A converter key point temperature prediction method and system based on improved Kalman filtering and related devices

Abstract

This invention provides a method, system, and related apparatus for predicting the critical temperature of a converter based on an improved Kalman filter, comprising the following steps: Step 1, acquiring historical temperature data of the converter under test within a set time period to form a temperature data matrix; Step 2, performing temperature difference calculation on the temperature data matrix to obtain a temperature difference matrix; Step 3, compressing the temperature difference matrix to obtain a compressed matrix; Step 4, predicting the temperature of the converter based on the obtained compressed matrix and combined with the improved Kalman filter to obtain a predicted temperature value for the converter. This invention can better adapt to the temperature variation characteristics of the converter under different operating conditions, providing more reliable temperature prediction results, thereby providing a more accurate basis for the temperature monitoring and control of the converter.

Description

Technical Field

[0001] This invention belongs to the field of temperature monitoring and control technology for wind turbine generators, specifically relating to a method, system, and related devices for predicting the temperature of key points of a converter based on an improved Kalman filter. Background Technology

[0002] A wind turbine is a device that generates electricity using wind energy. The converter is a crucial component that converts the alternating current (AC) generated by the wind turbine into direct current (DC) or AC power matched to the power grid. During converter operation, the temperature at key internal locations is a critical factor affecting its performance and lifespan, requiring effective monitoring and control. However, due to the unique environment, operating conditions, and cost limitations of data acquisition and analysis inherent in wind turbines, existing temperature data measurement and control methods suffer from problems such as a limited number of measurement points, low measurement accuracy, and overly simplistic control logic. There is an urgent need for a temperature prediction method based on mathematical models to unlock the deeper value of temperature data and improve temperature control capabilities without increasing hardware investment.

[0003] The existing mainstream converter temperature prediction methods mainly include temperature prediction methods based on least squares, temperature prediction methods based on neural networks, and temperature prediction methods based on Kalman filtering.

[0004] The least squares-based temperature prediction method is a linear regression-based temperature prediction method. It uses temperature measurement data and converter input / output data to solve a system of linear equations using the least squares method to obtain the temperature prediction values ​​for key points of the converter. Its advantage is its computational simplicity; its disadvantages are that it requires high-quality temperature measurement data and cannot handle nonlinear system models and measurement models.

[0005] The temperature prediction method based on neural networks is a nonlinear fitting-based approach. It uses a neural network to replace the system and measurement models, and through backpropagation, updates the neural network weights based on temperature measurement data and converter input / output data to obtain the temperature prediction values ​​for key points of the converter. Its advantage is its ability to handle nonlinear system and measurement models; its disadvantages include computational complexity and a susceptibility to overfitting.

[0006] The Kalman filter-based temperature prediction method is a minimum variance estimation-based temperature prediction method. It uses a Kalman filter to combine the system model and the measurement model, and iteratively updates the state estimate and Kalman gain matrix based on temperature measurement data and converter input / output data to obtain the temperature prediction value of key points of the converter. Its advantages include the ability to handle system noise and measurement noise; its disadvantages include high requirements for the parameters of the system model and measurement model, and its inability to adapt to changes in the weight of the measurement data.

[0007] In summary, it can be seen that the current mainstream prediction methods all have shortcomings to varying degrees, and further optimization through relevant algorithms is needed to improve their adaptability. Summary of the Invention

[0008] The purpose of this invention is to propose a method, system, and related device for predicting the temperature of key points of a converter based on an improved Kalman filter. By utilizing temperature measurement data and combining the system model and measurement model of the converter, the improved Kalman filter algorithm is used to estimate the temperature of key points of the converter, thereby enabling temperature monitoring and control of the converter and improving the operating efficiency and safety of wind turbine units.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a method for predicting the critical temperature of a converter based on an improved Kalman filter, comprising the following steps: Step 1: Obtain historical temperature data of the converter under test within a set time period to form a temperature data matrix; Step 2: Perform temperature difference calculation on the temperature data matrix to obtain the temperature difference matrix; Step 3: Compress the temperature difference matrix to obtain a compressed matrix; Step 4: Based on the obtained compression matrix, the temperature of the converter is estimated using an improved Kalman filter to obtain the converter temperature prediction value.

