not limited to, neural network models, deep learning models, or other models that can predict temperature. This application does not limit the model.

[0084] For example, the acquired time-series electrical parameters, feature images, and ambient temperature of the power conversion equipment can be input into a pre-trained model, which will then output the predicted temperature of the location to be measured.

[0085] During the training process, the sample electrical parameters corresponding to the sample unit / sample position, the sample feature image corresponding to the sample electrical parameters, and the sample ambient temperature can be used as samples. The sample temperature value corresponding to the sample unit / sample position can be used as the sample label to construct training samples and train the temperature prediction model.

[0086] The sample temperature value is the actual temperature at which the sample unit / sample location operates under the sample electrical parameters.

[0087] In some embodiments, step 130 may further include:

[0088] The target fusion features are obtained by fusing time-series electrical parameters, feature images, and ambient temperature.

[0089] The target fusion features are input into the temperature prediction model to obtain the predicted temperature output by the model; where,

[0090] The temperature prediction model is trained using sample fusion features as samples and sample temperature values ​​corresponding to the sample fusion features as sample labels.

[0091] In this embodiment, the target fusion feature is the feature sequence obtained by fusing data from different dimensions or categories, such as the feature obtained by fusing multi-source heterogeneous data.

[0092] In the process of building a temperature prediction model, the input dimension of the temperature prediction model can include the feature dimension after multimodal feature fusion, and the output dimension can include the temperature corresponding to the module to be measured.

[0093] The sample fusion feature is the feature obtained by fusing data such as sample electrical parameters and sample feature images corresponding to sample electrical parameters based on sample units / sample locations. The acquisition and processing methods of sample electrical parameters and sample feature images are similar to the processing methods of time-series electrical parameters and their corresponding feature images, and will not be described in detail here.

[0094] For sample units of the same category, multiple sample electrical parameters, sample feature images, and corresponding sample temperatures can be obtained. The corresponding values ​​are used as training samples to obtain a training set, which is then used to train the temperature prediction model.

[0095] In some embodiments, all training samples can be divided into a training set and a validation set in a ratio of 7:3 or 8:2. After training, the temperature prediction model is validated using the validation set until the temperature prediction model meets the required accuracy or precision.

[0096] During training, the loss between the predicted temperature and the actual temperature can be calculated through forward propagation of the network, and then the network parameters can be updated through backward propagation. By updating the network parameters through multiple rounds of iteration, the loss value between the predicted temperature and the actual temperature can be minimized or the preset index can be met.

[0097] Understandably, during the training process, individual training can be conducted based on each unit or different location included in the power conversion equipment to predict the predicted temperature corresponding to the corresponding unit or location.

[0098] In the R & D process, the inventor found that in the related art, the temperature of the inverter device unit is mainly evaluated by constructing the physical relationship between current, voltage and temperature, based on the current, voltage and other operating parameters at present. However, this method ignores the temporal characteristics of data such as current, voltage and temperature, and does not consider the temporal correlation of data during the evaluation of the temperature of the device unit, thus failing to make full use of the current and voltage data information, which affects the accuracy of the evaluated temperature.

[0099] In the present application, by converting the collected sequential electrical parameters into images of sequential data, the temporal correlation of the sequential electrical parameters can be extracted, and the hidden feature information and the spatio-temporal coupling characteristics between data can be fully explored, so as to present the data from different perspectives; on this basis, combining the feature images and sequential data that can characterize the temporal correlation of the sequential electrical parameters for temperature prediction can take into account more details, ensure the comprehensiveness of features, provide effective data support for the prediction effect, and thus improve the accuracy and authenticity of the predicted temperature obtained, making it closer to the true value.

[0100] According to the temperature prediction method of the power conversion device provided by the embodiment of the present application, by converting the collected sequential electrical parameters into images and predicting the temperature at the position to be measured based on the converted image features, sequential electrical parameters and ambient temperature, the temporal correlation of the sequential electrical parameters can be fully considered, and the accuracy and authenticity of the predicted temperature obtained can be improved, thereby effectively maintaining the normal operation of devices such as power conversion devices.

[0101] The implementation manner of step 120 will be described below.

