Insulation treatment method and system based on high-pressure-resistant coating type thermistor

Abstract

The invention relates to the technical field of thermistor insulation treatment, and discloses an insulation treatment method and system based on a high-pressure-resistant coating type thermistor. The method comprises the following steps: acquiring real-time resistance value data of a thermistor in a high-voltage environment, and generating a resistance dynamic response sequence; according to sequence amplitude change characteristics, insulation processing stages are divided, and resistance fluctuation threshold values of all the stages are marked; a resistance abnormal peak value in the stage is extracted, and a corresponding insulation compensation strategy is matched in combination with a preset coating material parameter library; and generating a dynamic coating control instruction based on the compensation strategy, and adjusting the insulation coating thickness distribution on the surface of the thermistor. According to the method, by capturing the dynamic change rule of the resistor, fine and dynamic adjustment of insulation treatment is achieved, the insulation reliability of the thermistor in the high-voltage environment can be improved, coating material waste is reduced, the equipment operation risk is reduced, and the method is suitable for high-voltage scenes such as an electric power system and new energy equipment.

Description

Technical Field

[0001] This invention relates to the field of thermistor insulation treatment technology, specifically to an insulation treatment method and system based on high-voltage coated thermistors. Background Technology

[0002] In power systems, new energy equipment, and industrial high-voltage detection devices, high-voltage coated thermistors serve as critical temperature monitoring and circuit protection components, and their insulation performance directly affects the stability and safety of the entire equipment operation. As the voltage requirements of related applications continue to increase, traditional insulation treatment methods for high-voltage coated thermistors are gradually revealing numerous shortcomings.

[0003] Currently, the industry commonly uses a fixed-thickness coating method for the insulation treatment of high-voltage coated thermistors. This involves pre-setting uniform insulation coating parameters based on experience to form a uniformly thick insulating coating on the thermistor surface. This method does not take into account the dynamic changes in the thermistor's resistance under actual high-voltage environments. When the thermistor is in a high-voltage environment with different voltage loads and different operating times, its internal temperature distribution will vary significantly, leading to non-uniform dynamic changes in resistance.

[0004] Traditional insulation treatment methods, lacking specific adjustments based on the dynamic resistance response sequence, fail to provide adequate insulation protection in areas with significant resistance fluctuations, while potentially leading to excessive material consumption in relatively stable resistance areas. Furthermore, traditional methods lack effective mechanisms for identifying and handling abnormal resistance peaks. In high-voltage environments, transient overvoltages or other anomalies can cause sudden, abnormal resistance peaks in the thermistors. Without timely matching of appropriate insulation compensation strategies, this can easily lead to insulation coating breakdown, causing equipment failure or even safety accidents.

[0005] Existing coating control methods mostly rely on static commands, which cannot adjust the coating thickness distribution based on the real-time dynamic resistance response of the thermistors. This results in a significant deviation between the insulation treatment effect and actual requirements. These problems not only reduce the service life and reliability of high-voltage coated thermistors but also increase equipment maintenance costs and operational risks, making it difficult to meet the high precision and adaptability requirements of current high-voltage equipment for thermistor insulation performance. Summary of the Invention

[0006] The purpose of this invention is to provide an insulation treatment method and system based on a high-voltage coated thermistor to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides an insulation treatment method based on a high-voltage coated thermistor, the method comprising:

[0008] Real-time resistance data of high-voltage coated thermistors are collected under high-voltage conditions to generate a dynamic resistance response sequence.

[0009] Based on the amplitude variation characteristics of the dynamic response sequence of resistance, the insulation treatment stages are divided and the resistance fluctuation threshold of each stage is marked.

[0010] Extract abnormal resistance peaks during the insulation treatment stage and match corresponding insulation compensation strategies based on a pre-set coating material parameter library.

[0011] Dynamic coating control commands are generated based on insulation compensation strategies to adjust the insulation coating thickness distribution on the surface of high-voltage coated thermistors.

[0012] Preferably, the specific steps for generating the dynamic response sequence of the resistance include:

[0013] The instantaneous resistance value of the high-voltage coated thermistor is obtained by a high-frequency sampling circuit, and the original resistance sequence is constructed in time order.

[0014] The original resistance sequence is normalized using a sliding window to eliminate the environmental temperature interference component.

[0015] The missing data points are filled in using piecewise linear interpolation to form a complete dynamic response sequence of the resistor.

[0016] Preferably, the specific steps for dividing the insulation treatment stages include:

[0017] Calculate the first-order absolute value of the dynamic response sequence of the resistor and identify the time boundary points of the differential abrupt change;

[0018] The dynamic response sequence of the resistor is divided into continuous sub-stages based on the timing boundary points;

[0019] The average resistance and standard deviation of each sub-stage are statistically analyzed, and sub-stages with an average resistance exceeding the benchmark value and a standard deviation greater than a preset threshold are marked as high-voltage sensitive stages.

[0020] Preferably, the specific steps of matching the corresponding insulation compensation strategy with the preset coating material parameter library include:

[0021] The abnormal resistance peak is compared with the breakdown voltage test data in the coating material parameter library to determine the material's withstand level.

[0022] Candidate coating materials are screened based on tolerance levels, and the optimal coating penetration depth is calculated in combination with the duration of the high-pressure sensitive stage.

[0023] Based on the optimal coating penetration depth, a ternary control parameter set including coating speed, spraying angle, and curing temperature is generated.

[0024] Preferably, the specific steps for adjusting the insulation coating thickness distribution on the surface of the high-voltage coated thermistor include:

[0025] Input the ternary control parameters into the multi-axis coating robot arm to drive the spray head to move along the surface of the thermistor in a variable trajectory.

[0026] Real-time monitoring of the change in dielectric constant of the coating layer, and feedback to correct the travel speed of the spray head;

[0027] When the dielectric constant reaches the target range, the infrared curing device is triggered to perform gradient heating treatment on the coating layer.

[0028] Preferably, the specific steps for providing feedback to correct the travel speed of the spray head include:

[0029] The real-time dielectric spectrum of the coating layer is collected by a dielectric sensor, and the amplitude attenuation rate of the characteristic frequency band is extracted.

[0030] The deviation between the amplitude attenuation rate and the preset safety threshold is compared to generate a speed adjustment coefficient;

[0031] The speed adjustment coefficient is superimposed on the painting speed component in the three-element control parameters to form a closed-loop control command.

