Electrochemical impedance spectroscopy-based plating adhesion evaluation method and system

Abstract

The application belongs to the technical field of electrochemical detection, and particularly relates to a plating layer adhesion force evaluation method and system based on electrochemical impedance spectroscopy. The method comprises the following steps: obtaining electrochemical impedance spectroscopy data and environmental factor data of the plating layer, forming interface state data and determining an initial reference state; extracting impedance spectroscopy features, combining the environmental factor data, a detection position identifier and the initial reference state to construct a correlation feature matrix; determining interface state distribution and interface abnormal information according to the correlation feature matrix; determining an adhesion force representation value based on an adhesion force calibration model, and calculating an adhesion force decay rate; updating a historical degradation sequence according to the interface abnormal information and the adhesion force decay rate, adjusting a preset prediction rule according to a clustering result, and outputting an adhesion force degradation monitoring result. The application can realize quantitative evaluation and degradation trend monitoring of the interface state of the plating layer.

Description

Technical Field

[0001] This application belongs to the field of electrochemical detection technology, specifically relating to a coating adhesion evaluation method and system based on electrochemical impedance spectroscopy. Background Technology

[0002] The quality of the coating directly affects the corrosion resistance, service life, and maintenance costs of metal components. In environments with high humidity, contact with corrosive media, and temperature fluctuations, the bonding between the coating and the substrate gradually deteriorates due to media penetration, interfacial corrosion, and microcrack propagation. Existing adhesion testing methods mostly rely on mechanical methods such as pull-out, cross-cutting, bending, or peeling. While these methods provide visual results, they are typically destructive, making it difficult to continuously monitor the same component and promptly identify early signs of adhesion degradation. Some online monitoring methods utilize electrical sensors to collect interfacial response data, but these often remain at the level of single impedance changes or threshold alarms, failing to comprehensively correlate impedance spectrum characteristics, environmental factors, and historical degradation processes. This results in test results being easily affected by temperature, humidity, the conductivity of the corrosive medium, and differences in measurement location.

[0003] In practical applications, the initial state of different detection locations varies, and the degradation rate of the same coating is also inconsistent at different corrosion stages. Without a baseline state, environmental correction, and historical pattern recognition, problems such as misjudgment, missed judgment, or unstable prediction are likely to occur. Summary of the Invention

[0004] To address the above issues, this application provides a coating adhesion evaluation method and system based on electrochemical impedance spectroscopy, which at least solves the problem of how to quantitatively monitor and predict coating adhesion degradation based on electrochemical impedance spectroscopy data and environmental factor data, and adaptively adjust the rules.

[0005] To achieve the above objectives, the technical solution adopted in this application is as follows: In a first aspect, this application provides a method for evaluating coating adhesion based on electrochemical impedance spectroscopy, the method comprising: Electrochemical impedance spectroscopy data and environmental factor data of the coating are acquired to form interface state data including detection location identification. The initial baseline state is determined by the interface state data of the initial acquisition period. Impedance spectrum features are extracted from interface state data, and an associated feature matrix is ​​constructed based on the impedance spectrum features, environmental factor data, detection location identifiers, and initial baseline state. The interface state distribution is determined based on the correlation feature matrix, and the interface anomaly information is determined based on the offset of the interface state distribution relative to the initial baseline state. The adhesion characterization value is determined based on the correlation feature matrix and adhesion calibration model, and the adhesion decay rate is determined based on the decrease in the adhesion characterization value and environmental factor data. The historical degradation sequence is updated based on interface anomaly information and adhesion decay rate. The preset prediction rules are adjusted based on the clustering results of the historical degradation sequence, and the adhesion degradation monitoring results are output.

[0006] In one possible implementation, electrochemical impedance spectroscopy data and environmental factor data of the coating are acquired to form interface state data including detection location identifiers. This includes: acquiring the impedance spectrum magnitude and impedance spectrum phase angle of the coating at multiple detection locations and multiple frequency points at the acquisition time, and acquiring the temperature, relative humidity, and conductivity of the corrosive medium corresponding to the acquisition time; and corroding the detection location identifier, acquisition time, impedance spectrum magnitude, impedance spectrum phase angle, temperature, relative humidity, and conductivity of the corrosive medium corresponding to each detection location to obtain interface state data.

[0007] In one possible implementation, impedance spectrum features are extracted from interface state data, including: determining the low-frequency impedance magnitude based on the impedance spectrum magnitude, determining the phase angle change based on the impedance spectrum phase angle at adjacent acquisition times and the same frequency points at the same detection location, and determining the slope of the impedance spectrum curve based on the impedance spectrum magnitude at adjacent frequency points at the same acquisition time at the same detection location; and determining the low-frequency impedance magnitude, phase angle change, and impedance spectrum curve slope as impedance spectrum features.

[0008] In one possible implementation, an associated feature matrix is ​​constructed based on impedance spectrum characteristics, environmental factor data, detection location identifiers, and initial reference states. This includes: determining the detection location to which the impedance spectrum characteristics belong based on the detection location identifiers, and determining the feature difference corresponding to the impedance spectrum characteristics based on the initial reference states; writing the impedance spectrum characteristics, environmental factor data, and feature difference corresponding to each detection location at the same acquisition time into the same matrix row of the associated feature matrix to obtain the associated feature matrix.

[0009] In one possible implementation, determining the interface state distribution based on the correlation feature matrix includes: determining the interface state evaluation value corresponding to each detection position based on the correlation feature matrix, and forming an interface state distribution based on the interface state evaluation value corresponding to each detection position; determining interface anomaly information based on the offset of the interface state distribution relative to the initial reference state includes: determining the initial state evaluation value corresponding to each detection position based on the initial reference state, calculating the difference between the interface state evaluation value corresponding to each detection position and the initial state evaluation value, and obtaining the offset; if the offset corresponding to any detection position is greater than the anomaly determination threshold, determining any detection position as an anomaly detection position, and generating interface anomaly information including the anomaly detection position.

[0010] In one possible implementation, the adhesion characterization value is determined based on the correlation feature matrix and the adhesion calibration model, including: obtaining an adhesion calibration model based on calibration samples, each calibration sample including impedance spectrum sample data, environmental factor sample data and mechanical adhesion test results; inputting the correlation feature matrix into the adhesion calibration model to obtain the adhesion characterization value, which is used to characterize the bonding state between the coating and the substrate.

[0011] In one possible implementation, the adhesion decay rate is determined based on the decrease in adhesion characterization value and environmental factor data, including: determining the basic decay rate based on the decrease in adhesion characterization value and the time interval corresponding to adjacent evaluation times; determining an environmental correction coefficient based on environmental factor data; and correcting the basic decay rate based on the environmental correction coefficient to obtain the adhesion decay rate.

