Artificial pollution test method in high-pressure environment

Abstract

The invention relates to the technical field of high-voltage engineering and insulation, in particular to an artificial pollution test method in a high-voltage environment, which comprises the following steps of: firstly, constructing a target deposition distribution function and a target response set on a parameterized grid on the surface of an insulator, and combining environment boundary condition time sequence sets to form a target deposition and response data set; layered deposition of soluble pollutants and insoluble particles is carried out in a power-on state through an electrode system time-varying electric field and particle charge control, and a surface state field is collected to generate an in-situ deposition state data set. Electrical and surface responses are obtained according to segmented loading and stepped flashover tests, and a median flashover voltage is estimated, so that an electrical and physical response data set is formed; and finally, carrying out closed-loop optimization on the time-varying trajectory, the environment program and the deposition spectrum shape of the electrode system based on the dynamic model within the leakage current constraint. According to the method, configuration reproducibility, threshold estimation stability and test rapid convergence are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of high-voltage engineering and insulation technology, in particular to a method for artificial contamination test under high-voltage environment. BACKGROUND

[0002] The power grid along the coast, sand and industrial belt section widely faces the pollution risk of superimposed soluble salts and insoluble particles. Contamination changes the surface conduction channel and creep path in the wetting and drying cycle, induces the increase of leakage current and flashover trip, and affects the power supply reliability and equipment life. The laboratory artificial contamination test is the basis for operation and maintenance grading, insulator type selection and anti-pollution strategy verification, but there are long-standing problems such as unrepeatable configuration, inconsistent statistical caliber and difficult quantification of safety boundary, which restrict the extrapolation of test results to engineering application. SUMMARY

[0003] In view of the many problems existing in the prior art, the present application provides a method for artificial contamination test under high-voltage environment. The present application defines target deposition and response based on parameterized grid, then writes configuration in layered form with time-varying electric field and charged particles in the powered state, subsequently obtains median flashover voltage through segmented loading and statistical estimation, and finally optimizes electrode trajectory, environmental program and deposition spectrum shape in the leakage current constraint, with the effect of repeatable target configuration, stable threshold estimation and fast convergence of test.

[0004] A method for artificial contamination test under high-voltage environment, comprising the following steps: Establishing a parameterized grid on the surface of an insulator, generating a target deposition distribution function and a target response set, merging an environmental boundary condition time sequence set to obtain a target deposition and response dataset; Forming a time-varying electric field distribution through an electrode system in the powered state, supplying soluble contaminant aerosol and insoluble particle aerosol to the surface of the insulator and implementing particle charging control, constructing a layered contaminant deposition structure according to the environmental boundary condition time sequence set and the supply switching, collecting surface state field data, and generating an in-situ deposition state dataset; Performing loading according to the environmental boundary condition time sequence set and voltage working condition changes, collecting electrical response time sequence and surface state field time sequence, obtaining median flashover voltage by flashover test and statistical method, calculating related increments, and generating an electro-matter response dataset; Within the leakage current constraint, based on the electro-matter response dataset and the target deposition and response dataset, taking the deviation index of deposition configuration and conductive connectivity, electrical response and median flashover voltage as the optimization target, establishing a dynamic model of surface state and migration behavior, updating the time-varying trajectory of the electrode system and the environmental boundary condition time sequence set, adjusting the contaminant spectrum parameter set, and outputting the converged parameter set.

[0005] Preferably, the environmental boundary condition time series set consists of relative humidity time series, temperature time series, wind speed time series and pollutant deposition flux time series, and is sampled and aligned with a uniform time reference to serve as input of the target deposition and response dataset.

[0006] Preferably, the electrode system is a multi-electrode shaped array, with electrode channel phase and amplitude set independently, and electrode channels arranged with fixed geometric spacing, and time-varying electric field distribution formed along the insulator axis and radial direction by sequential driving.

[0007] Preferably, the particle charging control adopts corona charging, with single positive or single negative polarity, and particle charge amount determined by Faraday cylinder combined with electrostatic gauge, and used to correct aerosol supply parameters.

[0008] Preferably, the construction sequence of the layered pollutant deposition structure is to first supply insoluble particle aerosol to form a particle skeleton, then increase relative humidity and supply soluble pollutant aerosol to form a coating layer in the wetting stage, and then reduce relative humidity to stabilize the layered structure in the drying stage.

[0009] Preferably, the surface state field data includes water film thickness field, surface equivalent conductance field and soluble pollutant surface concentration field, wherein the water film thickness field is obtained by non-contact optical measurement conversion, the surface equivalent conductance field is obtained by segmented measurement of multi-point electrode array and inversion, and the soluble pollutant surface concentration field is obtained by one of micro-elution conductance determination or spectral determination.

[0010] Preferably, the loading is performed in a segmented manner, and the loading sequence contains at least three types of segments among relative humidity step, temperature step, wind speed change, pollutant deposition flux pulse and voltage transient, and the start and end time of each segment is recorded with time markers and synchronized with the time stamp of the acquisition channel.

[0011] Preferably, the median flashover voltage is obtained by ascending and descending ladder flashover test, with consistent voltage step-up and down, and statistical regression method is used to estimate the results of multiple tests.

[0012] Preferably, the deviation index of deposition configuration and conductive connectivity is obtained by thresholding the surface equivalent conductance field to obtain the conduction map, and the deposition configuration difference is measured by optimal transport distance, and the conductive connectivity difference is measured by persistent homology bottleneck distance.

[0013] Preferably, the kinetic model of surface state and migration behavior is established by sparse structure identification, the candidate basis function set and state derivative are obtained by sparse regression to obtain the analytical form, the time-varying trajectory of the electrode system is updated by solving the gradient with the adjoint method, and the pollutant spectrum parameter set is adjusted by introducing the feasible region constraint of the upper limit of leakage current and the upper limit of leakage current change rate and using Bayesian optimization.

[0014] Compared with the prior art, the advantages and beneficial effects of the present application are that: By establishing the insulator surface parameterized grid and the target deposition distribution function and merging the environmental boundary condition time sequence set, integrated space-time alignment and configuration repeatability across devices and batches are realized.

[0015] By parallel driving of the electrode system time-varying electric field and particle charging control, layered deposition of soluble pollutants and insoluble particles is implemented in the powered state, realizing directional writing of conductive communication paths and programmable construction of spatial non-uniformity.

[0016] By segmented loading and stepwise flashover test and cooperating with statistical regression method, repeated estimation of median flashover voltage and causal attribution to environmental segments are realized.

[0017] By constructing a dynamic model of surface state and migration behavior within the leakage current constraint and performing closed-loop optimization, coordinated updating and rapid convergence of the electrode system time-varying trajectory, environmental boundary condition time sequence set and pollutant spectrum parameter set are realized.

[0018] By forming standardized fields of in-situ deposition state data sets and electrical and physical response data sets, an engineered data foundation is realized that is traceable, replayable and useful for model identification throughout the process. BRIEF DESCRIPTION OF DRAWINGS

[0019] Fig. 1 is a flowchart of the method of the present application; Fig. 2 is a schematic diagram of the data set relationship and closed-loop optimization in the present application. DETAILED DESCRIPTION

[0020] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure.

[0021] As Figs. 1-2 shown, a method for artificial contamination test under high voltage environment includes the following steps: Establishing an insulator surface parameterized grid, generating a target deposition distribution function and a target response set, merging an environmental boundary condition time sequence set, and obtaining a target deposition and response data set; This invention addresses the input-side unification and alignment issues in artificial pollution tests under high-voltage environments. It establishes a parameterized mesh on the insulator surface, generates a target deposition distribution function and a target response set, and merges the environmental boundary condition time series set to obtain the target deposition and response dataset. The complex three-dimensional umbrella skirt surface is "flattened" into a two-dimensional coordinate system, allowing any point to be mapped using coordinates along the creepage path. Coordinates along the direction of the busbar This unique representation ensures that field quantities such as water film thickness, salt density, and conductivity can be aligned and compared on a unified index. The method for constructing a parameterized mesh on the insulator surface includes: reading the insulator's CAD or laser scanning model, and then... (The sentence is incomplete and requires more context to translate accurately.) Sampling is performed along the radial generatrix in angular increments. Sampling, generation A regular mesh is used to record additional labels such as local curvature and skirt numbers, which can be used for visual mask and electric field sensitivity calculations. Its core lies in using the surface coordinates of the insulator surface as a unified benchmark, and using spatial non-uniformity templates and response indicators as constraints, to provide a data structure and index system that can be directly called upon for subsequent in-situ deposition, loading, and closed-loop optimization.

