Photovoltaic fire control temperature control prediction system

Abstract

The application discloses a kind of photovoltaic fire control temperature control prediction systems, belong to photovoltaic power generation technical field, it includes multi-source data acquisition module for obtaining photovoltaic surface temperature data, environmental monitoring data and ecological restoration monitoring data, generate multi-source monitoring dataset;Environment parameter fusion module is used to generate ecological correction factor and process multi-source monitoring dataset, generate comprehensive environmental parameter;Temperature trend prediction module is used to predict comprehensive environmental parameter, obtain temperature trend prediction value;Fire risk assessment module is used to risk assessment to temperature trend prediction value, generate fire risk index;Early warning control generation module is used to judge fire risk index, generate early warning control signal;Linkage response control module is used to adjust power output and start fire control unit.The application predicts temperature trend and assesses risk by fusing ecological restoration data, linkage regulates power and fire control, realizes active early warning and intelligent closed-loop control of photovoltaic fire.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation technology, and in particular to a photovoltaic fire-fighting temperature control and prediction system. Background Technology

[0002] Photovoltaic power generation, as a clean and renewable energy source, plays a crucial role in the global energy structure transformation. A photovoltaic power plant consists of numerous photovoltaic module arrays, inverters, combiner boxes, and other equipment. Among these, the operating temperature of the photovoltaic modules is a key factor affecting their power generation efficiency and lifespan. Excessive temperature not only reduces photoelectric conversion efficiency but also accelerates the aging of module materials, potentially leading to hot spot effects and even fires, posing a serious threat to power plant assets and personnel safety.

[0003] Existing photovoltaic power plant safety monitoring systems typically use temperature sensors or infrared thermal imaging technology to monitor the surface temperature of photovoltaic modules in real time. These systems usually have a fixed high-temperature alarm threshold. When the monitored real-time temperature exceeds this preset threshold, the system will trigger an alarm signal, prompting maintenance personnel to conduct an inspection, or in extreme cases, directly executing an emergency shutdown procedure. This monitoring method can detect high-temperature anomalies to a certain extent.

[0004] However, the aforementioned existing technical solutions have significant shortcomings. First, this alarm mechanism based on fixed thresholds is passive, only reacting after a high-temperature event has occurred, lacking foresight and failing to provide a time window for preventative maintenance and proactive intervention. Second, its environmental considerations are overly simplistic, typically relying only on macroscopic meteorological data and ignoring the microclimate impact of local ecological environments such as vegetation and water bodies on module heat dissipation, leading to biased temperature assessments. Finally, its response measures are relatively simple and crude, usually involving direct alarms or shutdowns, lacking the ability to classify and handle risks according to their severity, easily causing unnecessary power generation losses. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a photovoltaic fire-fighting temperature control and prediction system. This system employs a technical solution that integrates ecological restoration monitoring data to predict temperature trends, combines real-time operating status for risk coupling assessment, and links power output with fire-fighting units for tiered responses. This enables proactive, accurate prediction and intelligent closed-loop control of photovoltaic fire risks.

[0006] The above objectives can be achieved through the following approach: A photovoltaic fire-fighting temperature control and prediction system includes a multi-source data acquisition module for acquiring photovoltaic surface temperature data of photovoltaic modules, environmental monitoring data, and ecological restoration monitoring data to generate a multi-source monitoring dataset; an environmental parameter fusion module for performing quantitative analysis based on the ecological restoration monitoring data to generate an ecological correction factor, and using the ecological correction factor to fuse the multi-source monitoring dataset to generate comprehensive environmental parameters; a temperature trend prediction module for performing structured prediction processing based on the comprehensive environmental parameters to calculate a predicted temperature trend value; a fire risk assessment module for acquiring the current operating status parameters of the photovoltaic modules and performing risk coupling assessment in conjunction with the predicted temperature trend value to calculate and generate a fire risk index; an early warning control generation module for performing risk classification judgment based on the fire risk index and generating an early warning control signal; and a linkage response control module for adjusting the power output of the photovoltaic modules based on the early warning control signal and linkage with the fire-fighting control unit to respond.

[0007] Optionally, the multi-source data acquisition module includes: a component temperature acquisition unit, used to acquire temperature data through temperature sensors deployed on photovoltaic modules, and use the acquired surface temperature data as the first temperature data; a field environment acquisition unit, used to acquire ambient temperature, air humidity, wind speed and total solar radiation intensity through environmental sensors deployed in the photovoltaic field, forming environmental monitoring data, wherein the ambient temperature is used as the second temperature data; an ecological restoration acquisition unit, used to perform in-situ monitoring through ecological sensors deployed in the ecological restoration area, and acquire ecological restoration monitoring data including soil moisture and vegetation coverage; and a multi-source data integration unit, used to align and integrate the first temperature data, the environmental monitoring data and the ecological restoration monitoring data with timestamps to generate a multi-source monitoring dataset.

[0008] Optionally, the environmental parameter fusion module includes: an initial parameter fusion unit, used to standardize and weight the first temperature data and the environmental monitoring data in the multi-source monitoring dataset to generate initial fusion parameters; an ecological factor quantification unit, used to quantify the local environmental cooling effect based on the ecological restoration monitoring data to generate an ecological correction factor; and an environmental parameter correction unit, used to perform correction calculations on the initial fusion parameters based on the ecological correction factor to obtain comprehensive environmental parameters.

[0009] Optionally, the temperature trend prediction module includes: an environmental parameter decomposition unit, used to decompose the comprehensive environmental parameters to obtain a basic environmental component and an ecological correction component; a temperature baseline prediction unit, used to predict the temperature baseline based on the basic environmental component and calculate the basic temperature trend; a microclimate effect quantification unit, used to quantify the local microclimate effect based on the ecological correction component and generate a temperature correction amount; and a temperature prediction synthesis unit, used to combine the basic temperature trend and the temperature correction amount for joint analysis and calculation to obtain a temperature trend prediction value.

[0010] Optionally, the system further includes: collecting the latest operating data within a time period, the latest operating data including photovoltaic module operating records, fault event records, operating characteristic parameters, and abnormal temperature fluctuation characteristics; reconstructing the correlation between the basic environmental component and the ecological correction component based on the latest operating data to generate a dynamic correlation structure; dynamically adjusting the effect intensity of the ecological correction factor based on the dynamic correlation structure to generate an updated ecological correction factor; and using the updated ecological correction factor to collaboratively optimize the calculation logic of the temperature baseline prediction unit and the microclimate effect quantification unit to generate updated temperature trend prediction calculation rules.

[0011] Optionally, the fire risk assessment module includes: a real-time status acquisition unit, used to acquire the real-time current parameters and real-time voltage parameters of the photovoltaic module, and generate the current operating status parameters; an initial risk calculation unit, used to perform fusion calculation based on the temperature trend prediction value and the current operating status parameters, and calculate an initial risk value; and a risk normalization processing unit, used to acquire a historical risk benchmark, and use the historical risk benchmark to normalize the initial risk value, and generate a fire risk index.

