A coating control system for photovoltaic glass

Abstract

This invention relates to the field of automated operation and maintenance and cleaning technology for photovoltaic power plants, specifically a coating control system for photovoltaic glass. It includes modules for data acquisition, feature decoupling, risk assessment, and dynamic control. The system acquires motor ripple and micro-vibration signals. Its core is the use of transient triboelectric harmonic analysis to separate the dirt friction component from the substrate friction component and calculate a real-time wear risk index. Based on the logical relationship between the components and the index, the system adaptively generates continuous pressure or micro-pulse suspension control commands. This invention achieves real-time identification of film surface properties without optical sensors, ensuring a precise balance between decontamination and film protection during cleaning operations.

Description

Technical Field

[0001] This invention relates to the field of automated operation and maintenance and cleaning technology for photovoltaic power plants, specifically a coating control system for photovoltaic glass. Background Technology

[0002] With the widespread application of photovoltaic power generation technology, improving the light transmittance of the module is the key to ensuring power generation efficiency. Modern photovoltaic glass modules are generally coated with sol-gel porous silica antireflective film, which has a fine nanoporous structure and low mechanical strength. Currently, photovoltaic operation and maintenance mainly relies on automated mechanical cleaning equipment, which typically employs constant contact pressure or control mechanisms based on macroscopic force feedback. However, traditional cleaning methods struggle to distinguish the tribological characteristics of surface deposits and the coating substrate at the microscopic level in real time. When faced with highly adhesive contaminants such as bird droppings, existing technologies often fall into a zero-sum game between cleanliness and coating lifespan: maintaining high-intensity scrubbing can easily lead to cumulative mechanical damage or even peeling of the fragile coating layer; reducing contact pressure, on the other hand, fails to remove stubborn contaminants, resulting in long-term degradation of the module's light transmittance. Therefore, how to achieve precise removal and adaptive control of strongly adhesive dirt through physical feature decoupling while ensuring the integrity of the nanoscale coating structure has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0003] To solve the above-mentioned technical problems, the present invention provides a coating control system for photovoltaic glass. Specifically, the technical solution of the present invention includes: The data acquisition module is configured to acquire the operating status signal of the work execution mechanism on the surface of the target coating substrate. The operating status signal includes: the high-frequency ripple component of the drive motor current and the micro-vibration acceleration signal of the work arm. The feature decoupling module is configured to process the operating status signal using transient triboelectric analysis, map the interaction of the contact interface into a feature vector of the tribological impedance space, and separate the dirt friction component corresponding to the surface deposits and the substrate friction component corresponding to the coating layer. The risk assessment module is configured to calculate a real-time wear risk index that characterizes the damage trend of the target coating substrate surface based on the substrate friction component. The dynamic control module is configured to generate control commands to drive the operating actuator to perform either continuous pressure mode or micro-pulse suspension mode based on the logical relationship between the dirt friction component and the real-time wear risk index.

[0004] Preferably, the feature decoupling module is configured to separate the dirt friction component and the substrate friction component in the following manner: Call upon high-frequency ripple components and micro-vibration acceleration signals; Frequency domain decomposition of high-frequency ripple components and micro-vibration acceleration signals is performed using wavelet packet transform algorithm; The preset dirt-coating separation matrix is ​​invoked. The separation matrix defines the mapping relationship between different frequency ranges and physical interaction objects. Based on the separation matrix, signals falling into the high-frequency band abrupt change feature range are extracted as dirt friction components, and signals falling into the low-frequency band fundamental wave feature range are extracted as substrate friction components.

[0005] Preferably, the risk assessment module is configured to calculate the real-time wear risk index in the following manner: Obtain the real-time amplitude of the substrate friction component and the preset damage weighting coefficient; Calculate the product of the real-time amplitude and the damage weighting coefficient; The product is integrated over time, and the result of the integration is determined as a real-time wear risk index to quantify the cumulative wear of the target coating substrate during the cleaning process.