[0032] Preferably, the specific steps of the infrared curing device in performing gradient temperature treatment on the coating layer include:

[0033] Based on the glass transition temperature curve of the coating material, three stages of heating range are divided;

[0034] Maintain a constant temperature difference within each heating stage range and control the power of the infrared radiator to increase linearly.

[0035] When cross-linking reaction characteristics appear on the surface of the coating layer, switch to the next heating zone until curing is complete.

[0036] Preferably, the method further includes the following steps:

[0037] After curing, the final insulation impedance value of the thermistor is collected and its consistency is verified with the high voltage test standard value.

[0038] If the consistency check fails, the abnormal resistance peak value of the high-voltage sensitive stage is re-extracted and the insulation compensation strategy is iteratively optimized.

[0039] Preferably, the specific steps for iteratively optimizing the insulation compensation strategy include:

[0040] Increase the number of candidate materials in the coating material parameter library to expand the range of tolerance level matching;

[0041] Adjust the weight distribution ratio of the ternary control parameters, and prioritize optimizing the synergy between spraying angle and curing temperature;

[0042] The optimized insulation compensation strategy is stored in the historical decision database for later use.

[0043] Preferably, the present invention also includes an insulation treatment system based on a high-voltage coated thermistor, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of the insulation treatment method based on the high-voltage coated thermistor described above.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] By collecting real-time resistance data of thermistors under high-voltage conditions and generating a dynamic resistance response sequence, the resistance change pattern of thermistors under different high-voltage conditions can be accurately captured. This breaks the limitation of traditional insulation treatment relying on experience-based preset parameters, making the insulation treatment process more in line with the actual working state of thermistors.

[0046] By dividing the insulation treatment stages according to the amplitude change characteristics of the dynamic response sequence of the resistance and marking the resistance fluctuation threshold of each stage, the insulation treatment process can be finely divided. This allows subsequent insulation compensation and coating adjustments to be carried out in response to the resistance fluctuation characteristics of different stages, avoiding the insulation insufficiency or material waste that may occur in different working stages in the traditional fixed coating method, making the insulation treatment more targeted and reasonable.

[0047] By extracting abnormal resistance peaks during the insulation treatment stage and matching corresponding insulation compensation strategies with a pre-set coating material parameter library, this mechanism can quickly identify potential anomalies in high-voltage environments and provide timely and suitable insulation protection solutions for areas with abnormal resistance. This proactive compensation mechanism effectively solves the problem of delayed handling of anomalies in traditional methods, reduces the risk of insulation coating breakdown caused by abnormal resistance peaks, and improves the anti-interference capability and operational stability of thermistors in complex high-voltage environments.

[0048] Dynamic coating control commands are generated based on an insulation compensation strategy to adjust the insulation coating thickness distribution on the thermistor surface, achieving dynamic and precise coating process. Compared to traditional static coating control, dynamic coating control can flexibly adjust the coating thickness according to the real-time dynamic resistance response of the thermistor. The coating thickness is increased in areas with large resistance fluctuations and high insulation requirements, and appropriately reduced in areas with stable resistance and lower insulation requirements. This ensures the insulation performance of critical areas while avoiding unnecessary consumption of coating materials, thus reducing production costs.

[0049] The entire method organically combines dynamic resistance monitoring, stage division, anomaly handling, and dynamic coating to form a complete closed-loop insulation treatment system. This system enables the insulation treatment effect to be optimized in real time as the thermistor's operating state changes, significantly improving the insulation reliability and service life of high-voltage coated thermistors. It also reduces the frequency and cost of subsequent equipment maintenance, providing a stronger guarantee for the stable operation of high-voltage equipment. It is applicable to a variety of high-voltage application scenarios and has broad practicality and promotional value. Attached Figure Description

[0050] Figure 1 This is a schematic diagram illustrating the working principle of the insulation treatment method based on a high-voltage coated thermistor described in this invention.

[0051] Figure 2 A flowchart for generating the dynamic response sequence of the resistor;

[0052] Figure 3 A comparison and analysis diagram of resistance characteristics during high-voltage insulation;

[0053] Figure 4 A flowchart for matching insulation compensation strategies. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Please see Figure 1This invention provides an insulation treatment method based on a high-voltage coated thermistor. The method includes: acquiring real-time resistance data of the thermistor under high-voltage conditions; using a high-frequency sampling circuit to obtain instantaneous resistance values ​​and constructing an original resistance sequence in chronological order; then performing sliding window normalization on the original sequence to eliminate interference components caused by ambient temperature fluctuations; and finally using piecewise linear interpolation to fill in any missing data points, thereby forming a complete dynamic response sequence of the resistance. Based on the amplitude variation characteristics of this sequence, the absolute value of its first-order difference is calculated to identify the temporal boundary points of differential abrupt changes. The sequence is divided into continuous sub-stages, and the mean and standard deviation of the resistance in each sub-stage are statistically analyzed. Sub-stages with a mean exceeding a benchmark value and a standard deviation greater than a preset threshold are marked as high-voltage sensitive stages. Simultaneously, a resistance fluctuation threshold is set for each stage. During the high-voltage sensitive phase, abnormal resistance peaks are extracted and compared with breakdown voltage test data in a pre-defined coating material parameter library to determine the material withstand level. Candidate coating materials are then selected based on the withstand level, and the optimal coating penetration depth is calculated by considering the duration of the high-voltage sensitive phase. A ternary control parameter set, including coating speed, spraying angle, and curing temperature, is generated as an insulation compensation strategy. This ternary control parameter set is input into a multi-axis coating robot arm, driving the spray head to move along the surface of the thermistor in a variable trajectory. The change in the dielectric constant of the coating layer is monitored in real time, and feedback is used to correct the spray head's travel speed. When the dielectric constant reaches the target range, an infrared curing device is triggered to perform gradient heating on the coating layer, thereby dynamically adjusting the insulation coating thickness distribution.

[0056] Example 1: See Figure 2 The process of generating a dynamic resistance response sequence is a complete technical process involving precise data acquisition, signal conditioning, and sequence reconstruction. The quality of this reconstruction directly determines the accuracy of subsequent insulation processing stage divisions and compensation strategy formulation. The process begins with capturing the instantaneous resistance value of a high-voltage coated thermistor using a high-frequency sampling circuit. This high-frequency sampling circuit typically employs a Σ-Δ analog-to-digital converter architecture to achieve high resolution and high anti-interference capability. The sampling frequency must be set much higher than the highest frequency component of the high-voltage environmental excitation change to satisfy the Nyquist sampling theorem and avoid spectral aliasing. In practical applications, the sampling frequency can be dynamically configured according to the expected resistance change rate; for example, the sampling rate can be temporarily increased during a voltage surge to capture transient response details. The connection between the sampling circuit and the thermistor uses a four-wire Kelvin connection to eliminate measurement errors introduced by lead resistance. Each sampling point is accompanied by a high-precision timestamp to ensure strict timing alignment. The original resistance sequence constructed in this chronological order serves as the raw data basis for all subsequent processing operations.

