Method and device for detecting a fault of a hvdc valve

Abstract

The application discloses a kind of flexible straight converter valve fault detection method and device, the method includes: the detection data of flexible straight converter valve is collected, the detection data includes operating data and related data;The failure probability of each sub-module of the flexible straight converter valve is calculated;The fault tree of the flexible straight converter valve is established, and whether the flexible straight converter valve fails is determined according to the failure probability of each sub-module and the fault tree.This application can improve the reliability, accuracy and timeliness of flexible straight converter valve fault detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of flexible direct current valve fault detection, and in particular to a flexible direct current valve fault detection method and device. BACKGROUND

[0002] This section is intended to provide background information to facilitate a better understanding of embodiments of the present application described in the claims. The description herein does not constitute admission of prior art.

[0003] At present, as an important part of flexible direct current transmission projects, the operation reliability of the flexible direct current station directly affects the safe, stable and efficient operation of the overall flexible direct current power grid. At present, the fault detection of the flexible direct current valve mainly includes threshold alarm method, protection trigger alarm and defect prediction method based on model. However, the current flexible direct current valve fault detection has many problems: first, the accuracy of the artificial prediction through the change rule of the overall variable of the existing flexible direct current valve is low, second, the misjudgment rate of whether a defect occurs through a threshold of a variable is high, and third, the current prediction method is imperfect and can only accurately locate the fault root cause after disassembling the components after the fault occurs. SUMMARY

[0004] One object of the present application is to provide a flexible direct current valve fault detection method to improve the reliability, accuracy and timeliness of the flexible direct current valve fault detection. Another object of the present application is to provide a flexible direct current valve fault detection device. Still another object of the present application is to provide a computer device. Still another object of the present application is to provide a readable medium.

[0005] In order to achieve the above objects, one aspect of the present application discloses a flexible direct current valve fault detection method, comprising:

[0006] Collecting detection data of the flexible direct current valve, the detection data including operation data and related data;

[0007] Calculating the fault probability of each sub-module of the flexible direct current valve;

[0008] Establishing a fault tree of the flexible direct current valve, and determining whether the flexible direct current valve has a fault according to the fault probability of each sub-module and the fault tree.

[0009] Optionally, further comprising, after collecting the detection data of the flexible direct current valve:

[0010] Preprocessing the detection data of the flexible direct current valve.

[0011] Optionally, the preprocessing the detection data of the flexible direct current valve comprises:

[0012] The detection data is filled with missing values, abnormal value processing, data conversion, data feature analysis and category feature analysis;

[0013] According to the number of samples of each category obtained by the category feature analysis, it is determined whether there is a data imbalance problem, and if so, corresponding undersampling, oversampling or category weight adjustment processing is performed on the detection data.

[0014] Optionally, the calculation of the failure probability of each sub-module of the flexible HVDC valve comprises:

[0015] A plurality of decision trees are established for each sub-module of the flexible HVDC valve;

[0016] According to the plurality of decision trees, a random forest corresponding to each sub-module is formed;

[0017] Based on the prediction result of each decision tree in the random forest based on the detection data, the failure probability of the corresponding sub-module is obtained.

[0018] Optionally, the calculation of the failure probability of each sub-module of the flexible HVDC valve comprises:

[0019] A plurality of classifiers corresponding to a plurality of failure categories are respectively constructed for each sub-module of the flexible HVDC valve;

[0020] Through the plurality of classifiers, a corresponding failure detection result is predicted based on the detection data;

[0021] Based on the failure detection result of the plurality of classifiers, the failure probability of the corresponding sub-module is obtained.

[0022] Optionally, the establishment of the fault tree of the flexible HVDC valve according to the failure probability of each sub-module comprises:

[0023] According to the failure conditions corresponding to different failure categories of each sub-module of the flexible HVDC valve, at least one of a static fault tree and a dynamic fault tree is generated.

[0024] Optionally, the determination of whether the flexible HVDC valve fails according to the fault tree comprises:

[0025] The dynamic fault tree is converted into a corresponding Markov model based on the failure conditions;

[0026] The Markov model is simplified and analyzed to obtain the dynamic fault result of each dynamic fault tree by using logical rules and timing rules;

[0027] The static fault result of each sub-module is determined based on the detection data through the static fault tree.

