Random Forest Ensemble of Support Vector Regression for Solar Power ForecastingMohamed Abuella
This document describes a methodology for generating combined solar power forecasts using an ensemble of support vector regression models. The methodology includes:
1. Developing 24 individual SVR forecasting models from two different solar power datasets.
2. Using random forest regression to combine the forecasts from the 24 SVR models.
3. Evaluating the combined forecasts against individual model forecasts and a simple average combination, finding the ensemble approach improved accuracy in most months.
Hourly probabilistic solar power forecasts 3vMohamed Abuella
This document summarizes a presentation on hourly probabilistic solar power forecasting. It discusses:
1) The need for solar power forecasting to address the variability of solar generation. Various forecasting models and horizons are used.
2) The combined physical and statistical approach used to generate solar power forecasts, which includes solar plant modeling, numerical weather prediction, and statistical corrections.
3) The models evaluated include multiple linear regression, artificial neural networks, and support vector regression. Ensemble learning with random forests is used to combine the models' outputs.
4) Probabilistic forecasting methods evaluated include ensemble-based, analog ensemble, and persistence approaches. Different probability distributions are generated.
5)
Performed analysis on Temperature, Wind Speed, Humidity and Pressure data-sets and implemented decision tree & clustering to predict possibility of rain
Created graphs and plots using algorithms such as k-nearest neighbors, naïve bayes, decision tree and k means clustering
Khaled eskaf presentation predicting power consumption using genetic algorithmKhaled Eskaf
The document discusses predicting short-term electrical energy consumption using a dynamic model and genetic algorithm. It proposes extracting features from historical energy consumption time series data using a dynamic system model to determine impulse forces and damping factors. A genetic algorithm is then used to predict future consumption values based on these features, with the ability to continuously learn from new data online. The approach is evaluated using root-mean-square error and shown to achieve accurate predictions between 5x10-5 to 1x10-4 kilowatt hours.
This document discusses using deep learning models to predict weather events from numerical model data and satellite images. It describes three models: DeepRain predicts precipitation from weather radar images using convolutional LSTMs; DeepTC predicts tropical cyclone trajectories from numerical model data like temperature and pressure using convolutional LSTMs; and GlobeNet predicts typhoon tracks from satellite images and an autoencoder. The models show improved prediction accuracy over traditional methods, demonstrating the potential of deep learning for weather forecasting.
This document describes Siddharth Chaudhary's MSc research project on forecasting solar electricity generation using time series models. The research aims to 1) forecast solar generation in Delhi and Jodhpur, India, 2) evaluate the performance of forecasting models, and 3) compare potential solar generation between the two cities. Four time series models - TBATS, ARIMA, simple exponential smoothing, and Holt's method - are applied to solar radiation data from each city and their accuracy is assessed.
Short-term load forecasting with using multiple linear regression IJECEIAES
This document discusses short-term load forecasting using multiple linear regression. It summarizes the research method used, which involves developing a multiple linear regression model to predict electrical load based on variables like temperature, humidity, day of week, and previous load data. The model is trained on historical load and weather data from New York City over 9 years. Testing shows the model can predict load a day ahead with 5.15% mean absolute percentage error. Regression coefficients, t-statistics, and p-values indicate the trained model explains about 90% of the variation in load and the predictors are statistically significant. An example day-ahead hourly load forecast is provided for June 25, 2019.
Random Forest Ensemble of Support Vector Regression for Solar Power ForecastingMohamed Abuella
This document describes a methodology for generating combined solar power forecasts using an ensemble of support vector regression models. The methodology includes:
1. Developing 24 individual SVR forecasting models from two different solar power datasets.
2. Using random forest regression to combine the forecasts from the 24 SVR models.
3. Evaluating the combined forecasts against individual model forecasts and a simple average combination, finding the ensemble approach improved accuracy in most months.
Hourly probabilistic solar power forecasts 3vMohamed Abuella
This document summarizes a presentation on hourly probabilistic solar power forecasting. It discusses:
1) The need for solar power forecasting to address the variability of solar generation. Various forecasting models and horizons are used.
2) The combined physical and statistical approach used to generate solar power forecasts, which includes solar plant modeling, numerical weather prediction, and statistical corrections.
3) The models evaluated include multiple linear regression, artificial neural networks, and support vector regression. Ensemble learning with random forests is used to combine the models' outputs.
4) Probabilistic forecasting methods evaluated include ensemble-based, analog ensemble, and persistence approaches. Different probability distributions are generated.
5)
Performed analysis on Temperature, Wind Speed, Humidity and Pressure data-sets and implemented decision tree & clustering to predict possibility of rain
Created graphs and plots using algorithms such as k-nearest neighbors, naïve bayes, decision tree and k means clustering
Khaled eskaf presentation predicting power consumption using genetic algorithmKhaled Eskaf
The document discusses predicting short-term electrical energy consumption using a dynamic model and genetic algorithm. It proposes extracting features from historical energy consumption time series data using a dynamic system model to determine impulse forces and damping factors. A genetic algorithm is then used to predict future consumption values based on these features, with the ability to continuously learn from new data online. The approach is evaluated using root-mean-square error and shown to achieve accurate predictions between 5x10-5 to 1x10-4 kilowatt hours.
This document discusses using deep learning models to predict weather events from numerical model data and satellite images. It describes three models: DeepRain predicts precipitation from weather radar images using convolutional LSTMs; DeepTC predicts tropical cyclone trajectories from numerical model data like temperature and pressure using convolutional LSTMs; and GlobeNet predicts typhoon tracks from satellite images and an autoencoder. The models show improved prediction accuracy over traditional methods, demonstrating the potential of deep learning for weather forecasting.
This document describes Siddharth Chaudhary's MSc research project on forecasting solar electricity generation using time series models. The research aims to 1) forecast solar generation in Delhi and Jodhpur, India, 2) evaluate the performance of forecasting models, and 3) compare potential solar generation between the two cities. Four time series models - TBATS, ARIMA, simple exponential smoothing, and Holt's method - are applied to solar radiation data from each city and their accuracy is assessed.
Short-term load forecasting with using multiple linear regression IJECEIAES
This document discusses short-term load forecasting using multiple linear regression. It summarizes the research method used, which involves developing a multiple linear regression model to predict electrical load based on variables like temperature, humidity, day of week, and previous load data. The model is trained on historical load and weather data from New York City over 9 years. Testing shows the model can predict load a day ahead with 5.15% mean absolute percentage error. Regression coefficients, t-statistics, and p-values indicate the trained model explains about 90% of the variation in load and the predictors are statistically significant. An example day-ahead hourly load forecast is provided for June 25, 2019.
This document discusses applications of data mining techniques to predict mesoscale weather events like tornadoes and cloudbursts. It summarizes previous research that applied data mining methods like neural networks, support vector machines, and clustering to weather prediction. For tornado prediction, studies developed spatiotemporal models to identify relationships between storm variables. Other research used mesocyclone detection algorithms and neural networks to predict tornadoes. For cloudburst prediction, clustering relative humidity and divergence from numerical models provided early formation indications. The document also briefly explains ensemble forecasting, which runs multiple forecasts from slightly different initial conditions to sample forecast uncertainty.
