Multi-objective decision method for photovoltaic module scrapping, storage medium and equipment

Abstract

The application relates to a multi-target decision method for photovoltaic module scrapping treatment, a storage medium and equipment, and aims to solve the problem that the decision result is inaccurate when a decision on the treatment of a retired photovoltaic module is made in the prior art. To this end, the multi-target decision method comprises the following steps: acquiring multi-source state data of a photovoltaic module to be evaluated, a candidate treatment path and a technical constraint rule; constructing a multi-dimensional evaluation index based on the multi-source state data; pre-screening the candidate treatment path based on the technical constraint rule and outputting a feasible treatment path; constructing a multi-target optimization model based on the feasible treatment path; iteratively solving the multi-target optimization model based on the multi-dimensional evaluation index and a first algorithm to calculate a path comprehensive optimization index of each feasible treatment path; comparing the path comprehensive optimization index of each feasible treatment path and outputting an optimal treatment path and a decision report based on the comparison result. The application improves the accuracy of the decision result.

Description

Technical Field

[0001] This application relates to the field of photovoltaic technology, specifically providing a multi-objective decision-making method, storage medium, and device for the disposal of obsolete photovoltaic modules. Background Technology

[0002] With the rapid development of the global photovoltaic industry, a large number of early-stage photovoltaic power plants are gradually entering the middle and late stages of their service life, and the problem of mass retirement of photovoltaic modules is becoming increasingly prominent. Not all retired photovoltaic modules have lost their value; some modules still have the potential to continue operating, be repaired and reused, or be downgraded and reused, while others are more suitable for dismantling and recycling or direct scrapping.

[0003] Currently, decisions regarding the disposal of photovoltaic modules typically suffer from the following shortcomings: First, the decision-making criteria are often too simplistic, relying heavily on power attenuation rates or single test results as the standard, failing to comprehensively consider the coupled effects of multiple failure modes such as hot spots, microcracks, and insulation failure. Second, the evaluation models are disconnected from engineering constraints, often neglecting the technical boundary conditions such as insulation safety and structural integrity required for the continued operation or reuse of photovoltaic modules, potentially leading to decisions that are not feasible in engineering. Finally, there is insufficient consideration of the resource recycling value and environmental impact of photovoltaic modules, lacking a quantitative assessment of the entire life cycle value of the modules.

[0004] Therefore, how to scientifically and economically determine the optimal processing path for photovoltaic modules in different states has become an urgent problem to be solved. Summary of the Invention

[0005] This application aims to solve the aforementioned technical problems, namely, to address the issues in the prior art regarding the decision-making process for the disposal of decommissioned photovoltaic modules, such as the single evaluation dimension, lack of integration with engineering and technical constraints, and inaccurate decision-making results.

[0006] In a first aspect, this application provides a multi-objective decision-making method for the disposal of obsolete photovoltaic modules. The method includes: acquiring multi-source state data, candidate processing paths, and technical constraint rules for the photovoltaic modules to be evaluated; constructing a multi-dimensional evaluation index based on the multi-source state data; pre-screening the candidate processing paths based on the technical constraint rules and outputting feasible processing paths; constructing a multi-objective optimization model based on the feasible processing paths; iteratively solving the multi-objective optimization model based on the multi-dimensional evaluation index and a first algorithm to calculate a comprehensive path optimization index for each feasible processing path; comparing the comprehensive path optimization index of each feasible processing path, and outputting the optimal processing path and a decision report based on the comparison results.

[0007] By adopting the above technical solutions, this application overcomes the one-sidedness of relying solely on a single indicator for decision-making by constructing a multi-dimensional evaluation index based on multi-source state data, making the comprehensive state evaluation of photovoltaic modules more comprehensive and objective. By setting technical constraint rules to pre-screen candidate processing paths before optimizing the decision, processing paths that are not feasible or unsafe in engineering can be effectively eliminated, avoiding the problem of the evaluation model being out of touch with engineering practice and ensuring the technical feasibility of the final decision result. By constructing a multi-objective optimization model and calculating the comprehensive path optimization index, all feasible paths can be quantitatively compared from multiple dimensions such as technical state, risk, value, and environment, thereby scientifically outputting the optimal processing path, improving the accuracy and economic rationality of the decision result, and helping to maximize the full life cycle value of retired photovoltaic modules.

[0008] Optionally, the multi-source state data includes at least one of: electrical performance parameters, insulation parameters, infrared hot spot data, electroluminescence image data, appearance aging status, and service life; and / or the candidate processing paths include at least one of: continued operation path, post-repair operation path, downgraded reuse path, dismantling and recycling path, and direct scrapping path; and / or the technical constraint rules include at least one of: insulation constraint rules, hot spot constraint rules, structural integrity constraint rules, and maintenance economic boundary constraint rules. This application, by comprehensively considering the above-mentioned multi-source state data, provides a data foundation for accurate evaluation, ensuring the accuracy of the evaluation results. Furthermore, by pre-setting the above-mentioned candidate processing paths, this application ensures a complete and comprehensive decision-making scope, making the optimization results more practically instructive. In addition, by introducing the above-mentioned hard constraint rules derived from engineering practice and safety regulations, this application directly corresponds to the engineering safety baseline, ensuring that paths that do not meet safety conditions are eliminated in advance, thus ensuring the absolute safety and technical feasibility of the decision results.