[0102] In some embodiments, step 120 may include:

[0103] Performing sliding sampling on the sequential electrical parameters based on the target time window to obtain at least two sets of sequential data;

[0104] Converting at least two sets of sequential data into images to obtain at least one frame of feature images.

[0105] In this embodiment, the target time window is a sliding time window for data sampling.

[0106] The value of the target time window can be customized.

[0107] The value of the target time window should not exceed the time period covered by the collected sequential electrical parameters. [[ID=二十九]]

[0108] In the actual execution process, the target time window α can be set as: 0 < α < len; where len is the time period covered by the collected sequential electrical parameters.

[0109] Based on the target time window, the time series electrical parameters are sampled to obtain a set of time series data whose acquisition time is within the target time window. By performing image transformation on any two sets of time series data from the multiple sets of sampled time series data, the feature images corresponding to the two sets of time series data can be obtained.

[0110] When there are two or more sets of time-series data, multiple frames of feature images can be obtained.

[0111] Continue collecting time-series electrical parameters at 1-minute intervals. For example, when collecting time-series electrical parameters within 1 hour, the target time window can be set to 10 minutes, 15 minutes, or other values.

[0112] Taking a target time window of 15 minutes and a sliding step of 1 minute as an example, data sampling of time-series electrical parameters based on the target time window can yield approximately 45 sets of time-series data. Converting the sampled time-series data into images can yield approximately 45 frames of feature images.

[0113] Figure 2 illustrates a flowchart of converting time-series electrical parameters into a feature image. In actual execution, for example, the target time window x can be set to 4 minutes and the sliding step size to 1 minute. This allows sampling of multiple sets of time-series data, such as x1 to x4, x2 to x5, etc. By performing a cosine operation on any two sets of time-series data, such as using the time-series data corresponding to x1 to x4 as the horizontal axis and the time-series data corresponding to x2 to x5 as the vertical axis, a feature image containing multiple pixel blocks can be obtained. In this feature image, a pixel block is used to represent the correlation between the time-series electrical parameters at two acquisition times. For example, the first pixel block in the upper left corner of the feature image represents the correlation between the time-series electrical parameters at acquisition times x1 and x2.

[0114] In this embodiment, by setting a target time window to sample the time-series electrical parameters, the sampling accuracy and comprehensiveness can be improved, thereby improving the accuracy of the feature image obtained based on the sampled data, and further improving the accuracy of the temperature prediction results.

[0115] According to the temperature prediction method for power conversion equipment provided in this application embodiment, by performing sliding sampling and image transformation on the collected time-series electrical parameters through a target time window, more time-series correlation information can be retained, the information contained in the transformed feature image can be improved, and the accuracy and authenticity of the predicted temperature can be further improved, thereby effectively maintaining the normal operation of power conversion equipment and other equipment.

[0116] In some embodiments, performing image transformation on at least two sets of time-series data to obtain at least one frame of feature image may include:

[0117] Standardize the time series data of each group to obtain standard electrical parameters;

[0118] Convert standard electrical parameters to polar coordinates;

[0119] At least one feature image is obtained based on polar coordinates.

[0120] In this embodiment, standardization of each set of time series data may include normalizing the time series electrical parameters within each target time window, such as normalizing the time series electrical parameters to the [-1,1] interval, which facilitates subsequent calculations and improves calculation accuracy.

[0121] Let the time-series electrical parameters within the target time window be X = x1, x2, ..., x N For example, the standard electrical parameters obtained after standardization can be expressed as follows: Where N is the number of time-series electrical parameters within the target time window, N is a positive integer, and i is the i-th time-series electrical parameter.

[0122] The standardized electrical parameters are converted to polar coordinates using the following formula:

[0123]

[0124] Where, φ i These are the polar coordinates corresponding to the i-th standard electrical parameter; For the i-th standard electrical parameter; t i For x i The corresponding acquisition time; N is the number of time-series electrical parameters within the target time window.

[0125] The feature image can be obtained by obtaining the cosine of the angle between any two standard electrical parameters based on polar coordinates and expressing the cosine as a pixel value.