[0028] This application also discloses a fault detection device for a flexible DC converter valve, comprising:

[0029] The data acquisition module is used to collect the detection data of the flexible DC converter valve, which includes operating data and related data;

[0030] The submodule fault detection module is used to calculate the fault probability of each submodule of the flexible DC converter valve.

[0031] The converter valve fault determination module is used to establish a fault tree for the flexible DC converter valve and determine whether the flexible DC converter valve has failed based on the fault probability of each sub-module and the fault tree.

[0032] This application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0033] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0034] This application presents a method for fault detection of flexible DC-DC converter valves. The method collects detection data from the flexible DC-DC converter valve, including operational data and related data. It calculates the fault probability of each sub-module of the flexible DC-DC converter valve, establishes a fault tree for the valve, and determines whether the flexible DC-DC converter valve has malfunctioned based on the fault probability of each sub-module and the fault tree. Thus, this application calculates the fault probability of each sub-module of the flexible DC-DC converter valve using the detection data, establishes a fault tree for the valve, and determines whether the flexible DC-DC converter valve has malfunctioned based on the fault tree and the fault probability of each sub-module within it. Therefore, this application improves the accuracy of sub-module fault probability detection by first calculating the fault probability of each sub-module of the flexible DC-DC converter valve specifically, and then determines whether the flexible DC-DC converter valve as a whole has malfunctioned by using the constructed fault tree and the fault probabilities of each sub-module. This solves the problems of difficult identification, low accuracy, and poor timeliness of abnormal faults in flexible DC-DC converter valves. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0036] Figure 1A flowchart of a specific embodiment of the fault detection method for flexible DC converter valve of this application;

[0037] Figure 2 This is a flowchart of a specific embodiment S110 of the fault detection method for flexible DC converter valve in this application;

[0038] Figure 3 This is a schematic diagram of oversampling in a specific embodiment of the fault detection method for flexible DC converter valves of this application;

[0039] Figure 4 This is a flowchart illustrating one implementation of the fault probability of submodule S200 in the specific embodiment of the fault detection method for flexible DC converter valve of this application.

[0040] Figure 5 This is a schematic diagram of a random forest in a specific embodiment of the fault detection method for flexible DC converter valves in this application;

[0041] Figure 6 This is a flowchart illustrating another implementation of the failure probability of the S200 submodule in the specific embodiment of the fault detection method for the flexible DC converter valve of this application.

[0042] Figure 7 This is a flowchart of a specific embodiment S300 of the fault detection method for flexible DC converter valves in this application;

[0043] Figure 8 This is a schematic diagram of a dynamic fault tree for a specific embodiment of the fault detection method for flexible DC converter valves in this application;

[0044] Figure 9 This is a structural diagram of a specific embodiment of the fault detection device for flexible DC converter valves in this application;

[0045] Figure 10 A schematic diagram of a computer device suitable for implementing embodiments of the present invention is shown. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the embodiments of this application will be further described in detail below with reference to the accompanying drawings. Here, the illustrative embodiments and descriptions of this application are used to explain this application, but are not intended to limit this application.

[0047] Existing alarm strategies for flexible DC-DC converter valves typically employ threshold alarm methods, protection-triggered alarms, and model-based defect prediction methods. Threshold alarm methods monitor faults by setting thresholds. For example, over-temperature alarms monitor the internal temperature of the flexible DC-DC converter valve; when the temperature exceeds a set threshold, an alarm is triggered, potentially indicating overheating and requiring shutdown or load reduction. Insufficient submodule redundancy alarms are triggered when the existing redundancy falls below a design threshold. For protection-triggered alarms, flexible DC-DC converter valves are equipped with various protection functions, such as overcurrent and overvoltage protection. When these protections are triggered, alarm signals are generated. However, some protection-triggered alarms may require a certain response time after detecting a hazardous condition before the protection mechanism can be activated. This delay may prevent timely action when a hazard occurs, potentially causing equipment damage or personal injury. Furthermore, protection-triggered alarms may experience false alarms or missed alarms due to external interference or sensor malfunction.