[0009] Optionally, the step "constructing a multi-dimensional evaluation index based on the multi-source state data" includes: extracting a first-level feature vector, a second-level feature vector, and a third-level feature vector based on the multi-source state data; calculating a first evaluation index based on the first-level feature vector and a second algorithm, wherein the first evaluation index is used to quantitatively characterize the health status of the photovoltaic module; calculating a second evaluation index based on the second-level feature vector and a third algorithm, wherein the second evaluation index is used to quantitatively assess the future risk of the photovoltaic module; and calculating a third evaluation index based on the third-level feature vector and a fourth algorithm, wherein the third evaluation index is used to quantitatively assess the recycling value of the photovoltaic module. This application, through hierarchical feature extraction, transforms the original data into three evaluation indices with clear physical meaning, achieving a comprehensive decoupling of the photovoltaic module's status from the three dimensions of "health-risk-recycling," forming a complete value assessment chain for a comprehensive quantitative assessment of the photovoltaic module. This makes the decision-making basis more sufficient, the results more scientific, and improves the depth and breadth of the multi-objective decision-making method evaluation.

[0010] Optionally, the primary feature vector includes at least one of the following: power retention characteristics, insulation performance characteristics, hot spot characteristics, crack characteristics, appearance characteristics, PID sensitivity characteristics, historical performance degradation characteristics, junction box fault characteristics, and backsheet failure characteristics; and / or the secondary feature vector includes at least one of the following: hazardous substance content, material separation difficulty, and hazardous dust exposure index; and / or the tertiary feature vector includes at least one of the following: type, quality, recovery rate, purity, market price, and recycling cost of recyclable materials. Through the above settings, the unique degradation modes and failure types of photovoltaic modules are accurately captured, ensuring the professionalism and completeness of the health status assessment of photovoltaic modules, and making the recycling value assessment truly reliable.

[0011] Optionally, the second algorithm includes: HI = ∑ k ω k ·d k d k =(|x k1 -x k2 |) / x k2 Where HI is the first evaluation index, ω k d represents the weight coefficient of the k-th first-order eigenvector. k x represents the degradation degree of the k-th first-level eigenvector. k1 Let x be the measured value of the k-th first-order eigenvector. k2This is the expected value of the k-th first-level eigenvector. This application calculates the first evaluation index using a second algorithm, employing a degradation degree format. This standardizes features of different dimensions, facilitating weighted aggregation and achieving precise quantification of the health status of photovoltaic modules. It intuitively reflects the health level of photovoltaic modules, making it easier for engineers to understand and apply.

[0012] Optionally, the third algorithm includes: FRI = ∑ j λ j ·f(y j ), f(y j )= y j / y limit ; where FRI is the second evaluation index, λ j y is the weight coefficient of the j-th secondary feature vector. j y is the measured value of the j-th second-order feature vector; limit The regulatory standard limit is the value of the j-th secondary feature vector. This application employs a third algorithm to directly compare the secondary feature vector with the regulatory standard limit, thereby quantifying the relative severity of each risk indicator. This provides a legal basis for risk assessment, making it no longer a subjective judgment but an objective calculation with evidence and laws to follow, thus enhancing the compliance and authority of the multi-objective decision-making method.

[0013] Optionally, the fourth algorithm includes: RVI=(∑ i (m i ·η i ·γ i ·ρ i )-C r ) / P ref Among them, RVI is the third evaluation index, m i For the mass of the i-th type of recyclable material, η i Let γ be the recovery rate of the i-th type of recyclable material. i ρ is the purity correction factor for the i-th type of recyclable material. i Let C be the market price of the i-th type of recyclable material. r To recover costs, P ref This serves as a benchmark for reference value. By employing a fourth algorithm, this application can accurately calculate the net revenue of the dismantling and recycling path, and can dynamically respond to market price fluctuations and advancements in recycling technology, ensuring the accuracy of the calculation results of the third evaluation index.

[0014] Optionally, the first algorithm includes: DPI = α·HI - β·FRI + χ·RVI + δ·EVI + ε·NPV; where DPI is the comprehensive path optimization index, HI is the first evaluation index, α is the weight coefficient of the first evaluation index, FRI is the second evaluation index, β is the weight coefficient of the second evaluation index, RVI is the third evaluation index, χ is the weight coefficient of the third evaluation index, EVI is the environmental value index, δ is the weight coefficient of the environmental value index, NPV is the net present value index, and ε is the weight coefficient of the net present value index. This application, by employing the first algorithm, integrates health, risk, recycling value, environmental value, and net present value to achieve a unified "economic-environmental-technical" multi-objective, realizing scientific, transparent, and customizable decision-making.

[0015] In a second aspect, this application provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, it implements a multi-objective decision-making method for photovoltaic module obsolescence processing as described in any one of the above descriptions.

[0016] In a third aspect, this application provides an apparatus, characterized in that the apparatus comprises: a memory storing a computer program; and a processor communicatively connected to the memory, which, when invoked, executes the multi-objective decision-making method for photovoltaic module scrapping as described above. Attached Figure Description

[0017] The preferred embodiments of this application are described below with reference to the accompanying drawings, in which: Figure 1 This is a flowchart illustrating a multi-objective decision-making method for the disposal of obsolete photovoltaic modules provided in an embodiment of this application; Figure 2 This is a flowchart illustrating the process of constructing a multi-dimensional evaluation index in this application embodiment; Figure 3 This is a schematic diagram of the structure of a device provided in an embodiment of this application. Detailed Implementation

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

[0019] To address the technical problems in existing technologies regarding the decision-making process for decommissioned photovoltaic (PV) modules, such as a single evaluation dimension, lack of integration with engineering constraints, and inaccurate decision results, this application provides a multi-objective decision-making method for the disposal of PV modules. The method includes: acquiring multi-source state data of the PV modules to be evaluated, candidate processing paths, and technical constraint rules; constructing a multi-dimensional evaluation index based on the multi-source state data; pre-screening candidate processing paths based on the technical constraint rules and outputting feasible processing paths; constructing a multi-objective optimization model based on the feasible processing paths; iteratively solving the multi-objective optimization model based on the multi-dimensional evaluation index and a first algorithm to calculate the comprehensive path optimization index for each feasible processing path; comparing the comprehensive path optimization indices of each feasible processing path, and outputting the optimal processing path and a decision report based on the comparison results.

