Capacitance monitoring method and system for wall bushing of flexible direct current converter station

Abstract

The invention discloses a capacitance monitoring method and system for a wall bushing of a flexible direct current converter station, and relates to the technical field of online monitoring and electrical parameter identification of flexible direct current transmission equipment, and the method comprises the steps: arranging a capacitive voltage divider between an end screen of the wall bushing of the flexible direct current converter station and the ground, and constructing a series capacitance measurement structure; a primary voltage signal and an end screen current of the bushing are collected. Performing synchronous signal preprocessing on the primary voltage signal and the end screen current, and removing noise and extracting frequency domain characteristics by adopting wavelet transform and a filtering strategy; based on the frequency domain characteristics, calculating the coupling relation between sleeve structure parameters and capacitance, establishing a capacitance change identification model, and judging abnormal points in combination with temperature and humidity data. According to the method, the problems of static analysis lag, single variable misjudgment, insufficient spectrum resolution and the like are effectively solved, the real-time monitoring of the capacitance of the flexible direct-current wall bushing, the abnormal dynamic early warning and the systematic improvement of the model adaptability are realized, and the safe operation guarantee capability of equipment in the whole life cycle is remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of online monitoring and electrical parameter identification of flexible direct current (DC) transmission equipment, and in particular to a method and system for monitoring the capacitance of a wall bushing in a flexible DC converter station. Background Art

[0002] As a key form of next-generation HVDC transmission, Flexible DC transmission technology is gaining widespread application in areas such as long-distance power transmission and renewable energy grid integration, thanks to its advantages, such as strong converter controllability and adaptability. In Flexible DC commutation systems, wall bushings serve as the connection between AC and DC systems. Their insulation performance and operational stability are crucial to the safe operation of the system. In recent years, with the intensification of harmonic pollution and the continuous increase in system voltage levels, the health status monitoring of wall bushings has gradually evolved from offline testing to online monitoring. Research has shifted its focus to accurately acquiring key operating parameters, such as bushing capacitance, to enable dynamic identification and early warning control of insulation aging trends.

[0003] Although some literature has attempted to introduce fiber optic electric field sensors and harmonic analysis methods for bushing condition monitoring, several limitations persist in practical engineering scenarios. First, existing capacitance detection methods mostly rely on offline measurements or periodic calibration, which cannot capture capacitance variation trends in real time and cannot meet the requirements of flexible DC commutation systems for rapid response to high-frequency transient anomalies. Second, traditional capacitance estimation methods lack the ability to resist interference from harmonic components in high-frequency current disturbance environments, especially in power frequency environments. Furthermore, existing monitoring systems lack a fusion modeling mechanism for multi-source signals (such as structural parameters, ambient temperature and humidity, and frequency-domain energy characteristics), resulting in large model prediction errors and unreliable early warning results. More critically, traditional capacitance determination methods often rely on static threshold settings and cannot adapt to environmental changes. This is prone to misjudgments or missed detections, limiting the ability to early identify potential insulation degradation in actual operation. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing wall bushing capacitance monitoring method is unable to realize online capacitance estimation, lacks an efficient frequency domain feature extraction mechanism, and is difficult to cope with dynamic interference from environmental factors; and how to realize dynamic identification and abnormal warning of wall bushing capacitance based on the fusion of structural modeling and frequency domain features.

[0006] To solve the above technical problems, the present invention provides the following technical solution: a capacitance monitoring method for a wall bushing in a flexible DC converter station, comprising setting a capacitive voltage divider between the end screen of the wall bushing in the flexible DC converter station and the ground and constructing a series capacitance measurement structure to collect the bushing primary voltage signal and the end screen current.

[0007] The primary voltage signal and the end-screen current are synchronously preprocessed, and wavelet transform and filtering strategies are used to remove noise and extract frequency domain features.

[0008] The coupling relationship between bushing structural parameters and capacitance is calculated based on frequency domain characteristics, a capacitance change identification model is established, and abnormal points are identified by combining temperature and humidity data.

[0009] The frequency domain feature calculation of the coupling relationship between the bushing structural parameters and the capacitance includes: inputting the length, inner diameter, and dielectric constant of the wall bushing as structural parameters; constructing a capacitance estimation function based on an equivalent capacitance coupling coefficient model in combination with the frequency domain features; importing the model through a table lookup; the estimation function uses the spectral energy ratio and the main frequency change rate in the frequency domain features as independent variables; and using the capacitance change trend per unit length as the output result; and performing standardization and normalization processing in combination with historical data to form a standard coupling curve for subsequent abnormal point comparison.

