Back gain analysis method and system for double-sided photovoltaic module

Abstract

The invention relates to the field of double-sided photovoltaic, in particular to a back gain analysis method and system for a double-sided photovoltaic module. The method comprises the steps of obtaining a multi-source space-time dynamic data set of a target water body, performing water body reflection characteristic multi-dimensional diagnosis and site selection decision based on the multi-source space-time dynamic data set, and generating a site selection decision support information set; based on the site selection decision support information set, photovoltaic array geometric parameter multi-target collaborative optimization and shadow avoidance analysis facing back gain maximization are carried out, and an optimized array layout parameter set is generated; and based on the optimized array layout parameter set, performing integrated back gain quantitative evaluation and prospective strategy generation, and generating and outputting a double-sided photovoltaic back gain analysis report. In the double-sided photovoltaic back gain analysis process, the real power generation gain of the double-sided assembly on the water surface can be accurately predicted, and fine design of the photovoltaic array on geometric parameters and layout is supported.

Description

Technical Field

[0001] This application relates to the field of bifacial photovoltaics, and in particular to a method and system for back-side gain analysis of bifacial photovoltaic modules. Background Technology

[0002] In the field of bifacial photovoltaic power generation technology for water-facing deployment, bifacial photovoltaic modules, with their ability to generate electricity by reflecting light from their back side, have become a key technological path to increase power generation per unit area and reduce the levelized cost of electricity (LCOE), and have core value in promoting the transformation of the global energy structure and the intensive development of large-scale photovoltaic power plants.

[0003] However, existing back-side gain assessment methods for water body reflection environments typically treat water bodies as simple planes with fixed reflectivity, lacking synergistic integration and systematic quantitative analysis of multidimensional complex factors. This leads to a serious disconnect between the assessment model and the actual physical environment, making it impossible to accurately predict the actual power generation gain of bifacial modules on water surfaces, and also making it difficult to support the refined and optimized design of photovoltaic arrays in terms of geometric parameters and layout, thus limiting the upper limit of efficiency and return on investment of water surface photovoltaic power plants. Summary of the Invention

[0004] This application provides a method and system for back-side gain analysis of bifacial photovoltaic modules to solve the above-mentioned technical problems.

[0005] In a first aspect, this application provides a method for back-side gain analysis of bifacial photovoltaic modules, the method comprising: A multi-source spatiotemporal dynamic dataset of the target water body is acquired. Based on this dataset, multi-dimensional diagnosis of water body reflection characteristics and site selection decision are performed, generating a site selection decision support information set. Based on this support information set, multi-objective collaborative optimization of photovoltaic array geometric parameters and shadow avoidance analysis are performed to maximize back-side gain, generating an optimized array layout parameter set. Based on this optimized array layout parameter set, integrated back-side gain quantification and forward-looking strategy generation are performed, generating and outputting a bifacial photovoltaic back-side gain analysis report.

[0006] The above technical solutions elevate the evaluation of bifacial photovoltaic systems on water from extensive experience to full-chain quantitative analysis. Through precise diagnosis of reflective sources, customized array optimization, and integrated risk assessment, the accuracy of power generation prediction and project investment returns are systematically improved, providing a reliable scientific basis for project decision-making.

[0007] Optionally, the process of generating the site selection decision support information set includes: the multi-source spatiotemporal dynamic dataset includes water body substrate optical characteristics, water quality spatiotemporal variation characteristics, and water surface dynamic characteristics; based on the substrate optical characteristics and the water quality spatiotemporal variation characteristics, the inherent reflection and attenuation characteristics of the water body are analyzed to generate water body reflection characteristic analysis information; based on the water surface dynamic characteristics, the modulation effect of dynamic waves on water surface reflectivity is quantified to generate water surface reflectivity modulation information; and by combining the water body reflection characteristic analysis results and the water surface reflectivity modulation coefficient, the reflection efficiency and engineering suitability of multiple candidate water bodies are compared to generate the site selection decision support information set.

[0008] Optionally, the analysis of the inherent reflection and attenuation characteristics of the water body includes: identifying the material composition and color of the bottom and shoreline of the water body by analyzing the optical characteristics of the water body substrate, distinguishing between high-reflectivity substrates and low-reflectivity substrates, to determine whether the water body belongs to an ordinary lake or a high-reflectivity salt lake / crystallization lake, and obtaining a substrate type determination result; extracting the seasonal variation patterns and peak characteristics of water turbidity and chlorophyll concentration by analyzing the time series data of at least one complete year in the spatiotemporal variation characteristics of the water quality; coupling the substrate type determination result with the seasonal variation pattern, evaluating the attenuation of water body transmittance caused by changes in suspended matter concentration and periodic algal proliferation and its inherent weakening effect on reflected light intensity, and generating the water body reflection characteristic analysis information.

[0009] Optionally, the modulation effect of the quantified dynamic waves on water surface reflectivity includes: analyzing continuous monitoring data in the water surface dynamic characteristics to statistically analyze the distribution of water surface wave height and wave period characteristics under different wind levels; establishing a reflection path analysis model for incident sunlight on calm and undulating water surfaces based on the principle of solar position geometric optics; analyzing and comparing the proportion of effective reflected light flux loss that can be received by the back of the photovoltaic module caused by the dispersion of specular reflected light direction under different wave conditions, integrating all the effective reflected light flux loss proportions, and generating the water surface reflectivity modulation information.

[0010] Optionally, the comparison of the reflection efficiency and engineering suitability of multiple candidate water bodies includes: based on the water body reflection characteristic analysis information and the water surface reflectivity modulation information, coupling the apparent solar motion law and the seasonal characteristics of the water body, extracting and calculating the effective water surface reflectivity of each candidate water body under several representative seasons and several typical solar altitude angle scenarios within each season, generating a dynamic reflection efficiency map of each water body; combining the spatial geographic information of each water body with preset engineering constraints, conducting a preliminary screening of engineering feasibility, generating an engineering suitability classification map; performing spatial overlay analysis of the dynamic reflection efficiency map of each water body and the engineering suitability classification map, and quantitatively evaluating and ranking the comprehensive development potential of multiple candidate water bodies and their different sub-regions through weighted scoring or multi-criteria decision-making algorithms; based on the quantitative evaluation and ranking results, using them as the site selection decision support information set, which clearly identifies the water body with the best comprehensive conditions and its internal recommended priority construction areas.

[0011] Optionally, the process of generating the optimized array layout parameter set includes: extracting the dynamic reflectivity spectrum from the site selection decision support information set; converting the spatiotemporal reflectivity distribution of the water surface represented by this spectrum into reflected irradiance data that can be received on the back of the photovoltaic array based on physical optical transmission simulation technology; establishing a comprehensive optimization function for the components with the core optimization variables of component installation height, tilt angle, and north-south spacing of the array, aiming to maximize the back irradiance reception, minimize the front shading loss, and optimize the energy output per unit water surface area; using a multi-objective optimization algorithm to iteratively solve the comprehensive optimization function for the components, obtaining several recommended combinations of geometric parameters, and performing key period shadow analysis on each combination to ensure that the back shadow shading rate is lower than a preset shading threshold, and finally integrating the optimized array layout parameter set.

[0012] Optionally, the conversion into reflected irradiance intensity data that can be received on the back of the photovoltaic array includes: based on the dynamic reflectivity spectrum, analyzing the effective reflectivity matrix of the water surface under different representative seasons and typical solar altitude angles using a spatial interpolation algorithm; calculating the spatial distribution of reflected radiation intensity corresponding to the water surface based on the effective reflectivity matrix of the water surface and the solar direct irradiance calculated by the solar position astronomical algorithm; using the spatial distribution of reflected radiation intensity as the light source input, simulating the propagation path of reflected light in the three-dimensional space of the photovoltaic module array using ray tracing or geometric optics models, and calculating the reflected irradiance intensity data that can be received on the back of the photovoltaic array using the receiving plane on the back of each module as the target surface.

[0013] Optionally, the step of performing key-period shadow analysis for each combination to ensure that the back shadow shading rate is below a threshold includes: for each recommended combination of geometric parameters, constructing a three-dimensional geometric model of the photovoltaic array based on the component installation height, tilt angle, and north-south spacing of the array; selecting several key days that have a significant impact on the annual power generation of the system, and calculating the typical solar hour angle that may produce significant front-row shading effects on the key days; using the three-dimensional geometric model, calculating and visualizing the shadow projection of each combination of geometric parameters under the typical solar hour angle, quantitatively analyzing the proportion of the area of ​​the back row components that is shaded by the front row components or supporting structure, i.e., the back shadow shading rate; and selecting combinations among all combinations of geometric parameters where the back shadow shading rate is below the preset shading threshold as feasible solutions through shadow avoidance analysis, and integrating them into the optimized array layout parameter set.