[0022] In principle, the insulator surface is an irregular curved surface of revolution. If the deposition target is directly described on three-dimensional coordinates, subsequent deposition and measurement are difficult to align. This invention employs a dual-parameterization method along the creepage path direction and along the umbrella skirt generatrix direction to unfold the three-dimensional surface into a regular grid. Each grid cell is associated with a location index, local curvature label, and visualization mask, ensuring that any target field can be aligned with the in-situ measured water film thickness field and surface equivalent conductivity field under the same index. The spatial non-uniformity template is used to express the intensity zoning and connectivity expectation of the target deposition. It can be assigned values ​​according to structural regions such as the inner edge, outer edge, and ribs of the umbrella skirt, and allows for strip-like or patchy distributions along the creepage path.

[0023] The target deposition distribution function is defined as the surface density of contaminants per unit area and its component ratio, including the surface density of soluble contaminants, the surface density of insoluble particles, and their volume fraction labels. It also records the stratification relationship of the target and the surface energy constraint intention, facilitating subsequent shaping using a framework-first, then coating approach. Given the desired soluble salt surface density and insoluble particle surface density for each grid cell, it can be considered as a contaminant quality blueprint to be "written"; simultaneously, the relative weights of sections such as the outer edge zone and the back of the umbrella skirt are defined. Methods for constructing the target deposition distribution function include: statistically analyzing the distribution of salt spray and sand / dust contamination on actual lines or mapping these statistics to anti-fouling standards. On the grid, a desired surface density matrix is formed; a layering field can be added to indicate the process sequence of "skeleton first and then covering". The target response set is used to describe the desired electrical and surface state responses, including the median flashover voltage target, the characteristic label of the leakage current time series, and the spatial distribution target of the surface equivalent conductance. The environmental boundary condition time series set is used to describe the controllable and influential external driving of the deposition distribution and surface state during the test process, including the relative humidity time series, the temperature time series, the wind speed time series, and the pollutant deposition flux time series, and are aligned with a unified time base. The target deposition and response data set after merging fixes the field name, index, unit, and measurement method under a version number, ensuring that subsequent steps can seamlessly reference.

[0024] The purpose of merging is to fix the "desired deposition pattern" and the "external sequence that drives it to form" into the same data set, which is convenient for the subsequent controller to call by time segment. The environmental boundary conditions (humidity, temperature, wind speed, and pollutant flux sequence) can be adjusted in real time in the physical high-pressure test chamber; in numerical simulation, they are applied as boundary inputs to the CFD-electric field coupling model. Therefore, this step serves both physical high-pressure chamber tests and simulations: offline preparation → get a unified data set; online execution → the physical test replays the environmental sequence according to the data set, and the simulation loads the same sequence as the boundary.

[0025] In an embodiment, first, the geometric model and operating class of the test insulator are obtained, the grid row index is set according to the creepage path length, the grid column index is set according to the direction of the shed bus, and the grid spacing is selected according to the target resolution. A spatial non-uniformity template is generated for each grid cell, which is marked with a high-weight band in the neighborhood of the shed outer edge, a medium-weight band in the rib transition zone, and a low-weight band near the fitting. Then, the target deposition distribution function is given, specifically, a higher soluble pollutant target surface density and insoluble particle target surface density are assigned to the high-weight band, and the sequence and interface requirements are recorded in the layering field. Next, the target response set is constructed, and the median flashover voltage target value interval, the target characteristic label of the leakage current time series, and the target spatial distribution map of the surface equivalent conductance are given. Then, the environmental boundary condition time series set is prepared, including the relative humidity time series, the temperature time series, the wind speed time series, and the pollutant deposition flux time series, and the time stamp and grid index are aligned to obtain the target deposition and response data set. The data set is saved with fixed field names, including the target deposition distribution function, the spatial non-uniformity template, the target response set, and the environmental boundary condition time series set, and is accompanied by units and allowed value ranges, ensuring cross-device and cross-batch reuse.

[0026] The dataset fixes the geometry, target and working condition information in the same grid and time benchmark, significantly reducing the alignment error and artificial interpretation difference in the subsequent steps. The spatial non-uniformity template determines the expected position and intensity of non-uniform deposition before the test, avoiding the blindness of in-situ deposition stage. The target response set describes the electrical indicators and surface state indicators together, providing clear criteria for subsequent loading and optimization. The unified sampling and alignment of the environment boundary condition time sequence set makes the sequence working condition directly map to the grid index, facilitating the triggering of fragmented loading and synchronous acquisition.

[0027] For example, for a composite insulator with a line voltage level of 220, the creepage distance is 5600 mm, the number of grid rows along the creepage path direction can be set to 560, and the number of grid columns along the umbrella skirt bus direction is 120. The spatial non-uniformity template is set to high value band in the outer three rows, medium value band in the rib part four rows, and low value band in the ten rows adjacent to the fitting. The target deposition distribution function sets a higher target surface density of soluble contaminants and insoluble particles in the high value band, and requires the formation of an insoluble particle skeleton before forming a soluble contaminant coating layer in the stratification field. The target response set sets the target value interval of the medium value flashover voltage, and gives the fragment energy proportion label of the leakage current time sequence and the target spatial distribution graph of the surface equivalent conductance. The environment boundary condition time sequence set gives the relative humidity time sequence, temperature time sequence, wind speed time sequence and pollutant deposition flux time sequence with a sampling interval of 1 second. The target deposition and response dataset obtained by the above processing can be directly used as the unified input of in-situ deposition control, loading sequence design and closed-loop optimization, improving the reproducibility and statistical comparability of the test.

[0028] Preferably, the environment boundary condition time sequence set is composed of relative humidity time sequence, temperature time sequence, wind speed time sequence and pollutant deposition flux time sequence, and is sampled and aligned with a unified time benchmark as the input of the target deposition and response dataset.

[0029] The present application defines the environment boundary condition time sequence set as a group of relative humidity time sequence, temperature time sequence, wind speed time sequence and pollutant deposition flux time sequence, and samples and aligns with a unified time benchmark as the input of the target deposition and response dataset. The reason for using group time sequence instead of single point value is that the moisture absorption, dissolution, recrystallization and film connection of pollution show strong time dependence and path dependence, and the leakage current and flashover threshold in the high voltage loading stage are also sensitive to the wet-dry cycle, the lifting rate and the flux pulse height. Unifying the four types of external driving to the same time axis can provide a unique environmental index for each response record in the subsequent deposition control, loading segment design and parameter identification, avoiding alignment error and interpretation ambiguity.

[0030] In an implementation, a common clock is used as the time reference for the entire chain. Each sensor and execution node timestamps and caches data under the same time reference, and the data is aligned by the data aggregation node. Relative humidity and temperature sensors are placed at a representative height close to the surface of the insulator. Wind speed sensors are placed in the incoming flow direction and near the middle of the insulator. The deposition flux of pollutants is obtained by mass accumulation sensors or weighing of reference sampling plates. The flux is defined as the mass increment per unit area per unit time on the reference surface. To reduce the effects of the boundary layer and installation disturbance, thin rods and remote leads are used for sensor installation. Zero point and range calibration are performed before and after measurement. When the sampling interval and control step are consistent, the data is directly aligned. When they are inconsistent, the data is resampled and interpolated based on the common time axis. The interpolated section is marked with a mass identifier for subsequent analysis and screening. For sudden changes, a median strategy with forward and backward windows is used for consistency check. For missing data, linear or spline interpolation is used with clear labeling of the interpolation flag and the original missing length.

[0031] In terms of data structure, the time series set of environmental boundary conditions is fixed with uniform field names, uniform units, and uniform coordinate numbers. The time key is unique and monotonically increasing. After establishing the spatial index of the insulator surface parameterized grid, the mapping relationship from time to space is established, and the environmental driving at any time can generate the corresponding deposition and wetting boundary in the grid coordinates. The target deposition and response data set keeps the time key unchanged and uses relative humidity, temperature, wind speed, and pollutant deposition flux as exogenous variables in the subsequent control and identification process when loading the set, so that the time schedule of deposition control, the segment boundary of loading sequence, and the regression window of parameter identification can strictly correspond on the same time axis.

[0032] In Example 1, an environmental program including relative humidity, temperature, wind speed, and pollutant deposition flux is constructed. The relative humidity is increased to a high humidity platform at a fixed step per minute in the early stage, kept constant in the middle stage, and decreased to a dry level at a fixed step in the later stage. The temperature is maintained at a constant value, the wind speed is switched between two levels in a pulse form, and the pollutant deposition flux is set once before and after the wet impact. Each channel is sampled and aligned at the same second-level time reference, the missing section is not more than ten seconds and is filled by linear method, and the filling mark is recorded. After the set is used as the input of the target deposition and response data set, the deposition control triggers the coating stage of soluble pollutants at the high humidity platform, the loading sequence arranges the voltage transient segment after the flux pulse, the parameter identification extracts the change rate of water film and conductance using the window corresponding to the platform segment, and finally the model fitting and optimization are completed on the same time key.

