Photovoltaic intelligent monitoring and management system based on Internet of Things

Abstract

The invention relates to the technical field of photovoltaic intelligent operation and maintenance, in particular to a photovoltaic intelligent monitoring and management system based on the Internet of Things, and the system comprises a data collection module which obtains the original data flow of a sensing array disposed on a photovoltaic panel; the quantitative evaluation module determines a composite fouling adhesion index and a chemical dissolution characteristic index; the efficiency prediction module determines a predicted cleaning success rate based on the composite fouling adhesion index, the chemical dissolution characteristic index, the real-time temperature, the real-time humidity and a preset cleaning strategy control variable; the strategy generation module solves an optimal cleaning strategy corresponding to the minimization of the resource cost function; the model correction module obtains the actual cleaning efficiency after the optimal cleaning strategy is executed, and corrects the efficiency prediction module according to the deviation between the actual cleaning efficiency and the predicted cleaning success rate; according to the invention, the core prediction engine connecting the fouling characteristics and the cleaning strategy is constructed, and the robustness and accuracy of the prediction model are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic intelligent operation and maintenance technology, specifically to a photovoltaic intelligent monitoring and management system based on the Internet of Things. Background Technology

[0002] In the field of photovoltaic power plant operation and maintenance, cleaning photovoltaic panels is a core aspect of ensuring power generation efficiency. To maintain power generation, the operation and maintenance team needs to regularly remove various types of dirt and grime adhering to the surface of the photovoltaic panels, a process that involves the consumption of various resources such as water, electricity, and chemical cleaning agents. Current photovoltaic (PV) panel cleaning strategies primarily rely on fixed time periods or manual experience, lacking real-time, quantitative perception of the specific physicochemical characteristics of panel surface contamination. This approach cannot make refined strategy adjustments based on dynamic changes in actual operating conditions such as contamination adhesion strength, chemical composition, and environmental temperature and humidity. Due to the lack of predictive models for cleaning scheme effectiveness and cost optimization functions for resource input, existing methods struggle to achieve a balance between cleaning effectiveness and resource consumption, often leading to over- or under-cleaning, resulting in significant waste of resources such as water, electricity, and cleaning agents, and reducing the overall economic efficiency of operation and maintenance. Furthermore, traditional cleaning models are open-loop management systems, unable to utilize historical cleaning data to provide feedback and correction for future decisions, lacking adaptive and self-optimizing capabilities. Therefore, how to construct an intelligent management system capable of multi-dimensional perception, online quantification, efficiency prediction, and closed-loop optimization to achieve optimal allocation of PV cleaning resources is a pressing technical problem to be solved in this field. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a photovoltaic intelligent monitoring and management system based on the Internet of Things. Specifically, the technical solution of this invention includes: The data acquisition module is used to acquire the raw data stream of the sensor array deployed on the photovoltaic panel. The raw data stream includes: the attenuation signal and phase shift signal of the acoustic impedance sensor, the multi-channel reflectivity vector of the multispectral sensor, and the real-time temperature and real-time humidity of the environmental sensor. The quantitative evaluation module is used to determine the composite fouling adhesion index and chemical solubility index based on the raw data stream and preset calibration parameters. The performance prediction module is used to determine the predicted cleaning success rate based on the composite dirt adhesion index, chemical solubility index, real-time temperature, real-time humidity, and preset cleaning strategy control variables. The strategy generation module is used to solve for the optimal cleaning strategy that minimizes the resource cost function based on the predicted cleaning success rate and the preset cleaning efficiency threshold. The model correction module is used to obtain the actual cleaning efficiency after executing the optimal cleaning strategy, and to correct the efficiency prediction module based on the deviation between the actual cleaning efficiency and the predicted cleaning success rate.

[0004] Preferably, the quantitative evaluation module is used to determine the composite fouling adhesion index, including: Call the attenuation signal and phase shift signal in the original data stream, as well as the preset reference attenuation signal, reference phase shift signal and weighting coefficient of the clean photovoltaic panel surface; Based on the empirical model of acoustic impedance physics, the relative changes of the attenuation signal and the reference attenuation signal, as well as the relative changes of the phase shift signal and the reference phase shift signal, are weighted and summed. The weighted summation result was determined as the composite fouling adhesion index.

[0005] Preferably, the quantitative evaluation module is used to determine the chemical solubility characteristic index, including: Call the multi-channel reflectance vector in the original data stream, as well as the preset set of spectral channel indexes characterizing the absorption peaks of organic matter, the set of spectral channel indexes characterizing the absorption peaks of inorganic matter, and the sensitivity weighting coefficients; Based on an empirical proportional model, the sum of weighted reflectances representing organic matter in the multi-channel reflectance vectors is calculated separately. And the sum of weighted reflectance characterizing inorganic materials; The chemical solubility index is obtained by calculating the ratio of the sum of weighted reflectances characterizing organic matter to the sum of weighted reflectances characterizing inorganic matter.

[0006] Preferably, the performance prediction module is used to determine the predicted cleaning success rate, including: The fouling resistance index is determined based on the adhesion index, real-time temperature, real-time humidity, and preset model coefficients. The cleaning rate index is determined based on the chemical solubility characteristic index, cleaning strategy control variables, and preset rate coefficients. By combining the fouling resistance index, cleaning rate index, the duration of action in the cleaning strategy control variables, and the preset gain coefficient, the predicted cleaning success rate is determined based on the logistic growth model.

[0007] Preferably, the performance prediction module is used to determine the fouling resistance index, including: Call the adhesion index, real-time humidity, real-time temperature, and preset humidity normalization baseline value, temperature normalization baseline value and model coefficients; Based on an empirical linear regression model, the adhesion index, the ratio of real-time humidity to the humidity normalized baseline value, and the ratio of real-time temperature to the temperature normalized baseline value are linearly combined. The result of the linear combination was determined as the fouling resistance index.

