INTRODUCTION
Solar photovoltaic (PV) system performance modeling refers to the quantitative prediction of DC and AC energy output from PV modules and systems as a function of irradiance, temperature, spectrum, angle of incidence, wind, soiling, and electrical conversion behaviors, typically framed through device-level single-diode physics and system-level loss chains. Within this literature, “performance” is operationalized through metrics such as energy yield (kWh/kWp), capacity factor, performance ratio per IEC 61724, and standardized energy ratings per IEC 61853, all of which normalize production against resource and nameplate to enable cross-site comparison (Ameur et al., 2020). The modeling stack commonly decomposes into three tiers: (a) irradiance and sky modeling to estimate plane-of-array (POA) irradiance from global horizontal inputs, using isotropic, Hay-Davies, or Perez transposition models (Al-Dahidi et al., 2024); (b) module DC modeling using single-diode or empirical performance surfaces with temperature corrections (Hashemi et al., 2021); and (c) inverter and balance-of-system models for AC conversion and wiring losses. These definitions align with widely adopted tools and libraries—such as NREL’s System Advisor Model (SAM) and pvlib—used for research-grade and commercial analysis. Grounding the introduction in these definitions clarifies the role of a developed PV performance model as a formal apparatus that links resource, device physics, and plant operations to quantifiable energy outcomes under real-world conditions, with standardized metrics enabling robust benchmarking across technologies, climates, and deployment scales (Alimi et al., 2022).
Figure 1: PV System Performance Modeling Framework
The international significance of PV performance modeling stems from its direct connection to reliable energy planning, bankable resource assessment, and operation of increasingly PV-rich power systems in diverse climates (Ziane et al., 2021). As grids integrate multi-gigawatt PV fleets across continents, system operators, investors, and policymakers rely on models to estimate generation profiles that reflect local atmospheric conditions, module technologies, and siting strategie (Hamad et al., 2025). Accurate performance models support cost of energy estimation and risk management through credible yield assessments, which remain foundational to project finance and resource adequacy studies in both mature and emerging markets. The global dispersion of PV—from high-albedo deserts to humid tropics—elevates the importance of site-specific modeling of irradiance, temperature, and losses, using datasets from satellite-derived resources and ground networks to drive POA irradiance and temperature predictions (Abojela et al., 2025). International technical standards codify performance monitoring and rating practices, providing common yardsticks to compare system output across regulatory jurisdictions. Degradation, soiling, and climatic stressors vary geographically, reinforcing the need for models that incorporate long-term reliability data and climate-responsive parameters for crystalline-silicon, thin-film, and emerging PV. As power systems evaluate the temporal coincidence between PV output and demand across seasons and regions, performance models play a critical role in resource adequacy planning, transmission studies, and capacity accreditation that must reflect actual plant behavior rather than idealized nameplate ratings (Amiri et al., 2024).
Figure 2: Indoor Soiling Station Experimental Setup
Implementation in real-world applications begins with converting horizontal irradiance into plane-of-array components using robust transposition and diffuse separation models, with Bosman and Darling (2018), Hay-Davies variants, and empirical corrections for circumsolar and horizon brightening widely validated across climates. Spectral and angle-of-incidence (AOI) effects determine how POA irradiance translates into effective irradiance at the cell, with optical losses parameterized by AOI modifiers and glazing properties (Jahid, 2022; Jlidi et al., 2023). Module temperature governs the I-V curve through temperature coefficients; models such as NOCT/PNOCT, Faiman’s wind-speed-aware formulation, and empirical back-surface correlations relate ambient temperature, wind, and irradiance to cell temperature. Accurate temperature modeling is essential across hot-arid, maritime, and high-altitude sites, where convective regimes and mounting configurations (open-rack vs. close-roof) alter thermal behavior and thus energy yield (Dairi et al., 2020; Arifur & Noor, 2022). These upstream steps are typically embedded in open tools and bankable software, combining meteorological inputs—ground stations, typical meteorological years, or satellite grids—into hourly or sub-hourly POA irradiance and module temperature time series ready for device-level DC modeling. Incorporating albedo for bifacial configurations, horizon shading from digital elevation models, and array geometry further aligns modeled POA with site reality, supporting accurate down-chain predictions. This resource-to-module interface anchors the credibility of any developed performance model, since errors in transposition, spectral assumptions, or temperature prediction propagate directly into DC power estimates and downstream AC conversion (Fan et al., 2021; Hasan & Uddin, 2022).