[0033] The application does not repeat the mature measurement means, but emphasizes the consistency of time key and the stability of field name in application, emphasizes that the traceability record in the multi-channel case should be stored together with the data, including sampling interval, calibration time, missing measurement mark and interpolation method. Through the above principles and implementation modes, the environmental boundary condition time sequence set not only serves as an input, but also serves as an alignment skeleton for subsequent control, loading and identification, supporting the reproduction experiment and closed-loop convergence of the final deposition form and electrical response.

[0034] In the power-on state, a time-varying electric field distribution is formed through the electrode system, soluble contaminant aerosol and insoluble particle aerosol are supplied to the surface of the insulator, and particle charging control is implemented, a layered contaminant deposition structure is constructed according to the environmental boundary condition time sequence set and supply switching, surface state field data is collected, and an in-situ deposition state data set is generated; In the power-on state, a time-varying electric field distribution is formed through the electrode system, soluble contaminant aerosol and insoluble particle aerosol are supplied to the surface of the insulator, and particle charging control is implemented, a layered contaminant deposition structure is constructed according to the environmental boundary condition time sequence set and supply switching, surface state field data is collected, and an in-situ deposition state data set is generated;

[0035] The electrode system is composed of multiple channels, and the phase and amplitude of each channel are independently set. After being arranged around the test insulator, the electrode system is driven in sequence to form a high-field distribution moving along the creepage path direction and the umbrella skirt bus direction on the surface. The time-varying electric field serves two purposes. First, the normal electric field component is used to shape the migration and capture trajectory of charged particles near the wall surface. Second, the tangential electric field component is used to adjust the transient connection path of the thin water film, so that the deposition position is consistent with the subsequent conductive connection expectation. The particle charging control selects a single-pole charging method, and the charge parameter takes the mass-to-charge ratio as the core index, which is kept stable through online metering and closed-loop correction. The soluble contaminant aerosol serves as a precursor of soluble salt, which determines the surface equivalent conductance and the redissolution behavior in dry-wet cycles. The insoluble particle aerosol serves as a rough skeleton, providing nucleation sites and persistent topography.

[0036] The environmental boundary condition time series set includes a relative humidity time series, a temperature time series, a wind speed time series and a pollutant deposition flux time series. The relative humidity and the temperature determine the formation and evaporation of the thin water film, the wind speed determines the near-wall shear and the residence probability of incident particles, and the pollutant deposition flux gives the mass input per unit area per unit time. Through coupling with the supply switching, the insoluble particles are first deposited to form a skeleton, and then the soluble pollutants are supplied to complete the coating in the high relative humidity stage to realize the stratification rhythm. To avoid ambiguity in expression, the present application clearly defines the following terms: the water film thickness field is the spatial distribution of the thin water film thickness of each grid element on the surface; the surface equivalent conductance field is the spatial distribution of the unit length equivalent conductance of each grid element on the surface; the soluble pollutant surface concentration field is the spatial distribution of the surface density of the soluble salt on each grid element on the surface; and the insoluble particle skeleton density field is the spatial distribution of the surface density of the insoluble particles on each grid element on the surface. The above four fields together with the test metadata constitute the in-situ deposition state data set.

[0037] To highlight the core mechanism, the present application only gives three minimum evolution relationships in this step, which are used to guide the alignment and calibration of the control quantity and the observation quantity: , , wherein, represents the water film thickness, represents the surface density of the soluble pollutant, represents the surface density of the insoluble particle, represents the relative humidity time series, represents the surface normal electric field intensity, represents the near-wall wind speed, represents the soluble pollutant deposition flux time series, represents the insoluble particle deposition flux time series, is a non-negative coefficient related to the material and the geometry, is a dimensionless function of the thin film replenishment term under the joint influence of the electric field and the wind speed, is a dimensionless function of the dissolution and migration coupling term under the influence of the electric field. The specific values of the coefficients and the dimensionless functions are calibrated in the subsequent identification step. The present application does not expand the derivation here, but only for the corresponding relationship between the control quantity and the state quantity, to ensure the consistency of the data and the process.

[0038] In the aspect of collection strategy, the water film thickness field is measured by the oblique incidence reflection method or the structured light method, the surface equivalent conductance field is measured by the segmented electrode array, the soluble contaminant surface concentration field is obtained by the micro elution and conductance measurement or spectral inversion, and the insoluble particle skeleton density field is obtained by the surface scattering characteristics and the reference sampling.

[0039] In embodiment 2, the environmental program is set as a wet platform and drying, the insoluble particle skeleton is formed on the high relative humidity platform in the power-on state, and then the soluble contaminant flux is switched to complete coating. The channel phase of the electrode system is gradually scanned, so that the high field region trajectory moves along the outer edge of the umbrella skirt. At this time, the water film thickness field increases with the increase of the relative humidity, the surface equivalent conductance field forms a continuous band on the outer edge, and the soluble contaminant surface concentration field and the insoluble particle skeleton density field overlap in the same band. The in-situ deposition state data set is used to load the design after being checked, so that the subsequent electric-material response connection path is consistent with the deposition path.

[0040] Preferably, the electrode system is a multi-electrode shaped array, the electrode channel phase and the electrode channel amplitude are independently set, the electrode channels are arranged at a fixed geometric interval, and a time-varying electric field distribution moving along the axial and radial directions of the insulator is formed by sequence driving.

[0041] The present application adopts a multi-electrode shaped array as an electrode system, the channel phase and the channel amplitude are independently set, the channels are arranged at a fixed geometric interval around the insulator, and a time-varying electric field distribution continuously moving in the axial and radial directions of the insulator is formed by sequence driving. The normal component of the electric field distribution is used to control the capture of charged particles in the near-wall region, and the tangential component is used to guide the transient migration and connection of the thin water film, so as to realize the controllable formation of the layered contaminant deposition structure in the power-on state, and generate the in-situ deposition state data set with the spatially resolved surface state field data.

[0042] The effect of the multi-electrode shaped array can be regarded as the spatio-temporal shaping of the electric potential of the test space. The influence function of each electrode channel is obtained in advance in an offline solver, and then is superimposed in real time according to the channel phase and the channel amplitude in the test process. In order to clearly control the relationship between the control quantity and the state quantity, the present application gives the minimum field and flux coupling expression in this core link: , , , , , , is the electric field intensity vector, is the spatial position, is the time, the number of electrode channels, the nth time-varying driving potential of the channel, the nth channel influence function, the surface normal unit vector, the surface tangent unit vector, the normal and tangent field components, the normal and tangent field components, the insoluble particle deposition flux, the soluble contaminant deposition flux, the normal and tangent field components, the collection efficiency coefficient, the particle charge-to-mass ratio, the normal and tangent field components, the time series of volume fractions of insoluble particles and soluble contaminants in air, the surface equivalent conductance, the dry surface baseline conductance term, the coupling coefficient, the water film thickness, the surface soluble contaminant area density. The above expressions are used to guide the sequence driving strategy and supply switching strategy of the electrode system, and to check the physical consistency of the state field in the data acquisition stage.

[0043] In an implementation, the multi-electrode shaped array adopts a ring or spiral arrangement, and the ratio of the channel spacing to the outer diameter of the insulator is kept fixed to ensure the repeatability of the field shape. The channel phase adopts traveling wave coding to generate high-field region scanning along the axial direction, and the channel amplitude adopts envelope modulation to focus the high-field region in the radial direction. The time step of the sequence driving is consistent with the set of environmental boundary condition time series, and the driving state of the electrode system is written with the same time reference to the in-situ deposition state data set. The particle charging control adopts single polarity charging and is online checked by Faraday cylinder and electrometer, and the checking results are used to correct the supply mass flow and charge-to-mass ratio to ensure the coordination of flux and electric field. The acquisition link samples the surface state field at a high frequency at the moment of switching of the electrode system, the water film thickness field adopts oblique incidence reflection or structured light inversion, the surface equivalent conductance field adopts segmented electrode array scanning, the soluble contaminant surface concentration field adopts micro-elution conductance method or spectroscopy, and the insoluble particle skeleton density field is obtained by surface scattering inversion and benchmark sampling verification. All fields and electrode system, supply state are written with the same time key, thereby forming a replayable in-situ deposition state data set.

[0044] Example 3, set the number of channels to 24, using the uniform phase advance traveling wave encoding, so that the high field area along the creeping path continuously moving. Supply strategy for the first insoluble particles, then soluble contaminants, environmental boundary conditions time series set to high relative humidity platform plus dry. This way in the outer edge of the formation of a continuous skeleton band and in the same path to form a uniform coating, the surface equivalent conductance field is connected in the form of a band, the water film thickness field and the soluble contaminant surface concentration field in the in-situ deposition data set on the outer edge of the band overlap, the subsequent loading of the leakage current main channel and deposition path consistent, to verify the controllability of the electrode system on the connected path.