[0008] Preferably, the efficiency prediction module is used to determine the cleaning rate index, including: The chemical solubility index and cleaning strategy control variables are used. The cleaning strategy control variables include: water pressure, cleaning solution temperature, cleaning agent A concentration and cleaning agent B concentration. It also calls the preset pressure normalization reference value, temperature normalization reference value, physical cleaning rate coefficient, and chemical enhancement coefficient; Based on the physical cleaning rate coefficient, the ratio of water pressure to the normalized pressure reference value and the ratio of cleaning fluid temperature to the normalized temperature reference value are weighted and summed to obtain the physical cleaning rate term. Based on the chemical solubility index, the concentration of cleaning agent A, the concentration of cleaning agent B, and the chemical enhancement coefficient, a chemical enhancement factor is determined. The chemical enhancement factor is used to dynamically adjust the effectiveness of the concentrations of cleaning agent A and cleaning agent B. The cleaning rate index is determined by multiplying the physical cleaning rate term by the chemical enhancement factor.

[0009] Preferably, the strategy generation module is used to solve for the optimal cleaning strategy, including: Invoke cleaning strategy control variables, real-time temperature, and preset cost coefficients; Based on the linear cost accounting model, a resource cost function is constructed, which includes energy consumption cost and cleaning agent cost. Construct a constrained optimization objective, which aims to minimize the resource cost function and is constrained by the condition that the predicted cleaning success rate is not lower than the cleaning efficiency threshold. Solve the constraint optimization objective to determine the optimal cleaning strategy.

[0010] Preferably, the energy cost in the resource cost function includes the energy cost per unit temperature difference. Among them, when the cleaning fluid temperature is higher than the real-time temperature, the unit temperature difference energy consumption cost is calculated based on the difference between the cleaning fluid temperature and the real-time temperature. When the temperature of the cleaning fluid is no higher than the real-time temperature, the energy consumption cost per unit temperature difference is zero.

[0011] Preferably, the model correction module is used to correct the performance prediction module, including: The actual cleaning efficiency and the predicted cleaning success rate are compared. Calculate the deviation between actual cleaning efficiency and predicted cleaning success rate, and determine the loss function based on the deviation; Based on the gradient descent update rule, the gradient of the loss function with respect to the preset model coefficients, rate coefficients, and gain coefficients in the performance prediction module is calculated. Update the model coefficients, rate coefficients, and gain coefficients based on the gradient and the preset learning rate.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. Online, multidimensional quantification of fouling characteristics is achieved. This system acquires acoustic impedance and multispectral data through a sensor array. On the one hand, based on the acoustic impedance physical model, the relative changes in attenuation and phase shift signals are weighted and summed to obtain an adhesion index that can assess the adhesion strength. On the other hand, based on an empirical proportional model, the chemical solubility index is obtained by calculating the ratio of the sum of weighted reflectance representing organic and inorganic matter. This reduces the complexity of the original data to a key, quantifiable indicator that guides subsequent prediction and decision-making, thus achieving accurate perception of the physical and chemical characteristics of fouling. 2. A core prediction engine connecting soiling characteristics and cleaning strategies was constructed. This system innovatively adopts the logistic growth model to predict cleaning success rate. This model is driven by two core indices: a soiling resistance index that dynamically reflects environmental impact is constructed by linearly combining the adhesion index and real-time temperature and humidity; and a cleaning rate index that intelligently integrates water pressure, temperature, and the effectiveness of targeted cleaning agents is constructed by multiplying the physical cleaning rate term and the chemical enhancement factor. This design accurately fits the nonlinear cleaning process and greatly improves the robustness and accuracy of the prediction model. 3. Constrained optimization and precise cost control of cleaning decisions have been achieved. This system transforms the generation of cleaning strategies into an optimization problem with the goal of minimizing the resource cost function and the constraint that the predicted cleaning success rate is not lower than the efficiency threshold. This cost function is based on a linear model, accurately calculates the cost of cleaning agents and energy consumption, and makes a non-negative correction for energy consumption costs. Billing is based on temperature difference only when the temperature of the cleaning liquid is higher than the ambient temperature. This ensures that the system can always automatically solve for the most economical strategy combination while ensuring the cleaning effect. 4. An adaptive correction closed loop based on actual effect feedback was established. This system possesses online learning capabilities. After the cleaning strategy is implemented, the loss function is determined by calculating the deviation between the actual effectiveness and the predicted success rate. The system updates based on gradient descent rules, calculating the gradient of the loss function with respect to the preset model, rate, and gain coefficients in the effectiveness prediction module, and iteratively updating it in conjunction with the learning rate. This constructs a complete machine learning closed loop, enabling the system to continuously self-optimize and ensuring that its prediction model maintains high long-term accuracy and adaptability when facing new contamination or environmental changes. Attached Figure Description

[0013] 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

[0014] 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.

[0015] Example 1 Please see Figure 1 A photovoltaic intelligent monitoring and management system based on the Internet of Things includes: The data acquisition module is used to acquire the raw data stream of the sensor array deployed on the photovoltaic panel. The raw data stream includes: the attenuation signal and phase shift signal of the acoustic impedance sensor, the multi-channel reflectivity vector of the multispectral sensor, and the real-time temperature and real-time humidity of the environmental sensor. The quantitative evaluation module is used to determine the composite fouling adhesion index and chemical solubility index based on the raw data stream and preset calibration parameters. The performance prediction module is used to determine the predicted cleaning success rate based on the composite dirt adhesion index, chemical solubility index, real-time temperature, real-time humidity, and preset cleaning strategy control variables. The strategy generation module is used to solve for the optimal cleaning strategy that minimizes the resource cost function based on the predicted cleaning success rate and the preset cleaning efficiency threshold. The model correction module is used to obtain the actual cleaning efficiency after executing the optimal cleaning strategy, and to correct the efficiency prediction module based on the deviation between the actual cleaning efficiency and the predicted cleaning success rate.