[0048] Model-based methods essentially predict system output by constructing observers and determine defects based on the residuals between the predicted and observed values. These methods intuitively demonstrate the system response when defects occur, and mature research includes parameter estimation, state estimation, and equivalent space methods. However, as the structure and defect mechanisms of mechanical equipment become increasingly complex, random disturbances increase, and the relationships between input variables, implicit state variables, and output variables become more intricate, the complexity of modeling increases significantly. Furthermore, due to the increase in influencing factors, the reliability and accuracy of the model cannot be guaranteed. Therefore, model-based methods have many limitations in complex engineering applications and are no longer suitable for the current development trend of mechanical equipment. Based on the problems in existing technologies, this application first calculates the fault probability of each sub-module of the flexible DC converter valve in a targeted manner, improving the accuracy of fault probability detection for sub-modules. Then, by constructing a fault tree for the flexible DC converter valve and the fault probabilities of each sub-module, it determines whether the entire flexible DC converter valve has failed, solving the problems of difficult, low-accuracy, and poor timeliness in identifying abnormal faults in flexible DC converter valves.

[0049] According to one aspect of this application, this embodiment discloses a fault detection method for a flexible DC converter valve. For example... Figure 1 As shown, in this embodiment, the method includes:

[0050] S100: Collect detection data of the flexible DC converter valve, including operating data and related data.

[0051] It should be noted that the operating data may include at least one of the following: input voltage and current, output voltage and current, rated power, response time, harmonic content, and temperature. However, the operating data of the flexible DC converter valve alone cannot accurately predict whether a fault will occur; further data such as simulation factor data, equipment life test data, load-side data, and environmental data are required.

[0052] S200: Calculate the failure probability of each submodule of the flexible DC converter valve.

[0053] S300: Establish a fault tree for the flexible DC converter valve, and determine whether the flexible DC converter valve has failed based on the failure probability of each sub-module and the fault tree.

[0054] This application calculates the failure probability of each submodule of the flexible DC-DC converter valve based on test data, establishes a fault tree for the flexible DC-DC converter valve, and determines whether the flexible DC-DC converter valve has failed based on the fault tree and the failure probability of each submodule in the fault tree. Therefore, this application improves the accuracy of submodule failure probability detection by first calculating the failure probability of each submodule of the flexible DC-DC converter valve in a targeted manner, and then determines whether the flexible DC-DC converter valve as a whole has failed by using the constructed fault tree and the failure probabilities of each submodule, thus solving the problems of difficult, low-accuracy, and poor timeliness in identifying abnormal faults in flexible DC-DC converter valves.

[0055] In an optional implementation, the method further includes, after acquiring the detection data of the flexible DC converter valve:

[0056] S110: Preprocess the detection data of the flexible DC converter valve.

[0057] Specifically, and understandably, data preprocessing is a crucial step that impacts the performance and stability of the final integrated model. The goal of data preprocessing is to prepare the data so that different base models can be better trained and integrated. After obtaining the operational and related data of the flexible DC converter valve, a forward-looking, general understanding of the final merged data can be achieved, such as the data types and memory usage of indexes and columns. Furthermore, descriptive statistical summaries can be generated for numerical data, including data counts and percentiles, the number of data in each class within categorical data, and an understanding of the data distribution.

[0058] In alternative implementations, such as Figure 2 As shown, the preprocessing of the detection data of the flexible DC converter valve in step S110 includes:

[0059] S111: Perform missing value filling, outlier processing, data transformation, data feature analysis, and category feature analysis on the detected data.

[0060] S112: Determine whether there is a data imbalance problem based on the number of samples in each category obtained from the category feature analysis. If so, perform corresponding undersampling, oversampling or category weight adjustment processing on the detection data.

[0061] Data cleaning is the first step in data preprocessing. The acquired detection data may contain missing values, outliers, or erroneous values, all of which can negatively impact the performance of the fault detection model. Therefore, data cleaning is necessary. Cleaning mainly includes data preprocessing methods such as filling in missing values ​​and repairing outliers or removing abnormal samples. Depending on the needs of different sub-modules of the fault detection model, data processing methods such as data transformation, data feature analysis, and category feature analysis can be applied to the detection data, including feature scaling, data balancing, and data segmentation.