The document discusses wind speed prediction using the Weibull distribution and a hybrid Weibull-ANN technique. It presents the motivation for improved wind speed prediction due to the increasing use of wind energy. The Weibull distribution is described as a common statistical model used to analyze wind speed data. An artificial neural network model with backpropagation is also introduced for prediction. The document then analyzes wind speed data from Bhubaneswar using Weibull distributions and histograms to model the data distributions. Finally, it evaluates the hybrid Weibull-ANN technique for wind speed prediction performance.
Estimation of precipitation during the period of south west monsoonIAEME Publication
This document discusses a study that uses numerical methods in C programming to quantitatively estimate precipitation during the South West Monsoon period in India. The study collects temperature and precipitation data over 10 years and develops 2nd order equations to correlate temperature and computed precipitation for different monthly regions during the monsoon season. The equations are of the form Precipitation = C0 + C1*Temperature + C2*Temperature^2, where C0, C1, C2 are constants calculated using a C program. Segmenting the monsoon season into monthly regions provides better correlation between computed and actual precipitation compared to using the entire monsoon period. The study aims to quantitatively predict precipitation based on temperature to help farmers plan crop sowing.
IRJET - Intelligent Weather Forecasting using Machine Learning TechniquesIRJET Journal
This document discusses using machine learning techniques to forecast weather intelligently. It proposes using multi-target regression and recurrent neural network (RNN) models trained on historical weather data from Bangalore to predict future weather conditions like temperature, humidity, and precipitation. The data is first preprocessed before being fed to the models. The models are evaluated to accurately predict weather in the short term to help people like farmers and commuters without relying on expensive equipment.
Comparison of Solar Radiation Intensity Forecasting Using ANFIS and Multiple ...journalBEEI
This document compares the performance of two methods for forecasting solar radiation intensity: Adaptive Neuro Fuzzy Inference System (ANFIS) and Multiple Linear Regression (MLR). It uses weather data from Basel, Switzerland to test the methods. The ANFIS method uses a fuzzy inference system combined with neural networks, while MLR uses a mathematical approach. The performance of both methods is evaluated using root mean square error (RMSE) and mean absolute error (MAE) across different training/testing data compositions and time periods. The results show that ANFIS consistently provides lower error values than MLR, indicating it provides more accurate solar radiation forecasts.
Determination of wind energy potential of campus area of siirt universitymehmet şahin
In this study, wind energy potential of Siirt
University campus area is statistically examined by using the mean hourly wind speed data between 2014
and 2015 years which are measured by Vantage Pro2 device, located at the roof of the Engineering Faculty building with 6 m altitude. Weibull distribution
function and Rayleigh distribution function are used
as statistical approach to evaluate the wind data. Weibull distribution function is examined by using two different methods that are maximum likelihood estimation and Rayleigh method. The determination
coefficient (R2) and Root Mean Square Error (RMSE) values of these methods are compared. According the error analysis, it is indicated that the Rayleigh method
gives better results. Wind speed and wind power density are calculated in pursuance of Weibull distribution parameters. The results are evaluated as
monthly and annually. Hence, this preliminary study is made to determine the wind energy potential of Siirt University campus area.
an analysis of wind energy potential using weibull distributionWorking as a Lecturer
This document analyzes wind energy potential using the Weibull distribution. It discusses two studies that used the Weibull distribution to model wind speed and calculate parameters to estimate wind power potential. One study used a simulation model to describe wind turbine characteristics and power generated at a Sahara site in Algeria. The other calculated Weibull shape and scale factors using four methods and compared theoretical and observed probability density functions to determine the best fit. Both found the Weibull distribution directly influences estimates of wind power potential at a given location.
CFD down-scaling and online measurements for short-term wind power forecastingJean-Claude Meteodyn
Usually speaking, Forecast systems are classified : Intraday (Very Short term) is commonly Stochastic with online measurements while Extraday (Short term) is usually Deterministic based on NWP data. This work aims to breakdown these classifications, proposing a unique tool based on the unification of all these techniques.
VISUAL ANALYSIS OF ELECTRICITY DEMAND: ENERGY DASHBOARD GRAPHICS Graphical Da...Fatma ÇINAR
A real time interactive data management for Impulse and Response Analysis Technique using lattice and ggplot2 Graphical Packages embedded in R software has been employed. Average consumption, peak consumption and daily consumption data have been used while the temperature data is also employed to highlight the significance of relationship between consumption and the weather conditions. The demand for electricity by the factors affecting the demand with a multi-dimensional matrix graphics based on Energy Dashboard Software has been analysed leading to visualisation.
A WIND POWER PREDICTION METHOD BASED ON BAYESIAN FUSIONcsandit
The document proposes a wind power prediction method based on Bayesian fusion of multiple numerical weather predictions. It first establishes a relationship between wind speed and power using neural networks. It then analyzes the characteristics of wind speed forecasts from three independent weather sources. A Bayesian method is designed to fuse the wind speed forecasts, yielding a more accurate prediction than any single source. The fused wind speed is input to the neural network model to predict wind power at 15-minute intervals. Experimental results show the method improves accuracy of wind speed and power forecasting compared to using a single source.
This document discusses techniques for estimating solar energy. It begins by explaining the importance of accurate solar energy estimation for energy planners and utilities. It then describes three types of estimation horizons: very short term, short term, and long term prediction. Next, it outlines linear and nonlinear estimation techniques, including the Angstrom and Angstrom-Prescott models for linear regression. It focuses on the use of neural networks for short term estimation, noting their ability to model complex nonlinear relationships without extensive data or prior model specification. Finally, it provides examples of neural networks being used to estimate solar irradiance up to 24 hours ahead and assess solar potential in the Himalayan region.
This document provides a case study on forecasting monthly exceedance probabilities of solar radiation in Arizona. It discusses collecting solar radiation data from five stations in different locations in Arizona. The authors define exceedance probability as the probability of daily radiation being below an expected value. They use normalized distributions and simple linear regression to predict monthly exceedance probabilities and compare them to actual probabilities calculated from later test data. The document discusses setting up the model, including normalizing the data distributions and using data before 2011 to predict and data from 2011-2014 to test the predictions.
Photovoltaic Modules Performance Loss Evaluation for Nsukka, South East Niger...IJERA Editor
The Photovoltaic (PV) systems and technology offer excellent reliability when designed with the right implementation tools and based on good technical judgements of components that make up each of the critical sections of solar power system. The PV array is an essential section of a solar power system and it is expected to function to deliver pre – estimated power based on design estimations. There are factors that derail the performance of PV modules; the contributions of these factors are peculiar to specific sites of installation, hence the need to empirically evaluate and characterize installation sites before deployment of PV systems. This paper presents the characterization of Nsukka (South East, Nigeria) environment using decent instrumentation; and consequently highlights the power loss indicators for PV modules in the target site while presenting equally mitigable design.