[0012] Optionally, the early warning control generation module includes: a risk threshold acquisition unit, used to acquire a multi-level risk threshold system, the multi-level risk threshold system including at least two different levels of risk response thresholds; a risk level determination unit, used to compare the fire risk index with the multi-level risk threshold system level by level to determine the risk level; and an early warning signal generation unit, used to perform signal mapping according to the risk level to generate an early warning control signal containing corresponding control instructions.

[0013] Optionally, the acquisition of the multi-level risk threshold system includes: constructing a risk level sample library based on historical operation data and fire event records of the photovoltaic field, the type of photovoltaic modules, and the characteristics of the installation environment; performing cluster analysis based on the risk level sample library to divide it into multiple risk intervals, forming a multi-level risk threshold system; and dynamically calibrating the risk threshold system according to seasonal changes and module aging status.

[0014] Optionally, the linkage response control module includes: a power gradient control unit, used to match the corresponding power reduction coefficient according to the risk level in the early warning control signal, and gradually reduce the output power of the photovoltaic module; a fire-fighting graded activation unit, used to trigger the fire control unit to perform a stepped fire-fighting response based on the early warning control signal, and sequentially activate the cooling spray device and the flame retardant spray device; a response performance verification unit, used to collect power adjustment rate and fire-fighting execution status data in real time, and generate response performance evaluation indicators; and a dynamic adjustment unit, used to dynamically adjust the power reduction coefficient and the priority of the fire-fighting response according to the matching degree between the decay rate of the fire risk index and the response performance evaluation indicators.

[0015] Based on the same inventive concept, this invention also provides a photovoltaic fire-fighting temperature control prediction method. The method includes acquiring photovoltaic surface temperature data of photovoltaic modules, environmental monitoring data, and ecological restoration monitoring data to generate a multi-source monitoring dataset; performing quantitative analysis based on the ecological restoration monitoring data to generate an ecological correction factor, and using the ecological correction factor to fuse the multi-source monitoring dataset to generate comprehensive environmental parameters; performing structured prediction processing based on the comprehensive environmental parameters to calculate a predicted temperature trend value; acquiring the current operating status parameters of the photovoltaic modules and combining them with the predicted temperature trend value for risk coupling assessment to calculate a fire risk index; performing risk classification judgment based on the fire risk index to generate an early warning control signal; and adjusting the power output of the photovoltaic modules based on the early warning control signal and coordinating with the fire control unit for response.

[0016] Compared with the prior art, the present invention has the following advantages: 1. This invention, by introducing ecological restoration monitoring data and constructing an ecological correction factor to fuse multi-source environmental parameters, can precisely quantify the cooling effect of local microclimate on photovoltaic modules. This reduces model bias caused by traditional prediction methods relying solely on macroscopic meteorological data, making the description of the actual thermal environment of the modules more accurate, thereby improving the accuracy of temperature trend prediction and providing a data foundation for subsequent risk assessment.

[0017] 2. This invention establishes a forward-looking dynamic risk assessment mechanism that couples future temperature trend predictions with real-time operating status parameters of photovoltaic modules. This method surpasses the traditional alarm mode that relies solely on the current static temperature threshold, enabling earlier identification of high-risk operating conditions caused by the combined effects of high temperature and high electrical load. It achieves a shift from passive response to proactive prediction, improving the predictability and accuracy of fire risk identification.

[0018] 3. This invention designs an intelligent closed-loop linkage response control strategy. This strategy is based on a multi-level risk threshold system and can perform gradient power regulation and graded fire alarm activation. Simultaneously, the system improves the accuracy and economy of intervention measures through real-time verification of response performance and dynamic optimization of the control strategy. This reduces power generation losses caused by overreaction while ensuring decisive handling under high-risk conditions, achieving precise, efficient, and adaptive management of fire risks.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a framework diagram of a photovoltaic fire-fighting temperature control and prediction system according to an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the structure of a photovoltaic fire-fighting temperature control and prediction system according to an embodiment of the present invention.

[0023] Figure 3 This is a comparison chart of the adaptive optimization effects of embodiments of the present invention.

[0024] Figure 4 This is a three-dimensional schematic diagram of fire risk assessment according to an embodiment of the present invention.

[0025] Figure 5 This is a flowchart illustrating a photovoltaic fire-fighting temperature control prediction method according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Reference Figure 1One embodiment of the present invention proposes a photovoltaic fire-fighting temperature control prediction system, which uses integrated ecological restoration monitoring data to predict temperature trends and combines real-time working status for risk coupling assessment, and links power output with fire-fighting units for graded response. This technical solution enables proactive and accurate prediction and intelligent closed-loop control of photovoltaic fire risks.

[0028] The system described in this embodiment specifically includes: The multi-source data acquisition module is used to acquire photovoltaic surface temperature data, environmental monitoring data, and ecological restoration monitoring data of photovoltaic modules, and generate multi-source monitoring datasets; The environmental parameter fusion module is used to perform quantitative analysis based on the ecological restoration monitoring data, generate ecological correction factors, and use the ecological correction factors to fuse the multi-source monitoring dataset to generate comprehensive environmental parameters. The temperature trend prediction module is used to perform structured prediction processing based on the comprehensive environmental parameters and calculate the predicted temperature trend value. The fire risk assessment module is used to obtain the current operating status parameters of the photovoltaic module, and combine them with the temperature trend prediction value to perform risk coupling assessment and calculate and generate a fire risk index. The early warning control generation module is used to determine the risk level based on the fire risk index and generate an early warning control signal. The linkage response control module is used to adjust the power output of the photovoltaic module based on the early warning control signal and to link the fire control unit to respond.

[0029] Specifically, the system first comprehensively captures internal and external factors affecting the temperature of photovoltaic modules, including not only the module's own surface temperature and macro-environmental meteorological data, but also innovatively incorporates ecological restoration monitoring data. Next, the localized cooling effect of ecological restoration is quantified, generating an ecological correction factor to correct and integrate all monitoring data, forming a comprehensive environmental parameter that more accurately reflects the module's microenvironment. Based on this comprehensive environmental parameter, structured predictions are performed to calculate future temperature change trends, yielding a temperature trend prediction value. This forward-looking temperature trend prediction value is then coupled with the module's real-time operating status parameters for risk assessment, generating a dynamic and comprehensive fire risk index. Finally, based on this fire risk index, a risk level is determined, generating an early warning control signal and driving the linkage response control module to execute corresponding control strategies. This involves proactively adjusting the module's power output to cool it at the source and coordinating with the fire control unit to prepare for response, thus forming a closed-loop management system of perception, prediction, assessment, decision-making, and control. By introducing and quantitatively integrating ecological restoration monitoring data, the precision and accuracy of the perception of the actual operating thermal environment of photovoltaic modules are improved, thereby enhancing the reliability of temperature trend prediction. By coupling the predicted temperature trend with the real-time electrical status of the components, potential fire risks can be identified earlier, especially under extreme conditions of high temperature and high load. Furthermore, the generated early warning control signals can trigger coordinated responses from power adjustment and fire suppression units, proactively suppressing further risk development and preparing fire-fighting measures in advance. This improves the source control and proactive defense against photovoltaic fire risks, thereby enhancing the inherent safety level and intelligent operation and maintenance capabilities of photovoltaic power plants.