[0062] Because certain parameters of similar equipment from different manufacturers can vary, this significantly impacts anomaly detection models. Therefore, data processing is necessary based on factors such as different equipment manufacturers and conditions before being used in model fault detection. For example, the data processing methods for converter valves from different manufacturers typically differ depending on the equipment model, application, and technical characteristics. Different manufacturers' converters use different data formats and sampling frequencies, potentially employing sampling rates of 10 seconds or 5 seconds. Data format affects the data storage structure, and sampling frequency impacts data resolution. Therefore, data preprocessing requires appropriate parsing for different data formats and adjustments to unify different sampling frequencies. Data from different manufacturers requires different calibration methods to ensure accuracy. Therefore, after data processing such as filling missing values, handling outliers, data transformation, data feature analysis, and category feature analysis, the number of samples in each category obtained from the category feature analysis determines whether data imbalance exists. If so, corresponding undersampling, oversampling, or category weight adjustment processing is performed on the detection data to improve the accuracy of fault detection results.

[0063] In specific examples, some flexible DC converter valve sub-components suffer from data imbalance, meaning that the number of samples for some categories is significantly less than that for others. This causes the model to favor the majority class, impacting its performance. Therefore, oversampling can be used to address this data imbalance. For example... Figure 3 As shown, oversampling can be achieved through the following steps:

[0064] Sample x was selected from a small sample set of flexible DC converter valve sub-components. i and x j And the corresponding random numbers 0 < λ < 1, constructing a new sample x based on the relationship between the two samples. n =x i +λ(x j-x i The existing minority class samples are analyzed, and new samples are artificially synthesized based on the minority class samples and added to the dataset.

[0065] For each sample x in the minority class, calculate its distance to all samples in the minority class sample set using Euclidean distance to obtain its k nearest neighbors. A sampling ratio N is set based on the imbalance ratio to determine the sampling multiplier. For each minority class sample x, several samples are randomly selected from its k nearest neighbors. Assume the selected nearest neighbors are... For each randomly selected nearest neighbor Construct new samples from the original sample x using the following formula.

[0066]

[0067] In alternative implementations, such as Figure 4 As shown, S200 calculates the failure probability of each submodule of the flexible DC converter valve, including:

[0068] S211: Establish multiple decision trees for each sub-module of the flexible DC converter valve.

[0069] S212: Form a random forest corresponding to each sub-module based on the multiple decision trees.

[0070] S213: Based on the prediction results of each decision tree in the random forest based on the detection data, the failure probability of the corresponding sub-module is obtained.

[0071] Specifically, this application performs fault analysis on different sub-modules of the flexible DC converter valve. Each sub-module may have multiple different fault categories, and each fault category can be detected using multiple different fault detection models. Then, through ensemble learning, the fault probability of the sub-module for that fault category is determined based on multiple fault detection results. Finally, through ensemble learning, the fault probability of the sub-module and the overall fault probability of the flexible DC converter valve are obtained based on the fault probabilities of all fault categories of the sub-module, thus determining whether the flexible DC converter valve has failed.

[0072] The fault detection model can employ one or more models, such as simulation models, fault life prediction models, probabilistic models, and support vector machines. Each model outputs a probability estimate for a fault category, determining whether a fault corresponding to that category has occurred. Then, the final probability estimate for each fault category is obtained based on the probability estimates from multiple models using ensemble learning methods such as voting or averaging. This application utilizes ensemble learning technology to combine the prediction results of multiple independent models, thereby obtaining a more robust and accurate result regarding whether a fault has occurred in the flexible DC converter valve.

[0073] In this optional implementation, the failure probability of each submodule can be predicted by random forest, and then the result of whether the flexible DC converter valve as a whole has failed can be obtained by ensemble learning based on the failure probabilities of all submodules.

[0074] Random forests are essentially a collection of decision trees, each slightly different from the others. Although decision trees can be prone to overfitting, averaging the results from these trees helps reduce overfitting and maintain the predictive power of the trees. Figure 5 As shown, Tree0 to Tree4 all had poor classification results with the same data. Using random forest reduced the overfitting phenomenon.