This document discusses using a seasonal autoregressive integrated moving average (SARIMA) model to forecast precipitation in Mt. Kenya region. It fits various SARIMA models to monthly precipitation data from 1970 to 2011 and selects the best model with the lowest AIC and BIC values. The best model was found to be SARIMA(1,0,1)x(1,0,0)12, which had two statistically significant variables and passed diagnostic checks. Forecast accuracy statistics for this model, including ME, MSE, RMSE and MAE, indicated the SARIMA model provides a good method for precipitation forecasting in Mt. Kenya region.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
- In the past, experiments yielded limited data but today data is abundant and collected at low cost from each experiment
- Big data usually includes large data sets beyond the abilities of typical software to analyze in a reasonable time frame
- Data-driven models are based on analyzing relationships between input and output variables in a system using machine learning algorithms trained on representative data, without requiring explicit knowledge of the system's physical behavior
- Complexity of climate systems
- Climate modelling
- The need for modelling
- System thinking
- Analytical vs Numerical modeling
- Mathematical models
- Modeling process and model selection
- Model Uncertainty
- Modeling application and tools
Improved Kalman Filtered Neuro-Fuzzy Wind Speed Predictor For Real Data Set ...IJMER
This document presents three new models for short-term (24 hours ahead) wind speed forecasting for Egypt's northwestern coast based on real data collected from the site. The first model predicts wind speed using the same month of data from seven consecutive years. The second model predicts using only one month of data with a time series prediction scheme. The third model applies a discrete Kalman filter to one month of data first to reduce noise before prediction using an adaptive neuro-fuzzy inference system (ANFIS). The Kalman filtered data provided more accurate predictions with a 64% reduction in error compared to the first model.
The accurate prediction of solar irradiation has been
a leading problem for better energy scheduling approach.
Hence in this paper, an Artificial neural network based solar
irradiance is proposed for five days duration the data is
obtained from National Renewable Energy Laboratory, USA
and the simulation were performed using MATLAB 2013. It
was found that the neural model was able to predict the solar
irradiance with a mean square error of 0.0355.
Currently, the quality of wind measure of a site is assessed using Wind Power Density (WPD). This paper proposes to use a more credible metric namely, one we call the Wind Power Potential (WPP). While the former only uses wind speed information, the latter exploits both wind speed and wind direction distributions, and yields more credible estimates. The new measure of quality of a wind resource, the Wind Power Potential Evaluation (WPPE) model, investigates the effect of wind velocity distribution on the optimal net power generation of a farm. Bivariate normal distribution is used to characterize the stochastic variation of wind conditions (speed and direction). The net power generation for a particular farm size and installed capacity are maximized for different distributions of wind speed and wind direction, using the Unrestricted Wind Farm Layout Optimization (UWFLO) methodology. A response surface is constructed, using the recently developed Reliability Based Hybrid Functions (RBHF), to represent the computed maximum power generation as a function of the parameters of the wind velocity (speed and direction) distribution. To this end, for any farm site, we can (i) estimate the parameters of wind velocity distribution using recorded wind data, and (ii) predict the max- imum power generation for a specified farm size and capacity, using the developed response surface. The WPPE model is validated through recorded wind data at four differing stations obtained from the North Dakota Agricultural Weather Network (NDAWN). The results illustrate the variation of wind conditions and, subsequently, its influence on the quality of a wind resource.
A Post-processing Approach for Solar Power Combined Forecasts of Ramp EventsMohamed Abuella
This dissertation applies a post-processing approach to improve combined forecasts of solar power and solar power ramp events. The approach combines different forecasting models, then adjusts the combined forecasts to better predict ramp events. It develops a classification system for ramp event thresholds and evaluates performance using customized metrics. The approach provides probabilistic forecasts of solar power ramp events with uncertainty analysis.
Short Presentation: Mohamed abuella's Research HighlightsMohamed Abuella
- A post-processing approach combines and improves solar power forecasts and adjusts the combined forecasts in terms of ramp events.
- The approach classifies all possible thresholds and classes of ramp event forecasts and uses a customized cost function for imbalanced classification of ramp events.
- Suitable metrics are used for the feature selection process and performance evaluation, along with an uncertainty analysis for probabilistic forecasts of solar power ramp events.
This document discusses applications of data mining techniques to predict mesoscale weather events like tornadoes and cloudbursts. It summarizes previous research that applied data mining methods like neural networks, support vector machines, and clustering to weather prediction. For tornado prediction, studies developed spatiotemporal models to identify relationships between storm variables. Other research used mesocyclone detection algorithms and neural networks to predict tornadoes. For cloudburst prediction, clustering relative humidity and divergence from numerical models provided early formation indications. The document also briefly explains ensemble forecasting, which runs multiple forecasts from slightly different initial conditions to sample forecast uncertainty.
The document discusses wind speed prediction using the Weibull distribution and a hybrid Weibull-ANN technique. It presents the motivation for improved wind speed prediction due to the increasing use of wind energy. The Weibull distribution is described as a common statistical model used to analyze wind speed data. An artificial neural network model with backpropagation is also introduced for prediction. The document then analyzes wind speed data from Bhubaneswar using Weibull distributions and histograms to model the data distributions. Finally, it evaluates the hybrid Weibull-ANN technique for wind speed prediction performance.
Estimation of precipitation during the period of south west monsoonIAEME Publication
This document discusses a study that uses numerical methods in C programming to quantitatively estimate precipitation during the South West Monsoon period in India. The study collects temperature and precipitation data over 10 years and develops 2nd order equations to correlate temperature and computed precipitation for different monthly regions during the monsoon season. The equations are of the form Precipitation = C0 + C1*Temperature + C2*Temperature^2, where C0, C1, C2 are constants calculated using a C program. Segmenting the monsoon season into monthly regions provides better correlation between computed and actual precipitation compared to using the entire monsoon period. The study aims to quantitatively predict precipitation based on temperature to help farmers plan crop sowing.
IRJET - Intelligent Weather Forecasting using Machine Learning TechniquesIRJET Journal
This document discusses using machine learning techniques to forecast weather intelligently. It proposes using multi-target regression and recurrent neural network (RNN) models trained on historical weather data from Bangalore to predict future weather conditions like temperature, humidity, and precipitation. The data is first preprocessed before being fed to the models. The models are evaluated to accurately predict weather in the short term to help people like farmers and commuters without relying on expensive equipment.
Comparison of Solar Radiation Intensity Forecasting Using ANFIS and Multiple ...journalBEEI
This document compares the performance of two methods for forecasting solar radiation intensity: Adaptive Neuro Fuzzy Inference System (ANFIS) and Multiple Linear Regression (MLR). It uses weather data from Basel, Switzerland to test the methods. The ANFIS method uses a fuzzy inference system combined with neural networks, while MLR uses a mathematical approach. The performance of both methods is evaluated using root mean square error (RMSE) and mean absolute error (MAE) across different training/testing data compositions and time periods. The results show that ANFIS consistently provides lower error values than MLR, indicating it provides more accurate solar radiation forecasts.