[0020] like Figure 1 As shown, in one or more embodiments, this application provides a multi-objective decision-making method for the disposal of photovoltaic modules. Its core lies in decomposing the decision-making process for the disposal of photovoltaic modules into a series of logically clear, data-driven steps. This multi-objective decision-making method for the disposal of photovoltaic modules includes the following steps S10-S60.

[0021] Step S10: Obtain multi-source state data, candidate processing paths, and technical constraint rules for the photovoltaic modules to be evaluated. This provides comprehensive and fundamental input information for decision-making. The aforementioned information forms the cornerstone of decision-making, ensuring the comprehensiveness and accuracy of subsequent analysis. This method emphasizes the "multi-source" nature of the data, thereby enabling a more complete characterization of the module.

[0022] Specifically, multi-source condition data includes at least one of the following, but is not limited to: electrical performance parameters (e.g., power retention rate, open-circuit voltage, short-circuit current, maximum power point voltage, maximum power point current, fill factor, etc.), insulation parameters (e.g., insulation resistance, wet leakage current, etc.), infrared hotspot data (e.g., maximum temperature rise, hotspot area, etc.), electroluminescence image data (e.g., crack type, percentage of electroluminescence defect area, etc.), appearance aging conditions (e.g., appearance defects, backsheet cracking, EVA yellowing, junction box and connector abnormalities, etc.), and service life. Alternatively, multi-source condition data also includes: environmental data (e.g., humidity, temperature, salt spray level) and recycling parameters. This specific data collection allows the evaluation to delve into the physical mechanisms of module failure, enabling earlier and more accurate identification of potential problems in photovoltaic modules. By integrating this data, this application effectively solves the problem of misjudgment caused by the single evaluation basis in existing technologies, providing a data foundation for accurate evaluation.

[0023] Furthermore, candidate processing paths include, but are not limited to, at least one of the following: continued operation path, post-repair operation path, downgraded reuse path, dismantling and recycling path, and direct scrapping path. The aforementioned paths cover all possible processing methods for retired photovoltaic modules, from high-value utilization to end-of-life disposal. For example, the "downgraded reuse" path can utilize photovoltaic modules in applications with low performance requirements, such as off-grid lighting and agricultural photovoltaics, maximizing their value. By pre-setting the aforementioned candidate processing paths, this application ensures a complete and comprehensive decision-making scope, making the optimization results more practically instructive.

[0024] In addition, the technical constraint rules include, but are not limited to, at least one of the following: insulation constraint rules, hot spot constraint rules, structural integrity constraint rules, and maintenance economic boundary constraint rules. For example, the insulation constraint rule can be set to automatically exclude continued operation paths and degraded reuse paths when the insulation resistance of the photovoltaic module is below a certain safety threshold (such as the limit specified in the IEC standard). The hot spot constraint rule can be set to automatically exclude continued operation paths when the temperature rise of a hot spot detected by infrared detection exceeds a certain limit, potentially leading to high risks such as "bypass diode failure" or "exacerbated hot spot effect." When the temperature rise of a hot spot detected by infrared detection exceeds a higher limit, the post-repair operation path and degraded reuse path are automatically excluded. The structural integrity constraint rule can be set to exclude all reuse-related paths when the photovoltaic module exhibits defects affecting basic safety and protection performance, such as through-cracks or significant pulverization of the backsheet or severe deformation of the frame. The maintenance economic boundary constraint rule can be set to automatically exclude post-repair operation paths when the repair cost exceeds the sum of the new revenue and the net recovery revenue after repair. This application introduces the aforementioned hard constraints derived from engineering practice and safety standards, directly corresponding to the bottom line of engineering safety, ensuring that paths that do not meet safety conditions are eliminated in advance, thus ensuring the absolute safety and technical feasibility of the decision-making results.

[0025] Step S20: Construct a multi-dimensional evaluation index based on multi-source state data. This step is one of the core technologies of this method. It does not simply list the raw data, but rather uses data mining and feature engineering to extract complex underlying data into high-level semantic indices that directly reflect the core state of the module. By constructing a multi-dimensional evaluation system, such as assessing health status, future risks, and recycling value from different perspectives, it overcomes the one-sidedness of relying solely on a single indicator for decision-making, making the comprehensive state evaluation of photovoltaic modules more comprehensive and objective.

[0026] like Figure 2 As shown, in one or more embodiments, step S20 "constructing a multi-dimensional evaluation index based on multi-source state data" includes the following steps: S21, extracting first-level feature vectors, second-level feature vectors and third-level feature vectors based on multi-source state data respectively.

