Photovoltaic Output Power Research Articles

Evaluating and analyzing the performance of PV power output forecasting using different models of machine-learning techniques considering prediction accuracy

Solar energy as a clean, renewable, and sustainable energy source has considerable potential to meet global energy needs. However, the intermittent and uncertain character of the solar energy source makes the power balance management a very challenging task. To overcome these shortcomings, providing accurate information about future energy production enables better planning, scheduling, and ensures effective strategies to meet energy demands. The present paper aims to assess the performance of PV power output forecasting in PV systems using various machine learning models, such as artificial neural networks (ANN), linear regression (LR), random forests (RF), and Support Vector Machines (SVM). These learning algorithms are trained on two different datasets with different time steps: in the first one, a historical weather forecast with a one hour time step, and in the second one, a dataset of on-site measurements with a 5-minute time step. To provide a reliable estimation of prediction accuracy for different learning algorithms, a k-fold cross-validation (CV) is applied. Through a comparison analysis, an assessment of the accuracy of these algorithms based on various metrics such as RMSE, MAE, and MRE is performed, providing a detailed evaluation of their performance. Results obtained from this study demonstrate that the random forest algorithm (RF) outperformed other algorithms in predicting PV output, achieving the smallest prediction error, where the best values for RMSE, MRE, MAE, and R² for the weather dataset were 0.856 W, 0.256%, 0.364 W, and 0.99999, respectively, while thevalues for RMSE, MRE, MAE, and R² for the on-site measurements dataset were 8.525 W, 11.163%, 3.922 W, and 0.99922, respectively.

Advanced Automated Machine Learning Framework for Photovoltaic Power Output Prediction Using Environmental Parameters and SHAP Interpretability

Advanced Automated Machine Learning Framework for Photovoltaic Power Output Prediction Using Environmental Parameters and SHAP Interpretability

Photovoltaic Short-Term Output Power Forecast Model Based on Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise–Kernel Principal Component Analysis–Long Short-Term Memory

To solve the problem of photovoltaic power prediction in areas with large climate changes, this article proposes a hybrid Long Short-Term Memory method to improve the prediction accuracy and noise resistance. It combines the improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN) and kernel principal component analysis (KPCA) algorithm. The ICEEMDAN algorithm reduces the instability of the environmental factor sequence. The KPCA algorithm reduces the input dimensions of the model. LSTM performs dynamic time modeling of the multivariate feature sequences to predict the output PV power. The adaptability of the ICEEMDAN-KPCA-LSTM model is assessed with datasets from a PV plant in west China and evaluated by root mean squared error (RMSE), mean absolute percentage error (MAPE), and R-squared metrics. Using 70% of the datasets for output PV power estimation, the results show a good performance, with an RMSE of 4.3715, MAPE of 8.9264%, and R-squared value of 89.973%. By comparing with other prediction models, the ICEEMDAN-KPCA-LSTM photovoltaic output power model outperforms other models.

A Robust Salp Swarm Algorithm for Photovoltaic Maximum Power Point Tracking Under Partial Shading Conditions

Currently, numerous intelligent maximum power point tracking (MPPT) algorithms are capable of tackling the global optimization challenge of multi-peak photovoltaic output power under partial shading conditions, yet they often face issues such as slow convergence, low tracking precision, and substantial power fluctuations. To address these challenges, this paper introduces a hybrid algorithm that integrates an improved salp swarm algorithm (SSA) with the perturb and observe (P&O) method. Initially, the SSA is augmented with a dynamic spiral evolution mechanism and a Lévy flight strategy, expanding the search space and bolstering global search capabilities, which in turn enhances the tracking precision. Subsequently, the application of a Gaussian operator for distribution calculations allows for the adaptive adjustment of step sizes in each iteration, quickening convergence and diminishing power oscillations. Finally, the integration with P&O facilitates a meticulous search with a small step size, ensuring swift convergence and further mitigating post-convergence power oscillations. Both the simulations and the experimental results indicate that the proposed algorithm outperforms particle swarm optimization (PSO) and grey wolf optimization (GWO) in terms of convergence velocity, tracking precision, and the reduction in iteration power oscillation magnitude.