[0012] In one possible implementation, updating the historical degradation sequence based on interface anomaly information and adhesion decay rate includes: determining historical records matching the detection location identifier based on the detection location identifier; writing the interface anomaly information, adhesion decay rate, and environmental factor data corresponding to the evaluation time into the historical records according to the evaluation time to obtain the updated historical degradation sequence.

[0013] In one possible implementation, the preset prediction rules are adjusted based on the clustering results of the historical degradation sequences, including: clustering the historical degradation sequences based on interface anomaly information, adhesion decay rate, and environmental factor data in the historical degradation sequences to obtain clustering results; adjusting the feature weights, risk thresholds, and prediction time windows in the preset prediction rules based on the clustering results; and outputting adhesion degradation monitoring results, including: outputting adhesion degradation monitoring results under humid and corrosive scenarios based on the adjusted preset prediction rules.

[0014] Secondly, this application provides a coating adhesion evaluation system based on electrochemical impedance spectroscopy, used to implement a coating adhesion evaluation method based on electrochemical impedance spectroscopy. The system includes: The data acquisition module is used to acquire electrochemical impedance spectroscopy data and environmental factor data of the coating, and form interface state data including detection location identification. The initial reference state is determined by the interface state data of the initial acquisition period. The matrix construction module is used to extract impedance spectrum features from interface state data and construct an associated feature matrix based on impedance spectrum features, environmental factor data, detection location identifiers and initial reference states. The anomaly determination module is used to determine the interface state distribution based on the correlation feature matrix and to determine interface anomaly information based on the offset of the interface state distribution relative to the initial baseline state. The rate calculation module is used to determine the adhesion characterization value based on the correlation feature matrix and adhesion calibration model, and to determine the adhesion decay rate based on the decrease in the adhesion characterization value and environmental factor data. The sequence prediction module is used to update the historical degradation sequence based on interface anomaly information and adhesion decay rate, adjust the preset prediction rules based on the clustering results of the historical degradation sequence, and output the adhesion degradation monitoring results.

[0015] Compared with existing technologies, the advantages and beneficial effects of this application are as follows: By acquiring electrochemical impedance spectroscopy data and environmental factor data of the coating, and associating the detection location markers with the interface state data, continuous recording and spatial differentiation of the interface state at different detection locations were achieved.

[0016] By determining the initial baseline state from the interface state data of the initial acquisition period, the current interface state and the initial interface state can be compared, reducing the impact of initial differences at different locations on the judgment results.

[0017] By extracting impedance spectral features such as low-frequency impedance modulus, phase angle change, and impedance spectral curve slope, and combining them with environmental factor data to construct a correlation feature matrix, a unified expression of electrochemical response, corrosion environment, and spatial location was achieved.

[0018] By determining the interface state distribution based on the correlation feature matrix and determining the interface anomaly information based on the offset of the interface state distribution relative to the initial baseline state, the identification of local anomalies and early degradation signs is realized.

[0019] By determining the adhesion characterization value through an adhesion calibration model, and determining the adhesion decay rate based on the decrease in the adhesion characterization value and environmental factor data, a quantitative mapping effect from impedance spectrum response to the degree of adhesion degradation is achieved.

[0020] By adjusting the preset prediction rules through the clustering results of historical degradation sequences, the prediction parameters are updated as the degradation mode changes, making the adhesion degradation monitoring results more adaptable to long-term changes under humid and corrosive conditions. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the coating adhesion evaluation method based on electrochemical impedance spectroscopy in this application; Figure 2 This is a block diagram of the module composition of the coating adhesion evaluation system based on electrochemical impedance spectroscopy in this application. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solution, the present application will be described in detail below with reference to the embodiments. The description in this section is only exemplary and explanatory, and should not be used to limit the scope of protection of the present application in any way.

[0023] Electrochemical impedance spectroscopy (EIS) is a detection method that characterizes the electrochemical behavior of material interfaces by applying a small AC perturbation and measuring the current response. The results typically reflect charge transfer, media penetration, and interfacial polarization states between the coating, corrosive medium, and substrate interface as impedance spectral modulus, phase angle, and frequency response curves. For coating structures, interfacial porosity expansion, corrosive medium intrusion, and changes in bonding state cause variations in impedance response with frequency and time. Therefore, impedance spectral data from different detection locations can be collected using electrical sensors, and the interface state can be continuously characterized by incorporating environmental factors. Based on these characteristics, this application correlates electrochemical impedance spectroscopy data with environmental factor data, initial baseline conditions, and historical degradation sequences to form an evaluation method that reflects the degradation process of coating adhesion.

[0024] like Figure 1 As shown, a coating adhesion evaluation method based on electrochemical impedance spectroscopy is described, the method comprising: Electrochemical impedance spectroscopy data and environmental factor data of the coating are acquired to form interface state data including detection location identification. The initial baseline state is determined by the interface state data of the initial acquisition period. In this embodiment, for the coated component to be evaluated, an electrochemical impedance spectroscopy (EIS) acquisition terminal is arranged on the coating surface or in a detection area related to the coating interface state, and the detection area is divided into multiple detection locations. Each detection location is equipped with a unique detection location identifier to distinguish the impedance response and environmental exposure state under different spatial locations. The acquisition terminal acquires the electrochemical impedance spectroscopy data of the coating according to a set acquisition cycle, and simultaneously acquires environmental factor data. The electrochemical impedance spectroscopy data includes the impedance spectrum magnitude and impedance spectrum phase angle, and the environmental factor data includes the temperature, relative humidity, and conductivity of the corrosive medium corresponding to the same acquisition time. The system writes the detection location identifier, acquisition time, electrochemical impedance spectroscopy data, and environmental factor data into the same interface state record to form interface state data. For the interface state data that is in a stable state during the initial acquisition period, the system removes obviously abnormal records to obtain the initial reference state used to characterize the initial interface state of the coating. The initial reference state serves as the benchmark input for subsequent impedance spectroscopy feature comparison and interface anomaly judgment.

[0025] Electrochemical impedance spectroscopy (EIS) data and environmental factor data of the coating are acquired to form interface state data including detection location identifiers. This includes: acquiring the impedance spectrum magnitude and phase angle of the coating at multiple detection locations and multiple frequency points at the acquisition time, and acquiring the temperature, relative humidity, and conductivity of the corrosive medium at the corresponding acquisition time; and corroding the detection location identifier, acquisition time, impedance spectrum magnitude, impedance spectrum phase angle, temperature, relative humidity, and conductivity of the corrosive medium at each detection location to obtain interface state data.

[0026] In one embodiment, the data acquisition and correlation process is further defined. The coated component can be divided into detection locations based on its service location, corrosive medium contact area, coating thickness variation area, or historically prone-to-failure area. Each detection location corresponds to a detection location identifier, which can be formed by a combination of component number, area number, and sensor number, or can be automatically generated by the system according to the spatial order of the detection areas. The detection location identifier is not used as a simple record field, but rather as a spatial index for subsequent correlation of feature matrices, interface state distribution, and historical degradation sequences, enabling continuous tracking of data generated at the same detection location at different times and allowing for horizontal comparison of data between different detection locations.

