CN113162042B - MPPF capacitor failure assessment method in MMC system based on band energy - Google Patents

Abstract

The invention discloses a method for evaluating the failure of MPPF capacitors in an MMC system based on frequency band energy. First, the capacitor output voltage of the capacitor to be evaluated is obtained; then Fourier analysis is used to analyze the output voltage signal in the frequency domain to obtain a spectrum signal; and then the frequency domain is calculated. Calculate the power spectrum of the signal; integrate the high frequency band of the output voltage power spectrum of the capacitor to obtain the high frequency band energy of the output voltage; calculate the capacitance failure degree D according to the high frequency band energy; and finally judge whether the capacitor fails according to the capacitance failure degree D. According to the relationship that the larger the ESR value is, the larger the high frequency band energy of the capacitor output voltage is, the frequency band energy analysis is carried out by using the existing capacitor voltage signal in the MMC, and the current ESR change of the capacitor is obtained, and then the current failure state of the capacitor is judged. Capacitor failure detection can be accomplished with additional sensors, avoiding additional costs.

Description

A Failure Evaluation Method of MPPF Capacitors in MMC System Based on Band Energy

technical field

The invention relates to the technical field of capacitor state monitoring, in particular to a method for evaluating the failure of MPPF capacitors in an MMC system based on frequency band energy.

Background technique

Focusing on the strategic goals of building a low-carbon economy and building an energy Internet, and facing the increasing demand for electricity due to rapid economic development, renewable energy has emerged as the times require. High voltage direct current transmission (HVDC) has been widely used due to its advantages of long transmission distance and small system active power loss. According to the demand of low-carbon smart grid, the optimal design of large-capacity power electronic equipment in the future must comprehensively consider the aspects of cost, efficiency and reliability. Therefore, the reliability of HVDC system has become an urgent problem to be solved. Modular multilevel converter valve is the main converter topology of HVDC system, and its reliability research is very important for the stable operation of HVDC system.

Compared with the traditional two-pole voltage source converter, the large-capacity DC bus capacitors usually used for energy storage and filtering are distributed in each sub-module in the MMC (Modular Multilevel Converter). Capacitive energy storage is replaced, so the volume and weight of capacitors in MMC account for a large part of the system. MPPF (Metallized polypropylene film) capacitors are widely used in MMC due to their small size, strong AC current carrying capacity, and high tolerance. High voltage stress, high current stress in MMC and the fluctuation of the core temperature inside the capacitor caused by AC working conditions will accelerate the aging of the MPPF capacitor, and eventually lead to the failure of the MPPF capacitor. Capacitor failure will bring great risks to the operation of HVDC transmission system, so it is necessary to propose a failure detection method to evaluate the current health status of MPPF capacitors.

The topology of the MMC main circuit is shown in Figure 1. Point O is a zero potential reference point. The MMC consists of six bridge arms, each bridge arm is formed by N submodules (Submodule, SM) and a reactor L _arm in series, and the upper and lower bridge arms of each phase are combined into a phase unit. Each SM is a half-bridge unit, S ₁ and S ₂ are IGBT modules, D ₁ and D ₂ are anti-parallel diodes, C _SM is the DC side capacitor of the sub-module; i _arm is the bridge arm current, and u _SM is the output of the sub-module voltage, i _cap is the charging current of the capacitor, u _cap is the capacitor voltage, and the reference direction of each physical quantity is shown in the figure. Each sub-module is connected to the main circuit in series through the connection port, and the sum of the output voltages of each sub-module of the bridge arm is equal to the DC bus voltage U _dc . The MMC neutron module has three working modes and six working states, and the corresponding capacitor has two output states, as shown in Figure 2. When D1 is turned _on or S1 is turned _on , SM is in the cut-in mode, at this time the DC side capacitor is connected to the main circuit, and is charged and discharged through the bridge arm current, and the output voltage of the capacitor is ₊ u _cap ; S2 is turned _on or D2 When turned on, the SM is in the cut-off mode. At this time, the DC side capacitor is isolated, the voltage on both sides of the capacitor remains unchanged, and the output voltage of the capacitor is 0; when D ₁ is turned on or D ₂ is turned on, the SM is in the blocking mode. In this case It is a non-normal operating mode and is usually used to precharge capacitors or bypass capacitors in the event of a system failure.