[0010] Preferably, the temperature of the converter is predicted based on the obtained compression matrix and an improved Kalman filter. Specifically, the method is as follows: Based on the obtained compression matrix, the grid-side temperature prediction model and the machine-side temperature prediction model are constructed by combining the Karman filter, and the initial state estimate, initial estimation error covariance matrix, initial system noise covariance matrix and initial measurement noise covariance matrix are determined for the grid-side temperature prediction model and the machine-side temperature prediction model, respectively. By introducing a forgetting factor and a sliding window, and combining the initial system noise covariance matrix and the initial measurement noise covariance matrix, the system noise covariance matrix and the measurement noise covariance matrix at the current time k are calculated respectively. Based on the initial state estimate, the initial estimation error covariance matrix, the system noise covariance matrix, and the measurement noise covariance matrix, state estimation is performed by combining the measurement data within the current k-time, and the state estimate and estimation error covariance matrix at the current k-time corresponding to the network side and the machine side are obtained respectively. The state estimates at time k on the network side and the machine side are fused using dynamic model weights to obtain the final state estimate.

[0011] Preferably, the parameters of the grid-side temperature prediction model and the machine-side temperature prediction model are updated using online parameter identification.

[0012] Preferably, the dynamic model weights are obtained using the following method: Calculate the innovation sequence based on the state estimate at time k and the measurement data within time k. Estimate the covariance of innovation sequences; The model weights are dynamically adjusted based on the innovation sequence and the innovation sequence covariance.

[0013] Preferably, the determined initial state estimate and the initial estimation error covariance matrix are:

[0014]

[0015] in, This is the initial state estimate; This is the initial estimate of the error covariance matrix; Represents a diagonal matrix; All are initial diagonal parameters; These are the average values ​​of the temperature difference between the monitored IGBTs on the grid side (numbers 1 to 6) and the ambient temperature. Determined initial system noise covariance matrix and measurement noise covariance matrix:

[0016]

[0017] in, This is the initial system noise covariance matrix; This is the initial measurement noise covariance matrix.

[0018] Preferably, the system noise covariance matrix and the measurement noise covariance matrix at time k are calculated using the following formulas, specifically:

[0019]

[0020] in, Forgetting factor, To adjust the sliding window size; Kalman gain; For innovation sequence; These are measured values; State estimate; To measure the noise covariance matrix; The system noise covariance matrix; This is the measurement matrix.

[0021] Preferably, the state estimates at time k on the network side and the machine side are fused using the following formula to obtain the final state estimate:

[0022] in, The number of models in the vector; For the first The weights of each model; For the first State estimation of a model; It is an identity matrix.

[0023] Secondly, the present invention provides a converter critical point temperature prediction system based on improved Kalman filtering, comprising: The temperature matrix acquisition unit is used to acquire historical temperature data of the converter under test within a set time period and form a temperature data matrix. The temperature difference matrix acquisition unit is used to perform temperature difference calculation on the temperature data matrix to obtain the temperature difference matrix; The compression matrix acquisition unit is used to compress the temperature difference matrix to obtain the compression matrix; The temperature prediction unit is used to predict the temperature of the converter based on the obtained compression matrix and the improved Kalman filter, so as to obtain the converter temperature prediction value.

[0024] Thirdly, the present invention provides an electronic device including a processor and a memory, wherein the memory stores computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the method described thereon.

[0025] Fourthly, the present invention provides a computer program product, the computer program product including computer-executable instructions, which, when executed, implement any of the methods described herein.