[0027] Electrochemical impedance spectroscopy (EIS) data were acquired at multiple frequency points. These frequency points covered the high-frequency, mid-frequency, and low-frequency regions. The high-frequency region reflected solution resistance, lead contact, and surface electrolyte state; the mid-frequency region reflected coating shielding performance, porosity expansion, and dielectric permeation; and the low-frequency region reflected corrosion reactions and charge transfer processes at the coating-substrate interface. The system sampled at each detection location according to the same frequency sequence, obtaining the impedance spectrum magnitude and phase angle at that detection location at the acquisition time. The impedance spectrum magnitude represents the degree of resistance of the interface to the current response under AC disturbance, and the impedance spectrum phase angle represents the phase difference between the voltage and current responses; both together reflect the electrochemical state of the coating interface.

[0028] Environmental data and electrochemical impedance spectroscopy (EIS) data are bound together at the same acquisition time. Temperature can be obtained from a temperature sensor near the detection area, relative humidity from an ambient humidity sensor, and the conductivity of the corrosive medium can be obtained from the conductivity detection unit of the contact medium or condensation medium. For immersion, salt spray, or humid corrosive environments, the conductivity of the corrosive medium is used to characterize the ion migration ability in the medium; relative humidity is used to characterize the possibility of electrolyte film formation and persistence; and temperature is used to reflect the influence of environmental changes on the electrochemical reaction rate and material interface state. The acquisition time is uniformly recorded by the system clock or data acquisition controller, and data at the same acquisition time are considered synchronous acquisition results under the same environmental conditions.

[0029] When generating interface state data, the system records the detection location identifier, impedance spectrum magnitude, impedance spectrum phase angle, temperature, relative humidity, and corrosive medium conductivity for each detection location at each acquisition time into the same data record. If a detection location lacks an impedance spectrum magnitude or phase angle at a certain frequency point, the system marks that data record as incomplete and discards or supplements it before entering the initial reference state calculation. If environmental factor data lacks records at the same acquisition time, the system selects the nearest environmental record with an acquisition time difference less than the set synchronization error range for matching; if a match cannot be found, the interface state data at that acquisition time will not participate in the initial reference state determination. The synchronization error range can be set according to the acquisition cycle, usually less than half of one acquisition cycle, to avoid data mismatch caused by rapid environmental fluctuations.

[0030] The initial data acquisition period can be selected from a continuous acquisition window before the coating begins service, before corrosion exposure begins, or in the early stages of stable exposure. During the initial acquisition period, the system performs stability screening on the interface state data at each detection location. Screening criteria include the fluctuation range of the impedance spectrum modulus, the continuity of the impedance spectrum phase angle, the amplitude of temperature changes, the amplitude of relative humidity changes, and the amplitude of changes in the conductivity of the corrosive medium. Acquisition records exceeding the stable range are excluded as abnormal acquisition records. The stable range can be set based on historical test data of the same batch of coating samples, the accuracy of the testing equipment, and the fluctuation range of the actual service environment. The screened interface state data is used to determine the initial baseline state. The initial baseline state is saved separately for each detection location, ensuring that each detection location has an independent reference and avoiding the influence of differences in coating thickness, surface condition, or environmental contact conditions at different locations on subsequent judgments.

[0031] Impedance spectrum features are extracted from interface state data, and an associated feature matrix is ​​constructed based on the impedance spectrum features, environmental factor data, detection location identifiers, and initial baseline state. In this embodiment, after the interface state data enters the feature extraction stage, the system groups the electrochemical impedance spectroscopy (EIS) data according to the detection location identifier and acquisition time, enabling data from the same detection location at different acquisition times to form a continuous sequence. The system extracts the low-frequency impedance modulus, phase angle change, and impedance spectrum slope from each group of data as impedance spectral features characterizing the electrochemical state of the coating interface. The low-frequency impedance modulus reflects the interface corrosion reaction and charge transfer process, the phase angle change reflects the change in interface capacitance characteristics over time, and the impedance spectrum slope reflects the shape of the impedance spectrum as a function of frequency. The system combines the impedance spectral features with environmental factor data, detection location identifier, and initial baseline state to form a correlation feature matrix. This correlation feature matrix uses the detection location and acquisition time as indexes, allowing subsequent interface state distribution calculations to simultaneously utilize the current features, environmental conditions, and baseline differences.

[0032] Impedance spectrum features are extracted from interface state data, including: determining the low-frequency impedance magnitude based on the impedance spectrum magnitude, determining the phase angle change based on the impedance spectrum phase angle at adjacent acquisition times and the same frequency points at the same detection location, and determining the slope of the impedance spectrum curve based on the impedance spectrum magnitude at adjacent frequency points at the same acquisition time at the same detection location; the low-frequency impedance magnitude, phase angle change, and slope of the impedance spectrum curve are determined as impedance spectrum features.

[0033] In one embodiment, the impedance spectrum feature extraction process is further defined. The system reads data from multiple frequency points at the same detection location and at the same acquisition time, and arranges the impedance spectrum magnitude and impedance spectrum phase angle in ascending order of frequency. The low-frequency impedance magnitude can be selected from a preset low-frequency region, representing a frequency point, or the impedance spectrum magnitudes of multiple frequency points within the preset low-frequency region can be averaged. The preset low-frequency region is determined based on the frequency scanning range of the detection equipment and the interface response characteristics of the coating material. Typically, a low-frequency band that reflects the interface corrosion reaction and has a stable test signal is selected, while frequency points significantly affected by contact noise are not selected.

[0034] The phase angle change is used to represent the phase angle change at the same detection position at adjacent acquisition times. The calculation method can be expressed as follows: in, This is the change in phase angle. For the detection location number, For frequency point number, This is the data collection time sequence number. The detection location number is Frequency point number is The data collection time sequence number is Phase angle of the impedance spectrum at that time This refers to the impedance spectrum phase angle at the same detection location and frequency point at the previous acquisition time. The change in this phase angle is used to determine whether there is a continuous shift in the interface capacitance characteristics.

[0035] The slope of the impedance spectrum curve is used to represent the trend of impedance spectrum magnitude changes between adjacent frequency points at the same detection location and the same acquisition time. The calculation method can be expressed as follows: in, The slope of the impedance spectrum curve. The detection location number is Frequency point number is The data collection time sequence number is The impedance spectrum modulus at that time The detection location number is Frequency point number is The data collection time sequence number is The impedance spectrum modulus at that time and These represent the frequencies of two adjacent frequency points. The system writes the low-frequency impedance magnitude, phase angle change, and impedance spectrum slope into the feature record corresponding to the same detection location and the same acquisition time. If data for a certain frequency point is missing, the system prioritizes using supplementary data from the same detection location; if supplementary data cannot be obtained, that frequency point is not included in the impedance spectrum slope calculation to avoid missing values ​​from being included in subsequent matrix construction.