[0057] The original resistance sequence often contains various non-ideal components in its initial state, with interference caused by ambient temperature fluctuations being particularly significant because the intrinsic characteristics of thermistors make their output signal extremely sensitive to temperature changes. To eliminate this interference, a sliding window normalization process is needed for the original resistance sequence. Choosing the size of the sliding window is a trade-off; too small a window may lead to excessive smoothing and loss of the true high-voltage response characteristics, while too large a window may fail to effectively filter out low-frequency temperature drift. Typically, the window size can be set based on prior knowledge or preliminary data analysis to cover a number of data points spanning several temperature fluctuation cycles. Within each sliding window, the arithmetic mean and standard deviation of the resistance values ​​of the data points covered by the window are calculated. Then, the Z-score normalization algorithm is applied to subtract this local mean from each resistance value within the window and divide by the local standard deviation, thus transforming the data into standardized values ​​with a mean of zero and a standard deviation of one. This process effectively suppresses baseline drift caused by slow changes in ambient temperature, making the sequence more reflective of resistance changes directly induced by the high-voltage environment.

[0058] Due to the complexity of the high-voltage testing environment, data transmission links may experience momentary interruptions or be affected by strong electromagnetic pulses, resulting in sporadic or continuous missing data points in the original resistance sequence. If these missing points are not addressed, they will affect the continuity of the sequence and may cause subsequent analysis algorithms to fail. A piecewise linear interpolation method is used to fill these missing data points. This method requires accurately identifying the location of the missing point in the sequence and its adjacent valid data points. For a single missing point, the slope of the line connecting the two valid data points before and after it is calculated using the resistance values ​​and timestamps of the two points. Then, the corresponding resistance value is interpolated linearly based on the timestamp of the missing point. For the case of multiple consecutive missing data points, the consecutive missing segment is treated as a whole, with the first valid data point before and after it as the boundary. Linear interpolation is performed between these two points, and the resistance value increment is evenly distributed according to the time interval. This method assumes that the resistance change is approximately linear within a short time interval, which can maintain the basic rationality of the sequence change trend while ensuring filling efficiency. The sequence after missing value imputation still needs further verification of its physical rationality. This includes checking for non-physical jumps or abnormal smoothness introduced by interpolation. Sometimes, manual review or algorithmic judgment is required, taking into account the known characteristics of the high-voltage environment. The final dynamic resistance response sequence is a discrete-time sequence that is continuous in time, numerically normalized, and contains complete data points. This sequence is stored in array or time series data format, with each element containing a precise time index and the corresponding processed resistance value, providing clean and reliable input data for subsequent insulation processing stages. The entire construction process emphasizes the adaptability and robustness of the algorithm. For example, the sliding window size can be dynamically fine-tuned according to the sequence variance. A larger window is used during stable data periods to enhance smoothing, while a smaller window is automatically switched to preserve detailed features during periods of rapid data change. The interpolation algorithm may also introduce smoother methods such as spline interpolation as alternatives for handling boundary conditions to cope with special missing patterns. The output quality of the sequence construction stage directly affects the decision-making basis of the entire insulation processing system; therefore, each sub-step must be performed under strict error control and logical verification.

[0059] The implementation details of the high-frequency sampling circuit also include considerations of sampling timing and power consumption. A lower sampling rate can be used during non-critical periods to save system resources, while the highest sampling rate is immediately switched when a high-voltage start-up or resistance change rate exceeds a threshold. This adaptive sampling strategy relies on a real-time monitoring circuit. Sliding window normalization uses a recursive calculation method in actual computation to improve efficiency. When the window slides forward one point, it is not necessary to recalculate the statistics of the entire window. Instead, the statistical results of the previous window are combined with the values ​​of the newly added and removed points for incremental updates. This significantly reduces computational complexity, making it particularly suitable for long-term continuous monitoring scenarios. Piecewise linear interpolation also needs to consider boundary effects. For missing points at the beginning and end of the sequence, if there are not enough data points on one side, mirror expansion or simple repetition of boundary values ​​may be used for processing. Although this may introduce some error, it is usually within an acceptable range.

[0060] Example 2: The core of dividing the insulation treatment stage lies in the feature analysis and pattern recognition of the dynamic response sequence of resistance. The purpose is to decompose the continuous resistance change process into time intervals with different electrical characteristics. This process begins with a first-order difference operation on the sequence. The first-order difference is obtained by calculating the difference between the resistance values ​​of two adjacent data points in the sequence and taking the absolute value. This operation can effectively amplify the instantaneous rate of change of the resistance value, thus highlighting those critical points in the sequence where changes are drastic. Specifically, the system acquires the constructed dynamic response sequence of resistance, which is a complete data sequence after high-frequency sampling, sliding window normalization, and piecewise linear interpolation. The first-order difference operation generates a first-order difference absolute value sequence by traversing each data point in the sequence, calculating the difference between the resistance values ​​of two adjacent points, and taking the absolute value. For example, for the i-th point and the (i+1)-th point in the sequence, the operation is to calculate |R{i+1}-Ri|, where R represents the resistance value. This operation can effectively highlight the instantaneous rate of change of the resistance value because the magnitude of the difference value directly reflects the intensity of the resistance fluctuation per unit time. Subsequently, the system sets a dynamic threshold based on the statistical characteristics of the first-order difference sequence (such as the global mean and standard deviation), identifying points with difference values ​​significantly higher than the threshold as candidate abrupt change points. These points often correspond to critical events such as the application or removal of high-voltage conditions or changes in internal insulation status. To eliminate noise interference, the system also performs a neighborhood consistency check on the candidate points to ensure that abrupt change points form a continuous pattern within the local region. The finally confirmed time-series boundary points are used to divide the dynamic response sequence of the resistance into multiple sub-stages, providing a basis for marking subsequent high-voltage sensitive stages. This method amplifies the rate of change, ensuring that the division of insulation treatment stages is both accurate and reliable. The calculation process typically iterates through each data point of the sequence in a loop, generating a first-order difference absolute value sequence slightly shorter than the original sequence. This new sequence reflects the intensity of the change in the original resistance value within each sampling interval, and its peak points often correspond to the moments when critical events such as the application or removal of high-voltage conditions or minor breakdowns in the internal structure occur. Identifying temporal boundary points of differential mutations requires setting a reasonable threshold. This threshold is not fixed but depends on the overall statistical characteristics of the sequence. A dynamic thresholding method can typically be used based on the global distribution of the differential sequence; for example, the threshold can be set as the average of the differential sequence plus a certain multiple of the standard deviation. The first-order difference absolute value sequence is traversed, and points whose values ​​exceed the preset threshold are initially marked as candidate mutation points. However, due to signal noise or minor perturbations, these candidate points may contain spurious mutation points, requiring further verification and screening. The verification process usually employs a neighborhood consistency check, which examines whether the difference values ​​within a certain window before and after each candidate mutation point also maintain a high level. If only isolated spikes occur at individual points, they are likely caused by noise and should be eliminated. Only points that form continuous mutations within a local region are confirmed as valid temporal boundary points.These finally confirmed boundary points divide the resistive dynamic response sequence into multiple consecutive segments on the time axis.

