Forecasting Of Power Generation Research Articles

Solar power generation forecasting by a new hybrid cascaded extreme learning method with maximum relevance interaction gain feature selection

Solar power generation forecasting by a new hybrid cascaded extreme learning method with maximum relevance interaction gain feature selection

Mind the (spectral) gap: how the temporal resolution of wind data affects multi-decadal wind power forecasts

To forecast wind power generation in the scale of years to decades, outputs from climate models are often used. However, one major limitation of the data projected by these models is their coarse temporal resolution—usually not finer than three hours and sometimes as coarse as one month. Due to the non-linear relationship between wind speed and wind power, and the long forecast horizon considered, small changes in wind speed can result in big changes in projected wind power generation. Our study indicates that the distribution of observed 10 min wind speed data is relatively well preserved using three- or six-hourly instantaneous values. In contrast, daily or monthly values, as well as any averages, including three-hourly averages, are almost never capable of preserving the distribution of the underlying higher resolution data. Assuming that climate models behave in a similar manner to observations, our results indicate that output at three-hourly or six-hourly temporal resolution is high enough for multi-decadal wind power generation forecasting. In contrast, wind speed projections of lower temporal resolution, or averages over any time range, should be handled with care.

Effective mechanism for trading generation rights in the context of carbon emission rights

In the global context of intensive energy resource consumption and substantial greenhouse gas emissions, as the energy sector with the highest share of carbon emissions, the power generation industry is destined to play a central role in China's efforts to promote energy conservation and emission reduction. With the emergence of local carbon pilot projects and the implementation of a national carbon market, key emission units in the power industry, in addition to the electricity market business, also need to fulfill and pay carbon emission rights on schedule. The power generation rights are a market-based mechanism for clean energy units to replace conventional energy units in electricity generation, which can promote carbon emission reduction by incorporating the actual carbon emissions from conventional energy units. This paper establishes an optimized social welfare model that incorporates the interdependency between the thermal power generation unit market and the electric carbon market. It is based on the social welfare model of power generation rights and the initial allocation of carbon emission rights for the units, while considering the cost increase of excess emission under monthly forecasted power generation by the thermal power units. Taking a certain hydropower province as the simulation object, it was verified that the optimized model has a higher transaction rate of power generation rights, greater social welfare, and no reduction in unit revenue.

Hybrid deep learning and quantum-inspired neural network for day-ahead spatiotemporal wind speed forecasting

Wind is an essential, clean and sustainable renewable source of energy; however, wind speed is stochastic and intermittent. Accurate wind power generation forecasts are required to ensure that power generation can be scheduled economically and securely. This work proposes a hybrid deep learning technique that incorporates a quantum-inspired neural network to predict wind speeds 24 h in advance. An innovative neural network technique is implemented by cascading parallel convolutional neural networks (CNNs) with a long short-term memory (LSTM) and a quantum-inspired neural network (QINN). The proposed hybrid model is optimized using two iterative loops: the outer loop is implemented by using quantum particle swarm optimization (QPSO) to tune the structure of model and other hyperparameters/parameters. The inner loop uses an Adam optimizer to tune the weights and biases of the proposed model. Spatiotemporal wind speeds at various locations provide the 2D input data. Simulation results reveal that the proposed method outperforms other methods for 24 h-ahead wind speed forecasting.

Optimal cleaning schedule in solar PV using the biography-based helianthus optimization enabled coupled deep network

The research aims to develop a system that determines the optimal intervals for cleaning solar PV panels to maximize their efficiency and minimize power loss due to dust accumulation. The research utilizes data related to daily solar photovoltaic tariff rates, power generation, and conditions of dust accumulation on PV panels. To monitor the performance of PV panels, the research involves forecasting power generation. If the power generation falls below a predefined threshold, it indicates a need for cleaning. Dust particle loss conditions are analyzed using an optimized Support Vector Machine (SVM) coupled with a Deep Convolutional Neural Network (CNN) model. This model processes PV panel images to identify dusty conditions. The Deep CNN model extracts features from PV panel images, such as color-based statistical features and local gradient patterns. These features are crucial for identifying dusty conditions. The optimal cleaning schedule is determined based on the detection of dust particle loss conditions. Cleaning is initiated when the conditions warrant it. The novelty of this research lies in the development of a deep learning model for both detecting dust particle loss and forecasting power generation. This model is trained using data that includes daily tariff rates and dust particle loss information. The research proposes a new optimization method called “biography-based Helianthus optimization (HBBHO)” specifically designed to suit the characteristics of sunlight. This method is employed to schedule cleaning activities. The proposed approach achieves impressive accuracy, precision, sensitivity, and specificity, with values reaching 96.53% for most metrics, except for specificity, which is at 96.21%.

