A clcc converter fault early warning method and system

Abstract

The present application relates to a kind of CLCC converter fault early warning method and system, belong to power system equipment monitoring technical field, it solves the problems that existing CLCC converter fault monitoring is insufficient in the identification result explainability, commutation margin degradation is difficult to be quantified online, dominant risk parameter is difficult to be dynamically identified and the working condition adaptability is insufficient.It includes the acquisition of the multidimensional time series data of CLCC converter, then mechanism mapping model, obtains the current fault risk index and the risk contribution degree of each operating parameter;And disturbance is applied to operating parameter in multidimensional time series, obtain the fault risk index after disturbance;Based on the current fault risk index and the fault risk index after disturbance, obtain the dynamic sensitivity index of each operating parameter, and then obtain the key influence parameter set;Based on multidimensional time series data, identify the current operating condition, and then obtain the comprehensive fault risk value, and then obtain the fault early warning information.Realize the real-time fault early warning of CLCC converter.

Description

Technical Field

[0001] This invention relates to the field of power system equipment monitoring technology, and in particular to a CLCC converter fault early warning method and system. Background Technology

[0002] Controllable Line Commutated Converter (CLCC), as a novel converter solution integrating the advantages of traditional thyristor converter technology and flexible DC transmission technology, boasts superior performance in terms of large capacity, low loss, and controllable shutdown. It demonstrates significant engineering application value in high-capacity, long-distance power transmission and complex grid interconnection scenarios. However, with increasingly complex power system operating conditions, CLCC converters face the risk of commutation anomalies or even failures under conditions such as overload operation, critical commutation states, and external AC system disturbances. This poses a direct threat to the safe and stable operation of the power system. Therefore, real-time monitoring of the operating status of CLCC converters and effective early fault warning have become one of the key issues urgently needing to be addressed in the field of DC transmission technology.

[0003] Currently, research on converter operation status analysis and fault handling has made some progress. On the one hand, data-driven fault diagnosis methods are widely used. These methods typically utilize artificial intelligence algorithms such as deep learning to extract features and recognize patterns from multi-dimensional time-series data during converter operation, thereby identifying the fault type and location. On the other hand, some technical approaches focus on predicting and assessing equipment health status. They train multi-source sensor data by constructing machine learning models such as neural networks to output predicted equipment lifespan or fault occurrence probabilities. In addition, some studies take a mechanistic analysis approach, constructing multi-physics simulation models of converters to conduct coupled analysis of electromagnetic and temperature fields, in order to explore the internal physical response laws of various operating parameters during the fault formation process.

[0004] However, data-driven diagnostic methods primarily focus on end-to-end mapping between input operating data and fault categories and probabilities. While they can achieve state identification or risk prediction, they often struggle to explain the specific causes of commutation margin degradation, making it difficult to distinguish whether the increased risk stems from insufficient commutation voltage, valve current transfer lag, trigger control offset, insufficient turn-off angle margin, or accumulated thermal state of valve components. While multiphysics simulation-based mechanistic analysis methods accurately reflect internal physical processes, their high computational complexity limits their application to offline analysis scenarios, making direct integration into online real-time early warning systems difficult. Furthermore, existing methods generally lack a unified description and quantification of the coupling relationships between multiple operating parameters, hindering a systematic assessment of the specific impact of each parameter on converter fault risk. The lack of dynamic parameter sensitivity analysis mechanisms prevents effective identification of core parameters that dominate fault occurrence from numerous operating variables, thus limiting the improvement of early warning efficiency. In addition, most existing methods use fixed evaluation thresholds or static weighting systems, which fail to adaptively and dynamically adjust the early warning strategy according to changes in the actual operating conditions of the converter (such as overload, commutation criticality, or external disturbances). This results in significant deviations and uncertainties in the evaluation results under different scenarios, making it difficult to guarantee the robustness and reliability of the early warning system. Summary of the Invention

[0005] Based on the above analysis, the present invention aims to provide a CLCC converter fault early warning method and system to solve the problems of insufficient interpretability of end-to-end identification results, difficulty in online quantification of commutation margin degradation process, difficulty in dynamic identification of dominant risk parameters, and insufficient adaptability of fixed threshold or static weight evaluation methods to operating conditions in existing CLCC converter fault monitoring.

[0006] On one hand, embodiments of the present invention provide a fault early warning method for CLCC converters, comprising the following steps: The operating parameter data of the CLCC converter during the current warning period are collected and preprocessed to obtain multidimensional time series data; Based on the multidimensional time series and mechanism mapping model, the current fault risk index and the risk contribution of each operating parameter are obtained; and the operating parameters in the multidimensional time series are perturbed, and the perturbed fault risk index is obtained based on the mechanism mapping model. Based on the current fault risk indicators and the fault risk indicators after disturbance, the dynamic sensitivity indicators of each operating parameter are obtained, and then the set of key influencing parameters is obtained based on the dynamic sensitivity indicators. Based on the multidimensional time series data, the current operating conditions are identified, and then the weights of each key influencing parameter are obtained according to each key influencing parameter, its dynamic sensitivity index, and risk contribution. Based on each key influencing parameter and its corresponding weight, a comprehensive fault risk value is obtained, and then based on a preset risk threshold, fault warning information for the current warning period is obtained.

[0007] Furthermore, the operating parameters include commutation voltage, commutation current, firing angle, turn-off angle, current change rate, and temperature; the mechanism mapping model is expressed as: ; In the formula, Indicators representing failure risk This indicates the risk term related to the turn-off angle margin. This represents the multi-parameter normalized bias risk term. , These represent the weighting coefficients of the shut-off angle margin risk term and the multi-parameter normalized bias risk term, respectively.

[0008] Furthermore, the shut-off angle margin risk item Represented as: ; In the formula, Indicates the current shut-off angle. Indicates the dynamic safety shut-off angle. This indicates taking the maximum value; where, the dynamic safety shut-off angle is... Based on the preset minimum safe turn-off angle, dynamic corrections are made in conjunction with commutation voltage offset, commutation current load, commutation current change rate, and temperature.

