CN116184037A - Resonance capacitance state monitoring method of medium-voltage direct-current LLC resonant converter - Google Patents

Abstract

The invention relates to a method for monitoring the resonance capacitance state of a medium-voltage direct-current LLC resonant converter, which comprises the steps of closing a switching tube S of a full bridge at the initial time of stopping the LLC resonant converter ₁₁ ～S ₄₁ Then, the switching tube S is turned on after the closing process lasts 500ns ₃₁ And S is ₄₁ Forming a resonant cavity discharge loop independent of a rear stage, collecting time sequence data of resonant current in the resonant cavity, estimating the resonant capacitance by adopting a nonlinear parameter identification algorithm, sequentially performing operations of coding, initializing population, calculating individual fitness, selecting crossover variation evolution, iterating and optimizing during estimation, finally searching to obtain an optimal resonant current estimated value, and calculating to obtain the resonant capacitance estimated value based on the corresponding resonant frequency and resonant inductance. The invention is based on limited sampling dataThe discharge curve is used for estimating the resonance capacitance value, the cost is low, the reliability is high, and the nonlinear parameter identification algorithm is used for estimating the parameters, so that the estimation accuracy is high.

Description

Resonance capacitance state monitoring method of medium-voltage direct-current LLC resonant converter

Technical Field

The invention relates to the technical field of direct current converter state monitoring, in particular to a method for monitoring the state of a resonant capacitor of a medium-voltage direct current LLC resonant converter.

Background

The LLC resonant converter is a reliable and efficient medium-voltage direct-current converter structure, can be applied to medium-voltage direct-current occasions, and compared with a conventional resonator, the advantages of the LLC resonant converter can be summarized as the following points, and firstly, the LLC resonant converter has the adjusting capability of wide input voltage and wide output load; secondly, soft switching technology, namely zero voltage switching, is realized in the operation range, and the power supply efficiency is improved; third, parasitic elements including junction capacitance of the semiconductor device, leakage inductance and excitation inductance of the transformer in the loop can be used to realize zero-voltage switching.

In general, a typical LLC resonant topology includes four parts, namely a switching network, a resonant network, a transformer, and a rectifying and filtering network. The resonant cavity of the LLC resonant converter consists of a resonant capacitor and a resonant inductor, and the energy transmission of the converter is realized mainly through the resonant process of the resonant capacitor and the resonant inductor. Capacitors are one of the most prominent components of modern power electronic systems in failure rate, and face severe operating environments such as high temperature, high humidity, etc., and the limitation of heat dissipation volume in high power density power electronic systems also exacerbates the reliability challenges. In the resonance process, the resonance capacitor can cause unavoidable ageing of the device due to frequent charge and discharge. In particular, when LLC resonant converters based on silicon carbide devices operate to reduce losses by increasing the operating frequency, the risk of resonant capacitance aging failure is further increased. Then, once the resonant capacitor ages out, efficient and reliable operation of the LLC resonant converter is severely compromised. Improvements to address the above reliability issues can be broadly divided into three categories: firstly, improving a design link to improve the reliability of device design; secondly, based on the operated capacitor, the circuit operation conditions including topological structure, voltage, temperature, ripple current and the like are performed to achieve better robustness; third, status monitoring is implemented on the operating capacitor to ensure operation under reliable conditions and as a basis for discrimination for performing preventative maintenance. In view of feasibility and implementation value, achieving accurate capacitive operating state monitoring is of great importance for reliable operation of the system.

At present, the related research of capacitance monitoring is mainly based on a direct current link capacitor, and some resonant capacitance monitoring methods are not yet seen. The traditional direct current link capacitance monitoring method can be divided into a periodic small signal ripple-based method, an aperiodic large signal charge-discharge curve-based method and a black box model-based method according to principles. The artificial intelligence method based on the black box model is limited by difficult acquisition of a training data set and uncertain actual working conditions in actual application. Secondly, the method based on the periodic small signal ripple principle mainly comprises a circuit model method and a signal injection method. The establishment of an accurate model is the leading basis of a circuit model method, however, the extraction of voltage ripple signals is quite difficult to realize, more interference exists, the precision is difficult to ensure, and meanwhile, the change of working conditions requires the re-modeling and has poor universality. The signal injection method affects the normal operation of the system to a certain extent, and meanwhile, the burden of the control system is increased by introducing a large number of digital filters. In contrast, the non-periodic large signal charge-discharge curve does not need complex modeling and high-frequency sampling, the implementation condition is relatively simple, and the accuracy is good.

Disclosure of Invention

The invention aims to solve the technical problem of providing a resonant capacitor state monitoring method of a medium-voltage direct-current LLC resonant converter, which is simple and effective in realizing the state monitoring of the resonant capacitor in the LLC resonant converter by using a nonlinear parameter identification algorithm based on a non-periodic large signal principle.