[0075] Obtaining the failure probability based on random forest can be achieved through the following steps:

[0076] Step 1: Random Sampling: For each decision tree, samples are randomly drawn from the original training data with replacement to form a new training set. This is called "bootstrap sampling". This process results in the training set for each decision tree being slightly different because each sampling will contain different samples.

[0077] Step 2: Random Feature Selection: For each node's partition, a subset of features is randomly selected from the feature set. This helps reduce the correlation between trees, ensuring slight differences in feature selection for each tree. Random feature selection prevents certain features from dominating the overall prediction results of the random forest.

[0078] Step 3: Train the decision tree: Based on the random sampling and random feature selection described above, train a decision tree. The decision tree construction process can be based on the CART (Classification and Regression Tree) algorithm, which recursively divides the training set into subsets until a certain stopping condition is met (such as the number of node samples being too small).

[0079] Repeat steps 2 and 3: Repeat steps 2 and 3 multiple times to generate multiple independent decision trees.

[0080] Step 4: Analyze the Decision Tree Results: For classification problems, where the failure probability is determined by a classification algorithm, each decision tree will cast a prediction for a class. The final classification result can be determined through a majority vote, selecting the class predicted by the majority of trees as the final failure probability. For regression problems, each decision tree will provide a prediction. The final regression result can be obtained by averaging the predictions from all trees, which will serve as the final failure probability.

[0081] In alternative implementations, such as Figure 6 As shown, S200 calculates the failure probability of each submodule of the flexible DC converter valve, including:

[0082] S221: For each sub-module of the flexible DC converter valve, construct multiple classifiers corresponding to multiple fault categories.

[0083] S222: The corresponding fault detection result is obtained by predicting based on the detection data using the multiple classifiers.

[0084] S223: Obtain the fault probability of the corresponding sub-module based on the fault detection results of the multiple classifiers.

[0085] Specifically, in this optional implementation, the failure probability of the flexible DC converter valve submodule can also be determined by a classifier. The classifier for the submodule is equivalent to a fault detection model for that submodule, with each classifier obtaining the fault detection result based on the detection data of the corresponding submodule. For classification problems, the upper layer of the classifier can employ Stacking, using a voting approach where each classifier votes based on its own fault detection results. The classifier uses an absolute majority vote to obtain the failure probability of the submodule corresponding to the fault category; that is, for the final failure probability result, it must not only satisfy the condition of obtaining the most votes but also the condition of being voted for by more than half. If these conditions are not met, the prediction is directly rejected.

[0086] For regression problems, voting is clearly insufficient. An averaging method is needed to obtain the fault probability of a corresponding submodule for its fault category by averaging the fault detection results from multiple learners. More preferably, a weighted average can be used to calculate the fault probability of a submodule for its fault category by averaging the fault detection results from different classifiers based on preset weights. The preset weights can be determined according to...

[0087] It should be noted that the failure probability of the submodule in this application can also be obtained by using ensemble learning based on the failure detection results of all failure categories of the submodule, depending on the actual situation. For example, when the failure category is critical and affects the operation of the submodule, the failure detection result of the failure category directly determines the final failure probability of the submodule. Of course, the failure probability of the submodule can also be obtained by voting or averaging based on the failure detection results of all failure categories of the submodule.

[0088] In an optional implementation, S300 establishes a fault tree for the flexible DC converter valve based on the failure probability of each submodule, including:

[0089] S310: Generate at least one of a static fault tree and a dynamic fault tree based on the fault conditions corresponding to different fault categories of each submodule in the flexible DC converter valve.

[0090] Specifically, fault trees are a commonly used fault analysis method used to describe the logical relationships of system faults. Traditional fault trees are static, meaning they only consider the inherent probabilities and logical relationships of fault events. Dynamic fault trees, on the other hand, introduce a time dimension, allowing them to consider the evolution of fault events over time. In converter valve fault analysis, dynamic fault trees can more accurately describe the fault occurrence process, including time dependencies and fault triggering conditions.