[0009] Furthermore, applying perturbation to the operating parameters in the multidimensional time series involves performing the following steps for each operating parameter in the current operating parameter vector of the multidimensional time series: Based on the operating parameters, a virtual disturbance is applied to obtain the virtual parameter vector corresponding to the operating parameters. Then, based on the mechanism mapping model, the fault risk index after the operating parameters are disturbed is obtained. The virtual disturbance includes independent disturbance or joint disturbance. The independent disturbance method is to change only the value of the disturbed operating parameter, while the other operating parameters in the virtual parameter vector remain unchanged. The joint disturbance method is to apply a disturbance to the disturbed operating parameter and synchronously correct the operating parameters that are coupled with it according to the coupling constraints.

[0010] Furthermore, in the joint perturbation method: when the first... One operating parameter Applying perturbation At that time, the first one that has a coupling relationship with it One operating parameter According to the coupling coefficient Synchronous correction, its correction amount Represented as: .

[0011] Furthermore, the set of key influencing parameters is obtained through the following method: Based on the current fault risk indicators and the fault risk indicators after disturbance of each operating parameter, calculate the instantaneous sensitivity index of each operating parameter respectively. Based on the instantaneous sensitivity indices of each operating parameter in multiple consecutive early warning cycles, the dynamic sensitivity indices of each operating parameter are obtained. The key influencing parameters are selected by sorting or comparing the values ​​of the dynamic sensitivity indicators of each operating parameter according to their values, and then filtering them to obtain the set of key influencing parameters.

[0012] Furthermore, the instantaneous sensitivity index of each operating parameter is expressed as follows: ; in, ; In the formula, Indicates the first One operating parameter Instantaneous sensitivity index Indicates the first One operating parameter Changes in fault risk indicators caused by disturbances. Indicates the first One operating parameter Fault risk indicators after disturbance. Indicates the first One operating parameter The amount of disturbance applied.

[0013] Furthermore, the operating conditions include normal operating state, overload state, commutation critical state, and disturbance state; the weights of each key influencing parameter are expressed as follows: ; In the formula, Indicates the current operating condition category. Indicates the first The weights of the key influencing parameters , These represent the current operating condition categories. The corresponding number The first and second adjustment coefficients of the key influencing parameters Indicates the first The risk contribution of each key influencing parameter Indicates the first The dynamic sensitivity index corresponding to the key influencing parameters Indicates the number of key influencing parameters; , These represent the current operating condition categories. The corresponding number The first and second adjustment coefficients of the key influencing parameters Indicates the first The risk contribution of each key influencing parameter Indicates the first The dynamic sensitivity index corresponding to the key influencing parameters.

[0014] Furthermore, the comprehensive failure risk value Represented as: ; In the formula, , , These represent the linear weighting term, the nonlinear coupling term, and the risk trend term, respectively; where, The risk trend item Represented as: ; In the formula, This represents the weighting coefficient of the trend term. Indicates the length of the trend judgment window. Indicates the current warning period Fault risk indicators Indicates the current warning period The first Fault risk indicators for each early warning cycle.

[0015] On the other hand, the present invention provides a CLCC converter fault early warning system, comprising: The data acquisition and preprocessing module is used to acquire the operating parameter data of the CLCC converter during the current warning period and perform preprocessing to obtain multi-dimensional time series data. The fault risk assessment module is used to obtain the current fault risk index and the risk contribution of each operating parameter based on the multidimensional time series and mechanism mapping model; and to apply perturbation to the operating parameters in the multidimensional time series, and then obtain the perturbed fault risk index based on the mechanism mapping model. The sensitivity analysis module is used to obtain the dynamic sensitivity index of each operating parameter based on the current fault risk index and the fault risk index after disturbance, and then obtain the set of key influencing parameters based on the dynamic sensitivity index. The working condition identification and weight adaptation module is used to identify the current operating condition based on the multi-dimensional time series data, and then obtain the weight of each key influencing parameter according to each key influencing parameter, its dynamic sensitivity index, and risk contribution. The fault warning generation module is used to obtain a comprehensive fault risk value based on each key influencing parameter and its corresponding weight, and then obtain fault warning information for the current warning period based on a preset risk threshold.

[0016] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: This invention provides a fault early warning method and system for CLCC converters. 1. By constructing a fault risk model based on multi-parameter coupling relationship, online fault early warning and advanced risk identification during the operation of CLCC converter were realized. Compared with traditional post-fault analysis methods, it can identify potential fault risks in advance. 2. By conducting sensitivity analysis on operating parameters, key parameters that significantly impact fault risk can be identified, improving the interpretability of fault warning results; 3. By introducing an operating condition identification mechanism and adaptively adjusting the weights of key parameters according to different operating conditions, the adaptability of the early warning model in complex operating environments has been improved. 4. The results of multi-parameter coupling mechanism analysis are transformed into quantifiable fault risk indicators, and combined with threshold judgment to realize graded early warning, which improves the engineering application value and practicality of the method.

[0017] In this invention, the above-described technical solutions can be combined with each other to achieve more preferred combinations. Other features and advantages of this invention will be set forth in the following description, and some advantages may become apparent from the description or be learned by practicing the invention. The objects and other advantages of this invention can be realized and obtained from what is particularly pointed out in the description and drawings. Attached Figure Description

[0018] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts.

[0019] Figure 1 This is a flowchart illustrating the CLCC converter fault early warning method provided in Embodiment 1 of the present invention. Detailed Implementation

[0020] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0021] Example 1 A specific embodiment of the present invention discloses a fault early warning method for CLCC converters, such as... Figure 1 As shown, it includes the following steps: S1. Collect the operating parameter data of the CLCC converter during the current warning period, and preprocess it to obtain multi-dimensional time series data.

[0022] Specifically, the operating parameters include commutation voltage, commutation current, firing angle, turn-off angle, current change rate, and temperature. The data of each parameter constitute an operating parameter vector. Among them, the commutation voltage, commutation current, and temperature are directly acquired, the firing angle is a control quantity read by the trigger control unit, and the turn-off angle and current change rate are commutation characteristic quantities calculated based on the commutation voltage, commutation current, and triggering time. The sampling frequency of the commutation voltage, commutation current, firing angle, turn-off angle, and commutation current change rate is not less than 10kHz, and the sampling frequency of temperature is not less than 10Hz.

