A power semiconductor fault detection method and system based on multi-modal data

Abstract

The application relates to the field of power electronics, and relates to a power semiconductor fault detection method and system based on multi-modal data, which comprises the following steps: receiving a power semiconductor fault detection instruction, confirming a fault detection system based on the fault detection instruction, obtaining a historical parameter group set based on a historical parameter group storage unit, obtaining a normal parameter group set and a fault parameter group set based on a normalized historical parameter group set, obtaining a detection parameter group set based on a preset detection period, a preset detection frequency, a power semiconductor and a detection parameter group acquisition unit, obtaining a normalized detection parameter group set by using a normalization method and the detection parameter group set, transmitting the normalized detection parameter group set to a state evaluation unit, and evaluating the state of the power semiconductor based on the normal parameter group set, the fault parameter group set and the normalized detection parameter group set. The application can improve the efficiency and accuracy of power semiconductor fault detection.

Description

Technical Field

[0001] This invention relates to the field of power electronics technology, and in particular to a power semiconductor fault detection method and system based on multimodal data. Background Technology

[0002] Power semiconductor devices play a crucial role in modern power electronic systems and are widely used in various fields. The reliability and stability of power semiconductor devices directly affect the operating efficiency and safety of the entire power system. Therefore, timely and accurate detection of fault conditions in power semiconductor devices is of great significance for preventing equipment damage, reducing downtime, improving system reliability, and lowering maintenance costs.

[0003] Currently, traditional fault detection methods are mostly used to detect faults in power semiconductors. These methods mainly rely on monitoring single-type parameters, such as current and voltage. When faced with complex operating conditions and hidden faults, single-parameter monitoring often has limitations, failing to comprehensively reflect the operating status of power semiconductor devices and easily leading to misjudgments or missed diagnoses.

[0004] While traditional fault detection methods can detect faults to some extent, they typically require manual intervention for data collection and analysis, making them inefficient and prone to errors. As the complexity of power electronic systems continues to increase, traditional fault detection methods are struggling to meet the demands for high precision. Therefore, the efficiency and accuracy of power semiconductor fault detection need to be improved. Summary of the Invention

[0005] This invention provides a power semiconductor fault detection method and a computer-readable storage medium based on multimodal data, the main purpose of which is to improve the efficiency and accuracy of power semiconductor fault detection.

[0006] To achieve the above objectives, the present invention provides a power semiconductor fault detection method based on multimodal data, comprising:

[0007] The system receives a fault detection command from a power semiconductor and identifies a fault detection system based on the fault detection command. The fault detection system includes: a historical parameter group storage unit, a detection parameter group acquisition unit, and a status evaluation unit.

[0008] The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values.

[0009] Obtain a normalized historical parameter set based on the historical parameter set and a pre-built normalization method;

[0010] The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values.

[0011] Based on a preset detection period, a preset number of detections, a power semiconductor, and the detection parameter set acquisition unit, a detection parameter set is acquired. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values.

[0012] The normalization method and the detection parameter set are used to obtain a normalized detection parameter set, which is then transmitted to the state evaluation unit. When the state evaluation unit successfully receives the normalized detection parameter set, it evaluates the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the state is a normal state or a fault state.

[0013] Optionally, obtaining the historical parameter set based on the historical parameter set storage unit includes:

[0014] The initial historical parameter set is obtained based on the historical parameter set storage unit, and the initial historical parameter set is stored in the state evaluation unit to obtain the stored historical parameter set, wherein the initial historical parameter set includes multiple initial historical parameter sets.

[0015] Sort the initial historical parameter sets in the initial historical parameter set set and the stored historical parameter sets in the stored historical parameter set set according to the time sequence from front to back to obtain the initial sequence and the stored sequence.

[0016] An analysis node set is obtained using the initial sequence and the stored sequence. The analysis node set includes multiple analysis nodes, and each analysis node consists of an initial historical parameter set and a stored historical parameter set. The position of the initial historical parameter set in the initial sequence is the same as the position of the stored historical parameter set in the stored sequence. The following operations are performed on each analysis node in the analysis node set:

[0017] The initial verification value and the stored verification value are obtained by using the initial historical parameter group and the stored historical parameter group corresponding to the analysis node. The analysis node is updated using the initial verification value and the stored verification value to obtain the updated node. The updated nodes are then summarized to obtain the updated node set.

[0018] Determine whether the updated node set meets the pre-built verification conditions, wherein the verification conditions are as follows:

[0019]

[0020] Among them, a p ' represents the stored checksum corresponding to the p-th update node in the update node set, b p This represents the initial historical parameter set corresponding to the p-th update node in the update node set, and the initial historical parameter set is as follows: b p =(b p1 ,b p2 ,…,b pM ), where b p1 ,b p2 These represent the first and second historical parameter values ​​in the initial historical parameter set, respectively, where M represents the total number of historical parameter values ​​in the initial historical parameter set. p ' represents the stored historical parameter group corresponding to the p-th analysis node in the update node set, n represents the total number of update nodes in the update node set, and a p This represents the initial checksum corresponding to the p-th analysis node in the update node set;

[0021] If the verification condition is met, the stored historical parameter set is used as the historical parameter set; otherwise, the process returns to the step of storing the initial historical parameter set in the state evaluation unit until the verification condition is met.

[0022] Optionally, obtaining the initial verification value and the stored verification value using the initial historical parameter group and the stored historical parameter group corresponding to the analysis node includes:

[0023] The initial historical parameter set and the stored historical parameter set are converted using pre-built format conversion methods to obtain the initial historical sequence and the stored historical sequence.

[0024] The initial verification value is obtained based on the initial historical sequence and the pre-constructed hash function;

[0025] The storage verification value is obtained based on the stored historical sequence and the hash function.

[0026] Optionally, obtaining the normalized historical parameter set based on the historical parameter set and a pre-constructed normalization method includes:

[0027] A set of nodes to be processed is obtained based on a set of historical parameters. Nodes to be processed are then extracted sequentially from this set, and normalized nodes are obtained using the normalization method described below.

[0028]

[0029] Among them, X d This represents the value of the d-th parameter to be processed in the node to be processed. This indicates that the d-th parameter in the node to be processed has a negative value, max(|X d|) represents the maximum absolute value of all parameters to be processed in the node to be processed, min(|X) d |) represents the minimum absolute value of all parameters to be processed in the node to be processed, max() represents taking the maximum value, min() represents taking the minimum value, and X0 represents the normalized node;

[0030] After confirming that each node to be processed has a corresponding normalized node, the normalized nodes are aggregated to obtain a normalized node set. The normalized node set is then used to update the historical parameter set to obtain a normalized historical parameter set.

[0031] Optionally, obtaining the set of nodes to be processed based on the historical parameter set includes:

[0032] The parameter type set is obtained based on the historical parameter set, wherein the parameter type set includes multiple parameter types and the number of parameter types is M;

[0033] Perform the following operation on each parameter type in the parameter type set:

[0034] By utilizing parameter types, multiple parameter values ​​to be processed can be retrieved from the historical parameter set.

[0035] By aggregating multiple parameter values, the node to be processed is obtained;

[0036] After confirming that each parameter type has obtained a corresponding node to be processed, the nodes to be processed are summarized to obtain a set of nodes to be processed.

