Capacitance online monitoring method based on time-frequency domain joint model

Abstract

The invention discloses a capacitor on-line monitoring method based on a time-frequency domain joint model. The method is suitable for DC-Link capacitor state evaluation of power electronic equipment. Capacitive current is dynamically reconstructed through port current of an active inductor and a switching signal, and dependence on a special sensor is eliminated. An equivalent series resistance (RESR) is calculated based on a power loss model in a time domain, a 300Hz harmonic component is combined in a frequency domain to decouple a capacitance value, and synchronous high-precision monitoring of the RESR and the capacitance value is achieved. Simulation and experiments show that the RESR monitors error lt; 4%, capacitance error lt; and the monitoring period can be shortened to 30ms. According to the method, a main circuit does not need to be transformed, the method can be deployed in a DSP controller in an embedded mode, and a real-time non-intrusive state monitoring solution is provided for equipment such as an active inductor and a frequency converter.

Description

Technical Field

[0001] The present invention belongs to the technical field of power electronic equipment condition monitoring, and relates to an online capacitance monitoring method based on a time-frequency domain joint model, which is applicable to the health state assessment of capacitors in power electronic converters such as active inductors and frequency converters. Background Art

[0002] In recent years, the country has continuously increased policy support for the development of new energy, promoting the rapid development of power electronic technology. However, with the wide application of power electronic equipment in fields such as new energy grid-connected power generation, rail transit traction, and variable frequency speed regulation drive, its non-linear characteristics have led to increasingly prominent harmonic pollution, seriously affecting the power quality of the power grid. In the field of harmonic governance, traditional passive inductors have long been used as the main solution due to their simple structure, convenient maintenance, high reliability, etc. However, this solution has inherent limitations: on the one hand, its fixed cut-off frequency is difficult to adapt to the complex and changeable working conditions; on the other hand, large-capacity passive inductors face severe challenges in terms of volume, weight, and power density. To break through these limitations, the active inductor technology based on power converters has become a research hotspot in recent years. This technology realizes the dynamic adjustment of the inductance value through power electronic devices, significantly improving the flexibility of harmonic suppression. However, it is worth noting that the introduction of power electronic components in the power converter has also brought new reliability problems, which have greatly restricted the popularization and application of active inductor technology. The main components of the power conversion type active inductor include capacitors, power devices, and small-capacity filter inductors, etc. According to statistical data, power devices and capacitors are the main fault sources of power electronic converters, and the failure problems caused by the two account for more than 50%. Capacitors and power devices are not only key components in the active inductor module, but also the components that are most likely to age and fail. Their reliability is of great significance for ensuring the safe and stable operation of the active inductor and the power electronic conversion system. Therefore, it is necessary to deeply study the reliability guarantee mechanism of the active inductor and implement corresponding safety measures according to the operating state and characteristics of the device in the system to improve its operating reliability to meet the high reliability requirements of the active inductor.

[0003] Currently, the research on the health state monitoring of capacitors is generally based on two models. One is the equivalent circuit model of capacitors, and the other is to regard the whole system as a black box with multiple inputs and outputs. Among them, the first model can be further divided into state monitoring based on the principle of periodic small-signal ripple and the principle of non-periodic large-signal charge-discharge curves. The literature (Liu C, Deng F, Yu Q, et al. Submodule capacitance monitoring strategy for phase-shifted carrier pulsewidth-modulation-based modular multilevel converters[J]. IEEE Transactions on Industrial Electronics, 2020, 68(9): 8753-8767.) obtains the capacitance value by collecting the fundamental frequency information of the voltage signal and combining the response changes of the corresponding components before and after capacitor aging. The literature (Sundararajan P, Sathik M H M, Sasongko F, et al. Condition monitoring of DC-link capacitors using Goertzel algorithm for failure precursor parameter and temperature estimation[J]. IEEE Transactions on Power Electronics, 2019, 35(6): 6386-6396.) is based on the capacitor equivalent impedance model. The capacitor current is obtained by using a current sensor, and then the capacitance value is calculated by using the low-frequency components of the voltage and current. To avoid using a capacitor current sensor, the literature (Arya A, Ahmad M W, Agarwal N, et al. Capacitor impedance estimation utilizing dc-link voltage oscillations in single phase inverter[J]. IET Power Electronics, 2017, 10(9): 1046-1053.) indirectly obtains the capacitor current by analyzing the operating characteristics of a single-phase photovoltaic grid-connected inverter, using the switching signal and the grid-side inductor current signal, and calculates the capacitor impedance according to the second harmonic component of the fundamental frequency to evaluate the health state of the capacitor.The literature (Ahmad M W, Agarwal N, Kumar P N, et al. Low-frequency impedance monitoring and corresponding failure criteria for aluminum electrolytic capacitors [J]. IEEE Transactions on Industrial Electronics, 2017, 64(7): 5657-5666.) injects non-fundamental frequency signals into the system based on the signal injection method, and calculates the capacitance value according to the change components of the injected signals. However, such methods usually require additional capacitive current sensors, which will cause a certain degree of intrusion to the system in practical applications and increase the overall cost of the system. In addition, the existence of ESR is often ignored when estimating the capacitance value in the medium and low frequency bands by such methods, which will have a certain impact on the estimation accuracy of the capacitance value.

