Abstract:Solar energy is one of the most promising renewable energy resources. Forecasting photovoltaic power generation is an important way to increase photovoltaic penetration. However, the difficulty in qualifying the uncertainty of PV power generation, especially during hazy weather, makes forecasting challenging. This paper proposes a novel model to address the issue. We introduce a modified entropy to qualify uncertainty during hazy weather while clustering and attention mechanisms are employed to reduce computational costs and enhance forecasting accuracy, respectively. Hyperparameters were adjusted using an optimization algorithm. Experiments on two datasets related to hazy weather demonstrate that our model significantly improves forecasting accuracy compared to existing models.

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve This paper aims to address the uncertainty in photovoltaic (PV) power generation forecasting under smoggy weather conditions. Specifically: 1. **Importance of PV Power Generation Forecasting**: - PV power generation is a crucial component of renewable energy, and improving its forecasting accuracy is essential for increasing the penetration rate of PV power systems. - However, due to the influence of external factors, PV power generation has uncertainty, which is particularly significant under smoggy weather conditions. 2. **Limitations of Existing Models**: - Existing PV forecasting models are mainly divided into physical models, statistical models, and hybrid models. These models have certain limitations in dealing with uncertainty. - Physical models can provide high-precision forecasts but have high computational costs. - Statistical models rely on strict assumptions and are difficult to adapt to complex nonlinear relationships. - Hybrid models combine the advantages of multiple models but still have high computational costs. 3. **Special Challenges under Smoggy Weather**: - In smoggy weather, solar radiation is scattered, and particulate matter deposits on PV panels, leading to reduced solar radiation received by the PV panels and increased maintenance costs. - These factors increase the uncertainty of PV power generation, affecting the stability of the power system. ### Solution To address the above issues, the paper proposes a new model that includes the following aspects: 1. **Introduction of Improved Entropy (TEWPP)**: - Created a Tsallis entropy based on weighted permutation patterns (TEWPP) to quantify the uncertainty under smoggy weather conditions. - TEWPP quantifies uncertainty by analyzing pattern distribution, reflecting the distribution of PV power generation under smoggy weather. 2. **Hierarchical Clustering Method**: - Used a hierarchical clustering method based on TEWPP to reduce computational costs. - Hierarchical clustering can increase the similarity of elements within the same cluster, thereby reducing computational costs. 3. **Attention Mechanism**: - Introduced an attention mechanism to extract features and make predictions. - The attention mechanism can focus on key information, improving forecasting accuracy. 4. **Optimization Algorithm**: - Used the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to optimize hyperparameters, enhancing the robustness of the model. 5. **Experimental Validation**: - Conducted experiments on two datasets related to smoggy weather, and the results showed that the proposed model significantly improved forecasting accuracy. Through these methods, the paper aims to improve the forecasting accuracy of PV power generation under smoggy weather conditions, thereby enhancing the stability of the power system.

Short-Term Photovoltaic Forecasting Model for Qualifying Uncertainty during Hazy Weather

Probabilistic Forecasting of Photovoltaic Generation: an Efficient Statistical Approach

Hybrid Probabilistic Forecasting of Photovoltaic Power Generation Considering Weather Conditions

Nonparametric Probabilistic Forecasting of Regional Photovoltaic Power Based on Spatial Clustering and Combining Quantile Regression

A novel intra-hour PV power forecasting technique based on total-sky images

A Sky Image-Based Hybrid Deep Learning Model for Nonparametric Probabilistic Forecasting of Solar Irradiance

Prediction Interval Estimation and Deterministic Forecasting Model Using Ground-Based Sky Image

Short-Term Forecasting and Uncertainty Analysis of Photovoltaic Power Based on the FCM-WOA-BILSTM Model

Short-Term Photovoltaic Power Probabilistic Forecasting Based on Temporal Decomposition and Vine Copula

Hybrid Solar Forecasting Method Using Satellite Visible Images and Modified Convolutional Neural Networks

Forecasting and uncertainty analysis of day-ahead photovoltaic power using a novel forecasting method

A New Power Forecasting Method for Photovoltaic Plants under Hazy Conditions

Research on Probability Distribution of Short-Term Photovoltaic Output Forecast Error Based on Numerical Characteristic Clustering

A photovoltaic power estimation model based on the improved local cloud occlusion index algorithm considering sun block luminance

Pattern Recognition and Ultra-Short-Term Probabilistic Forecasting of Power Fluctuating in Aggregated Distributed Photovoltaics Clusters

Probabilistic forecasting of regional solar power incorporating weather pattern diversity

Forecasting of short-term photovoltaic power generation using combined interval type-2 Takagi-Sugeno-Kang fuzzy systems

Exploring Key Weather Factors from Analytical Modeling Toward Improved Solar Power Forecasting

A physical model with meteorological forecasting for hourly rooftop photovoltaic power prediction

Ultra-short-term Solar Power Forecasting Based on a Modified Clear Sky Model

Forecasting a Short-Term Photovoltaic Power Model Based on Improved Snake Optimization, Convolutional Neural Network, and Bidirectional Long Short-Term Memory Network