[0006] Preferably, the dynamic control module is configured as follows: It calls up the dirt friction component, real-time wear risk index, preset dirt stripping threshold, and preset wear safety threshold; If the real-time wear risk index is greater than or equal to the wear safety threshold, a control command to execute the micropulse suspension mode is generated to trigger the protection mechanism. If the real-time wear risk index is less than the wear safety threshold and the dirt friction component is greater than the dirt stripping threshold, a control command is generated to execute the continuous pressure mode to carry out the operation in the safe intensive washing zone.

[0007] Preferably, when generating control commands to execute the micropulse levitation mode, the dynamic control module is configured as follows: Generate discontinuous pulse width modulation torque commands; The brush in the drive mechanism passes through the target coating substrate surface at a specific resonant frequency in a small amplitude. The frequency matching effect is used to break down the surface deposits, while the average pressure of the brush on the coating layer is maintained at a level lower than that of the continuous pressure mode.

[0008] Preferred options also include: The health mapping module is configured to record data on the changes of substrate friction components with location, generate a triboelectric impedance thermogram, and map the coating aging distribution state on the target coating substrate surface based on the triboelectric impedance thermogram to output targeted coating repair and maintenance suggestions.

[0009] Preferably, the data acquisition module includes: The Hall current sensor is configured to acquire the stator current change of the drive motor at a preset sampling rate and transmit the acquisition results to the digital signal processor for fast Fourier transform calculation. The work actuator includes a high-response servo motor or voice coil motor, configured to respond to control commands to achieve millisecond-level pressure adjustment.

[0010] Preferably, the target coating substrate is a photovoltaic glass module with a sol-gel porous silica antireflective film.

[0011] Compared with the prior art, the present invention has the following beneficial effects: 1. This system acquires the high-frequency ripple of the drive motor current and the micro-vibration signal of the working arm through the data acquisition module. Using the transient friction harmonic analysis method and wavelet packet transform algorithm in the feature decoupling module, the originally chaotic mechanical operation signal is mapped into a feature vector in the tribological impedance space. This technical solution can separate the dirt friction component, which exhibits high-frequency abrupt change characteristics, and the substrate friction component, which exhibits low-frequency fundamental wave characteristics, from the complex background noise. Thus, without relying on expensive optical sensors, it realizes the real-time identification of the properties of the deposits on the surface of the nanoscale porous silica antireflection film, providing a precise physical basis for subsequent differentiated control. 2. Unlike traditional solutions that only focus on instantaneous pressure thresholds, this system introduces a risk assessment module. By integrating the product of the substrate friction component and the damage weighting coefficient over time, a real-time wear risk index is calculated. This mechanism can quantify the history of cumulative shear stress borne by the coating layer during the cleaning process. It can promptly identify the risk of cumulative damage caused by excessive residence time or repeated friction even if the instantaneous force value is safe. This ensures that the residual thickness of the photovoltaic glass coating after cleaning always meets the optical performance requirements, thus guaranteeing the power generation efficiency of the module throughout its entire life cycle. 3. For stubborn dirt such as bird droppings, this system automatically switches to micro-pulse suspension mode when it detects an increased risk of wear, generating discontinuous pulse width modulation torque commands. This mode uses frequency matching effect to drive the brush to resonate, using high-frequency vibration energy to break down the dirt structure, while controlling the average pressure of the brush on the coating layer to maintain an extremely low level. This indirect approach can remove strongly adhering contaminants while avoiding effective physical shearing of the sol-gel coating with low mechanical strength, significantly extending the decay period of the photovoltaic module's light transmittance. 4. This system integrates a health mapping module, which can simultaneously record the changes in substrate friction components with spatial location while performing cleaning operations, and generate a visualized triboelectric impedance thermogram. This technical solution utilizes the significant difference in triboelectric impedance between aged or damaged coatings and intact coatings to directly map the aging distribution of the coating on the surface of photovoltaic modules. This allows power plant operation and maintenance personnel to obtain module health data during routine cleaning without deploying additional specialized testing equipment, and to formulate maintenance strategies such as targeted film repair based on this data, reducing operation and maintenance costs and improving maintenance accuracy. Attached Figure Description