[0010] As a preferred solution of the capacitance monitoring method of the wall bushing of the flexible DC converter station described in the present invention, the provision of a capacitive voltage divider includes connecting two capacitors in series between the end screen and the ground to form a voltage divider measurement circuit, wherein the two capacitors are a main capacitor and a detection capacitor, the main capacitor is connected in parallel in the wall bushing body, the detection capacitor is connected to the voltage acquisition node, the voltage change across the main capacitor is measured, and a potential difference change curve between the end screen of the wall bushing and the ground is obtained, thereby realizing indirect derivation of the capacitance value.

[0011] As a preferred solution of the capacitance monitoring method for the wall bushing of a flexible DC converter station described in the present invention, the construction of a series capacitance measurement structure includes: arranging multiple broadband current sensors and high-voltage terminal voltage lead wires in a ring around the wall bushing housing; the high-voltage terminal voltage is led out through an insulated cable and connected to a voltage sampling circuit; the end-screen current is measured using a Rogowski coil and collected synchronously with the voltage signal; and a minimum sampling frequency threshold is set.

[0012] As a preferred embodiment of the capacitance monitoring method for wall bushings in flexible DC converter stations described herein, the synchronous signal preprocessing of the primary voltage signal and the end-panel current includes digital denoising of the voltage and current signals. The voltage signal is filtered using a bandpass filter based on frequency band filtering to remove power frequency interference, and the current signal is transformed using a wavelet transform to extract transient spectral components. After denoising, the voltage and current signals are time-domain aligned using timestamps.

[0013] As a preferred solution of the capacitance monitoring method of the wall bushing of the flexible DC converter station described in the present invention, the method of removing noise and extracting frequency domain features using wavelet transform and filtering strategy includes using Daubechies wavelet function for six-layer decomposition to extract high-frequency harmonic components, obtaining a spectrum distribution map through power spectral density analysis, and extracting characteristic point parameters within the target frequency band according to the set frequency band weight as input features of the subsequent capacitance coupling model. The energy proportion of each frequency band is used as an important dimension of the model feature vector to participate in the subsequent coupling relationship modeling.

[0014] As a preferred solution of the capacitance monitoring method of the wall bushing of the flexible DC converter station described in the present invention, the method comprises: extracting characteristic points within the target frequency band, dividing 1kHz to 10MHz into five sub-bands based on the frequency band division, and selecting the main peak frequency of the harmonic, the total energy of the frequency band, the maximum amplitude of the frequency band and the transient slope in each sub-band as characteristic point parameters. After the energy of each frequency band is normalized, a 20-dimensional feature vector is formed to describe the energy distribution characteristics of the high-frequency harmonics in the entire frequency domain. Before the feature vector is input into the capacitance coupling relationship model, a principal component analysis dimensionality reduction process is performed to reduce the modeling dimension and retain the key change information of the spectrum.

[0015] As a preferred embodiment of the capacitance monitoring method for wall bushings in flexible DC converter stations described in the present invention, the frequency-domain feature calculation of the coupling relationship between bushing structural parameters and capacitance includes obtaining the bushing's structural parameters, including its length, internal diameter, and dielectric properties, and establishing a reference capacitance per unit length based on the bushing's physical structure. Based on the high-frequency energy ratio and main frequency variation obtained from the frequency-domain feature extraction, a characteristic index describing changes in electrical behavior is constructed. This characteristic index is combined with the structural parameters and input into a preset estimation function. The estimation function is trained using historical operating data to obtain a set of weight parameters suitable for different bushing types, outputting an estimated capacitance per unit length at the current moment. The estimated capacitance variation trend is compared with a standard curve recorded during historical operation under healthy conditions, and all data is standardized. During the comparison, if there is a persistent deviation between the estimated curve and the standard curve, and the deviation exceeds a preset threshold, it indicates that the current bushing's capacitance characteristics have undergone an abnormal change. The time point and segment corresponding to the deviation are marked.

[0016] As a preferred solution of the capacitance monitoring method for the wall bushing of the flexible DC converter station described in the present invention, the method of establishing a capacitance change identification model and combining temperature and humidity data to determine abnormal points includes calculating the difference between the standardized capacitance change curve and the current measured curve, and setting temperature and humidity correction coefficients to dynamically adjust the difference threshold. The temperature and humidity are collected in real time by on-site environmental sensors, and the correction factor is calculated by regression analysis. If the difference exceeds the correction threshold, it is considered that the capacitance change is abnormal, the abnormal section is located, and the start and end positions of the abnormal area, the capacitance change rate and the corresponding timestamp are output.