[0061] Each continuous segment divided by boundary points is considered a sub-stage to be analyzed. Theoretically, each sub-stage should exhibit a relatively consistent resistance variation pattern, such as a stable period, a slow rise period, or a period of violent oscillation. For each sub-stage, the arithmetic mean of all resistance values ​​within it needs to be calculated to characterize the central resistance level of that stage, while its standard deviation is calculated to quantify the degree of resistance fluctuation. The mean reflects the average electrical stress level experienced by the thermistor during that time period, while the standard deviation reveals the stability of the resistance value. A larger standard deviation usually indicates that the resistance in that stage is more responsive to external high-voltage excitation or that an unstable insulation state exists. A baseline value is an important reference for marking high-voltage sensitive stages. This value is usually determined by long-term monitoring of the thermistor under stable conditions with no high voltage applied, or by combining the nominal resistance value of the thermistor model with historical normal operating data. The preset standard deviation threshold is an empirical or safety margin-based threshold used to determine whether the fluctuation is significant. The average resistance of a sub-stage is compared with a baseline value. If the average value is significantly higher than the baseline value, it indicates that the thermistor is under high load in that stage. Simultaneously, the standard deviation of that sub-stage is compared with a preset threshold. If the standard deviation is greater than the threshold, it means that the resistance fluctuates drastically. Only sub-stages that simultaneously meet both the conditions of an average value exceeding the baseline and a standard deviation exceeding the threshold are officially marked as high-voltage sensitive stages. These stages are key areas of concern due to weak insulation or potential breakdown, requiring special insulation compensation treatment.

[0062] The robustness of the stage partitioning algorithm is also reflected in its handling of boundary cases. For example, for short segments at the beginning and end of a sequence that are insufficient to form a complete analysis window, the algorithm will merge them with adjacent sub-stages or perform special processing based on their length and characteristics. The entire partitioning process can be offline or embedded in a real-time processing system. In real-time processing, a sliding window approach is used to continuously receive new resistance data and dynamically update the stage partitioning results. All information on the marked high-voltage sensitive stages, including their start time index, end time index, duration, average resistance within the stage, standard deviation, and maximum and minimum resistance values, is completely recorded to form a stage description list. This list serves as an important intermediate output, providing a clear time range and electrical characteristic basis for the next step of extracting abnormal resistance peaks and matching insulation compensation strategies. The accuracy of the partitioning needs to be continuously optimized through retrospective data verification, such as adjusting the parameters of the differential threshold or statistical criteria to make the partitioning results more consistent with the actual physical process of insulation state changes. The accuracy of temporal boundary point identification directly affects the accuracy of sub-stage segmentation. In practical applications, a multi-resolution analysis approach may be adopted, first using a large sliding window for coarse segmentation to identify major change stages, and then using a small window for fine boundary localization. The calculation of statistical features for sub-stages needs to consider the non-stationarity of the sequence. For longer sub-stages, internal segmentation statistics may be necessary to capture internal trend changes and prevent the features within long stages from being masked by averaging. The labeling criteria for high-voltage sensitive stages may not be absolute; sometimes fuzzy logic or the concept of confidence level are introduced to label sub-stages in a critical state and assign them a lower confidence weight for reference in subsequent processes. Metadata generated throughout the segmentation process, such as the threshold parameters used, the total number of segments, and the proportion of high-voltage sensitive stages, is also recorded by the system to assess the severity of the high-voltage test and the overall operating status of the thermistors. The execution efficiency of the stage segmentation algorithm is crucial for real-time applications; therefore, its implementation code is usually optimized, for example, by using incremental calculation of statistics and efficient data structures to store boundary point information to meet the system's real-time requirements. The intuitive visualization of the segmentation results is also part of the system, usually presented in the form of a time series graph. The graph clearly marks the boundaries of each sub-stage, especially the high-pressure sensitive stage, with different colors or markers, which facilitates intuitive understanding by operators and allows for manual intervention and correction when necessary.

[0063] See Figure 3In the insulation state analysis of high-voltage coated thermistors, the identification of the high-voltage sensitive stage relies on the comparison of the central tendency and distribution dispersion of the resistance values ​​at each stage. Specifically, the resistance characteristic comparison chart between the high-voltage sensitive stage and the non-sensitive stage uses a bar chart to display the average resistance value (dark gray bars) of each insulation stage, supplemented by error bars to indicate the resistance fluctuation range. The reference resistance value of 1000Ω is marked by a dashed line. The box plot of the resistance value distribution at each stage presents the distribution characteristics of the resistance values ​​at each stage. The box shows the quartile range, the whisker line indicates the normal value range, and the median line indicates the central tendency. The stage division covers high-voltage sensitive period 1, high-voltage sensitive period 2, stable period, slight fluctuation period, and recovery period. By comparing the deviation of the average resistance value of each stage from the reference value and the distribution width of the box plot, it is identified that the average resistance value of high-voltage sensitive period 1 rises to 1200Ω and the distribution dispersion is significant, indicating that the insulation response is active in this stage. In the parameter configuration, the reference resistance value is fixed at 1000Ω, the error bar is calculated based on the standard deviation of the resistance value at each stage, the box range of the box plot is defined by the 25th and 75th percentiles, and the outlier determination adopts the 1.5 times interquartile range rule.