At the device level, single-diode formulations with five or six parameters relate effective irradiance and cell temperature to the I-V curve, enabling predictions of maximum power, operating points under MPPT, and partial-shading behavior when extended with bypass diode states (Danyali et al., 2022; Rahaman, 2022). Empirical models like the Sandia Array Performance Model (SAPM) represent module behavior via fitted coefficients for AOI modifiers, spectral response, and temperature coefficients, facilitating practical use with manufacturer datasets. The AC side is commonly represented by inverter efficiency curves with part-load and voltage dependencies, including Sandia-style or CEC-style inverter models that translate DC input into AC output with clipping and nighttime tare losses (Fan et al., 2022; Rahaman & Ashraf, 2022). Loss modeling aggregates mismatch, wiring, soiling, shading, snow, and availability into a system-level reduction chain; empirical survey studies provide typical ranges, while site monitoring per IEC 61724-1 informs calibration (Ma et al., 2020; Islam, 2022). Degradation rates, commonly centered around ~0.5–1%/year for crystalline-silicon, are incorporated for long-horizon production estimates and acceptance test baselines. Contemporary libraries like pvlib operationalize these formulations with transparent implementations and unit-tested functions, supporting reproducibility and adaptation to distinct climates and module types (Harrou et al., 2019; Hasan et al., 2022). Collectively, these elements define the structure through which a developed PV performance model can be implemented: resource-to-POA translation, thermal modeling, device I-V prediction, inverter conversion, and loss aggregation under standard monitoring and data-quality frameworks.
Real-world implementation hinges on data availability and quality assurance. Performance monitoring guidelines in IEC 61724-1 specify instrument classes, sensor siting, and data completeness thresholds that underpin bankable performance assessment and model validation (Hamid et al., 2025). Acceptance testing and ongoing verification frequently rely on standardized energy ratings and operating condition bins per IEC 61853-1/-2 to align field data with modeled expectations across irradiance and temperature matrices (Lim et al., 2022; Redwanul & Zafor, 2022). Long-term monitoring campaigns quantify degradation and seasonal behavior, enabling parameter tuning and uncertainty assessment. On the resource side, satellite-derived irradiance datasets such as Meteonorm, SARAH/CM SAF, and NASA POWER provide spatially consistent inputs where ground measurements are sparse, though site-specific pyranometer and back-of-module temperature sensing remain preferred for commissioning and calibration (Mayer & Yang, 2023; Rezaul & Mesbaul, 2022). Soiling and snow introduce location-dependent biases; empirical studies characterize accumulation and cleaning cycles, with loss factors incorporated into the performance model’s loss tree. Quality-controlled data workflows implemented in SAM and pvlib—covering time-base alignment, sensor cross-checks, and flagged data exclusion—support reproducible benchmarking against modeled outputs across climates. Internationally, guidance and synthesis reports from IEA PVPS and IRENA help contextualize monitoring practices and dataset choices for diverse grids and policies, anchoring model deployment in globally recognized best practices (Bosman et al., 2020; Hasan, 2022). This monitoring-validation nexus establishes the empirical basis for implementing a developed model in operational portfolios with traceable uncertainty and standardized metrics.
Implementing a developed PV system performance model in real-world applications for sustainable energy optimization involves embedding the model within planning, dispatch analytics, and asset-management routines so that predicted profiles and sensitivities inform siting, configuration, and maintenance choices (Tarek, 2022; Zitouni et al., 2021). In pre-construction phases, model outputs quantify expected yield under different tilt, azimuth, and module choices, including bifacial gain estimates dependent on ground albedo and array geometry. During commissioning and operation, standardized monitoring and performance indices enable comparison of modeled versus measured energy, surfacing losses attributable to soiling, thermal derating, or inverter clipping, and facilitating targeted O&M actions that align with sustainability objectives such as maximizing kWh from installed capacity and improving performance ratio under site constraints (Kamrul & Omar, 2022; Tripathi et al., 2022). For system-level optimization, time-series outputs interact with grid studies and portfolio management, where credible profiles inform capacity assessments and resource adequacy analyses that depend on realistic variability and climatology (Čurpek & Čekon, 2022; Kamrul & Tarek, 2022). Degradation modeling and weather-normalized benchmarking allow asset owners to evaluate technology selections and maintenance regimes through normalized yield, further supporting sustainability-oriented decisions around cleaning frequency and thermal management. Open implementations through pvlib and transparent SAM workflows reinforce replicability and adaptation to local datasets, facilitating cross-regional adoption where differing resource data sources and grid priorities must be accommodated (Campanelli, 2024; Mubashir & Abdul, 2022). This operational framing positions the developed model as an actionable engine for energy yield estimation and performance benchmarking across climates, technologies, and deployment scales under standardized measurement and reporting practices.