[0017] As a preferred embodiment of the capacitance monitoring method for wall bushings in flexible DC converter stations described in the present invention, the regression analysis method includes selecting multiple sets of difference samples of ambient temperature, humidity, and capacitance curves under the same load conditions from historical bushing operation data, and performing a multivariate linear fit of the temperature sensitivity coefficient and the humidity sensitivity coefficient using the least squares method to obtain a regression equation. The temperature sensitivity coefficient represents the correction ratio for the difference amplitude caused by a 1°C change, and the humidity sensitivity coefficient represents the error tolerance for a 1% RH change.

[0018] Another object of the present invention is to provide a capacitance monitoring system for wall bushings in flexible DC converter stations. The system can calculate the coupling relationship between bushing structural parameters and capacitance based on frequency domain characteristics through a capacitance identification and judgment module, establish a capacitance change identification model, and determine abnormal points in combination with temperature and humidity data. This system solves the problems of existing wall bushing capacitance monitoring methods, such as the inability to achieve online capacitance estimation, the lack of an efficient frequency domain feature extraction mechanism, and the difficulty in coping with dynamic interference from environmental factors. Furthermore, the system addresses the issues of how to achieve dynamic identification and abnormality warning of wall bushing capacitance based on the fusion of structural modeling and frequency domain features.

[0019] As a preferred solution of the capacitance monitoring system of the wall bushing of the flexible DC converter station described in the present invention, it includes: a signal acquisition module, a frequency domain preprocessing and extraction module, and a capacitance identification and judgment module.

[0020] The signal acquisition module is used to set a capacitive voltage divider between the end screen of the wall bushing in the flexible DC converter station and the ground and build a series capacitance measurement structure to collect the bushing primary voltage signal and the end screen current.

[0021] The frequency domain preprocessing and extraction module is used to perform synchronous signal preprocessing on the primary voltage signal and the end-screen current, and adopts wavelet transform and filtering strategy to remove noise and extract frequency domain features.

[0022] The capacitance identification and judgment module is used to calculate the coupling relationship between the casing structure parameters and the capacitance based on the frequency domain characteristics, establish a capacitance change identification model and judge abnormal points in combination with temperature and humidity data.

[0023] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a method for monitoring the capacitance of a wall bushing in a flexible DC converter station.

[0024] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for monitoring the capacitance of a wall bushing in a flexible direct current converter station.

[0025] The present invention provides a method for monitoring the capacitance of a wall bushing in a flexible DC converter station. By placing a capacitive voltage divider between the bushing's terminal screen and ground, and constructing a series capacitance measurement structure, the method collects the primary voltage signal and the terminal screen current, achieving high-fidelity acquisition of electrical signals in the presence of high-frequency harmonics. This method provides stable and accurate basic signal support for subsequent frequency-domain analysis and modeling.

[0026] The collected voltage and current signals are filtered, wavelet decomposed and spectrally analyzed to extract frequency domain features and construct normalized feature vectors for subsequent modeling input, which effectively converts complex frequency domain changes into recognizable inputs for the structural model and improves the model's discrimination accuracy and robustness.

[0027] A capacitance estimation model is constructed based on structural parameters and frequency domain characteristics. Combined with temperature and humidity data, dynamic correction and regression analysis are used to identify and warn of capacitance anomalies. This system achieves high-resolution recognition of even small capacitance fluctuations and an environmentally adaptive anomaly warning mechanism, providing a basis for early intervention decisions for system operations and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0029] Figure 1 This is an overall flow chart of a method for monitoring the capacitance of a wall bushing in a flexible DC converter station provided by the first embodiment of the present invention.

[0030] Figure 2 This is an overall schematic diagram of a capacitance monitoring system for a wall bushing in a flexible DC converter station provided in a second embodiment of the present invention. DETAILED DESCRIPTION

[0031] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0032] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a method for monitoring the capacitance of a wall bushing in a flexible DC converter station, comprising:

[0033] S1: A capacitive voltage divider is installed between the end screen of the wall bushing in the flexible DC converter station and the ground, and a series capacitance measurement structure is constructed to collect the bushing primary voltage signal and the end screen current.