Developing Information Gateway for Intelligent Decision-Making and Stability in Solar Power Generation

Abstract This study has developed a specialized information gateway for inverters, utilizing relays, data converters, and single-board computers, among other components. Upon receiving data from the gateway, the server processes it to generate an intelligent solar monitoring system. The platform utilizes deep learning RNN and LSTM algorithms to forecast power generation at solar power plants, allowing for real-time monitoring of weather and power generation. By comparing actual and expected power generation data, the system can adjust equipment maintenance and cleaning schedules. It is designed to automatically alert staff members to take appropriate action when anomalous power generation data is continuously transmitted. Additionally, the system sends an alert for on-site inspection and removal of any abnormal situations to increase the stability of solar power generation. This study employs deep learning and IoT data collection to provide the knowledge necessary for intelligent decision-making and increased stability in solar power generation. JEL classification numbers: C43, F68, H41. Keywords: Solar energy, IoT monitoring, RNN, LSTN, Abnormal return.

Short-Term Wind Speed and Power Forecasting for Smart City Power Grid With a Hybrid Machine Learning Framework

To address foreseeable challenges during the penetration of wind energy into the power grid, including accurate wind power forecasting and smart power generation scheduling, this study proposes a novel short-term wind speed forecasting model, named EMD-KM-SXL, which is based on empirical mode decomposition (EMD), K-means clustering and machine learning, and a new two-stage short-term wind power forecasting model based on wind speed forecasting and wind power curve modeling. The former wind speed forecasting model regards historical wind speed observations as model inputs, and the latter power forecasting model utilizes knowledge augmentation, introducing wind power conversion relationship, environmental factors as well as wind power system status parameters. In the proposed wind speed forecasting model, three machine learning models, including support vector regressor, XGBoost regressor, and Lasso regressor, are employed to forecast three types of frequency components that are generated via EMD and K-means clustering. Then, the wind power curve model is utilized to compute potential outpower based on the predicted wind speed, which is regarded as the first stage of the proposed wind power forecasting model. In the second stage, environmental factors and wind power system status parameters are introduced and an artificial neural network model, considering preliminary predicted power, environmental factors, and wind power system status parameters as model inputs, is built to make final power prediction. Computational results show that proposed models achieve the best performance in terms of wind speed and power forecasting over different forecasting horizons ranging from 10 to 40 minutes, compared with benchmarking methods.

Energy supply-demand interaction model integrating uncertainty forecasting and peer-to-peer energy trading

With the penetration of large amounts of renewable energy resources into energy system, the interaction between energy supply and demand has become more complex and diverse. The complexity and diversity make it more difficult to achieve real-time, efficient, accurate and dynamic matching of energy supply and demand. Therefore, the study proposes an efficient energy supply-demand interaction model integrating uncertainty forecasting and peer-to-peer energy trading. First, to reduce the impact of supply and demand uncertainty on the energy supply and demand matching, gate recurrent unit and long short-term memory models are used to forecast power generation and consumption. Then, based on the results of forecasting, an energy supply-demand interaction model is proposed to assist the energy system in achieving dynamic energy supply-demand matching. Finally, the effectiveness of the proposed energy supply-demand interaction model has been verified through experiments. The proposed energy supply-demand interaction model that considers supply and demand uncertainty and economic benefits helps to better achieve transparent, efficient, stable, and sustainable matching of supply and demand. This study can reduce the impact of supply and demand uncertainty by forecasting power generation and consumption. In addition, this study considers the preferences of prosumers in their trading, reduces the cost of electricity for prosumers, and realizes the profitability of multiple subjects involved in the trading.