[0014] Optionally, the process of generating the bifacial photovoltaic back-side gain analysis report includes: using the optimized array layout parameter set as the core configuration, performing hourly performance simulation of the photovoltaic system throughout the year, quantifying the additional power generation obtained by the bifacial photovoltaic modules due to water surface reflection, i.e., the precise back-side gain value; analyzing the sensitivity of the precise back-side gain value to key input parameters such as reflectivity and installation height, identifying the main sources of gain risk and uncertainty; and combining the precise back-side gain value and sensitivity analysis conclusions to generate the bifacial photovoltaic back-side gain analysis report, which includes expected power generation improvement, economic assessment, and array optimization design suggestions.

[0015] Secondly, this application provides a back-side gain analysis system for bifacial photovoltaic modules, the system comprising: The site selection decision module is used to acquire a multi-source spatiotemporal dynamic dataset of the target water body, and based on the multi-source spatiotemporal dynamic dataset, to perform multi-dimensional diagnosis of water body reflection characteristics and site selection decision, generating a site selection decision support information set; the array optimization module is used to perform multi-objective collaborative optimization of photovoltaic array geometric parameters and shadow avoidance analysis oriented towards maximizing back-side gain based on the site selection decision support information set, generating an optimized array layout parameter set; the report generation module is used to perform integrated back-side gain quantitative evaluation and forward-looking strategy generation based on the optimized array layout parameter set, generating and outputting a bifacial photovoltaic back-side gain analysis report. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application; Figure 2 A flowchart illustrating a back-side gain analysis method for bifacial photovoltaic modules, provided as an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a back-side gain analysis system for bifacial photovoltaic modules provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0020] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0021] Existing back-side gain assessment methods for water body reflection environments typically treat water bodies as simple planes with fixed reflectivity, lacking synergistic integration and systematic quantitative analysis of multidimensional complex factors. This leads to a serious disconnect between the assessment model and the actual physical environment, making it impossible to accurately predict the actual power generation gain of bifacial modules on water surfaces, and also making it difficult to support the refined and optimized design of photovoltaic arrays in terms of geometric parameters and layout. This limits the upper limit of efficiency and return on investment for water surface photovoltaic power plants.

[0022] Based on this, this application provides a method and system for back-side gain analysis of bifacial photovoltaic modules. First, it integrates multi-source spatiotemporal data to perform multi-dimensional diagnosis of the water substrate, water quality, and wave dynamics, enabling scientific site selection based on reflection efficiency. Second, based on the dynamic reflection spectrum generated by the site selection, it performs multi-objective collaborative optimization and shadow avoidance analysis of the photovoltaic array's geometric parameters, aiming to maximize back-side gain and minimize shading. Finally, it uses the optimized layout parameters to conduct year-round refined simulation and sensitivity assessment, generating a comprehensive analysis report integrating quantitative benefit and risk strategies, which is then provided to photovoltaic project personnel. This solution elevates the evaluation of bifacial photovoltaic systems on water surfaces from extensive experience to full-chain quantitative analysis. Through precise reflection source diagnosis, customized array optimization, and integrated risk assessment, it systematically improves the accuracy of power generation prediction and project investment returns, providing a reliable scientific basis for project decision-making.

[0023] Figure 1 This is a schematic diagram illustrating an application scenario provided by this application. In the process of gain analysis on the back side of bifacial photovoltaic panels, the method provided in this application can accurately predict the actual power generation gain of bifacial modules on the water surface, supporting the refined design of photovoltaic arrays in terms of geometric parameters and layout.

[0024] Specifically, the method of this application is applied to any server that communicates with a multimodal water sensor group and obtains multi-source spatiotemporal dynamic datasets provided by the multimodal water sensor group through the server.

[0025] For specific implementation details, please refer to the following examples.

[0026] Figure 2 This is a flowchart illustrating a back-side gain analysis method for bifacial photovoltaic modules according to an embodiment of this application. The method of this embodiment can be applied to servers in the above-described scenario. Figure 2 As shown, the method includes: S201. Obtain a multi-source spatiotemporal dynamic dataset of the target water body, and based on the multi-source spatiotemporal dynamic dataset, perform multi-dimensional diagnosis of water body reflection characteristics and site selection decision, and generate a site selection decision support information set.

[0027] The target water body can be a natural or artificial water body whose surface reflectivity is to be evaluated for the construction of a double-sided photovoltaic power station, such as a lake, reservoir, salt lake, or waterlogged area in a coal mining subsidence zone.

[0028] Multi-source spatiotemporal dynamic datasets can be collections of observational data used to comprehensively characterize the optical, water quality, and dynamic properties of water bodies, with the data originating from a multi-modal sensor array for water bodies.

[0029] Multidimensional diagnosis of water body reflectivity and site selection decision can be an integrated analysis process. "Multidimensional diagnosis of water body reflectivity" refers to the systematic assessment of the ability of a water body surface to reflect sunlight from three dimensions: substrate composition (static inherent properties), water quality changes (dynamic attenuation factors), and water surface dynamics (dynamic modulation factors). "Site selection decision" is based on this diagnosis, combined with engineering conditions, to compare and select different candidate water bodies or their sub-regions.

[0030] The site selection decision support information set can be a comprehensive output of information including quantitative assessment results of water body reflection effectiveness, engineering suitability classification, and the final recommended site selection scheme.

[0031] Specifically, bifacial photovoltaic (PV) technology utilizes the back of the module to receive reflected light to enhance power generation, and its back-side gain potential is highly dependent on the reflectivity of the mounting surface. Water surfaces, as a vast and exploitable spatial resource, are an ideal scenario for deploying bifacial PV, with their core value lying in utilizing sunlight reflected from the water surface. However, existing preliminary assessment methods for water-based PV projects have significant flaws: they generally simplify water bodies as homogeneous planes with fixed reflectivity, completely ignoring the differences in optical substrates of different water bodies, seasonal dynamic changes in water quality, and the complex modulation effects of water surface morphology caused by wind and waves on reflectivity. This crude assumption leads to severely distorted back-side gain predictions, a lack of scientific basis for site selection, and decisions often based solely on water area and grid connection conditions, resulting in actual power generation revenue far lower than expected after project completion. This solution addresses this fundamental problem by forcibly introducing and integrating multi-source spatiotemporal dynamic datasets from remote sensing, hydrology, and meteorology to conduct a comprehensive diagnosis of water body reflectivity characteristics, from static substrates to dynamic modulation, and from inherent properties to spatiotemporal changes, and then makes scientific site selection decisions based on this analysis. This established a realistic and reliable physical model of the reflection source for the entire project, ensuring the accuracy of subsequent design and evaluation from the outset.

[0032] S202. Based on the site selection decision support information set, perform multi-objective collaborative optimization and shadow avoidance analysis of the photovoltaic array geometric parameters for maximizing back-side gain, and generate an optimized array layout parameter set.

[0033] The geometric parameters of a photovoltaic array can be key design variables that determine the spatial arrangement of the photovoltaic array. They typically include at least the module installation height (vertical distance from the water surface or ground), module tilt angle (angle between the module plane and the horizontal plane), and array north-south spacing (horizontal distance between the centers of the front and rear rows of modules).

[0034] Multi-objective collaborative optimization can be a process of simultaneously considering multiple, sometimes conflicting, optimization objectives and seeking the optimal balance point.

[0035] Shadow avoidance analysis can specifically refer to the analysis of shadows that may be cast on the back of bifacial modules. Since the back mainly receives reflected light from the front (water surface), the projection of the front modules or support structure may block the back of the rear modules, reducing their gain.

[0036] The optimized array layout parameter set can be a set of one or more sets of technically feasible and high-performance geometric parameter combinations obtained through the above optimization and shadow avoidance analysis.