[0045] Preferably, the particle charging control uses corona charging, the charging polarity is unipolar positive or unipolar negative, the particle charge is determined by Faraday cylinder and electrometer, and the determined value is used to correct the aerosol supply parameters.

[0046] In the artificial pollution test under high voltage environment, the particle charging control is used as a pre-adjustable link for deposition distribution and connectivity form. The corona charging device is used to charge the soluble contaminant aerosol and insoluble particle aerosol with unipolar positive or unipolar negative, so that the particles obtain a stable charge-to-mass ratio, and then enter the deposition area containing time-varying electric field to migrate directionally and capture near the wall along the electric field force. The online calibration of the charge-to-mass ratio is completed by the Faraday cylinder and the electrometer, and the determined value is used to close-loop correct the aerosol supply mass flow and the corona voltage, so as to realize the target deposition flux under the given normal electric field and air volume fraction conditions.

[0047] To establish a calculable mapping relationship between the control quantity and the observable, the present application gives two types of minimum necessary expressions in this core link. One is the determination and correction of the charge-to-mass ratio: , wherein represents the charge-to-mass ratio, with units of coulombs per kilogram; represents the current measured by the Faraday cylinder, with units of amperes; represents the aerosol mass flow entering the Faraday cylinder, with units of kilograms per second; represents the collection efficiency of the Faraday cylinder, which is a dimensionless coefficient, and its value is obtained by geometric and flow state calibration. The real-time calculation of is used as a feedback quantity into the joint controller of supply and charging, so that the target charge-to-mass ratio set can remain stable in the presence of disturbances. The second is the deposition flux estimation under the driving of near-wall electric migration: , wherein represents the deposition flux directed to the surface, with units of kilograms per square meter per second; represents the collection efficiency, which is a dimensionless coefficient; represents the surface normal electric field strength, with units of volts per meter; φi represents the volume fraction of the corresponding component in air, dimensionless. The target mass-to-charge ratio for the controller inverse solution is obtained as follows: wherein J represents the target deposition flux, in kg / m2 / s; μ represents the target mass-to-charge ratio.

[0048] In the embodiment, the corona charging device is arranged before the supply pipeline enters the climate chamber, and the electrode adopts a needle plate or needle ring geometry to avoid self-excited discharge across the insulator shell. The charging polarity is selected before the start of the test and remains unchanged throughout the test, so as to form a predictable electric migration direction after superimposing the time-varying electric field component. The supply pipeline adopts a conductive static hose, and there are not less than 2 grounding drainage points, and the pipeline length is not more than 2 meters, so as to reduce the leakage charge and return pulsation. The Faraday cylinder is arranged in the sampling branch, and the sampling flow is stably maintained by the mass flow controller. The sampling period of the sampling branch is not more than 10 minutes each time, and the zero point and range of the electrostatic meter are checked during the calibration process, and are written into the traceability log. The mass-to-charge ratio controller adopts a double-channel adjustment channel, one channel adjusts the corona voltage, and the other channel adjusts the mass flow of the spray. The two channels are sampled and output with the same time reference, and the control law adopts a proportional integral type correction: wherein J represents the target deposition flux, in kg / m2 / s; and Kp and Ki represent the proportional and integral gains Ts represents the sampling period. In order to avoid outliers caused by multiple charging of particles, the present application sets an electrostatic rectifier section after the charging chamber outlet, which reduces the charge fluctuation through weak field homogenization and straight section aerodynamic rectification, and uses the sliding median of the Faraday cylinder current as the feedback quantity.

[0049] This link works with the electric field shaping and the time sequence set of the environmental boundary conditions of the present application. When the normal electric field strength increases, the controller automatically reduces the mass-to-charge ratio or the mass flow of the spray to maintain the target deposition flux unchanged. When the volume fraction in the air rises, the controller reduces the corona voltage to avoid excessive charging leading to space charge shielding and nucleation anomalies. Since the Faraday cylinder and the electrostatic meter are in a unified time reference, their measured values are written into the control channel field in the in-situ deposition state data set together with the supply instructions for subsequent loading segments and statistical recognition.

[0050] In Example 4, single positive charging is used to form an outer connected skeleton, and the target deposition flux is set to kg / m2 / s, the surface normal electric field strength is V / m, the air volume fraction is , and the collection efficiency is The target mass-to-charge ratio is obtained from the inverse solution as follows: Milli-Coulombs per kilogram, the controller updates the corona voltage and mass flow at 1 second intervals, and the Faraday cup current stabilizes at Ampere, corresponding to a mass flow of kilograms per second, the in-situ deposition state dataset shows that the surface equivalent conductance field of the outer band continuously grows in accordance with the target non-uniformity template.

[0051] Preferably, the construction sequence of the layered pollutant deposition structure is to first supply the insoluble particle aerosol forming particle skeleton, then to increase the relative humidity in the wetting stage and supply the soluble pollutant aerosol forming coating layer, and then to reduce the relative humidity in the drying stage to stabilize the layered structure.

[0052] The present application writes the insoluble particles and the soluble pollutants on the surface of the insulator under the on-state in the order of the layered deposition. First, the insoluble particle skeleton is constructed to provide roughness and nucleation sites, then the relative humidity is increased in the wetting stage to supply the soluble pollutants, which are dissolved, redistributed and covered on the surface of the skeleton along the tangential electric field under the guidance of the thin water film, and finally the relative humidity is reduced in the drying stage to complete the recrystallization and pore consolidation, so as to obtain a layered pollutant deposition structure with controllable connectivity and stable morphology. This sequence synchronizes the three processes of particle near-wall capture induced by electric field, salt migration mediated by water film, and drying and solidification under the same time reference, ensuring that the subsequent loading and identification can reuse the same spatial and temporal index.

[0053] The core of the layered construction can be described by three minimum expressions. First, the effective occupancy of the skeleton increases with the surface area density of the insoluble particles, which can be approximately expressed as: wherein represents the effective occupancy, which is a dimensionless quantity, represents the surface area density of the insoluble particles, represents the convergence coefficient related to the particle morphology and surface energy. The occupancy and the spatial non-uniformity template together determine the initial region of the skeleton connectivity. Second, the time history of the coating thickness is jointly dominated by deposition and dissolution and redistribution in the wetting stage, which can be written as: wherein represents the local equivalent thickness of the soluble pollutants, represents the deposition conversion coefficient determined by geometry and porosity represents the time sequence of the soluble pollutant deposition flux, represents the volume fraction loss coefficient caused by drying, represents the time sequence of the relative humidity. Third, the surface equivalent conductance is related to the water film and the salt load and the connectivity threshold, which can be written as: wherein represents the surface equivalent conductance, represents the dry surface baseline conductance, a coupling coefficient representing the water film and salt load, a water film thickness, a surface soluble contaminant areal density, a step function taking value 1 when the quantity in the bracket is positive and 0 otherwise, a critical occupation threshold representing connectivity. The three equations link the skeleton connectivity, the coating growth and the conductance response together in the same spatial grid and time key, facilitating the mutual check between the measured and controlled quantities during the deposition execution.

[0054] The implementation points are three-fold. First, the skeleton stage only supplies insoluble particles and keeps the normal field component dominant, while the tangential component is used to move the high field region along the creep path, so that the skeleton is connected in a band or patch shape according to the spatial non-uniformity template. Second, the wetting stage raises the relative humidity to a range where a stable thin water film can be formed without causing runoff, while switching to soluble contaminant supply and increasing the proportion of the tangential component, so that the salt forms a continuous coating near the surface of the skeleton rather than cavity filling. Third, the drying stage reduces the relative humidity and the tangential component to suppress water film migration, allowing the coating layer that has been formed to recrystallize in the pores and achieve mechanical engagement with the skeleton, thereby improving shape retention without changing the topology of the skeleton.

[0055] In data acquisition, the three stages record the water film thickness field, the surface equivalent conductance field, the soluble contaminant surface concentration field and the insoluble particle skeleton density field at the same time reference, and simultaneously record the phase and amplitude trajectories of the electrode system and the supply switching time stamp. The focus of the skeleton stage is the low-frequency baseline of the insoluble particle skeleton density field and the surface equivalent conductance field, the focus of the wetting stage is the coupling increment of the water film thickness field and the soluble contaminant surface concentration field, and the focus of the drying stage is the spatial stability of the surface equivalent conductance field after the coating layer crystallizes.

[0056] Example 5 takes the sea salt scenario. The skeleton stage supplies insoluble particles in the outer high field band, and the water film thickness field remains at a low level, forming a band-shaped connection. The wetting stage maintains the relative humidity at a platform and switches to soluble contaminant supply, and the tangential component guides the coating to close along the outer edge, and the surface equivalent conductance field is raised on the same band. After reducing the relative humidity in the drying stage, the surface equivalent conductance field remains stable in a band shape, and the in-situ deposition state data set shows that the coating layer and the skeleton overlap well, meeting the target deposition distribution function and the spatial non-uniformity template.