Similar solution to the present invention [Reference: Sun Yuting. Condition monitoring of metallized polypropylene film capacitors in MMC [D]. Chongqing University, 2020]:

Considering the sub-module capacitor as the impedance, the impedance of the capacitor can be calculated from the formula (1). The simplest and fastest way to obtain the impedance is the Fourier decomposition to represent the voltage and current responses of the capacitor. Therefore, obtaining the impedance value at any frequency also obtains the capacitance value and ESR.

Select the fundamental frequency and double frequency components of the capacitor voltage and current after the Fourier transform to calculate the impedance, and the calculation formula is as (2).

Therefore, the steps to realize the state monitoring of capacitors in the MMC system in this scheme are: obtaining the voltage and current information of the capacitors, performing Fourier transformation to obtain the response at the fundamental frequency and the double frequency, and finally calculating the current capacitance value of the capacitor and ESR value.

The method in this document mainly judges the aging state of metallized film capacitors by calculating the capacitance value of the capacitor, and assists in judging the aging state by calculating the ESR value. However, the failure criterion of metallized film capacitors is usually a 5% drop in capacitance. , ESR increases twice. Therefore, ESR has a larger variation range than the capacitance value, and the monitoring capacitance value is prone to misjudgment due to environmental interference or sensor error, and the ESR value in this document is calculated by a complex formula, It puts forward higher requirements on the computing power of the processor, and is not suitable for use in the actual working system.

In addition, the existing state monitoring methods for power capacitors can be basically divided into the following three categories: 1. External circuit detection method; 2. Specific signal injection method; 3. Algorithmic calculation method. As shown in Figure 3.

1. Sensor detection method

A sensor is added to the circuit of the capacitor to be detected to detect the current and voltage information of the capacitor, and the current capacitance value and ESR value are determined by calculating the impedance and power of the capacitor and other information. Since the electrical information of the capacitor can be detected in real time, such methods have high real-time performance and high precision. However, there are a large number of sub-modules in the MMC system, and each sub-module contains a capacitor. Therefore, if this method is used to monitor the health status of the capacitor, a large number of sensor detection circuits need to be added, which will cost a lot of money, and the economic and practical value is not high. In addition, such methods are usually calculated according to the specific working principle of the capacitor in the circuit, so the portability of the capacitance state monitoring in the MMC system is poor.

2. Specific signal injection method

In order to avoid the high cost brought by the sensor detection method, scholars propose a specific signal injection method. By injecting a signal of a certain frequency into the circuit, the response of the capacitor under this frequency component is obtained to obtain information such as the current capacitance value and ESR value. However, the capacitor health information obtained by this method is too dependent on frequency, so there are two problems: one is that the capacitance ESR information obtained from the experiment cannot be compared with the simulated value because the capacitance detection is performed at different frequency points; the other is the capacitance The initial value of the capacitor given by the manufacturer is obtained through experiments at a frequency of 120Hz, and the capacitance value of the capacitor at different frequencies is different. Therefore, the capacitance value obtained by the specific signal injection method cannot be compared with that given in the technical manual. The initial value is compared to infer the health state.

3. Specific algorithm calculation

This kind of method is the emerging method of capacitance condition monitoring. The voltage and current conditions of the capacitor are processed through specific algorithms such as the least squares method and the Kalman filter method, and the current capacitor state information is calculated. Such methods do not require additional costs and are relatively easy to implement. However, since various algorithms have their own shortcomings, the current state monitoring methods using algorithmic computing all have shortcomings such as excessive computational complexity and complex computing processes.

