Capacitor health state assessment method and device, storage medium and computer equipment

Abstract

According to the capacitor health state evaluation method and device, the storage medium and the computer equipment provided by the invention, the capacitor health state evaluation model is determined, and the capacitor health state evaluation model is used for controlling the fault-free sample capacitor to perform multi-state operation, and obtaining the electrical parameters of the sample capacitor in the multi-state operation; the temperature rise value of the sample capacitor under the multi-state operation is obtained, according to the electrical parameters and the temperature rise value corresponding to the sample capacitor, a classification prediction algorithm based on XGBoost is adopted, and the electrical parameters comprise the capacitance, the dielectric loss angle tangent value and the equivalent series resistance value; and acquiring an electrical parameter and a temperature rise value of the to-be-evaluated capacitor, and inputting the electrical parameter and the temperature rise value into the capacitor health state evaluation model to obtain a health state evaluation result of the to-be-evaluated capacitor. Therefore, the blank of the multi-parameter fusion evaluation method for the health state of the capacitor for the flexible direct current converter valve is filled, and different health states of the capacitor can be quickly and accurately identified.

Description

Technical Field

[0001] The present application relates to the field of flexible direct current transmission technology, and in particular to a capacitor health status assessment method, device, storage medium and computer equipment. Background Art

[0002] The modular multilevel converter is the core equipment of flexible DC transmission technology. Its power submodule contains metallized film capacitors, which have the advantages of high energy storage density and good reliability. Under the action of long-term complex stress, the capacitor may suffer from self-healing discharge, electrochemical corrosion, electrode shedding, bulging and heating, and other degradation and failures. In order to determine the degree of degradation of the capacitor, it is necessary to effectively detect it. Since the metallized film capacitors are encapsulated inside the submodule, the number is large and difficult to disassemble, and testing them one by one will consume a lot of manpower and material resources. Therefore, a health status assessment method for metallized film capacitors is proposed to guide operation and maintenance.

[0003] At present, the fault monitoring methods for DC support capacitors are mainly focused on the calculation of single-frequency capacitance values. One is a DC support capacitor monitoring method, that is, during each opening time of the submodule to which the DC support capacitor belongs, the voltage value and current value of the DC support capacitor are obtained, and the capacitance value of the DC support capacitor during each opening time is calculated based on the voltage value and current value. The capacitance value during each opening time is mathematically statistically analyzed to obtain at least one mathematical statistical result of the capacitance value, and the working state of the submodule to which the DC support capacitor belongs is determined based on the mathematical statistical result of the capacitance value and the preset conditions; the other is a capacitor high-temperature welding reliability test method. When the capacitor does not melt and fall apart during the test process through the visual feedback mechanism, the test ends, indicating that the capacitor meets the existing reflow soldering process curve. Based on this, there is no DC support capacitor health status assessment method that integrates multi-parameter feature quantities in the prior art. Summary of the invention

[0004] The purpose of the present application is to solve at least one of the above-mentioned technical defects, especially the technical defect that there is no method for evaluating the health status of DC support capacitors by fusing multi-parameter feature quantities in the prior art.

[0005] In a first aspect, the present application provides a method for evaluating the health status of a capacitor, the method comprising:

[0006] Determine a capacitor health status assessment model, which is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining electrical parameters of the sample capacitor under the multi-state operation, and obtaining the temperature rise value of the sample capacitor under the multi-state operation, and using an XGBoost-based classification prediction algorithm according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, wherein the electrical parameters include capacitance, dielectric loss tangent value, and equivalent series resistance value;

[0007] Obtain the electrical parameters and temperature rise values ​​of the capacitor to be evaluated;

[0008] The electrical parameters and temperature rise values ​​of the capacitor to be evaluated are input into the capacitor health status evaluation model to obtain the health status evaluation result of the capacitor to be evaluated.

[0009] In one of the embodiments, one end of the sample capacitor is connected to a multi-frequency harmonic synthesis power source;

[0010] The process of controlling a fault-free sample capacitor for multi-state operation includes:

[0011] determining a plurality of preset percentages of the rated DC voltage corresponding to the sample capacitor;

[0012] Through CAN communication, the AC output voltage value of the multi-frequency synthesis power supply is set according to the voltage value corresponding to each preset percentage, so that the sample capacitor operates at different voltages.

[0013] In one embodiment, the process of obtaining electrical parameters of a sample capacitor under multi-state operation includes:

[0014] Obtaining electrical parameters of the sample capacitor in a healthy state;

[0015] The sample capacitor is subjected to each AC output voltage value of the multi-frequency synthetic power supply for a plurality of preset withstand voltage durations;

[0016] For each AC output voltage value, in each preset withstand voltage duration, if the duration of operation of the sample capacitor at the AC output voltage value reaches the preset withstand voltage duration, the electrical parameters of the sample capacitor are recorded after the temperature of the sample capacitor drops to room temperature.

[0017] In one embodiment, the process of obtaining electrical parameters of a sample capacitor under multi-state operation further includes:

[0018] For each AC output voltage value, in each preset withstand voltage duration, if the sample capacitor is damaged when operating at the AC output voltage value and the duration does not reach the preset withstand voltage duration, the current duration and the electrical parameters of the sample capacitor are recorded.