[0036] The associated feature matrix is ​​constructed based on impedance spectrum characteristics, environmental factor data, detection location identifiers, and initial reference states. This includes: determining the detection location to which the impedance spectrum characteristics belong based on the detection location identifiers, and determining the feature difference corresponding to the impedance spectrum characteristics based on the initial reference states; writing the impedance spectrum characteristics, environmental factor data, and feature difference corresponding to each detection location at the same acquisition time into the same matrix row of the associated feature matrix to obtain the associated feature matrix.

[0037] In one embodiment, the process of constructing the associated feature matrix is ​​further defined. After completing the impedance spectrum feature extraction, the system determines the detection location to which the impedance spectrum feature belongs based on the detection location identifier, and calls the reference features saved in the initial reference state for that detection location. The initial reference state may include the low-frequency impedance modulus reference value, phase angle reference value, and impedance spectrum curve slope reference value within the initial acquisition period. The reference features are determined using stable acquisition records. If there are obvious jump points or missing points within the initial acquisition period, the system excludes the record and does not use it as a source of reference features. In this way, different detection locations have independent references, which can avoid misjudging differences in coating thickness, local contact state, or edge area environment as degradation differences.

[0038] The characteristic difference is used to represent the degree of deviation of the current impedance spectrum characteristics from the initial reference state, and its calculation method can be expressed as: in, For characteristic differences, For the detection location number, This is the data collection time sequence number. This is the impedance spectrum characteristic number. The detection location number is The data collection time sequence number is The impedance spectrum characteristic number is The current impedance spectrum eigenvalues ​​at that time, This is the baseline characteristic value for the same detection location and the same impedance spectrum characteristics in the initial reference state. This characteristic difference is used to convert the current state of different detection locations into a deviation state relative to their respective references, providing a unified scale for subsequent interface state distribution and interface anomaly information judgment.

[0039] The correlation feature matrix organizes data according to detection location and acquisition time. Each matrix row corresponds to the interface state at a detection location and acquisition time. The matrix row includes the detection location identifier, acquisition time, low-frequency impedance modulus, phase angle change, impedance spectrum slope, temperature, relative humidity, corrosive medium conductivity, and characteristic differences corresponding to each impedance spectrum feature. The detection location identifier is used to locate the spatial source, the acquisition time is used to locate the temporal source, the impedance spectrum features are used to characterize the interface electrochemical response, environmental factor data is used to characterize humid corrosion conditions, and the characteristic differences are used to characterize the degree of deviation from the initial reference state. The system performs field integrity checks before writing to the matrix row. If the detection location identifier, acquisition time, or main impedance spectrum features are missing, the record is not included in the correlation feature matrix. If a certain environmental factor data is missing, the system can call data from nearby environmental sensors at the same acquisition time for substitution and record the substitution mark in the matrix row. Through this matrix structure, subsequent calculations can directly read spatial location, temporal sequence, impedance response, environmental conditions, and reference deviation information from the same data structure.

[0040] The interface state distribution is determined based on the correlation feature matrix, and the interface anomaly information is determined based on the offset of the interface state distribution relative to the initial baseline state. In this embodiment, after the associated feature matrix is ​​constructed, the system reads the matrix row corresponding to each detection location at the current acquisition time according to the detection location identifier, and obtains the impedance spectrum characteristics, environmental factor data, and feature differences relative to the initial reference state from the matrix row. The system performs dimensional consistency processing on various types of data, converting different physical quantities into comparable evaluation inputs, and then calculates the interface state evaluation value corresponding to each detection location according to the pre-configured evaluation weights. The interface state evaluation value is used to characterize the degree of interface degradation at the current acquisition time at the detection location. The system arranges the interface state evaluation values ​​of multiple detection locations according to the detection location identifier to form an interface state distribution. The system continues to call the initial state evaluation values ​​corresponding to each detection location in the initial reference state and calculates the offset of the current interface state evaluation value relative to the initial state evaluation value. When the offset exceeds the anomaly judgment threshold, the corresponding detection location is determined as an abnormal detection location, and the system generates interface anomaly information including the abnormal detection location, offset, and acquisition time.

[0041] The interface state distribution is determined based on the correlation feature matrix, including: determining the interface state evaluation value corresponding to each detection position based on the correlation feature matrix, and forming the interface state distribution based on the interface state evaluation value corresponding to each detection position; the interface anomaly information is determined based on the offset of the interface state distribution relative to the initial baseline state, including: determining the initial state evaluation value corresponding to each detection position based on the initial baseline state, calculating the difference between the interface state evaluation value corresponding to each detection position and the initial state evaluation value, and obtaining the offset; if the offset corresponding to any detection position is greater than the anomaly judgment threshold, any detection position is determined as an anomaly detection position, and interface anomaly information including the anomaly detection position is generated.

[0042] In one embodiment, the determination method for the interface state evaluation value, offset, and anomaly judgment threshold is further defined. Before calculating the interface state evaluation value, the system performs validity verification on the input fields in the associated feature matrix. Validity verification includes checking whether the detection location identifier exists, whether the acquisition time is continuous, whether the impedance spectrum characteristics are complete, whether the environmental factor data is within the device's measurement range, and whether the feature difference originates from the initial reference state of the same detection location. For matrix rows where the detection location identifier is missing or the acquisition time cannot be identified, the system does not participate in the interface state distribution calculation; for cases where a single environmental factor data is missing but neighboring sensor data exists at the same acquisition time, the system can use neighboring sensor data as a substitute and record the substitute source in the corresponding matrix row; for cases where low-frequency impedance modulus, phase angle change, or impedance spectrum curve slope is missing, the system marks the matrix row as a record to be retested and does not generate interface anomaly information to avoid misjudgment due to missing core electrochemical characteristics.

[0043] The interface status evaluation value can be calculated by weighting the normalized impedance spectrum characteristics and environmental factor data. The calculation method can be expressed as follows: in, The detection location number is The data collection time sequence number is The interface status evaluation value at that time. The number of input fields used in the evaluation. For the input field sequence number, The input field sequence number is The evaluation weight at that time The detection location number is The data collection time sequence number is The input field number is The normalized input values ​​are determined by the time. These normalized input values ​​can be derived from low-frequency impedance modulus, phase angle change, impedance spectrum slope, temperature, relative humidity, corrosive medium conductivity, and corresponding characteristic differences. Evaluation weights can be set based on the correlation between each input field in the calibration sample and the mechanical adhesion test results, or configured based on historical test data of similar coatings. For coating systems more sensitive to the decrease in low-frequency impedance modulus and changes in phase angle plateau, the evaluation weights corresponding to low-frequency impedance modulus and phase angle change can be increased; for application scenarios where the intensity of humid corrosion exposure changes significantly, the evaluation weights corresponding to relative humidity and corrosive medium conductivity can be increased.