[0012] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a structural diagram of the system of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0014] Example 1: Please see Figure 1 A coating control system for photovoltaic glass, comprising: The data acquisition module is configured to acquire the operating status signal of the work execution mechanism on the surface of the target coating substrate. The operating status signal includes: the high-frequency ripple component of the drive motor current and the micro-vibration acceleration signal of the work arm. The feature decoupling module is configured to process the operating status signal using transient triboelectric analysis, map the interaction of the contact interface into a feature vector of the tribological impedance space, and separate the dirt friction component corresponding to the surface deposits and the substrate friction component corresponding to the coating layer. The risk assessment module is configured to calculate a real-time wear risk index that characterizes the damage trend of the target coating substrate surface based on the substrate friction component. The dynamic control module is configured to generate control commands to drive the operating actuator to perform either continuous pressure mode or micro-pulse suspension mode based on the logical relationship between the dirt friction component and the real-time wear risk index.

[0015] This embodiment elaborates on the architecture logic and physical mapping mechanism of the above system, aiming to solve the deadlock contradiction between removing strongly adhering dirt and protecting the nanoscale antireflection coating in the existing photovoltaic cleaning technology. The system delves into the microscopic physical interaction level through the data acquisition module. The high-frequency ripple component of the drive motor current is not a macroscopic force feedback, but refers to the high-frequency harmonic superimposed on the fundamental wave in the motor stator current. This component carries the texture feature information between the brush and the contact surface. Together with the micro-vibration acceleration signal obtained by the accelerometer installed at the end of the working arm, they constitute a composite signal source reflecting the high-frequency mechanical vibration response of the contact interface. The feature decoupling module introduces transient triboelectric harmonic analysis, a signal processing method based on time-frequency analysis. This method aims to map the physical interactions of the contact interface into eigenvectors of the tribological impedance space, thereby separating the mixed signal in the orthogonal physical space into a dirt friction component exhibiting specific frequency abrupt changes and high impedance characteristics, and a substrate friction component exhibiting relatively stable fundamental wave characteristics. Based on this, the real-time wear risk index calculated by the risk assessment module is not an instantaneous pressure reading, but a quantitative indicator characterizing cumulative physical damage, equivalent to wear volume, used to characterize the integral assessment of the shear stress history of the coating layer. The dynamic control module constructs a coating stress threshold dynamic envelope controller, which switches between continuous pressure and micro-pulse suspension modes in milliseconds based on the decoupled physical quantities mentioned above. This embodiment digitizes the physical cleaning process into a triboelectric impedance signal, enabling microscopic differentiation of the physical properties of the object being cleaned. In the operation and maintenance scenario of photovoltaic power plants, this system breaks the zero-sum game between cleanliness and coating life, allowing the cleaning equipment to accurately remove strongly adhering dirt such as bird droppings without damaging the nanoscale porous structure, thereby significantly extending the decay cycle of the light transmittance of photovoltaic modules.

[0016] Example 2: The feature decoupling module is configured to separate the dirt friction component and the substrate friction component in the following manner: Call upon high-frequency ripple components and micro-vibration acceleration signals; Frequency domain decomposition of high-frequency ripple components and micro-vibration acceleration signals is performed using wavelet packet transform algorithm; The preset dirt-coating separation matrix is ​​invoked. The separation matrix defines the mapping relationship between different frequency ranges and physical interaction objects. Based on the separation matrix, signals falling into the high-frequency band abrupt change feature range are extracted as dirt friction components, and signals falling into the low-frequency band fundamental wave feature range are extracted as substrate friction components.