Determination of wind energy potential of campus area of siirt universitymehmet şahin
In this study, wind energy potential of Siirt
University campus area is statistically examined by using the mean hourly wind speed data between 2014
and 2015 years which are measured by Vantage Pro2 device, located at the roof of the Engineering Faculty building with 6 m altitude. Weibull distribution
function and Rayleigh distribution function are used
as statistical approach to evaluate the wind data. Weibull distribution function is examined by using two different methods that are maximum likelihood estimation and Rayleigh method. The determination
coefficient (R2) and Root Mean Square Error (RMSE) values of these methods are compared. According the error analysis, it is indicated that the Rayleigh method
gives better results. Wind speed and wind power density are calculated in pursuance of Weibull distribution parameters. The results are evaluated as
monthly and annually. Hence, this preliminary study is made to determine the wind energy potential of Siirt University campus area.
an analysis of wind energy potential using weibull distributionWorking as a Lecturer
This document analyzes wind energy potential using the Weibull distribution. It discusses two studies that used the Weibull distribution to model wind speed and calculate parameters to estimate wind power potential. One study used a simulation model to describe wind turbine characteristics and power generated at a Sahara site in Algeria. The other calculated Weibull shape and scale factors using four methods and compared theoretical and observed probability density functions to determine the best fit. Both found the Weibull distribution directly influences estimates of wind power potential at a given location.
CFD down-scaling and online measurements for short-term wind power forecastingJean-Claude Meteodyn
Usually speaking, Forecast systems are classified : Intraday (Very Short term) is commonly Stochastic with online measurements while Extraday (Short term) is usually Deterministic based on NWP data. This work aims to breakdown these classifications, proposing a unique tool based on the unification of all these techniques.
VISUAL ANALYSIS OF ELECTRICITY DEMAND: ENERGY DASHBOARD GRAPHICS Graphical Da...Fatma ÇINAR
A real time interactive data management for Impulse and Response Analysis Technique using lattice and ggplot2 Graphical Packages embedded in R software has been employed. Average consumption, peak consumption and daily consumption data have been used while the temperature data is also employed to highlight the significance of relationship between consumption and the weather conditions. The demand for electricity by the factors affecting the demand with a multi-dimensional matrix graphics based on Energy Dashboard Software has been analysed leading to visualisation.
A WIND POWER PREDICTION METHOD BASED ON BAYESIAN FUSIONcsandit
The document proposes a wind power prediction method based on Bayesian fusion of multiple numerical weather predictions. It first establishes a relationship between wind speed and power using neural networks. It then analyzes the characteristics of wind speed forecasts from three independent weather sources. A Bayesian method is designed to fuse the wind speed forecasts, yielding a more accurate prediction than any single source. The fused wind speed is input to the neural network model to predict wind power at 15-minute intervals. Experimental results show the method improves accuracy of wind speed and power forecasting compared to using a single source.
This document discusses techniques for estimating solar energy. It begins by explaining the importance of accurate solar energy estimation for energy planners and utilities. It then describes three types of estimation horizons: very short term, short term, and long term prediction. Next, it outlines linear and nonlinear estimation techniques, including the Angstrom and Angstrom-Prescott models for linear regression. It focuses on the use of neural networks for short term estimation, noting their ability to model complex nonlinear relationships without extensive data or prior model specification. Finally, it provides examples of neural networks being used to estimate solar irradiance up to 24 hours ahead and assess solar potential in the Himalayan region.
This document provides a case study on forecasting monthly exceedance probabilities of solar radiation in Arizona. It discusses collecting solar radiation data from five stations in different locations in Arizona. The authors define exceedance probability as the probability of daily radiation being below an expected value. They use normalized distributions and simple linear regression to predict monthly exceedance probabilities and compare them to actual probabilities calculated from later test data. The document discusses setting up the model, including normalizing the data distributions and using data before 2011 to predict and data from 2011-2014 to test the predictions.
Photovoltaic Modules Performance Loss Evaluation for Nsukka, South East Niger...IJERA Editor
The Photovoltaic (PV) systems and technology offer excellent reliability when designed with the right implementation tools and based on good technical judgements of components that make up each of the critical sections of solar power system. The PV array is an essential section of a solar power system and it is expected to function to deliver pre – estimated power based on design estimations. There are factors that derail the performance of PV modules; the contributions of these factors are peculiar to specific sites of installation, hence the need to empirically evaluate and characterize installation sites before deployment of PV systems. This paper presents the characterization of Nsukka (South East, Nigeria) environment using decent instrumentation; and consequently highlights the power loss indicators for PV modules in the target site while presenting equally mitigable design.
This document discusses using a seasonal autoregressive integrated moving average (SARIMA) model to forecast precipitation in Mt. Kenya region. It fits various SARIMA models to monthly precipitation data from 1970 to 2011 and selects the best model with the lowest AIC and BIC values. The best model was found to be SARIMA(1,0,1)x(1,0,0)12, which had two statistically significant variables and passed diagnostic checks. Forecast accuracy statistics for this model, including ME, MSE, RMSE and MAE, indicated the SARIMA model provides a good method for precipitation forecasting in Mt. Kenya region.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
- In the past, experiments yielded limited data but today data is abundant and collected at low cost from each experiment
- Big data usually includes large data sets beyond the abilities of typical software to analyze in a reasonable time frame
- Data-driven models are based on analyzing relationships between input and output variables in a system using machine learning algorithms trained on representative data, without requiring explicit knowledge of the system's physical behavior
- Complexity of climate systems
- Climate modelling
- The need for modelling
- System thinking
- Analytical vs Numerical modeling
- Mathematical models
- Modeling process and model selection
- Model Uncertainty
- Modeling application and tools
Improved Kalman Filtered Neuro-Fuzzy Wind Speed Predictor For Real Data Set ...IJMER
This document presents three new models for short-term (24 hours ahead) wind speed forecasting for Egypt's northwestern coast based on real data collected from the site. The first model predicts wind speed using the same month of data from seven consecutive years. The second model predicts using only one month of data with a time series prediction scheme. The third model applies a discrete Kalman filter to one month of data first to reduce noise before prediction using an adaptive neuro-fuzzy inference system (ANFIS). The Kalman filtered data provided more accurate predictions with a 64% reduction in error compared to the first model.
The accurate prediction of solar irradiation has been
a leading problem for better energy scheduling approach.
Hence in this paper, an Artificial neural network based solar
irradiance is proposed for five days duration the data is
obtained from National Renewable Energy Laboratory, USA
and the simulation were performed using MATLAB 2013. It
was found that the neural model was able to predict the solar
irradiance with a mean square error of 0.0355.
Currently, the quality of wind measure of a site is assessed using Wind Power Density (WPD). This paper proposes to use a more credible metric namely, one we call the Wind Power Potential (WPP). While the former only uses wind speed information, the latter exploits both wind speed and wind direction distributions, and yields more credible estimates. The new measure of quality of a wind resource, the Wind Power Potential Evaluation (WPPE) model, investigates the effect of wind velocity distribution on the optimal net power generation of a farm. Bivariate normal distribution is used to characterize the stochastic variation of wind conditions (speed and direction). The net power generation for a particular farm size and installed capacity are maximized for different distributions of wind speed and wind direction, using the Unrestricted Wind Farm Layout Optimization (UWFLO) methodology. A response surface is constructed, using the recently developed Reliability Based Hybrid Functions (RBHF), to represent the computed maximum power generation as a function of the parameters of the wind velocity (speed and direction) distribution. To this end, for any farm site, we can (i) estimate the parameters of wind velocity distribution using recorded wind data, and (ii) predict the max- imum power generation for a specified farm size and capacity, using the developed response surface. The WPPE model is validated through recorded wind data at four differing stations obtained from the North Dakota Agricultural Weather Network (NDAWN). The results illustrate the variation of wind conditions and, subsequently, its influence on the quality of a wind resource.