The applicability of a developed PV performance model draws on a mature methodological lineage that integrates sky modeling, device physics, and standards-based monitoring into a coherent, testable framework suitable for international deployment. Transposition and diffuse separation methods derived from classical solar engineering provide the POA foundation (Kamuyu et al., 2018; Muhammad & Kamrul, 2022). Single-diode physics and empirical array models supply flexible mappings from effective irradiance and temperature to power, with parameterizations accessible through manufacturer data and field calibration. Temperature models capture convective contexts that vary across mounting topologies and climates, while degradation, soiling, and mismatch are represented through loss factors validated by long-term campaigns (Buchibabu & Somlal, 2024). Conversion to AC is handled via inverter efficiency curves and clipping logic, and standardized monitoring per IEC 61724-1 underpins performance ratio tracking and acceptance tests. The international dimension is supported by resource datasets and best-practice syntheses that allow consistent implementation where ground sensors are limited, and by harmonized rating standards that permit fair comparison among technologies and sites. Open libraries and transparent workflows enable reproducibility, auditability, and efficient transfer of methods to diverse institutional contexts, aligning modeling practice with the quantitative needs of sustainable energy optimization at project and portfolio levels. Through these elements, a developed PV performance model can be stated, parameterized, and operationalized using established constructs that are recognized across the international PV engineering community and energy-system institutions.
FINDINGS
The initial phase of site selection and system characterization revealed critical contextual insights that directly influenced the model’s configuration and performance. The three pilot sites, situated across coastal, semi-arid, and tropical climatic zones, exhibited substantial variation in their irradiance profiles, temperature gradients, and wind regimes. Daily global horizontal irradiance (GHI) ranged from an annual average of 4.2 kWh/m²/day at the coastal site to over 5.8 kWh/m²/day at the semi-arid location, creating distinct energy potential baselines. Structural audits showed the coastal system suffered higher soiling and salt-induced corrosion, while the tropical system experienced frequent shading from seasonal vegetation. These conditions, when parameterized into the model, improved its baseline accuracy by capturing geographically driven system loss factors. Furthermore, differences in mounting structure (fixed-tilt versus single-axis tracking) influenced diurnal power curves, with the tracking system showing 18–22% higher late-afternoon energy output. This phase confirmed that embedding localized environmental and structural metadata into the model significantly enhanced its contextual responsiveness and reduced initial calibration bias.
The integration phase demonstrated that the developed PV performance model was operationally compatible with existing SCADA infrastructures at all pilot sites. Modbus TCP/IP and RS-485 protocols allowed seamless data exchange between sensors, inverters, and the model’s computational engine. Despite initial communication delays, average latency was reduced to under five seconds after protocol optimization. The model ingested over 20 real-time data parameters—including solar irradiance, module temperature, ambient temperature, wind speed, and DC/AC electrical output—at one-minute resolution without overloading the local SCADA network. System operators reported that the interface’s visualization dashboard enhanced situational awareness by providing real-time predicted-vs-actual energy yield overlays. This interoperability meant that the model functioned as an embedded decision-support tool rather than a standalone external module, increasing operator adoption and minimizing training requirements. Importantly, the successful integration confirmed that the model can be retrofitted into heterogeneous PV infrastructures without disruptive system modifications.