[0037] Optionally, obtaining the normal parameter set and the fault parameter set based on the normalized historical parameter set includes:

[0038] Using preset partitioning identifiers, the normalized historical parameter groups in the normalized historical parameter group set are identified as normal parameter groups or fault parameter groups.

[0039] The normal parameter group and the fault parameter group are summarized separately to obtain the normal parameter group set and the fault parameter group set.

[0040] Optionally, the step of acquiring the detection parameter set based on a preset detection time period, a preset number of detections, power semiconductor, and the detection parameter set acquisition unit includes:

[0041] Multiple detection times are calculated based on the detection period and the number of detections, using the following formula:

[0042]

[0043] Among them, t θ Let t represent the θ-th detection time among multiple detection times. minThe start time of the detection period is indicated by Δt, the detection period is indicated by Δt, and the number of detections is indicated by R (R is an integer greater than 1).

[0044] The detection parameter set is obtained based on multiple detection times, power semiconductors, and the detection parameter set acquisition unit.

[0045] Optionally, the step of evaluating the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set includes:

[0046] For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations:

[0047] The minimum normal evaluation value is obtained based on the normalized detection parameter set, the normal parameter set, and the pre-constructed evaluation value calculation formula, wherein the evaluation value calculation formula is as follows:

[0048]

[0049] Among them, l min y represents the minimum normal evaluation value among all normal parameter groups in the normalized detection parameter set and the normal parameter set. k y′ represents the k-th normalized detection parameter value in the normalized detection parameter set. k This represents the k-th normal parameter value in the normal parameter group; M represents the total number of normal parameter values ​​in the normal parameter group and the total number of normalized detection parameter values ​​in the normalized detection parameter group; and g represents the total number of normal parameter groups in the normal parameter group set.

[0050] The minimum fault assessment value is obtained based on the normalized detection parameter set, the fault parameter set, and the evaluation value calculation formula.

[0051] Compare the minimum normal assessment value with the minimum fault assessment value;

[0052] If the minimum normal assessment value is greater than the minimum fault assessment value, the normalized test parameter group is identified as the fault parameter group; otherwise, the normalized test parameter group is identified as the normal parameter group.

[0053] By summarizing the fault parameter group and the normal parameter group, the minimum evaluation value set is obtained;

[0054] The state of the power semiconductor is evaluated based on the minimum set of evaluation values.

[0055] Optionally, evaluating the state of the power semiconductor based on the minimum set of evaluation values ​​includes:

[0056] The number of normal parameter groups and the number of faulty parameter groups in the minimum evaluation value set are counted separately to obtain the number of normal parameters and the number of faulty parameters.

[0057] For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations:

[0058] Calculate the Euclidean distance of each normal parameter group in the normalized detection parameter group and normal parameter group set to obtain the evaluation normal distance subset. After each normalized detection parameter group in the normalized detection parameter group set has obtained the corresponding evaluation normal distance subset, summarize the evaluation normal distance subset to obtain the evaluation normal distance set. The evaluation normal distance subset includes multiple evaluation normal distances.

[0059] Calculate the mean of the normal distances in the assessment set to obtain the normal assessment mean;

[0060] Determine whether the number of normal operations, the number of faults, and the average value of normal operations meet the pre-built evaluation criteria, wherein the evaluation criteria are as follows:

[0061]

[0062] Where q1 represents the number of normal operations, q2 represents the number of faults, α represents the preset evaluation threshold, and p 1l p1 represents the l-th normal distance in the normal distance set, p2 represents the normal mean value, and β is a preset coefficient.

[0063] If the evaluation conditions are met, the power semiconductor is confirmed to be in a normal state; otherwise, the power semiconductor is confirmed to be in a fault state.

[0064] To achieve the above objectives, the present invention also provides a power semiconductor fault detection system based on multimodal data, comprising:

[0065] The historical parameter set acquisition module is used to receive fault detection instructions from power semiconductors and confirm the fault detection system based on the fault detection instructions. The fault detection system includes: a historical parameter set storage unit, a detection parameter set acquisition unit, and a status evaluation unit.

[0066] The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values.

[0067] The historical parameter set processing module is used to obtain a normalized historical parameter set based on the historical parameter set and a pre-built normalization method.

[0068] The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values.

[0069] The detection parameter set acquisition module is used to acquire a detection parameter set based on a preset detection period, a preset number of detections, power semiconductors, and the detection parameter set acquisition unit. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values.

[0070] The status assessment module is used to obtain a normalized detection parameter set using the normalization method and the detection parameter set, and transmit the normalized detection parameter set to the status assessment unit. When the status assessment unit successfully receives the normalized detection parameter set, it assesses the status of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the status is a normal status or a fault status.

[0071] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0072] A memory that stores at least one instruction; and a processor that executes the instructions stored in the memory to implement the power semiconductor fault detection method based on multimodal data described above.

[0073] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one instruction, which is executed by a processor in an electronic device to implement the power semiconductor fault detection method based on multimodal data described above.

[0074] To address the problems described in the background section, this invention receives a fault detection command from a power semiconductor and establishes a fault detection system based on the command. The fault detection system includes a historical parameter set storage unit, a detection parameter set acquisition unit, and a status evaluation unit. The historical parameter set storage unit acquires a set of historical parameter sets, each containing multiple historical parameter sets and M types of historical parameter values. This invention verifies whether anomalies have occurred in the initial historical parameter set after storage by checking conditions, thus obtaining an accurate historical parameter set and improving the accuracy of subsequent evaluations using the historical parameter set as a reference. To ensure the accuracy of power semiconductor status, this invention obtains a normalized historical parameter set based on a historical parameter set and a pre-constructed normalization method. Specifically, this invention selects between normal and fault parameter sets based on the normalized historical parameter set by determining whether the historical parameter values ​​in the node to be processed simultaneously contain both positive and negative values. The normal parameter set includes multiple normal parameter groups, each containing M types of normal parameter values. Similarly, the fault parameter set includes multiple fault parameter groups, each containing M types of fault parameter values. This invention divides the normalized historical parameter set into normal and fault parameter sets using a partitioning identifier. The data set, through classification, provides a clear data foundation for subsequent evaluation of the power semiconductor's state using normalized historical parameter sets. This invention acquires a detection parameter set based on a preset detection period, a preset number of detections, the power semiconductor, and the detection parameter set acquisition unit. The detection parameter set includes multiple detection parameter groups, and each detection parameter group includes M detection parameter values. Therefore, this invention uses the calculation formula to obtain the detection time, ensuring that the detection time is evenly distributed within the preset detection period, avoiding overly concentrated or sparse acquisition of detection parameter groups, and ensuring that the acquired detection parameter sets comprehensively reflect the power semiconductor's state within the detection period. To improve the accuracy of fault detection and status assessment, this invention utilizes the normalization method and detection parameter set to obtain a normalized detection parameter set, which is then transmitted to the status assessment unit. Upon successful receipt of the normalized detection parameter set, the status assessment unit assesses the power semiconductor's status based on the normal parameter set, fault parameter set, and the normalized detection parameter set. The status is either normal or faulty. This invention improves the accuracy and reliability of fault detection through dual verification of the ratio and distance mean and deviation between normal and fault parameter sets, achieving precise judgment of the power semiconductor's status. Therefore, this invention can improve the efficiency and accuracy of power semiconductor fault detection. Attached Figure Description

[0075] Figure 1 This is a flowchart illustrating a power semiconductor fault detection method based on multimodal data according to an embodiment of the present invention.