[0030] Optionally, the multi-source data acquisition module includes: The module temperature acquisition unit is used to acquire temperature data through temperature sensors deployed on the photovoltaic module, and to obtain photovoltaic surface temperature data as the first temperature data. The field environment data acquisition unit is used to collect ambient temperature, air humidity, wind speed and total solar radiation intensity through environmental sensors deployed in the photovoltaic field to form environmental monitoring data, wherein the ambient temperature is used as the second temperature data. The ecological restoration data collection unit is used to collect ecological restoration monitoring data, including soil moisture and vegetation cover, through in-situ monitoring using ecological sensors deployed in the ecological restoration area. The multi-source data integration unit is used to align and integrate the first temperature data, the environmental monitoring data, and the ecological restoration monitoring data with timestamps to generate a multi-source monitoring dataset.

[0031] Specifically, the photovoltaic (PV) surface temperature data is first acquired through a module temperature acquisition unit. This process requires deploying high-precision temperature sensors, such as thermocouples or platinum resistance thermometers, at key temperature measurement points on the PV module's backsheet to perform continuous contact temperature measurement. This captures the actual thermal state of the PV module under sunlight and electrical load in real time, obtaining the PV surface temperature data, which is defined as the primary temperature data. Secondly, a local meteorological monitoring network is constructed through an area environmental acquisition unit. Integrated environmental sensors are installed at representative locations within the PV field, including thermometers for measuring ambient temperature, hygrometers for measuring air humidity, anemometers for measuring wind speed, and radiometers for measuring total solar radiation intensity. These sensors simultaneously collect data, forming real-time environmental monitoring data reflecting the macroclimatic conditions of the PV field. The collected ambient temperature serves as the secondary temperature data. Subsequently, an ecological restoration acquisition unit conducts in-situ monitoring of the ecological restoration areas between or around the PV arrays. In-situ monitoring refers to direct on-site measurement without damaging the original environment. This ecological restoration data acquisition unit deploys soil moisture sensors to acquire soil moisture and uses vegetation index sensors or hyperspectral imaging equipment to calculate the normalized vegetation index (NZD), thereby quantifying vegetation cover. This data collectively constitutes ecological restoration monitoring data, which directly reflects the local microclimate's regulatory capacity. Finally, all acquired temperature data, environmental monitoring data, and ecological restoration monitoring data are transmitted to the multi-source data integration unit. The core task of this unit is timestamp alignment and integration. Due to potential differences in sampling frequencies and data transmission delays among different sensors, a unified Network Time Protocol (NTP) is used to synchronize the time of all acquisition devices, ensuring accurate timestamps for each data point. Upon receiving the data stream, interpolation or resampling algorithms are used to unify data from different sources onto the same time series, ultimately generating a structured multi-source monitoring dataset. This dataset includes the photovoltaic module's own temperature, macroscopic environmental parameters, and microscopic ecological influencing factors at each time point, providing a complete and synchronized data foundation for subsequent environmental parameter fusion and temperature prediction. By constructing a multi-source data acquisition system encompassing three dimensions—modules, site environment, and ecological restoration—the comprehensiveness and granularity of data input are enhanced. Compared to traditional methods that rely solely on macroscopic meteorological data or single module temperature, this approach quantifies and collects data on the local microclimate effects influencing photovoltaic module heat dissipation, specifically the cooling effect resulting from ecological restoration measures. This strengthens the ability to perceive dynamic changes in the local environment and improves the reliability and foresight of the entire photovoltaic fire protection temperature control prediction multi-source monitoring dataset.

[0032] Optionally, the environmental parameter fusion module includes: The initial parameter fusion unit is used to standardize and weight the first temperature data and the environmental monitoring data in the multi-source monitoring dataset to generate initial fusion parameters. An ecological factor quantification unit is used to quantify the local environmental cooling effect based on the ecological restoration monitoring data and generate ecological correction factors. An environmental parameter correction unit is used to perform correction calculations on the initial fusion parameters based on the ecological correction factor to obtain comprehensive environmental parameters.

[0033] Specifically, the initial parameter fusion unit first receives the first temperature data from the multi-source monitoring dataset, namely the photovoltaic surface temperature data, as well as environmental monitoring data, including ambient temperature, air humidity, wind speed, and total solar radiation intensity. Since these data have different physical dimensions and numerical ranges, direct processing would introduce bias. Therefore, they are first standardized using the zero-mean Z-score standardization method to eliminate the influence of dimensions and make them comparable. Next, based on the physical mechanisms of the impact of each environmental parameter on the temperature rise of the photovoltaic module and the results of historical data analysis, weights are assigned to each standardized parameter, and weighted fusion is performed to generate the initial fusion parameters. Subsequently, the ecological factor quantification unit specifically processes the ecological restoration monitoring data. Based on two core indicators—soil moisture and vegetation cover—the ecological factor quantification unit quantifies the local environmental cooling effect caused by vegetation transpiration and soil moisture evaporation. This effect is calculated by constructing a physical model or empirical formula to calculate the ecological correction factor. For calculating the ecological correction factor... ,have: ; in, This refers to standardized soil moisture data. This is the standardized vegetation cover data; and These are the soil moisture weighting coefficient and vegetation cover weighting coefficient, respectively, representing the contribution of soil moisture and vegetation cover to local cooling. Their values ​​were obtained through long-term microclimate observation or fluid dynamics simulation analysis of the photovoltaic field. This ecological correction factor is a dimensionless parameter used to correct for macroscopic environmental conditions. Finally, as... Figure 2As shown, the environmental parameter correction unit uses the ecological correction factor generated by the ecological factor quantification unit to correct the initial fusion parameters calculated by the initial parameter fusion unit. This correction typically employs multiplication to reflect the weakening effect of ecological cooling on the basic thermal environment, ultimately yielding comprehensive environmental parameters. These comprehensive environmental parameters not only include the influence of macroscopic meteorological conditions and the thermal state of the photovoltaic modules themselves, but also accurately account for the microclimate regulation effect brought about by local ecological restoration measures. By introducing the ecological correction factor to correct the initial fused environmental parameters, the problem of traditional methods only considering macroscopic meteorological data and ignoring the positive impact of on-site ecological restoration measures such as vegetation and water bodies on the local microclimate is reduced. This also reduces misjudgments caused by overestimating the ambient temperature, improving the accuracy and reliability of the entire temperature control prediction and fire risk assessment.