[0017] This embodiment is a further specification of the feature decoupling module separation logic in Embodiment 1, utilizing the frequency response differences of different material physical properties in the tribological impedance space for signal purification; the system calls the high-frequency ripple component. and micro-vibration acceleration signal As the original input, where, This represents the current ripple amplitude. The acceleration amplitude is used; the above signal is decomposed into multiple frequency domains using a wavelet packet transform algorithm. Compared with the traditional Fourier transform, this algorithm has higher time-frequency resolution, can finely divide the signal frequency band, and generate frequency band energy vectors. The system calls the preset dirt-coating separation matrix. This matrix is ​​essentially a pre-built classifier that defines the mapping relationship between high-frequency transient pulses corresponding to brittle dirt disintegration and low-frequency stable waveforms corresponding to flexible bristle slippage. Regarding matrices The specific construction method and threshold setting are based on the following: Technicians need to collect operational status signals of the actuator on both clean coated glass surfaces and typical contaminated surfaces in a laboratory environment beforehand; calculate the energy distribution of each frequency band using wavelet packet transform; and target the energy of the contaminated signal. This embodiment addresses the challenge of directly defining parameters in mixed signals by employing the differential energy extraction method to calculate the frequency bands within the background set. Energy upper limit threshold For each sample in the training set, calculate its total bandwidth energy. Define the overall fouling characteristic energy ratio. For each sample in the training set, calculate its , defined as the ratio of the sum of the energy exceeding the threshold in each frequency band to the total energy, as shown in the following formula: Fitting with Gaussian distribution Based on the statistical distribution in the training set, determine the confidence lower bound of dirt features and set the effective feature threshold. Set as ,in, for Mean of distribution for Standard deviation; if the threshold calculated for a certain frequency band This indicates that the frequency band can effectively distinguish between dirt and background noise with a 99.7% confidence level, so corresponding weights are set. Conversely, if it cannot be effectively distinguished, then set The resulting binarized or weighted matrix is ​​the preset separation matrix. In this calculation logic, the total number of frequency bands Depends on the number of wavelet packet decomposition levels ,follow The relationship; in this embodiment, the selected ,therefore ; Specifically, matrix The component calculation follows the formula below: in, For the separation matrix, For frequency band weighting, Total number of frequency bands; for The separation matrix; Indicates the first The weight of frequency bands on dirt characteristics This represents the weights of the basis features; , , The number of wavelet packet decomposition layers; And satisfy mutual exclusion constraints ; The first row of the weight vector corresponds to the dirt features, and the second row corresponds to the basis features; it should be noted that for any frequency band Weight and Satisfying the mutual exclusion constraint, i.e. Furthermore, the values ​​are strictly binary, 0 or 1, which respectively represent that the energy characteristics of this frequency band are completely attributed to dirt characteristics or completely attributed to substrate characteristics. Based on this matrix, the system performs component extraction, calculated using the following formula: and By introducing square root operations, the weighted energy value is restored to an equivalent voltage amplitude with linear physical meaning. This is to match the dimensional requirements of the subsequent mechanical model; among which, and The unit is volt (Volt). ), representing the equivalent signal amplitudes of the dirt friction component and the substrate friction component, respectively; Indicates the first The signal energy of the frequency band is calculated using the following formula: , The amplitude of the discrete signal within this frequency band, expressed in volts squared (V). This characterizes the signal power intensity of the corresponding frequency band; in, The equivalent signal amplitude of the dirt friction component is given. The equivalent signal amplitude of the substrate friction component is given. Indicates the first The energy of a frequency band is calculated as the sum of the squares of the signal amplitudes within that band, with the physical unit being the square of the volt (V). This characterizes the signal power intensity of the corresponding frequency band; Through the above calculations, the signal energy falling within the high-frequency abrupt change characteristic range is converted into an equivalent voltage amplitude as the dirt friction component. The signal energy falling into the low-frequency fundamental characteristic range is converted into an equivalent voltage amplitude as the substrate friction component. ; This embodiment solves the technical problem that the traditional current threshold method cannot distinguish the source of resistance by introducing wavelet packet transform and a separation matrix determined by statistical difference method. When facing complex field environments, this technology can accurately extract effective physical features from chaotic mechanical noise, ensuring that the system will not misjudge the resistance caused by dirt blockage as excessive pressure on the coating, thereby achieving accurate decoupling of the target object in the cleaning operation.

[0018] Example 3: The risk assessment module is configured to calculate the real-time wear risk index in the following way: Obtain the real-time amplitude of the substrate friction component and the preset damage weighting coefficient; Calculate the product of the real-time amplitude and the damage weighting coefficient; The product is integrated over time, and the result of the integration is determined as a real-time wear risk index to quantify the cumulative wear of the target coating substrate during the cleaning process.