[0091] This application combines ensemble learning with dynamic fault tree analysis, specifically integrating ensemble learning probability estimation with dynamic fault tree analysis, to obtain more comprehensive and accurate information in fault analysis. First, through ensemble learning probability estimation, the probability estimate for each fault category can be obtained, i.e., the likelihood of each fault occurring. Then, these probability estimates are incorporated into the dynamic fault tree, combining the time dimension and logical relationships to analyze and predict the faults of the converter valve.

[0092] In alternative implementations, such as Figure 7 As shown, S300 determines whether the flexible DC converter valve has malfunctioned based on the fault tree, including:

[0093] S321: Convert the dynamic fault tree into a corresponding Markov model based on the fault conditions.

[0094] S322: Simplify the Markov model using logical and timing rules and analyze the dynamic fault results of each dynamic fault tree.

[0095] S323: Determine the static fault results of each sub-module based on the detection data using the static fault tree.

[0096] Specifically, firstly, based on qualitative analysis, the time-independent parts of the flexible DC converter valve failure, such as stage control fault alarms and station control fault alarms, can be identified, and a static fault tree can be generated. Whether a fault has occurred in this static fault tree can be calculated based on the tree structure and the fault probability of the sub-modules. This is a conventional technical method in this field and will not be elaborated here.

[0097] Traditional fault trees are static, considering only the inherent probabilities and logical relationships of fault events. Dynamic fault trees, however, introduce a time dimension, allowing for the evolution of fault events over time. In converter valve fault analysis, dynamic fault trees can more accurately describe the fault occurrence process, including time dependencies and fault triggering conditions. For example, a fault in a converter valve submodule might have root causes such as a driver board failure or fiber optic communication anomalies. In complex systems, if a fault only occurs when related events occur in a specific order, dynamic fault trees are needed for modeling and system fault analysis. Based on the submodule fault probability, the probability of fault occurrence at time is calculated, and dynamic fault trees for different logic gates are generated. Dynamic fault trees can be formed based on the time and triggering conditions of each submodule in the fault. For example, Figure 8 A dynamic fault tree for a converter valve in a specific example is shown.

[0098] By combining the failure probabilities of sub-modules in the dynamic fault tree, the dynamic fault tree is converted into a corresponding Markov model. After simplifying the Markov model using logical and temporal rules, the analysis is performed to solve the occurrence probability of the top event of each dynamic fault tree, that is, whether the fault category corresponding to the dynamic fault tree has occurred, and then the occurrence probability of the top event (flexible DC converter valve failure) is obtained.

[0099] Starting with the fault state of the converter valve system, and tracing back along the Markov state transition process, all fault modes corresponding to the dynamic fault tree can be identified. Assume that at a certain time, the system is in state t... i The failure rate of the state is λ. i ,x(t i )=λ i By combining the analysis results of each link, the failure probability of the entire system can be obtained:

[0100] P{x(t n |x(t1)=λ1,...,x(t) n-1 )=λ n-1}=P{x(t n )|x(t n-1 )=λ n-1}

[0101] This invention addresses several problems in the detection of abnormal faults in existing flexible DC converter valve equipment: First, the accuracy of predictions based on the changing patterns of existing variables is very low. Second, the false positive rate is high when judging whether a defect has occurred based on a threshold of a certain variable. Third, current prediction methods are imperfect, and most can only accurately locate the root cause of the fault after the fault has occurred by disassembling the component. Based on the knowledge of fault models and fault trees of key sub-components of flexible DC converter valve equipment, this invention utilizes ensemble learning and dynamic fault tree methods to solve the problems of difficulty in identifying abnormal faults, low accuracy, and poor timeliness in flexible DC converter valves. Specifically, first, it utilizes existing isolated models and solves problems such as data imbalance; second, the method based on ensemble learning probability estimation and dynamic fault tree can better handle fault analysis problems in complex systems, improving the prediction and diagnosis capabilities of converter valve faults; third, it can not only consider the integration of multiple models, but also the time dependence and dynamic nature of fault occurrence, thus providing more comprehensive and accurate fault analysis results; fourth, the probability estimation results are incorporated into the dynamic fault tree, combining the time dimension and logical relationships to analyze and predict converter valve faults.