[0004] From the perspective of the energy storage of capacitors, many scholars have estimated the capacitance value based on the charge-discharge curve of capacitors. The literature (Buiatti G M, Martin-Ramos J A, Amaral A M R, et al. Condition monitoring of metallized polypropylene film capacitors in railway power trains [J]. IEEE Transactions on Instrumentation and Measurement, 2009, 58(10): 3796-3805.) addressed the problem of monitoring the DC-side capacitors in the locomotive traction system and proposed to estimate the capacitance value of the capacitor during the start-stop process of the locomotive. This method does not require any additional hardware circuit and only needs to write a simple monitoring program to complete accurate monitoring. The literature (Kim M, Sul S K, Lee J. Condition monitoring of DC-link capacitors in drive system for electric vehicles [C] / / 2012 IEEE Vehicle Power and Propulsion Conference. IEEE, 2012: 633-637.) studied the problem of capacitance value monitoring in the electric vehicle drive system. When the motor stops working, the capacitor and the inverter are isolated from the power supply, and at this time the capacitor starts to discharge through the motor winding. By controlling the inverter, the capacitor can discharge at a constant current, and the capacitance value can be calculated using the capacitor voltage discharge curve and the discharge current curve. Further, the literature (Li H, Xiang D, Han X, et al. High-accuracy capacitance monitoring of DC-link capacitor in VSI systems by LC resonance [J]. IEEE Transactions on Power Electronics, 2019, 34(12): 12200-12211.) proposed a capacitance discharge curve scheme based on LC resonance for adjustable speed drive systems. During the discharge of the capacitor, the capacitor, the A-phase and C-phase windings can form an LC resonance circuit, and the capacitance value is calculated by sampling the discharge trajectory.In the literature (Wu Y, Du X. A VEN condition monitoring method of DC-link capacitors for power converters[J]. IEEE Transactions on Industrial Electronics, 2018, 66(2): 1296-1306.), outside the independent system, a variable discharge network is designed to discharge the capacitor. The designed monitoring unit is an external circuit, which does not need to change the structure of the system itself and has high flexibility. The above method does not need to measure the small-signal ripple, but only needs to sample the non-periodic discharge curve, and the principle is simple and easy to implement. However, the applicable range of this kind of method is limited, only applicable to simple discharge models, and often only monitors the capacitance value, making it difficult to take into account the monitoring of both capacitance value and ESR.

[0005] In addition to the above types of methods, some scholars have established the connection between state parameters and system signals through intelligent algorithms to achieve capacitance monitoring. The literature (Rodriguez E, Gaowen L, Farivar GG, et al. Capacitor condition monitoring based on an adaptive observer of the low-frequency capacitor voltage ripples for modular multilevel converters[C] / / 2019 IEEE 4th International Future Energy Electronics Conference(IFEEC). IEEE, 2019:1-6.) proposed an online capacitance state monitoring method based on an adaptive observer of low-frequency capacitor voltage ripples, realizing non-invasive parameter identification of capacitance values; the literature (Soliman H, Abdelsalam I, Wang H, et al. Artificial neural network-based DC-link capacitance estimation in a diode-bridge front-end inverter system[C] / / 2017 IEEE 3rd International Future Energy Electronics Conference and ECCE Asia(IFEEC 2017-ECCE Asia). IEEE, 2017:196-201.) established a black-box model with multiple signal inputs and a single capacitance output based on neural network algorithms to obtain capacitance values. This type of method does not need to rely on specific frequency signals to calculate and obtain capacitance values, but the construction and training of capacitance estimation models based on intelligent algorithms require a large amount of data and time, which are not easy to implement in practical applications. Summary of the Invention