[0019] This embodiment details the algorithm model of the risk assessment module based on the cumulative effect of microscopic wear; this module calculates the real-time wear risk index using the following integral formula. : in, The real-time equivalent signal amplitude of the substrate friction component, in units of: This value is determined in advance by the real-time friction force between the brush and the coating interface. Proportional ; The force sensitivity coefficient for electromechanical conversion, unit: This is used to map voltage signals into equivalent frictional forces. Let be the integral variable, representing the time elapsed from the start of the cleaning operation to the current moment; This represents the current real-time sampling moment; The constant scanning speed of the brush along the surface of the photovoltaic glass, in units of: ; This is the damage weighting coefficient, used to map the equivalent frictional work to the wear volume; in, This refers to the substrate friction component separated by the feature decoupling module, and its physical unit is voltage amplitude. Regarding the integral parameter, Defined as a time integral dummy variable; Represents the start time of a single cleaning operation cycle; Defined as the current real-time sampling moment, this serves as the upper limit of integration, establishing the calculation window for wear accumulation; for the parameters... It originates from a pre-set database and its physical meaning is a conversion factor that maps equivalent frictional work to wear volume. The damage weighting coefficient is pre-calculated based on the following calibration model: in, For standard wear rate, Porosity factor; This formula reflects the physical laws governing the relationship between wear rate and material properties; specifically, The standard specific wear rate is the standard wear rate measured by the standard wear test, and the unit is 1. This parameter characterizes the volume loss per unit frictional work. The porosity factor is calculated based on the coating refractive index, and the coating refractive index is measured using an elliptic polarization spectrometer. Then use the formula The calculation shows that, The effective refractive index of the coating is determined using an ellipsometric spectrometer at a standard wavelength. The values ​​obtained from the measurement; The refractive index of dense silicon dioxide is taken as a standard constant. ; To achieve Vickers hardness, a nanoindenter is used to indent at a specific depth, such as... The following measurements were obtained, for example This calculation method quantifies the differences in tolerance between coatings with different hardness and porosity; the system acquires data in real time. and The product of the two is calculated, which represents the instantaneous damage rate; this product is then subjected to a change in time with respect to the operation period. The integral operation converts the instantaneous frictional force into accumulated wear work, and finally outputs the result. To quantify the cumulative damage during the cleaning process; This embodiment introduces a time integration calculation mechanism to effectively address the problem of coating thickness reduction caused by soft wear. In automated cleaning scenarios, even if the instantaneous pressure does not exceed the damage limit, the system can identify the risk if the brush stays at a certain point for too long, causing the cumulative wear index to exceed the standard. This mechanism ensures that the residual thickness of the photovoltaic glass coating after cleaning always meets the optical performance requirements.

[0020] Example 4: The dynamic control module is configured as follows: It calls up the dirt friction component, real-time wear risk index, preset dirt stripping threshold, and preset wear safety threshold; If the real-time wear risk index is greater than or equal to the wear safety threshold, a control command to execute the micropulse suspension mode is generated to trigger the protection mechanism. If the real-time wear risk index is less than the wear safety threshold and the dirt friction component is greater than the dirt stripping threshold, a control command is generated to execute the continuous pressure mode to carry out the operation in the safe intensive washing zone.

[0021] This embodiment describes the decision-making logic of the dynamic control module based on multi-dimensional parameters, which constructs an automated safe operation window; the module receives the dirt friction component in real time. and real-time wear risk index And read the preset dirt removal threshold. and wear safety threshold ; Regarding key thresholds and Setting logic and determination method: Wear safety threshold Determination: Reverse calibration was performed based on the optical performance attenuation limit; samples of coated glass from the same batch were selected for destructive wear tests, and the transmittance was monitored simultaneously. When the transmittance decreases relative to the initial value achieve When the industry standard allows for the maximum attenuation, the accumulated wear risk index at this point is recorded. Set a safety threshold , Introduction Safety margin; for example, calibration in experiments. Then set Dirt removal threshold Determination: Based on the system noise floor level setting; the equipment is operated on a clean glass surface to collect data. Calculate the mean value of the dirt friction component samples from each cycle. and standard deviation ;set up To ensure a false trigger rate of less than one in a million; for example, the measured average noise floor. , Then set ; The system executes logical judgments and responds to... This indicates that the coating is nearing the critical damage point, and the system immediately generates a control command to execute the micro-pulse levitation mode, which has the highest priority to forcibly cut off the damage source; in response to and This indicates that the system is currently in a state of dirt but the coating is safe. The system generates a control command to execute the continuous pressure mode, applying sufficient physical shear force to quickly remove the dirt. This embodiment achieves a leap from experience-based control to quantitative closed-loop control. In actual operation, this logic ensures that the system no longer relies on the subjective judgment of the operator, but maximizes cleaning efficiency while protecting the coating. Especially when dealing with unevenly distributed dirt, it can dynamically adjust the operation strategy and avoid the risk of asset impairment caused by blindly strong cleaning.