A Post-processing Approach for Solar Power Combined Forecasts of Ramp EventsMohamed Abuella
This dissertation applies a post-processing approach to improve combined forecasts of solar power and solar power ramp events. The approach combines different forecasting models, then adjusts the combined forecasts to better predict ramp events. It develops a classification system for ramp event thresholds and evaluates performance using customized metrics. The approach provides probabilistic forecasts of solar power ramp events with uncertainty analysis.
Short Presentation: Mohamed abuella's Research HighlightsMohamed Abuella
- A post-processing approach combines and improves solar power forecasts and adjusts the combined forecasts in terms of ramp events.
- The approach classifies all possible thresholds and classes of ramp event forecasts and uses a customized cost function for imbalanced classification of ramp events.
- Suitable metrics are used for the feature selection process and performance evaluation, along with an uncertainty analysis for probabilistic forecasts of solar power ramp events.
Predictive Analysis of Global Solar Radiation in Awka Using Statistical Error...ijtsrd
Information on the accessibility of solar radiation at a location is an imperative factor in choosing appropriate solar energy system and devices for several applications. Sunshine hours, rainfall, cloud cover, atmospheric pressure data measured in Awka 06.20°N, 07.00°E , Anambra state for a period of nine 2005– 2013 were used to create Angstrom type regression equations models for estimating the global solar radiation received on a horizontal surface in Awka. The results of the correlation were also tested for error using statistical test methods of the mean bias error, MBE, root mean square error, RMSE, and mean percentage error, MPE, to calculate the performance of the models. It was perceived that combination of parameters could be used to estimate the total solar radiation incident on a location. Nwokoye, A. O. C | Mbadugha, A. O "Predictive Analysis of Global Solar Radiation in Awka Using Statistical Error Indicators" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-5 | Issue-4 , June 2021, URL: https://www.ijtsrd.compapers/ijtsrd41310.pdf Paper URL: https://www.ijtsrd.comphysics/nanotechnology/41310/predictive-analysis-of-global-solar-radiation-in-awka-using-statistical-error-indicators/nwokoye-a-o-c
Comparative Study of Selective Locations (Different region) for Power Generat...ijceronline
This document presents a comparative study of solar energy potential at four locations in Gujarat, India. Monthly solar radiation data from 2009-2013 was collected from agricultural universities located in central Gujarat (Anand), north Gujarat (S.K. Nagar), south Gujarat (Navsari), and Saurashtra (Junagadh). The data was analyzed using the Angstrom-Prescott equation to calculate solar radiation from sunshine hours and the Weibull distribution to model the frequency distribution of solar radiation. A statistical analysis using the correlation coefficient found Junagadh had the highest solar energy potential (R^2=0.988658), followed by S.K. Nagar, Navsari,
The document summarizes a study that used the Angstrom-Prescott model to estimate global solar radiation based on monthly sunshine hours in Mubi, Nigeria. The researchers obtained meteorological data from 2009-2013, including average monthly sunshine hours. They calculated the Angstrom constants a and b to be 0.27 and 0.54, respectively. The model had a good correlation coefficient of 0.87 compared to measured data. The estimated solar radiation values can be used to design and analyze solar energy systems for the region.
This document provides an introduction and overview of machine learning applications for numerical weather prediction systems. It discusses how machine learning has been used in weather and climate modeling for over 25 years, with many successful applications developed. These include using machine learning for model initialization, within numerical weather models, and for post-processing. Several specific examples are described, such as using neural networks for satellite retrievals, emulating model physics like radiation schemes, and for nonlinear ensemble post-processing. Machine learning is presented as a versatile tool that can help address challenges around data handling, model resolution, and representing subgrid scale processes.
International Journal of Engineering Research and Development (IJERD)IJERD Editor
journal publishing, how to publish research paper, Call For research paper, international journal, publishing a paper, IJERD, journal of science and technology, how to get a research paper published, publishing a paper, publishing of journal, publishing of research paper, reserach and review articles, IJERD Journal, How to publish your research paper, publish research paper, open access engineering journal, Engineering journal, Mathemetics journal, Physics journal, Chemistry journal, Computer Engineering, Computer Science journal, how to submit your paper, peer reviw journal, indexed journal, reserach and review articles, engineering journal, www.ijerd.com, research journals,
yahoo journals, bing journals, International Journal of Engineering Research and Development, google journals, hard copy of journal
FourCastNet is a global weather forecasting model that provides accurate short- to medium-range predictions at an unprecedented high resolution of 0.25° using a data-driven deep learning approach. FourCastNet matches or exceeds the accuracy of ECMWF's numerical weather prediction model IFS for variables like surface winds and precipitation. It generates week-long forecasts nearly 45,000 times faster than IFS, enabling large ensembles for improved probabilistic forecasting. FourCastNet uses an adaptive Fourier neural operator with a vision transformer backbone to make the first high-resolution global weather predictions from a deep learning model.
Solar Photovoltaic Power Forecasting in Jordan using Artificial Neural NetworksIJECEIAES
In this paper, Artificial Neural Networks (ANNs) are used to study the correlations between solar irradiance and solar photovoltaic (PV) output power which can be used for the development of a real-time prediction model to predict the next day produced power. Solar irradiance records were measured by ASU weather station located on the campus of Applied Science Private University (ASU), Amman, Jordan and the solar PV power outputs were extracted from the installed 264KWp power plant at the university. Intensive training experiments were carried out on 19249 records of data to find the optimum NN configurations and the testing results show excellent overall performance in the prediction of next 24 hours output power in KW reaching a Root Mean Square Error (RMSE) value of 0.0721. This research shows that machine learning algorithms hold some promise for the prediction of power production based on various weather conditions and measures which help in the management of energy flows and the optimisation of integrating PV plants into power systems.
The document describes a new regression model developed to estimate global solar radiation using artificial neural networks. The model was developed based on the Angstrom-Prescott model and used only latitude and altitude as inputs to estimate monthly average daily global solar radiation. The neural network model with 10 hidden neurons was trained on data from 4 locations in North India. The model estimated regression coefficients a ranging from 0.209 to 0.222 and b ranging from 0.253 to 0.407. Validation tests showed good agreement with actual solar radiation values.
1. The document discusses using sky imagers for short-term solar forecasting, as traditional methods lack sufficient spatial and temporal resolution for small-scale applications.
2. The proposed sky imager forecast model involves 7 steps: image analysis, cloud detection, cloud projection, shadow projection, irradiance modeling, predicting cloud motion to generate forecasts, and PV power modeling.
3. Accurate cloud detection, projection, and shadow projection are challenging due to issues like cloud inhomogeneity, perspective errors with distance from camera, and sensitivity to errors in estimated cloud base height.