[0023] Specifically, preprocessing includes denoising, outlier removal, normalization, and time series alignment; among which, The noise reduction process employs low-pass filtering or moving average methods. Specifically, short-window filtering with a window length of 5 to 20 sampling points or low-pass filtering with a cutoff frequency of 500 Hz to 1 kHz is used for commutation voltage, commutation current, firing angle, turn-off angle, and commutation current change rate. Long-window smoothing with a window length of 50 to 200 sampling points is used for temperature. Outlier removal employs an anomaly detection method based on statistical distribution. Sampling points that deviate from the mean by more than 3 times the standard deviation within the sliding time window are identified as outliers and removed or corrected. Amplitude limiting verification is performed on abnormal data with a turn-off angle outside the range of 0° to 180° and a commutation current change rate exceeding the rated change rate range. Normalization employs minimum-maximum normalization or standard deviation normalization methods to unify all operating parameters into dimensionless values; Time series alignment is based on timestamp alignment using a unified time reference, which is the trigger moment, the start moment of commutation, or the inspiration moment of the commutation cycle. A sliding time window is used to segment the data.

[0024] Specifically, the warning period is a preset fixed time interval, which is used to control the triggering timing of two consecutive executions of the method in this embodiment.

[0025] Specifically, the multidimensional time series data is constructed based on sampling time or commutation period. If it is constructed based on sampling time, the multidimensional time series data includes the operating parameter vectors of the current sampling time and the previous N historical sampling times. If it is constructed based on commutation period, the multidimensional time series includes the operating parameter vectors corresponding to the current complete commutation period and the previous N historical complete commutation periods.

[0026] More specifically, N is determined based on the ratio of the transient process duration of the CLCC converter to the sampling period, and N ranges from 5 to 20; the value of N ensures that the total time span of the multidimensional time series covers the complete transient process duration of the CLCC converter.

[0027] More specifically, if based on the sampling time, when the current warning period is triggered, the latest written running parameter vector is extracted from the data cache as the running parameter vector of the current sampling time, and the N running parameter vectors written consecutively before are extracted as the running parameter vectors of the historical sampling times. The running parameter vector of the current sampling time and the running parameter vectors of the N historical sampling times are arranged in chronological order of sampling time to jointly constitute multidimensional time series data.

[0028] If based on the commutation cycle, when the current warning cycle is triggered, the operating parameter vector corresponding to the most recent complete commutation cycle that ended before the trigger time is extracted from the data cache as the operating parameter vector of the current commutation cycle, and the operating parameter vectors corresponding to the previous N consecutive complete commutation cycles are extracted as the operating parameter vectors of the historical commutation cycles. The operating parameter vector of the current commutation cycle and the operating parameter vectors of the N historical commutation cycles are arranged in chronological order of the commutation cycles, together forming multidimensional time series data. Among them, the operating parameter vector corresponding to each commutation cycle is composed of the average value of the operating parameters at each sampling time within the commutation cycle or the instantaneous value at the end of the commutation. The running parameter vectors in the data cache are written sequentially according to the sampling time or commutation cycle, and the writing frequency is not lower than the trigger frequency of the warning cycle, so as to ensure that there are at least N+1 valid running parameter vectors in the cache when the warning cycle is triggered.

[0029] S2. Based on the multidimensional time series and mechanism mapping model, obtain the current fault risk index and the risk contribution of each operating parameter; and apply a perturbation to the operating parameters in the multidimensional time series, and then obtain the perturbed fault risk index based on the mechanism mapping model.

[0030] Specifically, the mechanism mapping model establishes a mapping relationship between operating parameters and fault risks based on the commutation mechanism of CLCC converters, used to obtain fault risk indicators from multi-dimensional time series data. Specifically, the commutation margin of a CLCC converter is determined by the commutation voltage-time area. During commutation, the integral area of ​​the commutation voltage within the commutation overlap angle must be sufficient for the thyristor to complete carrier extraction and restore forward blocking capability. If this area is insufficient, commutation failure occurs. Therefore, a mechanism mapping model is established to characterize the impact of changes in various operating parameters on the converter's operating state and fault risks. The fault risk indicator is the commutation margin degradation risk, jointly characterized by commutation anomaly risk, voltage stability risk, and operating state anomaly risk.

[0031] In practice, the mechanism mapping model is represented as: ; In the formula, Indicators representing failure risk This indicates the risk term related to the turn-off angle margin. This represents the multi-parameter normalized bias risk term. , These represent the weighting coefficients of the shut-off angle margin risk term and the multi-parameter normalized bias risk term, respectively.

[0032] Understandably, fault risk indicators It is a dimensionless, non-negative value; the larger the value, the higher the risk of commutation margin degradation. This indicates that all operating parameters are within a safe range and there is no risk of commutation margin degradation.

[0033] Specifically, the weighting coefficients of the shut-off angle margin risk term and the multi-parameter normalized bias risk term satisfy... The weighting coefficients are set based on the degree to which the turn-off angle margin determines the success of commutation in the commutation physical mechanism of the CLCC converter. The weight of the turn-off angle margin risk term is... The weight of the deviation risk term is set to 0.6~0.8 for multi-parameter normalization. Taking a value of 0.2 to 0.4 allows the fault risk index to be dominated by the turn-off angle margin and supplemented by multiple parameter deviations, ensuring both sensitive response to the critical state of commutation and early detection of abnormal degradation of other parameters.

[0034] In practical implementation, the shut-off angle margin risk item Represented as: ; In the formula, Indicates the current shut-off angle. Indicates the dynamic safety shut-off angle. This indicates taking the maximum value; where the current shut-off angle γ is the shut-off angle value corresponding to the latest timestamp of the running parameter vector in the multidimensional time series data.

[0035] In practical implementation, the dynamic safety shut-off angle Based on the preset minimum safe turn-off angle, dynamic corrections are made in conjunction with commutation voltage offset, commutation current load, commutation current change rate, and temperature.