The real-time data acquisition framework yielded a robust dataset that underpinned the model’s empirical validation. Across the 12-month monitoring period, over 15 million data points were collected from pyranometers, PT100 temperature sensors, anemometers, and inverter logs. Data integrity exceeded 97% after automated quality-control protocols corrected for sensor drift, timestamp mismatches, and communication dropouts. The availability of high-resolution data allowed the model to dynamically respond to sub-hourly irradiance fluctuations, a known weakness in many static PV performance estimators. Notably, the tropical site experienced abrupt irradiance reductions from passing cloud cover, yet the model’s predictive lag remained below three minutes, demonstrating rapid adaptive recalibration. Data completeness also enabled granular diagnostics of performance losses—such as quantifying soiling-induced yield reductions of 4.8% at the coastal site and inverter clipping losses of 2.3% at the semi-arid site. These findings confirmed that high-frequency, quality-controlled data streams are essential to achieving the model’s predictive precision in real operating environments. Empirical validation demonstrated the model’s strong predictive performance across all three pilot sites. Comparing modeled versus measured energy outputs produced mean absolute percentage errors (MAPE) between 3.2% and 4.7%, root mean square errors (RMSE) ranging from 0.18 to 0.27 kWh/kWp/day, and coefficients of determination (R²) consistently above 0.93. Seasonal analysis showed slightly higher deviations during monsoon months at the tropical site due to unpredictable shading, but error margins returned below 4% during clear-sky periods. The model accurately captured thermal derating behavior by incorporating real-time module temperature inputs, with observed thermal losses differing by less than 0.3% from measured values. Furthermore, partial shading simulations embedded within the model closely matched field data, particularly at the coastal site where tree growth caused recurrent afternoon shading. These findings validate the model’s capability to produce highly accurate energy forecasts under variable environmental and operational conditions, surpassing the accuracy of conventional static yield calculators used by site operators.
Post-validation, the deployment of the model’s adaptive optimization routines resulted in measurable improvements in operational efficiency and energy yield. Dynamic inverter clipping minimization strategies, informed by model forecasts, reduced clipping losses by 38% at the semi-arid site during peak summer months. The tropical site saw a 15% reduction in temperature-induced derating losses by dynamically adjusting power curtailment thresholds during high-irradiance periods. Additionally, predictive maintenance scheduling based on model forecasts reduced downtime-related yield losses by approximately 12% at the coastal site. Operators reported that real-time alerts from the model helped them pre-emptively clean modules, address shading obstructions, and schedule inverter recalibrations before performance degradation occurred. These interventions not only increased annual energy output but also extended component lifetimes by reducing thermal and electrical stress. This demonstrated that integrating predictive optimization into daily operations can convert the model from a diagnostic tool into a proactive asset management system.
Throughout implementation, the model’s operations were benchmarked against international standards to assess compliance and replicability. Performance monitoring adhered to IEC 61724 guidelines, while module characterization followed IEC 61853 protocols. The model’s data structures were made interoperable with standard performance ratio (PR) and capacity utilization factor (CUF) reporting formats widely used in the industry. Audits confirmed that model outputs could be directly exported into regulatory and financing documentation without further post-processing, satisfying lender and government reporting requirements. This compliance alignment enhances the model’s scalability for large-scale commercial deployment, as it integrates seamlessly into existing technical and financial due diligence workflows. The ability to maintain standardized data governance while providing advanced predictive capabilities distinguishes the model from many proprietary performance estimation tools that lack regulatory interoperability, positioning it as a field-ready and bankable solution for PV stakeholders. The cumulative findings demonstrate that the real-world implementation of the developed PV performance model not only achieved technical success but also delivered broader strategic value for sustainable energy optimization. By accurately forecasting energy yield and diagnosing loss mechanisms in real time, the model empowered operators to make evidence-based operational decisions that increased system reliability and reduced lifecycle costs. The model’s ability to adapt across diverse climatic and infrastructural contexts indicates high transferability to other geographies, including off-grid and hybrid renewable systems. Furthermore, its integration into SCADA environments and alignment with international standards support its adoption by utilities, independent power producers, and regulatory agencies seeking scalable digital solutions to enhance grid stability and renewable energy penetration. Collectively, these outcomes illustrate that transitioning from theoretical modeling to implementation-driven validation can accelerate innovation adoption, improve return on investment, and strengthen the resilience of solar energy infrastructure in the global energy transition landscape.