[0019] In one embodiment, the process of obtaining the temperature rise value of the sample capacitor under multi-state operation includes:

[0020] Determine multiple temperature collection points of the sample capacitor, and determine the initial temperature of each temperature collection point of the sample capacitor, wherein the temperature collection points are three equally divided points at the upper position, three equally divided points at the middle position, and three equally divided points at the lower position of the sample capacitor;

[0021] For each AC output voltage value, when the sample capacitor runs at the AC output voltage value until the temperature is stable, the corresponding steady-state temperature is collected at each temperature collection point of the sample capacitor;

[0022] The temperature rise value of the sample capacitor is determined according to the initial temperature and the steady-state temperature corresponding to each temperature acquisition point.

[0023] In one embodiment, the process of establishing a capacitor health status assessment model includes:

[0024] Determining electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state, and determining electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state;

[0025] The XGBoost algorithm is used to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state and the healthy state label, and to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state and the unhealthy state label, so as to obtain a capacitor health status assessment model.

[0026] In one embodiment, the objective function of the XGBoost algorithm is:

[0027]

[0028] in, is the number of tree child nodes in the current regression tree, Used to control the number of leaf nodes. It is used to control the importance of a single regression tree and effectively prevent overfitting. is the first-order gradient sum of all samples of the current leaf node, is the second-order gradient sum of all samples of the current leaf node, is the weight of the current leaf node.

[0029] In a second aspect, the present application provides a capacitor health status assessment device, the device comprising:

[0030] A model determination module is used to determine a capacitor health status assessment model, wherein the capacitor health status assessment model is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining electrical parameters of the sample capacitor under multi-state operation, and obtaining temperature rise values ​​of the sample capacitor under multi-state operation, and using an XGBoost-based classification prediction algorithm according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, wherein the electrical parameters include capacitance, dielectric loss tangent value, and equivalent series resistance value;

[0031] A parameter acquisition module, used to obtain the electrical parameters and temperature rise value of the capacitor to be evaluated;

[0032] The evaluation result determination module is used to input the electrical parameters and temperature rise value of the capacitor to be evaluated into the capacitor health status evaluation model to obtain the health status evaluation result of the capacitor to be evaluated.

[0033] In a third aspect, the present application provides a storage medium: the storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the capacitor health status assessment method in any of the above embodiments.

[0034] In a fourth aspect, the present application provides a computer device, comprising: one or more processors, and a memory;

[0035] The memory stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the steps of the capacitor health status assessment method in any of the above embodiments are performed.

[0036] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0037] In the capacitor health status assessment method, device, storage medium and computer equipment provided in the present application, a capacitor health status assessment model is determined. The capacitor health status assessment model is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining the electrical parameters of the sample capacitor under multi-state operation, and obtaining the temperature rise value of the sample capacitor under multi-state operation, and according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, a classification prediction algorithm based on XGBoost is used, wherein the electrical parameters include capacitance, dielectric loss tangent value and equivalent series resistance value; the electrical parameters and temperature rise values ​​of the capacitor to be evaluated are obtained; the electrical parameters and temperature rise values ​​of the capacitor to be evaluated are input into the capacitor health status assessment model to obtain the health status assessment result of the capacitor to be evaluated. Compared with the traditional method that only relies on a single parameter for fault monitoring, this method integrates multiple electrical parameters such as capacitance, dielectric loss tangent value, equivalent series resistance value, and temperature rise value. By establishing a model through the XGBoost classification prediction algorithm, the health status of the capacitor can be more comprehensively evaluated, filling the gap in the multi-parameter fusion evaluation method for the health status of capacitors used in flexible DC converters. In addition, the multi-parameter fusion method can more accurately reflect the degree of degradation of the capacitor. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0039] Figure 1 A flow chart of a capacitor health status assessment method provided in an embodiment of the present application;

[0040] Figure 2 A schematic diagram of the structure of a capacitor health status assessment device provided in an embodiment of the present application;

[0041] Figure 3 A schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0043] The present application provides a capacitor health status assessment method. The following embodiments are described by taking the method applied to a computer device as an example. It can be understood that the computer device can be any device with data processing functions, including but not limited to a single server, a server cluster, a personal laptop computer, a desktop computer, etc. Figure 1 As shown, the method includes:

[0044] S101: Determine a capacitor health status assessment model.

[0045] The capacitor health status assessment model is established by controlling the fault-free sample capacitor to perform multi-state operation, obtaining the electrical parameters of the sample capacitor under multi-state operation, and obtaining the temperature rise value of the sample capacitor under multi-state operation. According to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, the classification prediction algorithm based on XGBoost is used, wherein the electrical parameters include capacitance, dielectric loss tangent value and equivalent series resistance value. Capacitance is a physical quantity that measures the ability of a capacitor to store charge, indicating the amount of charge that a capacitor can store under a given potential difference. The dielectric loss tangent value is a physical quantity that measures the degree of energy loss of the capacitor dielectric material under an AC electric field. In an AC circuit, a capacitor not only stores electrical energy, but also consumes part of the electrical energy due to the resistive component of the dielectric material. This part of the energy loss is called dielectric loss. The dielectric loss angle is the angle between the current vector and the voltage vector, and the dielectric loss tangent value is the tangent value of this angle, which reflects the insulation performance of the capacitor dielectric material. The equivalent series resistance value is a physical quantity that describes that in addition to the ideal capacitance characteristics, a capacitor also exhibits certain resistance characteristics in an AC circuit. This resistance characteristic is determined by factors such as the lead resistance, electrode resistance, dielectric resistance, and contact resistance inside the capacitor. In an AC circuit, the equivalent series resistance will cause additional energy loss during the charging and discharging process of the capacitor and affect the frequency response characteristics of the capacitor.