[0102] Based on the same principle, this application also discloses a fault detection device for a flexible DC converter valve. For example... Figure 9 As shown, in this embodiment, the flexible DC converter valve fault detection device includes a data acquisition module 11, a sub-module fault detection module 12, and a converter valve fault determination module 13.

[0103] The data acquisition module 11 is used to collect the detection data of the flexible DC converter valve, which includes operating data and related data.

[0104] The submodule fault detection module 12 is used to calculate the fault probability of each submodule of the flexible DC converter valve.

[0105] The converter valve fault determination module 13 is used to establish a fault tree for the flexible DC converter valve, and determine whether the flexible DC converter valve has failed based on the fault probability of each sub-module and the fault tree.

[0106] This application presents a method for fault detection of flexible DC-DC converter valves. The method collects detection data from the flexible DC-DC converter valve, including operational data and related data. It calculates the fault probability of each sub-module of the flexible DC-DC converter valve, establishes a fault tree for the valve, and determines whether the flexible DC-DC converter valve has malfunctioned based on the fault probability of each sub-module and the fault tree. Thus, this application calculates the fault probability of each sub-module of the flexible DC-DC converter valve using the detection data, establishes a fault tree for the valve, and determines whether the flexible DC-DC converter valve has malfunctioned based on the fault tree and the fault probability of each sub-module within it. Therefore, this application improves the accuracy of sub-module fault probability detection by first calculating the fault probability of each sub-module of the flexible DC-DC converter valve specifically, and then determines whether the flexible DC-DC converter valve as a whole has malfunctioned by using the constructed fault tree and the fault probabilities of each sub-module. This solves the problems of difficult identification, low accuracy, and poor timeliness of abnormal faults in flexible DC-DC converter valves.

[0107] In an optional implementation, the data acquisition module 11 is further configured to preprocess the detection data of the flexible DC converter valve after acquiring the detection data of the flexible DC converter valve.

[0108] In an optional implementation, the data acquisition module 11 is used to fill missing values, handle outliers, transform data, perform data feature analysis, and class feature analysis on the detection data; determine whether there is a data imbalance problem based on the number of samples in each class obtained from the class feature analysis, and if so, perform corresponding undersampling, oversampling, or class weight adjustment processing on the detection data.

[0109] In an optional implementation, the submodule fault detection module 12 is used to establish multiple decision trees for each submodule of the flexible DC converter valve; form a random forest corresponding to each submodule based on the multiple decision trees; and obtain the fault probability of the corresponding submodule based on the prediction result of each decision tree in the random forest based on the detection data.

[0110] In an optional implementation, the submodule fault detection module 12 is used to construct multiple classifiers corresponding to multiple fault categories for each submodule of the flexible DC converter valve; predict the corresponding fault detection result based on the detection data through the multiple classifiers; and obtain the fault probability of the corresponding submodule based on the fault detection results of the multiple classifiers.

[0111] In an optional implementation, the converter valve fault determination module 13 is used to generate at least one of a static fault tree and a dynamic fault tree based on the fault conditions corresponding to different fault categories of each sub-module in the flexible DC converter valve.

[0112] In an optional implementation, the converter valve fault determination module 13 is used to convert the dynamic fault tree into a corresponding Markov model based on the fault conditions; simplify the Markov model using logical rules and timing rules and analyze the dynamic fault results of each dynamic fault tree; and determine the static fault results of each sub-module based on the detection data through the static fault tree.

[0113] Since the principle by which this device solves the problem is similar to the methods described above, the implementation of this device can be found in the implementation of the methods, and will not be repeated here.

[0114] This application also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0115] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0116] Those skilled in the art will understand that the embodiments of this application can be provided as methods, systems, or computer programs, producing the systems, apparatuses, modules, or units described in the above embodiments. Specifically, they can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer device; specifically, a computer device can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0117] In a typical example, the computer device specifically includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method executed by the client as described above, or the method executed by the server as described above.

[0118] The following is for reference. Figure 10 It shows a schematic diagram of the structure of a computer device 600 suitable for implementing the embodiments of this application.