[0044] The offset is used to represent the degree of difference between the current interface state and the initial interface state, and it can be calculated as follows: in, The detection location number is The data collection time sequence number is Offset at time, This is the current interface status evaluation value. This refers to the initial state evaluation value for the same detection location under the initial baseline state. The initial state evaluation value is calculated from stable interface state data within the initial acquisition period. If there are abnormal fluctuation records within the initial acquisition period, these abnormal fluctuation records are not included in the calculation of the initial state evaluation value. When the offset is positive and continues to increase, it indicates that the interface state at that detection location is showing a degradation trend relative to the initial state. If the offset increases for a short period of time but subsequently recovers, the system records this situation as a temporary environmental disturbance candidate record and does not directly use it as a basis for judging continuous degradation.

[0045] The anomaly detection threshold can be determined based on the coating material type, the fluctuation range of the evaluation value during the initial acquisition period, and the repeatability error of the detection equipment. Specifically, the system can use the upper limit of the normal fluctuation of the interface state evaluation value during the initial acquisition period as the basic threshold, and set a safety margin in conjunction with the error range of the detection equipment. For scenarios with long-term stable humid and corrosive environments, the anomaly detection threshold can be set relatively strictly; for scenarios with frequent temperature and humidity changes or large fluctuations in the conductivity of the corrosive medium, the anomaly detection threshold needs to be corrected in conjunction with environmental factor data. When the offset corresponding to any detection location is greater than the anomaly detection threshold, the system determines that detection location as an anomaly detection location and generates interface anomaly information. The interface anomaly information includes the anomaly detection location, acquisition time, interface state evaluation value, initial state evaluation value, offset, and the triggered anomaly detection threshold. This information is written into the subsequent adhesion characterization value calculation and historical degradation sequence update process to limit the detection locations that need to be tracked and the time of anomaly occurrence.

[0046] The adhesion characterization value is determined based on the correlation feature matrix and adhesion calibration model, and the adhesion decay rate is determined based on the decrease in the adhesion characterization value and environmental factor data. In this embodiment, after obtaining the correlation feature matrix and interface anomaly information, the system calls the adhesion calibration model to quantitatively characterize the bonding state between the coating and the substrate. The adhesion calibration model is obtained based on pre-established calibration samples, which include impedance spectrum sample data, environmental factor sample data, and mechanical adhesion test results. The system inputs the impedance spectrum features, environmental factor data, detection location identifiers, and feature differences from the correlation feature matrix into the adhesion calibration model, and outputs the adhesion characterization value at the corresponding detection location at the evaluation time. The adhesion characterization value is not equivalent to the directly measured mechanical pull-off force, but is a calculation result used to reflect the bonding state of the coating interface. The system reads the adhesion characterization values ​​of two adjacent times according to the evaluation time, calculates the basic attenuation rate, and determines the environmental correction coefficient based on the environmental factor data. The environmental correction coefficient participates in the basic attenuation rate correction to obtain the adhesion attenuation rate used for subsequent historical degradation sequence updates.

[0047] The adhesion characterization value is determined based on the correlation feature matrix and the adhesion calibration model, including: obtaining the adhesion calibration model based on the calibration samples, each calibration sample including impedance spectrum sample data, environmental factor sample data and mechanical adhesion test results; inputting the correlation feature matrix into the adhesion calibration model to obtain the adhesion characterization value, which is used to characterize the bonding state between the coating and the substrate.

[0048] In one embodiment, the establishment of the adhesion calibration model and the determination of adhesion characterization values ​​are further defined. The establishment of the adhesion calibration model relies on a calibration sample library. The calibration sample library can be established using samples from the same batch before the coated components are put into formal operation, or it can be continuously supplemented from periodically sampled samples, accelerated corrosion samples, and on-site verification samples. Each calibration sample includes impedance spectrum sample data, environmental factor sample data, and mechanical adhesion test results obtained under the same test conditions for the same test object. The impedance spectrum sample data includes low-frequency impedance modulus, phase angle change, slope of the impedance spectrum curve, and corresponding characteristic differences. The environmental factor sample data includes temperature, relative humidity, and conductivity of the corrosive medium. The mechanical adhesion test results can be obtained by pull-off test, cross-cut test, or shear test. The mechanical adhesion test results generated by different test methods are graded or numerically unified before entering the calibration sample library, so that the adhesion calibration model can obtain output results under the same caliber.

[0049] The adhesion calibration model can employ conventional supervised learning models such as linear regression, support vector regression, random forest regression, or gradient boosting regression, or it can be implemented using lookup table interpolation. The model type is determined based on the number of calibration samples and on-site computing resources. When the number of samples is small, linear regression or lookup table interpolation can be used, facilitating parameter interpretation and on-site deployment. When the number of samples is large and there is a non-linear relationship between impedance spectrum characteristics and mechanical adhesion detection results, support vector regression or tree models can be used. During model training, impedance spectrum sample data and environmental factor sample data are used as inputs, and mechanical adhesion detection results are used as the calibration target. After training, the system uses independent validation samples to check the output error. The allowable range is determined based on the repeatability error of the mechanical adhesion detection results and the evaluation accuracy requirements. If the error exceeds the allowable range, the calibration sample library needs to be supplemented with samples or the model needs to be retrained.

[0050] During operation, the system no longer performs destructive mechanical testing on the online coating. Instead, it inputs the associated feature matrix into the adhesion calibration model to obtain adhesion characterization values. These values ​​can be represented as continuous numerical values ​​or mapped to several state levels. When using continuous numerical values, lower values ​​indicate weaker bonding between the coating and the substrate. When using state levels, the level boundaries are determined by the distribution of mechanical adhesion test results in the calibration sample library. To avoid interference from abnormal acquisition records, the system checks whether the associated feature matrix contains complete impedance spectrum features and environmental factor data before inputting it into the model. If the core impedance spectrum feature is missing, no adhesion characterization value is output at the current evaluation time. If a single environmental factor data point is missing, the system can call a nearby environmental record from the same evaluation time as a substitute and mark the substitute source in the calculation record. The output adhesion characterization value is bound and saved with the detection location identifier and evaluation time, providing continuous input for subsequent attenuation rate calculations.

[0051] The adhesion decay rate is determined based on the decrease in adhesion characterization value and environmental factor data, including: determining the basic decay rate based on the decrease in adhesion characterization value and the time interval between adjacent evaluation times; determining the environmental correction coefficient based on environmental factor data, and correcting the basic decay rate based on the environmental correction coefficient to obtain the adhesion decay rate.