[0034] Optionally, the temperature trend prediction module includes: An environmental parameter decomposition unit is used to decompose the comprehensive environmental parameters to obtain basic environmental components and ecological correction components. A temperature baseline prediction unit is used to predict the temperature baseline based on the basic environmental components and calculate the basic temperature trend. The microclimate effect quantification unit is used to quantify the local microclimate effect based on the ecological correction component and generate a temperature correction amount. The temperature prediction synthesis unit is used to perform joint analysis and calculation by combining the basic temperature trend and the temperature correction amount to obtain the temperature trend prediction value.

[0035] Specifically, the environmental parameter decomposition unit first performs inverse analysis on the input comprehensive environmental parameters, decomposing them into basic environmental components and ecological correction components. The basic environmental components mainly consist of initial fusion parameters, reflecting the combined influence of macro-meteorological factors such as total solar radiation intensity, ambient temperature, and wind speed. The ecological correction components, derived from ecological correction factors, characterize the microclimate regulation effect brought about by local ecological conditions such as soil moisture and vegetation cover. This decomposition aims to separate the dominant factors affecting temperature change from local correction factors for differentiated processing. Subsequently, the temperature baseline prediction unit predicts the temperature baseline based on the basic environmental components. The temperature baseline prediction unit uses a time-series prediction model, such as a Long Short-Term Memory (LSTM) network, which can construct a time-series prediction model by learning the nonlinear relationship between historical basic environmental components and photovoltaic surface temperature data. During the prediction phase, the currently and recently collected basic environmental components are input into the trained time-series prediction model, which outputs a baseline temperature trend over a future period without considering local ecological effects. Simultaneously, the microclimate effect quantification unit, based on the ecological correction components, accurately quantifies the cooling effect of local microclimates, generating temperature correction values. The microclimate effect quantification unit analyzes the relationship between ecological correction components and actual temperature reduction values ​​by establishing thermodynamic models or data-driven regression models. This allows for the calculation of the specific temperature correction amount that should occur relative to the baseline temperature trend under current ecological conditions. The model is built by selecting typical areas with different ecological characteristics within the photovoltaic field, simultaneously collecting historical data on ecological correction components such as standardized soil moisture and vegetation cover, and the actual temperature reduction values ​​of the modules. Based on this data, the energy balance equation is fitted using the least squares method or regression training is performed using algorithms such as random forests. This temperature correction amount is typically negative, indicating a cooling effect. Finally, the temperature prediction synthesis unit jointly analyzes and calculates the baseline temperature trend calculated by the temperature baseline prediction unit and the temperature correction amount generated by the microclimate effect quantification unit, usually through algebraic superposition, to obtain the final temperature trend prediction value. This is for calculating future... Temperature trend prediction at any time ,have: ; in, It is the base temperature trend obtained from the temperature baseline prediction unit; and This is a temperature correction calculated by the microclimate effect quantification unit based on the current ecological correction component. Through this synthetic calculation, a comprehensive temperature trend prediction value is obtained that reflects both macro-climate changes and accurately incorporates the impact of local microclimates. By decomposing the complex temperature prediction problem into two sub-problems—basic trend prediction and microclimate effect correction—the accuracy and interpretability of the prediction are improved. Furthermore, by separating and quantifying the cooling effect of ecological restoration measures, not only is the accuracy of predicting the future temperature of photovoltaic modules improved, but the role of local microclimate regulation is also clearly revealed, providing data support for assessing the effectiveness of ecological restoration.

[0036] Optionally, the system further includes: Collect the latest operating data within the time period, including the photovoltaic module's operating records, fault event records, operating characteristic parameters, and abnormal temperature fluctuation characteristics; Based on the latest operational data, the correlation between the basic environmental component and the ecological correction component is reconstructed to generate a dynamic correlation structure; Based on the dynamic correlation structure, the intensity of the ecological correction factor is dynamically adjusted to generate an updated ecological correction factor. The updated ecological correction factor is used to collaboratively optimize the calculation logic of the temperature baseline prediction unit and the microclimate effect quantification unit to generate updated temperature trend prediction calculation rules.

[0037] Specifically, the process begins with periodically collecting the latest operational data over a given period. This data constitutes a comprehensive description of the current health status and operational characteristics of the photovoltaic power plant. This includes operational records showing photovoltaic module power generation and conversion efficiency; fault event records documenting fault types and occurrence times such as hot spots, arcs, and short circuits; operational characteristic parameters reflecting module degradation and pollution loss; and abnormal temperature fluctuations identified from photovoltaic surface temperature data using anomaly detection algorithms that cannot be explained by normal environmental changes. Next, based on this latest operational data, the correlation between the basic environmental components and the ecological correction components is reconstructed. This process utilizes machine learning algorithms, such as incremental learning or transfer learning, to analyze the potential patterns revealed in the latest operational data. For example, when modules exhibit specific early fault characteristics, their temperature response to the basic environmental components becomes more sensitive, or the cooling effect of the ecological correction components is weakened. This analysis updates and generates a dynamic correlation structure, which can be an updated function model that more accurately describes the complex mechanism by which the macro-environment and micro-ecology interact with module temperature under the current module health state. Internally, it encapsulates dynamic mapping relationships or correction algorithms learned from the latest operational data between photovoltaic module health states (such as module aging degree, failure event frequency, or abnormal temperature fluctuation characteristics) and ecological correction components (such as soil moisture and vegetation cover) and cooling efficiency. Based on this dynamic correlation structure, the intensity of the ecological correction factor is dynamically adjusted. This means that the internal parameters of the ecological correction factor are no longer fixed when calculating it. For example, in the calculation formula of the ecological correction factor, the weight coefficients representing the contribution of soil moisture and vegetation cover will be adjusted according to the dynamic correlation structure to generate an updated ecological correction factor. These two coefficients are calculated based on the latest operating data, especially information such as fault event records and component aging status, through the dynamic correlation structure, reflecting the actual effectiveness of the ecological cooling effect under the current operating conditions. For example, when the latest operating data reveals that a specific photovoltaic module has slight aging or local hot spot risk, it may lead to a decrease in its sensitivity to ecological cooling. At this time, the dynamic correlation structure will output a correction factor accordingly to reduce the impact of ecological cooling. and The weights of the ecological correction factor are used to reflect the reduction in the effectiveness of the specific component due to ecological cooling. Finally, this updated ecological correction factor is used to collaboratively optimize the computational logic of the temperature baseline prediction unit and the microclimate effect quantification unit. This is not a simple parameter replacement, but rather the use of the updated ecological correction factor as a new constraint or feature input to retrain or fine-tune the models of these two core prediction units. This optimization process ensures that the prediction of the basic temperature trend and the quantification of local microclimate effects can synchronously adapt to the latest state of the system, ultimately generating more accurate and realistic updated temperature trend prediction calculation rules. Figure 3As shown in the figure, subplot (a) compares the predicted temperature with the actual temperature before and after optimization, with the shaded area representing the improvement. Subplot (b) compares the prediction error before and after optimization. By introducing a closed-loop adaptive learning mechanism, the temperature prediction system is endowed with the ability to continuously evolve, reducing the problem of decreased prediction accuracy of static models when facing photovoltaic module aging, performance degradation, and seasonal environmental changes. This enhances the long-term effectiveness and high accuracy of temperature trend prediction, thereby improving the reliability and foresight of fire risk early warning.