[0006] The object of the present invention is to provide an online capacitance monitoring method based on a time-frequency domain joint model, aiming to solve the problems existing in the prior art. By analyzing the operating principle of the active inductor, this method proposes a reconstruction strategy for capacitor current. Based on the reconstructed current, the equivalent series resistance R of the capacitor is estimated based on power loss in the time domain ESR and the capacitance value C of the capacitor is decoupled based on impedance characteristics in the frequency domain, realizing comprehensive monitoring of the capacitor state without the need for an additional capacitor current sensor. The object of the present invention is achieved through the following specific technical solutions.

[0007] An online capacitance monitoring method based on a time-frequency domain joint model, comprising the following steps:

[0008] S1 Based on the operating state of the active inductor, reconstruct the DC-Link capacitor current I through the port current I ind and the switch signal S n ; dc ;

[0009] S2 Calculate the equivalent series resistance R of the capacitor according to the reconstructed capacitor current and the AC component of the capacitor voltage in the time domain ESR ;

[0010] S3 Extract the components of the capacitor voltage and current at a fixed frequency in the frequency domain, and decouple and calculate the capacitance value C in combination with the R ESR value.

[0011] The online capacitance monitoring method provided by the present invention realizes non-invasive monitoring without a capacitor current sensor, and at the same time covers the full-parameter monitoring of the equivalent series resistance R ESR and the capacitance value C, improving the monitoring efficiency and accuracy.

[0012] Furthermore, the capacitance current reconstruction logic in step S1 is: when the active inductor is in the discharging mode, the capacitance current is equal to the port current; when in the charging mode, the capacitance current is equal to the negative value of the port current. Dynamically reconstruct the current through the switch state, eliminating the dependence on a dedicated sensor and reducing the hardware cost.

[0013] Even further, the reconstruction logic is implemented by the following formula:

[0014] I dc = I ind (S2·S3 - S1·S4)

[0015] In the formula, I ind is the active inductor port current value; I dc is the reconstructed value of the DC-Link capacitor current; S1, S2, S3, and S4 are the switch signals of the active inductor H-bridge.

[0016] Expressing the reconstruction process mathematically facilitates hardware implementation and algorithm programming.

[0017] Furthermore, the calculation formula of R ESR in step S2 is as follows:

[0018]

[0019] In the formula, P L is the total active power consumed by the capacitor; p C is the instantaneous power of the capacitor; τ is the integration time; u C,ac and i C,acThey are the AC sampling values of the capacitor voltage and current respectively; I C,rms is the effective value of the capacitor current.

[0020] By power integration, high-frequency noise interference is eliminated to improve the R ESR estimation accuracy.

[0021] Furthermore, a recursive low-pass filter is used to implement power averaging calculation, and the filtering coefficient k τ is set according to the steady-state characteristics of the system. This can optimize the use of computing resources and meet the requirements of real-time online monitoring.

[0022] Furthermore, in step S3, the impedance formula is used to decouple the capacitance value C:

[0023]

[0024] In the formula, |Z C | is the impedance of the DC-Link capacitor at a fixed frequency, and f is the fixed frequency.

[0025] Furthermore, the fixed frequency is selected as 300 Hz.

[0026] Furthermore, the monitoring error of the equivalent series resistance R ESR of the capacitor is less than 4%, and the monitoring error of the capacitance value C of the capacitor is less than 3.5%.

[0027] Furthermore, the on-line monitoring method of the capacitor is integrated into a DSP (Digital Signal Process) controller, and the monitoring period is configured to be 30 ms - 300 ms. This can achieve hardware embedded deployment and support dynamic adjustment of the monitoring frequency.