[0034] During the capacitance monitoring process of the wall bushing in a flexible DC converter station, the first step is to complete signal acquisition and the construction of a measurement structure. To this end, a set of capacitive voltage divider structures is set up between the end screen of the wall bushing and the ground to accurately capture the electrical signal and indirectly derive the capacitance. The capacitive voltage divider forms a measurement circuit by connecting two capacitors in series between the end screen and the ground wire. One capacitor is the main capacitor, arranged directly in parallel with the bushing body, and carries the majority of the voltage. The other is a detection capacitor, one end of which is connected to the end screen of the bushing and the other end is connected to the voltage acquisition node. The voltage borne by the detection capacitor is collected in real time by the measurement circuit and, combined with the known main capacitor parameters, indirectly converted into the change in the potential difference between the entire end screen and the ground, thus providing the basic input for subsequent capacitance estimation.

[0035] To further improve the spatial resolution of collected data and the integrity of current monitoring, several broadband current sensors are evenly distributed around the circumference of the wall bushing. These sensors are fixed in a ring configuration, ensuring comprehensive acquisition of induced current signals around the bushing from multiple angles. The current sensors are either capacitively coupled or Rogowski coil types, and their detection range covers a wide frequency band from power frequency to several megahertz. Simultaneously, a primary voltage signal is extracted from the high-voltage end of the bushing via an insulated lead wire and connected to the voltage sampling circuit via a dedicated voltage lead structure, ensuring safe isolation and effective acquisition of the high-voltage signal.

[0036] To ensure synchronization and sampling accuracy of voltage and current signals in the time domain, all acquisition channels are equipped with a unified clock trigger control. The signal acquisition frequency is set to a sampling rate of at least 10 megapoints per second, ensuring that complete waveform characteristics can be captured even in harmonic conditions containing high-frequency components. This acquisition frequency parameter is pre-set based on the rated operating voltage level of the converter station and the target monitoring frequency bandwidth, with a fixed lower limit that is no less than a multiple of the minimum bandwidth required for monitoring.

[0037] Through the above structural arrangement and sampling logic control, step S1 completes the synchronous high-fidelity acquisition of the primary voltage signal and the end-screen current signal of the wall bushing, and provides a high-integrity and computable signal basis for subsequent preprocessing and coupling calculations.

[0038] S2: Perform synchronous signal preprocessing on the primary voltage signal and the end-screen current, and use wavelet transform and filtering strategy to remove noise and extract frequency domain features.

[0039] The voltage and current signals are digitally denoised. A bandpass filter based on frequency filtering is used to remove power frequency interference from the voltage signal, while a wavelet transform is used to extract the transient spectral components of the current signal. After denoising, the voltage and current signals are time-aligned using timestamps.

[0040] Furthermore, in order to remove the power frequency interference in the voltage signal, the bandpass filter response function is defined as follows:

[0041]

[0042] Among them, H bp (f m ) represents the frequency f m The bandpass filter transfer function value. f m Represents the mth frequency component in the input signal. l Indicates the lower limit frequency of the bandpass filter. h Indicates the upper limit frequency of the bandpass filter. The setting range is to eliminate 50Hz power frequency interference and retain high-frequency harmonic components.

[0043] The Daubechies wavelet function is used for six-layer decomposition to extract high-frequency harmonic components. The spectrum distribution map is obtained through power spectral density analysis. The characteristic point parameters in the target frequency band are extracted according to the set frequency band weights as the input features of the subsequent capacitance coupling model. The energy proportion of each frequency band is used as an important dimension of the model feature vector to participate in the subsequent coupling relationship modeling.

[0044] Furthermore, the wavelet decomposition coefficient calculation is expressed as:

[0045]

[0046] Among them, W c (s n ,t p ) represents the nth level scale and time point t p The wavelet decomposition coefficients under s n Indicates the wavelet scale of the nth layer (corresponding to the frequency band). q Indicates the qth sampling channel in the original signal at time tp The numerical value of ψ n,q (s n ) represents the Daubechies wavelet basis function, at scale s n The response function of the qth channel on . Q represents the total number of channels.

[0047] Extracting feature points within the target frequency band involves dividing the 1kHz to 10MHz frequency band into five sub-bands based on frequency segmentation. Within each sub-band, the harmonic peak frequency, total energy, maximum amplitude, and transient slope are selected as feature point parameters. Each band energy is normalized to form a 20-dimensional feature vector that describes the energy distribution of high-frequency harmonics across the entire frequency domain.

[0048] Furthermore, the band-normalized energy vector calculation is expressed as:

[0049]

[0050] in, Represents the normalized energy value of the rth frequency band. E r Represents the total original energy value in the rth frequency band. The sum of the energy in all five frequency bands. r and u represent the subscripts of the frequency band numbers.