[0038] Optionally, the fire risk assessment module includes: The real-time status acquisition unit is used to acquire the real-time current parameters and real-time voltage parameters of the photovoltaic module and generate the current operating status parameters. An initial risk calculation unit is used to perform a fusion calculation based on the temperature trend prediction value and the current working status parameters to calculate the initial risk value. The risk normalization processing unit is used to obtain historical risk benchmarks and normalize the initial risk value using the historical risk benchmarks to generate a fire risk index.

[0039] Specifically, the electrical performance of the photovoltaic modules is first continuously monitored by a real-time status acquisition unit. High-precision sensors deployed in the photovoltaic string combiner box or the DC side of the inverter collect real-time current and voltage parameters from each monitoring unit. These two parameters are integrated into current operating status parameters, directly reflecting the current electrical load level and electrical stress of the modules. Next, an initial risk calculation unit performs a risk coupling assessment based on multi-dimensional information. The initial risk calculation unit receives the temperature trend prediction value output by the temperature trend prediction module, as well as the current operating status parameters generated by the real-time status acquisition unit. To unify parameters with different physical dimensions within the risk assessment framework, a multi-factor risk assessment model is used to calculate the initial risk value, comprehensively considering the contributions of temperature and electrical power to the risk. For calculating the initial risk value... ,have: ; in, This is a temperature trend forecast; The real-time power is calculated based on real-time current and real-time voltage parameters. and These are preset reference benchmark values, such as the rated operating temperature and rated power of photovoltaic modules under standard test conditions, used for parameter normalization. The basic risk coefficient is determined through statistical analysis and failure mode and impact analysis of historical failure data; and The index weights, representing the sensitivity of temperature and power to risk impact, can be calibrated through statistical analysis and failure mode and effects analysis of historical fault data. This risk assessment model reflects the synergistic effect of temperature and electrical load; that is, the risk posed by temperature rise under high power output is far greater than the same temperature rise under low power. Finally, the risk normalization unit standardizes the initial risk values ​​to make them more universal and comparable, generating the final fire risk index. First, a historical risk benchmark is obtained from the historical database. This benchmark is obtained through statistical analysis of long-term operating data of the photovoltaic field, defining the risk value distribution range under normal operating conditions, such as using the minimum and maximum values ​​or specific quantiles of historical risk values ​​as benchmarks. Subsequently, using this historical risk benchmark, normalization methods such as linear mapping are used to transform the fluctuating initial risk values ​​into a fixed, intuitive range, such as 0 to 100, thereby generating the fire risk index. The fire risk index provides a standardized decision-making basis for subsequent risk classification and early warning control. Figure 4 As shown, a three-dimensional surface illustrates how the fire risk index changes with predicted temperature and current power. Darker colors represent higher fire risk index values. The diagram uses a fire risk index of less than 60 as the "safe" range; 60-80 as the "caution" range; 80-95 as the "warning" range; and above 95 as the "danger" range. By deeply coupling forward-looking temperature predictions with real-time electrical load status, a dynamic, multi-dimensional fire risk assessment model is constructed. Compared to traditional alarm mechanisms that rely solely on static temperature thresholds, this model can more accurately identify high-risk conditions such as "high temperature and high load," realizing a shift from monitoring a single physical quantity to comprehensive risk assessment, thus improving the accuracy of fire risk identification and the effectiveness of early warning.

[0040] Optionally, the early warning control generation module includes: A risk threshold acquisition unit is used to acquire a multi-level risk threshold system, wherein the multi-level risk threshold system includes at least two different levels of risk response thresholds; The risk level determination unit is used to compare the fire risk index with the multi-level risk threshold system step by step to determine the risk level; The early warning signal generation unit is used to perform signal mapping according to the risk level and generate an early warning control signal containing corresponding control instructions.

[0041] Specifically, the first step is to acquire a pre-built, multi-level risk threshold system stored in a configuration library. This system is not a single alarm threshold but includes at least two different levels of risk response thresholds, defining four risk levels: "Safe," "Caution," "Warning," and "Danger." Each level corresponds to a threshold range for a fire risk index. For example, a fire risk index below 60 is defined as "Safe," between 60 and 80 as "Caution," between 80 and 95 as "Warning," and above 95 as "Danger." These thresholds form the benchmark for risk classification. The fire risk index, calculated in real-time by the fire risk assessment module, is then compared level by level with this multi-level risk threshold system, starting from the highest risk level and proceeding downwards. The system checks whether the current fire risk index exceeds the lower limit of the "Danger" level. If it does, the risk level is determined to be "Danger"; otherwise, it checks whether it exceeds the lower limit of the "Warning" level, and so on, until a unique risk level is determined. If the index is below the threshold of the lowest level, the status is determined to be "Safe." Once a risk level is determined, a signal mapping is performed based on that level to generate an early warning control signal. This mapping relationship is predefined in the system control logic. Each risk level corresponds to a specific set of control instructions and information content. For example, the "Caution" level might map to a signal containing the text message "Recommend strengthening inspection" and a log recording instruction; the "Warning" level might generate a signal containing instructions to "Execute power limitation" and "Activate Level 1 cooling"; the "Danger" level would generate a signal containing the highest priority instructions such as "Emergency output cut-off" and "Activate full fire response"; and the "Safe" level would operate normally without any restrictions. The final generated early warning control signal is a structured data packet that explicitly contains the identified risk level and the specific set of control instructions bound to it. This early warning control signal will be sent to the linkage response control module to trigger subsequent physical operations. By establishing a multi-level risk threshold system, refined and hierarchical management of fire risks is achieved. Different levels of intervention measures can be taken according to the severity of the risk, reducing overreactions at the early stages of a risk and causing unnecessary power generation losses, thus improving the intelligence level and operational economy of the entire fire temperature control system.

[0042] Optionally, the multi-level risk threshold acquisition system includes: Based on historical operational data and fire incident records of photovoltaic power plants, the types of photovoltaic modules, and the characteristics of the installation environment, a risk level sample library is constructed. Cluster analysis is performed on the risk level sample library to divide it into multiple risk intervals, forming a multi-level risk threshold system. The risk threshold system is dynamically calibrated based on seasonal changes and component aging status.