[0027] Specifically, the primary eigenvector refers to features concerning the current health status of photovoltaic (PV) modules. Primary eigenvectors include, but are not limited to, at least one of the following: power retention characteristics, insulation performance characteristics, hot spot characteristics, crack characteristics, appearance characteristics, PID sensitivity characteristics, historical performance degradation characteristics, junction box failure characteristics, and backsheet failure characteristics. These features are all recognized key health indicators in the PV field, covering multiple aspects such as electrical performance, safety performance, and structural performance. They accurately capture the unique degradation modes and failure types of PV modules, ensuring the professionalism of PV module health status assessments. Furthermore, the secondary eigenvector refers to features concerning the future risks and environmental impacts of PV modules. Secondary eigenvectors include, but are not limited to, at least one of the following: hazardous substance content (e.g., lead, cadmium), material separation difficulty (affecting energy consumption and secondary pollution during recycling), and hazardous dust exposure index (e.g., silicon-containing dust that may be generated during dismantling). These features allow decision-making to move beyond economic and technical considerations, incorporating environmental, health, and social responsibility (EHS) factors to ensure the completeness of health assessments. Finally, the tertiary eigenvector refers to features concerning the recycling value of PV modules. The third-level feature vector includes, but is not limited to, at least one of the following: type of recyclable material (e.g., glass, aluminum, silicon, silver, copper, etc.), quality, recovery rate, purity, market price, and recycling cost. These features directly determine the economic feasibility of the dismantling and recycling path. By modeling these dynamically changing parameters, the assessment of recycling value is no longer a static estimate, but a dynamic result closely linked to the market and technology. This comprehensive coverage of the economic factors of resource recycling ensures a true and reliable evaluation of recycling value.

[0028] S22. A first evaluation index is calculated based on the first-level feature vector and the second algorithm. The first evaluation index is used to quantitatively characterize the health status of the photovoltaic module. For example, the first evaluation index is used to quantitatively characterize the current comprehensive technical performance and remaining lifespan of the photovoltaic module.

[0029] Specifically, the primary feature vector mainly contains features that directly reflect the current performance and physical state of the module. These features are then integrated by the second algorithm to derive a comprehensive health score. This approach provides a clear quantitative basis for judging the health status of photovoltaic modules. Furthermore, the second algorithm can employ a specific mathematical model. The second algorithm includes HI = ∑ k ω k ·d k d k =(|x k1 -x k2 |) / x k2 ; where HI is the primary evaluation index, ω k Let d be the weight coefficient of the k-th first-order eigenvector. kLet x be the degradation degree of the k-th first-order eigenvector. k1 Let x be the measured value of the k-th first-order eigenvector. k2 Let be the expected value of the k-th first-level eigenvector. The formula above does not directly use the absolute value of the feature, but rather calculates the relative deviation (i.e., the degree of degradation d) between the measured value and the expected value (e.g., the factory nominal value or the average value of healthy components in the same batch). k This allows for a fairer evaluation of the degradation status of different characteristic parameters. For example, a 10% decrease in power and a 10% decrease in insulation resistance have different severity; through normalization and weighting, a more reasonable comprehensive evaluation can be made. This application calculates a first evaluation index using a second algorithm, adopting a degradation degree form (the degree of deviation of the measured value from the expected value). This standardizes and unifies characteristics of different dimensions, facilitating weighted synthesis and achieving accurate quantification of the health status of photovoltaic modules. It intuitively reflects the health level of photovoltaic modules, making it easier for engineers to understand and apply.

[0030] In one or more embodiments, the value of the first evaluation index HI can be controlled within the range of 0 to 1. Exemplarily, the value of the first evaluation index HI can be any one of 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, or 1. The specific values ​​of the first evaluation index HI given above are merely illustrative; any value within the range of 0 to 1 is within the scope of protection of this application. Furthermore, the smaller the value of the first evaluation index HI, the better the health status of the photovoltaic module.

[0031] In one or more examples, the weight coefficient ω of the k-th first-order eigenvector k The weighting factor can be controlled within the range of 0.01 to 0.3. For example, when the first-level feature vector is a power retention rate feature, the weighting factor can be 0.28. When the first-level feature vector is an insulation performance feature, the weighting factor can be 0.2. When the first-level feature vector is a hot spot feature, the weighting factor can be 0.16. When the first-level feature vector is a crack feature, the weighting factor can be 0.14. When the first-level feature vector is an appearance feature, the weighting factor can be 0.08. When the first-level feature vector is a backplane failure feature, the weighting factor can be 0.06. When the first-level feature vector is a junction box fault feature, the weighting factor can be 0.05. When the first-level feature vector is a historical performance degradation feature, the weighting factor can be 0.03. It should be noted that the above weighting factors are only illustrative examples. In practical applications, they can be adaptively adjusted according to component type, application scenario, and power plant operation and maintenance needs, enhancing the adaptability and flexibility of the method.

[0032] S23. A second evaluation index is calculated based on the secondary feature vector and the third algorithm. The second evaluation index is used to quantitatively assess the future risks of photovoltaic modules. For example, the second evaluation index is used to quantitatively assess the probability and severity of power drop and safety accidents caused by potential defects (such as microcrack propagation, hot spot deterioration, and insulation breakdown) during the continued operation or reuse of photovoltaic modules in the future.

[0033] Specifically, the secondary feature vector mainly contains features that foreshadow potential future problems, such as the environmental risks that hazardous substances in the material may pose during long-term use, or the risk of microcracks propagating under stress. The future risks of photovoltaic modules are then quantitatively assessed through the calculation of the third algorithm, adding a risk prediction dimension to the multi-objective decision-making method. Furthermore, the third algorithm can adopt the following model: FRI = ∑ j λ j ·f(y j ), f(y j )= y j / y limit ; where FRI is the second evaluation index, λ j Let y be the weight coefficient of the j-th second-order eigenvector. j Let y be the measured value of the j-th second-order eigenvector; limit Let y be the regulatory standard limit for the j-th secondary feature vector. The third algorithm directly links the measurement of future risk to the authoritative "regulatory standard limit". For example, when the secondary feature vector is the content of hazardous substances, y... limit It can be the maximum permissible concentration stipulated by national environmental protection regulations. When the measured value y... j When approaching or exceeding the limit, f(y) j The value of the risk factor will increase rapidly, thus significantly increasing the value of the second evaluation index (FRI), highlighting high-risk factors and providing a clear threshold basis for pre-screening. This application employs a third algorithm to directly compare the secondary feature vector with regulatory standard limits, thereby quantifying the relative severity of each risk indicator. This gives risk assessment a legal basis, making it no longer a subjective judgment but an objective calculation with evidence and legal procedures, enhancing the compliance and authority of the multi-objective decision-making method.