In order to solve the technical problems, the invention adopts the following technical methods: a resonance capacitance state monitoring method of a medium voltage direct current LLC resonant converter comprises the following steps:

step S1, closing a switching tube S of the full bridge at the initial time of shutdown of the LLC resonant converter ₁₁ ～S ₄₁ ；

Step S2, the switching tube S of the full bridge is turned on after the off state in step S1 is continued for 500ns ₃₁ And S is ₄₁ Forming a resonant cavity discharge loop independent of the rear stage;

s3, collecting time sequence data of resonant current in the resonant cavity, and estimating the resonant capacitance by adopting a nonlinear parameter identification algorithm;

step S301, encoding: coding the identification parameters such as initial voltage, initial phase angle, resonant frequency and attenuation coefficient into a quaternary vector;

step S302, initializing a population: defining upper and lower bounds of variables according to physical meanings of initial voltage, initial phase angle, resonant frequency and attenuation coefficient, and randomly generating a population;

step S303, calculate the fitness: based on initial voltage, initial phase angle, resonant frequency and attenuation coefficient in an individual, obtaining an estimated resonant current sequence according to an RLC discharge estimation model shown in the following formula (1), solving a second norm by performing difference between the estimated resonant current sequence and acquired resonant current time sequence data, and taking the reciprocal as fitness;

wherein I is _Lr For resonant current, I ₀ For the initial current signal amplitude of the discharge phase,

for the initial phase angle, ω is the resonant frequency, τ is the attenuation coefficient;

step S304, individual evolution: firstly, selecting a plurality of better individuals by utilizing the fitness of each individual based on a selection operator and eliminating a plurality of worse individuals; then randomly setting crossing points in the paired individual code strings according to the selection probability and exchanging genes with each other to form a new individual; finally, randomly selecting variation points according to variation probability, and randomly searching new individuals at the variation points;

step S305, iterative optimization: repeating steps S303 and S304, searching to obtain an optimal resonance current estimated value under the set iteration times, calculating to obtain a resonance capacitance estimated value based on the corresponding resonance frequency and the known resonance inductance value by adopting the following formula (2):

further, the method further comprises the step S4: the resonance capacitance estimation is corrected according to the influence of the ambient temperature.

Preferably, in step S4, the formula for correcting the estimated value of the resonance capacitance is as follows:

wherein alpha is _M 、β _M 、γ _M The temperature characteristic parameters of the capacitor are all temperature characteristic parameters of the capacitor, and for specific types of capacitors, the three temperature characteristic parameters can be measured through experiments; t (T) _a 、T _a，min 、T _a,max Respectively representing the actual operating temperature, the minimum operating temperature and the maximum operating temperature of the capacitor.

The invention provides a resonance capacitance state monitoring method of a medium-voltage direct-current LLC resonant converter, which mainly utilizes a nonlinear parameter identification algorithm based on a non-periodic large signal principle to process current data acquired during discharging of a resonant cavity of the LLC resonant converter so as to realize state monitoring of a resonant capacitance in the LLC resonant converter. Compared with the traditional capacitance state monitoring method, the method for estimating the resonance capacitance value based on the discharge curve of the limited sampling data is low in cost and high in reliability, and the method utilizes a basic group intelligent algorithm-genetic algorithm in a nonlinear parameter identification algorithm to estimate parameters, so that the estimation accuracy is high. In general, the method is simple and reliable, and can effectively realize the state monitoring of the resonant capacitor in the LLC resonant converter.

Drawings

Fig. 1 is a circuit diagram of a medium voltage dc LLC resonant converter in accordance with the invention;

FIG. 2 is a waveform diagram of resonant current during RLC discharge in accordance with the present invention;

FIG. 3 is a graph showing resonant current data points obtained at different sampling frequencies according to the present invention.

Detailed Description

The invention will be further described with reference to examples and drawings, to which reference is made, but which are not intended to limit the scope of the invention.

The invention relates to a medium-voltage direct-current LLC resonant converter which comprises a switch network, a resonant network, a transformer and a rectifying and filtering network, wherein the specific structure is shown in the figure 1, and V is shown in the figure _in 、R _L The input voltage and the load resistance of the LLC resonant converter are respectively. The input side of LLC resonant converter is parallelly connected with a switching tube Q ₁ And is connected in parallel with a capacitor C formed by DC-Link _in1 And bleeder resistor R _in1 An RC circuit is formed. The switching network of the LLC resonant converter comprises a power switching tube S ₁₁ ～S ₄₁ Body diode D ₁₁ ～D ₄₁ The resonant network comprises a resonant capacitor C _r1 Resonant inductance L _r1 And excitation inductance L _m1 The transformer is T ₁ (the turn ratio of the primary side to the secondary side is m: 1), the rectifying and filtering network comprises a rectifying diode D ₅₁ ～D ₈₁ Composed rectifier bridge and filter output capacitor C connected with the rectifier bridge in parallel _f1 。

Based on the medium-voltage direct-current LLC resonant converter, the invention provides a method for monitoring the resonant capacitance state of the medium-voltage direct-current LLC resonant converter, and the implementation process of the method is as follows.