[0064] Example 3; see Figure 4 Matching insulation compensation strategies and adjusting coating thickness distribution is a crucial process for translating electrical characteristics into physical actions. Its implementation relies on accurate interpretation of resistance anomaly signals and a deep understanding of coating materials and processes. Extracting abnormal resistance peaks within the insulation treatment stage is the first step. Peak detection algorithms are typically based on local extremum theory combined with the statistical characteristics of high-voltage sensitive stages. The algorithm traverses the resistance data segments marked as high-voltage sensitive stages, identifying data points that significantly deviate from the average resistance value of that stage. The degree of deviation is usually measured as a multiple of the standard deviation of the resistance in that stage; for example, a dynamic threshold is set, such as the mean plus three times the standard deviation. Points exceeding this threshold are initially identified as candidate abnormal peaks. To eliminate random noise interference, the algorithm also checks the neighborhood characteristics of candidate peak points, requiring a true abnormal peak to have a distinct convex shape rather than a single-point spike within a small time window before and after it. Finally, the confirmed abnormal peak points, their corresponding resistance values, and the time of occurrence are recorded to form an abnormal event list.

[0065] The core step in material selection is comparing the identified abnormal resistance peaks with the breakdown voltage test data in a pre-defined coating material parameter library. This library is a structured database where each record corresponds to a specific insulating coating material. Each record contains key performance parameters such as breakdown voltage, volume resistivity, dielectric constant, coefficient of thermal expansion, and aging characteristics under different temperatures and humidity levels. The comparison process involves converting the resistance value corresponding to each abnormal peak point into a voltage across its terminals using Ohm's law and the known test current. This voltage value is then compared one by one with the breakdown voltage test data for various materials in the parameter library (usually the average value minus a certain standard deviation to allow for a safety margin). The ratio of this voltage value to the material's breakdown voltage is calculated. Based on the ratio range, the material's tolerance level is divided into several discrete levels. For example, a ratio below 0.5 is defined as a high tolerance level, 0.5 to 0.8 as a medium level, and above 0.8 as a low level.

[0066] After selecting candidate coating materials based on their tolerance levels, the optimal coating penetration depth needs to be calculated by considering the duration of the high-voltage sensitive phase. Deeper penetration is not necessarily better; it needs to be sufficient to suppress abnormal discharges but not so thick as to affect the thermistor's response speed or cause mechanical stress problems. The calculation process considers the diffusion characteristics of the insulating material and the effective time of the high-voltage application, and their relationship can be expressed as:

[0067]

[0068] Where: d p The value represents the optimal coating penetration depth to be determined. It is a length-based quantity that determines the effective protective thickness the coating layer needs to achieve. The symbol D represents the diffusion coefficient of the selected candidate coating material. This coefficient is determined by the material's inherent properties and characterizes the ease with which the substance migrates within the insulating material. The symbol t... s This represents the total duration of the high-voltage sensitive stage, a value directly obtained from the stage division results. The proportionality coefficient k is a dimensionless empirical parameter whose value integrates various engineering factors such as electric field strength, safety margin, and interface bonding state.

[0069] Based on the calculated optimal coating penetration depth d pThe system needs to generate a set of executable ternary control parameters, specifically including coating speed, spray angle, and curing temperature. Coating speed primarily affects the amount of material deposited per unit time, and it has an inverse relationship with the target penetration depth. It is usually determined by finding a preset process curve or solving optimization equations. The spray angle determines the angle between the spray stream and the thermistor surface, affecting the coating's adhesion efficiency and contour shape, and is particularly important for uniform coverage of complex surfaces. The curing temperature directly relates to the reaction rate of the coating material's polymerization and cross-linking and the density of the final coating film. The temperature setting needs to refer to the technical data sheet provided by the material supplier and consider the balance between the substrate temperature and the curing oven temperature. Generating the ternary parameters is a multi-objective optimization process, requiring the simultaneous satisfaction of penetration depth requirements, coating efficiency, material utilization, and avoidance of defects such as sagging and orange peel.

[0070] The generated set of triplet control parameters is input into the control system of the multi-axis painting robot arm. The robot arm typically has six or more degrees of freedom, and its trajectory is pre-generated by a path planning algorithm based on the three-dimensional geometric model of the thermistor, or adaptively adjusted based on real-time sensor feedback. The spray head is precisely driven to move along the thermistor surface in a variable trajectory. In areas with a high probability of abnormal resistance peaks or weak areas identified based on historical failure analysis, the robot arm automatically reduces its travel speed and may use a smaller spray angle for localized reinforcement coating. In other areas, standard parameters are used to improve overall efficiency. Real-time monitoring is achieved through a dielectric constant sensor integrated near the spray head. This sensor measures the dielectric constant of the sprayed but uncured coating in a non-contact manner (e.g., using capacitance principles), and the measured value is fed back to the central controller. The central controller compares the real-time monitored dielectric constant with the target range determined based on the material and target penetration depth. If the monitored value deviates from the target range, for example, due to low dielectric constant caused by environmental humidity or batch differences in materials, the controller immediately calculates a correction amount. This correction amount is typically generated using a proportional-integral algorithm and converted into an adjustment command for the spray head's travel speed. The speed adjustment aims to change the coating amount per unit area, thereby indirectly controlling the coating thickness before drying and its corresponding dielectric properties. This feedback adjustment process continues, forming a rapid internal closed-loop control, striving to maintain the coating state within the target range during spraying. Once the system determines that the coating's dielectric constant has stably entered the target range and remained there for a predetermined time, it triggers a signal to start the next stage, the infrared curing device. The entire coating thickness adjustment process embodies a closed-loop automated control logic from electrical parameter sensing to physical action execution.