Feasibility Assessment of a Sustainable Building Applied Photovoltaic (BAPV) System With Solar Tracking Features Based on Techno-Economic Criteria: Malaysia Case Study

In line with the increase attention regarding building applied photovoltaic system, the research in enhancing its performance is also critical to be explored, one of them is the utilization of solar tracking system. The goal of this paper is to analyze the feasibility of using a tracking system for increasing the PV output power and analyzing the energy sustainability aspects on a building applied photovoltaic system. A feasibility study considering three scenarios of PV capacities and propose two options in either renewing or reusing the PV panels to extend its life time. Based on the results, the tracking system contributes to the improvement of 18% in the energy generation. On the technical aspects, renewing the PV panels with a large capacity of up to 30 kW-peak is the most feasible option to meet the electricity demand on the building. In the meantime, the reuse option with large capacity is the best with the net present value is about USD 24,639.22 or 34% more than the option of renewal the PV panels. In summary, reusing the panel is beneficial from the economic and environment standpoints as it represents sustainable development of power generation. Since the realistic feasibility cases and analyses used, the novelty in terms of the approach is beneficial for the future investigations on adding a tracking system, keeping or replacing the PV system for sustainable aspect while improving its output power.

An innovative hybrid model combining informer and K‐Means clustering methods for invisible multisite solar power estimation

AbstractThe employment of behind‐the‐meter solar photovoltaic (PV) systems has gained increasing popularity in recent years as more individuals and organizations aim to reduce their reliance on conventional grid‐connected power sources and take advantage of the environmental and economic benefits of solar power. However, precisely estimating the potential output of PV systems is a challenging task, since most of the PV systems used in residential properties have been installed behind the meter. Consequently, electric power companies are limited to accessing only the recorded net electricity consumption. This article introduces an innovative approach to estimate behind‐the‐meter PV power generation within a large region, utilizing a limited representative subset. The proposed framework integrates Missforest, that is, a robust tool for missing data imputation, with a hybrid application of K‐Means, Pearson Correlation Coefficient, and Principal Component Analysis, for the precise selection of representative PV sites. Additionally, it leverages the Informer model, a cutting‐edge deep learning‐based time series model, to link the relationship between the PV power generation at representative sites and the total PV power output on the entire region. To conduct a case study, the power output of 367 PV sites and solar radiation measured at 105 weather stations in Taiwan were collected and analyzed. The application of this comprehensive methodology demonstrates a notable advancement in the estimation of “invisible” PV power generation in comparison to other established techniques.

Multistep photovoltaic power forecasting based on multi-timescale fluctuation aggregation attention mechanism and contrastive learning

The integration of photovoltaic (PV) power into electrical grids introduces significant uncertainty due to the inherent volatility and intermittency of solar energy, underscoring the need for precise short and medium-term PV power forecasting. Despite the superior performance of Transformer-based time series methods, their application to PV power prediction remains suboptimal. In response to this deficiency, this paper proposes a novel attention mechanism that aggregates fluctuations across multiple time scales. This mechanism enhances the segmentation and extraction of nonlinear correlations between PV power outputs and meteorological factors, assigning variable weights to patterns of change across different time scales. Furthermore, a novel approach for selecting similar days is also developed based on contrastive learning, which enables self-supervised identification of similarities among PV power samples and enhances the model’s attention to local dynamic variations. Comparative analysis with eight state-of-the-art benchmark methods shows that the proposed MFA-attention model achieves lower prediction errors and improved effectiveness.© 2017 Elsevier Inc. All rights reserved.

Fine-grained simulation model for PV power output interval based on two-stage scenario clustering and dual-ensemble compatible learning