[0028] The present invention has the following beneficial technical effects: The monitoring method provided by the present invention is non-invasive and eliminates the need for current sensors; it can achieve full-parameter monitoring and synchronously obtain the equivalent series resistance R ESR of the capacitor and the capacitance value C; the monitoring accuracy is high, the error of the equivalent series resistance R ESR is <4%, and the error of the capacitance value C is <3.5%; it can perform real-time monitoring, and the DSP can achieve a response at the 30 ms level. Description of the Drawings

[0029] Figure 1 is the active inductor structure and control schematic diagram.

[0030] Figure 2 is the algorithm flow chart for estimating the equivalent series resistance R ESR of the capacitor based on the capacitor power loss.

[0031] Figure 3 is the algorithm flow chart of the online decoupling algorithm for state parameters based on the time-frequency domain joint model.

[0032] Figure 4 It is the comparison between the actual current and the reconstructed current waveforms.

[0033] Figure 5 It is the comparison between the actual current and the FFT analysis results of the reconstructed current.

[0034] Figure 6 It is the DC-Link capacitor voltage waveform and the FFT analysis results.

[0035] Figure 7 It is the R ESR simulation results based on the capacitor power loss.

[0036] Figure 8 It is the reconstructed waveform of the capacitor current in Group A.

[0037] Figure 9 It is the comparison between the FFT analysis results of the reconstructed current and the actual current of the capacitor in Group A.

[0038] Figure 10 It is the capacitor voltage waveform in Group A and its FFT analysis results.

[0039] Figure 11 It is the R ESR experimental results based on the capacitor power loss.

[0040] Figure 12 It is the measured values of the capacitors in each group at 300 Hz. Specific implementation manners

[0041] Next, in conjunction with the accompanying drawings of the specification, the technical solutions of the present invention will be clearly and completely described. Obviously, the described implementation manners are only a part of the implementation manners of the present invention, rather than all of them. Based on the implementation manners of the present invention, all other implementation manners obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0042] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be understood as a limitation of the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance, quantity or position.

[0043] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "installation", "connection", and "linkage" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0044] The active inductor structure and control principle are as Figure 1 shown. To meet the control requirements of the active inductor, the controller needs to collect the active inductor port current I ind , port voltage V ind , DC-Link capacitor voltage V dc , and finally generate a switching signal through hysteresis control to control the switching tube.

[0045] As Figure 1 known, the control system of the active inductor can complete the control requirements without sampling the capacitor current signal. However, the capacitor current signal is very important for the on-line state monitoring of the capacitor. Both it and the capacitor voltage signal contain the impedance information of the capacitor, that is, the state parameters of the capacitor to be monitored. To obtain the capacitor current signal, many documents obtain it by adding a current sensor to the capacitor circuit. However, adding an additional capacitor current sensor will damage the main circuit structure and increase the overall cost of the system. The increase in the cost of the active inductor will weaken its own advantages compared with the passive inductor and hinder its further promotion and application. To overcome the above problems and combined with the characteristics of the active inductor itself, the present invention proposes a strategy for reconstructing the DC-Link capacitor current of the active inductor, aiming to indirectly obtain the capacitor current through the existing signals of the system.

[0046] According to the above topological structure and control principle, it can be known that the capacitor current I dc is jointly determined by the active inductor port current I ind and the switching signal S n . Therefore, the capacitor current I ind and the switching signal S n can be used to reconstruct the capacitor current I dc , thus avoiding the use of an additional capacitor current sensor. The working modes of the active inductor are as follows:

[0047] Mode Ⅰ: When I ind_ref -I ind >h, that is, when the difference between the port current reference value and the actual value is greater than the upper limit h of the hysteresis width, S2 and S3 are turned on, and the capacitor discharges. At this time, I dc =I ind ;

[0048] Mode Ⅱ: When I ind_ref -I ind <-h, that is, when the difference between the port current reference value and the actual value is less than the lower limit -h of the loop width, S1 and S4 are turned on, and the capacitor is charged. At this time, I dc =-I ind ;

[0049] Mode Ⅲ: When -h < I ind_ref -I ind <h, that is, when the difference between the port current reference value and the actual value is within the loop width [-h, h], the switch state remains unchanged.

[0050] According to the above analysis, Mode Ⅲ is a continuation of Modes Ⅰ and Ⅱ. Therefore, there are two operating modes for the active inductor. Table 1 shows the logic of DC-Link capacitor current reconstruction.