[0057] Preferably, the surface state field data includes the water film thickness field, the surface equivalent conductance field and the soluble contaminant surface concentration field, wherein the water film thickness field is obtained by non-contact optical measurement conversion, the surface equivalent conductance field is obtained by segmented measurement of a multi-point electrode array and inversion, and the soluble contaminant surface concentration field is obtained by one of micro-elution conductometric determination or spectral determination.

[0058] The present application analyzes the state of the insulator surface in the conditions of power-on and segmented loading, and the surface state field data is composed of water film thickness field, surface equivalent conductance field and soluble contaminant surface concentration field, and is written into the in-situ deposition state data set and the electrical and physical response data set after alignment with the parameterized grid and the unified time reference. The three types of field quantities are obtained by non-contact or weak disturbance means, avoiding disturbance to the high-voltage boundary layer, thin water film and conductive communication path.

[0059] The water film thickness field is obtained based on non-contact optical measurement conversion. In the application of the present application, the oblique incidence reflection interference or structured light phase method is selected, and the local film thickness is converted by calibrating the reflection phase difference or the stripe phase at the same grid position. The core relationship can be written as , wherein represents the water film thickness, represents the central wavelength of the imaging light, represents the water film refractive index, represents the incident angle, represents the phase difference between the reference and the measured reflection signal. The above formula is only used to map the phase measurement to the minimum necessary relationship of thickness, and the specific phase acquisition and unwrapping are completed by the imaging system. In order to adapt to the curved surface of the insulator and the shielding of the shed, the present application pre-generates a visible mask on the parameterized grid and performs multi-view stitching; for the wetting and drying process changing with time, the unified time key is sampled and the stripe phase is unwrapped, so that the thickness field is continuous at the segment boundary. Effectively, the water film thickness field provides the spatial distribution of the measured wetting intensity, providing quantitative basis for explaining the changes of the leakage current segment energy and the median flashover voltage.

[0060] The surface equivalent conductance field is measured by a multi-point electrode array and obtained by inversion. The electrode array is arranged along the creeping path, and only a small number of electrodes are excited at the same time in each segment, reducing electromagnetic coupling and body heating. The measurement modeling adopts a linearized transfer relationship, which combines the measured port current and port voltage into an observation vector, and establishes a matrix mapping between the equivalent conductance vector discretized on the parameterized grid, and the inversion adopts a least square with regularization: , wherein represents the estimated equivalent conductance grid vector, represents the current voltage observation vector after splicing, represents the sensitivity matrix determined by geometry and electrode arrangement, represents the discrete smoothing operator, represents the regularization weight. To avoid the counter disturbance of the power-on measurement to the field shape, the present invention interleaves the electrode array driving with the main voltage program in time, and aligns the start and end of each scan with the ambient boundary condition time series set by the segment time marker. Effectively, the surface equivalent conductance field after inversion presents significant changes at the conduction path formation and breakdown location, consistent with the connectedness events after thresholding the conduction map, which can be directly used for the connectedness difference indicator and the target calculation of the closed-loop optimization.

[0061] The surface concentration field of the soluble contaminant is obtained by one of micro-elution conductance measurement or spectroscopy. Micro-elution uses a fixed-area micro-cavity or a hydrophilic membrane for short-time contact on the grid cell, to obtain the eluent conductance, which is then converted to the surface area density by the calibration curve, denoted as wherein represents the surface area density of the soluble contaminant, represents the eluent conductance; the spectroscopy converts the area density by the ratio or differential absorption at the known salt absorption peak, solving the quantitative problem of multi-component overlap. Both means are archived with one-to-one mapping of “sampling window - in-situ position - parameterized grid cell”; the time interval and sampling volume of repeated sampling at the same position are solidified in the metadata, ensuring cross-batch comparability. Effectively, the concentration field reveals the salt load difference under the same water film thickness field, which is a key complementary quantity to explain the anisotropy of the equivalent conductance field and the difference in electrical response.

[0062] The synchronization and alignment of the three types of field quantities are the key operations of the present invention. All measurements are sampled with a unified time reference and synchronized with the segment time marker; all spatial data are recorded on the parameterized grid with the same coordinates and accompanied by a visual mask and a quality label, and the missing points are handled by interpolation or resampling strategy and written into the quality flag. Thus, the water film thickness field, the surface equivalent conductance field and the surface concentration field of the soluble contaminant can be constructed in the same segment window to calculate the connectedness difference indicator and the coupled regression of the electrical and surface state.

[0063] In Example 6, the three types of field quantities are obtained in the wet platform segment. The water film thickness field shows that the thickness of the outer edge region increases, and the surface concentration field of the soluble contaminant also increases in the same region. The surface equivalent conductance field after inversion forms a continuous band of high values, which significantly corresponds to the energy of the leakage current segment of the subsequent voltage step segment.

[0064] According to the ambient boundary condition time series set and the voltage working condition change, the loading is performed, the electrical response time series and the surface state field time series are collected, the median flashover voltage is obtained by the flashover test and statistical method, the related increments are calculated, and the electrical-matter response data set is generated; The present application is driven by the time series set of environmental boundary conditions and the voltage profile change as a unified time base, loads the insulator which has completed layered deposition, obtains the time series of electrical response and the time series of surface state field, and obtains the median flashover voltage by flashover test and statistical method, calculates the related increment in the synchronous window, and generates the electrical-matter response data set. The data set is used for subsequent closed-loop optimization link to measure the deviation of the connectivity threshold, migration rate and target deposition distribution.

[0065] In principle, the time series set of environmental boundary conditions defines the time-varying trajectory of relative humidity, temperature, wind speed and pollutant deposition flux, and the voltage profile change defines the amplitude program and transient segment of the applied voltage. After alignment on the same time axis, the external excitation and response of each segment can be established in a one-to-one correspondence. The time series of electrical response at least includes the time series of leakage current and the flashover event mark synchronized therewith; the time series of surface state field at least includes the time series of water film thickness field and the time series of surface equivalent conductance field. In order to avoid conceptual ambiguity, this paper refers to the loading window marked by start and end time on the time axis as segment, and refers to the difference value of the same grid or the same channel at the end time and the start time of the segment as increment.

[0066] In the estimation of the median flashover voltage, the present application adopts the step flashover test and carries out statistical regression. A plurality of test points are formed by voltage stepping, and whether flashover occurs is recorded for each test point. Let the voltage be , and the flashover probability be . A probability model is established by using logistic regression: The median flashover voltage is defined as the voltage that makes the probability equal to one-half: , wherein and are regression coefficients. The median flashover voltage reflects the tolerance threshold under the given deposition configuration and environmental program, and is written as a core scalar in the electrical-matter response data set. The strength and connectivity of the electrical response are described by the segment energy and the peak equivalent indicators, and the leakage current segment energy is defined as: , wherein is the time series of leakage current, and are the start and end time of the segment. The increment of the surface state field is calculated in grid units, for example, the surface equivalent conductance field increment: , and the water film thickness field increment: , wherein is the spatial grid coordinate, and are the time-space distribution of surface equivalent conductance and water film thickness, respectively. All increments are written with the same time key and spatial index, ensuring the repeatability of cross-segment comparison and regression identification.

[0067] The implementation points are as follows. First, the programming of the loading procedure follows the order of "environment first, voltage second", that is, the surface wetting state is set by the relative humidity step or platform segment first, and then the voltage step or voltage transient is applied in the window. Second, the acquisition system is time-stamped with a time reference consistent with the time series set of environmental boundary conditions, the leakage current time series uses a high sampling rate channel, and the surface state field is obtained in snapshot mode synchronized with the voltage step, and if necessary, encrypted sampling is added at key segments. Third, the start and end time of all segments, the values of environmental quantities, the amplitude and polarity of the voltage, and the result marker of whether flashover or not are written into the dataset metadata together with the sensor quality flag, forming a replayable test record.

[0068] Example 7: Wetting platform plus voltage ladder. The relative humidity is raised to a stable platform before loading and remains constant, and the temperature and wind speed are constant. The voltage is raised in fixed steps until the first flashover occurs, and then repeated multiple times with the same step, obtaining samples with and without flashover. The median flashover voltage is obtained by logistic regression. The segment energy index shows that there is a significant corresponding relationship between the surface equivalent conductance field increment in the outer edge band region and the leakage current segment energy, verifying that the connected path written in the deposition stage is reproduced in the loading stage.