Photovoltaic (PV) power simulation is indispensable for the planning and operation of renewable-rich power systems. Among various simulation methods, data-driven indirect simulation methods are the most efficient, primarily involving scenario extraction and power output generation. Typical scenario extraction methods often feature coarse characterizations, which also cannot be expressed by traditional power output generation models lacking the ability of quantifying the uncertainty of the output results. This paper proposes a fine-grained simulation model for PV power output interval, i.e., upper bound and lower bound, based on two-stage clustering of multiple typical scenarios and dual-ensemble compatible learning. The first stage of scenario extraction, considering the impact of total daily irradiance on PV output, employs a line distance index for preliminary scenario division of daily PV output. These scenarios are further divided into five characteristic scenarios with an angular distance index considering the impact of intra-day irradiance variations on power output. PV power output interval is generated from dual-ensemble compatible learning including Stacking and Bagging, which are adaptively selected with a Bias-Variance Selection Index (BS) calculated using the value of scenario characteristics. Stacking and Bagging are driven by three neural network base learners with Bayesian layers, which can output the lower bound and upper bound with confidence at a certain level, thus quantifying the uncertainty of the result accurately. The actual data of Daxing and Gangjialing power stations are used in case study. Results form case study show that compared with the existing methods, the proposed method can improve the Power Coverage Rate (PCR) and Average Power Simulation Accuracy (APA) of uncertainty simulation results by 27.44% and 5.39%, respectively. Meanwhile, the Average Daily Maximum Power Simulation Accuracy (AMPA) is similar to that of existing methods.

Clearness index cluster analysis for photovoltaic weather classification based on solar irradiation measurement data: Theory and application

With the rapid expansion of photovoltaic (PV) power generation, accurately predicting and evaluating PV power output remains a critical challenge due to the inherent variability of solar irradiation. This paper proposes a novel weather classification approach based on the clearness index (CI), derived from solar irradiation data collected in Zhangbei, China, a region abundant in solar resources. Three clustering methods, which are Gaussian Mixture Model, K-Means, and Fuzzy C-Means (FCM), were applied to classify weather into three types: sunny, semi-cloudy, and cloudy. Our findings indicate that FCM consistently outperformed the other methods, particularly under variable conditions. By employing the classified data, short-term PV power forecasting is significantly enhanced. A hybrid CNN-LSTM model trained on the weather-clustered data achieved superior prediction accuracy compared to models trained without weather classification. Integrating weather classification into PV power prediction can notably reduce prediction errors and improve grid management. The proposed approach is recommended for future evaluations of solar applications in Zhangbei and other regions with similar climatic characteristics.

Enhanced operation of PVWPS based on advanced soft computing optimization techniques

This study introduces three soft computing (SC) optimization algorithms aimed at enhancing the efficiency of photovoltaic water pumping systems (PVWPS). These algorithms include the Gorilla Troop Algorithm (GTO), Honey Badger Algorithm (HBA), and Snake Algorithm (SAO). The goal of the SC optimizers is to maximize the output power of the PV array (PPV) and enhance the efficiency of the DC motor (η), thereby optimizing the water flow rate (Q) of the pumping system. The analytical modeling approach proposed in this study involves forecasting the optimal duty cycle (Dop) for a buck-boost converter, taking into account variables such as solar radiation (G) and ambient temperature (T). A comparative analysis is conducted between the suggested SC optimizers and analytical modeling. MATLAB simulation is employed to explore an adaptive neuro-fuzzy inference system (ANFIS) trained for the proposed system. The objective is to assess system performance and accuracy. Findings indicate a strong convergence between the analytical model and the simulation model utilizing SC optimizers. Moreover, the neuro-fuzzy system trained offline, coupled with the proposed SC optimizers, demonstrates superior performance compared to traditional control methods like perturb and observe (P&O) and incremental conductance (IC). This superiority is evident across various metrics including motor efficiency (η), photovoltaic (PV) output power (PPV), water flow rate (Q), and time response.

Accurate Method for Solar Power Generation Estimation for Different PV (Photovoltaic Panels) Technologies

In 2023, solar photovoltaic energy alone accounted for 75% of the global increase in renewable capacity. Moreover, this natural energy resource is the one that requires the least investment, which makes it accessible to developing countries. Increasing return on investment in these regions requires a particular evaluation of environmental parameters influencing PV systems performance. Higher temperatures decrease PV module efficiency and, as a result, their power output. Additionally, fluctuations in solar irradiance directly impact the energy generated by these systems. Consequently, it is essential for investors to improve accurate predictive models that assess the power generation capacity of photovoltaic systems under local environmental conditions. Therefore, accurate estimation of maximum power generation is then crucial for optimizing photovoltaic (PV) system performances and selecting suitable PV modules for specific climates. In this context, this study presents an experimental comparison of three maximum power prediction methods for four PV module types (amorphous silicon, monocrystalline silicon, micromorphous silicon, and polycrystalline silicon) under real outdoor conditions. Experimental data gathered over the course of a year are analyzed and processed for the four PV technologies. Three different methods taking into account environmental parameters are presented and analyzed. The first estimation method utilizes irradiance as the primary input parameter, while two additional methods incorporate ambient temperature and PV module temperature for enhanced accuracy. The performance of each method is evaluated using standard statistical metrics, including the root mean square error (RMSE) and coefficient of determination (R2). The results demonstrate the effectiveness of all three methods, with RMSE values ranging from 1.6 W to 3.8 W and R2 values consistently above 0.95. The most appropriate method for estimating PV power output is determined by the specific type of photovoltaic module and the availability of meteorological parameters. This study provides valuable insights for selecting an appropriate maximum power prediction method and choosing the most suitable PV module for a given climate.