[0016] This embodiment provides a photovoltaic intelligent monitoring and management system based on the Internet of Things. The system aims to solve the technical problems of existing photovoltaic cleaning strategies relying on manual experience or fixed cycles, resulting in resource waste and low cleaning efficiency. This system achieves precise management of photovoltaic panel contamination and optimal allocation of resources by constructing an intelligent closed loop from multi-dimensional perception, online quantification, efficiency prediction to closed-loop optimization. The data acquisition module aims to acquire raw contamination status data of the photovoltaic panel surface in real time and from multiple dimensions, providing raw input for subsequent quantitative evaluation. In this embodiment, this module is implemented through a sensor array deployed at key locations on the photovoltaic panel. This sensor array specifically includes: Acoustic impedance sensor: used to acquire attenuation signals characterizing the physical adhesion properties of fouling layers. and phase shift signal The attenuation signal and phase shift signal refer to the physical changes that occur when the sound wave passes through the interface between the fouling layer and the photovoltaic panel. The degree of these changes is highly correlated with the physical adhesion strength and thickness of the fouling layer. Multispectral sensor: used to acquire multichannel reflectance vectors characterizing the chemical composition of fouling layers. The multichannel reflectance vector refers to the reflectance readings of the fouling layer in multiple specific spectral channels. By analyzing the reflectance of specific absorption peaks, the proportion of its chemical composition can be inferred. Environmental sensor: Used to obtain the real-time temperature of the environment where the photovoltaic panel is located. With real-time humidity These two parameters will affect the adhesion characteristics of the dirt and the chemical reaction rate of the cleaning process; The quantitative evaluation module aims to transform the heterogeneous, raw data stream collected by the sensor array into a structured fouling characteristic index that can be used for engineering decision-making. Based on the raw data stream and preset calibration parameters, such as the baseline signal of the clean surface and weighting coefficients, this module decouples and quantifies the physical and chemical characteristics of fouling, ultimately determining two core indicators: Composite Fouling Adhesion Index The composite fouling adhesion index is a dimensionless parameter used to uniformly measure the physical adhesion strength between the fouling layer and the surface of the photovoltaic panel. The higher the value, the more stubborn the stain, and the more difficult it is to remove physically. Chemical solubility index The chemical solubility index is a dimensionless parameter used to characterize the relative proportion of organic and inorganic matter in fouling. The value will directly guide the subsequent mixing ratio of chemical cleaning agents; The performance prediction module aims to simulate the removal effect of different cleaning strategies on specific soiling before the actual cleaning action is performed. This module is a core prediction engine, and its inputs include: 1. The composite soiling adhesion index determined by the quantitative evaluation module. Chemical solubility index 2. Real-time temperature acquired by the data acquisition module and real-time humidity ; and 3. Pre-defined cleaning strategy control variables to be evaluated; cleaning strategy control variables This refers to the combination of parameters that describe a cleaning action, such as water pressure. Cleaning solution temperature Concentration of cleaning agents A / B and duration of action Based on the above inputs, this module determines the predicted cleaning success rate. Predicted cleaning success rate It refers to a probability value between 0 and 1, representing a specific strategy. In the current environment Remove specific stains The expected results; The strategy generation module aims to find the optimal solution that balances cleaning effectiveness and resource cost from a vast number of possible cleaning strategy combinations. This module constructs a constrained optimization problem based on the predicted cleaning success rate output by the efficiency prediction module. and preset cleaning efficiency threshold Construct constraints and a resource cost function. Cleaning efficiency threshold This refers to the minimum acceptable cleanliness standards pre-set by the operations and maintenance team based on actual operating conditions or power generation contract requirements. For example, This module aims to minimize the resource cost function, and uses... Given the constraints, solve the optimization problem to determine the optimal cleaning strategy. ; The model correction module aims to build a closed-loop feedback mechanism, enabling the system to adapt to environmental changes, variations in fouling types, or deviations in initial model parameters, continuously improving the accuracy of the prediction model; and ensuring the system executes the optimal cleaning strategy. Then, the module obtains the actual cleaning efficiency after executing the optimal cleaning strategy. For example, the efficiency can be assessed using a light transmittance sensor or power generation reading of the photovoltaic panels after cleaning; subsequently, the module calculates the actual cleaning efficiency. Compared with the predicted cleaning success rate Deviation between Based on this deviation, the module uses online learning algorithms in machine learning, such as gradient descent, to correct the internal model coefficients in the performance prediction module, so that the model is closer to the actual situation in the next prediction. This embodiment constructs a complete intelligent monitoring and management closed loop, from pollution perception to characteristic quantification, strategy prediction, optimal decision-making, and closed-loop correction. Compared with traditional cleaning methods that rely on experience or fixed cycles, this system can dynamically generate targeted cleaning strategies with the lowest resource consumption based on the real-time, in-situ physicochemical characteristics of photovoltaic panels. Its significant technical effect is that, while ensuring that the power generation efficiency is restored to the standard, it greatly reduces the consumption of water, electric heating, pressurization, and chemical cleaning agents, realizing refined, automated, and resource-optimized photovoltaic operation and maintenance.

[0017] Example 2 The quantitative assessment module is used to determine the composite fouling adhesion index, including: Call the attenuation signal and phase shift signal in the original data stream, as well as the preset reference attenuation signal, reference phase shift signal and weighting coefficient of the clean photovoltaic panel surface; Based on the empirical model of acoustic impedance physics, the relative changes of the attenuation signal and the reference attenuation signal, as well as the relative changes of the phase shift signal and the reference phase shift signal, are weighted and summed. The weighted summation result was determined as the composite fouling adhesion index.