[0046] The capacitor health status assessment model is a mathematical model used to assess the health status of capacitors. Health status usually refers to the ability of a capacitor to maintain its performance parameters within an acceptable range during operation. Multi-state operation refers to the operating state of a capacitor under different working conditions, such as different voltages, currents, temperatures, etc. By simulating or actually operating the capacitor under these different conditions, data on the performance changes of the capacitor can be collected. Electrical parameters include capacitance (C), dielectric loss tangent (tan ) and equivalent series resistance (ESR), are important indicators for evaluating capacitor performance, reflecting the capacitor's ability to store charge, energy loss, and internal resistance. The temperature rise value is the temperature rise caused by energy loss during the operation of the capacitor. The temperature rise value is an important indicator for evaluating the heat dissipation capacity and thermal stability of the capacitor. The XGBoost classification prediction algorithm is an efficient and widely used machine learning algorithm, especially suitable for classification and regression problems. It is based on the gradient boosting framework and improves prediction accuracy by building multiple weak classifiers (usually decision trees) and combining them into a strong classifier.

[0047] In this step, first, select a fault-free sample capacitor and perform multi-state operation under a controlled environment. In each state, record the electrical parameters such as the capacitance, dielectric loss tangent value, equivalent series resistance value, and temperature rise value of the capacitor. Use the XGBoost algorithm to train the model with electrical parameters and temperature rise values ​​as input features and the health status of the capacitor as the output target, such as normal, slightly degraded, severely degraded, etc. During the training process, it is necessary to adjust the parameters of XGBoost, such as the learning rate, the number of trees, the depth of the tree, etc., to optimize the performance of the model.

[0048] S102: Obtain electrical parameters and temperature rise values ​​of the capacitor to be evaluated.

[0049] In this step, you can use a tester to measure the electrical parameters of the capacitor. Connect the capacitor to be evaluated to the test port of the tester correctly, ensure that the connection is firm, and avoid measurement errors caused by poor contact. According to the specifications and test requirements of the capacitor, set appropriate test parameters on the tester, start the tester, measure according to the set parameters, and import the measured electrical parameter values ​​into the computer device.

[0050] In addition to the above-mentioned electrical parameter test equipment, temperature sensors and data acquisition systems are also required. Install the temperature sensor on or near the surface of the capacitor to ensure that it can accurately reflect the temperature changes of the capacitor. Start the power supply to start the capacitor working. At the same time, start the data acquisition system and start collecting data from the temperature sensor in real time. During the test, regularly record the data of the ambient temperature and the surface temperature of the capacitor. After the test, calculate the temperature rise value of the capacitor based on these data.

[0051] S103: Inputting the electrical parameters and temperature rise value of the capacitor to be evaluated into a capacitor health status evaluation model to obtain a health status evaluation result of the capacitor to be evaluated.

[0052] In this step, the electrical parameters and temperature rise values ​​of the capacitor to be evaluated are combined into an input vector. The length of this vector should be the same as the length of the feature vector used in model training, and the order of the features should also be consistent. The constructed input vector is passed to the capacitor health status assessment model, and the prediction function of the model is called to obtain the health status assessment result. For multi-class classification problems, the prediction result can be a probability distribution, and for binary or multi-classification problems, the prediction result can be a specific category label. If the model outputs a probability distribution, the category with the highest probability can be selected as the predicted health status. If the model outputs a specific category label, it can be directly mapped to the corresponding health status description.

[0053] In the above embodiment, a capacitor health status assessment model is determined. The capacitor health status assessment model is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining the electrical parameters of the sample capacitor under multi-state operation, and obtaining the temperature rise value of the sample capacitor under multi-state operation, and according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, a classification prediction algorithm based on XGBoost is used to establish the model, wherein the electrical parameters include capacitance, dielectric loss tangent value, and equivalent series resistance value; the electrical parameters and temperature rise value of the capacitor to be evaluated are obtained; the electrical parameters and temperature rise value of the capacitor to be evaluated are input into the capacitor health status assessment model to obtain the health status assessment result of the capacitor to be evaluated. Compared with the traditional method that relies on only a single parameter for fault monitoring, this method integrates multiple electrical parameters such as capacitance, dielectric loss tangent value, equivalent series resistance value, and temperature rise value. By establishing a model through the XGBoost classification prediction algorithm, the health status of the capacitor can be more comprehensively evaluated, and the multi-parameter fusion method can more accurately reflect the degree of degradation of the capacitor.

[0054] In one embodiment, one end of the sample capacitor is connected to a multi-frequency harmonic synthesis power supply;

[0055] The process of controlling a fault-free sample capacitor for multi-state operation includes:

[0056] determining a plurality of preset percentages of the rated DC voltage corresponding to the sample capacitor;

[0057] Through CAN communication, the AC output voltage value of the multi-frequency synthesis power supply is set according to the voltage value corresponding to each preset percentage, so that the sample capacitor operates at different voltages.