[0076] Figure 2 This is a functional block diagram of a power semiconductor fault detection system based on multimodal data provided in an embodiment of the present invention;

[0077] Figure 3 This is a schematic diagram of the structure of an electronic device that implements the power semiconductor fault detection method based on multimodal data, according to an embodiment of the present invention.

[0078] Explanation of reference numerals in the attached figures:

[0079] 1. Electronic device; 10. Processor; 11. Storage device; 12. Bus.

[0080] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0081] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0082] This application provides a power semiconductor fault detection method based on multimodal data. The execution entity of the power semiconductor fault detection method based on multimodal data includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the power semiconductor fault detection method based on multimodal data can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0083] Reference Figure 1 The diagram shown is a flowchart illustrating a power semiconductor fault detection method based on multimodal data according to an embodiment of the present invention. In this embodiment, the power semiconductor fault detection method based on multimodal data includes:

[0084] S1. Receive a fault detection command from a power semiconductor, and confirm a fault detection system based on the fault detection command. The fault detection system includes: a historical parameter group storage unit, a detection parameter group acquisition unit, and a status evaluation unit.

[0085] It should be understood that a power semiconductor is a semiconductor device used to control and convert electrical energy, enabling efficient transmission and conversion of electrical energy. A fault detection command is an instruction issued by personnel performing fault detection. A fault detection system refers to the software or application used to detect faults in power semiconductors. The historical parameter set storage unit is a storage module used to store historical parameter sets of the power semiconductor under normal or faulty conditions; the historical parameter set serves as a reference benchmark for subsequent fault detection and condition assessment. The detection parameter set acquisition unit is a module used to acquire detection parameter sets of the power semiconductor's current operating state; these detection parameter sets are compared with the historical parameter sets to determine whether the power semiconductor is faulty. The condition assessment unit is a module used for condition assessment; by comparing the acquired historical parameter sets and detection parameter sets, it evaluates whether the power semiconductor's operating state is normal or faulty.

[0086] S2. Obtain the historical parameter set based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values.

[0087] It should be explained that obtaining the historical parameter set based on the historical parameter set storage unit includes:

[0088] The initial historical parameter set is obtained based on the historical parameter set storage unit, and the initial historical parameter set is stored in the state evaluation unit to obtain the stored historical parameter set, wherein the initial historical parameter set includes multiple initial historical parameter sets.

[0089] Sort the initial historical parameter sets in the initial historical parameter set set and the stored historical parameter sets in the stored historical parameter set set according to the time sequence from front to back to obtain the initial sequence and the stored sequence.

[0090] An analysis node set is obtained using the initial sequence and the stored sequence. The analysis node set includes multiple analysis nodes, and each analysis node consists of an initial historical parameter set and a stored historical parameter set. The position of the initial historical parameter set in the initial sequence is the same as the position of the stored historical parameter set in the stored sequence. The following operations are performed on each analysis node in the analysis node set:

[0091] The initial verification value and the stored verification value are obtained by using the initial historical parameter group and the stored historical parameter group corresponding to the analysis node. The analysis node is updated using the initial verification value and the stored verification value to obtain the updated node. The updated nodes are then summarized to obtain the updated node set.

[0092] Determine whether the updated node set meets the pre-built verification conditions, wherein the verification conditions are as follows:

[0093]

[0094] Among them, a p ' represents the stored checksum corresponding to the p-th update node in the update node set, b p This represents the initial historical parameter set corresponding to the p-th update node in the update node set, and the initial historical parameter set is as follows: b p =(b p1 ,b p2 ,…,b pM ), where b p1 ,b p2 These represent the first and second historical parameter values ​​in the initial historical parameter set, respectively, where M represents the total number of historical parameter values ​​in the initial historical parameter set. p ' represents the stored historical parameter group corresponding to the p-th analysis node in the update node set, n represents the total number of update nodes in the update node set, and a p This represents the initial checksum corresponding to the p-th analysis node in the update node set;

[0095] If the verification condition is met, the stored historical parameter set is used as the historical parameter set; otherwise, the process returns to the step of storing the initial historical parameter set in the state evaluation unit until the verification condition is met.

[0096] It should be understood that a historical parameter set refers to a set of parameter values ​​containing M types of historical parameter values. Historical parameter values ​​refer to various parameter values ​​of power semiconductors recorded under normal operating or fault conditions, including but not limited to: collector current, gate voltage, and on-state saturation voltage drop. The initial historical parameter set refers to the set of historical parameter sets obtained from the historical parameter set storage unit, while the stored historical parameter set refers to the set of historical parameter sets stored in the condition assessment unit. The historical parameter set is the set of historical parameter sets confirmed based on the fulfillment of the test conditions and serves as a reference benchmark for subsequent fault detection and condition assessment. Optionally, before calculating the initial historical parameter set and the stored historical parameter set in the test conditions, the following may also be included:

[0097] The number of historical parameter values ​​in the initial historical parameter group and the number of historical parameter values ​​stored in the stored historical parameter group are obtained respectively to obtain the initial number and the stored number. After confirming that the initial number and the stored number are the same, the initial historical parameter vector and the stored historical parameter vector are obtained respectively using the initial historical parameter group and the stored historical parameter group. The detection conditions are constructed using the initial historical parameter vector and the stored historical parameter vector.

[0098] Understandably, before storing the initial set of historical parameters in the state evaluation unit to obtain the stored set of historical parameters, the process further includes:

[0099] Using a preset memory threshold, multiple target historical parameter sets are identified in the initial historical parameter set, and the memory values ​​corresponding to the target historical parameter sets are all less than or equal to the memory threshold, and the target historical parameter set includes one or more complete initial historical parameter sets.

[0100] Multiple target historical parameter sets are used to obtain multiple identifier historical parameter sets. The difference between the identifier historical parameter sets and the target historical parameter sets is that the identifier historical parameter sets identify preset identifier values.

[0101] Furthermore, the memory threshold can be a manually set memory value. The purpose of setting the memory threshold is to divide the initial historical parameter set so that when the stored historical parameter set is detected to be different from the initial historical parameter set, the target historical parameter set corresponding to the stored historical parameter set can be retrieved from the initial historical parameter set and retransmitted, thereby saving energy consumption during transmission. For example, if the initial historical parameter set includes 10 initial historical parameter sets, the preset memory threshold can divide the 10 initial historical parameter sets into 3 target historical parameter sets. The first target historical parameter set includes 3 initial historical parameter sets, the second target historical parameter set includes 4 initial historical parameter sets, and the third target historical parameter set includes 3 initial historical parameter sets. The first, second, and third target historical parameter sets can be identified as 1-target historical parameter set, 2-target historical parameter set, and 3-target historical parameter set, respectively, where 1, 2, and 3 are the identification values. If the test condition detects that the 2-target historical parameter set is invalid, the 2-target historical parameter set can be retransmitted to save the resources required during transmission.