[0051] Table 1 DC-Link Capacitor Current Reconstruction Logic Table

[0052]

[0053] Based on the DC-Link capacitor current reconstruction logic shown in the above table, the DC-Link capacitor current I dc can be calculated as:

[0054] I dc = I ind (S2·S3 - S1·S4) (1)

[0055] In the formula, I ind is the active inductor port current value; I dc is the reconstructed value of the DC-Link capacitor current; S1, S2, S3, and S4 are the switch signals of the active inductor H-bridge.

[0056] In order to meet the online monitoring of the DC-Link capacitor R ESR in the active inductor and even other power electronic converters, the present invention performs real-time online monitoring of R ESR based on the capacitor power loss. The active power loss P L of the capacitor can be expressed as:

[0057]

[0058] In the formula, ESR k represents the equivalent series resistance value under the kth current harmonic, and I RMS,k represents the effective value of the kth current harmonic component. Although the above formula can be used to calculate the R ESR value, it is actually difficult to obtain the R ESRValue. In addition, the failure determination criterion of the capacitor is usually based on a wide range of frequency regions without considering a specific frequency, so the above formula is not practical.

[0059] Actually, the R of the equivalent circuit of the capacitor ESR value will show slight frequency characteristics and increase at high frequencies due to eddy current effects. Generally, it only changes significantly in the lower frequency band and the extremely high frequency band, and the value tends to be stable in a relatively large range of intermediate frequencies. Most of the harmonic components of the current of the active inductor and the DC-Link capacitor of most power electronic converters fall within this range. Therefore, the present invention estimates R from the perspective of the power loss of the capacitor ESR value can accurately reflect the true R ESR value. Thus, there is the following formula:

[0060]

[0061] In the formula, P L is the total active power consumed by the capacitor; p C is the instantaneous power of the capacitor; τ is the integration time; u C,ac and i C,ac are the AC sampled values of the capacitor voltage and current respectively; I C,rms is the effective value of the capacitor current. Assuming that the DC loss caused by the leakage current is negligible, according to the above formula, we get:

[0062]

[0063] This method finally obtains a calculated value R ESR , which is very close to the actual R ESR , Figure 2 is the algorithm flow chart of the method.

[0064] Figure 2 gives the schematic diagram of the real-time online monitoring algorithm of R C,AC measured based on u C,AC and i ESR , where the average low-pass filter is implemented recursively, and k τ is the set filtering coefficient. This method does not require a specific band-pass filter, as long as the time constant of the average low-pass is high enough, that is, enough operations are performed to ensure accuracy. In addition, from Figure 2 it can be seen that the above operations are all multiplications and accumulations, which are easy to execute and time-consuming in the hardware system. The time-consuming division operation only needs to be executed once at the end of the entire monitoring period and will not occupy too much chip resources. In addition, since the aging of the capacitor is a long-term process, only by presetting the monitoring interval, the real-time online status monitoring of the capacitor under the steady-state operation condition of the active inductor can be completed.

[0065] In order to decouple the capacitance impedance at a specific frequency through the obtained R ESR value and obtain the capacitance value, the present invention takes the active inductor on the DC side of a three-phase uncontrolled rectifier system as an example to analyze and verify the proposed method. According to the working principle of the active inductor, the harmonic components of the DC-Link capacitor current of the active inductor mainly include low-frequency harmonic components such as the 6th and 12th harmonics from the three-phase uncontrolled rectifier circuit, as well as high-frequency switching frequency harmonics generated during its own switching process. Since hysteresis control is adopted and its switching frequency is not fixed, the high-frequency harmonic components are relatively dispersed and have small amplitudes. Considering the actual extraction difficulty and the accuracy of the final result, the present invention selects the 300Hz components of the capacitor voltage and capacitor current to calculate the capacitance impedance. Ignoring the equivalent series inductance ESL of the capacitor, according to the impedance characteristics of the capacitor, there is:

[0066]

[0067] Then the capacitance impedance at 300Hz is:

[0068]