[0071] The selection process for candidate materials may not be unique; sometimes several alternatives are provided along with their respective advantages and disadvantages. For example, one material may have a high tolerance level but also a high cost, while another material may have a slightly lower tolerance level but a more mature process. The system may allow operators to make the final choice based on specific factors such as cost and delivery time. There are strong coupling relationships between the ternary control parameters. For instance, changes in the spraying angle may affect the effective coating speed, while the curing temperature setting needs to be matched with the coating speed to prevent the accumulation of uncured material. Therefore, the parameter generation algorithm needs to decouple these interrelationships or adopt a collaborative optimization strategy. The path planning of the multi-axis coating robot arm needs to avoid collisions with the workpiece fixture or other peripheral equipment. The path optimization algorithm needs to consider dynamic constraints such as acceleration and jerk to ensure the smoothness of the spraying process. Specifically, path planning is based on a 3D geometric model of the thermistor. An algorithm pre-generates the motion trajectory of the spray head, taking into account the curved shape of the thermistor and potential weak points. During planning, the system simulates the range of motion of each joint of the robotic arm and uses collision detection algorithms (such as bounding boxes or distance queries) to check the minimum distance between the spray head and the gripper in real time. Once a potential collision risk is detected, the trajectory is immediately adjusted to bypass the obstacle. Furthermore, the path optimization algorithm comprehensively considers dynamic constraints, including acceleration and jerk (rate of change of acceleration) control. For example, between trajectory points, the algorithm limits the maximum acceleration and jerk value of the robotic arm to avoid sudden starts / stops or jitter, ensuring smooth spray head movement. This contributes to uniform coating distribution and reduces material waste. In implementation, the system dynamically adjusts the path based on ternary control parameters (coating speed, spray angle, and curing temperature). Low-speed, small-angle spraying is used in high-pressure sensitive areas to enhance coverage, while the path is optimized in stable areas to improve efficiency. The entire process is corrected in real time through closed-loop control, ensuring that path planning is both safe and efficient. Real-time measurement of dielectric constant is susceptible to interference from environmental electromagnetic noise. Sensor signals usually need to undergo anti-interference processing such as filtering and shielding, and the measured values ​​need to be temperature compensated to improve accuracy.

[0072] Example 4: The specific implementation of feedback correction for spray head travel speed and infrared curing is embodied in a highly dynamic closed-loop control system. This system finely adjusts the coating and curing processes based on the real-time dielectric properties during the coating layer formation process. The implementation begins with continuous monitoring of the wet coating layer using a dielectric sensor. This sensor typically employs a parallel-plate capacitive design, with its plates placed parallel to the thermistor coating surface at a small distance in a non-contact manner. The sensor operates under a high-frequency alternating electric field to reduce the influence of ionic conductivity and primarily senses the polarization characteristics of the material. The sensor continuously acquires the dielectric spectrum of the coating layer at a preset high sampling rate. The spectrum covers a wide frequency range to capture the relaxation behavior of the material at different frequencies. The acquired raw spectrum data includes amplitude and phase information. Extracting the amplitude attenuation rate of characteristic frequency bands from the acquired real-time dielectric spectrum is the core analysis step. The selection of characteristic frequency bands is based on the coating material currently used. Different polymer-based insulating materials exhibit significant dielectric loss peaks within specific frequency ranges due to molecular chain segment movement or interfacial polarization. For example, for a certain epoxy resin-based coating material, its characteristic frequency band may be set between 1 MHz and 10 MHz. Within this frequency band, the system calculates the negative slope of the dielectric spectrum amplitude as a function of frequency, serving as a quantitative indicator of the amplitude attenuation rate. During the calculation, the system uniformly selects several frequency points within this characteristic frequency band and fits the logarithmic relationship curve of the amplitude versus frequency at these points. The slope of this curve is defined as the amplitude attenuation rate at the current moment. This parameter can sensitively reflect the degree of molecular orientation, solvent residue state, and possible microporous structures within the coating layer.

[0073] Comparing the calculated real-time amplitude attenuation rate with a preset safety threshold forms the basis for generating control commands. The safety threshold is not a fixed single value but a dynamic range. Its lower limit typically corresponds to a state where excessively thin coating layers or insufficient cross-linking may lead to decreased insulation performance, while its upper limit corresponds to the risk of internal stress cracking caused by excessively thick coating layers or rapid solvent evaporation. The comparison process calculates the deviation between the real-time attenuation rate and the midpoint of the threshold range. This deviation is input into a proportional-integral controller, which outputs a speed adjustment coefficient based on the magnitude of the deviation and the duration of accumulation. This coefficient is a dimensionless multiplier factor. The speed adjustment coefficient is directly added to the coating speed component in the currently executed triplet control parameters. If the attenuation rate deviates from the threshold range, it indicates that the current coating rate is inappropriate. For example, an excessively high attenuation rate may mean that too much material is deposited per unit time and the solvent is not easily evaporating, requiring the generation of a coefficient less than 1 to reduce the spray head travel speed; conversely, a coefficient greater than 1 is generated to increase the speed. The final step in the entire feedback loop is to generate a closed-loop control command and drive the robotic arm to execute it. The newly calculated coating speed is recombined with the original spraying angle and curing temperature parameters to form an updated triplet command, which is sent to the servo driver of the multi-axis coating robotic arm via real-time Ethernet or a similar industrial bus. The driver immediately adjusts the speed of the corresponding joint motor to change the scanning speed of the spray head relative to the thermistor surface. This adjustment is performed smoothly to avoid uneven coating caused by sudden speed changes. The entire speed adjustment cycle, from sensor acquisition, data analysis, deviation calculation to command issuance and execution, is completed in a very short time to achieve near real-time control.

[0074] Once the system detects that the dielectric constant of the coating layer and the amplitude decay rate of its derived parameters have stabilized within the target range for a preset stabilization time, it automatically triggers the infrared curing device to start the gradient heating program. The activation of the infrared curing device is not simply placing the coated part at a high temperature, but rather a carefully designed multi-stage heating process based on the glass transition temperature curve of the coating material. The glass transition temperature curve, provided by the material supplier or obtained through prior differential scanning calorimetry testing, describes the temperature range from the glassy state to the elastic state. Based on this curve, the entire curing process is divided into three main stages: the first stage is from ambient temperature to a safe temperature below the glass transition initiation temperature; the purpose of this stage is to slowly evaporate residual solvent and ensure uniform heating of the coating to reduce thermal stress. The second stage is from the end temperature of the first stage to a temperature above the glass transition region; in this stage, polymer segments begin to gain sufficient mobility for rearrangement and initial crosslinking. The third stage involves further heating to the final temperature required for complete material curing and a brief holding period to allow the crosslinking reaction to proceed fully. Within each defined heating zone, the control system maintains a constant rate of temperature change, meaning the temperature rises linearly over time. This is achieved by precisely controlling the linear increase of the infrared radiator's emission power. The power control module receives a setpoint from the temperature controller, calculated based on the target heating curve and real-time feedback of the furnace temperature (measured by multiple thermocouples). It adjusts the current or voltage applied to the infrared lamps or plate heaters to ensure the output power increases strictly linearly. Infrared pyrometers or thermocouples installed inside the furnace continuously monitor temperature changes on the coating surface and feed the data back to the controller to form a closed-loop temperature control system.