Solar parking lot management: An IoT platform for smart charging EV fleets, using real-time data and production forecasts

The fast transition to the electrification of the energy system, combined with an exponential growth of the market share of electric vehicles, is leading to a tight interrelation between electric energy production and transportation, two prominent sectors in fossil fuels consumption and greenhouse gas emissions. Accelerating this process, the management of electric fluxes, aiming at optimizing production and demand coupling, plays a crucial role in reaching the net-zero emission target. The proposed software platform is designed to optimally manage the energy fluxes for a solar powered parking lot, serving a fleet of electric vehicles; the real-time knowledge of energy production and demand, in conjunction with forecasted power generation, allows the maximization of renewable energy self-consumption, thus reducing the exchange with the external grid. The software platform can work either in design mode, allowing the dimensioning of the various parking lot components, or in real-time mode managing instantaneously the energy balance. As a case study, it is tested on the 2019 parking lot mobility data of a research center, assuming a complete transformation of the then existing fleet of employees' cars to electric vehicles. A comparison of the resulting energy flows with those projected by an established commercial tool is performed, as well as a preliminary economic evaluation. Both consistency of the simulation results and favorable economics validate the presented smart charging algorithm and Internet of Things platform for the real-time energy management of a solar parking lot.

An intuitive unified hybrid approach for medium to long-term forecasting of different renewable energy sources

Forecasting of stochastic renewable energy resources is quintessential for effective planning, operations and management of the power systems. Existing literature contains ample studies on very short and short-term forecasting of renewables. But it’s still challenging to obtain high accuracy for medium-term and long-term forecasting. Therefore, an intuitive unified hybrid approach is proposed in this paper for medium to long-term forecasting of power generation from different weather-dependent renewables single-handedly by utilising their inherited periodic seasonal patterns iteratively year-on-year basis. Bi-LSTM (Bidirectional Long Short Term Memory) and ARIMA (Auto Regressive Integrated Moving Average) are utilised to construct the proposed approach with the aid of STL (Seasonal-Trend decomposition using Loess) decomposition and data pre-processing. The performance of proposed approach (STL-ARIMA-BiLSTM) is validated using seven recent datasets of wind, solar and hydro power. It yields accurate forecasts for a week-ahead to a year-ahead forecasting horizons. MAE (Mean Absolute Error) lying in range from 6.3% to 6.6%, 5.6% to 6.7% and 4.72% for a year-ahead forecast of wind, hydro and solar power, respectively, is obtained. It is established that these long-term forecast errors are even less than the short-term forecast errors of some of the existing studies demonstrating novelty and practicality of the proposed approach.

A dual spatio-temporal network for short-term wind power forecasting

The intricate spatial and temporal complexities inherent in wind speed data necessitate the utilization of spatio-temporal networks for accurate wind power generation forecasting. But these complexities are made worse by the constant presence of high-frequency disturbances and clear outliers in the spatial and temporal correlations of wind speed, which makes forecasting wind power very difficult. Additionally, the intricate topological architectures that underpin current models often result in superfluous computational operations and susceptibility to overfitting. In response to these formidable challenges, we introduce an innovative solution in the form of the dual spatio-temporal (DST) network. This paradigm, characterized by meticulous design, excels in acquiring precise and resilient spatio-temporal representations essential for wind power forecasting. This is possible because of the careful development of two types of spatial correlation layers: synchronous layers and asynchronous layers. Each type is fine-tuned to recognize different spatial patterns. Recognizing that spatial and temporal correlations change over time, our method explores spatio-temporal representations at many different scales. This makes the system more adaptable and robust. Notably, for effective noise mitigation over extended forecasting horizons, we judiciously integrate wavelet decomposition. Empirical validation, rigorously conducted across twelve wind power stations in Anhui Province, China, firmly attests to the efficacy of our proposed DST network. In summary, our innovation presents a promising avenue for advancing wind power forecasting, characterized by adept noise reduction and improved precision while concurrently optimizing computational efficiency in modeling intricate spatial–temporal dependencies.