SUMMARY OF THE INVENTION

In view of the above problems, the purpose of the present invention is to provide a method for evaluating the failure of MPPF capacitors in an MMC system based on frequency band energy. It has the advantages of simple calculation, simple application and strong portability. The technical solution is as follows:

A method for evaluating the failure of MPPF capacitors in an MMC system based on frequency band energy, comprising the following steps:

Step 1: Obtain the capacitor output voltage of the capacitor to be evaluated;

Step 2: Use Fourier analysis to perform frequency domain analysis on the capacitor output voltage signal to obtain the spectrum signal ΔV(ω);

Step 3: Calculate the power spectrum P(ω) of the output voltage frequency domain signal of the capacitor;

Step 4: Perform time integration on the high frequency band of the power spectrum P(ω) to obtain the high frequency band energy E(w) of the output voltage;

Step 5: Calculate the capacitance failure degree D:

D=(E(w)-E ₀ (w))/ΔE

Among them, E ₀ (w) is the initial high frequency band energy, and ΔE is the energy that increases when the capacitor fails;

Step 6: Determine whether the capacitor fails according to the capacitor failure degree D: if the capacitor failure degree D=1, the capacitor has failed, otherwise the capacitor is normal.

Further, the step 1 specifically includes:

The equivalent series resistance, capacitance and equivalent series inductance are connected in series to establish the equivalent model of the capacitor; then the voltage drop on the capacitor to be measured is the sum of the voltage drop on the resistor, capacitor and inductor:

v(t)= _vESR (t)+ _vC (t)+ _vL (t)

Among them, v _ESR (t) is the voltage drop generated by the equivalent series resistance, v _C (t) is the voltage drop of the capacitor part, and v _L (t) is the voltage drop on the equivalent series inductance;

Define loss tangent as the ratio of ESR to capacitor impedance:

Among them, ESR is the equivalent series resistance value of the capacitor under test, ω ₀ is the working angular frequency, which is 2*π*50; C is the capacitance value of the capacitor under test;

Ignoring the effect of the equivalent series inductance, the output voltage of the capacitor is expressed as follows:

Among them, ΔV(t) is the voltage drop generated by the equivalent series resistance, which is reflected as the increment of the capacitor voltage; i _arm is the bridge arm current; θ(tt ₀ ) is the step signal at t ₀ , and t ₀ is the capacitor charging Start time; _vc (t ₀ ) is the voltage drop of the capacitor part at time t ₀ .

Further, in step 5, the increment of the capacitor voltage is analyzed in the frequency domain, and the high frequency band energy expression is obtained as follows:

Among them, i _ac is the current flowing through the capacitor; ω is the angular frequency independent variable; δ is the dielectric loss angle; T is the integration time.

The beneficial effects of the present invention are:

1) According to the relationship that the larger the ESR value is, the larger the high frequency band energy of the capacitor output voltage is, the frequency band energy analysis is carried out by using the existing capacitor voltage signal in the MMC, and the current ESR change of the capacitor is obtained, and then the current failure state of the capacitor is judged.

2) The present invention can realize capacitor failure detection without adding additional sensors, avoiding extra cost, and the method can realize real-time detection of capacitor ESR value, which is suitable for capacitor health state detection in most switching circuits, and has strong universality.

3) The present invention establishes a failure detection model based on the high frequency band energy of the capacitor output voltage. The model can detect the high frequency band energy of the capacitor output voltage in real time, reflect the change of the capacitor ESR value, and qualitatively analyze the high frequency band energy and the failure degree of the capacitor. The relationship between.

4) The algorithm of the present invention is simple and easy to practice, and can provide a basis for the health management of capacitors in large-capacity power electronic equipment.

Description of drawings

Figure 1 is a topology diagram of the MMC main circuit and sub-modules.

Figure 2 is a diagram of six working states of the sub-module.

Figure 3 shows the research method for the condition monitoring of metallized film capacitors.

Figure 4 shows the capacitor voltage and bridge arm current in the MMC system.

FIG. 5 is an equivalent circuit diagram of a capacitor.

Figure 6 shows the output voltage waveform of the capacitor under different ESR.

Figure 7 shows the output voltage power spectrum when the capacitance value is reduced.

Figure 8 shows the output voltage power spectrum when the capacitor ESR increases.

Figure 9 shows the high frequency band energy of the capacitor voltage after the ESR is increased and the capacitance value is decreased.