[0102] For example, the testing process of the test conditions is introduced here using only the 1-target historical parameter set as an example: Assume that the 1-target historical parameter set is {20230301-(2.0A,5.0V), 20230303-(2.1A,5.0V), 20230305-(2.2A,5.1V)}, where 20230301, 20230303, and 20230305 all represent time, and the initial sequence corresponding to the 1-target historical parameter set is {(2.0A,5.0V), (2.1A,5.0V), (2.2A,5.1V)}. The set of stored historical parameters is {20230301-(2.0A,5.0V), 20230303-(2.2A,5.3V), 20230305-(2.2A,5.1V)}, and the corresponding stored sequences are {(2.0A,5.0V), (2.2A,5.3V), (2.2A,5.1V)}. Using the initial sequence and the stored sequences, three analysis nodes are obtained: the first analysis node {(2.0A,5.0V), (2.0A,5.0V)}, the second analysis node {(2.1A,5.0V), (2.2A,5.3V)}, and the third analysis node... Let's analyze the node {(2.2A, 5.1V), (2.2A, 5.1V)}. Taking the first analysis node {(2.1A, 5.0V), (2.1A, 5.0V)} as an example: using the initial historical parameter set (2.1A, 5.0V) and the stored historical parameter set (2.1A, 5.0V), we obtain the initial checksum and the stored checksum, both of which are A1. We then update the analysis node using the initial checksum A1 and the stored checksum A1, resulting in the first updated node {A1-(2.1A, 5.0V), A2-(2.1A, 5.0V)}. Similarly, we obtain the second updated node {B1-(2.1A, 5.0V)}. The first update node is {C1-(2.2A,5.1V)}, and the second update node is {C1-(2.2A,5.1V)}. Here, A1 represents that the initial check value and the stored check value are the same in the first update node, C1 represents that the initial check value and the stored check value are the same in the third update node, and B1 and B2 represent that the initial check value and the stored check value are different in the second update node. Finally, it is confirmed that the three update nodes in the example do not meet the verification conditions, indicating that the 1-target historical parameter set has data anomalies after storage, and it is necessary to return to the step of re-storing the 1-target historical parameter set until the verification conditions are met.

[0103] Further, the step of obtaining the initial verification value and the stored verification value using the initial historical parameter group and the stored historical parameter group corresponding to the analysis node includes:

[0104] The initial historical parameter set and the stored historical parameter set are converted using pre-built format conversion methods to obtain the initial historical sequence and the stored historical sequence.

[0105] The initial verification value is obtained based on the initial historical sequence and the pre-constructed hash function;

[0106] The storage verification value is obtained based on the stored historical sequence and the hash function.

[0107] Understandably, in order to obtain the initial verification value and stored verification value corresponding to the initial historical parameter group and the stored historical parameter group using the hash function, it is necessary to use a format conversion method to convert the initial historical parameter group and the stored historical parameter group into a format that the hash function can process. Optionally, the format conversion method can be to convert the initial historical parameter group and the stored historical parameter group into a string or a byte sequence. In this embodiment of the invention, the initial historical sequence and the stored historical sequence are obtained using the format conversion method. Optionally, JSON serialization is used as the format conversion method. The hash function is a function that can convert a string or byte sequence into a fixed-length hash value through a specific algorithm. If the input initial historical sequence and the stored historical sequence are the same, then the initial verification value and the stored verification value obtained using the hash function are the same; otherwise, the stored verification value is different from the initial verification value. Here, the initial verification value is the hash value obtained by calculating the initial historical sequence using the hash function, and the stored verification value is the hash value obtained by calculating the stored historical sequence using the hash function. The hash function includes, but is not limited to: the non-cryptographic hash function MD5, the cryptographic hash function SHA-256, and the cryptographic hash function SHA-3. The process of obtaining hash values ​​using hash functions and byte sequences is existing technology and will not be described in detail here. This embodiment of the invention verifies whether any parameter values ​​are abnormal after the initial historical parameter set is stored by checking the conditions, thus obtaining an accurate historical parameter set and improving the accuracy of subsequent evaluation of the power semiconductor state using the historical parameter set as a reference.

[0108] S3. Obtain the normalized historical parameter set based on the historical parameter set and the pre-constructed normalization method.

[0109] It should be explained that the process of obtaining the normalized historical parameter set based on the historical parameter set and the pre-constructed normalization method includes:

[0110] A set of nodes to be processed is obtained based on a set of historical parameters. Nodes to be processed are then extracted sequentially from this set, and normalized nodes are obtained using the normalization method described below.

[0111]

[0112] Among them, Xd This represents the value of the d-th parameter to be processed in the node to be processed. This indicates that the d-th parameter in the node to be processed has a negative value, max(|X d |) represents the maximum absolute value of all parameters to be processed in the node to be processed, min(|X) d |) represents the minimum absolute value of all parameters to be processed in the node to be processed, max() represents taking the maximum value, min() represents taking the minimum value, and X0 represents the normalized node;

[0113] After confirming that each node to be processed has a corresponding normalized node, the normalized nodes are aggregated to obtain a normalized node set. The normalized node set is then used to update the historical parameter set to obtain a normalized historical parameter set.

[0114] Furthermore, the step of obtaining the set of nodes to be processed based on the historical parameter set includes:

[0115] The parameter type set is obtained based on the historical parameter set, wherein the parameter type set includes multiple parameter types and the number of parameter types is M;

[0116] Perform the following operation on each parameter type in the parameter type set:

[0117] By utilizing parameter types, multiple parameter values ​​to be processed can be retrieved from the historical parameter set.

[0118] By aggregating multiple parameter values, the node to be processed is obtained;

[0119] After confirming that each parameter type has obtained a corresponding node to be processed, the nodes to be processed are summarized to obtain a set of nodes to be processed.

[0120] For example, if the historical parameter set is represented as {(2.1A, 5.0V), (2.0A, 5.0V), (2.2A, 5.1V)}, the corresponding parameter type set is (current parameter value, voltage parameter value). Using the current parameter value as the search condition, the node to be processed is retrieved from the historical parameter set as (2.1A, 2.0A, 2.2A). Similarly, the node to be processed corresponding to the voltage parameter value is (5.0V, 5.0V, 5.1V), and the set of nodes to be processed is {(2.1A, 2.0A, 2.2A), (5.0V, 5.0V, 5.1V)}. This embodiment of the invention processes the historical parameter set through a normalization method, which can unify parameter values ​​of different dimensions and magnitudes to the same range (such as [0,1] or [-1,1]). By eliminating the influence of dimensions, the consistency of the data is improved, which can enhance the accuracy of fault detection and condition assessment of power semiconductors.

[0121] It should be understood that, assuming the historical parameter set is represented as {(2.1A,5.0V), (2.0A,5.0V), (2.2A,5.1V)}, and the node set to be processed is {(2.1A, 2.0A, 2.2A), (5.0V, 5.0V, 5.1V)}, the normalization method is used to normalize the node set to be processed, resulting in a normalized node set of {(0.5, 0.0, 1.0), (0.0, 0.0, 1.0)}. The historical parameter set is then updated using the normalized node set {(0.5, 0.0, 1.0), (0.0, 0.0, 1.0)}, resulting in a normalized historical parameter set of {(0.5,0.0), (0.0,0.0), (1.0,1.0)}. It is common knowledge that power semiconductors may exhibit certain parameters that are positive under normal conditions and negative under fault conditions. Therefore, to accommodate the data characteristics of different parameter values, the specific normalization method is determined by judging whether both positive and negative values ​​exist simultaneously in the node to be processed. Optionally, if both positive and negative values ​​exist simultaneously, the Max-Abs normalization method can be used to normalize, preserving the positive and negative characteristics of the historical parameter values ​​themselves; otherwise, the Min-Max normalization method is used. This embodiment of the invention selects a suitable normalization method for the node to be processed by judging whether both positive and negative historical parameter values ​​exist simultaneously, ensuring that the normalized historical parameter values ​​have higher accuracy in subsequent evaluations.