[0036] Specifically, dynamic safety shut-off angle Represented as: ; in, ; ; ; ; In the formula, Indicates the minimum safe shut-off angle. , , , These represent the commutation voltage offset correction term, commutation current load correction term, commutation current change rate correction term, and temperature correction term, respectively. , , , These represent the sensitivity coefficients for the commutation voltage offset correction term, commutation current load correction term, commutation current rate of change correction term, and temperature correction term, respectively. Indicates the commutation reference voltage. Indicates the commutation rated current. This indicates the rate of change of the commutation rated current. Indicates reference temperature. Indicates the maximum permissible temperature. This represents the commutation current load factor threshold value. , , , These represent the current commutation voltage, commutation current, current rate of change, and temperature, respectively. The current commutation voltage, commutation current, current rate of change, and temperature are the commutation voltage, commutation current, current rate of change, and temperature values ​​corresponding to the latest timestamp of the operating parameter vector in the multidimensional time series data.

[0037] More specifically, minimum safe shut-off angle Determined based on the minimum turn-off angle required to restore blocking capability in the physical characteristics of the thyristor; more specifically, the minimum safe turn-off angle. Take 7°~10°.

[0038] More specifically, the sensitivity coefficient of the commutation voltage offset correction term. Based on the principle of commutation voltage-time area integration, the adjustment range of the turn-off angle margin by the voltage drop is controlled; preferably, Take a value of 1.0 to 3.0. Sensitivity coefficient for the commutation current load correction term. and commutation current load factor threshold Based on the physical relationship between the commutation overlap angle and the increase of current, the margin requirement is automatically increased under heavy load; preferably, Take a value of 0.5~2.0. Take a value of 0.6 to 0.8. Sensitivity coefficient for the commutation current rate of change correction term. Based on the effect of inductor-induced electromotive force in the commutation circuit, a margin is added in advance during transient disturbances; preferably... Take a value of 0.1 to 0.5. Sensitivity coefficient for the temperature correction term. Based on the growth curve of thyristor turn-off time with increasing temperature, the margin is increased more rapidly at high temperatures, thereby enabling... It can comprehensively reflect the differentiated impacts of various extreme operating conditions such as voltage dips, overloads, transient disturbances, and high temperatures on commutation margins, ensuring sufficient dynamic safety margins under all operating conditions to mitigate the risk of commutation failure; preferably, Take a value of 0.2 to 1.0.

[0039] In practical implementation, the multi-parameter normalization bias risk item Represented as: ; In the formula, Indicates the number of runtime parameters. Indicates the first The weighting coefficients of each operating parameter. Indicates the first The current values ​​of each running parameter, Indicates the first Reference values ​​for each running parameter. This represents the absolute value. Where, the th... The current value of the current running parameter is the value corresponding to the running parameter vector with the latest timestamp in the multidimensional time series data. The values ​​of the running parameters.

[0040] Specifically, the reference values ​​for each operating parameter are determined based on the rated operating parameters, control setpoints, or historical stable operating data. More specifically, the reference values ​​for commutation voltage and commutation current are taken as the rated values; the reference value for the firing angle is taken as the setpoint of the control system; the reference value for the turn-off angle is taken as the dynamic safe turn-off angle; the reference value for the rate of change of commutation current is taken as the theoretically calculated value under rated operating conditions or the average value within a historical stable operating period; and the reference value for temperature is taken as the median of the normal operating temperature range of the equipment. When using historical stable operating data to determine the reference values, a data segment of the CLCC converter with a load rate of 0.8 to 1.0, continuous operation time of not less than 24 hours, and no fault records is selected, and the arithmetic mean of each parameter within this data segment is taken as the corresponding reference value.

[0041] Specifically, the weighting coefficients of each operating parameter are determined comprehensively based on the parameter's basic weight, current operating conditions, and sensitivity results, characterizing the relative severity of each parameter deviation on commutation margin degradation. For example, the weighting coefficient for the turn-off angle is set to 0.30, the weighting coefficient for the commutation voltage to 0.20, the weighting coefficient for the firing angle to 0.15, the weighting coefficient for the commutation current to 0.15, the weighting coefficient for the commutation current change rate to 0.12, and the weighting coefficient for temperature to 0.08, ensuring that the sum of the weighting coefficients for all operating parameters is 1. These weighting coefficients can be readjusted before implementation of this method based on the specific CLCC converter parameters and operating conditions, and can be adjusted according to different stages of equipment operation.

[0042] Understandably, the multi-parameter normalized deviation risk term characterizes the overall degree of deviation of each operating parameter from its normal reference value; when the shut-off angle is still within the safe range but other operating parameters have already deviated abnormally, the multi-parameter normalized deviation risk term can capture potential risks in advance.

[0043] In practice, the risk contribution rate allocates the fault risk index according to the normalized deviation ratio of each operating parameter, representing the share of each parameter's deviation from the reference value in the total risk; and the sum of the risk contribution rates of all operating parameters equals the current fault risk index; the larger the risk contribution rate of a parameter, the higher its real-time hazard to the current risk.

[0044] Specifically, the risk contribution of each operating parameter is expressed as follows: ; In the formula, Indicates the first Risk contribution of each operating parameter Indicates the first The weighting coefficients of each operating parameter. Indicates the first The current values ​​of each running parameter, Indicates the first Reference values ​​for each running parameter.

[0045] During implementation, a perturbation is applied to the operating parameters in the multidimensional time series. This involves performing the following steps for each operating parameter in the current operating parameter vector of the multidimensional time series: Based on the operating parameters, a virtual disturbance is applied to obtain the virtual parameter vector corresponding to the operating parameters. Then, based on the mechanism mapping model, the fault risk index after the operating parameters are disturbed is obtained. The virtual disturbance includes independent disturbance methods or joint disturbance methods. The independent disturbance method changes only the value of the disturbed operating parameter, while the other operating parameters in the virtual parameter vector remain unchanged. The joint disturbance method applies a disturbance to the disturbed operating parameter and synchronously corrects the operating parameters that are coupled with it according to coupling constraints.

[0046] Specifically, for different operating parameters, either an independent perturbation method or a joint perturbation method is selected to perform virtual perturbation; when there is a physical coupling relationship between the operating parameter and other operating parameters, the joint perturbation method is used; when there is no physical coupling relationship between the operating parameter and other operating parameters, the independent perturbation method is used.

[0047] More specifically, under the joint disturbance mode, the change in the synchronously corrected associated operating parameters is attributed to the disturbed operating parameters, and the resulting post-disturbance fault risk index corresponds to the operating parameters, rather than to the synchronously corrected associated operating parameters.