[0058] Among them, the multi-frequency harmonic synthesis power supply is a power supply device that can generate multiple frequencies and / or harmonic components. It allows the user to adjust the waveform and frequency of the output voltage as needed to simulate different electrical environments or test conditions.

[0059] It can be understood that the various parameters of the capacitor need to be monitored by CAN communication connection, and CAN (Controller Area Network) is an ISO internationally standardized serial communication protocol. In order to reduce the number of wiring harnesses and use multiple LANs for high-speed communication of large amounts of data, the CAN communication protocol for the automotive field came into being. Due to its high performance and high reliability, CAN has been widely used in industrial automation, ships, medical equipment, industrial equipment, etc. The CAN controller determines whether it is a dominant level or an invisible level based on the potential difference between the two lines. At the same time, the CAN protocol has many superior features such as multi-master control, system flexibility, fast communication speed and long distance, error detection, error notification and error recovery functions, and is recognized as one of the most promising field buses. Therefore, this embodiment also uses CAN to interconnect detection equipment.

[0060] Specifically, first, the rated DC voltage of the sample capacitor needs to be known. Then, according to the test requirements, a series of preset voltage percentages are determined, such as 50%, 75%, 100%, 125%, etc. Using the CAN communication protocol, control instructions are written to convert the preset voltage percentages into specific voltage values. These control instructions are sent to the multi-frequency harmonic synthesis power supply to adjust its output voltage to the corresponding level. At each preset voltage level, the sample capacitor is allowed to run for a period of time. During this period, various performance parameters of the sample capacitor can be collected.

[0061] In one example, a fault-free capacitor is first placed in a constant temperature and humidity oven, with the temperature set to 55°C and the humidity set to 60%; the two ends of the capacitor are connected to a multi-frequency harmonic synthesis power supply and a high-voltage DC source; the high-voltage DC source outputs a DC voltage of 2800V, and the multi-frequency synthesis power supply outputs an adjustable AC voltage of 600V through an intermediate frequency transformer and a filter capacitor; a CAN communication connection is established between the computer and the multi-frequency synthesis power supply, and current sensors, voltage sensors, and multi-channel temperature sensors are installed to measure various parameters of the capacitor; the output voltage and output current limit value of the capacitor are set; a constant voltage and environmental operating condition are maintained to perform a capacitor withstand voltage test. Among them, the output voltage of the multi-frequency synthesis harmonic source is changed by a computer, specifically, the AC output voltage value is 8%, 12%, 16%, and 20% of the rated DC voltage of the capacitor as different operating conditions.

[0062] In this embodiment, the use of a multi-frequency harmonic synthesis power supply can simulate a more complex electrical environment, thereby more accurately reflecting the performance of the sample capacitor in actual applications. By operating the sample capacitor at different voltage percentages, multiple key points within its operating voltage range can be covered, which helps to fully reflect the performance of the capacitor at different voltage levels.

[0063] In one embodiment, the process of obtaining electrical parameters of a sample capacitor under multi-state operation includes:

[0064] Obtaining electrical parameters of the sample capacitor in a healthy state;

[0065] The sample capacitor is subjected to each AC output voltage value of the multi-frequency synthetic power supply for a plurality of preset withstand voltage durations;

[0066] For each AC output voltage value, in each preset withstand voltage duration, if the duration of operation of the sample capacitor at the AC output voltage value reaches the preset withstand voltage duration, the electrical parameters of the sample capacitor are recorded after the temperature of the sample capacitor drops to room temperature.

[0067] The preset withstand voltage time refers to the length of time that the sample capacitor needs to continue operating at each AC output voltage value, which is used to evaluate its stability and durability at that voltage.

[0068] Specifically, before the test begins, first ensure that the sample capacitor is in a healthy state, that is, it has not experienced any failure or degradation. Use appropriate measuring equipment to record the electrical parameters of the sample capacitor in a healthy state. According to the test requirements, set the output voltage value of the multi-frequency synthetic power supply and apply it to the sample capacitors one by one. For each AC output voltage value, let the sample capacitor run continuously for the preset withstand voltage time. At each AC output voltage value, when the duration of the sample capacitor's operation reaches the preset withstand voltage time, stop running and wait for its temperature to drop to room temperature. Use measuring equipment to record the electrical parameters of the sample capacitor after the temperature drops to room temperature.

[0069] In one example, a capacitor in a healthy state is connected to a high-voltage bridge, the capacitance, dielectric loss tangent and equivalent series resistance of the capacitor in the healthy state are measured, and the capacitance voltage and capacitance current at both ends of the capacitor are recorded through a current sensor and a voltage sensor; thereafter, when the withstand voltage test is carried out for 250h, 500h, 750h, 1000h, 1500h, 2000h, 3000h, 4000h, 5000h, 6000h, 7000h and 8000h, respectively, the capacitor is cooled and left to stand until room temperature, and the capacitance, dielectric loss tangent and equivalent series resistance of the capacitor after different withstand voltage test times are measured.