[0052] In one embodiment, the calculation and environmental correction method for the adhesion decay rate are further defined. The system calculates a base decay rate based on the adhesion characterization values ​​at adjacent evaluation times. The base decay rate represents the rate at which the bonding state between the coating and the substrate decreases between adjacent evaluation times, and its calculation method can be expressed as follows: in, The detection location number is The evaluation time sequence number is The basic decay rate at that time, The detection location number is The evaluation time sequence number is The adhesion characterization value at that time, This represents the adhesion characterization value at the same detection location at the previous evaluation time. For the evaluation time sequence number is The corresponding time, The input to this calculation relationship is the adhesion characterization value and the time interval between adjacent evaluation times, and the output is the base decay rate. If the current adhesion characterization value is higher than the previous evaluation time, the system will record the base decay rate as zero or mark the evaluation time as a recovery fluctuation record to avoid misinterpreting short-term measurement fluctuations as negative degradation.

[0053] The environmental correction factor is used to characterize the effect of a humid and corrosive environment on the rate of adhesion degradation. The system determines the environmental correction factor based on temperature, relative humidity, and the conductivity of the corrosive medium. The calculation method can be expressed as follows: in, The detection location number is The evaluation time sequence number is Environmental correction factor at that time, This is the normalized temperature value. This is the normalized value for relative humidity. This represents the normalized value of the conductivity of the corrosive medium. Temperature weighting, Weighted by relative humidity, The weights represent the conductivity of the corrosive medium. Normalized values ​​for temperature, relative humidity, and conductivity of the corrosive medium can be converted from the permissible range in the field or from historical sample ranges. Each weight can be determined based on the correlation between environmental factor data and mechanical adhesion test results in the calibration samples, or configured using corrosion test results of similar coating materials. If a certain environmental factor has a small impact on the degradation differences of the samples, its corresponding weight can be set to a smaller value.

[0054] The corrected adhesion decay rate can be expressed as: in, The detection location number is The evaluation time sequence number is The adhesion decay rate is calculated based on the baseline decay rate and environmental correction coefficients. The output is the adhesion decay rate used for updating historical degradation sequences. Before calculation, the system checks if the evaluation time interval meets the minimum time interval requirement. If the time interval is too short, the adhesion characterization value is easily affected by measurement noise, and the system does not calculate the adhesion decay rate. If the time interval is too long, the system marks the evaluation period as a low-temporal-resolution record and reduces the weight of this record in subsequent clustering. After the adhesion decay rate is calculated, the system saves it along with the detection location identifier, evaluation time, adhesion characterization value, environmental correction coefficient, and corresponding environmental factor data to ensure that subsequent historical degradation sequences can track the location, time, and environmental causes of degradation.

[0055] The historical degradation sequence is updated based on interface anomaly information and adhesion decay rate. The preset prediction rules are adjusted based on the clustering results of the historical degradation sequence, and the adhesion degradation monitoring results are output.

[0056] In this embodiment, after obtaining interface anomaly information and adhesion decay rate, the system retrieves the corresponding historical records according to the detection location identifier and writes the interface anomaly information, adhesion decay rate, and environmental factor data at the current evaluation time into the same historical record. The historical records are arranged according to the evaluation time, forming a historical degradation sequence that reflects the changes of the same detection location over time. When performing clustering processing on the historical degradation sequence, the system reads the anomaly location, offset, adhesion decay rate, and environmental factor data in the sequence and classifies sequences with similar trends into the same degradation category. The clustering results are used to adjust the feature weights, risk thresholds, and prediction time windows in the preset prediction rules. The adjusted preset prediction rules are used to output the adhesion degradation monitoring results under wet and corrosive scenarios. The monitoring results include risk locations, risk change trends, and subsequent monitoring cycles.

[0057] The historical degradation sequence is updated based on interface anomaly information and adhesion decay rate, including: determining historical records matching the detection location identifier based on the detection location identifier; writing interface anomaly information, adhesion decay rate, and environmental factor data corresponding to the evaluation time into the historical records according to the evaluation time to obtain the updated historical degradation sequence.

[0058] In one embodiment, the update process for the historical degradation sequence is further defined. The system uses the detection location identifier as the retrieval index for historical records, searching for historical records that match the detection location identifier in the database or local cache. If a matching historical record exists, the system appends the data of the current evaluation time to the matching historical record; if no matching historical record exists, the system creates a new historical record based on the detection location identifier, and uses the current evaluation time as the starting evaluation time of that historical record. The historical record includes at least the detection location identifier, evaluation time, interface anomaly information, adhesion decay rate, temperature, relative humidity, and conductivity of the corrosive medium. The interface anomaly information may include the anomaly detection location, offset, anomaly determination threshold, and anomaly persistence marker, wherein the anomaly persistence marker is used to distinguish between single fluctuations and continuous anomalies. The adhesion decay rate is used to represent the rate at which the bonding state at the detection location decreases between adjacent evaluation times, and environmental factor data is used to illustrate the humid corrosion conditions corresponding to this rate of decrease.

[0059] Before writing to the historical record, the system performs an integrity check on the data at the current evaluation time. If the detection location identifier is empty, the evaluation time is missing, or the adhesion decay rate cannot be calculated, the current record is not written to the historical degradation sequence, and a check mark is generated. If interface anomaly information exists but the adhesion decay rate is missing, the system retains the interface anomaly information and marks the evaluation time as a rate-missing record, so that subsequent clustering processing can identify that the record cannot be used for rate trend comparison. If the adhesion decay rate exists but the interface anomaly information is not triggered, the system still writes the record to the historical degradation sequence to identify situations where the anomaly judgment threshold has not been exceeded but there is a continuous decay trend. When environmental factor data is missing, the system can call records from nearby environmental sensors at the same evaluation time to supplement it; if it cannot be supplemented, the environmental factor field is set to a missing mark, and the participation weight of the record at that evaluation time is reduced in the clustering process.

[0060] The historical degradation sequence is arranged in ascending order of evaluation time. When multiple records exist at the same evaluation time, the system retains the record with higher data integrity or merges them according to the acquisition quality marker. The acquisition quality marker can be determined based on the integrity of impedance spectrum characteristics, the integrity of environmental factor data, and the synchronization error at the evaluation time. The system can also set a sequence update window, the length of which is set according to the coating service life, acquisition cycle, and the rate of change of the corrosion environment. For components with long-term stable service, the update window can cover a longer period; for scenarios with high humidity, high salt, or frequent wet-dry cycles, the update window is appropriately shortened so that the preset prediction rules can respond more quickly to recent degradation trends. The updated historical degradation sequence enters the clustering process as the basic data for judging the degradation mode and adjusting the prediction parameters.

[0061] The preset prediction rules are adjusted based on the clustering results of historical degradation sequences, including: clustering historical degradation sequences based on interface anomaly information, adhesion decay rate and environmental factor data to obtain clustering results; adjusting feature weights, risk thresholds and prediction time windows in the preset prediction rules based on the clustering results; and outputting adhesion degradation monitoring results, including: outputting adhesion degradation monitoring results under humid and corrosive scenarios based on the adjusted preset prediction rules.