[0051] Before the eigenvector is input into the capacitance coupling relationship model, a principal component analysis dimensionality reduction process is performed to reduce the modeling dimension and retain the key spectrum change information.

[0052] Furthermore, the principal component analysis feature dimensionality reduction is expressed as:

[0053]

[0054] Among them, V j Represents the j-th dimension of the principal component vector. ω j,k Represents the weight coefficient projected from the normalized energy dimension k to the principal component dimension j. represents the kth component of the normalized energy vector. j represents the dimension number of the principal component after dimensionality reduction. k represents the dimension number of the frequency band energy before dimensionality reduction. K represents the dimension of the original feature vector (20 in the present invention).

[0055] It should be noted that step S2 of the present invention realizes the accurate extraction of high-frequency harmonics and the fusion of structural features by constructing a multi-stage frequency domain signal processing chain. First, bandpass filtering and six-layer wavelet decomposition are combined to effectively eliminate power frequency interference and capture transient high-frequency characteristics. Secondly, a multi-dimensional feature vector based on frequency band energy ratio, main frequency, etc. is designed, and normalization and principal component dimensionality reduction operations are introduced to ensure that the model input retains key information and reduces dimensional complexity. This step breaks through the problems of low resolution and single characteristic structure of spectrum analysis in the existing technology, realizes the unification of high-frequency noise robust extraction and structural coupling modeling, and improves the sensitivity and engineering adaptability of wall bushing capacitance monitoring.

[0056] S3: Calculate the coupling relationship between bushing structural parameters and capacitance based on frequency domain characteristics, establish a capacitance change identification model, and combine temperature and humidity data to determine abnormal points.

[0057] The structural parameters of the wall bushing, including its length, internal diameter, and dielectric properties, are obtained. A reference capacitance per unit length is established based on the bushing's physical structure. Based on the high-frequency energy percentage and main frequency variations obtained from frequency-domain feature extraction, a characteristic index describing changes in electrical behavior is constructed. This characteristic index, combined with the structural parameters, is then input into a pre-set estimation function.

[0058] Furthermore, the capacitance per unit length estimation function is expressed as:

[0059] C est (t r )=w1·S len +w2·S dia +w3·S ε +w4·F ratio (t r )+w5·F shift (t r )+b

[0060] Among them, C est (t r ) indicates that at time t r The estimated capacitance per unit length at the moment. S len Indicates the actual length of the wall bushing. S dia Indicates the inner diameter of the wall bushing. ε Indicates the relative dielectric constant of the casing material. F ratio (t r ) represents the proportion of high-frequency energy at the current moment. F shift (t r ) represents the rate of change of the main frequency compared to the historical curve. w1, w2, w3, w4, and w5 are the weight coefficients obtained from model training, which are used to adjust the ratio of different inputs. b represents the bias constant term.

[0061] The estimation function is trained using historical operating data to obtain a set of weight parameters suitable for different bushing types, outputting an estimated capacitance per unit length at the current moment. The estimated capacitance trend is compared with the standard curve under healthy conditions recorded during historical operation, and all data is standardized. During this comparison, if there is a persistent deviation between the estimated and standard curves, and the deviation exceeds the set judgment threshold, it indicates that the capacitance characteristics of the current bushing have undergone an abnormal change. The time point and segment corresponding to the deviation are marked.

[0062] Furthermore, the standardized interpolation is expressed as:

[0063]

[0064] Among them, D norm (t r ) represents the current time t r The normalized difference of C est (t r ) represents the current estimated capacitance per unit length. C ref (t r ) represents the standard capacitance curve value under historical health conditions. The numerator represents the capacitance deviation, and the denominator represents the standard reference. The overall measurement is the deviation ratio between the estimated result and the standard reference.

[0065] The difference between the standardized capacitance change curve and the current measured curve is calculated, and the temperature and humidity correction coefficients are set to dynamically adjust the difference threshold. The temperature and humidity are collected in real time by the on-site environmental sensor, and the correction factor is calculated through regression analysis.