[0034] In one or more examples, a higher value for the second evaluation index (FRI) indicates a higher risk of continued operation or reuse of the photovoltaic modules. Furthermore, the weight coefficient λ of the j-th secondary eigenvector... j It can be controlled within the range of 0.01 to 0.4. Weighting coefficient λ j It can be adjusted according to risk appetite, balancing conservative and flexible decision-making.

[0035] S24. A third evaluation index is calculated based on the three-level feature vector and the fourth algorithm. The third evaluation index is used to quantitatively assess the recycling value of photovoltaic modules. For example, the third evaluation index is used to quantitatively assess the net economic value that can be obtained by dismantling and recycling photovoltaic modules.

[0036] Specifically, the third-level feature vector mainly contains features related to the material composition and recycling economics of photovoltaic modules. The recycling value of photovoltaic modules is then evaluated through calculation using a fourth algorithm, integrating the concept of a circular economy into the decision-making model. Furthermore, the fourth algorithm can adopt the following model: RVI = (∑ i (m i ·η i ·γ i ·ρ i )-C r ) / P ref Among them, RVI is the third evaluation index, m i For the mass of the i-th type of recyclable material, η i Let γ be the recovery rate of the i-th type of recyclable material. i ρ is the purity correction factor for the i-th type of recyclable material. i Let C be the market price of the i-th type of recyclable material. r To recover costs, P ref Using this as a benchmark, the above calculation formula constructs a complete recycling value accounting model. It considers not only the revenue from recycled materials (based on mass m) but also... i Recovery rate η i Market price ρ i The decision also introduced a purity correction factor γ. i This reflects the impact of recycling technology on material value, minus the total recycling cost C. r Finally, by referring to benchmark P ref (For example, the cost of new photovoltaic modules) is normalized. This application, by employing a fourth algorithm, can accurately calculate the net revenue of the dismantling and recycling path, and can dynamically respond to market price fluctuations and advancements in recycling technology, ensuring the accuracy of the calculation results of the third evaluation index, RVI.

[0037] It should be noted that the fourth algorithm introduces a purity correction coefficient γ. i and market price ρ i This allows the evaluation of recycling value to be dynamically adjusted based on material purity, recycling rate, market price, and dismantling cost, making the evaluation of dismantling and recycling paths more realistic and closer to reality.

[0038] This application transforms raw data into three evaluation indices with clear physical meaning through hierarchical feature extraction, achieving a comprehensive decoupling of the photovoltaic module status from the three dimensions of "health-risk-recycling" and forming a complete value assessment chain to conduct a comprehensive quantitative assessment of photovoltaic modules. This makes the decision-making basis more sufficient and the results more scientific, and improves the depth and breadth of the multi-objective decision-making method evaluation.

[0039] Step S30: Pre-screen the candidate processing paths based on the technical constraint rules and output the feasible processing paths.

[0040] Specifically, in practical engineering, some processing paths may perform well in economic models, but they may be completely unfeasible due to safety red lines (e.g., insufficient insulation performance) or technical limitations (e.g., severe structural damage that cannot be repaired). This application avoids invalid calculations in the subsequent optimization process by pre-eliminating infeasible processing paths, and ensures that the final output decision is safe, reliable, and implementable in engineering practice, thus guaranteeing the engineering feasibility of the decision results.

[0041] Step S40: Construct a multi-objective optimization model based on feasible processing paths.

[0042] Specifically, the above steps formalize and model complex decision-making problems, making it possible to solve them using mathematical tools. This multi-objective optimization model can simultaneously consider multiple conflicting or related objectives such as economic benefits, environmental impact, resource utilization, and implementation risks, demonstrating the comprehensiveness and forward-looking nature of multi-objective decision-making methods.

[0043] Step S50: Iteratively solve the multi-objective optimization model based on the multi-dimensional evaluation index and the first algorithm to calculate the path comprehensive optimization index for each feasible processing path.

[0044] Specifically, in this step, the multi-dimensional evaluation index constructed in the previous steps serves as the core input, quantitatively supporting the solution process of the optimization model. The first algorithm, acting as the solution engine, iteratively calculates and weighs various objectives, ultimately providing a comprehensive "score" for each feasible path, namely the path comprehensive optimization index. This quantification process makes the merits of different processing paths intuitively comparable. Furthermore, the first algorithm, used to calculate the final path comprehensive optimization index (DPI), can employ a linear weighted model. The first algorithm includes: DPI = α·HI - β·FRI + χ·RVI + δ·EVI + ε·NPV; where DPI is the path comprehensive optimization index, HI is the first evaluation index, α is the weight coefficient of the first evaluation index, FRI is the second evaluation index, β is the weight coefficient of the second evaluation index, RVI is the third evaluation index, χ is the weight coefficient of the third evaluation index, EVI is the environmental value index, δ is the weight coefficient of the environmental value index, NPV is the net present value index, and ε is the weight coefficient of the net present value index. The above calculation formula is the core of the entire decision-making method. In the above formula, the previously calculated first evaluation index HI, second evaluation index FRI, and third evaluation index RVI are used as inputs, and additional evaluation dimensions are introduced: Environmental Value Index EVI (e.g., the quantified value of carbon emission reduction achieved through continued power generation or material recycling) and Net Present Value Index NPV (a financial assessment of future cash flows from different treatment paths). The positive and negative signs in the formula reflect the attributes of different indicators: HI, RVI, EVI, and NPV are benefit-oriented indicators, the higher the better, hence positive; while FRI is a cost- or risk-oriented indicator, the higher the worse, hence negative. The weighting coefficients α, β, χ, δ, and ε can be set by staff according to decision-making preferences (e.g., prioritizing economic or environmental benefits), providing good flexibility and interpretability. Through the first algorithm, a complex multi-objective decision-making problem is transformed into a clear and computable comprehensive index ranking problem, which integrates health, risk, recycling value, environmental value, and net present value, achieving a unified "economic-environment-technology" multi-objective, and realizing scientific, transparent, and customizable decision-making.