[0062] In one embodiment, the clustering process of historical degradation sequences and the adjustment of preset prediction rules are further defined. The preset prediction rules are configured before system operation based on the coating material type, testing equipment accuracy, calibration sample library, and historical maintenance records. Feature weights are used to determine the participation ratio of impedance spectrum characteristics, environmental factor data, and offsets in risk assessment; initial values ​​can be determined based on the correlation between each input field in the calibration sample and the mechanical adhesion test results. Risk thresholds are used to determine whether the adhesion degradation risk reaches the warning condition; initial values ​​can be determined based on the allowable lower limit of the mechanical adhesion test results, the normal fluctuation range of the initial baseline state, and the repeatability error of the testing equipment. The prediction time window is used to limit the time length covered by the degradation trend prediction; initial values ​​can be determined based on the acquisition cycle, coating service life, and the rate of change of the corrosion environment.

[0063] The system extracts sequence features for clustering from historical degradation sequences. These features may include the mean offset, duration of offset increase, mean adhesion decay rate, peak adhesion decay rate, mean temperature, mean relative humidity, and mean conductivity of the corrosive medium. For historical degradation sequences of varying lengths, the system can align them according to the evaluation time or extract statistical features within a fixed time window. Sequences with significant missing evaluation times are not directly included in clustering; instead, they are categorized as needing supplementation to prevent incomplete sequences from affecting the clustering results.

[0064] The similarity between historical degradation sequences can be represented by a weighted distance, which can be calculated as follows: in, For the first The historical degradation sequence and the first Sequence distance between historical degenerate sequences For the first Offset statistics of historical degradation sequences For the first Offset statistics of historical degradation sequences For the first Statistical values ​​of adhesion decay rate of historical degradation sequences. For the first Statistical values ​​of adhesion decay rate of historical degradation sequences. For the first Environmental factor exposure values ​​for a historical degradation sequence For the first Environmental factor exposure values ​​for a historical degradation sequence For offset weights, As the adhesion decay rate weight, Environmental factor exposure weights are assigned. The offset statistics can be the mean or maximum value within the evaluation window, the adhesion decay rate statistics can be the mean or peak value within the evaluation window, and the environmental factor exposure values ​​can be determined based on the normalized statistical results of temperature, relative humidity, and corrosive medium conductivity within the evaluation window. Each weight is set based on calibration samples or historical maintenance records. When the adhesion decay rate is more sensitive to failure assessment, its weight is set to a larger value; when fluctuations in environmental factors significantly affect the degradation process, the environmental factor exposure weight is increased accordingly.

[0065] The system can use hierarchical clustering, cluster center iteration, or distance threshold-based grouping to obtain clustering results. The clustering results include at least three categories: slow decay, persistent anomalies, and environment-triggered acceleration. Slow decay corresponds to historical degradation sequences with low adhesion decay rates and slowly increasing offsets; persistent anomalies correspond to historical degradation sequences where offsets continuously exceed the anomaly threshold and the adhesion decay rate remains high; environment-triggered acceleration corresponds to historical degradation sequences where the adhesion decay rate increases significantly with rising relative humidity or the conductivity of the corrosive medium. Preset prediction rules are adjusted based on the clustering results. When a sequence belongs to the persistent anomaly category, the system lowers the risk threshold and shortens the prediction time window; when a sequence belongs to the environment-triggered acceleration category, the system increases the proportion of environmental factor data in the feature weights and sets the monitoring period to a shorter period; when a sequence belongs to the slow decay category, the system maintains a longer prediction time window and outputs monitoring results according to the normal cycle.

[0066] The adhesion degradation monitoring results can include the detection location identifier, current risk category, adhesion decay rate, triggered cluster category, prediction time window, and suggested retesting cycle. Before outputting the monitoring results, the system checks whether the current historical degradation sequence meets the minimum record count requirement. The minimum record count requirement is determined based on the number of continuous evaluation moments required for clustering processing, the acquisition cycle, and the prediction time window of the preset prediction rule. If the number of records is insufficient, the system outputs results pending further observation and does not directly output a high-risk conclusion. When the number of records meets the requirement and the cluster category is stable, the system outputs the adhesion degradation monitoring results under humid and corrosive scenarios based on the adjusted preset prediction rule, enabling the monitoring results to reflect the degradation state of the same detection location under the combined effects of impedance spectrum changes, adhesion decay, and environmental exposure.

[0067] like Figure 2 As shown, a coating adhesion evaluation system based on electrochemical impedance spectroscopy is used to implement a coating adhesion evaluation method based on electrochemical impedance spectroscopy. The system includes: The data acquisition module is used to acquire electrochemical impedance spectroscopy data and environmental factor data of the coating, forming interface state data including detection location identifiers. The initial reference state is determined by the interface state data of the initial acquisition period. The data acquisition module consists of an electrochemical impedance spectroscopy acquisition electrode, an impedance measurement circuit, an environmental sensor, an analog front-end circuit, an analog-to-digital converter, an acquisition controller, and a data buffer unit. It is used to acquire impedance spectrum magnitude, impedance spectrum phase angle, temperature, relative humidity, and conductivity of the corrosive medium at different detection locations, and bind the acquired data with the detection location identifiers.

[0068] The matrix construction module is used to extract impedance spectrum features from interface state data and construct an associated feature matrix based on impedance spectrum features, environmental factor data, detection location identifier, and initial reference state. The matrix construction module consists of a digital signal processor, a feature extraction processing unit, a data normalization unit, a reference state storage unit, and a matrix cache unit. It is used to extract impedance spectrum features from interface state data and write the impedance spectrum features, environmental factor data, detection location identifier, and initial reference state into the associated feature matrix.

[0069] The anomaly determination module is used to determine the interface state distribution based on the correlation feature matrix and to determine interface anomaly information based on the offset of the interface state distribution relative to the initial reference state. The anomaly determination module consists of a state evaluation processor, a threshold storage unit, a comparison and judgment circuit, and an anomaly recording unit. It is used to calculate the interface state distribution based on the correlation feature matrix and compare the interface state distribution with the initial reference state to obtain interface anomaly information.

[0070] The rate calculation module is used to determine the adhesion characterization value based on the correlation feature matrix and the adhesion calibration model, and to determine the adhesion decay rate based on the decrease in the adhesion characterization value and environmental factor data. The rate calculation module consists of a model operation processor, a calibration parameter storage unit, a timing calculation unit and an environmental correction calculation unit. It is used to call the adhesion calibration model to determine the adhesion characterization value, and to calculate the adhesion decay rate based on the decrease in the adhesion characterization value and environmental factor data.

[0071] The sequence prediction module is used to update historical degradation sequences based on interface anomaly information and adhesion decay rate, adjust preset prediction rules based on the clustering results of historical degradation sequences, and output adhesion degradation monitoring results. The sequence prediction module consists of a historical data storage unit, a clustering operation processor, a prediction rule storage unit, a rule update processing unit, and a result output interface.