[0069] Preferably, the loading is performed in a segmented manner, and the loading sequence includes at least three types of segments of relative humidity step, temperature step, wind speed change, pollutant deposition flux pulse, and voltage transient. The start and end time of each segment is recorded with a time marker and synchronized with the time stamp of the acquisition channel.

[0070] The present application performs loading in a segmented manner, arranges the time series set of environmental boundary conditions and voltage working condition changes as a group of atomized segments, and synchronously records them with the time stamp of the acquisition channel in the same time reference. The purpose of segmented loading is to split the effects of wetting and drying, convection and diffusion, flux pulse and electromigration, and transient electric stress on the connected path into identifiable causal units, thereby supporting the estimation of the median flashover voltage and parameter identification at the statistical and dynamic levels at the same time.

[0071] In the application context of the present application, a segment refers to a control window with clear start and end time on a unified time axis, and each segment only changes a few exogenous quantities, thereby maintaining identifiability. The segment types include relative humidity step, temperature step, wind speed change, pollutant deposition flux pulse, and voltage transient. Each segment is written into the metadata with a time marker, and the time marker and the time stamp of the acquisition channel use the same public clock. The time drift is corrected at the acquisition end with a time correction message and again at the convergence end with an alignment algorithm. This ensures that the electrical response time series and the surface state field time series have a repeatable registration relationship at the segment boundaries.

[0072] For ease of control and analysis, a segment indicator function is used to formally describe the loading procedure:

[0073] where is an indicator function of the th segment, and are the segment start and end times, respectively. The environmental channels are given in the form of segment-wise superposition, for example, the relative humidity time series: , the wind speed time series: , the pollutant deposition flux time series: , and the voltage program: where is the relative humidity time series, is the wind speed time series, is the pollutant deposition flux time series, is the voltage program time series, with quantities with subscript being baseline values, and quantities with subscript being increments applied within the segment, is the set of relative humidity step segment indices, is the set of wind speed change segment indices, is the set of pollutant deposition flux pulse segment indices, is the set of voltage transient segment indices. The above expressions are not for derivation, but for explicit control of the mapping of quantities to segment boundaries, ensuring that the windows for subsequent response calculation and statistical estimation are consistent.

[0074] The response side acquires the leakage current time series and the surface state field time series with the same time reference. The electrical response strength within a segment is described by the segment energy and the peak, defined as: where is the leakage current time series. The surface state field is computed in increments of the grid, with the surface equivalent conductance field increment: and the water film thickness field increment: where is the spatial grid coordinate, is the time-space distribution of the surface equivalent conductance, is the time-space distribution of the water film thickness. For each segment, a response vector consisting of and and is constructed, along with the control increments and and and within the segment, written into the data structure of electrical and physical responses, forming the segment-level record of the electro-physical response.

[0075] The median flashover voltage is estimated by the step flashover test combined with statistical methods. Several segments in the voltage program are set as voltage step segments, and whether flashover occurs in each step segment is recorded. The flashover probability is modeled by logistic regression, and the median flashover voltage is solved to represent the tolerance threshold of the current deposition configuration and environmental program. Since all segments are recorded with a unified time marker, the estimation of the median flashover voltage can be associated with a specific environmental combination segment, facilitating the separation of the contribution of humidity and wind speed to the threshold.

[0076] The implementation points include three aspects. First, the arrangement strategy follows the principle of few variable changes, and only a few exogenous variables are changed in each segment, and the rest remain the same, avoiding confusion. Second, the synchronization strategy uses a common clock at the control end and the collection end at the same time, and the time marker falls on the sampling grid. If there is a cross-grid error, the two adjacent points are aligned by interpolation in the data aggregation link, and the alignment method is retained in the metadata. Third, the quality control strategy writes the sensor state and missing flag in each segment, and adds a quality label to the resampling and interpolation segment to ensure the consistency of subsequent statistical screening.

[0077] In embodiment 8, three types of segments of relative humidity step, wind speed change and voltage transient are selected. The process is to first increase the relative humidity to the platform, then apply a wind speed pulse on the platform, and finally apply a short-time voltage transient in the window. The start and end time of each type of segment is recorded in seconds, and the leakage current time series and surface state field time series have synchronized data in the three segment windows. The results show that the water film thickness field in the wind speed pulse segment has a negative increment, the leakage current segment in the voltage transient segment has a significant increase in energy, and the median flashover voltage increases compared with the baseline without wind speed pulse.

[0078] Within the leakage current constraint, based on the electrical-material response data set and the target deposition and response data set, the kinetic model of surface state and migration behavior is established with the deviation index of deposition configuration and conductive connectivity, electrical response and median flashover voltage as the optimization target, the time-varying trajectory of the electrode system and the time series set of environmental boundary conditions are updated, the pollutant spectrum parameter set is adjusted, and the converged parameter set is output.

[0079] Within the leakage current constraint, the present application uses the electrical-material response data set and the target deposition and response data set to construct a closed-loop optimization problem, and jointly corrects the time-varying trajectory of the electrode system, the time series set of environmental boundary conditions and the pollutant spectrum parameter set until the converged parameter set is obtained. The core idea is to use the kinetic model of surface state and migration behavior as a forward predictor, and to use the deposition configuration distance, conductive connectivity deviation, electrical response deviation and median flashover voltage deviation as target items to solve the minimization problem under the condition of ensuring the safety boundary of leakage current.

[0080] The kinetic model takes the water film thickness field, the soluble contaminant surface area density field and the insoluble particle surface area density field as state variables, takes the environmental boundary condition time series set and the electrode system time-varying trajectory as well as the contaminant deposition flux as exogenous driving, and takes the surface equivalent conductance field as a derived variable. In order to ensure the consistency of symbols, the following minimum expressions of state and derived variables are defined for coupling control variables and observables: , , , wherein, is the water film thickness, is the surface area density of soluble contaminant, is the surface area density of insoluble particle, is the relative humidity time series, is the surface normal electric field intensity, is the near-wall wind speed, and are the deposition flux time series of soluble contaminant and insoluble particle respectively, is the surface equivalent conductance, is the dry surface baseline conductance, is the skeleton effective occupancy, is the connectivity threshold, is the indicator function, and are the dimensionless coupling functions, is the non-negative coefficient. The numerical values of the coefficients and functions are obtained by identifying the electro-physical response data set in the calibration window. This set of equations is not used to derive the theoretical upper limit, but to generate the predicted quantities in the optimization loop, which can be aligned with the measured quantities item by item.

[0081] The optimization objective adopts the weighted sum of multiple indicators and is written as: wherein, represents the control variable set that needs to be updated, including the parameters of the electrode system time-varying trajectory, the incremental parameters of the environmental boundary condition time series set and the contaminant spectrum parameter set, is the deposition configuration distance, which is used to measure the spatial difference between the predicted soluble contaminant surface area density field and the target deposition distribution function, is the conductive connectivity deviation index, which is used to measure the difference between the thresholded surface equivalent conductance field connectivity structure and the target connectivity expectation, is the leakage current time series difference, which can be quantified by segment energy and peak value and aligned with the target response set, is the predicted median flashover voltage, is the target median flashover voltage, is the non-negative weight. The constraint condition is written as , wherein is a leakage current time series, is a leakage current upper limit, is a leakage current rate of change upper limit. The constraints are set within the boundaries of test safety and insulator thermal loading, as the feasible region of the optimization process.

[0082] The solution procedure is an iteration in discrete time domain. First, read the piecewise data of the electro-physical response dataset and the target deposition and response dataset, initialize the state field and weights on a unified time base and a unified grid. Second, use the kinetic model and the current control variable set to make a forward prediction, generate the water film thickness field, the soluble contaminant surface area density field, the insoluble particle surface area density field, and the surface equivalent conductivity field, and predict the median flashover voltage by the voltage program. Third, calculate the deposition configuration distance, the electrically conductive connectivity deviation, the leakage current time series difference, and the median flashover voltage deviation, form the cost function and its gradient information. The gradient can be obtained by the adjoint method or by perturbation estimation. Fourth, update the control variable set in the feasible region projection mode, the update of the time-varying trajectory of the electrode system takes the parameterized trajectory of the phase and amplitude as the variable, the update of the time series set of the environmental boundary condition takes the piecewise amplitude and the piecewise time length as the variable, and the update of the contaminant spectrum parameter set takes the particle size distribution, the charge-to-mass ratio, the volume fraction of soluble contaminants, and the volume fraction of insoluble particles as the variable. Fifth, check the constraints piece by piece, if any piece violates the leakage current upper limit or the rate of change upper limit, scale the update step in that piece and rewrite the control variable set. Sixth, judge convergence, if the cost function does not decrease enough or the control variable changes less than the threshold, output the converged parameter set.