[0022] Example 5: When generating control commands to execute the micropulse levitation mode, the dynamic control module is configured as follows: Generate discontinuous pulse width modulation torque commands; The brush in the drive mechanism passes through the target coating substrate surface at a specific resonant frequency in a small amplitude. The frequency matching effect is used to break down the surface deposits, while the average pressure of the brush on the coating layer is maintained at a level lower than that of the continuous pressure mode.

[0023] This embodiment details the specific technical implementation of the micropulse levitation mode, which is a core means of dealing with high-risk scenarios. When this mode is triggered, the dynamic control module generates a discontinuous pulse width modulation torque command. Unlike the constant output in the continuous mode, the motor receives a high-frequency pulse signal at this time. This command drives the brush in the work execution mechanism to operate at a specific resonant frequency. A small amplitude passes through the substrate surface at this frequency The selection aims to match the inherent frequencies or structural weakness frequencies of typical dirt such as bird droppings crystals, and its determination method is based on the following model: in, The target resonant frequency, in units of: ; The equivalent stiffness of the dirt is expressed in units of: The in-situ indentation test using a micro mechanical probe was used to determine the result. For equivalent mass of dirt, unit: Estimate based on the volume and density of the area in contact with the dirt; Regarding the specific values ​​and calibration methods for key parameters: For Type-A solidified bird droppings commonly found in photovoltaic power plants, in-situ indentation testing was conducted using a micro mechanical probe to measure its equivalent stiffness. The average value is The equivalent mass of the fouling that participates in the resonance Based on the volume density of the contact area, for a diameter ,thickness Typical dirt spots, calibrated quality for Substituting the above values ​​into the formula, we get: Therefore, the system sets the micropulse frequency to The nearby frequency sweep range; During this process, the system utilizes the frequency matching effect to induce mechanical resonance within the dirt, and controls the pulse duty cycle. This increases the average pressure of the brush on the coating layer. Maintained at a level significantly lower than that of the continuous pressure mode; in this model, the pulse peak force Defined as the maximum instantaneous normal force applied to the contact surface by the motor at the moment of pulse conduction, its value is calibrated in real time by the peak current fed back by the current loop; to achieve accurate calibration, this embodiment introduces the following electromechanical conversion formula: in, Represents peak pulse force, unit: ; Motor torque constant, unit: ; Peak current, unit: ; For transmission efficiency; This is the reduction ratio; The effective lever arm length of the boom, in units of: It should be noted that this formula is a static baseline model. In actual millisecond-level high-speed dynamic control, the controller can further introduce an inertial compensation term. Perform dynamic correction; It should be noted that this formula is based on the steady-state torque balance principle and serves as the benchmark model for micropulse control. In actual millisecond-level high-speed dynamic control, the controller can further introduce an inertial compensation term. Perform dynamic correction; According to the torque balance principle, the motor output torque is converted into an end force via a lever arm; therefore, the lever arm length is located in the denominator. Among these parameters, the voice coil motor is selected. To address the issue of dimensional inconsistency between pressure and force, this embodiment introduces the effective contact area between the brush and the substrate. The normalization calculation is performed using the following formula: in, The average pressure exerted by the brush on the coating, in units of: or ; For peak pulse force; The effective contact area between the brush and the coating surface, in units of: ; The duty cycle of the micropulse drive signal, with a value range of... The controller determines the risk based on the current cumulative damage value. Dynamically calculated to ensure average contact pressure It remains below the critical damage threshold; Experimental verification data: To verify the above technical effects, a period of [duration missing] was conducted at a photovoltaic power station in Northwest China. A comparative test was conducted over one month; the experimental group used the micro-pulse suspension mode of this system, while the control group used traditional constant force cleaning; the test results showed that the transmittance decay period of the control group components was shorter. Transmittance is defined as light transmittance. decline The average time required is The experimental group, under the same conditions, had a transmittance decay period of [day / day]. Extended to Furthermore, microscopic examination showed that the scratch density on the coated surface had decreased. This confirms the precise decontamination and low-damage characteristics of this mode due to the frequency matching effect. This embodiment utilizes the principle of resonant disintegration to replace traditional friction peeling. In extreme scenarios where stubborn dirt and fragile coatings coexist, this mode achieves a "hitting the bull from behind the mountain" effect, that is, it uses high-frequency vibration energy to pulverize dirt under extremely low average contact pressure. Thus, while removing the attached substances, it generates almost no effective shear force against the flexible coating, resolving the fundamental conflict between powerful decontamination and coating protection.