[0037] Specifically, the geometric layout parameters of photovoltaic arrays are core design variables affecting their energy capture efficiency. In conventional ground-mounted power plants, optimization mainly focuses on direct sunlight and sky-scattered light received on the front of the modules, while avoiding inter-row shading. However, in the case of bifacial photovoltaic systems on water, the energy source becomes exceptionally complex: the back of the modules primarily receives reflected light from the water in front, with specific spatiotemporal distribution and incident direction. This transforms the optimization objective from maximizing the front side to a multi-objective collaborative problem that simultaneously balances back-side gain, front-side shading loss, and energy output per unit area. Existing design tools and methods lack the ability to model this complex coupling relationship, typically relying on experience from ground-mounted power plants or performing single-point trials, failing to achieve a systematically optimal performance solution. Furthermore, the unique shadows cast by the front-row modules and supports onto the back of the rear-row modules in the water-surface scenario are a blind spot not covered by traditional shadow analysis; neglecting them will directly erode back-side gain. This solution addresses this design dilemma by constructing a multi-objective collaborative optimization function centered on back-side gain based on refined reflection source information (site selection decision support information set) provided upstream, and integrating back-side shadow avoidance analysis. This enables deep adaptation and global optimization of array geometry parameters to site-specific reflection conditions, resulting in truly customized high-performance layout solutions.

[0038] S203. Based on the optimized array layout parameter set, perform integrated back-side gain quantification evaluation and forward-looking strategy generation, and generate and output a bifacial photovoltaic back-side gain analysis report.

[0039] Integrated back-side gain quantification assessment can be a comprehensive assessment process. "Integrated" is reflected in the fact that the assessment process not only calculates the gain value, but also integrates sensitivity analysis (assessing the impact of changes in key input parameters on the gain) and uncertainty analysis. "Quantification assessment" refers to using high-performance simulation tools (such as PVsyst based on physical optics models, self-developed simulation programs, etc.) to simulate the operation of the photovoltaic system throughout a typical meteorological year or a specific period, with "optimized array layout parameter set" as input, and accurately calculate the additional power generation caused by water surface reflection (usually expressed as a percentage or absolute power).

[0040] Forward-looking strategy generation can be based on the results of quantitative assessment and sensitivity analysis, and can propose predictive suggestions and countermeasures for future stages such as project planning, financing, construction and operation.

[0041] A bifacial photovoltaic back-side gain analysis report can be a final, comprehensive output document; it is not only a performance prediction report, but also a decision support report.

[0042] Specifically, accurate power generation assessment and risk assessment are the ultimate basis for investment decisions and performance guarantees for photovoltaic projects. Currently, the revenue assessment reports for bifacial photovoltaic projects on water surfaces are often perfunctory, providing only a single gain prediction value based on simplified reflectivity assumptions and standard array configurations. This assessment neither forms a closed-loop verification with the aforementioned refined site selection and customized design, nor does it provide sensitivity analysis for uncertainties in key input parameters. This results in low reliability of the prediction results, making it impossible to effectively identify and manage the project's main technical risks (such as water quality deterioration and reflectivity prediction deviations), and failing to provide reliable decision support for owners and investors. This solution addresses this shortcoming by requiring integrated, hourly performance simulation throughout the year based on the scientifically optimized array layout parameter set produced in the aforementioned steps. This is not only the final performance quantification but also a closed-loop verification of the overall solution design effectiveness. More importantly, this step mandates sensitivity analysis and forward-looking strategy generation, thereby elevating a simple prediction report into a comprehensive decision support document that includes accurate revenue expectations, quantified identification of key risks, and specific response recommendations.

[0043] The approach provided in this embodiment first integrates multi-source spatiotemporal data to conduct multi-dimensional diagnosis of the water body substrate, water quality, and wave dynamics, enabling scientific site selection based on reflection efficiency. Second, based on the dynamic reflection spectrum generated by the site selection, multi-objective collaborative optimization and shadow avoidance analysis are performed on the geometric parameters of the photovoltaic array, aiming to maximize back-side gain and minimize shading. Finally, the optimized layout parameters are used for year-round refined simulation and sensitivity assessment, generating a comprehensive analysis report integrating quantitative benefit and risk strategies, which is then provided to photovoltaic project personnel. This solution elevates the evaluation of bifacial photovoltaic systems on water surfaces from extensive experience to full-chain quantitative analysis. Through precise reflection source diagnosis, customized array optimization, and integrated risk assessment, it systematically improves the accuracy of power generation prediction and project investment returns, providing a reliable scientific basis for project decision-making.

[0044] In some embodiments, the multi-source spatiotemporal dynamic dataset includes the optical characteristics of the water body substrate, the spatiotemporal variation characteristics of water quality, and the dynamic characteristics of the water surface. Based on the substrate optical characteristics and the spatiotemporal variation characteristics of water quality, the inherent reflection and attenuation characteristics of the water body are analyzed to generate water body reflection characteristic analysis information. Based on the dynamic characteristics of the water surface, the modulation effect of dynamic waves on the water surface reflectivity is quantified to generate water surface reflectivity modulation information. By combining the water body reflection characteristic analysis results with the water surface reflectivity modulation coefficient, the reflection efficiency and engineering suitability of multiple candidate water bodies are compared to generate a site selection decision support information set.

[0045] The optical characteristics of a water body substrate can be relatively stable optical properties determined by the material composition, color, and reflectivity of the bottom and coastal areas of the water body. For example, white saline-alkali land or light-colored sandy substrates will significantly enhance the reflectivity of the water body, while dark silt or rocky substrates will have the opposite effect.

[0046] The spatiotemporal variation characteristics of water quality can be seen in the patterns of change of key water quality parameters such as suspended solids concentration (turbidity) and chlorophyll a concentration (characterizing algae content) with factors such as season, rainfall, and human activities.

[0047] The dynamic characteristics of the water surface can be the statistical distribution and variation of wave parameters (such as significant wave height and wave period) driven by the wind field. Waves can destroy the mirror properties of the water surface, causing reflected light to diffuse.

[0048] Information on water body reflection characteristics analysis can be a quantitative and structured diagnostic conclusion about the inherent reflection potential and optical attenuation characteristics of water bodies generated by coupling analysis of the optical characteristics of the substrate and the spatiotemporal variation characteristics of water quality.

[0049] Water surface reflectivity modulation information can be obtained by analyzing the dynamic characteristics of the water surface and quantitatively assessing the loss ratio or modulation coefficient of the effective reflectivity of the water surface caused by dynamic waves relative to the theoretical value of a calm water surface.

[0050] The comparison between reflection performance and engineering suitability can be a comprehensive decision-making process. "Reflection performance" is a quantitative indicator of the potential performance of water as a reflective light source, calculated based on optical diagnostic results. "Engineering suitability" is a development feasibility classification derived from non-optical constraints such as water topography, water depth, geology, transportation, and ecological red lines.

[0051] Specifically, traditional surface photovoltaic (PV) site selection heavily relies on land costs and grid connection distance, completely ignoring the differences in optical performance of water as a "reflective light source," and failing to consider its dynamic attenuation (e.g., algal blooms) and physical modulation (e.g., wind and waves). This leads to a failure of backside gain in projects built in turbid or turbulent waters, resulting in a return on investment far below expectations. To address these issues, this step first utilizes remote sensing image processing software (such as ENVI) to perform atmospheric correction and water body extraction on acquired satellite images. Supervised classification and spectral indices (e.g., NDWI) are used to identify substrate types (e.g., areas with reflectance > 0.4 are identified as "high-reflectance saline-alkali substrates"). Simultaneously, time-series water quality monitoring data is imported, and time-series analysis methods (e.g., STL decomposition) are used to extract the seasonal patterns of turbidity and chlorophyll concentration (e.g., the chlorophyll concentration of a certain lake reaches a peak of 50 μg / L in August each year). These two methods are coupled to generate water body reflectance characteristic analysis information (e.g., "The inherent reflectance of this water body reaches 0.3 during the winter high-reflectance substrate dominance period, and attenuates to 0.1 during the summer algal bloom period"). Secondly, wave buoy data is analyzed to statistically determine the wave height probability distribution across different wind speed ranges (e.g., at wind speeds of 3-5 m / s, the main wave height is 0.1-0.3 meters). Based on geometric optics principles, a micro-element normal vector distribution model for undulating water surfaces is established (commonly using the Cox-Munk model). This model calculates the proportion of luminous flux that can be received by the back of a fixed-angle component due to the expansion of the reflected light cone angle at a specific solar altitude angle, thereby generating water surface reflectivity modulation information (e.g., "Under average annual conditions, waves cause a 20% loss in effective reflectivity at this point").