[0075] Monitoring the presence of cross-linking reaction characteristics on the coating surface is the basis for determining the stage switching. Cross-linking reactions are usually accompanied by small exothermic peaks, changes in surface gloss, or infrared absorption spectral characteristics. The system can determine the degree of reaction by monitoring the weakening or strengthening of specific chemical bond characteristic absorption peaks using an online infrared spectrometer. When the system detects and stabilizes a characteristic signal representing the main reaction of the current stage (such as the completion of solvent evaporation or the cross-linking degree reaching a certain preset level), it automatically switches the control parameters to the set values ​​of the next heating range. This process requires no manual intervention until all three heating stages are completed, ultimately obtaining a fully cured, high-quality insulating coating with controllable internal stress. Refer to Table 1, which shows the key parameter settings for each stage in a simplified infrared curing gradient heating process.

[0076] Table 1: Parameters of Infrared Curing Gradient Heating Process

[0077]

[0078] Calibration of dielectric sensors is crucial and is typically performed before the start of each production day using a standard template with a known dielectric constant to eliminate drift. The selection of characteristic frequency bands sometimes requires fine-tuning based on the specific batch of coating material used, as there may be subtle differences in additives or molecular weight distribution between different batches. The parameters of the proportional-integral controller need to be tuned according to the specific coating material, the dynamic characteristics of the robotic arm, and the sensor response time to avoid system oscillation or sluggish response. Maintaining temperature uniformity within the infrared curing oven is a challenge, and consistency in the thermal field distribution is usually ensured through multi-zone independent airflow control and optimized radiator layout. In addition to being based on characteristic response signals, the logic for stage switching also sets a maximum dwell time as a safety backup to prevent the system from indefinitely waiting at a certain stage due to sensor failure.

[0079] Example 5: The consistency verification and iterative optimization process constitutes a self-improvement mechanism for the insulation treatment method. Its purpose is to verify the effectiveness of a single treatment and continuously improve the overall process level based on the results. After the infrared curing device completes the gradient heating treatment and the coating layer cools sufficiently to room temperature, a high-precision impedance analyzer is used to collect the final insulation impedance value of the thermistor. The measurement process must be carried out under controlled environmental conditions to eliminate the influence of temperature and humidity fluctuations. The measurement frequency is usually selected from the industry standard recommended 1kHz or the frequency point specified in the specific product specification. The measurement voltage is set to a safe value far below the breakdown voltage of the coating layer to avoid damage to the finished product. The contact between the probe and the thermistor electrode must be good and highly repeatable. Each measurement usually involves multiple readings, and the average value is taken as the final insulation impedance value of the component. The consistency verification of the measured final insulation impedance value with the preset high-voltage test standard value is the basis for determining whether the treatment is qualified. The high-voltage test standard value comes from the product technical specifications of this type of thermistor or applicable international standards. This standard value is usually a minimum requirement. The verification method is not a simple scalar comparison, but rather employs a statistical test approach. It calculates the ratio of the measured impedance value to the standard value and examines whether this ratio falls within a tolerance range determined based on measurement uncertainty and process capability analysis. For example, the tolerance range might be set between 0.95 and 1.05 times the ratio. Simultaneously, the verification process may also incorporate confidence intervals established based on historical qualified batch data for auxiliary judgment. If the measured value is not only greater than the standard value but also falls within or exceeds the confidence interval of historical qualified data, it is judged as a high-quality pass. If the measured value is greater than the standard value but near the lower limit of the confidence interval, it is judged as a borderline pass, but still indicates a risk. If the measured value is lower than the standard value, the consistency verification is judged as a failure. When the consistency verification fails, the system initiates a diagnostic and replanning process, re-extracting all recorded abnormal resistance peaks of the component during the high-voltage sensitive phase from the stored data. This extraction may use different, more sensitive detection parameters than the initial processing, such as lowering the peak value judgment threshold to capture minor anomalies with smaller amplitudes but potentially cumulative effects. At the same time, the system will retrieve the log data of the entire processing process, including the boundaries of the stage division, the identification of the coating materials used, the specific values ​​of the ternary control parameters, and the temperature curve of the curing process, and analyze these data in conjunction with the results of failure.

[0080] The iterative optimization strategy for insulation compensation is based on failure analysis. The primary direction of optimization is to increase the variety of candidate materials in the coating material parameter library to expand the matching range of withstand levels. Expanding the material library is not simply about increasing the number of records, but rather about introducing new materials in a targeted manner based on the characteristics exhibited by the failure. For example, if the failure mode is early breakdown, a nano-filled composite material with higher dielectric strength might be introduced; if the failure mode is impedance reduction under humid and hot conditions, a fluorocarbon coating material with stronger hydrophobicity might be added. For each new material added, its complete performance parameters, especially breakdown voltage, volume resistivity, dielectric loss factor, and aging data under different environmental conditions, need to be entered into the database, and its correspondence with the withstand levels of various abnormal voltage peaks needs to be recalculated.

[0081] Another key aspect of the optimization process is adjusting the weight distribution of the ternary control parameters. In the initial strategy, the weights of the coating speed, spraying angle, and curing temperature may be set equally or based on general experience. The optimization algorithm analyzes the data processed and the verification results to determine which factor has a more significant impact on the final insulation performance, thereby adjusting the weight distribution. For example, if the analysis reveals a significant thin layer at the corners of the thermistor, the weight of the spraying angle parameter may be increased. This allows the optimization algorithm to focus more on optimizing the synergy between the spraying angle and curing temperature when recalculating parameters, ensuring uniform coverage of complex geometric surfaces. Weight adjustments are typically explored using gradient-based optimization algorithms or design of experiments. Storing the optimized insulation compensation strategy in a historical decision database is a crucial step in forming organizational knowledge assets. The stored content includes not only the new material selection list and the updated ternary control parameters and their weights, but also the original data background that led to this optimization, namely the dynamic resistance response sequence of the failed component, the identified high-voltage sensitive phase, abnormal peak information, and the results of failed consistency verification. The database is organized in a traceable structure. Each record contains a timestamp, component serial number, processing result, and strategy version number. When dealing with thermistors of similar models or with similar failure modes, the system can first query the database to retrieve strategy records that have successfully solved similar problems in the past as a reference for the initial strategy, thereby realizing case-based reasoning and learning.