[0043] Specifically, the first step is to construct a comprehensive risk level sample library based on historical data. This library's data sources include long-term historical operational data from photovoltaic (PV) sites, such as temperature, power, and environmental parameters, as well as detailed fire event records containing the status of various parameters prior to the event. Simultaneously, it integrates basic information about PV modules, such as module type, manufacturer, and rated parameters, as well as the characteristics of their installation environment, such as installation tilt angle, ventilation conditions, and surrounding obstructions. By associating and labeling this multi-dimensional data, each historical time point is assigned a risk label, such as categorized as safe, low-risk, medium-risk, or high-risk based on whether a fire occurred and the severity of the event, thus constructing the risk level sample library. Next, cluster analysis is performed based on this constructed risk level sample library. Cluster analysis is an unsupervised learning method designed to discover natural group structures within data. Algorithms such as K-Means or DBSCAN are used to calculate and analyze the distribution of fire risk indices for all samples in the risk level sample library. The algorithm automatically groups samples with similar risk index distributions together, forming different clusters. Each cluster represents a risk interval, and the boundaries between clusters naturally constitute the risk level divisions. By interpreting and calibrating these clusters, multiple risk intervals can be defined, thus forming an initial multi-level risk threshold system. Finally, to ensure the long-term effectiveness of the threshold system, a dynamic calibration mechanism is introduced. This mechanism takes into account the significant impact of seasonal changes on the thermal performance of photovoltaic modules, as well as the aging and degradation effects of modules over time. It automatically triggers the calibration process periodically, such as quarterly or annually. During the calibration process, the latest operational and environmental data can be used to reassess the risk distribution characteristics under the current season and module aging conditions. For example, in summer, when the overall temperature is high, the risk threshold may need to be appropriately increased to avoid frequent false alarms; while for severely aged modules, their heat resistance performance decreases, and the corresponding risk threshold needs to be decreased. Through this dynamic calibration, the multi-level risk threshold system is continuously updated, improving the accuracy of reflecting the true risk status of photovoltaic fields at different life stages and under external environments. By constructing and maintaining a multi-level risk threshold system through a closed-loop process of data-driven, intelligent analysis and dynamic calibration, the scientificity and adaptability of risk classification standards are improved. At the same time, the dynamic calibration mechanism enables the system to evolve on its own and proactively adapt to dynamic changes such as seasonal changes and equipment aging, reducing the impact of the accuracy of early warning judgments decaying with time and environmental changes.

[0044] Optionally, the linkage response control module includes: The power gradient control unit is used to match the corresponding power reduction coefficient according to the risk level in the warning control signal and gradually reduce the output power of the photovoltaic module. The fire-fighting graded activation unit is used to trigger the fire control unit to perform a tiered fire-fighting response based on the early warning control signal, thereby activating the cooling spray device and the flame retardant spray device in sequence. The response performance verification unit is used to collect power adjustment rate and fire protection execution status data in real time and generate response performance evaluation indicators. The dynamic adjustment unit is used to dynamically adjust the power reduction coefficient and the priority of the fire response based on the matching degree between the decay rate of the fire risk index and the response performance evaluation index.

[0045] Specifically, upon receiving a warning control signal containing the risk level, the power gradient control unit immediately executes a power adjustment strategy. This control unit has a pre-set mapping table between risk levels and power reduction coefficients. For example, a "Caution" level might correspond to a 5% power reduction coefficient, a "Warning" level to 20%, a "Danger" level to 100% (emergency shutdown), and a "Safe" level to 0% (normal operation). Based on the received risk level, the unit automatically matches and applies the corresponding power reduction coefficient, sending control commands to the photovoltaic inverter to gradually reduce or directly cut off the output power of the photovoltaic modules. The adjusted output power is then calculated. ,have: ; in, This represents the current output power. Risk level The corresponding power decline factor; The risk level is determined by a gradient power control system designed to proactively suppress further temperature increases by reducing the electrical heat generated by components. Simultaneously, the fire-fighting graded activation unit, based on the warning control signal, triggers a tiered fire response, controlling different types of fire-fighting equipment within the site. Under the warning control signal, at lower risk levels, such as the "warning" level, only cooling spray devices near high-temperature components may be activated to absorb heat through water evaporation, providing physical cooling. When the risk level escalates to "danger," flame retardant spray devices are further triggered to actively cover the component surface, chemically and physically isolating oxygen and heat sources to prevent the occurrence or spread of fire. This graded activation strategy improves the rational use of fire-fighting resources, achieving a balance between cost-effectiveness and safety. To assess the effectiveness of the control measures, a response performance verification unit operates in parallel. This unit collects real-time data on the actual rate and magnitude of power adjustments, as well as fire-fighting execution status data, such as whether sprinkler heads are operating normally and whether the spray flow rate meets standards, through a data acquisition interface. Based on this real-time feedback data, a series of response performance evaluation indicators are calculated and generated, such as power reduction response time, deviation between actual power reduction ratio and command value, and success rate of fire-fighting device operation. These response performance evaluation indicators quantify the execution effect of the linkage response control. Finally, the dynamic adjustment unit constitutes the intelligent core of the entire response control. The dynamic adjustment unit continuously monitors the changing trend of the fire risk index, especially its decay rate after control measures are implemented. Simultaneously, it performs a matching analysis between this decay rate and the response performance evaluation indicators generated by the response performance verification unit. If it finds that the risk index is decreasing slowly, but the response performance evaluation indicators show that power adjustment and fire-fighting execution are both in place, it will determine that the current control strength is insufficient. Based on this judgment, the dynamic adjustment unit will dynamically adjust the power reduction coefficient and the priority of the fire response, such as automatically increasing the initial value of the power reduction coefficient in the next warning of the same level, or advancing the triggering time of the fire response. The adjusted power reduction coefficient is calculated... Priority of adjusted fire response ,have: ; in, The target risk decay rate; The risk index when control measures are implemented; for Risk index after a certain period of time; The observation time interval; It is a function for finding the extreme values; The learning rate coefficient is usually determined through simulation debugging or historical response data playback optimization, and then fine-tuned through actual operation feedback; The fire response triggering timing before optimization; To adjust the step size coefficient, the setting depends on the system's sensitivity analysis to power control deviation and fire response delay, and is usually determined through control loop simulation or field tests; The deviation threshold is used. Through this continuous self-assessment and dynamic adjustment, the response strategy is learned and optimized online, abandoning the single, passive response mode of traditional fire protection systems. Instead, it adopts an advanced control strategy that is proactive, hierarchical, and capable of self-optimization, reducing unnecessary power generation losses and resource waste. The control strategy can be continuously iterated and improved based on the actual handling effect.

[0046] Based on the same inventive concept, such as Figure 5 As shown, the present invention also provides a photovoltaic fire-fighting temperature control prediction method, the method comprising: Acquire photovoltaic surface temperature data, environmental monitoring data, and ecological restoration monitoring data of photovoltaic modules to generate a multi-source monitoring dataset; Based on the ecological restoration monitoring data, quantitative analysis is performed to generate ecological correction factors, and the ecological correction factors are used to fuse the multi-source monitoring dataset to generate comprehensive environmental parameters. Based on the comprehensive environmental parameters, a structured prediction process is performed to calculate the predicted temperature trend value. Obtain the current operating status parameters of the photovoltaic module, and combine them with the temperature trend prediction value to perform a risk coupling assessment and calculate and generate a fire risk index. Based on the fire risk index, a risk classification judgment is made, and an early warning control signal is generated. The power output of the photovoltaic module is adjusted based on the early warning control signal, and the fire control unit is linked to respond.