[0083] The key points of implementation are the coupled update of the three types of controls and the alignment of data. The time-varying trajectory of the electrode system directly affects the ratio of the surface normal electric field to the tangential electric field, and then acts on the water film replenishment and particle capture; the time series set of the environmental boundary condition determines the rhythm of wetting, drying, and convection; the contaminant spectrum parameter set controls the injection intensity and charging ability of soluble contaminants and insoluble particles. The three are adjusted under the unified time key, which can avoid local optimization caused by adjusting only a single path.

[0084] In Example 9, the target is to improve the median flashover voltage. The initial condition is that the outer edge is band-shaped and connected, and the electrical response is relatively strong. The optimization process reduces the proportion of the tangential component of the time-varying trajectory of the electrode system, shortens the duration of the relative humidity platform, and reduces the volume fraction of soluble contaminants, so that the electrically conductive connectivity deviation index decreases, the leakage current piece energy decreases, and the median flashover voltage increases, finally satisfying the constraints and outputting the converged parameter set.

[0085] Preferably, the median flashover voltage is obtained by the ascending and descending step flashover test, the voltage steps of ascending and descending are consistent, and the statistical regression method is used to estimate the results of multiple tests.

[0086] The present application takes the ascending and descending step flashover test as the core, aligns the environmental boundary condition time series set with the voltage working condition change for execution, obtains flashover event data through the two paths of repeated ascending and descending steps, and estimates the median flashover voltage by using a statistical regression method. The method is used for quantifying the tolerance threshold under the layered pollutant deposition structure in the application of the present application, and provides a traceable scalar target and confidence information for closed-loop optimization.

[0087] In principle, the voltage program is discretized into several voltage steps, the step amplitude is consistent, and the step residence time is consistent within the sequence. Each step only changes the voltage amplitude, and the remaining exogenous quantity is given by the environmental boundary condition time series set and locked in the preset segment, ensuring causal identification. For each step, the binary result of whether flashover occurs or not is recorded, as well as the electrical response time series and surface state field time series synchronized therewith. The voltage is defined as , and the flashover probability under the voltage is . The probability-voltage relationship is established by using a logistic regression: , wherein is the regression coefficient. The median flashover voltage is defined as the voltage that makes the probability equal to 0.5: , in order to avoid systematic bias caused by path dependence, the ascending and descending steps use the same voltage step and residence time, and are included in the same regression when estimating, and a “path” indicator is introduced for covariate correction if necessary. If flashover does not occur at the lowest step or is certain to flashover at the highest step, the boundary samples are processed by using a censored likelihood in regression to ensure unbiased estimation.

[0088] In the embodiment, the voltage steps are generated by a programmable power supply, the step size , and the step residence time are set before the test and remain unchanged. The flashover event criterion adopts a double-threshold determination: the absolute value of the leakage current exceeds the threshold and the duration is not less than , or the voltage appears a relative drop exceeding the threshold within the time window . The event indicator is denoted as . In order to ensure time alignment, the start and end time of the voltage step is written into the metadata with a time marker, and the leakage current time series and the surface state field time series are recorded at the same time key, and the segment energy and increment are calculated at the step boundary, respectively, as the diagnostic quantity after regression but not the regression independent variable. After each round of ascending step, a fixed time recovery section is executed, followed by a descending step; the starting step and the ending step of the two paths are consistent, forming a symmetrical sampling design.

[0089] The connection with the rest of the invention is reflected in two points. First, the step dwell time is consistent with the length of the segment in the time series set of environmental boundary conditions, so that the voltage step and the relative humidity step, wind speed change, and pollutant deposition flux pulse form a one-to-one corresponding combined segment. Second, the flashover criterion is synchronized with the segment, so that the estimation of the median flashover voltage can be clearly attributed to a specific environmental combination, thereby supporting the subsequent decomposition analysis of the deposition configuration and the conductivity connectivity deviation.

[0090] The data processing procedure includes four steps. Step 1, segment the leakage current time series and the surface state field time series according to the steps, and calculate the segment energy and field increment of each step as quality control indicators. Step 2, organize the step voltage and flashover indicator of the rising and falling paths to form an independent and identically distributed sample set. Step 3, use maximum likelihood estimation to estimate the logistic regression coefficients , if necessary, add path indicators and test their significance. Step 4, calculate the median flashover voltage and its variance approximation, and write the segment quality label and deletion information into the electrical-object response dataset.

[0091] Example 10, for reproduction experiments. With the rated voltage as the scale, the initial step is set to , the highest step is set to , the step is , and the dwell time is seconds. First, perform the rising step and then perform the falling step, and the time series set of environmental boundary conditions maintains the relative humidity platform, constant temperature, and constant wind speed. A total of 30 step samples are obtained for the two paths, of which 14 samples occur flashover. The logistic regression obtains , , and thus . The segment energy distribution shows that the leakage current segment energy is significantly larger in samples with a high proportion of outer edge connectivity, and the regression residual is also relatively increased, indicating that the influence of connectivity on the threshold needs to be measured separately in the closed-loop optimization.

[0092] Preferably, the deviation indicators of the deposition configuration and the conductive connectivity are obtained based on the thresholding of the surface equivalent conductance field to generate the conduction graph, the deposition configuration difference is measured using the optimal transport distance, and the conductive connectivity difference is measured using the persistent homology bottleneck distance.

[0093] In the electrical and surface state linkage evaluation, the present invention unifies "deposition configuration difference" and "conductive connectivity difference" into calculable deviation indicators. The core approach is to first threshold the surface equivalent conductance field to generate the conduction graph, and then measure the deposition configuration and conductive connectivity respectively. Through this sequence, the continuous field information is mapped into comparable objects in the topological and geometric dual domains, facilitating the use as target quantities in closed-loop optimization.

[0094] First, the field-to-map mapping is performed. The surface equivalent conductance field on the same time key is chosen as the input, and the corresponding field comes from the electrical and physical response dataset. The surface equivalent conductance expressed in the grid coordinates is denoted as , , where represents the spatial index on the insulator surface parameterized grid. A threshold value is set as The set of conducting regions is obtained after The adjacency relationship is established on the parameterized grid in the four-neighborhood or eight-neighborhood, and the conducting map is constructed , where is the set of grid nodes satisfying the threshold condition, is the set of edges between adjacent conducting nodes. The threshold is determined by the principle of maximum correlation with the leakage current segment energy: the correlation coefficient between the connected component features under different thresholds and the leakage current segment energy is calculated on a set of candidate thresholds, and the threshold with the maximum correlation is selected as the working threshold. This approach is an operational detail in the field, ensuring the consistency of the conducting map and the electrical response. The deposition configuration difference is used to measure the spatial mismatch between the surface soluble contaminant area density and the target deposition distribution. The normalized soluble contaminant surface area density field is denoted as , and the normalized result of the target deposition distribution function on the same grid is denoted as . Both are considered as probability measures defined on the surface parameter domain. The first-order optimal transport distance is used to calculate the deposition configuration difference with the surface geodesic distance as the cost metric: , where represents the parameter domain of the insulator surface, represents the geodesic distance obtained by weighting along the creeping path and the umbrella skirt busbar direction on the parameter domain, represents all coupling sets with and as edges, represents another position point on the insulator surface parameter domain . The calculation is implemented under the mask of the conducting map, i.e., only integrating in the conducting area to highlight the deposition mismatch related to electrical communication. This distance is sensitive to "misplacement", "shift", and "void" in space, and can directly reflect the geometric difference between the deposition amount and the target.

[0095] The conductive connectivity difference is used to measure the deviation of the topological structure of the conducting map from the target connectivity expectation. The topological evolution is constructed by super-level set filtering of the surface equivalent conductance field, and the threshold is gradually swept from high to low. The connected components and loops of the conducting area under each threshold are calculated to obtain the corresponding persistent map. The measured field persistent map is denoted as , and the persistent map corresponding to the target connectivity template is denoted as . The difference between the two is measured by the bottleneck distance: wherein is the bottleneck distance between two persistent graphs, is the infinity norm. The bottleneck distance is sensitive to the topological events of "channel merging" "bridging" "splitting", can stably depict the generation and disappearance of conductive paths, and is not sensitive to local noise, and is suitable for cross-batch comparison.

[0096] The two types of indicators are used in closed-loop optimization in the following way: the and are added to the cost function as independent target items, corresponding to the deposition geometry and conductive topology respectively, and are calculated under the same time key and the same grid, avoiding data alignment errors. In order to form a unified scale with the electrical response, dimensionless processing is adopted, and the two items are each divided by a reference value, which comes from the no-load baseline or historical median level.