[0048] More specifically, when an operating parameter has a necessary causal relationship or synchronous change relationship with other operating parameters based on the switching mechanism, circuit equations or physical laws of the CLCC converter, it is determined that there is a physical coupling relationship between the operating parameter and other operating parameters.

[0049] Preferably, in this embodiment, based on the commutation mechanism of the CLCC converter, the following parameter combinations are determined to have a physical coupling relationship: commutation voltage and commutation current change rate, commutation current and commutation current change rate, commutation current and temperature, and firing angle and turn-off angle.

[0050] In practice, the independent disturbance method adopts: The disturbance amount is set according to the set ratio of the corresponding operating parameter's rated value, and applied to the operating parameter in the current operating parameter vector to obtain the virtual parameter vector; Alternatively, the disturbance amount can be set according to the historical fluctuation range of the operating parameter and applied to the operating parameter in the current operating parameter vector to obtain a virtual parameter vector.

[0051] Preferably, the disturbance amount is set according to the rated value ratio for commutation voltage, firing angle and turn-off angle; the disturbance amount is set according to the historical fluctuation range for commutation current, commutation current change rate and temperature.

[0052] Specifically, the ratio is set at 1% to 10% of the rated value.

[0053] Specifically, the disturbance amount is set according to the historical fluctuation range of this operating parameter as follows: Obtain historical operating data for the operating parameter within a preset historical time window, and obtain the standard deviation of the historical operating data. Multiply the standard deviation of the historical operating data by the preset scaling factor to obtain the disturbance of the operating parameter.

[0054] Specifically, the preset proportional coefficient is set to 0.1 to 0.5, depending on the parameter type and operating conditions. For parameters that change frequently (such as voltage and current), a smaller value (such as 0.1 to 0.2) is used to avoid excessive disturbances that could cause the virtual parameter vector to deviate from the physical reasonable range. For parameters that change gradually (such as temperature), a larger value (such as 0.3 to 0.5) can be used to ensure that the disturbance produces an observable response.

[0055] Specifically, the preset historical time window is the continuous running data of the most recent 24 hours.

[0056] In specific implementation, the joint perturbation method is as follows: when the first... One operating parameter Applying perturbation At that time, the first one that has a coupling relationship with it One operating parameter According to the coupling coefficient Synchronous correction, its correction amount Represented as: .

[0057] Specifically, coupling coefficient The coupling coefficient is determined based on the correlation of historical operating data, the switching mechanism model, or engineering experience. More specifically, for the coupling relationship between commutation voltage and commutation current change rate, commutation current and commutation current change rate, and firing angle and turn-off angle, the coupling coefficient is taken as 0.5 to 1.0; for the coupling relationship between commutation current and temperature, the coupling coefficient is taken as 0.1 to 0.3.

[0058] S3. Based on the current fault risk indicators and the fault risk indicators after disturbance, obtain the dynamic sensitivity indicators of each operating parameter, and then obtain the set of key influencing parameters based on the dynamic sensitivity indicators.

[0059] During implementation, the set of key impact parameters is obtained through the following methods: S31. Based on the current fault risk index and the fault risk index after disturbance of each operating parameter, calculate the instantaneous sensitivity index of each operating parameter.

[0060] In practical implementation, the instantaneous sensitivity index of each operating parameter is expressed as follows: ; in, ; In the formula, Indicates the first One operating parameter Instantaneous sensitivity index Indicates the first One operating parameter Changes in fault risk indicators caused by disturbances. Indicates the first One operating parameter Fault risk indicators after disturbance. Indicates the first One operating parameter The amount of disturbance applied.

[0061] S32. Based on the instantaneous sensitivity index of each operating parameter in multiple consecutive early warning cycles, the dynamic sensitivity index of each operating parameter is obtained.

[0062] In practice, the instantaneous sensitivity of each operating parameter for M consecutive early warning cycles is statistically processed to obtain the dynamic sensitivity index; wherein, the value of M is determined according to the ratio of the transient process duration of the CLCC converter to the sampling period or commutation period, and preferably M is 5 to 20.

[0063] Specifically, the statistical processing employs the average, maximum, or time-decay weighted method. It is understandable that statistical processing reduces the impact of single disturbance fluctuations on the screening results, improving the stability of the screening of key influencing parameters.

[0064] S33. Sort or compare the values ​​of the dynamic sensitivity indicators of each operating parameter to obtain a set of key influencing parameters.

[0065] Specifically, the dynamic sensitivity index of each operating parameter is normalized, and the operating parameters whose normalized dynamic sensitivity index is greater than or equal to the preset sensitivity threshold are added to the key influencing parameter set; or, all operating parameters are sorted from largest to smallest according to their dynamic sensitivity index, and the top K operating parameters are selected to form the key influencing parameter set.

[0066] More specifically, the preset sensitivity threshold is determined based on the sensitivity distribution in historical stable operation samples, equipment operation experience, or sensitivity ranking results, and is set to 0.2 to 0.3; K is 2 to 3, in order to exclude low-contribution parameters and improve the stability and interpretability of the screening of key impact parameters.

[0067] It should be noted that when the fault risk index is 0, sensitivity calculation is not performed, and the sensitivity of all operating parameters is directly determined to be zero, and the set of key influencing parameters is an empty set.

[0068] S4. Identify the current operating conditions based on the multidimensional time series data, and then obtain the weights of each key influencing parameter according to each key influencing parameter, its dynamic sensitivity index, and risk contribution.

[0069] In specific implementation, the operating conditions include normal operating state, overload state, commutation critical state, and disturbance state; among them, the overload state is determined based on the commutation current, the commutation critical state is determined based on the turn-off angle, and the disturbance state is determined based on the commutation voltage fluctuation amplitude, commutation current fluctuation amplitude, or commutation current change rate.

[0070] Specifically, the criteria for determining an overload state are as follows: like If the condition is such that the load is too high, then it is considered an overload condition; among which, Indicates the current commutation current. Indicates the rated commutation current. This indicates the preset overload threshold.

[0071] More specifically, overload threshold Based on the rated current parameters of the CLCC converter, the control and protection settings, and the safety limits in the operating procedures, the optimal value is 0.9 to 1.0.