[0026] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a method for predicting the temperature of key points in a converter based on an improved Kalman filter. By optimizing the algorithm and deeply mining the value of temperature data, it effectively improves the accuracy and adaptability of temperature prediction for key points in the converter. This method enhances the algorithm's responsiveness to different operating conditions by optimizing the determination of initial state estimates and combining a multi-model fusion strategy, making the temperature prediction results more reliable. An online parameter identification mechanism is introduced to achieve dynamic updates of model parameters, thereby improving the accuracy and stability of long-term predictions. Simultaneously, by using a forgetting factor and sliding window technology to adjust the covariance matrix of system noise and measurement noise in real time, it overcomes the limitations of traditional Kalman filters in terms of sensitivity to model parameters and inability to adapt to changes in data weights, significantly improving the robustness of the algorithm in complex environments. Furthermore, the use of dynamic weight fusion of grid-side and turbine-side temperature predictions further optimizes the comprehensiveness of state estimation. These improvements enable the model to more accurately reflect the temperature change characteristics of the converter in actual operation, providing a scientific basis for temperature control strategies. This not only helps prevent equipment overheating failures and extend the service life of the converter, but also improves the overall operating efficiency and safety of wind turbine units, possessing significant engineering application value.

[0027] In summary, the present invention can better adapt to the temperature variation characteristics of converters under different operating conditions, and provide more reliable temperature prediction results, thereby providing a more accurate basis for temperature monitoring and control of converters. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating an embodiment of the present invention. Detailed Implementation

[0029] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0030] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0031] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0032] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0033] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0034] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0035] Example 1 This embodiment presents a method for predicting the critical temperature of a converter based on an improved Kalman filter, comprising the following steps: Step 1: Obtain historical temperature data of the converter under test within a set time period to form a temperature data matrix; Step 2: Perform temperature difference calculation on the temperature data matrix to obtain the temperature difference matrix; Step 3: Compress the temperature difference matrix to obtain a compressed matrix; Step 4: Based on the obtained compression matrix, the temperature of the converter is estimated using an improved Kalman filter to obtain the converter temperature prediction value.

[0036] In this embodiment, the temperature of the converter is predicted based on the obtained compression matrix and an improved Kalman filter. Specifically, the method is as follows: Based on the obtained compression matrix, the grid-side temperature prediction model and the machine-side temperature prediction model are constructed by combining the Karman filter, and the initial state estimate, initial estimation error covariance matrix, initial system noise covariance matrix and initial measurement noise covariance matrix are determined for the grid-side temperature prediction model and the machine-side temperature prediction model, respectively. By introducing a forgetting factor and a sliding window, and combining the initial system noise covariance matrix and the initial measurement noise covariance matrix, the system noise covariance matrix and the measurement noise covariance matrix at the current time k are calculated respectively. Based on the initial state estimate, the initial estimation error covariance matrix, the system noise covariance matrix, and the measurement noise covariance matrix, state estimation is performed by combining the measurement data within the current k-time, and the state estimate and estimation error covariance matrix at the current k-time corresponding to the network side and the machine side are obtained respectively. The state estimates at time k on the network side and the machine side are fused using dynamic model weights to obtain the final state estimate.

[0037] Example 2 This embodiment presents a method for predicting the critical temperature of a converter based on an improved Kalman filter, comprising the following steps: Step 1: Collect temperature monitoring data from 14 temperature measurement points of the converter over a period of time (two PT100 sensors monitoring the temperature of the left and right cabinets of the frequency converter respectively; 12 NTCs monitoring the temperature of 6 IGBT modules on the grid side and the machine side, with 2 NTC measurement points corresponding to each IGBT module), and form a temperature data matrix T;

[0038] The temperature data matrix T is a 15-column n-row matrix, where each row represents the temperature measurement value of each sensor at a certain moment, the first column is the time data, and the other columns correspond to sensors at different locations. For time data, This is the environmental monitoring temperature data for the inverter network side (left cabinet); For the environmental monitoring temperature data of the inverter unit side (right cabinet); and This refers to the temperature data from two monitoring points of IGBT 1 on the network side (left cabinet). and This refers to the temperature data from two monitoring points of IGBT 2 on the network side (left cabinet), and so on. and Temperature data for two monitoring points of IGBT 6 on the machine side (right cabinet); The value of n is not less than 43200, and the acquisition frequency is 1 / 60Hz, meaning that the temperature data matrix T is at least one set of converter monitoring temperatures for 30 days.