[0069] In the formula, |Z C_300 | is the impedance of the DC-Link capacitor at 300Hz; X C_300 is the capacitive reactance of the DC-Link capacitor at 300Hz; |V dc_300 | is the amplitude of the 300Hz harmonic component of the DC-Link capacitor voltage; |I dc_300 | is the amplitude of the 300Hz harmonic component of the DC-Link capacitor current. The capacitive reactance X C_300 of the DC-Link capacitor is calculated according to the following formula:

[0070]

[0071] When the operating frequency of the capacitor migrates from the middle frequency band to the low frequency band, in some cases, the fluctuation amplitude of the equivalent series resistance is significantly lower than the growth trend of the capacitive reactance, that is, the capacitive reactance of the capacitor at 300Hz is much larger than R ESR , and at this time, R ESR can be ignored, and the capacitance value can be directly calculated by formula (7). However, due to the differences in the specific design and model of the capacitor, especially when using ordinary aluminum electrolytic capacitors, its R ESR value is relatively high, and directly ignoring R ESR will cause a large error in the estimated capacitance value in many cases.

[0072] To overcome the above problems, the present invention proposes a generally applicable capacitor state monitoring method, that is, R obtained by estimating the capacitor power loss ESRDecouple the impedance represented by the value pair in formula (6) to obtain the capacitance value at a specific frequency. This method can simultaneously monitor the capacitance value and R of the capacitor, and without a capacitive current sensor, it is a real-time online non-invasive DC-Link capacitance monitoring method. In the active inductor on the DC side of a three-phase uncontrolled rectifier system, the flow chart of the online decoupling algorithm for state parameters based on the joint time-frequency domain model is as ESR shown. Figure 3

[0073] To verify the effectiveness of the proposed capacitive current reconstruction strategy and the online decoupling method for capacitive state parameters, a simulation model of the active inductor on the DC side of a three-phase uncontrolled rectifier system is built on the MATLAB / Simulink platform for verification. The system simulation parameters are shown in Table 2.

[0074] Table 2 Specific parameters of the simulation platform for the active inductor on the DC side of a three-phase uncontrolled rectifier system

[0075]

[0076] According to the capacitive current reconstruction strategy described above, the DC-Link capacitive current is obtained by multiplying the sampled value of the port current of the active inductor by the switching state. The waveform comparison of the actual DC-Link capacitive current and the reconstructed DC-Link capacitive current is as Figure 4 shown, and the waveform Fourier analysis comparison is as Figure 5 shown. It can be seen from the figure that the waveforms of the actual current and the reconstructed current are almost the same, and the main harmonic components are basically the same, which proves the feasibility of the above reconstruction strategy. Therefore, the reconstructed capacitive current can replace the actual capacitive current for calculation.

[0077] According to Figure 5 the FFT analysis results of the reconstructed capacitive current and the actual current shown, the amplitude of the 300Hz harmonic component of the reconstructed capacitive current is 5.249A, and the amplitude of the 300Hz harmonic component of the actual current is 5.243A. The error between the two can be ignored in the actual estimation process. To obtain the impedance of the DC-Link capacitor at 300Hz, the FFT analysis of the DC-Link capacitor voltage waveform is also performed. The DC-Link capacitor voltage waveform and its FFT analysis results are as Figure 6 shown.

[0078] It can be seen from Figure 6 that the DC-Link capacitor voltage is mainly composed of a DC component and a small amount of 300Hz component, and the amplitude of the 300Hz harmonic component is 1.125V. Through the above analysis and calculation, the components of the capacitor voltage and current at 300Hz can be obtained respectively, so as to obtain the capacitance impedance at 300Hz, combined with the R estimated based on power loss in the time domain ESR ​The capacitance value can be decoupled from the value. The simulation and estimation results of the capacitor ESR value are as follows Figure 7 shown.

[0079] From Figure 7 it can be seen that from the start of the monitoring program at 0.3 s to obtaining the steady-state value of R ESR , it takes about 30 ms. The steady-state value of R ESR is about 30.25 mΩ, and the error from the simulation set value of R ESR is 0.83%. The entire monitoring process has low computational requirements, which is beneficial for implementation in the hardware system. According to the above simulation results, the capacitance value is decoupled and calculated using Equation (6). Now, the estimation results, the true value set by the system, and their relative errors are recorded in Table 3.