Forecast of clean energy generation in China based on new information priority nonlinear grey Bernoulli model.

The forecast of clean energy power generation is of major prominence to energy structure adjustment and the realization of sustainable economic development in China. In order to scientifically predict clean energy power generation data, a structure-adaptive nonlinear grey Bernoulli model submitted to the new information priority criterion (abbreviated as IANGBM) is established. Firstly, an improved conformable fractional accumulation operator that conforms to the priority of new information is proposed, which can effectively extract the information from small samples. Then, IANGBM is derived from the Bernoulli differential equation, and the perturbation bound theory proves that this model is suitable for the analysis of small sample data. In addition, the grey wolf optimization algorithm is utilized to optimize the model parameters to make the model more adaptable and generalized. To verify the superiority of the model, two cases consisting of wind and nuclear power generation prediction are implemented by comparing eight benchmark models involving IANGBM, GM, FGM, FANGBM, LR, SVM, BPNN, and LSTM. The experiment results demonstrate that the proposed model achieves higher prediction accuracy compared to the other seven competing models. Finally, the future nuclear and wind power generation from 2023 to 2030 are predicted by adopting the IANGBM(1,1) model. For the next 8 years, nuclear power generation will maintain stable development, while wind energy power generation will grow rapidly.

Day-Ahead Parametric Probabilistic Forecasting of Wind and Solar Power Generation Using Bounded Probability Distributions and Hybrid Neural Networks

Day-Ahead Parametric Probabilistic Forecasting of Wind and Solar Power Generation Using Bounded Probability Distributions and Hybrid Neural Networks

Short-term forecasting of photovoltaic power generation

Short-term forecasting of photovoltaic power generation

Exploring the Gradient Boosting and LSTM for Power Distribution-based Time Series Analysis

Electrical energy production is an important function in a country that has abundant demand for society. Therefore, power supply forecasting is very significant to keep the balance of power consumption and production and should not be in short supply at any cost. Thus, Forecasting and modelling time series analysis of power production becomes important in prediction. We explore the synergy between Gradient Boosting and Long Short-Term Memory (LSTM) networks in the context of time series analysis. Specifically, we investigate how XGBoost and LightGBM, well-known gradient boosting frameworks, complement LSTM, a recurrent neural network architecture. By combining these methods, we aim to uncover deeper insights and elevate predictive potential for time series forecasting. In this paper, we examine the machine learning methods of forecasting the power generation of 33-year time series data. In the conclusion, we put forward a comparative study between the outcomes attained from the use of LSTM and Gradient-boosting tree-based algorithms XGBoost and LightGBM. The results show that the XGBoost model outclasses the other models with a low error value in forecasting power generation.

Harnessing climate variables for predicting PV power output: A backpropagation neural network analysis in a subtropical climate region

This study examines the influence of weather conditions on the electricity production of photovoltaic (PV) panels in the Conghua district, a region in Southern China distinguished by its subtropical mountainous climate. The area's high humidity levels, high temperatures, and substantial rainfall often lead to predominantly cloudy skies, which alter solar irradiation patterns as a result. Utilizing the EnergyPlus simulation tool, the power generation of PV panels on rooftop application was simulated. A structured classification method is adopted to sort out weather conditions into categories of sunny, cloudy, and rainy. Through examining the influence of these varied weather scenarios on power generation, unique factors affecting electricity outputs are identified. As a result, this study illuminates the relationships between potential weather variables and PV power generation across each weather category. Subsequently, a back propagation neural network (BPNN) model is utilized to explore the relationship between weather categories and PV generation. The high coefficient of determination value (R2 > 0.95) from the comparison of simulated and predicted PV generation validates the model accuracy. This research enables the development of updated dataset and enriches the prediction of rooftop PV power generation. It equips designers and owners with a reliable, user-friendly, and real-time tool for PV power generation forecasting, and provides a working platform for PV performance evaluation during the design phase of PV projects. The insights derived from the BPNN model could guide the optimal design and implementation of PV systems in similar climates. Furthermore, these findings advance sustainable energy solutions in regions affected by complex weather patterns.