[0119] like Figure 10As shown, the computer device 600 includes a central processing unit (CPU) 601, which can perform various appropriate tasks and processes based on programs stored in read-only memory (ROM) 602 or programs loaded from storage section 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the system 600. The CPU 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0120] The following components are connected to I / O interface 605: an input section 606 including a keyboard, mouse, etc.; an output section 607 including a cathode ray tube (CRT), liquid crystal feedback (LCD), etc., and speakers, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to I / O interface 606 as needed. A removable medium 611, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 610 as needed so that computer programs read from it can be installed in storage section 608 as needed.

[0121] In particular, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program tangibly embodied on a machine-readable medium, the computer program including program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 609, and / or installed from removable medium 611.

[0122] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0123] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0124] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0125] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0126] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0127] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0128] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0129] This application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0130] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0131] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A fault detection method for a flexible DC converter valve, characterized in that, include: Collect test data of the flexible DC converter valve, including operating data and related data; Calculate the failure probability of each submodule of the flexible DC converter valve; Establish a fault tree for the flexible DC converter valve, and determine whether the flexible DC converter valve has failed based on the failure probability of each sub-module and the fault tree; The calculation of the failure probability of each submodule of the flexible DC converter valve includes: For each submodule of the flexible DC converter valve, multiple classifiers are constructed corresponding to multiple fault categories. The corresponding fault detection results are obtained by using the multiple classifiers to predict based on the detection data; The failure probability of the corresponding sub-module is obtained based on the failure detection results of the multiple classifiers; wherein, the failure probability of the sub-module corresponding to the failure category is obtained by the classifier using an absolute majority vote.

2. The fault detection method for a flexible DC converter valve according to claim 1, characterized in that, Further, this includes collecting test data from the flexible DC converter valve, followed by: The detection data of the flexible DC converter valve is preprocessed.

3. The fault detection method for a flexible DC converter valve according to claim 2, characterized in that, The preprocessing of the detection data of the flexible DC converter valve includes: The detection data is subjected to missing value filling, outlier processing, data transformation, data feature analysis, and category feature analysis. Based on the number of samples in each category obtained from the category feature analysis, determine whether there is a data imbalance problem. If so, perform corresponding undersampling, oversampling, or category weight adjustment processing on the detection data.

4. The fault detection method for a flexible DC converter valve according to claim 1, characterized in that, The calculation of the failure probability of each submodule of the flexible DC converter valve includes: Multiple decision trees are established for each sub-module of the flexible DC converter valve; A random forest corresponding to each submodule is formed based on the multiple decision trees; The failure probability of the corresponding submodule is obtained based on the prediction results of each decision tree in the random forest based on the detection data.

5. The fault detection method for a flexible DC converter valve according to claim 1, characterized in that, The step of establishing a fault tree for the flexible DC converter valve based on the failure probability of each sub-module includes: At least one of a static fault tree and a dynamic fault tree is generated based on the fault conditions corresponding to different fault categories of each submodule in the flexible DC converter valve.

6. The fault detection method for a flexible DC converter valve according to claim 5, characterized in that, Determining whether the flexible DC converter valve has failed based on the fault tree includes: The dynamic fault tree is converted into a corresponding Markov model based on the fault conditions; The Markov model is simplified using logical rules and timing rules, and the dynamic fault results of each dynamic fault tree are obtained through analysis. The static fault results of each submodule are determined based on the detection data using the static fault tree.

7. A fault detection device for a flexible DC converter valve, characterized in that, include: The data acquisition module is used to collect the detection data of the flexible DC converter valve, which includes operating data and related data; The submodule fault detection module is used to calculate the fault probability of each submodule of the flexible DC converter valve. A converter valve fault determination module is used to establish a fault tree for the flexible DC converter valve and determine whether the flexible DC converter valve has failed based on the fault probability of each sub-module and the fault tree. The calculation of the failure probability of each submodule of the flexible DC converter valve includes: For each submodule of the flexible DC converter valve, multiple classifiers are constructed corresponding to multiple fault categories. The corresponding fault detection results are obtained by using the multiple classifiers to predict based on the detection data; The failure probability of the corresponding sub-module is obtained based on the failure detection results of the multiple classifiers; wherein, the failure probability of the sub-module corresponding to the failure category is obtained by the classifier using an absolute majority vote.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method of any one of claims 1 to 6.

Images