[0082] The iterative process may not be a one-off event. If the newly processed parts still fail the consistency check after the initial optimization, the system will initiate a second iteration. This iteration may involve more fundamental parameter adjustments, such as modifying the algorithm parameters for stage division, redefining the boundary thresholds of the tolerance level, or even introducing a more complex coating path planning algorithm. Each iteration cycle generates new data records and stores them in the historical database, enabling the system's decision-making capabilities to gradually increase as the number of processed parts increases. As a result, the entire insulation processing method evolves from a fixed process flow into an intelligent system with adaptive and learning capabilities.

[0083] The cooling time before impedance measurement needs to be strictly defined, as residual heat can affect impedance readings. Typically, components are required to stand for a sufficient period in a standard laboratory environment to reach thermal equilibrium. The tolerance range for consistency verification needs to comprehensively consider the product's safety margin requirements, the accuracy of the measurement system, and production economics. An overly strict range can lead to unnecessary rework, while an overly lenient range may overlook potential defects. When re-extracting abnormal peaks, the system may perform more in-depth signal analysis on the resistance waveform during high-voltage sensitive periods, such as short-time Fourier transform, to check for the presence of discharge signals at specific frequencies. Maintaining the material parameter library is an ongoing process, requiring keeping pace with material suppliers' technology, updating material data promptly, or phasing out older materials whose performance no longer meets requirements. The algorithm for adjusting weight allocation ratios needs to have boundary constraints set to prevent the optimization process from getting bogged down in over-adjusting one parameter while neglecting the balance of other parameters. The historical decision database requires regular data maintenance and archiving, cleaning up invalid or outdated records, and building efficient indexes to support fast queries.

[0084] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof 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 process, method, article, or apparatus.

[0085] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An insulation treatment method based on a high-voltage coated thermistor, characterized in that, Includes the following steps: Real-time resistance data of high-voltage coated thermistors are collected under high-voltage conditions to generate a dynamic resistance response sequence. Based on the amplitude variation characteristics of the dynamic response sequence of resistance, the insulation treatment stages are divided and the resistance fluctuation threshold of each stage is marked. Extract abnormal resistance peaks during the insulation treatment stage and match corresponding insulation compensation strategies based on a pre-set coating material parameter library. Dynamic coating control commands are generated based on insulation compensation strategies to adjust the insulation coating thickness distribution on the surface of high-voltage coated thermistors.

2. The insulation treatment method based on a high-voltage coated thermistor according to claim 1, characterized in that, The specific steps for generating the dynamic response sequence of the resistor include: The instantaneous resistance value of the high-voltage coated thermistor is obtained by a high-frequency sampling circuit, and the original resistance sequence is constructed in time order. The original resistance sequence is normalized using a sliding window to eliminate the environmental temperature interference component. The missing data points are filled in using piecewise linear interpolation to form a complete dynamic response sequence of the resistor.

3. The insulation treatment method based on a high-voltage coated thermistor according to claim 2, characterized in that, The specific steps for dividing the insulation treatment stages include: Calculate the first-order absolute value of the dynamic response sequence of the resistor and identify the time boundary points of the differential abrupt change; The dynamic response sequence of the resistor is divided into continuous sub-stages based on the timing boundary points; The average resistance and standard deviation of each sub-stage are statistically analyzed, and sub-stages with an average resistance exceeding the benchmark value and a standard deviation greater than a preset threshold are marked as high-voltage sensitive stages.

4. The insulation treatment method based on a high-voltage coated thermistor according to claim 3, characterized in that, The specific steps for matching the corresponding insulation compensation strategy with the preset coating material parameter library include: The abnormal resistance peak is compared with the breakdown voltage test data in the coating material parameter library to determine the material's withstand level. Candidate coating materials are screened based on tolerance levels, and the optimal coating penetration depth is calculated in combination with the duration of the high-pressure sensitive stage. Based on the optimal coating penetration depth, a ternary control parameter set including coating speed, spraying angle, and curing temperature is generated.

5. The insulation treatment method based on a high-voltage coated thermistor according to claim 4, characterized in that, The specific steps for adjusting the insulation coating thickness distribution on the surface of the high-voltage coated thermistor include: Input the ternary control parameters into the multi-axis coating robot arm to drive the spray head to move along the surface of the thermistor in a variable trajectory. Real-time monitoring of the change in dielectric constant of the coating layer, and feedback to correct the travel speed of the spray head; When the dielectric constant reaches the target range, the infrared curing device is triggered to perform gradient heating treatment on the coating layer.

6. The insulation treatment method based on a high-voltage coated thermistor according to claim 5, characterized in that, The specific steps for providing feedback to correct the travel speed of the spray head include: The real-time dielectric spectrum of the coating layer is collected by a dielectric sensor, and the amplitude attenuation rate of the characteristic frequency band is extracted. The deviation between the amplitude attenuation rate and the preset safety threshold is compared to generate a speed adjustment coefficient; The speed adjustment coefficient is superimposed on the painting speed component in the three-element control parameters to form a closed-loop control command.

7. The insulation treatment method based on a high-voltage coated thermistor according to claim 6, characterized in that, The specific steps of the infrared curing device for gradient heating of the coating layer include: Based on the glass transition temperature curve of the coating material, three stages of heating range are divided; Maintain a constant temperature difference within each heating stage range and control the power of the infrared radiator to increase linearly. When cross-linking reaction characteristics appear on the surface of the coating layer, switch to the next heating zone until curing is complete.

8. The insulation treatment method based on a high-voltage coated thermistor according to claim 7, characterized in that, It also includes the following steps: After curing, the final insulation impedance value of the thermistor is collected and its consistency is verified with the high voltage test standard value. If the consistency check fails, the abnormal resistance peak value of the high-voltage sensitive stage is re-extracted and the insulation compensation strategy is iteratively optimized.

9. The insulation treatment method based on a high-voltage coated thermistor according to claim 8, characterized in that, The specific steps for iteratively optimizing the insulation compensation strategy include: Increase the number of candidate materials in the coating material parameter library to expand the range of tolerance level matching; Adjust the weight distribution ratio of the ternary control parameters and prioritize optimizing the synergy between spraying angle and curing temperature; The optimized insulation compensation strategy is stored in the historical decision database for later use.

10. An insulation treatment system based on a high-voltage coated thermistor, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the insulation treatment method based on high-voltage coated thermistors as described in any one of claims 1 to 9.

Images