[0122] S4. Obtain the normal parameter set and the fault parameter set based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values.

[0123] It should be explained that obtaining the normal parameter set and the fault parameter set based on the normalized historical parameter set includes:

[0124] Using preset partitioning identifiers, the normalized historical parameter groups in the normalized historical parameter group set are identified as normal parameter groups or fault parameter groups.

[0125] The normal parameter group and the fault parameter group are summarized separately to obtain the normal parameter group set and the fault parameter group set.

[0126] Understandably, the partitioning identifier is a mark that can distinguish historical parameter groups in the normalized historical parameter set; here, the partitioning identifier is either normal or faulty. It is easy to understand that before acquiring the historical parameter set, the acquired historical parameter set can be marked based on the semiconductor's operating state. For example, historical parameter sets acquired when the semiconductor is in a faulty state can be marked as the 0-historical parameter set. Therefore, the normalized historical parameter set can be divided into a normal parameter set and a faulty parameter set using the partitioning identifier.

[0127] For example, the normalized historical parameter set is {(0.5,0.0), (0.0,0.0), (1.0,1.0)}. Assuming that the initial historical parameter sets corresponding to the normalized historical parameter sets (0.5,0.0), (0.0,0.0), and (1.0,1.0) are stored in the historical parameter set storage unit in states of normal, normal, and fault, respectively, the initial historical parameter sets in the normal state are identified as the normal parameter set, and the initial historical parameter sets in the fault state are identified as the fault parameter set. Thus, the normal parameter set is {normal-(0.5,0.0), normal-(0.0,0.0)}, and the fault parameter set is {fault-(1.0,1.0)}. This embodiment of the invention divides the normalized historical parameter set into a normal parameter set and a fault parameter set by using identifiers, providing a clear data foundation for subsequent evaluation of the power semiconductor's state using the normalized historical parameter set.

[0128] S5. Based on the preset detection period, preset detection number, power semiconductor, and the detection parameter group acquisition unit, a detection parameter group set is acquired, wherein the detection parameter group set includes multiple detection parameter groups, and each detection parameter group includes M detection parameter values.

[0129] Further, the acquisition of the detection parameter set based on the preset detection time period, the preset number of detections, the power semiconductor, and the detection parameter set acquisition unit includes:

[0130] Multiple detection times are calculated based on the detection period and the number of detections, using the following formula:

[0131]

[0132] Among them, t θ Let t represent the θ-th detection time among multiple detection times. min The start time of the detection period is indicated by Δt, the detection period is indicated by Δt, and the number of detections is indicated by R (R is an integer greater than 1).

[0133] The detection parameter set is obtained based on multiple detection times, power semiconductors, and the detection parameter set acquisition unit.

[0134] It should be explained that the detection period refers to the time interval for parameter detection of the power semiconductor, and the starting time of this time interval is t. min The end time is t min +Δt. The number of detections is the number of times the parameters are detected within the detection period. The detection parameter set refers to the set of all detection parameter sets obtained through multiple detections within a preset detection period. A detection parameter set is a set of parameter values ​​containing M types of detection parameter values. Detection parameter values ​​refer to the various detection parameter values ​​of the power semiconductor collected at a certain detection moment. Detection parameter values ​​include, but are not limited to: collector current, gate voltage, and on-state saturation voltage drop.

[0135] For example, suppose the detection time is set to 10:00 to 11:00 AM on a certain day, with the starting time being t. min The detection time is set from 10:00 to 11:00, with a Δt of 60 minutes. Six detections need to be performed within the detection period. Using the aforementioned formula, the six detection times are calculated as follows: 10:05, 10:15, 10:25, 10:35, 10:45, and 10:55. This embodiment of the invention utilizes the aforementioned calculation formula to obtain the detection times, ensuring that the detection times are evenly distributed within the preset detection period. This avoids overly concentrated or sparse collection of detection parameter sets, ensuring that the collected detection parameter sets can comprehensively reflect the operating status of the power semiconductor within the detection period, thus improving the accuracy of fault detection and status assessment.

[0136] S6. Obtain a normalized detection parameter set using the normalization method and the detection parameter set, and transmit the normalized detection parameter set to the state evaluation unit. When the state evaluation unit successfully receives the normalized detection parameter set, it evaluates the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the state is a normal state or a fault state.

[0137] It is understandable that the method for obtaining the normalized detection parameter set is the same as the method for obtaining the normalized historical parameter set and can achieve the same effect, so it will not be described again here.

[0138] In detail, the evaluation of the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set includes:

[0139] For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations:

[0140] The minimum normal evaluation value is obtained based on the normalized detection parameter set, the normal parameter set, and the pre-constructed evaluation value calculation formula, wherein the evaluation value calculation formula is as follows:

[0141]

[0142] Among them, l min y represents the minimum normal evaluation value among all normal parameter groups in the normalized detection parameter set and the normal parameter set. k y′ represents the k-th normalized detection parameter value in the normalized detection parameter set. k This represents the k-th normal parameter value in the normal parameter group; M represents the total number of normal parameter values ​​in the normal parameter group and the total number of normalized detection parameter values ​​in the normalized detection parameter group; and g represents the total number of normal parameter groups in the normal parameter group set.

[0143] The minimum fault assessment value is obtained based on the normalized detection parameter set, the fault parameter set, and the evaluation value calculation formula.

[0144] Compare the minimum normal assessment value with the minimum fault assessment value;

[0145] If the minimum normal assessment value is greater than the minimum fault assessment value, the normalized test parameter group is identified as the fault parameter group; otherwise, the normalized test parameter group is identified as the normal parameter group.

[0146] By summarizing the fault parameter group and the normal parameter group, the minimum evaluation value set is obtained;

[0147] The state of the power semiconductor is evaluated based on the minimum set of evaluation values.

[0148] It should be understood that the minimum normal evaluation value is the minimum distance between the normalized test parameter set and all normal parameter sets in the normal parameter set. This distance is obtained through the evaluation value calculation formula, which calculates the distance between the test parameter set and each normal parameter set, and then takes the minimum value among these distances as the minimum normal evaluation value. Similarly, the minimum fault evaluation value is the minimum distance between the normalized test parameter set and all fault parameter sets in the fault parameter set. A normalized detection parameter set corresponds to a minimum normal evaluation value and a minimum fault evaluation value. By comparing the minimum normal evaluation value and the minimum fault evaluation value, it can be determined whether the normalized detection parameter set is closer to a normal state or a fault state. If the minimum normal evaluation value is greater than the minimum fault evaluation value, it indicates that the normalized detection parameter set is more similar to the fault parameter set, that is, the state of the power semiconductor corresponding to the normalized detection parameter set is closer to a fault state. Therefore, the normalized detection parameter set should be identified as a fault parameter set. Otherwise, the normalized detection parameter set should be identified as a normal parameter set. The method for identifying a normal or fault parameter set is the same as the method for obtaining the target historical parameter set, and will not be repeated here.