[0070] As you can understand, a high voltage bridge is a measuring device used to accurately measure the electrical parameters of electrical components such as resistance, inductance, and capacitance. This equipment is very important in electrical engineering and physics research, especially in the analysis and maintenance of high voltage systems. A high voltage bridge usually includes a series of precision measuring components and circuits that can safely operate in a high voltage environment. It can provide very accurate measurements to help engineers and researchers evaluate and optimize the performance of electrical equipment. By comparing the degree of imbalance of the test circuit with a known standard resistor, the high voltage bridge can accurately measure the value of an unknown resistor, which is essential to ensure the safe and efficient operation of electrical equipment.

[0071] In this embodiment, by testing the sample capacitor at different AC output voltage values ​​and different withstand voltage durations, its performance under different working conditions can be fully obtained. At each AC output voltage value, the sample capacitor is allowed to continue to run for a preset withstand voltage duration, and the electrical parameters are recorded after its temperature drops to room temperature, which can reduce the test error caused by temperature fluctuations or other external factors, and help improve the accuracy and reliability of the test.

[0072] In one embodiment, the process of obtaining electrical parameters of the sample capacitor under multi-state operation further includes:

[0073] For each AC output voltage value, in each preset withstand voltage duration, if the sample capacitor is damaged when operating at the AC output voltage value and the duration does not reach the preset withstand voltage duration, the current duration and the electrical parameters of the sample capacitor are recorded.

[0074] Specifically, for each AC output voltage value, the operating status of the sample capacitor is continuously monitored. If the sample capacitor is damaged, such as short circuit, open circuit, leakage, etc., before the preset withstand voltage duration is reached, the test is stopped immediately, and the current duration and electrical parameters before or during damage are recorded.

[0075] In this embodiment, the record of capacitor damage is added to make the test process more comprehensive. Not only the performance of the capacitor under normal operating conditions is considered, but also its behavior under extreme or abnormal conditions. The record of electrical parameters before damage can provide important information about the degradation of capacitor performance.

[0076] In one embodiment, the process of obtaining the temperature rise value of the sample capacitor under multi-state operation includes:

[0077] Determine multiple temperature collection points of the sample capacitor, and determine the initial temperature of each temperature collection point of the sample capacitor, wherein the temperature collection points are three equally divided points at the upper position, three equally divided points at the middle position, and three equally divided points at the lower position of the sample capacitor;

[0078] For each AC output voltage value, when the sample capacitor runs at the AC output voltage value until the temperature is stable, the corresponding steady-state temperature is collected at each temperature collection point of the sample capacitor;

[0079] The temperature rise value of the sample capacitor is determined according to the initial temperature and the steady-state temperature corresponding to each temperature acquisition point.

[0080] Specifically, multiple temperature collection points are selected on the sample capacitor, and these points should be evenly distributed at the high, middle and low positions of the capacitor to reflect the temperature distribution of the capacitor at different positions. For example, three equally divided points at the high position, three equally divided points at the middle position and three equally divided points at the low position can be selected, for a total of nine temperature collection points. Before the test begins, use a suitable temperature measurement device to measure and record the initial temperature of each temperature collection point. Connect the sample capacitor to a multi-frequency synthesis power supply and set a specific AC output voltage value. Start the power supply and let the sample capacitor run at this voltage until its temperature reaches a stable state. When the temperature of the sample capacitor is stable, use the same temperature measurement device to collect the corresponding steady-state temperature at each temperature collection point. For each temperature collection point, calculate the temperature rise value, that is, the difference between the steady-state temperature and the initial temperature, based on its initial temperature and steady-state temperature.

[0081] In one example, nine temperature collection points are set, and the system is operated at a set voltage until the temperature reaches a stable state and the temperature is collected. Collection points 1 to 3 are respectively three equally divided points at the high position of the wide-surface box of the whole capacitor, and for a single capacitor, three equally divided points at the top end face; collection points 4 to 6 are respectively three equally divided points at the middle position of the wide-surface box of the whole capacitor, and for a single capacitor, three equally divided points at the axial direction of the wide surface of the film; collection points 7 to 9 are respectively three equally divided points at the low position of the wide-surface box of the whole capacitor, and for a single capacitor, three equally divided points at the bottom end face. It can be understood that a single metalized film capacitor is mainly composed of metalized polypropylene or other types of plastic films. These films are evenly coated with a thin layer of metal on one or both sides, and the metals commonly used are aluminum or zinc. These metal films act as electrodes, and plastic films act as dielectric materials. The two ends of the metallized film capacitor are connected to the lead-out terminals through metallization evaporation technology, and then these superimposed films are tightly wound into a cylindrical shape or folded into a flat shape and placed in a protective shell. This structural design makes the metallized film capacitor have good self-healing properties, improves the capacitor's voltage resistance and reliability, and is widely used in power filtering, energy storage, frequency conversion and other fields. The whole machine capacitor is arranged in a staggered manner, with tin braided copper strips to connect several capacitors in parallel, and is potted in a stainless steel box with insulating materials such as polyurethane. When the capacitor deteriorates, the dielectric loss and metal loss will increase, resulting in an increase in the capacitor's heat loss, and ultimately an increase in the capacitor's steady-state temperature.

[0082] In this embodiment, by setting multiple temperature collection points at different positions of the sample capacitor, the temperature distribution of the capacitor during operation can be more comprehensively understood. By separately calculating and analyzing the temperature rise value of each temperature collection point, the thermal performance difference of the capacitor at different positions can be identified.