This document describes the generation of a typical meteorological solar radiation year (TMY) for Armidale, New South Wales, Australia using 23 years of daily global solar radiation data. The Finkelstein-Schafer statistical method was used to select the most representative year of data for each month based on how closely its cumulative frequency distribution matched the long-term monthly average. The resulting typical year showed monthly average radiation values ranging from a low of 10.41 MJ/m2 in June to a high of 25.88 MJ/m2 in December. Comparison of the TMY data to the long-term monthly averages showed good agreement, indicating the TMY successfully captured typical solar conditions for Armidale
The document presents a new framework for solar irradiance forecasting that uses dynamic feature selection based on deep reinforcement learning. The framework first uses feature extraction and clustering techniques to identify representative weather data from a large dataset. It then uses deep reinforcement learning to develop a new feature selection method that identifies the minimum features required to accurately forecast solar irradiance from the representative data. The selected features are adapted to different weather conditions. Case studies show the model significantly reduces the data required for accurate irradiance forecasting under different weather patterns.
Numerical weather prediction has greatly improved forecast accuracy over the past 50 years. Prior to 1955, forecasting was subjective and not very skillful, relying on extrapolating weather patterns. Developments like computers, satellites, radar, and the establishment of observation networks allowed creating numerical models based on atmospheric equations. Ensemble prediction now provides probabilistic forecasts capturing forecast uncertainty by running models with different initializations. While resolution has increased, the chaotic nature of the atmosphere means uncertainty remains, requiring probabilistic forecasts over single predictions.
This document describes the generation of a typical meteorological year (TMY) of solar irradiance data on tilted surfaces for Armidale, New South Wales, Australia. It utilizes 23 years of daily solar radiation measurements from 1990 to 2012 to select the most representative months using the Finkelstein-Schafer statistical method. Models are used to estimate hourly solar radiation on tilted surfaces at angles of 15°, 30°, 45°, 60°, and 75° based on the typical meteorological year horizontal surface data. Tables of the estimated typical solar irradiance values are generated for each day of the year on the tilted surfaces, providing important input data for solar energy system design and performance modeling in Armidale.
This document discusses using clear sky irradiance and temperature models to normalize PV system performance data and mitigate the effects of sensor drift in degradation analysis. It introduces clear sky models for irradiance and temperature that can be used instead of on-site sensor data. Examples are given showing how normalization with clear sky models (PRCS) can eliminate artificial shifts and trends seen when using sensor data (PRSTC) due to sensor problems. Both static and dynamic clear sky models are evaluated, finding static models perform similarly to more complex dynamic models. The clear sky normalization approach trades some precision for accuracy and stability in degradation analysis.
Forecasting commercial services from IrSOLaV - Luis Martin (IRSOLAV)IrSOLaV Pomares
IrSOLaV is a spin-off company founded in 2007 that provides solar radiation data and forecasting services using satellite imagery. It has participated in over 3.4GW of solar energy projects worldwide. IrSOLaV's products include time series data, reports, and maps of solar radiation. It also provides short-term (72 hours) and nowcasting (6 hours) forecasts of solar radiation. IrSOLaV conducts research partnerships and has offices in Spain, India, and Chile.
Using statistical and machine learning techniques to forecast the PV solar power, which can be implemented for: • Managing the economic dispatch, unit commitment, and trading of PV solar power generations with other conventional generations; • Using with situational awareness tools to manage the ramp limitation; Optimal energy management of energy storage systems; • Voltage regulator settings on feeders with PV distributed generation.
Similar to Hourly probabilistic solar power forecasts (20)
The document is a presentation by Mohamed Abuella that outlines his background, motivation, and research experience. It includes:
- An introduction to Abuella's background in electrical engineering and interests in areas like AI and sustainability.
- His motivation for professional advancement through collaboration and personal advancement through better alignment with interests.
- An overview of his research experience including projects in Libya on electrical systems and a master's thesis on optimal power flow optimization using particle swarm methods.
The document is a presentation by Mohamed Abuella about himself and his research. It includes:
- An introduction to Abuella's background and interests in electrical engineering, power systems analysis, renewable energy integration, and artificial intelligence.
- An outline of his research experience including projects in Libya on electrical systems and a master's thesis on optimal power flow with wind power.
- Details of his PhD thesis on solar power forecasts and developing a method to adjust hour-ahead forecasts using day-ahead forecasts and ramp rates.
- His work applying machine learning techniques like random forests to improve probabilistic solar power forecasts and evaluating the improvements.
- Other research on modeling solar and wind energy potential in Libya and
Vessel Path Identification in Short-Sea ShippingMohamed Abuella
This document discusses clustering ship path data from two vessels, Cinderella II and an unnamed vessel, into meaningful segments. For Cinderella II data containing over 200,000 data points over 124 trips from July to November 2022 in the Baltic Sea, various clustering algorithms were applied including k-means, Gaussian mixture models, and hierarchical clustering using a distance matrix, achieving up to 100% accuracy. The unnamed vessel data containing over 39,000 data points over 1,754 trips was clustered into three groups. Random forest classification was also able to classify over 200,000 test points from Cinderella II into five path classes with 100% precision, recall and F1-score.
Data Analytics for Improving Energy Efficiency in Short-Sea ShippingMohamed Abuella
Descriptive, Predictive, and Prescriptive Analytics for Improving Energy Efficiency in Short-Sea Shipping. This project presents the major findings of the data analytics for improving the fuel consumption of a passenger ship in Archipelago of Gothenburg, Sweden.
Wind energy analysis for some locations in libyaMohamed Abuella
Modeling, Planning, and Analysis for Wind Energy in Libya.
17 Locations have been chosen for Modeling, Planning, and Analysis of Wind Energy for Different Locations in Libya, ...they are: Tripoli, Misurata, Tarhunah, Ghanima, Msallata, Zuwara, Gharyan, Sirte, Benghazi, Magrun, Derna, Houn, Gadamis, Sabha, Kufra, Tobruk, Jaghbub
Also, a comparison with Berlin in Germany has been conducted alongside the Libyan locations.
Planning and Analysis for Solar Energy in LibyaMohamed Abuella
This document analyzes solar energy potential across nine locations in Libya using typical meteorological year data. It finds that southern Libyan locations generally yield more solar energy but northern locations still have good potential compared to other world regions. The average net capacity factor across Libya is about 30%. Variability is higher for coastal locations than southern Sahara locations. Aggregating solar resources across multiple locations reduces variability and slightly increases capacity factors, demonstrating the benefits of geographic diversification for solar power deployment in Libya.
Adjusting post processing approach for very short-term solar power forecastsMohamed Abuella
This document presents research on adjusting post-processing approaches for very short-term solar power forecasts. It begins with motivating the need for solar power forecasts, then describes the methodology used which includes data collection from 3 PV sites, data preparation involving cleaning, imputation and detrending, and forecasting models including persistence, ARIMA, NAR, ANN and ELM. The models' forecasts are combined using random forest, and the combined forecasts are adjusted using a novel approach involving fitting to target-horizon and double target-horizon forecasts and their ramp rates. The adjusted forecasts and a simple average approach are evaluated on 9 cases across locations and horizons using RMSE. The adjusting approach achieves lower RMSE than the simple average
The presentation models and analyzes wind and solar energy resources for four US locations: Charlotte NC, Boston MA, Boulder CO, and Tucson AZ. Data on wind speed, solar irradiance, and other factors are used to calculate the energy output and net capacity factor for each location. Boston generally has the highest wind energy potential while Tucson has the greatest solar energy potential. Further work could involve modeling renewable energy backed by energy storage systems.