[0149] Furthermore, the evaluation of the state of the power semiconductor based on the minimum set of evaluation values ​​includes:

[0150] The number of normal parameter groups and the number of faulty parameter groups in the minimum evaluation value set are counted separately to obtain the number of normal parameters and the number of faulty parameters.

[0151] For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations:

[0152] Calculate the Euclidean distance of each normal parameter group in the normalized detection parameter group and normal parameter group set to obtain the evaluation normal distance subset. After each normalized detection parameter group in the normalized detection parameter group set has obtained the corresponding evaluation normal distance subset, summarize the evaluation normal distance subset to obtain the evaluation normal distance set. The evaluation normal distance subset includes multiple evaluation normal distances.

[0153] Calculate the mean of the normal distances in the assessment set to obtain the normal assessment mean;

[0154] Determine whether the number of normal operations, the number of faults, and the average value of normal operations meet the pre-built evaluation criteria, wherein the evaluation criteria are as follows:

[0155]

[0156] Where q1 represents the number of normal operations, q2 represents the number of faults, α represents the preset evaluation threshold, and p 1l p1 represents the l-th normal distance in the normal distance set, p2 represents the normal mean value, and β is a preset coefficient.

[0157] If the evaluation conditions are met, the power semiconductor is confirmed to be in a normal state; otherwise, the power semiconductor is confirmed to be in a fault state.

[0158] Understandably, the evaluation normal distance subset is the set of Euclidean distances between a certain normalized detection parameter set in the normalized detection parameter set and every normal parameter set in the normal parameter set, and the evaluation normal distance set is the set of all evaluation normal distance subsets in the normalized detection parameter set. This represents the proportion of normal quantities to the sum of normal quantities and faulty quantities. If this proportion is greater than a set evaluation threshold, the power semiconductor is considered to be closer to a normal state. This represents the maximum relative deviation between each normal distance in the normal distance set and the normal average. If the maximum relative deviation is less than or equal to a preset coefficient, the normalized detection parameter set is considered to have a high similarity to the normal parameter set, and the power semiconductor is more likely to be in a normal state. When both conditions are met, the power semiconductor is considered to be in a normal state; otherwise, the power semiconductor is confirmed to be in a fault state.

[0159] To address the problems described in the background section, this invention receives a fault detection command from a power semiconductor and establishes a fault detection system based on the command. The fault detection system includes a historical parameter set storage unit, a detection parameter set acquisition unit, and a status evaluation unit. The historical parameter set storage unit acquires a set of historical parameter sets, each containing multiple historical parameter sets and M types of historical parameter values. This invention verifies whether anomalies have occurred in the initial historical parameter set after storage by checking conditions, thus obtaining an accurate historical parameter set and improving the accuracy of subsequent evaluations using the historical parameter set as a reference. To ensure the accuracy of power semiconductor status, this invention obtains a normalized historical parameter set based on a historical parameter set and a pre-constructed normalization method. Specifically, this invention selects between normal and fault parameter sets based on the normalized historical parameter set by determining whether the historical parameter values ​​in the node to be processed simultaneously contain both positive and negative values. The normal parameter set includes multiple normal parameter groups, each containing M types of normal parameter values. Similarly, the fault parameter set includes multiple fault parameter groups, each containing M types of fault parameter values. This invention divides the normalized historical parameter set into normal and fault parameter sets using a partitioning identifier. The data set, through classification, provides a clear data foundation for subsequent evaluation of the power semiconductor's state using normalized historical parameter sets. This invention acquires a detection parameter set based on a preset detection period, a preset number of detections, the power semiconductor, and the detection parameter set acquisition unit. The detection parameter set includes multiple detection parameter groups, and each detection parameter group includes M detection parameter values. Therefore, this invention uses the calculation formula to obtain the detection time, ensuring that the detection time is evenly distributed within the preset detection period, avoiding overly concentrated or sparse acquisition of detection parameter groups, and ensuring that the acquired detection parameter sets comprehensively reflect the power semiconductor's state within the detection period. To improve the accuracy of fault detection and status assessment, this invention utilizes the normalization method and detection parameter set to obtain a normalized detection parameter set, which is then transmitted to the status assessment unit. Upon successful receipt of the normalized detection parameter set, the status assessment unit assesses the power semiconductor's status based on the normal parameter set, fault parameter set, and the normalized detection parameter set. The status is either normal or faulty. This invention improves the accuracy and reliability of fault detection through dual verification of the ratio and distance mean and deviation between normal and fault parameter sets, achieving precise judgment of the power semiconductor's status. Therefore, this invention can improve the efficiency and accuracy of power semiconductor fault detection.

[0160] like Figure 2 The diagram shown is a functional block diagram of a power semiconductor fault detection system based on multimodal data provided in an embodiment of the present invention.

[0161] The power semiconductor fault detection system 100 based on multimodal data described in this invention can be installed in an electronic device. Depending on the functions implemented, the power semiconductor fault detection system 100 based on multimodal data may include a historical parameter set acquisition module 101, a historical parameter set processing module 102, a detection parameter set acquisition module 103, and a status evaluation module 104. The module described in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0162] The historical parameter set acquisition module 101 is used to receive a fault detection command from a power semiconductor and confirm a fault detection system based on the fault detection command. The fault detection system includes: a historical parameter set storage unit, a detection parameter set acquisition unit, and a status evaluation unit.

[0163] The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values.

[0164] The historical parameter set processing module 102 is used to obtain a normalized historical parameter set based on the historical parameter set and a pre-constructed normalization method.

[0165] The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values.

[0166] The detection parameter set acquisition module 103 is used to acquire a detection parameter set based on a preset detection period, a preset number of detections, power semiconductors, and the detection parameter set acquisition unit. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values.

[0167] The state assessment module 104 is used to obtain a normalized detection parameter set using the normalization method and the detection parameter set, and transmit the normalized detection parameter set to the state assessment unit. When the state assessment unit successfully receives the normalized detection parameter set, it assesses the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the state is a normal state or a fault state.

[0168] In detail, the modules in the power semiconductor fault detection system 100 based on multimodal data described in this embodiment of the invention employ the same methods as described above during use. Figure 1The method used is the same as the power semiconductor fault detection method based on multimodal data described in the article, and can produce the same technical effect, so it will not be repeated here.

[0169] like Figure 3 The diagram shown is a structural schematic of an electronic device that implements a power semiconductor fault detection method based on multimodal data, according to an embodiment of the present invention.

[0170] The electronic device 1 may include a processor 10, a memory 11 and a bus 12, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a power semiconductor fault detection method program based on multimodal data.

[0171] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 includes both internal storage units and external storage devices of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as the code of a power semiconductor fault detection method program based on multimodal data, but also to temporarily store data that has been output or will be output.

[0172] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules stored in the memory 11 (e.g., a power semiconductor fault detection method program based on multimodal data) and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0173] The bus 12 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus 12 can be divided into an address bus, a data bus, a control bus, etc. The bus 12 is configured to realize the connection and communication between the memory 11 and at least one processor 10, etc.