[0052] The method provided in this embodiment quantitatively integrates the multi-dimensional dynamic diagnosis of water body optical properties with engineering constraints, providing scientific and accurate decision support for the site selection of bifacial photovoltaic power stations on water surfaces. It avoids the falsely high back-side gain prediction caused by ignoring water quality degradation and wind and wave modulation, and can identify sites with more stable optical performance and better comprehensive development conditions throughout the entire life cycle from multiple candidate water bodies, thus ensuring the technical feasibility and economic viability of the project from the source.

[0053] In some embodiments, by analyzing the optical characteristics of the water body substrate, the material composition and color of the bottom and shoreline of the water body are identified, and high-reflectivity substrates and low-reflectivity substrates are distinguished to determine whether the water body belongs to an ordinary lake or a high-reflectivity salt lake / crystallization lake, thus obtaining a substrate type determination result; by analyzing the temporal data of at least one complete year in the spatiotemporal variation characteristics of water quality, the seasonal variation patterns and peak characteristics of water turbidity and chlorophyll concentration are extracted; the substrate type determination result is coupled with the seasonal variation pattern to evaluate the water body light transmittance attenuation caused by changes in suspended matter concentration and periodic algal proliferation and its inherent weakening effect on reflected light intensity, generating water body reflectance characteristic analysis information.

[0054] High reflectivity substrates can generally refer to materials with high albedo, such as white salt crusts (with a reflectivity of over 0.5), light-colored sandy sediments, or substrates composed of calcium carbonate crystal layers.

[0055] Low reflectivity substrates can refer to substrates composed of dark silt, organic matter, or rock.

[0056] Seasonal variation patterns and peak characteristics can be the patterns of seasonal fluctuations of key water quality parameters (turbidity and chlorophyll concentration) extracted from time series data of at least one full year, including the period, amplitude, and time and value of extreme values ​​(peak values).

[0057] The inherent attenuation effect can be an unavoidable loss of reflected light intensity caused by the absorption and scattering of light by substances inside the water body (suspended matter, algae). This attenuation is determined by the physicochemical properties of the water body itself and is unrelated to external dynamic conditions (such as waves).

[0058] Specifically, traditional methods often set the water reflectance to a fixed value (e.g., 0.1), completely ignoring its significant spatial and temporal heterogeneity. In reality, a salt lake with a white salt crust at the bottom (inherent reflectance up to 0.4) is fundamentally different from a deep reservoir with a black muddy bottom (inherent reflectance may be only 0.05); even within the same water body, reflectance decreases sharply during summer algal blooms. Without precisely distinguishing the substrate and quantifying the spatiotemporal patterns of water quality degradation, any subsequent backside gain calculations will be based on severe distortion. To address these issues, this step first uses high spatial resolution multispectral satellite imagery (e.g., Sentinel-2) to perform atmospheric correction and land-water separation on the target water body area. Subsequently, the spectral reflectance curves of the water body pixels are analyzed, and the substrate type is identified based on the differences in reflectance characteristics in the near-infrared and short-wave infrared bands using spectral angle mapping or supervised classification methods. For example, pixel areas with a consistently high reflectance in the shortwave infrared band (e.g., SWIR1) above 0.3 are identified as "high-reflectance salt lake / crystallization lake" substrates, while areas with a reflectance below 0.1 are identified as "ordinary lake surface (low-reflectance silt substrate)". Secondly, at least one year of MODIS or Sentinel-3 satellite remote sensing inversion product data for the water body is obtained, and daily / weekly time series of chlorophyll a concentration and total suspended matter concentration (as a proxy indicator of turbidity) are extracted. Time series decomposition methods (e.g., seasonal decomposition) are used to separate the trend term, seasonal term, and residual term, thereby clearly extracting the seasonal variation patterns and peak characteristics of water quality parameters. For example, it is identified that the chlorophyll concentration reaches its mean peak in August each year (e.g., 35 μg / L). Finally, the two are coupled: for a water body identified as having a "high-reflectivity saline lake substrate," the effective reflectivity of the water surface is assessed during the high-algae season (e.g., August) because the water's light transmittance decreases significantly due to algal proliferation. This results in the absorption and scattering of the strong reflected light originally contributed by the high-reflectivity substrate by the algae, reducing the reflectivity to 0.15 (e.g., from the theoretical substrate reflectivity of 0.4). This comprehensive analysis conclusion is output as "Water Body Reflectivity Analysis Information."

[0059] The method provided in this embodiment enables in-depth analysis of the inherent reflection characteristics of water bodies from "static type identification" to "dynamic attenuation quantification," providing accurate background information of reflection sources with spatiotemporal resolution that far exceeds fixed empirical values, thus laying a solid and reliable physical foundation for the entire back-side gain evaluation system.

[0060] In some embodiments, by analyzing continuous monitoring data in the dynamic characteristics of the water surface, the distribution of water surface wave height and wave period characteristics under different wind levels is statistically analyzed; based on the geometric optics principle of solar position, a reflection path analysis model of calm water surface and undulating water surface on incident sunlight is established; according to the reflection path analysis model, the proportion of effective reflected light flux loss that can be received by the back of photovoltaic module caused by the dispersion of specular reflected light direction under different wave conditions is analyzed and compared, and all effective reflected light flux loss proportions are integrated to generate water surface reflectivity modulation information.

[0061] A reflection path analysis model can refer to a physical model based on the principles of geometric optics, used to describe and calculate the direction of sunlight after reflection from a water surface. For calm water surfaces, this model simplifies to the standard law of specular reflection, and the direction of the reflected light is uniquely determined.

[0062] The effective reflected light flux loss ratio can be defined as the ratio (usually less than 1) between the actual reflected light flux that can be received by the back of a photovoltaic module at a specific tilt angle and position after the water surface is no longer an ideal mirror due to waves, causing the reflected light direction to diffuse, and the theoretical light flux that the back of the module should receive when the water surface is completely calm (ideal mirror reflection).

[0063] Specifically, traditional assessments either simply assume the water surface is a calm mirror or use only a single empirical coefficient to broadly estimate the wave impact, which seriously deviates from physical reality. Wind-generated waves tilt the water surface, causing a significant increase in the reflective cone angle of the mirror, and a large amount of reflected light cannot be captured by the back of the module due to directional deviation. Ignoring this effect will systematically overestimate the back gain. For example, under common wind conditions, waves may cause a loss of more than 20% in effective reflected light flux. If this modulation effect is not quantified, the power generation forecast will be seriously inflated, leading to incorrect investment decisions. To address the above problems, this step first processes continuous time-series water surface dynamic characteristic data (such as wave height and wave period sequences) from wave buoys or numerical models. Statistical methods (such as quantile statistics and probability density function fitting) are used to analyze the joint characteristic distribution of wave height and wave period in different wind speed ranges (such as 0-3 m / s, 3-6 m / s). For example, statistics show that when the wind speed is 4 m / s, the most likely wave height is 0.15 meters, corresponding to a main wave period of 1.2 seconds. Subsequently, based on this wave condition statistics, a Cox-Munk-type geometric optics model is applied to characterize the wave surface as a set of micro-facets that obey a probability density function with a specific slope distribution. For a given solar position (elevation angle, azimuth angle), the local specular reflection direction of each micro-facet is calculated, and the spatial distribution of all reflected rays is statistically analyzed, forming a "reflection cone" with a certain angular radius centered on the theoretical specular reflection direction. Next, a three-dimensional spatial geometric intersection calculation is performed between this "reflection cone" and the back of the photovoltaic module (defined as a spatial receiving plane based on its preset tilt angle and azimuth angle) to determine the proportion of light falling within the solid angle range of the receiving plane. For example, it is calculated that when the solar elevation angle is 30 degrees, for the aforementioned wave condition, the back of the module can only receive about 78% of the luminous flux within this "reflection cone," i.e., the loss ratio is 22%. Finally, the above calculations are repeated for different representative solar elevation angles and different wind speeds (corresponding to different wave conditions), integrating them to obtain a series of "condition-loss ratio" correspondences, constituting complete water surface reflectivity modulation information.

[0064] The method provided in this embodiment transforms the previously ignored or roughly estimated wave scattering effect into an accurate luminous flux loss ratio related to the sun's position and wave conditions. This fundamentally corrects the accuracy of the reflected irradiance input and provides key modulation parameters that can withstand physical testing for subsequent back-side gain assessment, avoiding systematic prediction bias caused by model distortion.