[0097] Example 11, for verification of a belt-shaped connected target. The target connection template defines a continuous belt at the outer edge, and the target deposition distribution function assigns a higher weight to the belt. After performing deposition under the wetting platform, the surface equivalent conductance field is obtained. Thresholding generates the conduction graph and calculates the two indicators. The results show that the deposition configuration difference is small and the connectivity difference is large, indicating that the deposition amount has been in place but the conduction path has been locally disconnected. Subsequent closed-loop optimization adjusts only the tangential field ratio of the electrode system and the wetting segment length without increasing the total deposition amount, so that the conduction graph appears continuous bridging, the bottleneck distance decreases, and the median flashover voltage rises to the target interval.

[0098] Preferably, the kinetic model of surface state and migration behavior is established by sparse structure identification, the candidate basis function set and state derivative are obtained by sparse regression to obtain an analytical form, the time-varying trajectory of the electrode system is updated after the gradient is calculated by the adjoint method, and the parameter set of the contaminant spectrum is adjusted by introducing the feasible region constraints of the upper limit of the leakage current and the upper limit of the leakage current rate and adopting Bayesian optimization.

[0099] The present application is aimed at the closed-loop link of artificial contamination test under high voltage environment, around the link of "observable - identifiable - controllable", first, the kinetic model of surface state and migration behavior is established by sparse structure identification, second, the gradient of the time-varying trajectory of the electrode system is calculated and updated by the adjoint method, and finally, the parameter set of the contaminant spectrum is adjusted within the feasible region of the upper limit of the leakage current and the upper limit of the leakage current rate by Bayesian optimization, so that the deposition configuration difference, the conductive connectivity difference, the electrical response difference and the median flashover voltage deviation are simultaneously reduced.

[0100] The kinetic model takes the water film thickness field, the soluble contaminant surface density field and the insoluble particle surface density field as state variables, takes the time sequence set of environmental boundary conditions and the time-varying trajectory of the electrode system and the deposition flux as exogenous driving, and takes the surface equivalent conductance field as a derived quantity. Let the state vector be , the exogenous vector is . In the sparse structure identification, a candidate basis function set containing constant term, linear term, cross term and physical heuristic term is constructed , the state derivative is obtained using denoising derivative estimation, and then sparse regression is performed to obtain the analytical form , where is the sparse coefficient vector. The state variable meanings are as follows: is the water film thickness; is the surface area density of soluble pollutants; is the surface area density of insoluble particles. The exogenous variable meanings are as follows: is the relative humidity time series; is the temperature time series; is the near-wall wind speed time series; is the surface normal electric field intensity time series; and are the deposition flux time series of soluble pollutants and insoluble particles respectively. In order to ensure physical consistency, non-negative constraints are imposed on part of the coefficients, and sign priors are imposed on diffusion and convection related terms, and cross validation and left-out segment test are used to verify robustness. The surface equivalent conductance is given by the minimum necessary coupling , where is the surface equivalent conductance, is the dry surface baseline conductance, is the coupling coefficient. The above formula is used to map the state variable to the measurable electrical variable, which is convenient for alignment with the electrical and physical response data set.

[0101] The time-varying trajectory of the electrode system is represented in a parameterized manner, and the driving potential of each electrode channel is written as a spline coefficient or an orthogonal basis expansion, and the summary parameters are denoted as . The target quantity is a multi-objective cost function , where is the deposition configuration difference index, is the conductive connectivity difference index, is the leakage current time series difference index, is the predicted median flashover voltage, is the target median flashover voltage, is the non-negative weight. is obtained by adjoint method . Let the dynamic equation be , the Lagrange quantity is , the adjoint equation is written as , , where For adjoint variables, state trajectories are obtained by forward integration, and adjoint trajectories are obtained by backward integration, and then the gradients of time-varying trajectory parameters of the electrode system are obtained for line search or quasi-Newton update. The process is aligned with the data field item by item under a unified time reference, ensuring that each update can be played back and checked.

[0102] The pollutant spectrum parameter set includes particle size distribution parameters, mass-to-charge ratio, soluble pollutant volume fraction, and insoluble particle volume fraction. The feasible region constraint is applied within the test safety boundary , , where is a time series of leakage current, is an upper limit of leakage current, is an upper limit of leakage current rate of change. Based on the Gaussian process, a proxy model of the target and the constraint is constructed, and the next set of spectrum parameters is selected within the feasible region using the constrained expected improvement criterion, and the iteration is updated after observing the response. By multiplying the acquisition function of the feasible probability and the expected improvement, safety and performance are jointly included in the decision.

[0103] Embodiment 12 gives the gradient update of the time-varying trajectory of the electrode system. Taking the edge band target as an example, the initial tangential field ratio is too large, resulting in a high difference index of conductive connectivity. After calculating the gradient according to the above-mentioned adjoint step, the phase difference of the two main channels is reduced and the residence time in the high field area is shortened. The next round of forward prediction shows that the bottleneck distance decreases, the leakage current segment energy decreases, and the median flashover voltage converges to the target interval.

[0104] The above is only an embodiment of the present application and does not limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A method for artificial contamination testing under high pressure, characterized in that, Includes the following steps: A parameterized mesh for the insulator surface is established, the target deposition distribution function and target response set are generated, and the environmental boundary condition time series set is merged to obtain the target deposition and response dataset; In the energized state, a time-varying electric field distribution is formed through the electrode system, soluble pollutant aerosols and insoluble particulate aerosols are supplied to the surface of the insulator and particulate charge control is implemented. Based on the time series set of environmental boundary conditions and the supply switching, a layered pollutant deposition structure is constructed, surface state field data is collected, and an in-situ deposition state dataset is generated. Loading is performed based on the time series set of environmental boundary conditions and voltage condition changes. The time series of electrical response and surface state field are collected. The median flashover voltage is obtained by flashover test and statistical methods. The relevant increments are calculated to generate an electro-physical response dataset. Within the leakage current constraint, based on the electrophysiological response dataset and the target deposition and response dataset, a dynamic model of surface state and migration behavior is established with the deviation indices of deposition configuration and conductive connectivity, and electrical response and median flashover voltage as optimization objectives. The time series set of time-varying trajectory of electrode system and environmental boundary conditions is updated, the pollutant spectrum parameter set is adjusted, and the convergence parameter set is output.

2. The method according to claim 1, characterized in that, The environmental boundary condition time series set consists of relative humidity time series, temperature time series, wind speed time series and pollutant deposition flux time series, and is sampled and aligned with a unified time reference as input to the target deposition and response dataset.

3. The method according to claim 1, characterized in that, The electrode system is a multi-electrode shaping array. The phase and amplitude of the electrode channels are set independently. The electrode channels are arranged with a fixed geometric spacing. A time-varying electric field distribution that moves along the axial and radial directions of the insulator is formed by sequential driving.

4. The method according to claim 1, characterized in that, Particle charge control employs corona charging, with the charge polarity being either unipolar positive or unipolar negative. The particle charge is measured using a Faraday cylinder in conjunction with an electrostatic meter, and the measured values ​​are used to correct the aerosol supply parameters.

5. The method according to claim 1, characterized in that, The construction sequence of the stratified pollutant deposition structure is as follows: first, insoluble particulate aerosols are supplied to form a particulate skeleton; then, in the wetting stage, the relative humidity is increased and soluble pollutant aerosols are supplied to form a coating layer; and finally, in the drying stage, the relative humidity is reduced to stabilize the stratified structure.

6. The method according to claim 1, characterized in that, The surface state field data includes the water film thickness field, the surface equivalent conductivity field, and the soluble pollutant surface concentration field. The water film thickness field is obtained by conversion through non-contact optical measurement, the surface equivalent conductivity field is obtained by segmented measurement through a multi-point electrode array and inversion, and the soluble pollutant surface concentration field is obtained by either micro-elution conductivity measurement or spectral measurement.

7. The method according to claim 1, characterized in that, Loading is performed in segments. The loading sequence includes at least three types of segments: relative humidity step, temperature step, wind speed change, pollutant deposition flux pulse, and voltage transient. The start and end times of each segment are recorded with timestamps and synchronized with the timestamps of the acquisition channels.

8. The method according to claim 1, characterized in that, The median flashover voltage was obtained through a step-up flashover test, with the voltage steps for both rising and falling being consistent. Statistical regression was used to estimate the results of multiple tests.

9. The method according to claim 1, characterized in that, The deviation index between deposition configuration and electrical connectivity is measured after obtaining the conduction map by thresholding the surface equivalent conductivity field. The difference in deposition configuration is measured by the optimal transmission distance, and the difference in electrical connectivity is measured by the persistent coherence bottleneck distance.

10. The method according to claim 1, characterized in that, The dynamic model of surface state and migration behavior is established through sparse structure identification. The candidate basis function set and the state derivative are sparsely regressed to obtain the analytical form. The time-varying trajectory of the electrode system is updated after obtaining the gradient by the adjoint method. The set of pollutant spectrum parameters is adjusted by introducing feasible region constraints of upper limit of leakage current and upper limit of leakage current change rate and using Bayesian optimization.

Images