Feasibility and Potential Assessment of Wind Resource a Case Study in North Shewa Zone, Amhara, Ethiopia

The assessment of wind resource potential and feasibility is critical for generating and forecasting power generation, as wellas resource identification. In Ethiopia, the majority of the country lacks a wind atlas, making it difficult to determine the availability of sources. Seven different areas (Debre Berhan, Alem Ketema, Mehal Meda, Eneware, Gundo Meskel, Majete and Shewa Robit) were investigated. For the work data was collected from various sources and analyzed using MATLAB software. The basic sources of data that were obtained nationally were from the NMA, which is the National Metrological Agency found in Addis Ababa, and were obtained centrally from each local delegate’s registered report from a height of 2 meters and 10 meters in each listed districts above. According to the results analysis, the average wind speed at most sites is reasonable and 4 m/s at a height of 10 meters, and some of the case study sites have an average wind speed of less than 3 m/s. The extrapolation prediction method produces more realistic results at 30 and 50 meters; for example, when 10 meter is extrapolated to 30 and 50 meters, the wind power densities are 75.2 w/m2, 300.9 w/m2, and 680.5 w/m2, respectively. Similarly, the average yearly energy density for 10 meter, 30 meter, and 50 meter is 2110.8, 4122.6, and 8219.9 Kwh/m2/year, respectively. As per the international standard for wind power and wind speed classification, Eneware and Mehal Meda are categorized under class 7, whereas Debre Berhan is categorized under class 3, while the remaining sites such as Shewarobit, Gunde Meskel, Alem Ketema, and Majete are classified under class 1 for the majority of the year.

Effective LSTMs with seasonal-trend decomposition and adaptive learning and niching-based backtracking search algorithm for time series forecasting

Long short-term memory faces challenges in information mining and parameter selection due to inherent uncertainty and randomness. In this study, we propose a novel hybrid model called ‘STL-ALN_BSA-LSTM’ that aims to tackle this issue. In the proposed model, data is first prepossessed using seasonal-trend decomposition to extract informative features. Next, we propose an adaptive learning and niching-based backtracking search algorithm to enhance the robustness and learning ability of the original backtracking search algorithm, which affords a balance between exploration and exploitation. Subsequently, the weights and thresholds within each LSTM are optimized by our optimization algorithm automatically, leading to improved interpretability and accuracy of the model. Finally, the proposed optimization algorithm was evaluated using five benchmark functions, and the hybrid model was compared with sixteen baselines on two real-world datasets. For rainfall forecasting, the hybrid model achieved the root mean square error of 7.3772, mean absolute error of 2.6187, and R2 of 0.8388. For power generation forecasting, our model outperformed the second-best model by 29.6%, 34.4%, and 16 ‰, respectively, as in the above metrics. These results demonstrate that the proposed hybrid model can attain higher accuracy in real-life applications than previous models, which contributes in-depth insights into parameter optimization for time series forecasting, and provides valuable suggestions in resource planning, risk reduction, and decision-making. The core code of this work can be found from https://github.com/zjuml/STL-ALNBSA-LSTM.

Deep Learning for Variable Renewable Energy: A Systematic Review

In recent years, both fields, AI and VRE, have received increasing attention in scientific research. Thus, this article’s purpose is to investigate the potential of DL-based applications on VRE and as such provide an introduction to and structured overview of the field. First, we conduct a systematic literature review of the application of Artificial Intelligence (AI), especially Deep Learning (DL), on the integration of Variable Renewable Energy (VRE). Subsequently, we provide a comprehensive overview of specific DL-based solution approaches and evaluate their applicability, including a survey of the most applied and best suited DL architectures. We identify ten DL-based approaches to support the integration of VRE in modern power systems. We find (I) solar PV and wind power generation forecasting, (II) system scheduling and grid management, and (III) intelligent condition monitoring as three high potential application areas.

Accurate four-hour-ahead probabilistic forecast of photovoltaic power generation based on multiple meteorological variables-aided intelligent optimization of numeric weather prediction data

Accurate four-hour-ahead probabilistic forecast of photovoltaic power generation based on multiple meteorological variables-aided intelligent optimization of numeric weather prediction data