[0065] In some embodiments, based on the analysis information of water body reflection characteristics and the modulation information of water surface reflectivity, the apparent motion law of the sun and the seasonal characteristics of the water body are coupled to extract and calculate the effective reflectivity of the water surface of each candidate water body under several representative seasons and several typical solar altitude angle scenarios within each season, generating a dynamic reflection efficiency map of each water body; combined with the spatial geographic information of each water body and the preset engineering constraints, a preliminary screening of engineering feasibility is conducted to generate an engineering suitability classification map; the dynamic reflection efficiency map of each water body and the engineering suitability classification map are spatially overlaid and analyzed, and the comprehensive development potential of multiple candidate water bodies and their internal different sub-regions is quantitatively evaluated and ranked through weighted scoring or multi-criteria decision-making algorithms; based on the quantitative evaluation and ranking results, a site selection decision support information set is used, in which the water body with the best comprehensive conditions and its internal recommended priority construction areas are clearly identified.

[0066] Dynamic reflectivity map is a spatiotemporal multidimensional data visualization and quantification product that comprehensively presents the spatial distribution of effective reflectivity of water surface calculated for each candidate water body in different representative seasons (such as spring, summer, autumn and winter) and under several typical solar altitude angle scenarios (such as the sun's position at 9:00, 12:00 and 15:00) within each season in the form of a map.

[0067] Pre-set engineering constraints can be a set of quantifiable screening criteria set during the early site selection stage of a photovoltaic project, based on engineering construction experience, safety regulations, and cost control principles, to quantitatively assess whether a specific body of water is suitable for installing a floating photovoltaic array.

[0068] Preliminary feasibility screening can be an automated analysis process based on a geographic information system (GIS) platform, which involves spatially overlaying and condition-judging the spatial geographic information of the water body (such as digital elevation models, shoreline vectors, and water depth point data) with preset engineering constraints.

[0069] A suitability grading map is a hierarchical visualization map generated based on geographic information system (GIS) spatial analysis, used to characterize the feasibility of engineering construction in different water areas.

[0070] Spatial overlay analysis refers to a method in a GIS environment that performs pixel-level or vector unit-level overlay operations on two or more spatial data layers with the same geographic coordinate system (specifically referring to the derived index layer of the dynamic reflectivity map and the engineering suitability grading map) to comprehensively evaluate the combination of multiple attributes on each geographic location unit (such as a 10m×10m grid) within a unified spatial framework.

[0071] Weighted scoring or multi-criteria decision-making algorithms can be a mathematical decision-making framework that integrates multiple evaluation criteria with different dimensions and importance (such as "reflection effectiveness" and "engineering suitability") into a unified and comparable comprehensive score.

[0072] Specifically, traditional site selection often separates optical potential assessment from engineering feasibility assessment, either choosing "optical highlands" with high reflectivity but insufficient water depth or poor geological conditions, or "engineering lowlands" with convenient construction but poor reflective background, failing to find a comprehensive optimal solution. To address these issues, this step first, based on the previously generated water body reflectivity analysis information (including substrate type and seasonal water quality attenuation) and water surface reflectivity modulation information (wave loss ratio), selects four representative days—the spring equinox, summer solstice, autumn equinox, and winter solstice—for each candidate water body, and calculates three typical solar altitude angles (e.g., 30°, 50°, 70°) for each day. For each "seasonal-solar altitude angle" combination, in GIS software, based on the inherent reflectivity characteristics of each point, the corresponding seasonal attenuation coefficient and the corresponding wave condition modulation coefficient are superimposed to calculate and generate the effective reflectivity spatial distribution matrix under that scenario. All matrices are combined to form the dynamic reflectivity performance map of the water body. Simultaneously, in another GIS workspace, data such as water body boundaries, DEM, water depth measurement points, and ecological protection zone vectors were imported. Constraints were set (e.g., water depth > 2.5m, slope < 10°, distance from shoreline < 500m, and not within the ecological red line). Raster Calculator or Weighted Overlay tools were used to determine the conditions and assign grading values, generating an engineering suitability grading map. Subsequently, the "annual effective reflectance" index layer extracted from the performance map was spatially overlaid with the engineering suitability grading map. For each spatial unit (raster cell), a multi-criteria decision algorithm such as Weighted Linear Combination (WLC) was used to assign weights to reflectance performance (e.g., weight 0.6) and engineering suitability level (e.g., weight 0.4), calculating the comprehensive score for each cell (e.g., a reflectance performance value of 0.3 maps to a score of 80, and an engineering level of "preferred suitability" maps to a score of 90, then the comprehensive score = 800.6 + 900.4 = 84). Finally, all candidate water bodies and their internal sub-regions are ranked according to the comprehensive score, generating a site selection decision support information set that clearly identifies the optimal water body and priority construction blocks.

[0073] The method provided in this embodiment introduces spatial overlay analysis and quantitative decision-making of spatiotemporal dynamic reflection efficiency maps and multi-constraint engineering suitability classification. This elevates site selection decisions from qualitative judgments based on single advantages or experience to a multi-objective, full-space-coverage quantitative optimization process. This ensures that the final recommended site achieves the optimal balance between energy benefits and engineering implementation, significantly reducing the technical and economic risks of project site selection.

[0074] In some embodiments, a dynamic reflectivity spectrum is extracted from the site selection decision support information set. Based on physical optical transmission simulation technology, the spatiotemporal reflectivity distribution of the water surface represented by this spectrum is transformed into reflected irradiance data that can be received on the back of the photovoltaic array. Based on the reflected irradiance data, a comprehensive optimization function for the module is established with the module installation height, tilt angle, and north-south spacing of the array as the core optimization variables. The objective is to maximize the back irradiance reception, minimize the front shading loss, and optimize the energy output per unit water surface area. A multi-objective optimization algorithm is used to iteratively solve the comprehensive optimization function for the module to obtain several recommended combinations of geometric parameters. Shadow analysis is performed on each combination during key periods to ensure that the back shading rate is lower than a preset shading threshold. Finally, the array layout parameter set is integrated and optimized.

[0075] Physical optics transmission simulation technology is a technique that simulates the entire process of light (electromagnetic wave) propagation, reflection, absorption and scattering in a three-dimensional digital environment based on physical principles such as geometric optics or ray tracing.

[0076] The spatiotemporal reflectance distribution of water surface can be a multi-objective mathematical optimization model. Its inputs are the geometric parameters of the photovoltaic array (installation height H, tilt angle θ, north-south spacing D), and its outputs are multiple performance indicators related to back gain, front shading, and land utilization.

[0077] Reflected irradiance data can be a dataset represented in the form of a spatial distribution matrix, calculated using physical optical transmission simulation techniques (such as ray tracing). The value of each data unit (pixel or grid) represents the solar radiation energy received per unit area and per unit time on the back receiving plane of the corresponding component in the photovoltaic array under specific spatiotemporal conditions (such as a solar altitude angle in a certain season). The unit is usually W / m².

[0078] The integrated optimization function of the components can be a multi-objective mathematical function model with the geometric parameters of the photovoltaic array (installation height H, tilt angle θ, north-south spacing D) as independent variables, and multiple technical and economic indicators as objectives, such as maximizing the back irradiance reception, minimizing the front shading loss, and optimizing the energy output per unit water surface area.

[0079] A multi-objective optimization algorithm can be an iterative algorithm used to solve mathematical problems with multiple (usually conflicting) optimization objectives.

[0080] The preset occlusion threshold can be an upper limit (e.g., 5%) of the back shadow occlusion rate that is set manually in shadow avoidance analysis.