[0072] It should be noted that, in this document, the terms "comprising," "including," and any other variations are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Specific examples have been used in this document to illustrate the principles and implementation methods of the technical solutions of this application. The above examples are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are merely preferred embodiments of this application. It should be pointed out that, due to the limitations of written expression and the objective existence of infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes, or combinations, or the direct application of the concept and technical solutions of this application to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A method for evaluating coating adhesion based on electrochemical impedance spectroscopy, characterized in that, The method includes: Electrochemical impedance spectroscopy data and environmental factor data of the coating are acquired to form interface state data including detection location identifiers. The initial reference state is determined by the interface state data of the initial acquisition period. Impedance spectrum features are extracted from the interface state data, and an associated feature matrix is ​​constructed based on the impedance spectrum features, the environmental factor data, the detection location identifier, and the initial reference state. The interface state distribution is determined based on the correlation feature matrix, and interface anomaly information is determined based on the offset of the interface state distribution relative to the initial reference state. The adhesion characterization value is determined based on the correlation feature matrix and the adhesion calibration model, and the adhesion decay rate is determined based on the decrease in the adhesion characterization value and the environmental factor data. The historical degradation sequence is updated based on the interface anomaly information and the adhesion decay rate. The preset prediction rule is adjusted based on the clustering results of the historical degradation sequence, and the adhesion degradation monitoring results are output.

2. The method according to claim 1, characterized in that, The process of acquiring electrochemical impedance spectroscopy data and environmental factor data of the coating to form interface state data including detection location markers includes: The impedance spectrum modulus and phase angle of the coating at multiple detection locations and multiple frequency points are collected at the time of collection, and the temperature, relative humidity and conductivity of the corrosive medium corresponding to the time of collection are obtained. The interface state data is obtained by associating the detection location identifier, acquisition time, impedance spectrum modulus, impedance spectrum phase angle, temperature, relative humidity, and conductivity of the corrosive medium corresponding to each detection location.

3. The method according to claim 2, characterized in that, The extraction of impedance spectrum features from the interface state data includes: The low-frequency impedance modulus is determined based on the impedance spectrum modulus, the phase angle change is determined based on the impedance spectrum phase angle at adjacent acquisition times and the same frequency points at the same detection location, and the slope of the impedance spectrum curve is determined based on the impedance spectrum modulus at adjacent frequency points at the same acquisition time at the same detection location. The low-frequency impedance modulus, the phase angle change, and the slope of the impedance spectrum curve are determined as the impedance spectrum characteristics.

4. The method according to claim 3, characterized in that, The construction of the correlation feature matrix based on the impedance spectrum characteristics, the environmental factor data, the detection location identifier, and the initial reference state includes: The detection location to which the impedance spectrum feature belongs is determined based on the detection location identifier, and the feature difference value corresponding to the impedance spectrum feature is determined based on the initial reference state. The impedance spectrum characteristics, environmental factor data, and feature differences corresponding to each detection location at the same acquisition time are written into the same matrix row of the correlation feature matrix to obtain the correlation feature matrix.

5. The method according to claim 2, characterized in that, The step of determining the interface state distribution based on the correlation feature matrix includes: determining the interface state evaluation value corresponding to each detection position based on the correlation feature matrix, and forming the interface state distribution based on the interface state evaluation value corresponding to each detection position. The step of determining interface anomaly information based on the offset of the interface state distribution relative to the initial reference state includes: Based on the initial reference state, determine the initial state evaluation value corresponding to each detection position, calculate the difference between the interface state evaluation value corresponding to each detection position and the initial state evaluation value, and obtain the offset; If the offset corresponding to any of the detection positions is greater than the anomaly determination threshold, the detection position is determined as an anomaly detection position, and interface anomaly information including the anomaly detection position is generated.

6. The method according to claim 1, characterized in that, The determination of adhesion characterization values ​​based on the correlation feature matrix and adhesion calibration model includes: Obtain the adhesion calibration model based on the calibration samples, each of the calibration samples including impedance spectrum sample data, environmental factor sample data and mechanical adhesion test results; The correlation feature matrix is ​​input into the adhesion calibration model to obtain the adhesion characterization value, which is used to characterize the bonding state between the coating and the substrate.

7. The method according to claim 6, characterized in that, The step of determining the adhesion decay rate based on the decrease in the adhesion characterization value and the environmental factor data includes: The basic attenuation rate is determined based on the decrease in the adhesion characterization value and the time interval corresponding to adjacent evaluation times; An environmental correction coefficient is determined based on the environmental factor data, and the base attenuation rate is corrected based on the environmental correction coefficient to obtain the adhesion attenuation rate.

8. The method according to claim 7, characterized in that, The step of updating the historical degradation sequence based on the interface anomaly information and the adhesion decay rate includes: Based on the detection location identifier, determine the historical records that match the detection location identifier; The interface anomaly information, the adhesion decay rate, and the environmental factor data corresponding to the evaluation time are written into the historical record according to the evaluation time to obtain the updated historical degradation sequence.

9. The method according to claim 8, characterized in that, The step of adjusting the preset prediction rules based on the clustering results of the historical degradation sequences includes: Based on the interface anomaly information, adhesion decay rate, and environmental factor data in the historical degradation sequence, the historical degradation sequence is clustered to obtain the clustering result; Adjust the feature weights, risk thresholds, and prediction time windows in the preset prediction rules based on the clustering results; The output adhesion degradation monitoring results include: outputting the adhesion degradation monitoring results under wet and corrosive scenarios based on the adjusted preset prediction rules.

10. A coating adhesion evaluation system based on electrochemical impedance spectroscopy, used to implement the coating adhesion evaluation method based on electrochemical impedance spectroscopy according to any one of claims 1 to 9, characterized in that, The system includes: The data acquisition module is used to acquire electrochemical impedance spectroscopy data and environmental factor data of the coating, and form interface state data including detection location identifiers. The initial reference state is determined by the interface state data during the initial acquisition period. The matrix construction module is used to extract impedance spectrum features from the interface state data and construct an associated feature matrix based on the impedance spectrum features, the environmental factor data, the detection location identifier, and the initial reference state. An anomaly determination module is used to determine the interface state distribution based on the associated feature matrix, and to determine interface anomaly information based on the offset of the interface state distribution relative to the initial reference state. The rate calculation module is used to determine the adhesion characterization value based on the correlation feature matrix and the adhesion calibration model, and to determine the adhesion decay rate based on the decrease in the adhesion characterization value and the environmental factor data. The sequence prediction module is used to update the historical degradation sequence based on the interface anomaly information and the adhesion decay rate, adjust the preset prediction rules based on the clustering results of the historical degradation sequence, and output the adhesion degradation monitoring results.

Images