[0066] Furthermore, the temperature and humidity dynamic correction threshold is calculated and expressed as:

[0067] T adj (t r )=T base +β temp ·V temp (t r )+β hum ·V hum (t r )

[0068] Among them, T adj (t r ) represents the current time t r The corrected dynamic judgment threshold. base Indicates the basic fixed threshold, used for static difference judgment. temp Indicates the temperature correction sensitivity coefficient, which indicates the influence ratio of each 1℃ temperature change on the threshold. V temp (tr ) represents the actual ambient temperature at the current time point. hum Represents the humidity correction sensitivity coefficient, which indicates the impact of each 1% RH humidity change on the threshold. V hum (t r ) represents the actual ambient humidity at the current time point.

[0069] If the difference exceeds the correction threshold, it is considered an abnormal capacitance change. The abnormal section is located, and the start and end positions of the abnormal section, the capacitance change rate, and the corresponding timestamp are output. The regression analysis method involves selecting multiple sets of difference samples of the ambient temperature, humidity, and capacitance curves under the same load conditions from the bushing operation history data. The least squares method is used to perform multivariate linear fitting of the temperature sensitivity coefficient and the humidity sensitivity coefficient to obtain the regression equation.

[0070] Furthermore, the temperature and humidity correction factor regression model is expressed as:

[0071]

[0072] in, Represents the standardized difference of the mth group of historical samples. (m) Indicates the ambient temperature corresponding to this group of samples. (m) Indicates the ambient humidity corresponding to the sample group. α is the intercept constant of the regression equation. ε (m) It represents the residual term of the mth group of samples and the regression error.

[0073] The temperature sensitivity coefficient indicates the correction ratio of the difference amplitude caused by each 1°C change, and the humidity sensitivity coefficient indicates the error tolerance change amplitude caused by each 1% RH change.

[0074] It should be noted that step S3 of the present invention constructs a capacitance estimation model by fusing structural parameters and frequency domain features, thereby realizing dynamic estimation and anomaly identification of the capacitance per unit length of the wall bushing. A multi-dimensional feature input function is used to model the capacitance change trend, and a standardized curve comparison mechanism is introduced to ensure that the estimation result is consistent with the historical health status. In addition, a dynamic threshold adjustment method based on temperature and humidity factors is proposed, and regression analysis is used to adaptively correct the judgment criteria, thereby effectively avoiding false alarms caused by environmental fluctuations. This step breaks through the limitations of the existing technology that only relies on static thresholds or single indicators for judgment, and realizes high-precision anomaly detection capabilities with multi-parameter coupling and electrical-environmental collaboration.

[0075] Example 2, reference Figure 2 , which is an embodiment of the present invention, provides a capacitance monitoring system for a wall bushing in a flexible DC converter station, including a signal acquisition module 100, a frequency domain preprocessing and extraction module 200, and a capacitance identification and judgment module 300.

[0076] S4: The signal acquisition module 100 is used to set a capacitive voltage divider between the end screen of the wall bushing of the flexible DC converter station and the ground and build a series capacitance measurement structure to collect the bushing primary voltage signal and the end screen current.

[0077] It includes a capacitive voltage divider arrangement submodule 101 and a high-frequency electrical signal acquisition submodule 102 .

[0078] Furthermore, the capacitor divider placement submodule 101 is used to configure a high-voltage adapter capacitor between the terminal shield of the flexible DC converter station's wall bushing and the ground. This also creates a series capacitance measurement channel, which extracts the high-voltage side signal to reduce its amplitude for easier sampling. The high-frequency electrical signal acquisition submodule 102 is used to synchronously collect the primary-side voltage signal and the terminal shield ground current of the wall bushing, using a broadband current transformer to achieve full-band response to high-frequency components.

[0079] It should be noted that the capacitive voltage divider placement submodule 101 provides secure high-voltage signal access for the entire signal acquisition module and is the foundation for the system's non-invasive capacitance monitoring. The high-frequency electrical signal acquisition submodule 102 ensures the consistency and bandwidth integrity of voltage and current signal acquisition, providing raw data support for subsequent frequency domain feature extraction.

[0080] It should also be noted that the signal acquisition module 100 is a prerequisite for the system to perform frequency domain modeling and state analysis, and the quality of its acquired signals has a direct impact on subsequent recognition results.

[0081] S5: The frequency domain preprocessing and extraction module 200 is used to perform synchronous signal preprocessing on the primary voltage signal and the end-screen current, using wavelet transform and filtering strategies to remove noise and extract frequency domain features.

[0082] It includes a signal filtering and normalization submodule 201 and a frequency domain feature extraction submodule 202 .