[0126] For example, a custom inner product can be represented in the following form:

[0127]

[0128] Where n is the nth standard electrical parameter; φ n Let n be the nth polar coordinate; 1 ≤ n ≤ N, and n is an integer.

[0129] According to the temperature prediction method for power conversion equipment provided in the embodiments of this application, time series data is converted into images using the Gram angle field method, which can capture the local and global features of time series data and is applicable to various types of time series data. This allows the position and gray value of each pixel to reflect the relationship between time series data points, thus better reflecting the structure, periodicity, and trend of the time series.

[0130] In some embodiments, step 130 may further include:

[0131] Obtain the simulation temperature corresponding to the test location output by the physical model; the physical model is a pre-constructed temperature simulation model of the components at the test location of the power conversion equipment;

[0132] Based on time-series electrical parameters, feature images, ambient temperature, and simulated temperature, a temperature prediction model is used to predict the temperature.

[0133] In this embodiment, the physical model is a pre-built simulation model to simulate the actual operation of the location to be tested.

[0134] The simulated temperature can include the average temperature of the simulated location within a preset time period.

[0135] The duration of the preset time period can be user-defined, and the preset time period should be a time period that is close to the target time window.

[0136] For example, the end time of the preset time period can be the last acquisition time in the target time window.

[0137] Given the known material properties of structural components or parts, the simulation temperature obtained through system simulation can provide a rough temperature range.

[0138] The ambient temperature can be the average of the actual ambient temperature within a preset time period, or the actual ambient temperature at a certain moment within the preset time period, such as the ambient temperature at the start of the preset time period.

[0139] In actual implementation, when making temperature predictions, in addition to considering time-series electrical parameters and feature images, data from multiple dimensions such as simulated temperature and ambient temperature can be integrated to further improve the accuracy and realism of the prediction results.

[0140] When the input feature dimension of the temperature prediction model also includes simulated temperature, the training samples during the training process can also include: the sample simulated temperature output by the physical model corresponding to the sample unit / sample position, etc. By training the temperature prediction model by comprehensively considering the sample electrical parameters, sample feature images, sample ambient temperature, sample simulated temperature output by the physical model corresponding to the sample unit, sample feature images and corresponding sample temperatures, the model accuracy and precision can be further improved.

[0141] According to the temperature prediction method for power conversion equipment provided in the embodiments of this application, by training a temperature prediction model to perform temperature prediction based on target fusion features, it is possible to learn the correlation between data of different dimensions, make full use of each feature data, and improve the accuracy of prediction results through multi-source heterogeneous feature data fusion and learning, so that the temperature estimation of each unit of the power conversion equipment is more accurate, stable and reliable; and the operation is simple and convenient and easy to implement.

[0142] In some embodiments, feature fusion of time-series electrical parameters, feature images, and ambient temperature to obtain target fused features may further include:

[0143] The target fused features are obtained by fusing time-series electrical parameters, feature images, ambient temperature, and simulated temperature.

[0144] In some embodiments, feature fusion of time-series electrical parameters, feature images, ambient temperature, and simulated temperature to obtain target fused features may further include:

[0145] Feature extraction was performed on time-series electrical parameters, feature images, ambient temperature, and simulated temperature. The extracted features were then fused to obtain the target fused features.

[0146] As shown in Figure 5, taking time-series electrical parameters including current and voltage as an example, after collecting current and voltage data at multiple acquisition times, the acquired data is sampled using a target time window, and the sampling results are converted from time-series data to images to obtain multi-frame feature images corresponding to current and multi-frame feature images corresponding to voltage.

[0147] The simulated temperature of the location to be measured is obtained by simulating the physical model corresponding to the location to be measured.

[0148] Then, feature extraction and feature fusion are performed on time-series data such as current, voltage, and ambient temperature, multi-frame feature images corresponding to current, multi-frame feature images corresponding to voltage, and multi-modal data such as simulated temperature to obtain target fused features. The target fused features are then input into the temperature prediction model, which predicts and outputs the predicted temperature T at the location to be measured.

[0149] In some embodiments, the physical model can be constructed based on the following steps:

[0150] A physical model is constructed based on the performance parameters of the location to be tested within a preset time period, the operating parameters of the location to be tested under normal working conditions, and the ambient temperature.