[0018] This embodiment specifically defines the determination of the composite fouling adhesion index in the quantitative evaluation module. This method aims to provide an accurate, online, non-destructive method for quantifying dirt adhesion. In calculation At that time, the quantization evaluation module calls the attenuated signal from the original data stream. and phase shift signal Simultaneously, the module retrieves a preset reference attenuation signal from the clean photovoltaic panel surface from the system configuration. Reference phase shift signal and weighting coefficients ; Reference attenuation signal and reference phase shift signal This refers to the acoustic impedance sensor reading obtained when calibrated on a clean photovoltaic panel surface, i.e., without any contamination. Its source is the reference value that was measured and stored in advance during the system initialization or laboratory calibration phase. Weighting coefficient This refers to a dimensionless parameter used to balance the contributions of the attenuation signal and phase shift signal to the total adhesion; its source is determined through offline calibration experiments; specifically, the preparation... Groups with different rated adhesion forces are denoted as ,in The contaminated samples were measured using a standard pull-out test, and their corresponding calibrated attenuation signals were measured respectively. and calibration phase shift signal Based on this Group The data is fitted using multiple regression analysis or least squares method to find the solution that makes the following formula calculated. and Weighting coefficients that exhibit the best linear correlation and And satisfy ; The module performs calculations based on an empirical model that combines the physical principles of acoustic impedance with calibration data fitting; the presence of a fouling layer in this model alters the propagation characteristics of sound waves at the interface, leading to... and Relative to the baseline value and An offset occurs; in this embodiment, this offset is quantified as an adhesion index using the following formula. : in, Real-time attenuation / phase shift signal, sourced from the data acquisition module; The reference attenuation / phase shift signal and weighting coefficients are derived from preset calibration parameters; That is, the relative change between the attenuated signal and the reference attenuated signal; This refers to the relative change between the phase-shifted signal and the reference phase-shifted signal. The module calculates a weighted sum of the two dimensionless relative changes and determines the result as the composite fouling adhesion index. Both terms on the right-hand side of the equation are products of dimensionless relative changes and dimensionless weights; therefore, their sum is... It is also a dimensionless parameter, with consistent dimensions; This embodiment provides an accurate online, non-destructive, and quantitative method for evaluating fouling adhesion. By fusing the relative changes of two acoustic signals—attenuation and phase shift—and performing a weighted summation, two different dimensions of physical measurement indicators are integrated into a single, robust composite fouling adhesion index. This solves the problem that traditional methods cannot assess the physical adhesion strength of fouling in real time, providing a key, quantifiable input for the subsequent performance prediction module to accurately calculate fouling resistance.

[0019] Example 3 The quantitative evaluation module is used to determine the chemical solubility property index, including: Call the multi-channel reflectance vector in the original data stream, as well as the preset set of spectral channel indexes characterizing the absorption peaks of organic matter, the set of spectral channel indexes characterizing the absorption peaks of inorganic matter, and the sensitivity weighting coefficients; Based on an empirical proportional model, the sum of weighted reflectances representing organic matter in the multi-channel reflectance vectors is calculated separately. And the sum of weighted reflectance characterizing inorganic materials; The chemical solubility index is obtained by calculating the ratio of the sum of weighted reflectances characterizing organic matter to the sum of weighted reflectances characterizing inorganic matter.

[0020] This embodiment further defines the determination of the chemical solubility characteristic index in the quantitative evaluation module. This method aims to provide an online approach for distinguishing between organic and inorganic components of contaminated chemicals. In calculation At that time, the quantization evaluation module calls the multi-channel reflectance vector in the original data stream. Simultaneously, the module calls the preset set of spectral channel indices characterizing the absorption peaks of organic compounds. Spectral channel index set characterizing absorption peaks of inorganic substances and sensitivity weighting coefficient ; Spectral Channel Index Set : refers to the numbering of a specific wavelength channel pre-selected based on the spectral absorption characteristics of common pollutants, such as CH bonds representing organic matter, Si-O bonds or OH bonds representing inorganic matter or water. The source is a pre-set standard pollutant spectral library. Sensitivity weighting coefficient This refers to the calibration coefficient for the corresponding spectral channel, used to correct for sensitivity differences between different channels; it is obtained through multivariate calibration experiments; specifically, the preparation... The group has a known ratio of organic to inorganic components, denoted as . ,in Standard soiled samples were used, and their corresponding multi-channel reflectance vectors were measured respectively. Based on this Group The data, after being calibrated using multivariate methods such as partial least squares regression, is used to solve for the following formula: and Sensitivity weighting coefficients with optimal relevance and ; The module is based on an empirical proportional model, and its technical motivation stems from the Beer-Lambert law in spectral analysis, which states that the content of a specific component is related to the absorbance / reflectance at a specific wavelength. The calculation is performed using the following formula: in, Chemical solubility index, dimensionless; This usually indicates that the soiling is mainly composed of organic matter. This indicates that the fraction is primarily composed of inorganic substances; to ensure robustness of calculations, a constant is added to the denominator. It is a preset non-negative minimum value used to prevent division by zero errors, for example... This is to avoid computational overflow caused by extremely weak signals from inorganic materials. : The first vector or The reflectivity value of the channel comes from the data acquisition module; The spectral channel index set and sensitivity weighting coefficients are derived from preset calibration parameters; Calculate the sum of the weighted reflectances representing organic matter in the multi-channel reflectance vector, which is the numerator of the formula. ; And the sum of the weighted reflectances characterizing inorganic materials, which is the denominator of the formula. ; The chemical solubility index is obtained by calculating the ratio of the sum of weighted reflectances characterizing organic matter to the sum of weighted reflectances characterizing inorganic matter. ; In the formula, All are dimensionless reflectance. It is also a dimensionless calibration coefficient, therefore both the numerator and denominator are dimensionless quantities, and their ratio is... It is also a dimensionless parameter, with consistent dimensions; This embodiment provides a method for online and rapid differentiation of fouling chemical components; through this empirical proportional model, complex multispectral vector data is reduced to a single, intuitive chemical solubility characteristic index. The index This clearly reflects whether the contamination is biased towards water-soluble inorganic matter or solvent-requiring organic matter, providing a core decision-making basis for the subsequent performance prediction module to calculate the cleaning rate and for the strategy generation module to intelligently adjust the concentration ratio of cleaning agent A (organic) and cleaning agent B (inorganic).

[0021] The performance prediction module is used to determine the predicted cleaning success rate, including: The fouling resistance index is determined based on the adhesion index, real-time temperature, real-time humidity, and preset model coefficients. The cleaning rate index is determined based on the chemical solubility characteristic index, cleaning strategy control variables, and preset rate coefficients. By combining the fouling resistance index, cleaning rate index, the duration of action in the cleaning strategy control variables, and the preset gain coefficient, the predicted cleaning success rate is determined based on the logistic growth model.