[0072] Specifically, the criteria for determining the critical state of commutation are as follows: If the current shut-off angle is less than the dynamic safety shut-off angle, or the difference between the dynamic safety shut-off angle and the current shut-off angle is greater than or equal to the preset margin threshold, then it is judged to be in a critical commutation state.

[0073] More specifically, the commutation margin threshold is set based on the minimum turn-off angle required by the thyristor and a certain safety margin is added; preferably, it can be set to 3° to 5°.

[0074] Specifically, the criteria for determining the disturbance state are as follows: If the fluctuation range of the current commutation voltage within the time window is greater than or equal to the preset voltage disturbance threshold, or the fluctuation range of the current commutation current within the time window is greater than or equal to the preset current disturbance threshold, or the rate of change of the current commutation current is greater than or equal to the preset rate of change of current threshold, then it is judged as a disturbance state.

[0075] More specifically, the voltage disturbance threshold and current disturbance threshold are set according to the sensitivity of the CLCC converter to AC system faults and the disturbance rejection capability of the control and protection system; preferably, the voltage disturbance threshold and current disturbance threshold are taken as 5% to 10% of the corresponding rated values. The current change rate threshold is set according to the maximum change rate of the commutation current and the safety margin during normal operation of the CLCC converter; preferably, it is taken as 1.2 to 1.5 times the rated current change rate.

[0076] Specifically, when the conditions for overload, commutation criticality, and disturbance are not met, the system is judged to be in normal operating condition.

[0077] In practice, when the overload state, commutation critical state, and disturbance state are simultaneously met, the operating condition category is determined in order of priority: commutation critical state, disturbance state, overload state, and normal operating state. In other words, if multiple criteria are met at the same time, the state with the highest priority is taken as the current operating condition.

[0078] Understandably, the commutation critical state has the highest priority because insufficient turn-off angle directly threatens the commutation safety of the CLCC converter.

[0079] During implementation, the weights of each key influencing parameter are expressed as follows: ; In the formula, Indicates the current operating condition category. Indicates the first The weights of the key influencing parameters , These represent the current operating condition categories. The corresponding number The first and second adjustment coefficients of the key influencing parameters Indicates the first The risk contribution of each key influencing parameter Indicates the first The dynamic sensitivity index corresponding to the key influencing parameters Indicates the number of key influencing parameters; , These represent the current operating condition categories. The corresponding number The first and second adjustment coefficients of the key influencing parameters Indicates the first The risk contribution of each key influencing parameter Indicates the first The dynamic sensitivity index corresponding to the key influencing parameters.

[0080] In practice, the first and second adjustment coefficients are set according to the following rules based on the current operating conditions: Under commutation critical conditions, the first and second adjustment coefficients of the turn-off angle, commutation voltage, and firing angle are taken to be greater than 1, while the adjustment coefficients of other parameters are taken to be less than or equal to 1. Under overload conditions, the first and second adjustment coefficients for commutation current, commutation current change rate, and temperature are taken to be greater than 1, while the adjustment coefficients for other parameters are taken to be less than or equal to 1. Under disturbance conditions, the first and second adjustment coefficients for commutation voltage, commutation current, and commutation current change rate are taken to be greater than 1, while the adjustment coefficients for other parameters are taken to be less than or equal to 1. Under normal operating conditions, the first and second adjustment coefficients of all parameters are both set to 1.

[0081] Specifically, under the critical commutation state, the first and second adjustment coefficients of the turn-off angle are taken as 1.5 to 2.0, the first and second adjustment coefficients of the commutation voltage are taken as 1.2 to 1.5, and the first and second adjustment coefficients of the firing angle are taken as 1.0 to 1.3; Under overload conditions, the first and second adjustment coefficients of the commutation current are taken as 1.5 to 2.0, and the first and second adjustment coefficients of the commutation current change rate and temperature are taken as 1.2 to 1.5. Under disturbance conditions, the first and second adjustment coefficients of the commutation voltage are taken as 1.5 to 2.0, and the first and second adjustment coefficients of the commutation current and the rate of change of the commutation current are taken as 1.3 to 1.8.

[0082] It should be noted that when the set of key impact parameters is empty, the weight of each key impact parameter is set to zero.

[0083] S5. Based on each key influencing parameter and its corresponding weight, a comprehensive fault risk value is obtained, and then based on a preset risk threshold, fault warning information for the current warning period is obtained.

[0084] During implementation, the comprehensive failure risk value Represented as: ; In the formula, , , These represent the linear weighting term, the nonlinear coupling term, and the risk trend term, respectively.

[0085] In practical implementation, the linear weighting term Represented as: ; In the formula, Indicates the first Key Influencing Parameters The corresponding risk contribution function maps the current value of each operating parameter to a normalized risk contribution value. The more the parameter deviates from the safe or normal range, the larger the function value.

[0086] Specifically, the risk contribution function can be implemented in different forms, such as linear mapping, threshold mapping, or exponential mapping, based on the physical characteristics and safety boundaries of each parameter. The function value should be such that when the parameter is within the safe or normal range, the function value tends to 0, and the larger the parameter deviates from the safe or normal range, the more monotonically the function value tends to 1.

[0087] For example, for the shut-off angle, using a margin-based mapping, the risk contribution function is expressed as: ; For the commutation voltage, using a deviation-based mapping, the risk contribution function is expressed as: ; For the commutation current, using a mapping with dead time, the risk contribution function is expressed as: ; In the formula, The dead zone threshold value of the commutation current is determined based on the short-time overload capacity of the CLCC converter. It is preferably taken as 0.8 to 0.9 times the rated current to make use of the equipment's allowable overload margin and avoid normal load fluctuations being misjudged as risks. For temperature, using an accelerated mapping, the risk contribution function is expressed as: ; In the formula, This represents the dead zone threshold value of the commutation current, which is determined based on the upper limit of the normal operating temperature of the CLCC converter.

[0088] In practical implementation, nonlinear coupling terms Represented as: ; In the formula, Indicates the first The key influencing parameter and the first The coupling weights between key influencing parameters are used to characterize the amplification effect of simultaneous changes in two parameters on the overall failure risk.