[0045] In one or more embodiments, the weighting coefficients α, β, χ, δ, and ε can all be controlled within the range of 0.01 to 0.4.

[0046] It should be noted that the key to the Path Optimization Index (DPI) lies not in the formula itself, but in the fact that the DPI is a path decision quantity built on the photovoltaic module's specific technical indices (HI, FRI, RVI), which fully reflects the coupling relationship between the unique failure modes and resource characteristics of photovoltaic modules.

[0047] Step S60: Compare the comprehensive optimization index of each feasible processing path, and output the optimal processing path and decision report based on the comparison results.

[0048] Specifically, the path with the highest comprehensive optimization index is the optimal processing path recommended by the system. Simultaneously, the generated decision report can include detailed information such as the photovoltaic module's health level, risk level, resource recycling value level, recommended processing path, economic calculation results, and unrecommended processing paths and their reasons. This provides staff with transparent and traceable decision-making basis, enhancing the scientific rigor and credibility of the decisions.

[0049] In one or more embodiments, the multi-objective decision-making method for photovoltaic module scrapping in this application not only makes refined decisions for individual photovoltaic modules, but also makes group decisions for large batches of photovoltaic modules, which greatly improves asset management and decision-making efficiency in the scenario of decommissioning large photovoltaic power plants.

[0050] On the other hand, in one or more embodiments, this application also provides a storage medium (not shown in the figures) storing a computer program, which, when executed by a processor, implements the multi-objective decision-making method for photovoltaic module scrapping as described in any of the above embodiments. The storage medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination thereof. This allows the solution of this application to be conveniently distributed, installed, and deployed in the form of a software product.

[0051] On the other hand, such as Figure 3 As shown, in one or more embodiments, this application also provides a device 100, which can be a server, industrial control computer, personal computer, or dedicated decision analyzer, but is not limited thereto. Device 100 includes a memory 110, a processor 120, and a display 130. The memory 110 can be a non-volatile storage medium, such as a hard disk, solid-state drive (SSD), or flash memory. The memory 110 stores computer programs. The processor 120 can be a central processing unit (CPU) or a microcontroller (MCU). The processor 120 is communicatively connected to the memory 110 and, when calling the computer program, executes the multi-objective decision-making method for photovoltaic module scrapping as described in any of the above embodiments. The display 130 is communicatively connected to the memory 110 and the processor 120, and is used to display the relevant GUI interactive interface of the multi-objective decision-making method for photovoltaic module scrapping.

[0052] To illustrate the technical solution of this application more specifically, a complete embodiment integrating the above-mentioned optimal technical features will be presented below. This embodiment will demonstrate the entire process of processing a specific decommissioned photovoltaic module, from raw data input to final decision output.

[0053] Suppose a crystalline silicon photovoltaic module has been in service at a photovoltaic power station for 11 years and a decision needs to be made regarding its disposal. First, multi-source status data is obtained. Through on-site testing and historical data queries, the following information is obtained: power retention rate is 78.6% (relative to the factory nominal value); insulation resistance test result is 50MΩ, meeting the safety lower limit for continued operation under a specific voltage; an infrared thermal imager detected a hot spot on its surface, with a maximum temperature rise of 19℃; high-resolution electroluminescence (EL) testing revealed severe cracks (such as Class III or higher cracks) in its internal cells, covering a cumulative area of ​​12% of the total area; visual inspection revealed slight cracks in the backsheet, but not penetrating, and the EVA encapsulation material showed no obvious yellowing; the junction box and its cable connectors appeared intact, with no signs of overheating. Simultaneously, database queries show that this module model does not contain high-risk hazardous substances (such as cadmium), and data on the quality, current market price, recycling rate, and recycling cost of its main recyclable materials (glass, aluminum frame, silicon, copper, silver) are also obtained.

[0054] Secondly, based on the above data, three core evaluation indices are constructed. First, the first evaluation index (HI) is constructed. The system extracts power retention rate, insulation performance, hot spot severity, EL crack ratio, and backsheet aging level as primary feature vectors. Then, the second algorithm is applied, i.e., HI = ∑ k ω k ·d k d k =(|x k1 -x k2 |) / x k2 For example, for the power retention characteristic, its expected value x k2 100%, measured value x k1 If the degradation rate is 78.6%, then its degradation degree d k The result is (100-78.6) / 100 = 0.214. After performing similar degradation calculations and normalization on all primary characteristics, the HI value of the module is obtained by weighted summation according to preset weighting coefficients (e.g., the weighting coefficient for power retention is 0.28, the weighting coefficient for insulation performance is 0.20, etc.), which intuitively indicates that the health status of the photovoltaic module is at a medium-to-high level.