Multi‐level interval rolling warning method for distributed photovoltaic fluctuation events

AbstractThe power fluctuation of distributed photovoltaic (PV) systems significantly impacts the balance of the power system, leading to risks like PV curtailment and load shedding. This paper proposes a multi‐level rolling warning method for distributed PV power fluctuation (DPPF) based on interval analysis, aiming to establish a framework for proactively mitigating the potential adverse effects of fluctuations in distributed PV systems. Firstly, the power control mechanism to deal with DPPF is clarified, and warning levels are defined to determine the range of fluctuations that can be controlled by different power control measures. Secondly, based on the probability density of DPPF, the probabilities of each warning level are obtained by integrating the probability densities within each warning range. Finally, the differences in the forecasting accuracy of PV power fluctuations at different time scales are analysed, and the rolling warning of DPPF is achieved by periodically updating PV power output to adjust the warning results. Simulation results demonstrate that the proposed method identifies the thresholds for each warning range and provides warnings for different system operating conditions and PV power fluctuation events, confirming its effectiveness and applicability.

Performance Analysis of On-Grid Solar Power Plant under Various Irradiation and Load Condition

This research aims to analyze the effect of irradiation and load variations on the performance of the 1 kWp on-grid solar power plant at Ujung Pandang State Polytechnic. The research was conducted for 7 days with component testing on May 8, 2024, irradiation variation testing on May 15-17 and June 5-6, 2024, and load variation testing on June 8, 2024. MATLAB GUI was used to theoretically calculate power, PV and GTI efficiency, Yf, Yr, PR, apparent power, and load reactive power. In addition, the MATLAB GUI also displays graphs of the effect of irradiation variations on the performance of solar power plants. PSIM simulation was used to compare the PV power generated with the same irradiation and temperature. The results show that irradiation greatly affects the performance of the solar power plant, where increasing irradiation increases the output power, PV and GTI efficiency, and export power to the grid, which results in an increase in Yf, Yr, and PR. PSIM simulations support these findings by showing a positive correlation between irradiation and PV output power. Load variation also affects the Pf value, where resistive (R) loads have the best Pf (1), while inductive-capacitive (LC) loads have the worst Pf (0.37) due to the high reactive power drawn. Thus, variations in load type affect power consumption and system performance.

Optimal capacity configuration of joint system considering uncertainty of wind and photovoltaic power and demand response

AbstractTo reduce phenomenon of abandoning wind and photovoltaic power, improve the limitations of traditional methods in dealing with uncertainty of wind and photovoltaic power and system planning, and improve the optimal configuration of resources, an optimal capacity configuration of joint system considering uncertainty of wind and photovoltaic power and demand response is proposed. Firstly, using probability distribution information of wind and photovoltaic power output, the distance between actual probability distribution and forecast probability distribution is constrained based on the 1‐norm and ∞‐norm. A fuzzy set considering uncertainty probability distribution is constructed, and a two‐stage distributed robust planning model is established. The first stage involves optimizing joint system capacity for scenarios with the lowest probability of wind and photovoltaic power; the second stage builds on capacity optimization scheme from the first stage and aims to minimize operating costs through simulation optimization. Secondly, column and constraint generation is used to solve the model. Finally, constructing an example based on actual data from a power grid in Northeast China for simulation and analysis, the results show that the method achieves a balanced optimization of robustness and economy, effectively reduces carbon emissions and improves ability of the system to consume wind and photovoltaic power.