[0089] Specifically, the coupling weight is determined based on the switching mechanism of the CLCC converter, the correlation coefficient of historical operating data, or the consistency of sensitivity changes. For parameter combinations with physical coupling relationships, a coupling weight greater than 0 is set; for parameter combinations without physical coupling relationships, the coupling weight is set to 0. Preferably, the coupling weight for parameter combinations with physical coupling relationships is between 0.5 and 1.0.

[0090] In practice, risk trend items Represented as: ; In the formula, This represents the weighting coefficient of the trend term. Indicates the length of the trend judgment window. Indicates the current warning period Fault risk indicators Indicates the current warning period The first Fault risk indicators for each early warning cycle.

[0091] Specifically, the trend weighting coefficient is used to control the amplification of the overall risk by the upward trend of risk. If the value is too large, it is easily affected by noise interference, leading to false alarms; if it is too small, the trend warning effect is not obvious. It is adjusted comprehensively based on historical operating data and anti-interference requirements, and preferably ranges from 0.1 to 0.5. The trend judgment window length is used to determine how long of historical data to observe to judge whether the risk continues to rise. If the value is too short, it is easily affected by instantaneous fluctuations; if it is too long, the response is lagging. It is set according to the typical time scale of CLCC converter fault evolution and the warning lead time requirement, and preferably ranges from 3 to 10 warning cycles.

[0092] Preferably, if the number of historical fault risk indicators is less than At any given time, the risk trend item is 0. The risk trend item calculation will be enabled after enough historical data has been accumulated.

[0093] Understandably, the introduction of the risk trend item allows the early warning system to focus not only on the current level of risk but also on how rapidly the risk is changing. Even if the current risk value has not yet exceeded the high-level threshold, if it is rising rapidly, the risk trend item can provide a higher-level warning in advance, thus achieving proactive early warning.

[0094] Specifically, the fault warning information includes: comprehensive fault risk value, risk level, and at least one key influencing parameter with the greatest risk contribution and its dynamic sensitivity value.

[0095] Specifically, the risk levels, in descending order of urgency, include normal status, Level 1 warning signal, Level 2 warning signal, and Level 3 warning signal.

[0096] During implementation, the risk thresholds include a first risk threshold, a second risk threshold, and a third risk threshold, with the first risk threshold being lower than the second risk threshold, and the second risk threshold being lower than the third risk threshold; fault warning information is output in a tiered manner according to the following rules: When the overall fault risk value is less than the first risk threshold, it is determined to be in a normal state and no warning signal is output. When the overall fault risk value is greater than or equal to the first risk threshold and less than the second risk threshold, a first-level early warning signal is output. When the overall fault risk value is greater than or equal to the second risk threshold and less than the third risk threshold, a level two early warning signal is output. When the comprehensive fault risk value is greater than or equal to the third risk threshold, a level three early warning signal is output.

[0097] Specifically, the first risk threshold, the second risk threshold, and the third risk threshold are set based on the CLCC converter's operating procedures, safety margins, historical stable operating data, or historical fault samples.

[0098] Preferably, when the comprehensive fault risk value continues to rise over multiple consecutive warning periods, the warning level is raised, such as by one level; when the comprehensive fault risk value is lower than the exit threshold corresponding to the current level for multiple consecutive warning periods, the warning level is lowered, such as by one level.

[0099] It should be noted that steps S1 to S5 are executed online in a loop according to the preset warning cycle. In each warning cycle, the warning information is recalculated and updated based on the updated multidimensional time series data.

[0100] Compared with existing technologies, the fault early warning method for CLCC converters provided in this embodiment realizes online fault early warning and advanced risk identification during the operation of CLCC converters by constructing a fault risk model based on multi-parameter coupling relationships. Compared with traditional post-fault analysis methods, it can identify potential fault risks in advance. By performing sensitivity analysis on operating parameters, it can identify key parameters that significantly affect fault risks, improving the interpretability of fault early warning results. By introducing an operating condition identification mechanism and adaptively adjusting the weights of key parameters according to different operating conditions, it improves the adaptability of the early warning model in complex operating environments. The analysis results of multi-parameter coupling mechanisms are transformed into quantifiable fault risk indicators, and combined with threshold judgment to achieve hierarchical early warning, improving the engineering application value and practicality of the method.

[0101] Example 2 A specific embodiment of the present invention discloses a CLCC converter fault early warning system, comprising: The data acquisition and preprocessing module is used to acquire the operating parameter data of the CLCC converter during the current warning period and perform preprocessing to obtain multi-dimensional time series data. The fault risk assessment module is used to obtain the current fault risk index and the risk contribution of each operating parameter based on the multidimensional time series and mechanism mapping model; and to apply perturbation to the operating parameters in the multidimensional time series, and then obtain the perturbed fault risk index based on the mechanism mapping model. The sensitivity analysis module is used to obtain the dynamic sensitivity index of each operating parameter based on the current fault risk index and the fault risk index after disturbance, and then obtain the set of key influencing parameters based on the dynamic sensitivity index. The working condition identification and weight adaptation module is used to identify the current operating condition based on the multi-dimensional time series data, and then obtain the weight of each key influencing parameter according to each key influencing parameter, its dynamic sensitivity index, and risk contribution. The fault warning generation module is used to obtain a comprehensive fault risk value based on each key influencing parameter and its corresponding weight, and then obtain fault warning information for the current warning period based on a preset risk threshold.

[0102] The specific implementation process of this invention can be found in the above method embodiments, and will not be repeated here.

[0103] Since this embodiment is based on the same principle as the above method embodiments, this system also has the corresponding technical effects of the above method embodiments.

[0104] Example 3 One specific embodiment of the present invention discloses an electronic device, comprising: The processor and memory, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method steps of Embodiment 1, including: collecting and preprocessing CLCC converter operating parameter data; constructing a mapping model between operating parameters and fault risks; calculating the sensitivity index of each operating parameter based on parameter disturbances and screening key influencing parameters; identifying the current operating condition and adaptively adjusting the weights of key influencing parameters; calculating a comprehensive fault risk value based on key influencing parameters and their weights and outputting early warning information.

[0105] Example 4 A specific embodiment of the present invention discloses a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it is used to implement the method steps of Embodiment 1, including: constructing a fault risk model based on the operating mechanism of a CLCC converter; performing disturbance analysis on operating parameters and calculating sensitivity indices; screening key influencing parameters and constructing parameter weight functions; adaptively adjusting the weights of key influencing parameters according to operating conditions; and conducting fault risk assessment based on the key influencing parameters and their weights and outputting early warning results.