[0055] Secondly, a second evaluation index (FRI) is constructed. The system extracts hazardous substance content, material separation difficulty, etc., as secondary feature vectors. A third algorithm is then applied, i.e., FRI = ∑ j λ j ·f(y j ), f(y j )= y j / y limit Because this component does not contain high-risk hazardous substances, its yj Well below the regulatory limit y limit Therefore, the risk score for this item is very low. After comprehensively considering the long-term risks that hot spots (temperature rise of 19℃) may bring and the risk of crack propagation, the FRI value of the photovoltaic module is calculated to be 0.34 through weighted calculation, indicating that its future risk is low.

[0056] Secondly, a third evaluation index (RVI) is constructed. The mass m of glass, aluminum, silicon, silver, and copper is extracted. i Market price ρ i Recovery rate η i These are used as third-level feature vectors. The fourth algorithm is applied, i.e., RVI = (∑ i (m i ·η i ·γ i ·ρ i )-C r ) / P ref Substituting the parameters into the formula, its net recovery value is calculated, and then compared and normalized with the cost of the new component, resulting in an RVI value of 0.41, indicating that it has a moderate resource recovery value.

[0057] Secondly, after calculating the three core indices HI=0.68, FRI=0.34, and RVI=0.41, the system enters the path pre-screening stage. Candidate processing paths include: continued operation path, post-repair operation path, downgraded reuse path, dismantling and recycling path, and direct scrapping path. The system applies technical constraint rules for checking: insulation performance (50MΩ) meets the lower operating limit, rule passed; hot spot temperature rise (19℃) does not exceed the threshold prohibiting continued operation (e.g., 20℃), rule passed; the proportion of severe EL crack area (12%) is lower than the threshold prohibiting downgraded reuse (e.g., 15%), rule passed. Therefore, the continued operation path, post-repair operation path, downgraded reuse path, dismantling and recycling path, and direct scrapping path are all considered technically feasible.

[0058] Finally, the system comprehensively optimizes all feasible paths. The first algorithm is applied: DPI = α·HI - β·FRI + χ·RVI + δ·EVI + ε·NPV. For each processing path, the system evaluates its corresponding EVI (Environmental Value) and NPV (Net Present Value). For example, a path that continues operating has a higher NPV, but its EVI is generally lower considering performance degradation; a path that requires maintenance (such as replacing junction boxes or reinforcing backplanes) requires initial investment, affecting its NPV, but may extend its lifespan and increase its EVI; the NPV of a dismantling and recycling path mainly comes from material sales revenue, while its EVI comes from the carbon emissions reduced by resource recycling. Substituting HI, FRI, RVI, and the EVI and NPV of each processing path into the formula, and using preset weighting coefficients (α, β, χ, δ, ε), the DPI value of each path is calculated. For example, the calculation results are as follows: the DPI value for the continued operation path is 0.57, the DPI value for the post-repair operation path is 0.66, the DPI value for the downgraded reuse path is 0.61, the DPI value for the dismantling and recycling path is 0.49, and the DPI value for the direct scrapping path is 0.20. By comparing the DPI values ​​of all processing paths, the system found that the post-repair operation path, with a score of 0.66, has the highest score. Therefore, the final decision report output by the system will clearly recommend "continued operation after repair" as the optimal processing path, and will include a detailed analysis process, including the HI, FRI, and RVI scores of the component, as well as the reasons for the lower scores of other paths (e.g., although the continued operation path is feasible, its long-term risks and benefits are not as good as post-repair; although the dismantling and recycling path has recycling value, its overall benefits are not as high as those of continued power generation, etc.). Through this complete, end-to-end quantitative analysis process, the method described in this application finds an optimal outcome for the component that balances technology, economy, risk, and value, demonstrating its scientific nature and practicality.

[0059] The application scenarios of this application are not limited to the refined decision-making of individual photovoltaic modules. It can better leverage its high efficiency when handling batches of decommissioned modules in large-scale photovoltaic power plants. In a typical application scenario, a power plant with tens of thousands of modules about to be decommissioned needs to formulate an overall handling strategy. Performing the aforementioned detailed analysis on each module would be extremely time-consuming and costly. In this case, the group decision-making mechanism provided by this method can be used. First, rapid and critical feature data is collected from all photovoltaic modules to be processed within the power plant. For example, batch scanning is performed using drones equipped with infrared and EL devices to obtain core data such as the power retention rate, maximum hot spot temperature rise, and EL crack area ratio of each photovoltaic module. Then, these photovoltaic modules are used as data points and input into a clustering algorithm model (e.g., the K-Means algorithm). The system automatically divides thousands of modules into several "groups" with similar states based on these key features. For example, three groups might be clustered: Group G1, characterized by "high power retention rate, no obvious hot spots, and minor cracks," can be defined as the "healthy group"; Group G2, characterized by "medium power retention rate, presence of low-temperature hot spots, and moderate cracks," can be defined as the "sub-healthy group"; and Group G3, characterized by "low power retention rate, presence of high-temperature hot spots, and severe cracks," can be defined as the "critical group." Subsequently, the decision-making process is no longer based on individual components but on the group as a whole. For each group, several representative samples can be extracted for detailed testing, or the group's characteristic mean (centroid) can be directly used to calculate the group's average HI, FRI, and RVI. Then, these average indices are substituted into the subsequent decision-making process (path pre-screening and comprehensive optimization) to derive a unified optimal treatment recommendation for the entire group. Ultimately, the decision might be as follows: for the thousands of components in group G1 (healthy group), a unified "continue to operate" strategy will be adopted; for the components in group G2 (sub-healthy group), "repair and then operate" or "downgrade and reuse" will be recommended; and for the components in group G3 (critically ill group), a unified "dismantling and recycling" path will be adopted. This "clustering and grouping first, then representative decision-making" model simplifies the complexity of large-scale decision-making, improves processing efficiency while ensuring decision quality, and provides owners of large photovoltaic assets with a decommissioning management tool. Furthermore, this method can also be applied to various scenarios such as photovoltaic asset transaction valuation, insurance claim assessment, and intelligent sorting lines in recycling plants.