Figure 10 shows the process of on-line failure detection of capacitors.

Detailed ways

The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The current and voltage of the capacitor in the MMC system are shown in Figure 4. When the trigger signal is 1, the capacitor is charged or discharged with the direction of the bridge arm current, and the voltage increases or decreases. When the trigger signal is 0, the capacitor is in a cut-out state and the voltage remains unchanged. The bridge arm current flows through the capacitor, including the DC component and the AC component, so the capacitor voltage is also the superposition of the DC component and the AC component.

The equivalent circuit diagram of the capacitor is shown in Figure 5. In the circuit analysis, the equivalent model of the capacitor is often considered to be an equivalent series resistance, a capacitor and an equivalent series inductance in series, and the voltage drop on the capacitor is the sum of the voltage drop on the resistor, capacitor and inductor. , as shown in formula (3). δ is the loss angle, and the tangent of the defined loss angle is the ratio of the ESR to the capacitor impedance, as shown in (4).

v(t)=v _ESR (t)+v _C (t)+v _L (t) (3)

The MMC simulation platform is built in MATLAB/Simulink, and the output waveform after the capacitor ESR value increases is shown in Figure 6. The waveforms show that the capacitors are connected in series with 0.01Ω, 0.03Ω and 0.05Ω resistors to simulate the output waveform of the increased ESR. It can be seen that compared with the output voltage of the capacitor in the healthy state, the increase of ESR will not affect the output of the capacitor The low frequency part of the voltage, but it will bring a large step amount at the moment of capacitor charging, that is, high frequency harmonics are generated. The reason for the step amount is analyzed as follows:

If the capacitor starts to charge at time t ₀ , the current flowing through the capacitor in a very short time is as follows:

Among them, θ(t) is the step signal at time t ₀ .

Ignoring the effect of the equivalent series inductance, the output voltage of the capacitor can be expressed as follows:

It can be seen that when the capacitor is in a healthy state and the ESR increases sharply due to the failure of the capacitor, the voltage drop caused by the equivalent series resistance cannot be ignored. The increase of the ESR value of the MPPF capacitor will cause the high frequency component of the output voltage to increase. Therefore, the high frequency component of the output voltage of the MPPF capacitor can be used to detect the change of the ESR value, thereby realizing the failure detection of the MPPF capacitor.

In the MMC system, the switching on and off of the sub-modules is determined by the capacitor voltage sequence of all the sub-modules of the current bridge arm, so the charging and discharging time of the capacitor is uncertain, which brings difficulties to the time domain analysis of the capacitor voltage. If Fourier analysis is used to analyze the voltage in the frequency domain, it can be seen from the above description that the increase of ESR will bring a small increment to almost every frequency point in the high frequency band of the capacitor output voltage. Monitoring, it is easy to be affected by environmental disturbances, resulting in errors in the measurement results. Therefore, the present invention uses the high frequency band energy of the output voltage of the capacitor as the characteristic quantity to detect the failure of the capacitor. The relationship between the frequency domain expression of a signal and its frequency band energy is shown in (7):

Perform frequency domain analysis on the increment of capacitor voltage in the above analysis, and the frequency band energy expression can be obtained as shown in (8):

It can be seen that the sharp increase of ESR after MPPF failure will lead to the increase of high-frequency harmonics of the capacitor output voltage, and the high-frequency energy will increase with the increase of ESR. In the process of capacitor failure, the decrease of capacitance value and the increase of ESR occur at the same time. Therefore, a simulation model is established in MATLAB/Simulink to carry out the high frequency band energy of the output voltage of the capacitor under the conditions of decreasing capacitance value and increasing ESR, respectively. The output voltage power spectrum analysis results As shown in Figures 7 and 8. It can be seen that when the ESR of the capacitor increases, the output voltage and power of the capacitor obviously appear layered, and the output power increases with the increase of the ESR; while when the capacitance value decreases, the output power is very small and is hardly affected by the capacitance value. Integrate the high frequency band of the output voltage and power of the capacitor to obtain the high frequency band energy of the output voltage. The analysis results are shown in Figure 9. The voltage high frequency band energy increases obviously with the increase of ESR, and the influence on the high frequency band energy after the capacitance value decreases can be ignored. Therefore, monitoring the high frequency band energy of the output voltage of the capacitor can effectively reflect the failure state of the capacitor. The MPPF capacitor failure detection process is shown in Figure 10. The specific process is as follows:

Step 1: Obtain the capacitor output voltage of the capacitor to be evaluated; the expression of the capacitor output voltage is shown in equation (6).