[0151] In this embodiment, the performance parameters are parameters such as the material properties of the location to be tested, such as internal resistance.

[0152] Normal operating conditions can be characterized by stable operation.

[0153] Operating parameters may include voltage, current, and power.

[0154] Operating parameters can be acquired through a signal acquisition device.

[0155] In some embodiments, when the operating parameter is power, the power can be calculated by multiplying the voltage and the current.

[0156] Taking IGBT units and diodes as examples, the physical model can be constructed using the following formula:

[0157] T J_IGBT(avg) =R th(j-r)_IGBT *P IGBT +T r

[0158] T J_diode(avg) =R th(j-r)_diode *P diode +T r

[0159] Among them, T J_IGBT(avg) The simulation temperature for the IGBT; R th(j-r)_IGBT P is the internal resistance of the IGBT; IBBT T represents the power of the IGBT under normal operating conditions. r Let be the ambient temperature at time r; r is the start time of the preset time period, and j is the end time of the preset time period.

[0160] T J_diode(avg) R represents the simulation temperature of the diode. th(j-r)_diode P is the internal resistance of the diode; diode T represents the power of the diode under normal operating conditions. r Let be the ambient temperature at time r.

[0161] It should be noted that the physical models for different types of units are constructed in a similar way and can be set according to actual needs, which will not be elaborated here.

[0162] According to the temperature prediction method for power conversion equipment provided in the embodiments of this application, temperature prediction is performed by combining multimodal data such as time-series electrical parameters, feature images, ambient temperature, and simulated temperature output by the physical model. The simulated temperature provides a basic reference value, which can reduce the prediction deviation of the temperature prediction model and improve the accuracy and authenticity of the prediction results.

[0163] The following section explains the temperature prediction model and the update method for the target time window.

[0164] In some embodiments, the temperature prediction model may be updated based on the following steps:

[0165] Under the condition of iterative update, the network parameters of the temperature prediction model are updated based on the predicted temperature and the actual temperature at the location to be measured.

[0166] In this embodiment, the iterative update conditions can be user-defined and include: time, data volume, prediction accuracy, or other triggering conditions.

[0167] For example, if the time since the last update of the temperature prediction model is not less than the target time, it is determined that the iterative update condition is met.

[0168] The target duration can be set to one week, half a month, or one month, etc., and this application does not limit it.

[0169] For example, when the accumulated data such as the timing electrical parameters and ambient temperature at the measured location reach the trigger threshold, it is determined that the iterative update conditions are met.

[0170] The trigger threshold can be customized based on actual needs, such as setting it to 1000 or 2000 records.

[0171] For example, if the prediction index corresponding to the temperature prediction model changes or the prediction accuracy decreases, but the accuracy index increases, it can be determined that the conditions for iterative update are met, and the model can be iterated and optimized online.

[0172] Of course, in some embodiments, multiple triggering conditions can be combined to determine whether the iterative update conditions are met, and this application does not limit this.

[0173] The predicted temperature is the temperature at the location to be measured under certain time-series electrical parameters and ambient temperature, as predicted by the temperature prediction model in practical applications. The actual temperature is the true temperature at the location to be measured under the same time-series electrical parameters and ambient temperature. Based on the difference between the predicted and actual temperatures, the current accuracy of the temperature prediction model can be calculated. Based on the current accuracy and accuracy indicators, the network parameters in the temperature prediction model, such as the weight values ​​of each node, can be adjusted to update and optimize the temperature prediction model.

[0174] Taking an inverter as an example, during actual operation, data such as DC-side current, DC-side voltage, and ambient temperature can be monitored in real time. When the iterative update condition is triggered, data can be collected by directionally collecting the actual temperature of each unit in the inverter and inputting the actual temperature into the temperature prediction model. The network parameters of the temperature prediction model, such as the weight values ​​of nodes, are updated with the prediction index as the target until a certain number of iterations or the prediction index is reached, thereby ending the current iterative update.

[0175] The iterative update method of the model is similar to that of the model training method, and this application does not limit it here.