[0022] This embodiment further defines the cleaning success rate predicted by the performance prediction module. The detailed three-step logic; this predictive model is a core custom model that connects soiling characteristics with cleaning strategies to predict cleaning results; The performance prediction module is based on the adhesion index calculated by the quantitative evaluation module. Real-time temperature Real-time humidity and preset model coefficients Determine the fouling resistance index ; Fouling Resistance Index It refers to a dimensionless parameter that integrates the physical adhesion strength of dirt and environmental factors such as temperature and humidity, used to quantify the inherent difficulty of removing current dirt; The module is based on the chemical solubility index calculated by the quantitative evaluation module. Control variables for the cleaning strategy to be evaluated and preset rate coefficient Determine the cleaning rate index ; Cleaning rate index This refers to a parameter used to quantify the effectiveness of a specific cleaning strategy per unit time, with its physical dimensions set as follows: That is, the percentage of cleanliness achieved per second; The module combines the fouling resistance index obtained in the first step. The cleaning rate index obtained in the second step The duration of action in the control variables of the cleaning strategy and preset gain coefficient Based on the logistic growth model, the predicted cleaning success rate is determined. ; Gain coefficient This refers to a dimensionless parameter used to control the steepness of the logistic function curve; its origin is derived from other model coefficients, such as... Together, they are obtained through regression training using historical clean data; during system operation, these coefficients will be dynamically updated by the model correction module; The technical motivation behind this Logistic / Sigmoid model is that cleaning effects are not simply linearly cumulative, but rather have a resistance threshold and a cleaning saturation effect; the cleaning action must overcome... The resistance represented by the logistic function, and the cleaning effect will eventually saturate around 100%; The model formula used in this embodiment can well describe this S-shaped growth curve, and is as follows: in, : Predicted cleaning success rate, dimensionless (0-1); Cleaning rate index, with dimensions of ; : Action time, dimensionless ; : Fouling resistance index, a dimensionless parameter; Gain coefficient, a preset parameter, is a dimensionless parameter; The core of this model lies in calculating net cleaning power, i.e. ;where Fc\tau This represents the total cleaning effect and is dimensionless. Representing fouling resistance, it is also dimensionless; therefore, the exponential term... The whole is dimensionless. It is also dimensionless, yet has consistent dimensions; when the total action is greater than the fouling resistance, the net efficacy is positive, and the cleaning success rate is high. The net effectiveness approaches 1; conversely, when the amount of action is insufficient to overcome the resistance, the net effectiveness is negative. Approaching 0; This embodiment constructs a system capable of connecting to fouling characteristics. With cleaning strategy The core prediction engine abstracts the complex cleaning process into an adversarial relationship between resistance and rate, and uses a logistic growth model for nonlinear fitting, enabling the system to adapt to any strategy. Any soiling and environment Cleaning effect Performing quantitative predictions is the prerequisite and foundation for the subsequent policy generation module to perform constraint optimization.

[0023] The performance prediction module is used to determine the fouling resistance index, including: Call the adhesion index, real-time humidity, real-time temperature, and preset humidity normalization baseline value, temperature normalization baseline value and model coefficients; Based on an empirical linear regression model, the adhesion index, the ratio of real-time humidity to the humidity normalized baseline value, and the ratio of real-time temperature to the temperature normalized baseline value are linearly combined. The result of the linear combination was determined as the fouling resistance index.

[0024] This embodiment specifically defines the determination of the fouling resistance index in the performance prediction module. The way; In calculation At that time, the performance prediction module calls the adhesion index from the previous step. Real-time humidity Real-time temperature and the preset humidity normalized reference value from the system configuration. Temperature normalized reference value and model coefficients ; Normalized baseline value : refers to the reference value used for dimensionless transformation of environmental parameters, for example , 25°C, its source is a preset value; Model coefficients : Refers to the dimensionless fitting coefficient, which is obtained by training a historical clean dataset using multiple linear regression; this dataset contains Each set of historical data includes a historical adhesion index. Historical humidity Historical temperature And the corresponding historical fouling resistance, calibrated experimentally, is denoted as ,in These coefficients were fitted using regression analysis. So that the following formula can be calculated and The residual is the smallest; The calculation employs an empirical linear regression model, the technical motivation of which is to comprehensively assess the combined impact of physical adhesion and environmental factors on cleaning difficulty. The model formula is as follows: in, : Fouling resistance index, a dimensionless parameter; Adhesion index, derived from the quantitative evaluation module, is a dimensionless parameter; The ratio of real-time humidity to the normalized baseline humidity value is a dimensionless parameter. The ratio of real-time temperature to the normalized temperature reference value is a dimensionless parameter. Model coefficients, derived from preset parameters, are dimensionless parameters; The model uses the adhesion index Humidity normalized ratio and temperature normalization ratio A linear combination is performed; all terms on the right side of the formula are dimensionless, therefore It is dimensionless, yet has consistent dimensions; based on physical understanding, the coefficients... Adhesion and The higher the humidity, the stronger the adhesion, the more difficult it is to remove dirt, and the greater the resistance. The temperature is expected to be positive; the higher the temperature, the softer the adhering material may become, resulting in lower resistance, hence the negative sign. The module determines the result of this linear combination as the fouling resistance index. ; This embodiment provides a more comprehensive definition of fouling resistance that is closer to actual working conditions; it considers not only the physical adhesion of the fouling itself. It also innovatively incorporates environmental factors The resistance model was incorporated through normalization and linear regression; this allows the system to automatically identify and quantify situations where cleaning resistance increases significantly, such as in low-temperature and high-humidity environments, thereby greatly improving efficiency. The accuracy of the index and Robustness of the prediction model.