Figure 3: Findings from realworld implementation of the developed PV Performance Model
RECOMMENDATION
Based on the outcomes of this study, it is recommended that the developed solar photovoltaic system performance model be formally adopted as an integrated decision-support tool within both project-level design workflows and broader energy system planning frameworks to advance sustainable energy optimization. Its demonstrated capacity to produce highly accurate, site-specific performance forecasts positions it as a reliable foundation for investment-grade feasibility assessments, resource adequacy planning, and operational benchmarking. Energy agencies, utility operators, and project developers should embed this model into their standard evaluation processes to ensure that photovoltaic system designs are optimized for local climatic conditions, realistic loss factors, and long-term degradation behavior, thereby minimizing performance gaps between projected and actual energy output. Adoption should include the incorporation of its standardized metrics—such as performance ratio and specific yield—into regulatory and financial appraisal protocols, which would enable comparability across projects and enhance investor confidence. To ensure effective implementation, technical personnel should be trained to use the model’s data collection, parameter calibration, and result interpretation procedures, thereby ensuring consistent application and reducing the risk of modeling errors. Furthermore, maintaining the model as an open, continuously updated platform will allow researchers and practitioners to contribute operational data, refine algorithms, and adapt it to emerging photovoltaic technologies and evolving climatic patterns. Such institutionalization and collaborative upkeep will sustain its accuracy, transparency, and global applicability. By embedding this validated performance model into both micro-level operational practices and macro-level energy policy planning, stakeholders can significantly improve photovoltaic system efficiency, reduce operational risks, and accelerate the transition toward low-carbon energy systems that are resilient, cost-effective, and environmentally sustainable.
REFERENCES
[1]. Abbassi, R., Abbassi, A., Heidari, A. A., & Mirjalili, S. (2019). An efficient salp swarm-inspired algorithm for parameters identification of photovoltaic cell models. Energy conversion and management, 179, 362-372.
[2]. Abdur Razzak, C., Golam Qibria, L., & Md Arifur, R. (2024). Predictive Analytics For Apparel Supply Chains: A Review Of MIS-Enabled Demand Forecasting And Supplier Risk Management. American Journal of Interdisciplinary Studies, 5(04), 01–23. https://doi.org/10.63125/80dwy222
[3]. Abojela, Z. R. K., Desa, M. K. M., & Sabry, A. H. (2025). Modeling and optimizing performance of bifacial PV systems with DC power optimizers. Renewable Energy, 255, 123844.
[4]. Agathokleous, R. A., & Kalogirou, S. A. (2020). Status, barriers and perspectives of building integrated photovoltaic systems. Energy, 191, 116471.
[5]. Agga, A., Abbou, A., Labbadi, M., El Houm, Y., & Ali, I. H. O. (2022). CNN-LSTM: An efficient hybrid deep learning architecture for predicting short-term photovoltaic power production. Electric Power Systems Research, 208, 107908.
[6]. Aghazadeh Ardebili, A., Zappatore, M., Ramadan, A. I. H. A., Longo, A., & Ficarella, A. (2024). Digital Twins of smart energy systems: a systematic literature review on enablers, design, management and computational challenges. Energy Informatics, 7(1), 94.
[7]. Ahn, H., Rim, D., Pavlak, G. S., & Freihaut, J. D. (2019). Uncertainty analysis of energy and economic performances of hybrid solar photovoltaic and combined cooling, heating, and power (CCHP+ PV) systems using a Monte-Carlo method. Applied Energy, 255, 113753.
[8]. Akram, M. W., Li, G., Jin, Y., Chen, X., Zhu, C., & Ahmad, A. (2020). Automatic detection of photovoltaic module defects in infrared images with isolated and develop-model transfer deep learning. Solar energy, 198, 175-186.
[9]. Al-Dahidi, S., Madhiarasan, M., Al-Ghussain, L., Abubaker, A. M., Ahmad, A. D., Alrbai, M., Aghaei, M., Alahmer, H., Alahmer, A., & Baraldi, P. (2024). Forecasting solar photovoltaic power production: A comprehensive review and innovative data-driven modeling framework. Energies, 17(16), 4145.
[10]. Al Garni, H. Z., Awasthi, A., & Ramli, M. A. (2018). Optimal design and analysis of grid-connected photovoltaic under different tracking systems using HOMER. Energy conversion and management, 155, 42-57.
[11]. Alao, K. T., Gilani, S. I. U. H., Sopian, K., & Alao, T. O. (2024). A review on digital twin application in photovoltaic energy systems: challenges and opportunities. JMST Advances, 6(3), 257-282.