[0176] According to the temperature prediction method for power conversion equipment provided in the embodiments of this application, by setting iterative update conditions, the temperature prediction model is updated based on real-time data of the power conversion equipment during operation when the iterative update conditions are met. This enables online updating and dynamic adjustment of the temperature prediction model, thereby adapting to different environmental changes to achieve higher temperature estimation results, improving the prediction accuracy and precision of the temperature prediction model, and enhancing the environmental adaptability of the temperature prediction model.

[0177] In some embodiments, the target time window may be updated based on the following steps:

[0178] Under the condition of satisfying the iterative update, adjust the target time window based on the target step size;

[0179] Based on the adjusted target time window, the following steps are performed: "Converting time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images; based on the time-series electrical parameters, feature images, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the predicted temperature of the location to be measured."

[0180] Once the accuracy metric is met based on the predicted temperature and the actual temperature at the location to be measured, the iterative update stops.

[0181] If the accuracy target is not met based on the predicted temperature and the actual temperature at the location to be measured, the target time window is adjusted based on the target step size.

[0182] In this embodiment, when setting a target time window for the first time, such as during the training phase of the temperature prediction model, a value can be randomly selected from the time period covered by the collected time-series electrical parameters as the initial value of the target time window, so as to sample electrical parameters based on the initial value and thus complete the training of the temperature prediction model.

[0183] In subsequent applications, if the iterative update conditions are met, the target time window can be updated. This can be achieved using a step-by-step increment method, as detailed below:

[0184] The target time window is adjusted using the formula: α = α + δ, where δ is the target step size for each adjustment.

[0185] It is understood that the target step size is a small value, and the target step size can remain consistent in each adjustment process, or it can change dynamically. This application does not impose any restrictions on this.

[0186] Based on the adjusted target time window, the time-series electrical parameters are sampled, and steps 120 to 130 are repeated to obtain the predicted temperature output by the temperature prediction model. Based on the difference between the predicted temperature and the actual temperature, it is determined whether the current prediction result meets the accuracy index. If it does, the iteration can be stopped, and the adjusted target time window is determined as the target time window in the actual application process before the next iteration update.

[0187] If the accuracy target is not met, the target time window can be adjusted based on the target step size, and steps 110 to 130 can be repeated until the accuracy target is met. Then, the target time window after the last adjustment is determined as the target time window in the actual application process before the next iteration update.

[0188] In some embodiments, during the process of triggering the iterative update condition, the target time window and the network parameters of the temperature prediction model can be adjusted simultaneously to achieve the update of the model parameters of the temperature prediction model and the update of the target time window.

[0189] According to the temperature prediction method for power conversion equipment provided in the embodiments of this application, by setting iterative update conditions to dynamically adjust the target time window when the iterative update conditions are met, the prediction accuracy and precision of the temperature prediction model can be improved, and the environmental adaptability of the temperature prediction model can be enhanced.

[0190] The temperature prediction method for power conversion equipment provided in this application can be executed by a temperature prediction device for power conversion equipment. This application uses the example of a temperature prediction device for power conversion equipment executing the temperature prediction method to illustrate the temperature prediction device for power conversion equipment provided in this application.

[0191] This application also provides a temperature prediction device for power conversion equipment.

[0192] As shown in Figure 6, the temperature prediction device of the power conversion equipment includes: a first processing module 610, a second processing module 620 and a third processing module 630.

[0193] The first processing module 610 is used to acquire the time-series electrical parameters of the location under test in the power conversion equipment at multiple acquisition times.

[0194] The second processing module 620 is used to convert time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images.

[0195] The third processing module 630 is used to predict the temperature of the location to be measured by using a temperature prediction model based on time-series electrical parameters, feature images, and the ambient temperature of the power conversion equipment. The temperature prediction model is constructed based on a multi-modal fusion algorithm.

[0196] According to the temperature prediction device for power conversion equipment provided in the embodiments of this application, by performing image conversion on the collected time-series electrical parameters, and predicting the temperature of the location to be measured based on the converted image features, time-series electrical parameters and ambient temperature, the device can fully consider the temporal correlation of the time-series electrical parameters, improve the accuracy and authenticity of the predicted temperature, and thus effectively maintain the normal operation of power conversion equipment and other equipment.