Step S1, closing a switching tube S of the full bridge at the initial time of shutdown of the LLC resonant converter ₁₁ ～S ₄₁ Under the specific implementation condition, the working mode of the LLC resonant converter at the initial time of shutdown has uncertainty, and the switching-off process of the switching tube lasts about 200ns, so that the LLC resonant converter is avoided from the viewpoint of safe operationThe working mode is free from direct switching, all the switching tubes of the full bridge are closed as an excessive stage, the duration of the stage is set to be 500ns, and the energy of the resonant cavity is stored in the resonant inductance and the resonant capacitance at the moment.

Step S2, after the closing phase of step S1 is completed, switching on the switching tube S of the full bridge of the LLC resonant converter ₃₁ And S is ₄₁ Resonance capacitor C _in1 Resonant inductance L _r1 And excitation inductance L _m1 And a switch tube S ₃₁ And S is ₄₁ The on-resistance of (a) forms a closed loop, and generates an RLC resonance process, RLC discharge is performed in the resonant cavity, the discharge waveform is a sine wave with attenuated amplitude, as shown in fig. 3, and the discharge estimation model is shown in the following formula (1):

wherein I is _Lr For resonant current, I ₀ For the initial current signal amplitude of the discharge phase,

for the initial phase angle ω is the resonant frequency and τ is the decay coefficient.

And S3, collecting time sequence data of resonant current in the resonant cavity, and estimating the resonant capacitance by adopting a nonlinear parameter identification algorithm.

The invention collects time sequence data of resonance current in the discharging process at a certain sampling frequency, as shown in figure 3, the data points of the resonance current are different at different sampling frequencies, and the invention performs capacitance parameter identification according to the collected data points. The traditional parameter identification algorithm comprises a least square method, an augmented least square method, a gradient approximation method and a maximum likelihood estimation method, has the defects of being limited to a linear system, single-point searching, being easy to be limited depending on an initial point, and the nonlinear system identification algorithm comprises a group intelligent algorithm and a neural network, has the advantages of strong multi-point searching/nonlinear fitting capacity and is suitable for processing a nonlinear system. In view of the above, the invention selects the nonlinear parameter identification algorithm to identify the capacitance parameters of the collected data points, the identification parameters comprise four parameters including initial voltage, initial phase angle, resonant frequency and attenuation coefficient, the limited number of parameters is considered, and the range of the parameter definition domain is clear by combining with the practical physical meaning, so the invention adopts the basic group intelligent algorithm-genetic algorithm to carry out efficient solution on the capacitance parameters, and the method is as follows.

Step S301, encoding. The group intelligent algorithm is to operate actual decision variables of the optimization problem, the decision variables are expressed into serial structure data through coding, identification parameters including initial voltage, initial phase angle, resonant frequency and attenuation coefficient are expressed as quaternary vectors, the initial voltage is set to be the amplitude of a sine wave under normal operation conditions and is related to the operation conditions, then the initial phase angle is set to be 0, the resonant frequency is obtained according to the resonant inductance value and the resonant capacitance value of the initial design of the resonant cavity, and the attenuation coefficient is obtained according to the resonant inductance value and the loop resistance value of the initial design.

Step S302, initializing a population. And defining upper and lower bounds of variables according to physical meanings of identification parameters such as initial voltage, initial phase angle, resonant frequency and attenuation coefficient, and randomly generating a population.

Step S303, calculate the fitness. Based on initial voltage, initial phase angle, resonant frequency and attenuation coefficient in an individual, substituting the initial voltage, initial phase angle, resonant frequency and attenuation coefficient into an RLC discharge estimation model shown in formula (1) to obtain an estimated resonant current sequence, based on a time sequence of acquired data, performing difference solution on the estimated resonant current sequence and the acquired resonant current time sequence data to obtain a second norm, taking the inverse as fitness, wherein the more accurate the estimated resonant current parameter is, the larger the fitness is.

Step S304, individual evolution, including selection, crossover and mutation.

Selecting: the process uses the adaptation value of each individual to eliminate some worse individuals and select some better individuals, namely, select each group of solutions with relatively more accurate parameter estimation, so as to perform the next step of crossing and mutation operation.