[0081] Specifically, traditional photovoltaic array designs are typically based on fixed ground reflection models or simplified empirical formulas. Their layout parameters (such as tilt angle and spacing) are mostly optimized to maximize front-side irradiance, completely ignoring the unique and complex nature of receiving reflected light from dynamic water surfaces on the back side. Simply applying these formulas can lead to excessively large array spacing, wasting water resources, or excessively small spacing, causing severe shading on the front side. Simultaneously, the module installation height, a crucial parameter for back-side gain (directly affecting the solid angle of the received reflected light), is often overlooked. To address these issues, this step first uses ray tracing software (such as Radiance or a self-developed simulator) to simulate physical optical transmission. Specifically, the effective reflectivity distribution of the water surface extracted from the dynamic reflection efficiency spectrum for a specific season (such as the summer solstice) and a specific solar altitude angle (such as 50° at noon) is mapped onto a fine three-dimensional water surface mesh model. Each mesh cell is assigned a corresponding reflectivity value and is considered a secondary light source emitting reflected light into the hemispherical space. Simultaneously, based on an initial set of geometric parameters (e.g., H=2 meters, θ=20°, D=8 meters), a precise 3D model of the photovoltaic array is constructed in the software, designating the back of the module as the irradiance receiving surface. The software emits millions of "rays" from the water surface grid, tracks their propagation paths, counts whether each ray hits the back of the module, and accumulates the energy carried, thereby calculating the reflected irradiance received on the back of the array under this set of parameters (e.g., the average irradiance reaches 18% of the direct irradiance on the front). Furthermore, based on the above simulation capabilities, a new set of geometric parameters (H, θ, D) is generated in each iteration of the optimization algorithm (e.g., NSGA-II). For each set of parameters, the aforementioned ray tracing simulation is performed to calculate three objective function values: F1 (annualized back-side irradiance, obtained by weighted integration of simulation results from representative times), F2 (the proportion of the area projected by the front-row components onto the front of the next-row components at typical times on key days, and annualized to obtain shading loss), and F3 (the net value of F1 and F2 divided by the array's footprint). The algorithm iterates hundreds to thousands of times, ultimately outputting a set of Pareto optimal solutions. For example, Scheme A (H=2.5m, θ=15°, D=10m) prioritizes high back-side gain; Scheme B (H=1.8m, θ=25°, D=6.5m) prioritizes high land utilization. Finally, an independent shading analysis is performed on each scheme in the Pareto solution set to ensure that its back-side shading rate is below a preset threshold (e.g., 5%) at all key times (e.g., 9 AM on the winter solstice). The selected schemes are then integrated into the optimized array layout parameter set.

[0082] The approach provided in this embodiment breaks through the limitations of traditional single-objective optimization or empirical design, systematically balancing the contradiction between maximizing back-side gain, minimizing front-side shading, and maximizing space utilization efficiency. It provides a physically verifiable and globally optimized array layout scheme for bifacial photovoltaic floating power stations, ensuring the expected power generation revenue and investment efficiency of the project from the design source.

[0083] In some embodiments, based on the dynamic reflectivity map, the effective reflectivity matrix of the water surface under different representative seasons and typical solar altitude angles is analyzed by spatial interpolation algorithm; according to the effective reflectivity matrix of the water surface, combined with the solar direct irradiance calculated by the solar position astronomical algorithm, the spatial distribution of reflected radiation intensity corresponding to the water surface is calculated; the spatial distribution of reflected radiation intensity is used as the light source input, and the propagation path of reflected light in the three-dimensional space of the photovoltaic module array is simulated by using ray tracing or geometric optics model, and the received plane on the back of each module is used as the target surface to calculate the reflected irradiance intensity data that can be received on the back of the photovoltaic array.

[0084] Spatial interpolation algorithms are mathematical methods used to estimate the values ​​of unknown points in a continuous spatial field based on observations of known discrete points in space. Common algorithms include Kriging interpolation or inverse distance weighted interpolation (IDW).

[0085] The effective reflectivity matrix of the water surface can be a regular two-dimensional array data structure obtained by processing it through a spatial interpolation algorithm.

[0086] Solar position astronomical algorithms can be based on precise astronomical models (such as the SPA (Solar Position Algorithm) algorithm) to calculate the apparent position of the sun (including solar altitude angle and solar azimuth angle) at a specific geographical location (latitude and longitude), date and time, and often can simultaneously output the calculation program or model of the solar direct irradiance (DNI) at the top of the atmosphere or under specific atmospheric conditions at that location.

[0087] The spatial distribution of reflected radiation intensity can be obtained by multiplying each element of the effective reflectivity matrix of the water surface by the solar direct irradiance (DNI) calculated by the solar position astronomical algorithm at the corresponding time, resulting in a new two-dimensional spatial distribution matrix.

[0088] Specifically, traditional methods for estimating back-side gain often simplify the entire water body into a single, uniform reflectance value, using simplified cosine projection or empirical view factors for calculation. This approach completely obliterates the spatial non-uniformity of water surface reflectance (such as the difference in reflectance between shallow nearshore areas and deep central lake areas) and the complex three-dimensional geometric relationships of reflected light propagation, leading to severely distorted calculation results. This makes it impossible to accurately assess the differences in back-side irradiance received by components at different locations within the array, thus affecting the accuracy of optimized layout. To address these issues, this step first extracts all effective reflectance sampling point data for a specific simulation scenario (e.g., a representative summer day with a solar altitude angle of 50°) from the dynamic reflectance performance map. In GIS or professional scientific computing software (such as MATLAB's `scatteredInterpolant` function), the Kriging interpolation algorithm is used to interpolate these discrete point data onto a high-resolution regular grid covering the entire water surface (e.g., generating a 1000-row × 2000-column floating-point matrix with a grid size of 0.5m × 0.5m) to obtain the effective reflectivity matrix of the water surface under this scenario. Then, the solar position astronomical algorithm library (such as the `get_solarposition` and `get_clearsky` functions in NREL's `pvlib`) is called, inputting the power station's geographical coordinates and the specific date and time corresponding to this simulation scenario (e.g., 10:30 AM on July 15th) to calculate the precise solar altitude angle (e.g., 49.8°) and direct solar irradiance (e.g., 800 W / m²). Next, scalar multiplication of the matrix is ​​performed: the irradiance value (800 W / m²) is multiplied by each element of the effective reflectivity matrix of the water surface to generate a spatial distribution matrix of reflected radiation intensity at the same resolution. Finally, in ray tracing simulation software (such as Radiance), this intensity distribution matrix is ​​assigned to the 3D water surface model, defining it as an extended light source with spatially varying intensity. Simultaneously, a precise 3D model of the photovoltaic array is created in the software, and the material on the back of the modules is set to a perfect Lambertian receiver. A Monte Carlo ray tracing simulation is then initiated, emitting a sufficient number of random rays (e.g., 5 million rays) from the water surface light source. The software automatically calculates whether each ray, after possible atmospheric attenuation, hits the back of a module and accumulates its energy contribution. After the simulation, the software outputs the irradiance value received by each grid cell on the back of the module; summing these values ​​yields the dataset of reflected irradiance intensity that the array can receive on the back under this scenario.

[0089] The method provided in this embodiment overcomes the spatial distortion problem of traditional homogenization models, enabling the evaluation of back-side gain to truly reflect the combined effect of spatial heterogeneity of water surface reflectivity and three-dimensional light propagation geometry, providing a solid and reliable physical input data foundation for subsequent multi-objective optimization of array layout.

[0090] In some embodiments, for each recommended combination of geometric parameters, a three-dimensional geometric model of the photovoltaic array is constructed based on the component installation height, tilt angle, and north-south spacing of the array; several key days that have a significant impact on the annual power generation of the system are selected, and the typical solar hour angle that may produce significant front-row shading effects on the key days is calculated; using the three-dimensional geometric model, the shadow projection of each combination of geometric parameters under the typical solar hour angle is calculated and visualized, and the proportion of the area of ​​the back of the rear components that is shaded by the shadows of the front components or supporting structures is quantitatively analyzed, i.e., the back shadow shading rate; the combination of all geometric parameter combinations in which the back shadow shading rate is lower than the preset shading threshold is selected as a feasible solution through shadow avoidance analysis and integrated into the optimized array layout parameter set.

[0091] A three-dimensional geometric model can be a digital three-dimensional solid model that is precisely constructed using computer-aided design (CAD) software or parametric modeling scripts, based on a combination of geometric parameters such as the physical dimensions of the photovoltaic module (e.g., 1.96 meters long and 1.0 meter wide), the module installation height (H), tilt angle (θ), and the north-south spacing (D) of the array. It contains the spatial position and size information of the module and supporting structure (e.g., purlins and columns).

[0092] A critical day can be a specific date in a year that has the most adverse or representative impact on the shading of the front and back of a photovoltaic array due to special changes in the solar altitude angle and azimuth angle.

[0093] Typical solar hour angles can be the solar hour angles (or local times) corresponding to specific time points selected for shading analysis on key days. These times are usually chosen when the solar altitude angle is low and the shadow is long, such as 9:00 am (solar hour angle is about -45°) and 3:00 pm (solar hour angle is about +45°) on the winter solstice, in order to assess the most severe potential shading.