[0174] Figure 3 Only electronic devices with components are shown; those skilled in the art will understand that... Figure 3 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0175] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0176] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0177] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired or wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0178] The power semiconductor fault detection method program based on multimodal data stored in the memory 11 of the electronic device 1 is a combination of multiple instructions. When run in the processor 10, it can achieve the following:

[0179] The system receives a fault detection command from a power semiconductor and identifies a fault detection system based on the fault detection command. The fault detection system includes: a historical parameter group storage unit, a detection parameter group acquisition unit, and a status evaluation unit.

[0180] The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values.

[0181] Obtain a normalized historical parameter set based on the historical parameter set and a pre-built normalization method;

[0182] The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values.

[0183] Based on a preset detection period, a preset number of detections, a power semiconductor, and the detection parameter set acquisition unit, a detection parameter set is acquired. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values.

[0184] The normalization method and the detection parameter set are used to obtain a normalized detection parameter set, which is then transmitted to the state evaluation unit. When the state evaluation unit successfully receives the normalized detection parameter set, it evaluates the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the state is a normal state or a fault state.

[0185] Specifically, the processor 10's implementation method for the above instructions can be found in [reference needed]. Figures 1 to 3 The descriptions of the relevant steps in the corresponding embodiments are not repeated here.

[0186] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0187] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0188] The system receives a fault detection command from a power semiconductor and identifies a fault detection system based on the fault detection command. The fault detection system includes: a historical parameter group storage unit, a detection parameter group acquisition unit, and a status evaluation unit.

[0189] The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values.

[0190] Obtain a normalized historical parameter set based on the historical parameter set and a pre-built normalization method;

[0191] The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values.

[0192] Based on a preset detection period, a preset number of detections, a power semiconductor, and the detection parameter set acquisition unit, a detection parameter set is acquired. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values.

[0193] The normalization method and the detection parameter set are used to obtain a normalized detection parameter set, which is then transmitted to the state evaluation unit. When the state evaluation unit successfully receives the normalized detection parameter set, it evaluates the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the state is a normal state or a fault state.

[0194] In the embodiments provided by this invention, it should be understood that the disclosed devices, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and actual implementations may have other classification methods.

[0195] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0196] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0197] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0198] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A power semiconductor fault detection method based on multimodal data, characterized in that, The method includes: The system receives a fault detection command from a power semiconductor and identifies a fault detection system based on the fault detection command. The fault detection system includes: a historical parameter group storage unit, a detection parameter group acquisition unit, and a status evaluation unit. The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values. Obtain a normalized historical parameter set based on the historical parameter set and a pre-built normalization method; The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values. Based on a preset detection period, a preset number of detections, a power semiconductor, and the detection parameter set acquisition unit, a detection parameter set is acquired. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values. The normalization method and the detection parameter set are used to obtain a normalized detection parameter set, which is then transmitted to the state evaluation unit. When the state evaluation unit successfully receives the normalized detection parameter set, it evaluates the state of the power semiconductor based on the normal parameter set, the fault parameter set, and the normalized detection parameter set, wherein the state is a normal state or a fault state. The evaluation of the power semiconductor's state based on the normal parameter set, fault parameter set, and normalized detection parameter set includes: For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations: The minimum normal evaluation value is obtained based on the normalized detection parameter set, the normal parameter set, and the pre-constructed evaluation value calculation formula, wherein the evaluation value calculation formula is as follows: in, This represents the minimum normal evaluation value among all normal parameter groups in the normal parameter group set and the normal parameter group set. Indicates the first normalized detection parameter group A normalized detection parameter value, Indicates the first parameter in the normal parameter group A normal parameter value, This indicates that there are a total of [number] parameters in the normal parameter group. There are a total of normal parameter values ​​and normalized detection parameter groups. A normalized detection parameter value, This indicates that there are a total of [number] normal parameter groups in the set. A normal parameter group; The minimum fault assessment value is obtained based on the normalized detection parameter set, the fault parameter set, and the evaluation value calculation formula. Compare the minimum normal assessment value with the minimum fault assessment value; If the minimum normal assessment value is greater than the minimum fault assessment value, the normalized test parameter group is identified as the fault parameter group; otherwise, the normalized test parameter group is identified as the normal parameter group. By summarizing the fault parameter group and the normal parameter group, the minimum evaluation value set is obtained; The state of the power semiconductor is evaluated based on the minimum set of evaluation values; The evaluation of the state of the power semiconductor based on the minimum evaluation value set includes: The number of normal parameter groups and the number of faulty parameter groups in the minimum evaluation value set are counted separately to obtain the number of normal parameters and the number of faulty parameters. For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations: Calculate the Euclidean distance of each normal parameter group in the normalized detection parameter group and normal parameter group set to obtain the evaluation normal distance subset. After each normalized detection parameter group in the normalized detection parameter group set has obtained the corresponding evaluation normal distance subset, summarize the evaluation normal distance subset to obtain the evaluation normal distance set. The evaluation normal distance subset includes multiple evaluation normal distances. Calculate the mean of the normal distances in the assessment set to obtain the normal assessment mean; Determine whether the number of normal operations, the number of faults, and the average value of normal operations meet the pre-built evaluation criteria, wherein the evaluation criteria are as follows: in, Indicates normal quantity. Indicates the number of faults. This indicates the preset evaluation threshold. Indicates the evaluation of the normal distance set. An assessment of normal distance, This indicates the normal mean of the assessment. These are preset coefficients; If the evaluation conditions are met, the power semiconductor is confirmed to be in a normal state; otherwise, the power semiconductor is confirmed to be in a fault state.

2. The power semiconductor fault detection method based on multimodal data as described in claim 1, characterized in that, The process of obtaining the historical parameter set based on the historical parameter set storage unit includes: The initial historical parameter set is obtained based on the historical parameter set storage unit, and the initial historical parameter set is stored in the state evaluation unit to obtain the stored historical parameter set, wherein the initial historical parameter set includes multiple initial historical parameter sets. Sort the initial historical parameter sets in the initial historical parameter set set and the stored historical parameter sets in the stored historical parameter set set according to the time sequence from front to back to obtain the initial sequence and the stored sequence. An analysis node set is obtained using the initial sequence and the stored sequence. The analysis node set includes multiple analysis nodes, and each analysis node consists of an initial historical parameter set and a stored historical parameter set. The position of the initial historical parameter set in the initial sequence is the same as the position of the stored historical parameter set in the stored sequence. The following operations are performed on each analysis node in the analysis node set: The initial verification value and the stored verification value are obtained by using the initial historical parameter group and the stored historical parameter group corresponding to the analysis node. The analysis node is updated using the initial verification value and the stored verification value to obtain the updated node. The updated nodes are then summarized to obtain the updated node set. Determine whether the updated node set meets the pre-built verification conditions, wherein the verification conditions are as follows: in, Indicates the first node in the updated node set. The stored verification value corresponding to each update node Indicates the first node in the updated node set. The initial historical parameter group corresponding to each update node, and the initial historical parameter group is as follows: ,in, These represent the first and second historical parameter values ​​in the initial historical parameter set, respectively. This indicates that there are a total of [number] in the initial history parameter set. Historical parameter values, Indicates the first node in the updated node set. The stored historical parameter group corresponding to each analysis node This indicates that the update node set has a total of One update node, Indicates the first node in the updated node set. The initial verification value corresponding to each analysis node; If the verification condition is met, the stored historical parameter set is used as the historical parameter set; otherwise, the process returns to the step of storing the initial historical parameter set in the state evaluation unit until the verification condition is met.