[0024] Example 6: Also includes: The health mapping module is configured to record data on the changes of substrate friction components with location, generate a triboelectric impedance thermogram, and map the coating aging distribution state on the target coating substrate surface based on the triboelectric impedance thermogram to output targeted coating repair and maintenance suggestions.

[0025] This embodiment expands the system's functionality by introducing a health mapping module to achieve digital surface inspection of components; during the cleaning process, the module simultaneously records the substrate friction components. Position coordinates of the work execution mechanism By combining spatial coordinates with the corresponding friction component amplitudes, a tribological impedance thermogram is generated. Since the roughened or peeled surface of the aged coating has a significantly different tribological impedance than the intact coating, the abnormal areas on the thermogram directly correspond to the damaged areas of the coating. Furthermore, the system maps the coating aging distribution state on the target coating substrate surface based on the thermogram and outputs targeted repair and maintenance suggestions based on this. This embodiment transforms the cleaning process into a full-body checkup; photovoltaic power plants do not need to deploy additional expensive optical inspection equipment, and can obtain health data of the component surface simply by using the routine cleaning process. This mechanism provides low-cost, high-frequency data support for the refined operation and maintenance of power plants, realizing the transformation from passive maintenance to predictive maintenance.

[0026] Example 7: The data acquisition module includes: The Hall current sensor is configured to acquire the stator current change of the drive motor at a preset sampling rate and transmit the acquisition results to the digital signal processor for fast Fourier transform calculation. The work actuator includes a high-response servo motor or voice coil motor, configured to respond to control commands to achieve millisecond-level pressure adjustment.

[0027] This embodiment clarifies the key hardware configuration required to implement the above algorithm; the data acquisition module uses a Hall current sensor and is configured to operate at a preset sampling rate, for example, greater than or equal to... The stator current change of the drive motor is collected, and its non-contact measurement characteristics and high frequency response capability are used to capture weak current ripples. The collected results are transmitted to the digital signal processor for fast Fourier transform calculation. The working actuator is selected from high-response servo motors or voice coil motors with extremely low electrical time constant and mechanical inertia, and configured to respond to control commands to achieve millisecond-level pressure adjustment. The hardware selection in this embodiment forms the physical basis for the algorithm's implementation; the high sampling rate sensor ensures that the ripple signal is not distorted, and the edge computing capability meets the real-time requirements, while the high-response motor ensures that the control system can keep up with the algorithm's decision speed, that is, complete the torque switching the instant that dirt appears, thereby ensuring the effective execution of the entire closed-loop system.

[0028] Example 8: The target coating substrate is a photovoltaic glass module with a sol-gel porous silica antireflective film.

[0029] This embodiment defines the specific application of this system; the target coating substrate is defined as a photovoltaic glass module with a sol-gel porous silica antireflective film. The coating is a nanoporous structure layer formed on the glass surface by a chemical wet process. Its characteristics are that it can significantly improve light transmittance, but its mechanical strength is much lower than that of the glass substrate, and it is easily damaged in traditional mechanical cleaning. This embodiment emphasizes the targeted nature of the technical solution; the aforementioned friction decoupling algorithm and micro-pulse suspension mode are designed specifically for the unique material properties of this porous, low-hardness, and high-optical-value material. For such high-value and fragile coated glass, this system provides the optimal solution under the current technical approach that balances cleaning efficiency and asset protection.