[0083] Furthermore, the signal filtering and normalization submodule 201 performs bandpass filtering on the collected voltage and current signals to suppress power frequency interference and low-frequency noise, and uses a normalization algorithm to unify the amplitude range. The frequency domain feature extraction submodule 202 performs multi-layer wavelet decomposition and Fourier analysis to extract frequency domain features such as harmonic energy and main frequency drift, and constructs a multidimensional feature vector as subsequent input.

[0084] It should be noted that the signal filtering and normalization submodule 201 provides a clean, standardized input signal, improving the accuracy and comparability of feature calculations. The frequency domain feature extraction submodule 202 is the core of the recognition model input construction. The extracted spectral distribution index is highly correlated with capacitance changes.

[0085] It should also be noted that the frequency domain preprocessing extraction module 200 is an intermediate link connecting the two stages of acquisition and recognition, and its output characteristics directly determine the resolution and response speed of the modeling results.

[0086] S6: The capacitance identification and judgment module 300 is used to calculate the coupling relationship between the bushing structure parameters and the capacitance based on the frequency domain characteristics, establish a capacitance change identification model, and determine abnormal points in combination with the temperature and humidity data.

[0087] It includes a capacitance estimation modeling submodule 301 and an environmental factor correction judgment submodule 302 .

[0088] Furthermore, the capacitance estimation modeling submodule 301 combines the wall bushing structural parameters with the frequency domain eigenvectors to construct a capacitance estimation function per unit length. This function then uses historical operational data to train a regression model to generate a time-varying curve of the estimated value. The environmental factor correction and judgment submodule 302 uses data collected by on-site temperature and humidity sensors to calculate the temperature and humidity sensitivity coefficient based on regression analysis, dynamically adjust the judgment threshold, and determine abnormal capacitance changes.

[0089] It should be noted that the capacitance estimation modeling submodule 301 establishes a coupling channel between electrical signal changes and physical parameters, ensuring that the recognition results are physically interpretable and adaptable to specific scenarios. The environmental factor correction judgment submodule 302 effectively eliminates the perturbations of external environmental changes on the judgment threshold, improving the stability and accuracy of the judgment.

[0090] It should also be noted that the capacitance identification and judgment module 300 is a key link in realizing intelligent early warning and abnormality location in this system, and the output results can be directly used for operation and maintenance strategy optimization and fault intervention response.

Claims

1. A method for monitoring the capacitance of a wall bushing in a flexible DC converter station, characterized in that: include: A capacitive voltage divider is installed between the end screen of the wall bushing in the flexible DC converter station and the ground, and a series capacitance measurement structure is constructed to collect the bushing primary voltage signal and the end screen current. Perform synchronous signal preprocessing on the primary voltage signal and the end-screen current, using wavelet transform and filtering strategies to remove noise and extract frequency domain features; The coupling relationship between bushing structural parameters and capacitance is calculated based on frequency domain characteristics, a capacitance change identification model is established, and abnormal points are identified by combining temperature and humidity data. The frequency domain feature calculation of the coupling relationship between the bushing structural parameters and the capacitance includes: inputting the length, inner diameter, and dielectric constant of the wall bushing as structural parameters; constructing a capacitance estimation function based on an equivalent capacitance coupling coefficient model in combination with the frequency domain features; importing the model through a table lookup; the estimation function uses the spectral energy ratio and the main frequency change rate in the frequency domain features as independent variables; and using the capacitance change trend per unit length as the output result; and performing standardization and normalization processing in combination with historical data to form a standard coupling curve for subsequent abnormal point comparison.

2. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 1, wherein: The provision of a capacitive voltage divider comprises: Two capacitors are connected in series between the end screen and the ground to form a voltage divider measurement circuit. The two capacitors are the main capacitor and the detection capacitor. The main capacitor is connected in parallel in the wall bushing body, and the detection capacitor is connected to the voltage acquisition node. The voltage change across the main capacitor is measured to obtain the potential difference change curve between the end screen of the wall bushing and the ground, thereby realizing indirect deduction of the capacitance value.

3. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 1 or 2, wherein: The construction of the series capacitance measurement structure includes: Multiple broadband current sensors and high-voltage voltage lead-out lines are arranged in a ring around the wall bushing casing. The high-voltage voltage is led out through an insulated cable and connected to the voltage sampling circuit. The end-screen current is measured through a Rogowski coil and collected synchronously with the voltage signal to set the minimum threshold of the sampling frequency.

4. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 3, wherein: The synchronous signal preprocessing of the primary voltage signal and the end-screen current includes: The voltage and current signals are digitally denoised. The voltage signal is filtered using a bandpass filter based on frequency band filtering to remove the power frequency interference component, and the current signal is transformed using wavelet transform to extract the transient spectrum component. After the denoising process is completed, the time domain alignment of the voltage and current signals is completed by timestamp marking.

5. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 4, wherein: The use of wavelet transform and filtering strategy to remove noise and extract frequency domain features includes: The Daubechies wavelet function is used for six-layer decomposition to extract high-frequency harmonic components. The spectrum distribution map is obtained through power spectral density analysis. The characteristic point parameters in the target frequency band are extracted according to the set frequency band weights as the input features of the subsequent capacitance coupling model. The energy proportion of each frequency band is used as an important dimension of the model feature vector to participate in the subsequent coupling relationship modeling.

6. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 5, wherein: The extracting of feature points within the target frequency band includes: Based on the frequency band division, 1kHz to 10MHz is divided into five sub-bands. In each sub-band, the main peak frequency of the harmonic, the total energy of the band, the maximum amplitude of the band, and the transient slope are selected as the characteristic point parameters. After the energy of each band is normalized, a 20-dimensional feature vector is formed to describe the energy distribution characteristics of the high-frequency harmonics in the entire frequency domain. Before the eigenvector is input into the capacitance coupling relationship model, a principal component analysis dimensionality reduction process is performed to reduce the modeling dimension and retain the key spectrum change information.

7. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 6, wherein: The frequency domain characteristic calculation of the coupling relationship between the bushing structure parameters and the capacitance includes: Obtain the structural parameters of the wall bushing, including the length, inner diameter, and dielectric properties of the bushing material, and establish a reference capacitance value per unit length based on the physical structure of the wall bushing; Based on the high-frequency energy ratio and main frequency changes obtained from frequency domain feature extraction, a characteristic index describing the change in electrical behavior is constructed. The characteristic index is combined with structural parameters and input into a preset estimation function. The estimation function is trained using historical operating data to obtain a set of weight parameters suitable for different types of bushings, and outputs an estimated value of the capacitance per unit length at the current moment. The estimated capacitance change trend is compared with the standard curve under healthy conditions recorded during historical operation, and all data are standardized; During the comparison process, if there is a persistent deviation between the estimated curve and the standard curve, and the deviation exceeds the set judgment threshold, the capacitance characteristics of the current bushing have undergone abnormal changes; Mark the time points and segments corresponding to the deviations.

8. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 7, wherein: The establishment of a capacitance change recognition model and the combination of temperature and humidity data to determine abnormal points include: The difference between the standardized capacitance change curve and the current measured curve is calculated, and the temperature and humidity correction coefficients are set to dynamically adjust the difference threshold. The temperature and humidity are collected in real time by the on-site environmental sensor. The correction factor is calculated through regression analysis. If the difference exceeds the correction threshold, it is considered that the capacitance change is abnormal, and the abnormal section is located. The start and end positions of the abnormal area, the capacitance change rate and the corresponding timestamp are output.

9. The method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to claim 8, wherein: The regression analysis method includes: From the bushing operation history data, multiple sets of ambient temperature, humidity and capacitance curve difference samples under the same load conditions were selected. The temperature sensitivity coefficient and humidity sensitivity coefficient were fitted with multivariate linear fitting using the least squares method to obtain the regression equation. The temperature sensitivity coefficient indicates the correction ratio of the difference amplitude caused by each 1°C change, and the humidity sensitivity coefficient indicates the error tolerance change amplitude caused by each 1% RH change.

10. A capacitance monitoring system for a wall bushing in a flexible DC converter station, characterized by: It comprises a signal acquisition module (100), a frequency domain preprocessing and extraction module (200), and a capacitance identification and judgment module (300); The signal acquisition module (100) is used to set a capacitive voltage divider between the end screen of the wall bushing of the flexible DC converter station and the ground and to construct a series capacitance measurement structure to collect the bushing primary voltage signal and the end screen current; The frequency domain preprocessing and extraction module (200) is used to perform synchronous signal preprocessing on the primary voltage signal and the end-screen current, and adopt wavelet transform and filtering strategies to remove noise and extract frequency domain features; The capacitance identification and judgment module (300) is used to calculate the coupling relationship between the casing structure parameters and the capacitance based on frequency domain characteristics, establish a capacitance change identification model, and judge abnormal points in combination with temperature and humidity data.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to any one of claims 1 to 9 are implemented.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for monitoring the capacitance of a wall bushing in a flexible DC converter station according to any one of claims 1 to 9 are implemented.

Images