Research on stacking ensemble method for day-ahead ultra-short-term prediction of photovoltaic power

The intermittency and uncertainty of photovoltaic power output make it difficult to match real-time electricity demand, and accurate prediction of photovoltaic power is the basis for improving photovoltaic energy consumption. This article proposes a new stacking ensemble forecast model based on the time series generative adversarial network (TimeGAN-SEFM), to improve the accuracy of ultra-short-term forecasting of photovoltaic power. Firstly, a dual layer feature weighted clustering weighted fuzzy C-means clustering algorithm is proposed to accurately partition the original dataset according to weather types. Secondly, in the proposed TimeGAN-SEFM, a Particle Swarm Optimization based BP neural network is used as a meta learner to integrate three types of base learners, thereby improving the structural diversity of the model. The dataset for each type of weather is proportionally divided into multiple sub-datasets for training base learners. In response to the limitation that the amount of data used for training a single base learner is relatively small after multiple splitting of the original data, a TimeGAN augmentation network is introduced to perform multi-dimensional data augmentation on the sub-training sets, which improves the size and diversity of the dataset used for training base learners, thereby improving the predictive performance of the model. Finally, the predictive performance of the proposed TimeGAN-SEFM was evaluated using measurement data from a 20 MW photovoltaic power station in North China. The results show that TimeGAN-SEFM effectively improves the accuracy of photovoltaic prediction and helps promote real-time matching of photovoltaic energy and power demand, thereby improving the utilization of photovoltaic energy.

Collaborative Research for the Development of Localized Solar PV Output Forecasting Models for the Philippines Using Geospatial Data

Abstract. This paper presents a collaborative effort to develop localized solar photovoltaic (PV) power output (PPV) forecasting models for the Philippines using geospatial data. It underlines the importance of solar energy in the country and discusses the opportunities and challenges associated with PPV forecasting. Project SINAG, a two-year research project, aimed to develop solar PV output forecasting models through a collaborative approach with academic institutions, solar energy industries, and government agencies. Actual PPV data from 43 solar PV installations were analyzed alongside meteorological data from the PAGASA weather bureau, ERA5, AHI-8, and FY- 4A. These datasets were filtered based on a one-year period to ensure quality. The study employed SARIMAX, LSTM, and XGBoost models individually and in hybrid models to develop the forecasting models. Model performance was evaluated using root mean square error (RMSE), mean absolute error (MAE), and mean absolute percentage error (MAPE). In a case study in Baguio City, the SARIMAX model exhibited strong seasonal dependence, providing more accurate forecasts in dry seasons than in wet seasons. Additionally, the forecasting accuracy of each model (SARIMAX, LSTM, and XGBoost) varied based on the month and location of the installation, emphasizing the need for local and season-based PPV forecasting models. Despite implementation challenges, such as collaboration arrangements, bureaucratic barriers, and budget constraints, the project produced thirteen research publications and provided data for three student theses. This paper also demonstrated diverse engagements and contributions that emphasize the significance of collaborative research in conducting nationwide-scale data-driven projects.

Thermal management enhancement of building-integrated photovoltaic systems using coupled heat pipe and evaporative porous clay cooler

The building-integrated photovoltaic (BIPV) panel systems often lead to a notable decline in PV panels efficiency and lifetime due to their temperature rise because of inadequate cooling. This study investigates the performance of innovative cooling strategy of using a coupled heat pipes and porous clay cooler and compare it with other related three BIPV cooling systems configurations including a conventional BIPV cooling with and without air gap and a pure evaporative porous clay cooling. The aim is to improve the PV panel cooling, ultimately reduce the PV temperature and enhance the overall performance of the BIPV system as well as reducing the building indoor temperature and boosting energy efficiency within the building. The system model, comprising sets of transient equations for the different system configurations, was solved using MATLAB and confirmed with previous experimental findings. The results indicate that comparing with the traditional BIPV system, using BIPV/Clay cooling systems and the hybrid BIPV/Clay-heat pipe cooling systems achieved peak PV temperature reductions of up to 14 °C and 14.7 °C, respectively. Additionally, these cooling methods led to a maximum interior room temperature reduction of approximately 14 °C and an average decrease of 8 °C. Moreover, the hybrid BIPV/Clay-heat pipe cooling system demonstrates superior performance compared to traditional BIPV system, achieving the highest improvements in PV electrical efficiency, output power, and exergy efficiency, with gains of 7.8 %, 6.4 %, and 8.4 %, respectively. Further, when the hybrid BIPV/Clay-heat pipe cooling system was employed, the clay cooling efficiency and clay exergy efficiency values improved on average by 30.2 % and 29.7 %, respectively, compared to the BIPV/Clay cooling system in addition to the increase of the lifetime of the PV panels due to isolating it from the contact with the water/water vapor of the clay cooling system. However, this hybrid approach resulted in a rise in electricity production costs from 0.077 $/kWh to roughly 0.138 $/kWh and extended the payback time from 7.38 years to 13.6 years. But, despite its initial economic drawbacks, the hybrid BIPV/Clay-HP cooling system demonstrates considerable effectiveness and long lifetime compared to other cooling configurations.