3. The power semiconductor fault detection method based on multimodal data as described in claim 2, characterized in that, The process of obtaining the initial verification value and the stored verification value using the initial historical parameter set and the stored historical parameter set corresponding to the analysis node includes: The initial historical parameter set and the stored historical parameter set are converted using pre-built format conversion methods to obtain the initial historical sequence and the stored historical sequence. The initial verification value is obtained based on the initial historical sequence and the pre-constructed hash function; The storage verification value is obtained based on the stored historical sequence and the hash function.

4. The power semiconductor fault detection method based on multimodal data as described in claim 3, characterized in that, The method for obtaining a normalized historical parameter set based on the historical parameter set and a pre-constructed normalization method includes: A set of nodes to be processed is obtained based on a set of historical parameters. Nodes to be processed are then extracted sequentially from this set, and normalized nodes are obtained using the normalization method described below. in, Indicates the first node among the nodes to be processed One parameter value to be processed, ∃ <0 indicates that the node in the pending process is the first one. The value of the parameter to be processed is negative. This represents the maximum absolute value of all parameter values ​​to be processed in the node to be processed. This represents the minimum absolute value of all parameter values ​​to be processed in the node to be processed. This indicates taking the maximum value. This indicates taking the minimum value. Represents a normalized node; After confirming that each node to be processed has a corresponding normalized node, the normalized nodes are aggregated to obtain a normalized node set. The normalized node set is then used to update the historical parameter set to obtain a normalized historical parameter set.

5. The power semiconductor fault detection method based on multimodal data as described in claim 4, characterized in that, The process of obtaining the set of nodes to be processed based on historical parameter sets includes: The parameter type set is obtained based on the historical parameter set, wherein the parameter type set includes multiple parameter types and the number of parameter types is M; Perform the following operation on each parameter type in the parameter type set: By utilizing parameter types, multiple parameter values ​​to be processed can be retrieved from the historical parameter set. By aggregating multiple parameter values, the node to be processed is obtained; After confirming that each parameter type has obtained a corresponding node to be processed, the nodes to be processed are summarized to obtain a set of nodes to be processed.

6. The power semiconductor fault detection method based on multimodal data as described in claim 5, characterized in that, The process of obtaining normal parameter sets and fault parameter sets based on normalized historical parameter sets includes: Using preset partitioning identifiers, the normalized historical parameter groups in the normalized historical parameter group set are identified as normal parameter groups or fault parameter groups. The normal parameter group and the fault parameter group are summarized separately to obtain the normal parameter group set and the fault parameter group set.

7. The power semiconductor fault detection method based on multimodal data as described in claim 6, characterized in that, The process of acquiring a set of detection parameters based on a preset detection period, a preset number of detections, power semiconductors, and the detection parameter set acquisition unit includes: Multiple detection times are calculated based on the detection period and the number of detections, using the following formula: in, Indicates the first detection time among multiple detection times. Each detection moment, This indicates the start time of the detection period. Indicates the detection period. The number of detections is represented by R (where R is an integer greater than 1). The detection parameter set is obtained based on multiple detection times, power semiconductors, and the detection parameter set acquisition unit.

8. A power semiconductor fault detection system based on multimodal data, characterized in that, The system includes: The historical parameter set acquisition module is used to receive fault detection instructions from power semiconductors and confirm the fault detection system based on the fault detection instructions. The fault detection system includes: a historical parameter set storage unit, a detection parameter set acquisition unit, and a status evaluation unit. The historical parameter set is obtained based on the historical parameter set storage unit. The historical parameter set includes multiple historical parameter sets, and each historical parameter set includes M kinds of historical parameter values. The historical parameter set processing module is used to obtain a normalized historical parameter set based on the historical parameter set and a pre-built normalization method. The normal parameter set and the fault parameter set are obtained based on the normalized historical parameter set. The normal parameter set includes multiple normal parameter sets, and each normal parameter set includes M normal parameter values. The fault parameter set includes multiple fault parameter sets, and each fault parameter set includes M fault parameter values. The detection parameter set acquisition module is used to acquire a detection parameter set based on a preset detection period, a preset number of detections, power semiconductors, and the detection parameter set acquisition unit. The detection parameter set includes multiple detection parameter sets, and each detection parameter set includes M detection parameter values. The status assessment module is used to obtain a normalized detection parameter set using the normalization method and the detection parameter set, and transmit the normalized detection parameter set to the status assessment unit. When the status assessment unit successfully receives the normalized detection parameter set, it assesses the status of the power semiconductor based on the normal parameter set, the fault parameter set and the normalized detection parameter set, wherein the status is a normal status or a fault status. The evaluation of the power semiconductor's state based on the normal parameter set, fault parameter set, and normalized detection parameter set includes: For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations: The minimum normal evaluation value is obtained based on the normalized detection parameter set, the normal parameter set, and the pre-constructed evaluation value calculation formula, wherein the evaluation value calculation formula is as follows: in, This represents the minimum normal evaluation value among all normal parameter groups in the normal parameter group set and the normal parameter group set. Indicates the first normalized detection parameter group A normalized detection parameter value, Indicates the first parameter in the normal parameter group A normal parameter value, This indicates that there are a total of [number] parameters in the normal parameter group. There are a total of normal parameter values ​​and normalized detection parameter groups. A normalized detection parameter value, This indicates that there are a total of [number] normal parameter groups in the set. A normal parameter group; The minimum fault assessment value is obtained based on the normalized detection parameter set, the fault parameter set, and the evaluation value calculation formula. Compare the minimum normal assessment value with the minimum fault assessment value; If the minimum normal assessment value is greater than the minimum fault assessment value, the normalized test parameter group is identified as the fault parameter group; otherwise, the normalized test parameter group is identified as the normal parameter group. By summarizing the fault parameter group and the normal parameter group, the minimum evaluation value set is obtained; The state of the power semiconductor is evaluated based on the minimum set of evaluation values; The evaluation of the state of the power semiconductor based on the minimum evaluation value set includes: The number of normal parameter groups and the number of faulty parameter groups in the minimum evaluation value set are counted separately to obtain the number of normal parameters and the number of faulty parameters. For each normalized detection parameter group in the normalized detection parameter group set, perform the following operations: Calculate the Euclidean distance of each normal parameter group in the normalized detection parameter group and normal parameter group set to obtain the evaluation normal distance subset. After each normalized detection parameter group in the normalized detection parameter group set has obtained the corresponding evaluation normal distance subset, summarize the evaluation normal distance subset to obtain the evaluation normal distance set. The evaluation normal distance subset includes multiple evaluation normal distances. Calculate the mean of the normal distances in the assessment set to obtain the normal assessment mean; Determine whether the number of normal operations, the number of faults, and the average value of normal operations meet the pre-built evaluation criteria, wherein the evaluation criteria are as follows: in, Indicates normal quantity. Indicates the number of faults. This indicates the preset evaluation threshold. Indicates the evaluation of the normal distance set. An assessment of normal distance, This indicates the normal mean of the assessment. These are preset coefficients; If the evaluation conditions are met, the power semiconductor is confirmed to be in a normal state; otherwise, the power semiconductor is confirmed to be in a fault state.

Images