[0039] Step 2: Perform temperature difference calculation on the temperature data in matrix T obtained in Step 1 to obtain the temperature difference matrix. ;

[0040] matrix It is an n x 7 matrix, with the first column containing time data. The value represents the difference between the relatively high temperature data of the monitoring point of IGBT No. 1 on the grid side (left cabinet) and the ambient temperature of the left cabinet. The value represents the difference between the relatively high temperature data of the monitoring point of IGBT 2 on the grid side (left cabinet) and the ambient temperature of the left cabinet, and so on. The difference between the relatively high temperature data of the monitoring point of IGBT 6 on the machine side (right cabinet) and the ambient temperature of the left cabinet is calculated as follows:

[0041]

[0042]

[0043]

[0044]

[0045]

[0046]

[0047] In the above formula, m corresponds to a certain monitoring time, taking IGBT No. 1 on the network side (left cabinet) as an example. For example, The larger of the two monitored temperatures at that moment. Let the ambient temperature on the network side (left cabinet) at that moment be... This represents the maximum temperature difference at that moment; Similarly, taking IGBT #5 on the machine side (left cabinet) as an example, i.e. For example, The larger of the two monitored temperatures at that moment. If the ambient temperature is at the machine side (right cabinet) at that time, then This represents the maximum temperature difference at that moment; The rest of the calculation methods are the same and will not be repeated here; Step 3: Since the temperature data is gradually changing, the temperature difference matrix obtained in Step 2 is processed. The compression process is performed to obtain a compression matrix. ,have:

[0048]

[0049]

[0050]

[0051] As can be seen from the above calculation method, for Within 10 minutes Number( For the net side, The average value of the temperature difference between the IGBT monitoring temperature (on the machine side) and the ambient temperature; for The average temperature difference between the IGBT monitoring unit 1 on the internal network side (left cabinet) and the ambient temperature at 10-minute intervals. for The average temperature difference between the internal network side (left cabinet) IGBT 2 and the ambient temperature at each 10-minute interval, and so on; Step 4, based on the compression matrix obtained in Step 3 Determine the initial state estimate and the initial estimation error covariance matrix Then we have:

[0052]

[0053] In the above formula, The initial state estimate is obtained by selecting the matrix from step 3. The first row of temperature data is arranged in order; this value is relatively accurate. Represents a diagonal matrix. These are the initial diagonal parameters; In this embodiment, middle For a unified value, denoted as , For a unified value, denoted as ,but:

[0054] Step 5: Set the initial system noise covariance matrix With measurement noise covariance matrix Given that the temperature sensor value is 0.2℃, then:

[0055]

[0056] Step 6: Perform state estimation based on the initial state estimate obtained in Step 4. Based on the above system model, the state estimate at time k is calculated. and the estimated error covariance matrix The details are as follows:

[0057]

[0058] In the above formula, The state estimate for the current time k is represented by the estimated temperature of the converter key point. Let k be the prior state estimate at the current time k, representing the temperature predicted based on information from the previous time. The Kalman gain matrix at time k determines the degree of influence of the measured value on the estimated value; The actual temperature measurement value at the current time k; The measurement matrix maps the state space to the measurement space. Let be the estimation error covariance matrix at the current time k, representing the uncertainty of the estimated value; Let k be the prior estimate error covariance matrix at the current time k; It is the identity matrix; in, and It consists of the predicted state estimate and the prediction error covariance matrix. It is the Kalman gain matrix, calculated using the following formula:

[0059]

[0060]

[0061] in: This is the state transition matrix, describing how the temperature state transitions from time k-1 to time k; For the control input matrix, describe how the control inputs affect the temperature state; For control input at time k-1, such as the operation of the cooling system; Let be the system process noise covariance matrix at time k, representing the uncertainty of the system model; Let k be the measurement noise covariance matrix, representing the uncertainty of the measurement process; Through the above formula It can be estimated from the initial state value First, derive... And then through the formula Derivation By repeating the above formula, we can derive... and ; and The derivation method is as follows:

[0062]