[0080] Table 3 Simulation Results of DC-Link Capacitor State Monitoring Based on the Joint Time-Frequency Domain Model

[0081]

[0082] From Table 3, it can be seen that the estimation error of the R ESR value is 0.83%, which verifies the effectiveness of the online monitoring algorithm of R ESR based on power loss. The simulation error mainly depends on the time length of the convergence of the monitoring program operation. In fact, a good accuracy can be obtained in a relatively short time, avoiding excessive data calculation. In addition, the estimation error of the capacitance value is 0.072%. The simulation results prove that the proposed monitoring method can accurately monitor the R ESR value and decouple the capacitance value from the specific frequency impedance, realizing the comprehensive online monitoring of the capacitance value and the equivalent series resistance.

[0083] To verify the effectiveness of the capacitance state monitoring method based on the joint time-frequency domain model proposed above, a state monitoring experiment is carried out on the DC-side active inductor of the three-phase uncontrolled rectifier system under the online operating condition. The specific parameters are shown in Table 4. The DC-Link capacitor is composed of aluminum electrolytic capacitors with the model 100PX1000MEFC18X35.5, and the nominal parameters are 100 V / 1000 μF. The aging of the capacitor is simulated by removing the parallel sub-capacitors.

[0084] First, to verify the effectiveness of the proposed DC-Link capacitor current reconstruction strategy, the required active inductor port current signal data is collected through the existing port current sensor on the experimental platform, and the switch sequence information is obtained through the DSP. Then, a program is written according to the reconstruction logic in Table 1 to process the data. Finally, the complete reconstructed current waveform of Group A capacitors is obtained as Figure 8As shown, it can be seen that the reconstructed current waveform is basically consistent with the waveform verified by simulation, and it is initially judged to conform to the theoretical analysis results in the previous text. Further, it is compared with the FFT analysis results of the measured capacitor current, and the comparison results are as Figure 9 shown. From Figure 9 it can be seen that the amplitudes of the harmonic components of the reconstructed current and the measured current basically match, verifying the feasibility of the method, and the current reconstruction of the DC-Link capacitor has a good effect.

[0085] While verifying the effectiveness of the capacitor current reconstruction strategy, the harmonic components of the reconstructed current of each group of capacitors at 300 Hz are obtained. According to the monitoring principle of the time-domain model, in order to obtain the impedance of the DC-Link capacitor at 300 Hz, it is necessary to extract the 300 Hz component of the capacitor voltage. The DC-Link capacitor voltage waveform is collected through the active inductor DC-side capacitor voltage sampling circuit as shown in Figure 10 (a). The main body is a 30 V DC component with a ripple of about 2 V. The recorded capacitor voltage waveform data is imported into MATLAB for FFT analysis, and the results are as shown in Figure 10 (b).

[0086] The 300 Hz components of the DC-Link capacitor voltage and the reconstructed current obtained from the above analysis and calculation are statistically sorted, and the results are shown in Table 4. According to the data in this table, the impedance at 300 Hz can be calculated. According to the impedance characteristics of the capacitor, its impedance is composed of the capacitance value and the equivalent series resistance value. Next, the online monitoring method of R ESR based on the capacitor power loss will be verified to obtain R ESR , and then the capacitance value of the capacitor is decoupled from the capacitor impedance at this frequency.

[0087] Table 4 Extraction results of the 300 Hz components of the capacitor voltage and current of each group

[0088]

[0089] After verifying the accuracy of the capacitor current reconstruction strategy, the reconstructed value can be used to monitor the R ESR value online based on the steady-state loss of the capacitor under steady-state operating conditions. The collected capacitor voltage and current data are processed by the integral and low-pass filter algorithms in the DSP for a certain period of time, and finally the estimated stable value is output. The estimated results of the R ESR value of the DC-Link capacitor under the three connection methods are as shown in Figure 11 .

[0090] From Figure 11 it can be seen that from the start of the monitoring program to obtaining R ESRThe steady-state value can be calculated in the controller in about 300 ms, which can respond well to the monitoring program and has low requirements for hardware resources. The capacitors of each group were measured by an LCR meter at a frequency of 300 Hz, and the measured results are as Figure 12 shown. The comparison between the online monitoring results of R ESR and the measured values by LCR is presented in Table 5.