Step 2: Use Fourier analysis to perform frequency domain analysis on the capacitor output voltage signal to obtain the spectral signal ΔV(ω).

Step 3: Calculate the power spectrum P(ω) of the capacitor output voltage frequency domain signal.

Step 4: Perform time integration on the high frequency band of the power spectrum P(ω) to obtain the high frequency band energy E(w) of the output voltage; the expression of the high frequency band energy of the output voltage is shown in equation (8).

Step 5: Calculate the capacitance failure degree D.

D=(E(w)-E ₀ (w))/ΔE

where E ₀ (w) is the initial high frequency band energy, and ΔE is the increased energy when the capacitor fails.

Step 6: Determine whether the capacitor fails according to the capacitor failure degree D: if the capacitor failure degree D=1, the capacitor has failed, otherwise the capacitor is normal.

Claims (3)

1. A failure evaluation method for an MPPF capacitor in an MMC system based on band energy is characterized by comprising the following steps:

step 1: acquiring the capacitance output voltage of an MPPF capacitor to be evaluated, wherein the MPPF capacitor is a metalized polypropylene film capacitor;

and 2, step: performing frequency domain analysis on the capacitor output voltage signal by using Fourier analysis to obtain a frequency spectrum signal delta V (omega);

and step 3: calculating a power spectrum P (omega) of the capacitor output voltage frequency domain signal;

and 4, step 4: performing time integration on the high-frequency band of the power spectrum P (omega) to obtain output voltage high-frequency band energy E (w);

and 5: calculating the capacitance failure degree D:

D＝(E(w)-E ₀ (w))/ΔE

wherein E is ₀ (w) is the initial high-band energy, which is the high-band energy of the capacitor in a healthy state; Δ E is the energy added when the capacitor fails;

step 6: judging whether the capacitor fails according to the capacitor failure degree D: if the capacitor failure rate D is 1, the capacitor fails, otherwise, the capacitor is normal.

2. The MPPF capacitor failure evaluation method in band energy based MMC system of claim 1, wherein step 1 specifically includes:

connecting the equivalent series resistance, the equivalent series capacitance and the equivalent series inductance in series to establish a capacitor equivalent model; the voltage drop on the capacitor to be measured is the sum of the voltage drops on the resistor, the capacitor and the inductor:

v(t)＝v _ESR (t)+v _C (t)+v _L (t)

wherein v is _ESR (t) is the voltage drop, v, produced by the equivalent series resistance _C (t) partial voltage drop of capacitance, v _L (t) is the voltage drop across the equivalent series inductance;

define the loss tangent as the ratio of ESR to capacitance impedance:

wherein, ESR is equivalent series resistance value, omega, of the capacitor to be measured ₀ Taking 2 × pi × 50 as the working angular frequency; c is the capacitance value of the capacitor to be measured;

neglecting the effect of the equivalent series inductance, the capacitor output voltage is expressed as follows:

wherein i _arm Is the bridge arm current; theta (t-t) ₀ ) Is t ₀ Step signal of time, t ₀ Starting time for charging the capacitor; v. of _c (t ₀ ) Is t ₀ The capacitance part of the moment drops.

3. The MPPF capacitor failure evaluation method in the MMC system based on band energy of claim 2, characterized in that, in step 4, the increment of the capacitance voltage is analyzed in frequency domain, obtaining the expression of the high-band energy as follows:

wherein i _ac Is the current flowing through the capacitor; omega is an angular frequency independent variable; delta is the dielectric loss angle; t is the integration time.

Images