[0063] In the above formula, Forgetting factor, To adjust the sliding window size; The Kalman gain at time k; The innovation sequence at time k; The value measured at time i; The state estimate at time i; Step 7: Based on Step 6, a multi-model fusion strategy is introduced to better adapt to the temperature variation characteristics of the converter under different operating conditions, namely:

[0064] In the above formula, In this embodiment, the number of models in the vector is [number]. The number is 2, corresponding to the network side and the machine side respectively; For the first The weights of each model; For the first State estimation of each model; weights The system is dynamically adjusted based on the prediction error of each model, as follows:

[0065] In the above formula, For the first The prediction error of each model; This is the corresponding covariance matrix; Step 8: To improve the accuracy of the system model, online parameter identification of the temperature prediction model is introduced based on the above steps, as follows:

[0066] In the above formula, Let k be the model parameter vector at time k; Let be the regression matrix at time k; The learning rate enables the system model to adaptively adjust over time, improving long-term prediction accuracy.

[0067] In this embodiment, The initial setting is empty. Approximation using the Jacobian matrix Set to 0.1; Step 9: Dynamically adjust the weights of the measurement data based on the steps described above. First, calculate the innovation sequence:

[0068] For innovation sequences, it represents the difference between measured and predicted values; Next, estimate the covariance of the innovation sequence:

[0069] The covariance of the innovation series is used to assess the reliability of the innovation series. Finally, the iteration continues until the predetermined number of iterations is reached or the deviation between the predicted and actual values ​​is reached, i.e., the covariance of the innovation sequence satisfies the convergence condition, and the final state estimate is output. This is the estimated temperature of the converter.

[0070] Step 10, Temperature control system integration; In this embodiment, a temperature control system is set up for predicted heat dissipation control, mainly composed of a temperature controller, a temperature setting module and a heat sink, which is used to perform real-time temperature control based on the output of the temperature prediction model.

[0071] Example 3 This embodiment provides a converter critical point temperature prediction system based on improved Kalman filtering, including: The temperature matrix acquisition unit is used to acquire historical temperature data of the converter under test within a set time period and form a temperature data matrix. The temperature difference matrix acquisition unit is used to perform temperature difference calculation on the temperature data matrix to obtain the temperature difference matrix; The compression matrix acquisition unit is used to compress the temperature difference matrix to obtain the compression matrix; The temperature prediction unit is used to predict the temperature of the converter based on the obtained compression matrix and the improved Kalman filter, so as to obtain the converter temperature prediction value.

[0072] Example 4 This embodiment also provides a computing device. The computing device includes a bus, a processor, a memory, and a communication interface. The processor, memory, and communication interface communicate with each other via the bus. The computing device can be a server or a terminal device. It should be understood that this application does not limit the number of processors and memory in the computing device.

[0073] A bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, a bus can include a path for transmitting information between various components of a computing device (e.g., memory, processor, communication interfaces).

[0074] The processor may include any one or more of the following: central processing unit (CPU), graphics processing unit (GPU), tensor processing unit (TPU), application specific integrated circuit (ASIC), field-programmable gate array (FPGA), microprocessor (MP), or digital signal processor (DSP).

[0075] Memory can include volatile memory, such as random access memory (RAM). Processors can also include non-volatile memory. volatile memory, such as read-only memory (ROM). ROM (memory only), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0076] The memory stores executable program code, which the processor executes to implement the functions of the aforementioned units, thereby achieving, for example, the method described in Embodiment 1. That is, the memory may store instructions for the methods and functions relating to the computing device in any of the above embodiments.

[0077] The communication interface uses transceiver modules such as, but not limited to, network interface cards and transceivers to enable communication between computing devices and other devices or communication networks.

[0078] Example 5 This embodiment also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, cause the processor to perform the methods and functions of the computing device involved in any of the above embodiments.

[0079] Generally, the various embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software, which can be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of this disclosure are shown and described as block diagrams, flowcharts, or represented using some other illustration, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as, as non-limiting examples, in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0080] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for predicting the temperature of key points in a converter based on an improved Kalman filter, characterized in that, Includes the following steps: Step 1: Obtain historical temperature data of the converter under test within a set time period to form a temperature data matrix; Step 2: Perform temperature difference calculation on the temperature data matrix to obtain the temperature difference matrix; Step 3: Compress the temperature difference matrix to obtain a compressed matrix; Step 4: Based on the obtained compression matrix, the temperature of the converter is estimated using an improved Kalman filter to obtain the converter temperature prediction value.