[0047] To verify the feasibility of this invention in practice, it was applied to a large-scale ground-mounted photovoltaic power station. This power station, located in an ecological restoration area, faces the problem that traditional temperature control systems cannot accurately assess the local microclimate cooling effect caused by vegetation restoration, leading to inaccurate temperature predictions and misjudgments of fire risk. The power station aims to utilize this invention to achieve accurate prediction of photovoltaic module temperature and proactive, tiered fire risk control by integrating ecological monitoring data.

[0048] In this embodiment, the photovoltaic power station deployed the photovoltaic fire-fighting temperature control and prediction system described in this invention. Thermocouple temperature sensors were deployed on the photovoltaic modules; environmental sensors were installed within the site to collect ambient temperature, humidity, wind speed, and total solar radiation intensity; and soil moisture sensors and spectral sensors for calculating vegetation cover were deployed in the ecological restoration area between the photovoltaic arrays. The experiment lasted for four months, covering the climate change cycle from mid-spring to late summer, and the system continuously monitored, predicted, and verified the operation status of the entire power station.

[0049] On a typical summer afternoon at 14:00 on July 15, 2025, the site's environmental data acquisition unit measured an ambient temperature of 36℃ and a total solar radiation intensity of 980W / m². 2 The system focused on monitoring two areas with significant ecological differences: Area A and Area B. Area A had high vegetation coverage (85%) and soil moisture (40%), with an ecological correction factor of 0.92 calculated by the ecological factor quantification unit, indicating a significant cooling effect. Area B, a recently constructed area, had sparse vegetation (15%) and soil moisture of only 10%, with a calculated ecological correction factor of 0.98, indicating a weak cooling effect. After correction using this factor, the environmental parameter fusion module predicted the component temperature 30 minutes later using the temperature trend prediction module. For Area A, the predicted temperature was 78℃; while for Area B, the predicted temperature was as high as 86℃. Actual measurement results showed that the component temperature in Area A was 79℃, and in Area B it was 87℃, with prediction errors within 2%, demonstrating that introducing the ecological correction factor can improve the accuracy of predictions.

[0050] At 14:30 on the same day, the photovoltaic modules in Area B, due to continuous high temperature and high radiation operation, were detected by the real-time status acquisition unit to be operating at near full power. The fire risk assessment module, combining the predicted temperature trend of 86℃ and high-power operating parameters, calculated a fire risk index of 88 using a risk coupling assessment model. After obtaining this index, the early warning control generation module compared it with a multi-level risk threshold system, such as: <60 for safe level, 60-80 for caution level, 80-95 for warning level, and >95 for danger level, determining the current risk level as "warning". Subsequently, the system generated an early warning control signal containing corresponding control instructions. Upon receiving the signal, the linkage response control module matched the 20% power reduction coefficient corresponding to the "warning" level and issued a command to the inverter in that area to reduce the output power of the photovoltaic modules by 20%; simultaneously, the fire-fighting classification activation unit triggered the cooling spray device in that area for active physical cooling.

[0051] The system also demonstrates its dynamic adaptive optimization capability. When operating to the 4th month, the system collected the latest operation data of some photovoltaic modules put into operation in the early stage, and found that their operation characteristic parameters have slightly attenuated, and the occurrence frequency of abnormal temperature fluctuation characteristics has increased slightly. Based on these data, the system reconstructed the correlation between the basic environmental component and the ecological correction component, and generated a dynamic correlation structure. This structure shows that these aged modules have increased sensitivity to high temperature, and the cooling effect of the ecological correction factor is weakened on their surface. Therefore, the system dynamically adjusted the action intensity of the ecological correction factor applied to these modules, and collaboratively optimized the temperature prediction calculation rule. After optimization, under a similar environmental condition, the temperature prediction value of the system for the aged modules is increased compared with that before optimization, which is closer to the actual monitoring value, thereby realizing earlier risk warning.

[0052] From the perspective of data comparison, the prediction system of the present invention has advantages in the accuracy of fire risk identification and the initiative of response. After introducing ecological data, the prediction accuracy of the system for module temperature is improved compared with the traditional model, especially in areas with significant microclimate differences. The fire risk index can comprehensively reflect the superimposed risk of "high temperature" and "high load", and its effectiveness of risk identification is higher than that of single temperature threshold alarm. The hierarchical linked response control strategy, on the premise of ensuring safety, reduces the power generation loss caused by excessive response compared with the "one-size-fits-all" emergency shutdown scheme. The adaptive optimization mechanism of the system improves its capability of maintaining high-precision prediction throughout the entire life cycle of the photovoltaic power station.

[0053] It should be noted that the electrical connections between the above-mentioned various units do not necessarily represent direct connection of lines, and indirect connection methods are applicable to the embodiments of the present invention as long as the object of the present invention is achieved. What is described above is only exemplary embodiments of the present invention, and the scope of the present invention cannot be limited thereby.

[0054] That is, all equivalent changes and modifications made in accordance with the teachings of the present invention still fall within the scope covered by the present invention. Those skilled in the art will readily conceive of other embodiments of the present invention after considering the description and the disclosure of the practice of the present invention. The present application is intended to cover any variations, uses, or adaptive changes of the present invention, and these variations, uses, or adaptive changes follow the general principles of the present invention and include common general knowledge or conventional technical means in the technical field that are not recorded in the present invention.

Claims

1. A photovoltaic fire-fighting temperature control and prediction system, characterized in that, The system includes: The multi-source data acquisition module is used to acquire photovoltaic surface temperature data, environmental monitoring data, and ecological restoration monitoring data of photovoltaic modules, and generate a multi-source monitoring dataset. The ecological restoration monitoring data includes soil moisture and vegetation coverage. The environmental parameter fusion module is used to perform quantitative analysis based on the ecological restoration monitoring data, generate ecological correction factors, and use the ecological correction factors to fuse the multi-source monitoring dataset to generate comprehensive environmental parameters. The temperature trend prediction module is used to perform structured prediction processing based on the comprehensive environmental parameters and calculate the predicted temperature trend value. The fire risk assessment module is used to obtain the current operating status parameters of the photovoltaic module, and combine them with the temperature trend prediction value to perform risk coupling assessment and calculate and generate a fire risk index. The early warning control generation module is used to determine the risk level based on the fire risk index and generate an early warning control signal. The linkage response control module is used to adjust the power output of the photovoltaic module based on the early warning control signal and to link the fire control unit to respond. The environmental parameter fusion module includes: The initial parameter fusion unit is used to standardize and weight the photovoltaic surface temperature data and the environmental monitoring data in the multi-source monitoring dataset to generate initial fusion parameters. An ecological factor quantification unit is used to quantify the local environmental cooling effect based on the ecological restoration monitoring data and generate ecological correction factors. An environmental parameter correction unit is used to perform correction calculations on the initial fusion parameters based on the ecological correction factor to obtain comprehensive environmental parameters. The temperature trend prediction module includes: An environmental parameter decomposition unit is used to decompose the comprehensive environmental parameters to obtain basic environmental components and ecological correction components. A temperature baseline prediction unit is used to predict the temperature baseline based on the basic environmental components and calculate the basic temperature trend. The microclimate effect quantification unit is used to quantify the local microclimate effect based on the ecological correction component and generate a temperature correction amount. The temperature prediction synthesis unit is used to perform joint analysis and calculation by combining the basic temperature trend and the temperature correction amount to obtain the temperature trend prediction value.