[0094] The back shading rate can be calculated by shadow projection at a specific typical solar hour angle, and is the ratio of the area of ​​the back of the rear photovoltaic module (i.e. the receiving side facing the water) covered by the shadows of the front module, adjacent modules, or supporting structure to the total area of ​​the back of the module.

[0095] Specifically, in bifacial photovoltaic (PV) hydroelectric power stations, the back side receives reflected light from the water surface. If the back side is shaded, the gain effect will be significantly reduced or even completely lost. Traditional array layout optimization often only considers front-side shading or simplifies shading, especially neglecting specific analysis of back-side shading. Without rigorous shading analysis during critical periods, optimized parameter combinations (such as excessively small array spacing) may cause large-area shading of the back-side modules by the front-side modules during low solar altitude angles in winter, severely weakening the expected back-side gain and resulting in actual power generation far below the design value. To address these issues, this step first uses parametric modeling tools (such as Rhino+Grasshopper or Python's py3d library) to automatically construct the corresponding 3D geometric model for each set of geometric parameter combinations recommended by the multi-objective optimization algorithm (e.g., scheme A: H=2.5 meters, θ=15°, D=10 meters). The model strictly adheres to the module dimensions and installation parameters, accurately arranging all module rows and including a support purlin model with a cross-sectional size of 0.1 meters × 0.15 meters. Next, based on the geographical coordinates of the power station location (e.g., 30° North latitude), the winter solstice, summer solstice, and vernal equinox are selected as key days. For each key day, the typical solar hour angle that may produce significant shadows is calculated using a solar position algorithm; for example, 9:00 and 15:00 local true solar time are selected for the winter solstice. Then, for each scenario of "parameter combination - key day - typical hour angle", shadow projection calculation is performed. Specifically, using ray casting or projective geometry algorithms: based on the solar vector direction (elevation angle and azimuth angle) at that hour angle, the outlines of all objects that may cause shading (edges of front-row components, purlins) in the 3D model are projected onto the back receiving plane of the rear-row components along the opposite direction of sunlight. By calculating the Boolean intersection of these projected outlines and the rectangular area on the back of the components, the area of ​​the shading area is accurately obtained, and then the back shadow shading rate of the component at that moment is calculated (for example, at 9:00 on the winter solstice, 6.2% of the area on the back of the second row of components in scheme A is covered by the shadow of the lower edge of the first row of components). Finally, iterate through all scenarios and check whether the back shadow occlusion rate of each set of parameters is lower than the preset occlusion threshold (e.g., 5%) at all preset key time angles. Select parameter combinations that meet all constraints (e.g., scheme A passes, while a certain scheme B with D=8 meters has an occlusion rate of 7.5% at 9:00 on the winter solstice and is eliminated) and integrate them into the final optimized array layout parameter set.

[0096] The method provided in this embodiment ensures that the recommended array layout can keep the back shadow shading within an acceptable range at any critical time, thereby guaranteeing the stable realization of the back gain of the bifacial modules, avoiding the loss of back power generation potential due to design oversights, and significantly improving the reliability of power plant power generation prediction and investment security.

[0097] In some embodiments, the photovoltaic system is simulated hourly throughout the year with the optimized array layout parameter set as the core configuration. The additional power generation obtained by the bifacial photovoltaic module due to water surface reflection is quantitatively calculated, i.e., the accurate back gain value. The sensitivity of the accurate back gain value to key input parameters such as reflectivity and installation height is analyzed to identify the main sources of gain risk and uncertainty. Based on the accurate back gain value and the sensitivity analysis conclusions, a bifacial photovoltaic back gain analysis report is generated, which includes the expected power generation improvement, economic assessment, and array optimization design suggestions.

[0098] Hourly performance simulation throughout the year refers to the process of using professional photovoltaic system simulation software or platforms (such as PVsyst, SAM) with optimized array layout parameter sets (including installation height, tilt angle, spacing and corresponding component models, inverter parameters, etc.) as the core input, combined with typical meteorological year (TMY) data of the power station location, and simulating the power generation of the photovoltaic system every hour in the 8760 hours of the year with a time step of one hour.

[0099] The precise back-side gain value can be obtained through hourly performance simulations throughout the year. It is the difference between the total power generation obtained by using bifacial photovoltaic modules and considering water surface reflection under the exact same environment and system configuration, and the power generation obtained by a single-sided photovoltaic module assuming the same tilt angle and receiving sunlight only from the front.

[0100] Sensitivity analysis is an analytical method that involves systematically changing the values ​​of key input parameters of a model to observe and quantify the magnitude and patterns of changes in the model's output (in this case, the precise backside gain value).

[0101] Key input parameters can be input variables that significantly affect the calculation results of the accurate back-side gain value in the back-side gain analysis model and have a certain degree of uncertainty or variability.

[0102] Gain risk and uncertainty sources can be the main factors identified by sensitivity analysis that cause the accurate back-side gain value to be lower than expected or to fluctuate significantly. For example, the analysis may reveal that the gain value is extremely sensitive to reflectivity (a 10% decrease in reflectivity leads to a 15% decrease in gain) and relatively insensitive to changes in installation height within a certain range (a 0.5-meter change in height results in a gain change of <3%), thus identifying "reflectivity prediction accuracy" as the core risk.

[0103] Specifically, traditional evaluations of floating photovoltaic (PV) projects often stop at theoretical gain estimations or single-scenario simulations. Their reports lack accurate power generation data based on real-time weather conditions throughout the year, and fail to quantify the actual impact of uncertainties in key assumptions (such as reflectivity) on returns. This results in insufficient basis for investment decisions and an inability to answer core risk questions such as "What will the returns be in the worst-case scenario?" and "Which design parameter is most critical?" To address these issues, this step first inputs the final recommended scheme from the optimized array layout parameter set (e.g., scheme C: H=2.2 meters, θ=18°, D=9 meters) and its corresponding component and inverter models as core configurations into the PVsyst PV simulation software. In the software, besides configuring regular meteorological files, the most crucial step is to import the monthly (or more refined) reflectivity data sequence of the water surface, defined by dynamic reflectivity performance maps and water surface reflectivity modulation information, through its "user-defined reflectivity model" function. The array geometry is then precisely set according to the optimized array layout parameter set. After the software performs hourly performance simulations throughout the year, it outputs an annual total power generation report for the bifacial system (e.g., 12.5 million kWh per year for a single power station). Next, under identical simulation settings, the module model was replaced with an equivalent single-sided module model, and the simulation was run again to obtain a baseline power generation (e.g., 11.5 million kWh). Subtracting the baseline from the baseline yielded the precise backside gain value (1 million kWh / year). Subsequently, sensitivity analysis was performed: multiple scenarios were created in PVsyst, for example, reducing the input average water surface reflectivity by 10% (from 0.25 to 0.225), and resimulating to observe the gain change (potentially decreasing to 850,000 kWh); or simply reducing the module installation height from 2.2 meters to 1.7 meters and observing the gain change (potentially decreasing to 920,000 kWh). By comparing the fluctuation range of the gain value under different parameter changes, reflectivity was identified as the most sensitive key input parameter, i.e., the main source of gain risk. Finally, all the above results were integrated: the precise back gain value was converted into the expected power generation increase (8.7%), and an economic assessment was completed by combining local electricity prices (e.g., 0.45 yuan / kWh) and cost data (the IRR was calculated to be 9.2%). Based on the sensitivity conclusions and optimization process, specific array optimization design suggestions were extracted. These contents were then structured and compiled to generate the final bifacial photovoltaic back gain analysis report.

[0104] The method provided in this embodiment reveals the key driving factors and potential risks behind the returns, thus providing authoritative evidence with both quantitative support and qualitative insights for project investment decisions, engineering design optimization, and subsequent operation, greatly improving the scientific nature and reliability of project development.

[0105] Figure 3 A schematic diagram of a back-side gain analysis system for bifacial photovoltaic modules provided in an embodiment of this application is shown below. Figure 3As shown, a back-side gain analysis system 300 for bifacial photovoltaic modules in this embodiment includes: a site selection decision module 301, an array optimization module 302, and a report generation module 303.

[0106] The site selection decision module 301 is used to acquire a multi-source spatiotemporal dynamic dataset of the target water body, perform multi-dimensional diagnosis of water body reflection characteristics and site selection decision based on the multi-source spatiotemporal dynamic dataset, and generate a site selection decision support information set. The optimized array module 302 is used to perform multi-objective collaborative optimization and shadow avoidance analysis of the photovoltaic array geometric parameters for maximizing back-side gain based on the site selection decision support information set, and generate an optimized array layout parameter set. The report generation module 303 is used to perform integrated back-side gain quantification evaluation and forward-looking strategy generation based on the optimized array layout parameter set, and generate and output a bifacial photovoltaic back-side gain analysis report.