2. The method for predicting the critical temperature of a converter based on an improved Kalman filter according to claim 1, characterized in that, Based on the obtained compression matrix, the temperature of the converter is predicted using an improved Kalman filter. The specific method is as follows: Based on the obtained compression matrix, the grid-side temperature prediction model and the machine-side temperature prediction model are constructed by combining the Karman filter, and the initial state estimate, initial estimation error covariance matrix, initial system noise covariance matrix and initial measurement noise covariance matrix are determined for the grid-side temperature prediction model and the machine-side temperature prediction model, respectively. By introducing a forgetting factor and a sliding window, and combining the initial system noise covariance matrix and the initial measurement noise covariance matrix, the system noise covariance matrix and the measurement noise covariance matrix at the current time k are calculated respectively. Based on the initial state estimate, the initial estimation error covariance matrix, the system noise covariance matrix, and the measurement noise covariance matrix, state estimation is performed by combining the measurement data within the current k-time, and the state estimate and estimation error covariance matrix at the current k-time corresponding to the network side and the machine side are obtained respectively. The state estimates at time k on the network side and the machine side are fused using dynamic model weights to obtain the final state estimate.

3. The method for predicting the temperature of key points in a converter based on an improved Kalman filter according to claim 2, characterized in that, The parameters of the grid-side temperature prediction model and the machine-side temperature prediction model were updated using online parameter identification.

4. The method for predicting the temperature of key points of a converter based on an improved Kalman filter according to claim 2, characterized in that, The specific method for obtaining dynamic model weights is as follows: Calculate the innovation sequence based on the state estimate at time k and the measurement data within time k. Estimate the covariance of innovation sequences; The model weights are dynamically adjusted based on the innovation sequence and the innovation sequence covariance.

5. The method for predicting the critical temperature of a converter based on an improved Kalman filter according to claim 2, characterized in that, Determined initial state estimates and initial estimation error covariance matrix: in, This is the initial state estimate; This is the initial estimate of the error covariance matrix; Represents a diagonal matrix; All are initial diagonal parameters; These are the average values ​​of the temperature difference between the monitored IGBTs on the grid side (numbers 1 to 6) and the ambient temperature. Determined initial system noise covariance matrix and measurement noise covariance matrix: in, This is the initial system noise covariance matrix; This is the initial measurement noise covariance matrix.

6. The method for predicting the critical temperature of a converter based on an improved Kalman filter according to claim 2, characterized in that, The system noise covariance matrix and measurement noise covariance matrix at time k are calculated using the following formulas. The specific method is as follows: in, Forgetting factor, To adjust the sliding window size; Kalman gain; For innovation sequence; These are measured values; State estimate; To measure the noise covariance matrix; The system noise covariance matrix; This is the measurement matrix.

7. The method for predicting the critical temperature of a converter based on an improved Kalman filter according to claim 2, characterized in that, The final state estimate is obtained by fusing the state estimates from the network side and the machine side at time k using the following formula: in, The number of models in the vector; For the first The weights of each model; For the first State estimation of a model; It is an identity matrix.

8. A converter key point temperature prediction system based on improved Kalman filtering, characterized in that, include: The temperature matrix acquisition unit is used to acquire historical temperature data of the converter under test within a set time period and form a temperature data matrix. The temperature difference matrix acquisition unit is used to perform temperature difference calculation on the temperature data matrix to obtain the temperature difference matrix; The compression matrix acquisition unit is used to compress the temperature difference matrix to obtain the compression matrix; The temperature prediction unit is used to predict the temperature of the converter based on the obtained compression matrix and the improved Kalman filter, so as to obtain the converter temperature prediction value.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer instructions that, when executed by the processor, cause the electronic device to perform the method of any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes computer-executable instructions that, when executed, implement the method of any one of claims 1 to 7.

Images