2. The photovoltaic fire-fighting temperature control and prediction system according to claim 1, characterized in that, The multi-source data acquisition module includes: The module temperature acquisition unit is used to acquire temperature data through temperature sensors deployed on the photovoltaic module, and to obtain photovoltaic surface temperature data as the first temperature data. The field environment data acquisition unit is used to collect ambient temperature, air humidity, wind speed and total solar radiation intensity through environmental sensors deployed in the photovoltaic field to form environmental monitoring data, wherein the ambient temperature is used as the second temperature data. The ecological restoration data collection unit is used to collect ecological restoration monitoring data, including soil moisture and vegetation cover, through in-situ monitoring using ecological sensors deployed in the ecological restoration area. The multi-source data integration unit is used to align and integrate the first temperature data, the environmental monitoring data, and the ecological restoration monitoring data with timestamps to generate a multi-source monitoring dataset.

3. The photovoltaic fire-fighting temperature control and prediction system according to claim 1, characterized in that, The system also includes: Collect the latest operating data within the time period, including the photovoltaic module's operating records, fault event records, operating characteristic parameters, and abnormal temperature fluctuation characteristics; Based on the latest operational data, the correlation between the basic environmental component and the ecological correction component is reconstructed to generate a dynamic correlation structure; Based on the dynamic correlation structure, the intensity of the ecological correction factor is dynamically adjusted to generate an updated ecological correction factor. The updated ecological correction factor is used to collaboratively optimize the calculation logic of the temperature baseline prediction unit and the microclimate effect quantification unit to generate updated temperature trend prediction calculation rules.

4. The photovoltaic fire-fighting temperature control and prediction system according to claim 1, characterized in that, The fire risk assessment module includes: The real-time status acquisition unit is used to acquire the real-time current parameters and real-time voltage parameters of the photovoltaic module and generate the current operating status parameters. An initial risk calculation unit is used to perform a fusion calculation based on the temperature trend prediction value and the current working status parameters to calculate the initial risk value. The risk normalization processing unit is used to obtain historical risk benchmarks and normalize the initial risk value using the historical risk benchmarks to generate a fire risk index.

5. The photovoltaic fire-fighting temperature control and prediction system according to claim 1, characterized in that, The early warning control generation module includes: A risk threshold acquisition unit is used to acquire a multi-level risk threshold system, wherein the multi-level risk threshold system includes at least two different levels of risk response thresholds; The risk level determination unit is used to compare the fire risk index with the multi-level risk threshold system step by step to determine the risk level; The early warning signal generation unit is used to perform signal mapping according to the risk level and generate an early warning control signal containing corresponding control instructions.

6. A photovoltaic fire-fighting temperature control and prediction system according to claim 5, characterized in that, The multi-level risk threshold acquisition system includes: Based on historical operational data and fire incident records of photovoltaic power plants, the types of photovoltaic modules, and the characteristics of the installation environment, a risk level sample library is constructed. Cluster analysis is performed on the risk level sample library to divide it into multiple risk intervals, forming a multi-level risk threshold system. The risk threshold system is dynamically calibrated based on seasonal changes and component aging status.

7. A photovoltaic fire-fighting temperature control and prediction system according to claim 5, characterized in that, The linkage response control module includes: The power gradient control unit is used to match the corresponding power reduction coefficient according to the risk level in the warning control signal and gradually reduce the output power of the photovoltaic module. The fire-fighting graded activation unit is used to trigger the fire control unit to perform a tiered fire-fighting response based on the early warning control signal, thereby activating the cooling spray device and the flame retardant spray device in sequence. The response performance verification unit is used to collect power adjustment rate and fire protection execution status data in real time and generate response performance evaluation indicators. The dynamic adjustment unit is used to dynamically adjust the power reduction coefficient and the priority of the fire response based on the matching degree between the decay rate of the fire risk index and the response performance evaluation index.

8. A photovoltaic fire-fighting temperature control prediction method, employing a photovoltaic fire-fighting temperature control prediction system as described in any one of claims 1-7, characterized in that, The method includes: Acquire photovoltaic surface temperature data, environmental monitoring data, and ecological restoration monitoring data of photovoltaic modules to generate a multi-source monitoring dataset. The ecological restoration monitoring data includes soil moisture and vegetation coverage. Based on the ecological restoration monitoring data, quantitative analysis is performed to generate ecological correction factors, and the ecological correction factors are used to fuse the multi-source monitoring dataset to generate comprehensive environmental parameters. Based on the comprehensive environmental parameters, a structured prediction process is performed to calculate the predicted temperature trend value. Obtain the current operating status parameters of the photovoltaic module, and combine them with the temperature trend prediction value to perform a risk coupling assessment and calculate and generate a fire risk index. Based on the fire risk index, a risk classification judgment is made, and an early warning control signal is generated. The power output of the photovoltaic module is adjusted based on the early warning control signal, and the fire control unit is activated in response. The environmental parameter fusion module includes: The initial parameter fusion unit is used to standardize and weight the photovoltaic surface temperature data and the environmental monitoring data in the multi-source monitoring dataset to generate initial fusion parameters. An ecological factor quantification unit is used to quantify the local environmental cooling effect based on the ecological restoration monitoring data and generate ecological correction factors. An environmental parameter correction unit is used to perform correction calculations on the initial fusion parameters based on the ecological correction factor to obtain comprehensive environmental parameters. The temperature trend prediction module includes: An environmental parameter decomposition unit is used to decompose the comprehensive environmental parameters to obtain basic environmental components and ecological correction components. A temperature baseline prediction unit is used to predict the temperature baseline based on the basic environmental components and calculate the basic temperature trend. The microclimate effect quantification unit is used to quantify the local microclimate effect based on the ecological correction component and generate a temperature correction amount. The temperature prediction synthesis unit is used to perform joint analysis and calculation by combining the basic temperature trend and the temperature correction amount to obtain the temperature trend prediction value.

Images