[0083] In one embodiment, the process of establishing a capacitor health status assessment model includes:

[0084] Determining electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state, and determining electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state;

[0085] The XGBoost algorithm is used to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state and the healthy state label, and to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state and the unhealthy state label, so as to obtain a capacitor health status assessment model.

[0086] Specifically, the electrical parameters and temperature rise values ​​of the sample capacitors are divided into healthy states and unhealthy states, and are marked as healthy state labels and unhealthy state labels, respectively. Afterwards, the collected data can be cleaned, outliers, missing values, etc. can be removed, and the data can be normalized or standardized so that parameters of different dimensions can be compared in the same model. The features that have the greatest impact on the health status assessment of the capacitor are selected from the electrical parameters and temperature rise values. Using the XGBoost algorithm, a correspondence is established between the sample data of the healthy state and the healthy state label. Similarly, a correspondence is established between the sample data of the unhealthy state and the unhealthy state label. The performance of the model is optimized by adjusting the parameters of XGBoost. Furthermore, methods such as cross-validation can be used to evaluate the accuracy and generalization ability of the model, and the model parameters can be adjusted according to the evaluation results until satisfactory performance is achieved.

[0087] In this embodiment, by using the XGBoost algorithm, it is possible to fully utilize the information in the sample data and improve the accuracy of the capacitor health status assessment.

[0088] In one embodiment, the objective function of the XGBoost algorithm is:

[0089]

[0090] in, is the number of tree child nodes in the current regression tree, Used to control the number of leaf nodes. It is used to control the importance of a single regression tree and effectively prevent overfitting. is the first-order gradient sum of all samples of the current leaf node, is the second-order gradient sum of all samples of the current leaf node, is the weight of the current leaf node.

[0091] Specifically, XGBoost is an improvement on the boosting algorithm based on GBDT. The objective functions of XGBoost and GBDT are both to minimize the objective function. The difference between the two is that XGBoost adds a regular term to the objective function to constrain the objective function. The regular term represents the complexity of the base decision regression tree. The objective function is as follows.

[0092]

[0093] In the formula is a regular term, which is used to prevent overfitting. and is a hyperparameter used to control the intensity of the penalty. is the number of tree child nodes in the current regression tree, Used to control the number of leaf nodes. Sum the squares of each node. It is used to control the importance of a single regression tree and effectively prevent overfitting.

[0094] The difference between the two lies in the minimum loss function The solutions are different. Since different loss functions are used to solve different objectives (classification or regression), in order to obtain a general solution to the loss function, GBDT only performs a first-order Taylor expansion on the loss function, while XGBoost performs a second-order Taylor expansion, as shown below.

[0095]

[0096] In the formula For the front The predicted value of a tree, For the current The predicted value of the tree, and The loss function is decomposed into leaf nodes, and the loss value of all samples in the current tree leaf node is calculated. Combining the regular penalty term with the second-order Taylor expansion formula, the objective function can be written as:

[0097]

[0098] In the formula is the first-order gradient sum of all samples of the current leaf node, is the sum of the second-order gradients of all samples of the current leaf node. The first-order and second-order gradients are constants when solving specific problems, so they can be directly summed and used and express.

[0099] When solving the optimal solution of the current tree, only each leaf node needs to be solved The minimum value of the tree, and finally the optimal solution of the entire tree can be obtained by adding up the loss of each leaf node. It can be observed that the objective function of the second-order Taylor expansion is a quadratic equation with a minimum value when the opening is upward. The formula method can be used to solve the minimum value. The final objective function is shown below.

[0100]

[0101] In terms of learning rate, XGBoost multiplies the weight of the leaf node by the learning rate after one iteration, in order to weaken the influence of each tree on the result and to expand the learning space. At the same time, XGBoost draws on the random forest algorithm and supports column sampling, which can not only reduce overfitting, but also reduce the amount of calculation, which is also a feature of XGBoost that is different from traditional GBDT.

[0102] In summary, the boosting strategy of XGBoost is similar to that of GBDT. The common point between the two is that after the newly added base classifier, the difference between the predicted value and the true value is fitted. The difference is that the GBDT fitting direction is along the negative gradient direction ( ), only uses the first-order Taylor expansion information, while XGBoost performs the second-order Taylor expansion on the loss function. Both algorithms support custom loss functions. The difference is that XGBoost requires the first-order and second-order gradient information of the loss function, while GBDT only needs the first-order gradient information. Compared with GBDT, XGBoost has more accurate fitting direction and faster speed, and is more suitable for engineering implementation. Finally, the evaluation model of multi-parameter fusion feature quantity and capacitor health status is obtained.

[0103] The classification function of XGBoost is required for capacitor health assessment. The label item is the degradation degree of the capacitor, which is set to healthy state, slightly degraded, moderately degraded, and severely degraded, respectively, and is set to 0~3; the data item is 208 sets of feature data obtained after 13 sets of withstand voltage time for capacitors at 4 output voltage levels. The parameters for setting XGBoost are shown in the table.