Study of NEPLAN Software for Load Flow and Short Faults Analysis تصميم شبكة ت...Mohamed Abuella
Thesis for: Bachelor of Technology, College of Industrial Technology, Misurata, Libya Advisor: Mohamed Shetwan
تصميم شبكة توزيع كهربائية لمشروع وحدات سكنية باستخدام نيبلان
مشروع تخرج لنيل درجة البكالوريس التقني
كلية التقنية الصناعية
مصراتة، ليبيا
اشراف/ أ. محمد شتوان
Mohamed abuella's cloud of key skills & interestsMohamed Abuella
Mohamed Abuella's document outlines his key skills and interests. It appears to be a resume or profile highlighting areas of expertise such as cloud computing. Overall, the document provides an overview of Mohamed Abuella's qualifications and experiences.
Qualifying combined solar power forecasts in ramp events' perspectiveMohamed Abuella
Applying a qualifying metric for solar power forecasts to assess their capability to predict the ramp events, especially by the combined forecasts, then those forecasts can be implemented for: Managing high ramp-rates of PV solar power generation; Optimal energy management of energy storage systems; Voltage regulator settings on feeders with PV distributed generation.
Study the Effects of Renewable Resources on Electric Grid FrequencyMohamed Abuella
This document summarizes a study on the effects of renewable resources on electric grid frequency. It discusses how variable generation from renewables like wind and solar can cause frequency deviation due to mismatches between generation and demand. Frequency deviation can impact customer equipment and cause the grid to shed loads or generators. Case studies on islands like Bornholm, Sicily, and Hawaii show how high renewable penetration leads to larger frequency drops from smaller generation losses. Various mitigation techniques are discussed, including controlling inertial response in wind turbines, curtailing wind power to provide governor-like frequency response, retrofitting solar inverters, using energy storage, and modifying AGC control schemes.
Study of using particle swarm for optimal power flowMohamed Abuella
The document describes a study using particle swarm optimization to solve the optimal power flow problem in power systems including wind power generators. The study models the intermittent nature of wind power using a Weibull probability distribution and transforms wind speed to power probabilities. The optimal power flow problem is formulated as a mathematical optimization problem minimizing generation costs subject to constraints. Particle swarm optimization is applied to solve the optimal power flow problem, with the algorithm updating particles based on personal and global best positions to iteratively search for the minimum cost solution satisfying all constraints.
1) The presentation discusses the smart grid, including its motivation due to issues with the current electric grid like fossil fuel scarcity and reliability concerns. It outlines the history and definitions of the smart grid.
2) Key requirements and characteristics of the smart grid are described, such as advanced monitoring and communication infrastructure to integrate renewable energy and provide two-way energy and information flow.
3) A case study of a smart grid deployment in Boulder, Colorado is summarized, including installing smart meters and fiber optic lines to provide customers with energy usage monitoring and more pricing options. Results showed potential cost savings and standardization needs.
Using recycled concrete aggregates (RCA) for pavements is crucial to achieving sustainability. Implementing RCA for new pavement can minimize carbon footprint, conserve natural resources, reduce harmful emissions, and lower life cycle costs. Compared to natural aggregate (NA), RCA pavement has fewer comprehensive studies and sustainability assessments.
Embedded machine learning-based road conditions and driving behavior monitoringIJECEIAES
Car accident rates have increased in recent years, resulting in losses in human lives, properties, and other financial costs. An embedded machine learning-based system is developed to address this critical issue. The system can monitor road conditions, detect driving patterns, and identify aggressive driving behaviors. The system is based on neural networks trained on a comprehensive dataset of driving events, driving styles, and road conditions. The system effectively detects potential risks and helps mitigate the frequency and impact of accidents. The primary goal is to ensure the safety of drivers and vehicles. Collecting data involved gathering information on three key road events: normal street and normal drive, speed bumps, circular yellow speed bumps, and three aggressive driving actions: sudden start, sudden stop, and sudden entry. The gathered data is processed and analyzed using a machine learning system designed for limited power and memory devices. The developed system resulted in 91.9% accuracy, 93.6% precision, and 92% recall. The achieved inference time on an Arduino Nano 33 BLE Sense with a 32-bit CPU running at 64 MHz is 34 ms and requires 2.6 kB peak RAM and 139.9 kB program flash memory, making it suitable for resource-constrained embedded systems.
Comparative analysis between traditional aquaponics and reconstructed aquapon...bijceesjournal
The aquaponic system of planting is a method that does not require soil usage. It is a method that only needs water, fish, lava rocks (a substitute for soil), and plants. Aquaponic systems are sustainable and environmentally friendly. Its use not only helps to plant in small spaces but also helps reduce artificial chemical use and minimizes excess water use, as aquaponics consumes 90% less water than soil-based gardening. The study applied a descriptive and experimental design to assess and compare conventional and reconstructed aquaponic methods for reproducing tomatoes. The researchers created an observation checklist to determine the significant factors of the study. The study aims to determine the significant difference between traditional aquaponics and reconstructed aquaponics systems propagating tomatoes in terms of height, weight, girth, and number of fruits. The reconstructed aquaponics system’s higher growth yield results in a much more nourished crop than the traditional aquaponics system. It is superior in its number of fruits, height, weight, and girth measurement. Moreover, the reconstructed aquaponics system is proven to eliminate all the hindrances present in the traditional aquaponics system, which are overcrowding of fish, algae growth, pest problems, contaminated water, and dead fish.
Harnessing WebAssembly for Real-time Stateless Streaming PipelinesChristina Lin
Traditionally, dealing with real-time data pipelines has involved significant overhead, even for straightforward tasks like data transformation or masking. However, in this talk, we’ll venture into the dynamic realm of WebAssembly (WASM) and discover how it can revolutionize the creation of stateless streaming pipelines within a Kafka (Redpanda) broker. These pipelines are adept at managing low-latency, high-data-volume scenarios.
DEEP LEARNING FOR SMART GRID INTRUSION DETECTION: A HYBRID CNN-LSTM-BASED MODELgerogepatton
As digital technology becomes more deeply embedded in power systems, protecting the communication
networks of Smart Grids (SG) has emerged as a critical concern. Distributed Network Protocol 3 (DNP3)
represents a multi-tiered application layer protocol extensively utilized in Supervisory Control and Data
Acquisition (SCADA)-based smart grids to facilitate real-time data gathering and control functionalities.
Robust Intrusion Detection Systems (IDS) are necessary for early threat detection and mitigation because
of the interconnection of these networks, which makes them vulnerable to a variety of cyberattacks. To
solve this issue, this paper develops a hybrid Deep Learning (DL) model specifically designed for intrusion
detection in smart grids. The proposed approach is a combination of the Convolutional Neural Network
(CNN) and the Long-Short-Term Memory algorithms (LSTM). We employed a recent intrusion detection
dataset (DNP3), which focuses on unauthorized commands and Denial of Service (DoS) cyberattacks, to
train and test our model. The results of our experiments show that our CNN-LSTM method is much better
at finding smart grid intrusions than other deep learning algorithms used for classification. In addition,
our proposed approach improves accuracy, precision, recall, and F1 score, achieving a high detection
accuracy rate of 99.50%.