[0197] In some embodiments, the second processing module 620 may also be used for:

[0198] Sliding sampling of time-series electrical parameters is performed based on the target time window to obtain at least two sets of time-series data;

[0199] At least two sets of time-series data are transformed into images to obtain at least one frame of feature image.

[0200] In some embodiments, the second processing module 620 may also be used for:

[0201] Standardize the time series data of each group to obtain standard electrical parameters;

[0202] Convert standard electrical parameters to polar coordinates;

[0203] At least one feature image is obtained based on polar coordinates.

[0204] In some embodiments, the device may further include a fourth processing module for:

[0205] Under the condition of satisfying the iterative update, adjust the target time window based on the target step size;

[0206] Based on the adjusted target time window, the following steps are performed: "Converting time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images; based on the time-series electrical parameters, feature images, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the predicted temperature of the location to be measured."

[0207] Once the accuracy metric is met based on the predicted temperature and the actual temperature at the location to be measured, the iterative update stops.

[0208] If the accuracy target is not met based on the predicted temperature and the actual temperature at the location to be measured, the target time window is adjusted based on the target step size.

[0209] In some embodiments, the third processing module 630 can also be used for:

[0210] Obtain the simulation temperature corresponding to the test location output by the physical model; the physical model is a pre-constructed temperature simulation model of the components at the test location of the power conversion equipment;

[0211] Based on time-series electrical parameters, feature images, ambient temperature, and simulated temperature, a temperature prediction model is used to predict the temperature.

[0212] In some embodiments, the device may further include a fifth processing module for:

[0213] A physical model is constructed based on the performance parameters of the location to be tested within a preset time period, the operating parameters of the location to be tested under normal working conditions, and the ambient temperature.

[0214] In some embodiments, the device may further include a sixth processing module for:

[0215] Under the condition of iterative update, the network parameters of the temperature prediction model are updated based on the predicted temperature and the actual temperature at the location to be measured.

[0216] The temperature prediction device of the power conversion equipment in this application embodiment can be an electronic device or a component of an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, handheld computer, in-vehicle electronic 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.

[0217] The temperature prediction device for the power conversion equipment in this embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this embodiment does not specifically limit its use.

[0218] The temperature prediction device for power conversion equipment provided in this application embodiment can realize the various processes implemented in the method embodiments of Figures 1 to 5. To avoid repetition, it will not be described again here.

[0219] This application also provides a power conversion device.

[0220] In some embodiments, the power conversion device may include an inverter.

[0221] The power conversion device performs temperature prediction based on the temperature prediction method for power conversion devices described in any of the above embodiments.

[0222] According to the power conversion equipment provided in the embodiments of this application, the collected time-series electrical parameters are sampled and image converted through a target time window. Based on the converted feature image and time-series electrical parameters, the temperature of the location to be measured is predicted. This can fully consider the temporal correlation of the time-series electrical parameters, improve the accuracy and authenticity of the predicted temperature, and thus effectively maintain the normal operation of the power conversion equipment and other equipment.

[0223] In some embodiments, as shown in FIG7, this application embodiment also provides an electronic device 700, including a processor 701, a memory 702, and a computer program stored in the memory 702 and executable on the processor 701. When the program is executed by the processor 701, it implements the various processes of the temperature prediction method embodiment of the power conversion device described above and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0224] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0225] 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 temperature prediction method embodiment for the power conversion device described above and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0226] The processor is the processor in the electronic 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.

[0227] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the temperature prediction method for the power conversion device described above.

[0228] The processor is the processor in the electronic 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.

[0229] 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 temperature prediction method embodiment of the power conversion device described above, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0230] 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.

[0231] 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.

[0232] 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 prior art, 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.

[0233] 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.

[0234] 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.

[0235] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for predicting the temperature of a power conversion device, characterized in that, include: Acquire the time-series electrical parameters of the location to be measured in the power conversion equipment at multiple acquisition times; The time-series electrical parameters are converted into feature images according to a preset time-to-image processing method; The time-to-image processing method is a method for converting time-series data into feature images; based on the time-series electrical parameters, the feature images, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the predicted temperature of the location to be measured; the temperature prediction model is constructed based on a multimodal fusion algorithm.