[0060] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0061] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0062] The embodiments, implementation methods, and related technical features of this application can be combined and substituted for each other without conflict.

[0063] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. Without departing from the principles of this application, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of this application.

Claims

1. A multi-objective decision-making method for the disposal of obsolete photovoltaic modules, characterized in that, include: Obtain multi-source state data, candidate processing paths, and technical constraint rules for the photovoltaic modules to be evaluated; A multi-dimensional evaluation index is constructed based on the multi-source state data; Based on the aforementioned technical constraint rules, the candidate processing paths are pre-screened, and feasible processing paths are output. A multi-objective optimization model is constructed based on the feasible processing path; Based on the multi-dimensional evaluation index and the first algorithm, the multi-objective optimization model is iteratively solved to calculate the path comprehensive optimization index for each feasible processing path. The comprehensive optimization index of each feasible processing path is compared, and the optimal processing path and decision report are output based on the comparison results.

2. The multi-objective decision-making method for photovoltaic module scrapping according to claim 1, characterized in that, The multi-source status data includes at least one of the following: electrical performance parameters, insulation parameters, infrared hot spot data, electroluminescence image data, appearance aging status, and service life; and / or The candidate processing paths include at least one of the following: continued operation path, post-repair operation path, downgraded reuse path, dismantling and recycling path, and direct scrapping path; and / or The technical constraint rules include at least one of the following: insulation constraint rules, hot spot constraint rules, structural integrity constraint rules, and maintenance economic boundary constraint rules.

3. The multi-objective decision-making method for photovoltaic module scrapping according to claim 1 or 2, characterized in that, The step "Constructing a multi-dimensional evaluation index based on the multi-source state data" includes: Based on the multi-source state data, first-level feature vectors, second-level feature vectors, and third-level feature vectors are extracted respectively; The first evaluation index is calculated based on the first-level feature vector and the second algorithm. The first evaluation index is used to quantitatively characterize the health status of the photovoltaic module. The second evaluation index is calculated based on the second-level feature vector and the third algorithm. The second evaluation index is used to quantitatively assess the future risks of photovoltaic modules. The third evaluation index is calculated based on the three-level feature vector and the fourth algorithm. The third evaluation index is used to quantitatively evaluate the recycling value of photovoltaic modules.

4. The multi-objective decision-making method for photovoltaic module scrapping according to claim 3, characterized in that, The primary feature vector includes at least one of the following: power retention rate feature, insulation performance feature, hot spot feature, crack feature, appearance feature, PID sensitivity feature, historical performance degradation feature, junction box fault feature, and backplane failure feature; and / or The secondary feature vector includes at least one of: hazardous substance content, material separation difficulty, and hazardous dust exposure index; and / or The three-level feature vector includes at least one of the following: type of recyclable material, quality, recovery rate, purity, market price, and recycling cost.

5. The multi-objective decision-making method for photovoltaic module scrapping according to claim 4, characterized in that, The second algorithm includes: HI = ∑ k ω k ·d k d k =(|x k1 -x k2 |) / x k2 ; Where HI is the first evaluation index, ω k d represents the weight coefficient of the k-th first-order eigenvector. k x represents the degradation degree of the k-th first-level eigenvector. k1 Let x be the measured value of the k-th first-order eigenvector. k2 This is the expected value of the k-th first-level eigenvector.

6. The multi-objective decision-making method for photovoltaic module scrapping according to claim 4, characterized in that, The third algorithm includes: FRI=∑ j λ j ·f(y j ), f(y j )= y j / y limit ; Wherein, FRI is the second evaluation index, λ j y is the weight coefficient of the j-th secondary feature vector. j y is the measured value of the j-th second-order feature vector; limit The regulatory standard limit for the j-th secondary feature vector is given.

7. The multi-objective decision-making method for photovoltaic module scrapping according to claim 4, characterized in that, The fourth algorithm includes: RVI=(∑ i (m i ·η i ·γ i ·ρ i )-C r ) / P ref ; Among them, RVI is the third evaluation index, m i For the mass of the i-th type of recyclable material, η i Let γ be the recovery rate of the i-th type of recyclable material. i ρ is the purity correction factor for the i-th type of recyclable material. i Let C be the market price of the i-th type of recyclable material. r To recover costs, P ref As a benchmark for reference value.

8. The multi-objective decision-making method for photovoltaic module scrapping according to claim 1, characterized in that, The first algorithm includes: DPI = α·HI - β·FRI + χ·RVI + δ·EVI + ε·NPV; Wherein, DPI is the comprehensive route optimization index, HI is the first evaluation index, α is the weight coefficient of the first evaluation index, FRI is the second evaluation index, β is the weight coefficient of the second evaluation index, RVI is the third evaluation index, χ is the weight coefficient of the third evaluation index, EVI is the environmental value index, δ is the weight coefficient of the environmental value index, NPV is the net present value index, and ε is the weight coefficient of the net present value index.

9. A storage medium, characterized in that, A computer program is stored on the storage medium, and when the computer program is executed by a processor, it implements a multi-objective decision-making method for the disposal of photovoltaic modules according to any one of claims 1-8.

10. A device, characterized in that, The device includes: a memory storing a computer program; and A processor, which is communicatively connected to the memory, executes the multi-objective decision-making method for photovoltaic module obsolescence processing according to any one of claims 1-8 when the computer program is invoked.

Images