[0104]

[0105] In the table, objective is the problem type, multi:softmax indicates a multi-classification problem; num_class is the number of multi-classification categories, and is used together with multi:softmax; gamma is a parameter that controls whether to perform post-pruning, and the larger it is, the more conservative it is; max_depth is used to build the depth of the tree, and the larger it is, the easier it is to overfit; lamda is the L2 regularization term parameter that controls the model complexity weight value, and the larger the parameter is, the less likely the model is to overfit; subsample is a randomly sampled training sample; colsample_bytree is a column sampling performed when generating a tree; eta is the learning rate; and nthread is the number of CPU threads.

[0106] When using XGBoost to classify and identify health status, the validation scale is set to 0.2, which means that 160 sets of feature data and their label items are randomly selected as training sets, accounting for 77% of the total data set; the remaining 48 sets of feature data and their label items are used as test sets, accounting for 23% of the total data set.

[0107] The following is a description of the capacitor health status assessment device provided in the embodiment of the present application. The capacitor health status assessment device described below and the capacitor health status assessment method described above can be referred to in correspondence with each other. Figure 2 As shown, the present application provides a capacitor health status assessment device, the device comprising:

[0108] The model determination module 201 is used to determine a capacitor health status assessment model, which is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining electrical parameters of the sample capacitor under the multi-state operation, and obtaining the temperature rise value of the sample capacitor under the multi-state operation, and adopting a classification prediction algorithm based on XGBoost according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, wherein the electrical parameters include capacitance, dielectric loss tangent value, and equivalent series resistance value;

[0109] A parameter acquisition module 202, used to acquire electrical parameters and temperature rise values ​​of the capacitor to be evaluated;

[0110] The evaluation result determination module 203 is used to input the electrical parameters and temperature rise value of the capacitor to be evaluated into the capacitor health status evaluation model to obtain the health status evaluation result of the capacitor to be evaluated.

[0111] In one embodiment, one end of the sample capacitor is connected to a multi-frequency harmonic synthesis power supply;

[0112] The model determination module 201 includes:

[0113] A percentage determination unit, used to determine a plurality of preset percentages of the rated DC voltage corresponding to the sample capacitor;

[0114] The voltage setting unit is used to set the AC output voltage value of the multi-frequency synthesis power supply according to the voltage values ​​corresponding to each preset percentage through CAN communication, so that the sample capacitor operates at different voltages.

[0115] In one embodiment, the model determination module 201 includes:

[0116] A first electrical parameter acquisition unit, used to acquire electrical parameters of the sample capacitor in a healthy state;

[0117] A withstand voltage test unit, used to test the sample capacitor at each AC output voltage value of the multi-frequency synthetic power supply for a plurality of preset withstand voltage durations;

[0118] The second electrical parameter acquisition unit is used to record the electrical parameters of the sample capacitor after the temperature of the sample capacitor drops to room temperature for each AC output voltage value and in each preset withstand voltage duration if the duration for which the sample capacitor operates at the AC output voltage value reaches the preset withstand voltage duration.

[0119] In one embodiment, the model determination module 201 further includes:

[0120] The third electrical parameter acquisition unit is used to record the current duration and the electrical parameters of the sample capacitor for each AC output voltage value and each preset withstand voltage duration if the sample capacitor is damaged when operating under the AC output voltage value and the duration does not reach the preset withstand voltage duration.

[0121] In one embodiment, the model determination module 201 includes:

[0122] A temperature collection point determination unit is used to determine multiple temperature collection points of the sample capacitor and determine the initial temperature of each temperature collection point of the sample capacitor, wherein the temperature collection points are three equally divided points at the upper position, three equally divided points at the middle position, and three equally divided points at the lower position of the sample capacitor;

[0123] A steady-state temperature acquisition unit, for acquiring the corresponding steady-state temperature at each temperature acquisition point of the sample capacitor for each AC output voltage value when the sample capacitor runs at the AC output voltage value until the temperature is stable;

[0124] The temperature rise value determination unit is used to determine the temperature rise value of the sample capacitor according to the initial temperature and the steady-state temperature corresponding to each temperature collection point.

[0125] In one embodiment, the model determination module 201 includes:

[0126] A parameter determination unit, used to determine electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state, and to determine electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state;

[0127] The model building unit is used to use the XGBoost algorithm to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state and the healthy state label, and to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state and the unhealthy state label, so as to obtain a capacitor health status assessment model.

[0128] In one embodiment, the objective function of the XGBoost algorithm is:

[0129]

[0130] in, is the number of tree child nodes in the current regression tree, Used to control the number of leaf nodes. It is used to control the importance of a single regression tree and effectively prevent overfitting. is the first-order gradient sum of all samples of the current leaf node, is the second-order gradient sum of all samples of the current leaf node, is the weight of the current leaf node.

[0131] In one embodiment, the present application further provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the capacitor health status assessment method as described in any of the above embodiments.

[0132] In one embodiment, the present application further provides a computer device having computer-readable instructions stored therein. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the capacitor health status assessment method as described in any one of the above embodiments.

[0133] Indicatively, Figure 3 As shown, Figure 3 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 may be provided as a server. Figure 3 The computer device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by a memory 301 for storing instructions executable by the processing component 302, such as an application. The application stored in the memory 301 may include one or more modules each corresponding to a set of instructions. In addition, the processing component 302 is configured to execute instructions to perform the capacitor health status assessment method of any of the above embodiments.

[0134] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server TM, Mac OS X TM, Unix TM, Linux TM, Free BSD TM, or the like.