Crossing: the crossover operator is realized by adopting a single-point crossover method, namely, a crossover point is randomly set in the individual code strings paired in pairs according to the selection probability, and then partial genes of the two paired individuals are mutually exchanged at the crossover point, namely, any variable in the two groups of solutions, namely, initial voltage, initial phase angle, resonant frequency and attenuation coefficient, are exchanged, so that two new individuals are formed.

Variation: the operation is to randomly select a variation point according to the variation probability, and randomly search for a new individual at the variation point.

Step S305, iterative optimization: repeating steps S303 and S304, searching to obtain an optimal resonance current estimated value under the set iteration times, determining a good resonance inductance value based on the corresponding resonance frequency and the power supply equipment design, and calculating to obtain a resonance capacitance estimated value by adopting the following formula (2):

step S4, correcting the estimated value C of the resonance capacitor _r : since the operating frequency of the resonant capacitor in the LLC resonant converter is determined based on the fixed resonant frequency, the effect of the change in the switching frequency can be eliminated from the correction, and only the effect of the ambient temperature can be considered, and the capacitance estimation value C obtained in step S304 can be obtained by using the following formula (3) _r And (3) correcting:

wherein alpha is _M 、β _M 、γ _M The temperature characteristic parameters of the capacitor are all temperature characteristic parameters of the capacitor, and for specific types of capacitors, the three temperature characteristic parameters can be measured through experiments; t (T) _a 、T _a，min 、T _a,max Respectively representing the actual operating temperature, the minimum operating temperature and the maximum operating temperature of the capacitor.

The foregoing embodiments are preferred embodiments of the present invention, and in addition, the present invention may be implemented in other ways, and any obvious substitution is within the scope of the present invention without departing from the concept of the present invention.

In order to facilitate understanding of the improvements of the present invention over the prior art, some of the figures and descriptions of the present invention have been simplified, and some other elements have been omitted from this document for clarity, as will be appreciated by those of ordinary skill in the art.

Claims (3)

1. A method for monitoring the resonance capacitance state of a medium-voltage direct-current LLC resonant converter is characterized by comprising the following steps:

step S1, closing a switching tube S of the full bridge at the initial time of shutdown of the LLC resonant converter ₁₁ ～S ₄₁ ；

Step S2, the switching tube S of the full bridge is turned on after the closing process in step S1 is continued for 500ns ₃₁ And S is ₄₁ Forming a resonant cavity discharge loop independent of the rear stage;

s3, collecting time sequence data of resonant current in the resonant cavity, and estimating the resonant capacitance by adopting a nonlinear parameter identification algorithm;

step S301, encoding: coding the identification parameters such as initial voltage, initial phase angle, resonant frequency and attenuation coefficient into a quaternary vector;

step S302, initializing a population: defining upper and lower bounds of variables according to physical meanings of initial voltage, initial phase angle, resonant frequency and attenuation coefficient, and randomly generating a population;

step S303, calculate the fitness: based on initial voltage, initial phase angle, resonant frequency and attenuation coefficient in an individual, obtaining an estimated resonant current sequence according to an RLC discharge estimation model shown in the following formula (1), solving a second norm by performing difference between the estimated resonant current sequence and acquired resonant current time sequence data, and taking the reciprocal as fitness;

wherein I is _Lr For resonant current, I ₀ For the initial current signal amplitude of the discharge phase,

for the initial phase angle, ω is the resonant frequency, τ is the attenuation coefficient;

step S304, individual evolution: firstly, selecting a plurality of better individuals by utilizing the fitness of each individual based on a selection operator and eliminating a plurality of worse individuals; then randomly setting crossing points in the paired individual code strings according to the selection probability and exchanging genes with each other to form a new individual; finally, randomly selecting variation points according to variation probability, and randomly searching new individuals at the variation points;

step S305, iterative optimization: repeating steps S303 and S304, searching to obtain an optimal resonance current estimated value under the set iteration times, calculating to obtain a resonance capacitance estimated value based on the corresponding resonance frequency and the known resonance inductance value by adopting the following formula (2):

2. the method for monitoring the resonant capacitance state of a medium voltage direct current LLC resonant converter in accordance with claim 1, wherein: and step S4, correcting the estimated value of the resonance capacitance according to the influence of the ambient temperature.

3. The method for monitoring the resonant capacitance state of a medium voltage direct current LLC resonant converter in accordance with claim 2, wherein: in the step S4, the formula for correcting the estimated value of the resonance capacitance is as follows:

wherein alpha is _M 、β _M 、γ _M The temperature characteristic parameters of the capacitor are all temperature characteristic parameters of the capacitor, and for specific types of capacitors, the three temperature characteristic parameters can be measured through experiments; t (T) _a 、T _a，min 、T _a,max Respectively representing the actual operating temperature, the minimum operating temperature and the maximum operating temperature of the capacitor.

Images