[0091] Table 5 Online monitoring results of R ESR based on the capacitive power loss

[0092]

[0093] According to the above theoretical analysis, to achieve a comprehensive monitoring of the state parameters of the DC-Link capacitor, based on obtaining the R ESR value in the time-domain model, the capacitive impedance is further decoupled and calculated in the frequency domain to obtain the capacitance value of the DC-Link capacitor at 300 Hz. Finally, the capacitance value and R ESR at 300 Hz are successfully extracted under steady-state conditions. The comparison between the final monitoring results and the measured values by LCR is shown in Table 6.

[0094] Table 6 Comparison between the online decoupling experiment results and the measured values of the capacitance

[0095]

[0096] The present invention addresses the problem of capacitor monitoring for an active inductor under on-line operating conditions, and proposes a capacitor monitoring method based on a time-frequency domain joint model. According to the relationship between the port current of the active inductor and the capacitor current, this method constructs a reconstruction strategy for the capacitor current. Then, based on the existing capacitor voltage signal and the reconstructed current signal, R ESR is estimated based on power loss in the time domain, and the capacitance value at a specific frequency is decoupled based on the obtained R ESR value and impedance characteristics in the frequency domain. Finally, the present invention verifies the effectiveness of the proposed method in the experimental platform of the DC-side active inductor of a three-phase uncontrolled rectifier system. The error of the equivalent series resistance R ESR is <4%, and the error of the capacitance C is <3.5%. This method can simultaneously achieve non-invasive monitoring of the capacitance value and R ESR , and without any additional hardware circuit and cost, it can meet the requirements of the monitoring accuracy of capacitors in practical engineering applications.

[0097] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention without departing from the principles and purposes of the present invention. The protection scope of the present invention is defined by the claims and their equivalent technical solutions.

Claims

1. An on-line capacitance monitoring method based on a time-frequency domain joint model, characterized in that The following steps are involved: S1 reconstructs the DC-Link capacitor current I dc based on the operating state of the active inductor, through the port current I ind and the switching signal S n ; ind and the switching signal S n reconstructs the DC-Link capacitor current I dc ; S2 calculates the equivalent series resistance R of the capacitor according to the AC components of the reconstructed capacitor current and the capacitor voltage in the time domain ESR ; S3 extracts the components of the capacitor voltage and current at a fixed frequency in the frequency domain, and decouples and calculates the capacitance value C in combination with the R ESR value.

2. The capacitance on-line monitoring method according to claim 1, wherein The capacitor current reconstruction logic of step S1 is: when the active inductor is in the discharge mode, the capacitor current is equal to the port current; when it is in the charging mode, the capacitor current is equal to the negative value of the port current.

3. The on-line capacitance monitoring method according to claim 2, wherein The reconstruction logic is implemented by the following formula: I dc = I ind (S2·S3 - S1·S4) Where, I ind is the active inductor port current value; I dc is the reconstructed value of the DC-Link capacitor current; S1, S2, S3, and S4 are the switching signals of the active inductor H-bridge.

4. The on-line capacitance monitoring method according to claim 1, characterized in that, In step S2, R ESR The calculation formula is as follows: Where, P L is the total active power consumed by the capacitor; p C is the instantaneous power of the capacitor; τ is the integration time; u C,ac and i C,ac are the AC sampled values of the capacitor voltage and current respectively; I C,rms is the effective value of the capacitor current.

5. The on-line capacitance monitoring method according to claim 4, wherein The power averaging calculation is implemented using a recursive low-pass filter, and the filter coefficient k τ is set according to the steady-state characteristics of the system.

6. The on-line capacitance monitoring method according to claim 1, characterized in that, In step S3, the impedance formula is used to decouple the capacitance C: where |Z C | is the impedance of the DC-Link capacitor at the fixed frequency, and f is the fixed frequency.

7. The on-line capacitance monitoring method according to claim 6, characterized in that, The fixed frequency is selected as 300 Hz.

8. The on-line capacitance monitoring method according to claim 1, characterized in that Capacitive equivalent series resistance R ESR The monitoring error is less than 4%, and the monitoring error of the capacitance value C of the capacitor is less than 3.5%.

9. The capacitance on-line monitoring method according to claim 1, wherein The capacitance online monitoring method is integrated into a DSP controller, and the monitoring period is configured to be 30ms-300ms.

Images