[0025] The efficiency prediction module is used to determine the cleaning rate index, including: The chemical solubility index and cleaning strategy control variables are used. The cleaning strategy control variables include: water pressure, cleaning solution temperature, cleaning agent A concentration and cleaning agent B concentration. It also calls the preset pressure normalization reference value, temperature normalization reference value, physical cleaning rate coefficient, and chemical enhancement coefficient; Based on the physical cleaning rate coefficient, the ratio of water pressure to the normalized pressure reference value and the ratio of cleaning fluid temperature to the normalized temperature reference value are weighted and summed to obtain the physical cleaning rate term. Based on the chemical solubility index, the concentration of cleaning agent A, the concentration of cleaning agent B, and the chemical enhancement coefficient, a chemical enhancement factor is determined. The chemical enhancement factor is used to dynamically adjust the effectiveness of the concentrations of cleaning agent A and cleaning agent B. The cleaning rate index is determined by multiplying the physical cleaning rate term by the chemical enhancement factor.

[0026] This embodiment specifically defines the determination of the cleaning rate index in the efficiency prediction module. The model employs an empirical multiplicative interaction model, designed to reflect the synergistic effect of physical and chemical cleaning. In calculation At that time, the performance prediction module calls the chemical solubility characteristic index from the previous step. And the cleaning strategy control variables to be evaluated; in this embodiment, the cleaning strategy control variables Specifically defined as including: water pressure Cleaning solution temperature Concentration of cleaning agent A and the concentration of cleaning agent B ; Cleaning agent A and cleaning agent B refer to special cleaning agents for organic and inorganic substances, respectively. At the same time, the module calls the preset pressure normalization baseline value. Temperature normalized reference value Physical cleaning rate coefficient and chemical enhancement coefficient ; Normalized baseline value : Reference values ​​used for dimensionless pressure and temperature, preset; Rate / Enhancement Factor Its source is obtained through regression training on a historical cleaning dataset; this dataset contains Each set of historical data includes historical fouling characteristics. Historical cleaning strategies used And the corresponding, experimentally calibrated historical cleaning rate is denoted as ,in These coefficients were fitted using nonlinear regression. So that the following formula can be calculated and The residual is the smallest; To ensure dimensional consistency, the physical cleaning rate coefficient The dimensions are set as Chemical enhancement coefficient The dimension is set to the reciprocal of concentration, for example... Assuming The dimensions are , The calculation formula is as follows: Obtain the physical cleaning rate item: The first part of the calculation formula Defined as the physical cleaning rate term; This is based on the physical cleaning rate coefficient ( The ratio of water pressure to the normalized reference value of pressure ( ), the ratio of the cleaning fluid temperature to the normalized reference temperature value ( ), and perform a weighted summation; Dimensional check: and All are dimensionless. Dimensions are Therefore, the dimension of the physical cleaning rate term is ; Identify the chemical enhancement factor: The second part of the calculation formula Defined as a chemical enhancing factor; This factor is based on the chemical solubility index. Concentration of cleaning agent A Concentration of cleaning agent B and chemical enhancement coefficient ( To determine; Dimensional check: ( ) ( ) Dimensionless Dimensionless; ( ) ( ) Dimensionless Dimensionless; the constant 1 is also dimensionless; therefore, the chemical enhancement factor as a whole is dimensionless. The core innovation of this design lies in the effectiveness of using chemical enhancement factors to dynamically adjust the concentrations of cleaning agent A and cleaning agent B; specifically: When the soiling is predominantly organic... If the value is high, then High item weight, When the item weight is low, at this time Contribution of organic solvents Enlarged; When the fouling is predominantly inorganic... If the value is low, then High item weight, When the item weight is low, at this time Contribution of inorganic solvents Enlarged; Determine the cleaning rate index: Based on the physical cleaning rate term and the chemical enhancement factor respectively The overall cleaning rate index is determined by multiplying the product of the dimensionless product and the product of the two. Its final dimension is ; This embodiment implements the physical strategy through a sophisticated multiplicative interaction model that ensures dimensional consistency. Chemical strategies The intelligent integration; especially the design of chemical enhancement factors, which utilizes the chemical solubility index As a dynamic weight, the targeted cleaning agent was automatically amplified. or The effect of this, while suppressing the contribution of mismatched cleaning agents; makes it possible Indices can accurately reflect specific strategies For specific stains It achieves the true cleaning rate, avoiding the waste of resources and potential chemical residues caused by blindly using two cleaning agents.

[0027] The strategy generation module is used to solve for the optimal cleaning strategy, including: Invoke cleaning strategy control variables, real-time temperature, and preset cost coefficients; Based on the linear cost accounting model, a resource cost function is constructed, which includes energy consumption cost and cleaning agent cost. Construct a constrained optimization objective, which aims to minimize the resource cost function and is constrained by the condition that the predicted cleaning success rate is not lower than the cleaning efficiency threshold. Solve the constraint optimization objective to determine the optimal cleaning strategy.