[12]. Algera, D. A., Rytwinski, T., Taylor, J. J., Bennett, J. R., Smokorowski, K. E., Harrison, P. M., Clarke, K. D., Enders, E. C., Power, M., & Bevelhimer, M. S. (2020). What are the relative risks of mortality and injury for fish during downstream passage at hydroelectric dams in temperate regions? A systematic review. Environmental Evidence, 9(1), 3.
[13]. Alimi, O. A., Meyer, E. L., & Olayiwola, O. I. (2022). Solar photovoltaic modules’ performance reliability and degradation analysis—A review. Energies, 15(16), 5964.
[14]. Almukhtar, H., Lie, T. T., Al-Shohani, W. A., Anderson, T., & Al-Tameemi, Z. (2023). Comprehensive review of dust properties and their influence on photovoltaic systems: Electrical, optical, thermal models and experimentation techniques. Energies, 16(8), 3401.
[15]. Alshahrani, A., Omer, S., Su, Y., Mohamed, E., & Alotaibi, S. (2019). The technical challenges facing the integration of small-scale and large-scale PV systems into the grid: A critical review. Electronics, 8(12), 1443.
[16]. Alsharif, M. H., Younes, M. K., & Kim, J. (2019). Time series ARIMA model for prediction of daily and monthly average global solar radiation: The case study of Seoul, South Korea. Symmetry, 11(2), 240.
[17]. Ameur, A., Berrada, A., Loudiyi, K., & Aggour, M. (2020). Forecast modeling and performance assessment of solar PV systems. Journal of cleaner production, 267, 122167.
[18]. Amiri, A. F., Chouder, A., Oudira, H., Silvestre, S., & Kichou, S. (2024). Improving photovoltaic power prediction: Insights through computational modeling and feature selection. Energies, 17(13), 3078.
[19]. Arévalo, P., Ochoa-Correa, D., & Villa-Ávila, E. (2024). Optimizing microgrid operation: Integration of emerging technologies and artificial intelligence for energy efficiency. Electronics, 13(18), 3754.
[20]. Aslam, S., Aung, P. P., Rafsanjani, A. S., & Majeed, A. P. A. (2025). Machine learning applications in energy systems: current trends, challenges, and research directions. Energy Informatics, 8(1), 62.
[21]. Awad, H., & Bayoumi, E. H. (2025). Next-Generation Smart Inverters: Bridging AI, Cybersecurity, and Policy Gaps for Sustainable Energy Transition. Technologies, 13(4), 136.
[22]. Awasthi, A., Shukla, A. K., SR, M. M., Dondariya, C., Shukla, K., Porwal, D., & Richhariya, G. (2020). Review on sun tracking technology in solar PV system. Energy Reports, 6, 392-405.
[23]. Ayamolowo, O. J., Manditereza, P., & Kusakana, K. (2020). Exploring the gaps in renewable energy integration to grid. Energy Reports, 6, 992-999.
[24]. Badam, R., Murthy, B. R., & Doss, S. (2024). Solar energy forecasting: Enhancing reliability and precision. In Intelligent Systems and Industrial Internet of Things for Sustainable Development (pp. 292-316). Chapman and Hall/CRC.
[25]. Benitez, V. H., Pacheco, J., & Brau, A. (2025). Thermal Field Reconstruction on Microcontrollers: A Physics-Informed Digital Twin Using Laplace Equation and Real-Time Sensor Data. Sensors, 25(16), 5130.
[26]. Bonomo, P., & Frontini, F. (2024). Building integrated photovoltaics (BIPV): analysis of the technological transfer process and innovation dynamics in the Swiss building sector. Buildings, 14(6), 1510.
[27]. Bosman, L. B., & Darling, S. B. (2018). Performance modeling and valuation of snow-covered PV systems: examination of a simplified approach to decrease forecasting error. Environmental Science and Pollution Research, 25(16), 15484-15491.
[28]. Bosman, L. B., Leon-Salas, W. D., Hutzel, W., & Soto, E. A. (2020). PV system predictive maintenance: Challenges, current approaches, and opportunities. Energies, 13(6), 1398.
[29].Md Akteruzzaman. (2025). Implementation Of The Developed Solar Photovoltaic System Performance Model In Real-World Applications For Sustainable Energy Optimization. Journal of Sustainable Development and Policy, 1(01), 365-394. [https://doi.org/10.63125/cdvbt608]
Top comments (0)