2. The temperature prediction method for power conversion equipment according to claim 1, characterized in that, The step of converting the time-series electrical parameters into feature images according to a preset time-to-image processing method includes: performing sliding sampling on the time-series electrical parameters based on a target time window to obtain at least two sets of time-series data; and performing image conversion on the at least two sets of time-series data to obtain at least one frame of feature image.

3. The temperature prediction method for power conversion equipment according to claim 2, characterized in that, The step of performing image transformation on the at least two sets of time-series data to obtain at least one frame of feature image includes: standardizing each set of time-series data to obtain standard electrical parameters; converting the standard electrical parameters into polar coordinates; and obtaining the at least one frame of feature image based on the polar coordinates.

4. The temperature prediction method for power conversion equipment according to claim 2, characterized in that, The target time window is updated based on the following steps: When the iterative update conditions are met, the target time window is adjusted based on the target step size; based on the adjusted target time window, the steps of "converting the time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images; based on the time-series electrical parameters, the feature images, and the ambient temperature of the power conversion equipment, a temperature prediction model is used to predict the predicted temperature of the location to be measured" are executed; if the accuracy index is met based on the predicted temperature and the actual temperature of the location to be measured, the iterative update is stopped; if the accuracy index is not met based on the predicted temperature and the actual temperature of the location to be measured, the target time window is adjusted again based on the target step size.

5. The temperature prediction method for power conversion equipment according to any one of claims 1-4, characterized in that, The step of predicting the temperature of the test location based on the time-series electrical parameters, the feature image, and the ambient temperature of the power conversion equipment using a temperature prediction model includes: obtaining the simulation temperature corresponding to the test location output by a physical model; the physical model is a pre-constructed temperature simulation model of the components at the test location of the power conversion equipment; and predicting the temperature based on the time-series electrical parameters, the feature image, the ambient temperature, and the simulation temperature using the temperature prediction model.

6. The temperature prediction method for power conversion equipment according to claim 5, characterized in that, The physical model is constructed based on the following steps: the physical model is constructed based on the performance parameters of the location to be tested within a preset time period, the operating parameters of the location to be tested under normal operating conditions, and the ambient temperature.

7. The temperature prediction method for power conversion equipment according to any one of claims 1-4, characterized in that, The temperature prediction model is updated based on the following steps: under the condition of satisfying the iterative update, the network parameters of the temperature prediction model are updated based on the predicted temperature and the actual temperature of the location to be measured.

8. A temperature prediction device for power conversion equipment, characterized in that, include: The first processing module is used to acquire the time-series electrical parameters of the location to be measured in the power conversion equipment at multiple acquisition times. The second processing module is used to convert the time-series electrical parameters into feature images according to a preset time-to-image processing method; the time-to-image processing method is a method for converting time-series data into feature images; the third processing module is used to predict the predicted temperature of the location to be measured by using a temperature prediction model based on the time-series electrical parameters, the feature images and the ambient temperature of the power conversion equipment; the temperature prediction model is constructed based on a multimodal fusion algorithm.

9. The temperature prediction device for power conversion equipment according to claim 8, characterized in that, The second processing module is further configured to: perform sliding sampling on the time-series electrical parameters based on the target time window to obtain at least two sets of time-series data; and perform image conversion on the at least two sets of time-series data to obtain at least one frame of feature image.

10. The temperature prediction device for power conversion equipment according to claim 9, characterized in that, The second processing module is further configured to: standardize the time-series data of each group to obtain standard electrical parameters; convert the standard electrical parameters into polar coordinates; and obtain the at least one frame feature image based on the polar coordinates.

11. A power conversion device, characterized in that, The power conversion equipment uses a temperature prediction method based on any one of claims 1-7 to predict the temperature.

12. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the temperature prediction method for power conversion equipment as described in any one of claims 1-7.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the temperature prediction method for the power conversion equipment as described in any one of claims 1-7.

Images