[0106] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0107] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A fault early warning method for a CLCC converter, characterized in that, Includes the following steps: The operating parameter data of the CLCC converter during the current warning period are collected and preprocessed to obtain multidimensional time series data; Based on the multidimensional time series and mechanism mapping model, the current failure risk indicators and the risk contribution of each operating parameter are obtained. The operating parameters in the multidimensional time series are perturbed, and the fault risk index after perturbation is obtained based on the mechanism mapping model. Based on the current fault risk indicators and the fault risk indicators after disturbance, the dynamic sensitivity indicators of each operating parameter are obtained, and then the set of key influencing parameters is obtained based on the dynamic sensitivity indicators. Based on the multidimensional time series data, the current operating conditions are identified, and then the weights of each key influencing parameter are obtained according to each key influencing parameter, its dynamic sensitivity index, and risk contribution. Based on each key influencing parameter and its corresponding weight, a comprehensive fault risk value is obtained, and then based on a preset risk threshold, fault warning information for the current warning period is obtained.

2. The CLCC converter fault early warning method according to claim 1, characterized in that, The operating parameters include commutation voltage, commutation current, firing angle, turn-off angle, current change rate, and temperature; the mechanism mapping model is expressed as: ； In the formula, Indicators representing failure risk This indicates the risk term related to the turn-off angle margin. This represents the multi-parameter normalized bias risk term. , These represent the weighting coefficients of the shut-off angle margin risk term and the multi-parameter normalized bias risk term, respectively.

3. The CLCC converter fault early warning method according to claim 2, characterized in that, The shut-off angle margin risk item Represented as: ； In the formula, Indicates the current shut-off angle. Indicates the dynamic safety shut-off angle. This indicates taking the maximum value; where, the dynamic safety shut-off angle is... Based on the preset minimum safe turn-off angle, dynamic corrections are made in conjunction with commutation voltage offset, commutation current load, commutation current change rate, and temperature.

4. The CLCC converter fault early warning method according to claim 1, characterized in that, Applying perturbation to the operating parameters in the multidimensional time series involves performing the following steps for each operating parameter in the current operating parameter vector of the multidimensional time series: Based on the operating parameters, a virtual disturbance is applied to obtain the virtual parameter vector corresponding to the operating parameters. Then, based on the mechanism mapping model, the fault risk index after the operating parameters are disturbed is obtained. The virtual disturbance includes independent disturbance or joint disturbance. The independent disturbance method is to change only the value of the disturbed operating parameter, while the other operating parameters in the virtual parameter vector remain unchanged. The joint disturbance method is to apply a disturbance to the disturbed operating parameter and synchronously correct the operating parameters that are coupled with it according to the coupling constraints.

5. The CLCC converter fault early warning method according to claim 4, characterized in that, In the aforementioned joint perturbation method: when the first... One operating parameter Applying perturbation At that time, the first one that has a coupling relationship with it One operating parameter According to the coupling coefficient Synchronous correction, its correction amount Represented as: 。 6. The CLCC converter fault early warning method according to claim 2, characterized in that, The set of key influencing parameters is obtained through the following methods: Based on the current fault risk indicators and the fault risk indicators after disturbance of each operating parameter, calculate the instantaneous sensitivity index of each operating parameter respectively. Based on the instantaneous sensitivity indices of each operating parameter in multiple consecutive early warning cycles, the dynamic sensitivity indices of each operating parameter are obtained. The key influencing parameters are selected by sorting or comparing the values ​​of the dynamic sensitivity indicators of each operating parameter according to their values, and then filtering them to obtain the set of key influencing parameters.

7. The CLCC converter fault early warning method according to claim 6, characterized in that, The instantaneous sensitivity index of each operating parameter is expressed as follows: ； in, ; In the formula, Indicates the first One operating parameter Instantaneous sensitivity index Indicates the first One operating parameter Changes in fault risk indicators caused by disturbances. Indicates the first One operating parameter Fault risk indicators after disturbance. Indicates the first One operating parameter The amount of disturbance applied.

8. The CLCC converter fault early warning method according to claim 1, characterized in that, The operating conditions include normal operation, overload, commutation criticality, and disturbance; the weights of each key influencing parameter are expressed as follows: ； In the formula, Indicates the current operating condition category. Indicates the first The weights of the key influencing parameters , These represent the current operating condition categories. The corresponding number The first and second adjustment coefficients of the key influencing parameters Indicates the first The risk contribution of each key influencing parameter Indicates the first The dynamic sensitivity index corresponding to the key influencing parameters This indicates the number of key influencing parameters.

9. The CLCC converter fault early warning method according to claim 1, characterized in that, Overall Failure Risk Value Represented as: ； In the formula, , , These represent the linear weighting term, the nonlinear coupling term, and the risk trend term, respectively; where, The risk trend item Represented as: ； In the formula, This represents the weighting coefficient of the trend term. Indicates the length of the trend judgment window. Indicates the current warning period Fault risk indicators Indicates the current warning period The first Fault risk indicators for each early warning cycle.

10. A CLCC converter fault early warning system, characterized in that, include: The data acquisition and preprocessing module is used to acquire the operating parameter data of the CLCC converter during the current warning period and perform preprocessing to obtain multi-dimensional time series data. The fault risk assessment module is used to obtain the current fault risk index and the risk contribution of each operating parameter based on the multidimensional time series and mechanism mapping model. The operating parameters in the multidimensional time series are perturbed, and the fault risk index after perturbation is obtained based on the mechanism mapping model. The sensitivity analysis module is used to obtain the dynamic sensitivity index of each operating parameter based on the current fault risk index and the fault risk index after disturbance, and then obtain the set of key influencing parameters based on the dynamic sensitivity index. The working condition identification and weight adaptation module is used to identify the current operating condition based on the multi-dimensional time series data, and then obtain the weight of each key influencing parameter according to each key influencing parameter, its dynamic sensitivity index, and risk contribution. The fault warning generation module is used to obtain a comprehensive fault risk value based on each key influencing parameter and its corresponding weight, and then obtain fault warning information for the current warning period based on a preset risk threshold.

Images