Direct Normal Irradiance Prediction-Based Optimum Interval Tilt Angles for Enhancement of Energy Output, Levelized Cost of Energy, and CO2 Emission in a Grid-Connected Photovoltaic System

Abstract The tilt angle of photovoltaic (PV) panels is a crucial determinant of their performance and can be adjusted using different tracking methods. Periodically changing the tilt angle strikes a practical balance between efficiency and cost. This work introduces a bi-directional long short-term memory (Bi-LSTM)-based direct normal irradiance (DNI) prediction to estimate the time intervals for the tilt angle adjustments. DNI prediction involves 22-year (2000–2022) historical time series data and the Bi-LSTM deep learning model to predict DNI at different time frames for the location Madurai, India. Using the predicted DNI, tilt angle-based DNI is mapped using the tilt angle correlation through a nearest neighborhood interpolation method. DNI potential over a specific period is utilized to find the optimum time intervals for the tilt angle adjustments. The simulation study of this work is implemented with a 5 kW grid-connected solar PV system using pvsyst software. The effectiveness of the proposed methodology is evaluated based on the improvements in power output, levelized cost of energy (LCOE), and carbon emission reductions and compared with other existing methods. The results showed that using the proposed optimal tilt angle intervals led to a 10.31% increase in PV output power, the lowest LCOE at 3.61 c/kW h, and 8.363 tCO2/year carbon emissions.

Distributed transaction method for multi-energy systems clusters considering uncertainties in photovoltaic power output

Abstract In light of the emerging trend of distributed transactions within the multi-energy system(MES) cluster, this paper proposes a distributed transaction method for MES that takes into account the uncertainties in photovoltaic (PV) power output. This approach aims to achieve optimized resource scheduling within the MESs, as well as coordinated optimization of transactions between different MESs. Firstly, the uncertainties in PV power output within the MES are addressed using the distributionally robust chance constraint. Subsequently, the optimization problem of the MES cluster is decomposed using the alternating direction multiplier method, thereby establishing a distributed transaction model for the MES cluster that considers PV power output uncertainties. Finally, the case study demonstrates that the proposed model can balance the individual interests of each MES with the overall interests of the MES cluster. It effectively reduces the operating costs of the MESs and determines optimal operation strategies for them.

A distributed photovoltaic short-term power forecasting model based on lightweight AI for edge computing

Abstract In recent years, an immense amount of distributed photovoltaics are integrated into low-voltage distribution network, producing a substantial volume of operational data. The centralized cloud data center cannot process massive amounts of data precisely and promptly. Therefore, the operational status of distributed photovoltaic systems in low-voltage distribution network becomes difficult to predict. However, edge computing in the distribution network enables local data processing to improve the forecasting service’s real-time reliability. In this regard, this paper proposes a distributed photovoltaic short-term power forecasting model based on lightweight AI algorithms. Firstly, based on the Pearson correlation coefficient method, an analysis is conducted on the historical operational data in the network to extract important meteorological features correlated with the photovoltaic power output. Secondly, a distributed photovoltaic power forecasting model for the distribution network is constructed based on the Xception and attention mechanism. Finally, the model is trained using pruning, which removes redundant parts of the model, resulting in a compact and efficient forecasting model. By validating real-world datasets, the results demonstrate that the model presented in this article has a smaller size and higher forecasting accuracy than other state-of-the-art forecasting models.