[0107] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A method for back-side gain analysis of bifacial photovoltaic modules, characterized in that, include: A multi-source spatiotemporal dynamic dataset of the target water body is acquired. Based on the multi-source spatiotemporal dynamic dataset, a multi-dimensional diagnosis of the water body's reflection characteristics and a site selection decision are performed, generating a site selection decision support information set. Based on the site selection decision support information set, multi-objective collaborative optimization and shadow avoidance analysis of the geometric parameters of the photovoltaic array aimed at maximizing back-side gain are performed to generate an optimized array layout parameter set. Based on the optimized array layout parameter set, an integrated back-side gain quantification evaluation and forward-looking strategy generation are performed, generating and outputting a bifacial photovoltaic back-side gain analysis report.

2. The method according to claim 1, characterized in that, The process of generating the location decision support information set includes: The multi-source spatiotemporal dynamic dataset includes optical features of the water body substrate, spatiotemporal variation features of water quality, and dynamic features of the water surface. Based on the optical characteristics of the substrate and the spatiotemporal variation characteristics of the water quality, the inherent reflection and attenuation characteristics of the water body are analyzed to generate water body reflection characteristic analysis information. Based on the aforementioned water surface dynamic characteristics, the modulation effect of dynamic waves on water surface reflectivity is quantified to generate water surface reflectivity modulation information. By combining the analysis results of the water body reflection characteristics with the water surface reflectivity modulation coefficient, the reflection performance and engineering suitability of multiple candidate water bodies are compared, and the site selection decision support information set is generated.

3. The method according to claim 2, characterized in that, The analysis of the inherent reflection and attenuation characteristics of the water body includes: By analyzing the optical characteristics of the water body substrate, the material composition and color of the bottom and shore of the water body are identified, and high reflectivity substrates and low reflectivity substrates are distinguished to determine whether the water body belongs to an ordinary lake or a high reflectivity salt lake / crystallized lake, and the substrate type determination result is obtained. By analyzing the time series data of at least one full year in the aforementioned spatiotemporal variation characteristics of water quality, the seasonal variation patterns and peak characteristics of water turbidity and chlorophyll concentration are extracted. The substrate type determination result is coupled with the seasonal variation pattern to evaluate the water body light transmittance attenuation caused by changes in suspended matter concentration and periodic algal proliferation, as well as its inherent weakening effect on reflected light intensity, thereby generating the water body reflectance characteristic analysis information.

4. The method according to claim 3, characterized in that, The modulation effect of the quantized dynamic waves on the water surface reflectivity includes: By analyzing the continuous monitoring data in the water surface dynamic characteristics, the distribution of water surface wave height and wave period characteristics under different wind force levels was statistically analyzed. Based on the principles of solar position geometric optics, a reflection path analysis model for incident sunlight on calm and undulating water surfaces is established. Based on the reflection path analysis model, the effective reflected light flux loss ratio that can be received by the back of the photovoltaic module is analyzed and compared due to the dispersion of the specular reflected light direction under different wave conditions. All the effective reflected light flux loss ratios are integrated to generate the water surface reflectivity modulation information.

5. The method according to claim 4, characterized in that, The comparison of the reflection performance and engineering suitability of multiple candidate water bodies includes: Based on the water body reflection characteristic analysis information and the water surface reflectivity modulation information, the apparent motion law of the sun and the seasonal characteristics of the water body are coupled to extract and calculate the effective reflectivity of each candidate water body in several representative seasons and several typical solar altitude angle scenarios in each season, and generate dynamic reflection efficiency maps of each water body. By combining the spatial geographic information of each water body with the preset engineering constraints, a preliminary screening of engineering feasibility is conducted, and an engineering suitability classification map is generated. The dynamic reflection efficiency maps of each water body are spatially overlaid with the engineering suitability classification map. The comprehensive development potential of multiple candidate water bodies and their different sub-regions is quantitatively evaluated and ranked by weighted scoring or multi-criteria decision-making algorithms. Based on the quantitative assessment and ranking results, the site selection decision support information set is used to clearly identify the water body with the best comprehensive conditions and the areas within it that are recommended for priority construction.

6. The method according to claim 5, characterized in that, The process of generating the optimized array layout parameter set includes: The dynamic reflectivity spectrum is extracted from the site selection decision support information set. Based on physical optics transmission simulation technology, the spatiotemporal reflectivity distribution of the water surface represented by this spectrum is converted into reflected irradiance data that can be received on the back of the photovoltaic array. Based on the reflected irradiance data, a comprehensive optimization function for the components is established with the component installation height, tilt angle and array north-south spacing as the core optimization variables. The objective is to maximize the back irradiance reception, minimize the front shading loss and optimize the energy output per unit water surface area. A multi-objective optimization algorithm is used to iteratively solve the comprehensive optimization function of the components to obtain several recommended combinations of geometric parameters. For each combination, shadow analysis is performed during key periods to ensure that the shadow occlusion rate on the back side is lower than the preset occlusion threshold. Finally, the optimized array layout parameter set is integrated.

7. The method according to claim 6, characterized in that, The conversion into reflected irradiance intensity data that can be received on the back of the photovoltaic array includes: Based on the dynamic reflectivity map, the effective reflectivity matrix of the water surface under different representative seasons and typical solar altitude angles is analyzed by spatial interpolation algorithm. Based on the effective reflectivity matrix of the water surface and the solar direct irradiance calculated by the solar position astronomical algorithm, the spatial distribution of reflected radiation intensity corresponding to the water surface is calculated. Using the spatial distribution of reflected radiation intensity as the light source input, the propagation path of reflected light in the three-dimensional space of the photovoltaic module array is simulated using ray tracing or geometric optics models. With the receiving plane on the back of each module as the target surface, the reflected irradiance data that can be received on the back of the photovoltaic array is calculated.

8. The method according to claim 6, characterized in that, The step of performing key-time period shadow analysis on each combination to ensure that the back shadow occlusion rate is below a threshold includes: For each recommended combination of geometric parameters, a three-dimensional geometric model of the photovoltaic array is constructed based on the component installation height, tilt angle, and north-south spacing of the array. Select several key days that have a significant impact on the system's annual power generation, and calculate the typical solar hour angles that may produce significant front-row shading effects on the key days; Using the three-dimensional geometric model, the shadow projection of each geometric parameter combination under the typical solar hour angle is calculated and visualized, and the area ratio of the rear component back side shaded by the shadow of the front component or support structure is quantitatively analyzed, that is, the rear shadow shading rate. The combinations of geometric parameters in which the back shadow occlusion rate is lower than the preset occlusion threshold are selected as feasible solutions through shadow avoidance analysis and integrated into the optimized array layout parameter set.

9. The method according to claim 6, characterized in that, The process of generating the bifacial photovoltaic back-side gain analysis report includes: Using the optimized array layout parameter set as the core configuration, the photovoltaic system's performance is simulated hourly throughout the year, and the additional power generation obtained by the bifacial photovoltaic modules due to water surface reflection is quantitatively calculated, i.e., the precise back gain value. The sensitivity of the precise back-side gain value to key input parameters such as reflectivity and installation height was analyzed to identify the main sources of gain risk and uncertainty. Based on the accurate back-side gain values ​​and sensitivity analysis conclusions, a bifacial photovoltaic back-side gain analysis report is generated, which includes expected power generation improvement, economic assessment, and array optimization design recommendations.

10. A back-side gain analysis system for bifacial photovoltaic modules, characterized in that, The method applied to any one of claims 1-9 includes: The site selection decision module is used to acquire a multi-source spatiotemporal dynamic dataset of the target water body, perform multi-dimensional diagnosis of water body reflection characteristics and site selection decision based on the multi-source spatiotemporal dynamic dataset, and generate a site selection decision support information set. The optimized array module is used to perform multi-objective collaborative optimization and shadow avoidance analysis of the geometric parameters of the photovoltaic array aimed at maximizing the back-side gain based on the site selection decision support information set, and generate an optimized array layout parameter set. The report generation module is used to perform integrated back-side gain quantification evaluation and forward-looking strategy generation based on the optimized array layout parameter set, and generate and output a bifacial photovoltaic back-side gain analysis report.

Images