[0028] This embodiment specifically defines the optimal cleaning strategy corresponding to minimizing the resource cost function in the strategy generation module. The way; To perform the solution, the strategy generation module constructs a resource cost function. The steps are as follows: The module calls the cleaning strategy control variables to be optimized. Real-time temperature Data comes from the data acquisition module and preset cost coefficients. ; Cost coefficient This refers to the unit cost of various resources, which is pre-set based on actual electricity prices and cleaning agent purchase prices; to ensure consistency of dimensions, its dimensions are set as follows: For example: yuan Pa s , For example: yuan K s , For example: yuan s Assuming Dimensions are ; Based on the linear cost accounting model, the module constructs a resource cost function. ; Resource cost function Refers to the execution strategy The total economic cost required is expressed in yuan; in this embodiment, the function includes energy costs and cleaning agent costs, and its specific form is as follows: Dimensional check: Energy cost item 1: (Yuan Pa s ) (Pa) (s) Yuan; Energy cost item 2: (Yuan K s ) (K) (s) Yuan; Cleaning agent cost item: ( (Yuan s ) (%)+ (Yuan s ) (%) (s) Yuan; All terms in the formula are in units of yuan, can be added together, and have consistent dimensions. First item The total energy cost includes the cost of water pump pressurization and the cost of water heating; Second item Total cleaning agent cost; After constructing the cost function, the module constructs a constraint optimization objective; The goal of this optimization problem is to find a strategy that minimizes cost and achieves the desired results. ; In this embodiment, the constrained optimization objective is defined as: Objective function: That is, the objective is to minimize the resource cost function; Constraints: That is, the constraint is that the predicted cleaning success rate is not lower than the cleaning efficiency threshold. The module solves for the constrained optimization objective; The system employs numerical optimization algorithms, such as Sequential Least Squares Programming (SQP), or intelligent algorithms like Genetic Algorithms and Particle Swarm Optimization, to solve the above problems. The final solution... This refers to the determined optimal cleaning strategy; the choice of specific algorithm can be based on the efficiency requirements, the continuity and discreteness of the decision variables, and the cost function. and constraints The nonlinearity and nonconvexity intensity are adapted to achieve a balance between computational resources and global optimization. This embodiment transforms the cleaning decision from a fuzzy, experience-based operational problem into a precise, solvable constrained optimization problem; it does so by constructing a quantifiable cost function that ensures dimensional consistency. and performance constraints This system can ensure that the cleaning effect is no less than that of other systems. Under the rigid premise, automatically find the strategy combination that minimizes the total cost of resources such as water, electricity, and cleaning agents. This enables precise cost control and optimal resource allocation in the cleaning process.

[0029] The energy cost in the resource cost function includes the energy cost per unit temperature difference. Among them, when the cleaning fluid temperature is higher than the real-time temperature, the unit temperature difference energy consumption cost is calculated based on the difference between the cleaning fluid temperature and the real-time temperature. When the temperature of the cleaning fluid is no higher than the real-time temperature, the energy consumption cost per unit temperature difference is zero.

[0030] This embodiment specifically defines the resource cost function. A key detail of energy consumption costs is how heating costs are calculated; The energy cost in the resource cost function specifically includes the energy cost per unit temperature difference; this cost item is... The mathematical expression in is ; The function refers to taking 0 and... The larger value in; The technical motivation behind this model is to correct the simple linear model. The logical flaw of potential negative costs in the process makes the cost accounting model more consistent with physical reality; the specific logic is as follows: When the cleaning fluid temperature Greater than real-time temperature Time: at this time , The unit temperature difference energy consumption cost is calculated based on the difference between the cleaning fluid temperature and the real-time temperature; the total heating cost is... This means the system needs to consume energy to move water from ambient temperature. Heated to The actual costs incurred; When the cleaning fluid temperature No greater than the real-time temperature Time: i.e. ;at this time , The energy cost per unit temperature difference is zero; this means that the system uses ambient temperature water or colder water for cleaning, such as at night or in winter, without requiring additional energy for heating, so the heating cost is 0. This embodiment introduces... This nonlinear function precisely defines the triggering conditions and calculation logic for heating costs; this non-negativity correction ensures that the cost term always conforms to the physical meaning that costs cannot be negative, making the cost function... This ensures greater accuracy, thereby guaranteeing that the policy generation module obtains the optimal policy. Reliability and economy.

[0031] Example 9: The model correction module, used to correct the performance prediction module, includes: The actual cleaning efficiency and the predicted cleaning success rate are compared. Calculate the deviation between actual cleaning efficiency and predicted cleaning success rate, and determine the loss function based on the deviation; Based on the gradient descent update rule, the gradient of the loss function with respect to the preset model coefficients, rate coefficients, and gain coefficients in the performance prediction module is calculated. Update the model coefficients, rate coefficients, and gain coefficients based on the gradient and the preset learning rate.

[0032] This embodiment specifically defines the closed-loop feedback mechanism of the model correction module for correcting the performance prediction module; this mechanism adopts the standard gradient descent update rule of online machine learning to achieve model adaptation and self-evolution; In the optimal strategy After the process is completed and the actual cleaning effect is obtained, the model correction module performs the following correction process: Module calls include, for example, the actual cleaning efficiency measured by a light transmittance sensor. The predicted cleaning success rate used in this forecast The data originates from the performance prediction module, and the deviation between the two is calculated. ; Based on this bias, the module determines a loss function. In this embodiment, the mean squared error (MSE) is used as the loss function: loss function Used to quantify the accuracy of model predictions. The smaller the value, the more accurate the prediction; because and All are dimensionless. It is also dimensionless; To achieve model correction, the module calculates the loss function based on the gradient descent update rule. The set of all preset coefficients to be trained in the performance prediction module is denoted as . gradient ; Coefficient set This embodiment includes: Model coefficients in the model ; Rate coefficients in the model ; Gain coefficients in the model ; The gradient Through the chain rule, from via The formula is obtained by backpropagation calculation of the Logistic Stearns model, and its form is: ; After calculating the gradient, the module uses that gradient... and the preset learning rate Update the above model coefficients, rate coefficients, and gain coefficients; Learning rate It is a preset, dimensionless tuning parameter, such as 0.01, used to control the step size of each correction. The formula for the update rule is as follows: in, This is the current coefficient. These are the updated coefficients; to ensure dimensional consistency, this update rule requires correction terms. Dimensions and coefficients The dimensions must match. For non-dimensional coefficients, such as rate coefficients, the corresponding gradient terms need to be adapted to the dimensions before updating to ensure that the update operation is physically valid. For example, the dimensionless nature of the correction term can be achieved by multiplying the gradient term by a reference value with the same dimensions as the coefficient to be updated, or by dividing it by the typical scale value of the coefficient. This embodiment constructs a complete online machine learning closed loop based on actual effect feedback; by updating the rules through gradient descent, the system can utilize data from each actual cleaning process. Yes, it automatically and iteratively fine-tunes its core prediction model. Internal coefficients This gives the system a strong adaptive capability, enabling it to continuously learn the characteristics of new types of fouling or the impact of environmental changes, ensuring that the performance prediction module evolves over time and always maintains high accuracy in its predictions and the effectiveness of its optimal strategy. This closed-loop correction mechanism based on actual data feedback also effectively compensates for the loss of physical fidelity caused by the simplification of calculations in the various empirical models in the performance prediction module, ensuring that the final prediction effect of the model can continuously approach the real physical process through a data-driven approach.