[0030] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A coating control system for photovoltaic glass, characterized in that, include: The data acquisition module is configured to acquire the operating status signal of the work execution mechanism on the surface of the target coating substrate. The operating status signal includes: the high-frequency ripple component of the drive motor current and the micro-vibration acceleration signal of the work arm. The feature decoupling module is configured to process the operating status signal using transient triboelectric analysis, map the interaction of the contact interface into a feature vector of the tribological impedance space, and separate the dirt friction component corresponding to the surface deposits and the substrate friction component corresponding to the coating layer. The risk assessment module is configured to calculate a real-time wear risk index that characterizes the damage trend of the target coating substrate surface based on the substrate friction component. The dynamic control module is configured to generate control commands to drive the operating actuator to perform either continuous pressure mode or micro-pulse suspension mode based on the logical relationship between the dirt friction component and the real-time wear risk index.

2. The coating control system for photovoltaic glass according to claim 1, characterized in that, The feature decoupling module is configured to separate the dirt friction component and the substrate friction component in the following manner: Call upon high-frequency ripple components and micro-vibration acceleration signals; Frequency domain decomposition of high-frequency ripple components and micro-vibration acceleration signals is performed using wavelet packet transform algorithm; The preset dirt-coating separation matrix is ​​invoked. The separation matrix defines the mapping relationship between different frequency ranges and physical interaction objects. Based on the separation matrix, signals falling into the high-frequency band abrupt change feature range are extracted as dirt friction components, and signals falling into the low-frequency band fundamental wave feature range are extracted as substrate friction components.

3. A coating control system for photovoltaic glass according to claim 2, characterized in that, The risk assessment module is configured to calculate the real-time wear risk index in the following manner: Obtain the real-time amplitude of the substrate friction component and the preset damage weighting coefficient; Calculate the product of the real-time amplitude and the damage weighting coefficient; The product is integrated over time, and the result of the integration is determined as a real-time wear risk index to quantify the cumulative wear of the target coating substrate during the cleaning process.

4. A coating control system for photovoltaic glass according to claim 3, characterized in that, The dynamic control module is configured as follows: It calls up the dirt friction component, real-time wear risk index, preset dirt stripping threshold, and preset wear safety threshold; If the real-time wear risk index is greater than or equal to the wear safety threshold, a control command to execute the micropulse suspension mode is generated to trigger the protection mechanism. If the real-time wear risk index is less than the wear safety threshold and the dirt friction component is greater than the dirt stripping threshold, a control command is generated to execute the continuous pressure mode to carry out the operation in the safe intensive washing zone.

5. A coating control system for photovoltaic glass according to claim 4, characterized in that, When generating control commands to execute the micropulse levitation mode, the dynamic control module is configured as follows: Generate discontinuous pulse width modulation torque commands; The brush in the drive mechanism passes through the target coating substrate surface at a specific resonant frequency in a small amplitude. The frequency matching effect is used to break down the surface deposits, while the average pressure of the brush on the coating layer is maintained at a level lower than that of the continuous pressure mode.

6. A coating control system for photovoltaic glass according to claim 1, characterized in that, Also includes: The health mapping module is configured to record data on the changes of substrate friction components with location, generate a triboelectric impedance thermogram, and map the coating aging distribution state on the target coating substrate surface based on the triboelectric impedance thermogram to output targeted coating repair and maintenance suggestions.

7. A coating control system for photovoltaic glass according to claim 1, characterized in that, The data acquisition module includes: The Hall current sensor is configured to acquire the stator current change of the drive motor at a preset sampling rate and transmit the acquisition results to the digital signal processor for fast Fourier transform calculation. The work actuator includes a high-response servo motor or voice coil motor, configured to respond to control commands to achieve millisecond-level pressure adjustment.

8. A coating control system for photovoltaic glass according to claim 1, characterized in that, The target coating substrate is a photovoltaic glass module with a sol-gel porous silica antireflective film.

Images