[0135] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0136] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the term "include", "comprise" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not clearly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or equipment including the elements. Herein, "one", "one", "said", "the" and "it" may also include plural forms, unless the context clearly indicates another way. A plurality refers to at least two cases, such as 2, 3, 5 or 8, etc. "And / or" includes any and all combinations of the relevant listed items.

[0137] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can refer to each other.

[0138] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A capacitor health status assessment method, characterized in that: The method comprises: Determine a capacitor health status assessment model, wherein the capacitor health status assessment model is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining electrical parameters of the sample capacitor under the multi-state operation, and obtaining a temperature rise value of the sample capacitor under the multi-state operation, and adopting an XGBoost-based classification prediction algorithm according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, wherein the electrical parameters include capacitance, dielectric loss tangent value, and equivalent series resistance value; Obtain the electrical parameters and temperature rise values ​​of the capacitor to be evaluated; The electrical parameters and the temperature rise value of the capacitor to be evaluated are input into the capacitor health status evaluation model to obtain the health status evaluation result of the capacitor to be evaluated.

2. The capacitor health status assessment method according to claim 1, characterized in that: One end of the sample capacitor is connected to a multi-frequency harmonic synthesis power supply; The process of controlling a fault-free sample capacitor to perform multi-state operation includes: Determining a plurality of preset percentages of the rated DC voltage corresponding to the sample capacitor; Through CAN communication, the AC output voltage value of the multi-frequency synthetic power supply is set according to the voltage value corresponding to each of the preset percentages, so that the sample capacitor operates at different voltages.

3. The capacitor health status assessment method according to claim 2, characterized in that: The process of obtaining the electrical parameters of the sample capacitor under multi-state operation includes: Acquiring electrical parameters of the sample capacitor in a healthy state; The sample capacitor is subjected to a plurality of preset withstand voltage durations at each AC output voltage value of the multi-frequency synthetic power supply; For each of the AC output voltage values, in each of the preset withstand voltage durations, if the duration of operation of the sample capacitor at the AC output voltage value reaches the preset withstand voltage duration, the electrical parameters of the sample capacitor are recorded after the temperature of the sample capacitor drops to room temperature.

4. The capacitor health status assessment method according to claim 3, characterized in that: The process of obtaining the electrical parameters of the sample capacitor under multi-state operation also includes: For each of the AC output voltage values, in each of the preset withstand voltage durations, if the sample capacitor is damaged when operating at the AC output voltage value and the duration does not reach the preset withstand voltage duration, the current duration and the electrical parameters of the sample capacitor are recorded.

5. The capacitor health status assessment method according to claim 2, characterized in that: The process of obtaining the temperature rise value of the sample capacitor under multi-state operation includes: Determine a plurality of temperature acquisition points of the sample capacitor, and determine an initial temperature of each of the temperature acquisition points of the sample capacitor, wherein the temperature acquisition points are respectively three equally divided points at a high position, three equally divided points at a middle position, and three equally divided points at a low position of the sample capacitor; For each of the AC output voltage values, when the sample capacitor runs at the AC output voltage value until the temperature is stable, collecting the corresponding steady-state temperature at each of the temperature collection points of the sample capacitor; The temperature rise value of the sample capacitor is determined according to the initial temperature and the steady-state temperature corresponding to each temperature collection point.

6. The capacitor health status assessment method according to claim 1, characterized in that: The process of establishing the capacitor health status assessment model includes: Determining electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state, and determining electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state; The XGBoost algorithm is used to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in a healthy state and the healthy state label, and to establish a correspondence between the electrical parameters and temperature rise values ​​of the sample capacitor in an unhealthy state and the unhealthy state label, so as to obtain the capacitor health state assessment model.

7. The capacitor health status assessment method according to claim 6, characterized in that: The objective function of the XGBoost algorithm is: in, is the number of tree child nodes in the current regression tree, Used to control the number of leaf nodes. It is used to control the importance of a single regression tree and effectively prevent overfitting. is the first-order gradient sum of all samples of the current leaf node, is the second-order gradient sum of all samples of the current leaf node, is the weight of the current leaf node.

8. A capacitor health status assessment device, characterized in that: The device comprises: A model determination module, used to determine a capacitor health status assessment model, wherein the capacitor health status assessment model is established by controlling a fault-free sample capacitor to perform multi-state operation, obtaining electrical parameters of the sample capacitor under the multi-state operation, and obtaining a temperature rise value of the sample capacitor under the multi-state operation, and adopting an XGBoost-based classification prediction algorithm according to the electrical parameters and temperature rise values ​​corresponding to the sample capacitor, wherein the electrical parameters include capacitance, dielectric loss tangent value, and equivalent series resistance value; A parameter acquisition module, used to obtain the electrical parameters and temperature rise value of the capacitor to be evaluated; The evaluation result determination module is used to input the electrical parameters and temperature rise value of the capacitor to be evaluated into the capacitor health status evaluation model to obtain the health status evaluation result of the capacitor to be evaluated.

9. A storage medium, characterized in that: The storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the capacitor health status assessment method according to any one of claims 1 to 7.

10. A computer device, characterized in that: include: one or more processors, and memory; The memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the capacitor health status assessment method according to any one of claims 1 to 7 are performed.

Images