[0033] 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 photovoltaic intelligent monitoring and management system based on the Internet of Things, characterized in that, include: The data acquisition module is used to acquire the raw data stream of the sensor array deployed on the photovoltaic panel. The raw data stream includes: the attenuation signal and phase shift signal of the acoustic impedance sensor, the multi-channel reflectivity vector of the multispectral sensor, and the real-time temperature and real-time humidity of the environmental sensor. The quantitative evaluation module is used to determine the composite fouling adhesion index and chemical solubility index based on the raw data stream and preset calibration parameters. The performance prediction module is used to determine the predicted cleaning success rate based on the composite dirt adhesion index, chemical solubility index, real-time temperature, real-time humidity, and preset cleaning strategy control variables. The strategy generation module is used to solve for the optimal cleaning strategy that minimizes the resource cost function based on the predicted cleaning success rate and the preset cleaning efficiency threshold. The model correction module is used to obtain the actual cleaning efficiency after executing the optimal cleaning strategy, and to correct the efficiency prediction module based on the deviation between the actual cleaning efficiency and the predicted cleaning success rate.

2. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 1, characterized in that, The quantitative evaluation module is used to determine the composite fouling adhesion index, including: Call the attenuation signal and phase shift signal in the original data stream, as well as the preset reference attenuation signal, reference phase shift signal and weighting coefficient of the clean photovoltaic panel surface; Based on the empirical model of acoustic impedance physics, the relative changes of the attenuation signal and the reference attenuation signal, as well as the relative changes of the phase shift signal and the reference phase shift signal, are weighted and summed. The weighted summation result was determined as the composite fouling adhesion index.

3. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 2, characterized in that, The quantitative evaluation module is used to determine the chemical solubility characteristic index, including: Call the multi-channel reflectance vector in the original data stream, as well as the preset set of spectral channel indexes characterizing the absorption peaks of organic matter, the set of spectral channel indexes characterizing the absorption peaks of inorganic matter, and the sensitivity weighting coefficients; Based on an empirical proportional model, the sum of weighted reflectances representing organic matter in the multi-channel reflectance vectors is calculated separately. And the sum of weighted reflectance characterizing inorganic materials; The chemical solubility index is obtained by calculating the ratio of the sum of weighted reflectances characterizing organic matter to the sum of weighted reflectances characterizing inorganic matter.

4. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 3, characterized in that, The performance prediction module is used to determine the predicted cleaning success rate, including: The fouling resistance index is determined based on the adhesion index, real-time temperature, real-time humidity, and preset model coefficients. The cleaning rate index is determined based on the chemical solubility characteristic index, cleaning strategy control variables, and preset rate coefficients. By combining the fouling resistance index, cleaning rate index, the duration of action in the cleaning strategy control variables, and the preset gain coefficient, the predicted cleaning success rate is determined based on the logistic growth model.

5. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 4, characterized in that, The performance prediction module is used to determine the fouling resistance index, including: Call the adhesion index, real-time humidity, real-time temperature, and preset humidity normalization baseline value, temperature normalization baseline value and model coefficients; Based on an empirical linear regression model, the adhesion index, the ratio of real-time humidity to the humidity normalized baseline value, and the ratio of real-time temperature to the temperature normalized baseline value are linearly combined. The result of the linear combination was determined as the fouling resistance index.

6. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 5, characterized in that, The performance prediction module is used to determine the cleaning rate index, including: The chemical solubility index and cleaning strategy control variables are used. The cleaning strategy control variables include: water pressure, cleaning solution temperature, cleaning agent A concentration and cleaning agent B concentration. It also calls the preset pressure normalization reference value, temperature normalization reference value, physical cleaning rate coefficient, and chemical enhancement coefficient; Based on the physical cleaning rate coefficient, the ratio of water pressure to the normalized pressure reference value and the ratio of cleaning fluid temperature to the normalized temperature reference value are weighted and summed to obtain the physical cleaning rate term. Based on the chemical solubility index, the concentration of cleaning agent A, the concentration of cleaning agent B, and the chemical enhancement coefficient, a chemical enhancement factor is determined. The chemical enhancement factor is used to dynamically adjust the effectiveness of the concentrations of cleaning agent A and cleaning agent B. The cleaning rate index is determined by multiplying the physical cleaning rate term by the chemical enhancement factor.

7. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 6, characterized in that, The strategy generation module is used to solve for the optimal cleaning strategy, including: Invoke cleaning strategy control variables, real-time temperature, and preset cost coefficients; Based on the linear cost accounting model, a resource cost function is constructed, which includes energy consumption cost and cleaning agent cost. Construct a constrained optimization objective, which aims to minimize the resource cost function and is constrained by the condition that the predicted cleaning success rate is not lower than the cleaning efficiency threshold. Solve the constraint optimization objective to determine the optimal cleaning strategy.

8. The photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 7, characterized in that, The energy cost in the resource cost function includes the energy cost per unit temperature difference. Among them, when the cleaning fluid temperature is higher than the real-time temperature, the unit temperature difference energy consumption cost is calculated based on the difference between the cleaning fluid temperature and the real-time temperature. When the temperature of the cleaning fluid is no higher than the real-time temperature, the energy consumption cost per unit temperature difference is zero.

9. A photovoltaic intelligent monitoring and management system based on the Internet of Things according to claim 8, characterized in that, The model correction module is used to correct the performance prediction module, including: The actual cleaning efficiency and the predicted cleaning success rate are compared. Calculate the deviation between actual cleaning efficiency and predicted cleaning success rate, and determine the loss function based on the deviation; Based on the gradient descent update rule, the gradient of the loss function with respect to the preset model coefficients, rate coefficients, and gain coefficients in the performance prediction module is calculated. Update the model coefficients, rate coefficients, and gain coefficients based on the gradient and the preset learning rate.

Images