Eric Nizeyimana's document 2006 from gicumbi to ttc nyamata handball play
Hourly probabilistic solar power forecasts
1. 17-19 September 2017 - 49th North American Power Symposium
Hourly Probabilistic Solar Power Forecasts
Energy Production and Infrastructure Center
Department of Electrical and Computer Engineering
University of North Carolina at Charlotte
Mohamed Abuella
Prof. Badrul Chowdhury
September 18, 2017
2. 17-19 September 2017 - 49th North American Power Symposium
2
Presentation Outline
Overview of Solar Power
Forecasting
Hourly Probabilistic
Forecasting of Solar Power
3. 17-19 September 2017 - 49th North American Power Symposium
3
Renewables Generations
(Wind and Solar) are Too
Variable
High Efficiency and
Large Energy Storage
Still not Exist
Reducing
Cost
and Pollution
Why
Forecast?Variable Generations (V.G.) Forecasting
4. 17-19 September 2017 - 49th North American Power Symposium
VG forecasting in US. electric utilities and ISO,
such as CAISO, ERCOT, MISO, ISO-NE, NYISO,…etc.
Botterud, J. Wang, V. Miranda, and R. J. Bessa, “Wind power forecasting in US electricity markets,” The Electricity Journal, vol. 23, no. 3, pp. 71–82, 2010.
Elke Lorenz, “Solar Resource Forecasting” International Solar Energy Society (ISES) Webinar, 2016.
Voyant, C., Notton, G., Kalogirou, S., Nivet, M. L., Paoli, C., Motte, F., & Fouilloy, A. (2017). Machine learning methods for solar radiation forecasting: A
review. Renewable Energy, 105, 569-582.
Forecast Horizons
Post operating reserve
requirements
Clear DA market
using SCUC/SCED
Rebidding
for RAC Post-DA RAC
using SCUC
Prepare and
submit DA bids
Clear RT market using
SCED (every 5min)
Intraday RAC
using SCUC
Prepare and
submit RT bids Post results (RT
energy and
reserves)
Post results (DA
energy and
reserves)
Day Ahead:
Operating Day:
11:00 16:00 17:00
-30min
Operating hour
DA: Day Ahead.
RT: Real Time.
SCUC: Security Constrained Unit
Commitment.
SCED: Security Constrained
Economic Dispatch.
RAC: Reliability Assessment
Commitment.
Wind / Solar
Power
Forecasting
Intra-hour Intra-day Day ahead
Forecasting horizon 15 min to 2 h 1 h to 6 h 1 day to 3 day
Granularity-Time step 30 s to 5 min hourly hourly
Related to
Ramping events,
variability related
to operations
Load following
forecasting
Unit
commitment,
transmission
scheduling, day
ahead markets
Forecasting Models Total Sky Imager and/or time series
Satellite Imagery and/or NWP
Variable Generations (V.G.) Forecasting
Forecast horizons, forecasting models and
the related applications of VG forecasts
4
Where
Forecast?
5. 17-19 September 2017 - 49th North American Power Symposium
Flowchart of the Combined Approach (Physical and
Statistical) of the Solar Power Forecasting
;
Solar Plant and
Terrain
Characteristics
Numerical Weather Prediction
(NWP)
Atmospheric variables
SCADA Data
Spatial Refinement
Local Roughness, topology.
Atmospheric Stability
Solar Power Plant
Modeling
Solar Power Conversion Equation/s
Model Output Statistics (MOS)
Systematic Error Correction
Solar Generation Forecast
Conversion to Power
Downscaling
Regression
Extrapolation
Solar Power Forecasting
5
How
Forecast?
6. 17-19 September 2017 - 49th North American Power Symposium
Flowchart of the Solar Power Forecasting
6
Pre-Processing
Outlier detection and data
cleansing
Feature engineering
Weather Data
Solar irradiance
Temperature
Cloud coverage
Humidity
...etc.
PV System Data
Measured PV power output
Location and modules type,
orientation, tilt,..etc. Forecasting Models
Persistence model
Statistical models
Artificial intelligence models
Point forecast
Post-Processing
Ensemble
Analog ensemble
Probabilistic forecast
Combining the models’ outcomes
by ensemble learning
Hourly Probabilistic Forecasting of Solar Power
7. 17-19 September 2017 - 49th North American Power Symposium
Data Description:
PV solar system is near Canberra, Australia. The panel type is Solarfun SF160-24-1M195,
consisting of 8 panels, its nominal power of (1560W), and panel orientation 38° clockwise from
the north, with panel tilt (of 36°). The historical observed solar power data are normalized to the
rated capacity (i.e., 1560W).
https://crowdanalytix.com/contests/global-energy-forecasting-competition-2014-probabilistic-solar-power-forecasting
http://www.ecmwf.int (European Centre for Medium-Range Weather Forecasts)
Training
Testing
Hourly Probabilistic Forecasting of Solar Power
The weather forecast data and the available
measured solar power data from April 2012 to
May 2014.
No. Month Year
1 April 2012
2 May 2012
3 June 2012
4 July 2012
5 August 2012
6 September 2012
7 October 2012
8 November 2012
9 December 2012
10 January 2013
11 February 2013
12 March 2013
13 April 2013
14 May 2013
15 June 2013
16 July 2013
17 August 2013
18 September 2013
19 October 2013
20 November 2013
21 December 2013
22 January 2014
23 February 2014
24 March 2014
25 April 2014
26 May 2014
Weather Variables
No. Variable Name
1 Cloud Water Content
2 Cloud Ice Content
3 Surface Pressure
4 Relative Humidity
5 Cloud Cover
6 10m-U Wind
7 10m-V Wind
8 2-m Temperature
9 Surface solar radiation down
10 Surface thermal radiation down
11 Top net solar radiation
12 Total precipitation
Weather predictions are produced by the global numerical weather prediction system (ECMWF).
7
8. 17-19 September 2017 - 49th North American Power Symposium
Multiple Linear Regression (MLR) Analysis, Artificial Neural Networks (ANN), and Support
Vector Regression (SVR) are deployed for the short-term solar power forecasting.
Flowchart of the forecasting models
New Input X
Predicted Output Y
Flowchart of solar forecasting model building steps
Building
the Model
Training Validation Testing
Forecasting Models
Parametric and Nonparametric Models
T. Hastie, R. Tibshirani, J. Friedman, and others, The elements of statistical learning, 2nd Edition. Springer-Verlag New York,
2009. 8
9. 17-19 September 2017 - 49th North American Power Symposium
General diagram of combining different models
Model B
Model A
Model C
Model N
Method of
Combining the
Models
Combined Forecasts
Individual
forecasting
models
Combining Various Models
Methods of
Combining The
Models
Random forest (RF) is chosen to be the ensemble learning
method for combining the various models’ outcomes.
Fcomb=WA*MA+ WB*MB + WC*MC ….+ WN*MN
WN is a weight is assigned to the outcome of a model MN
Ensemble Forecasts
T. Hastie, R. Tibshirani, J. Friedman, and others, The elements of statistical learning, 2nd
Edition. Springer-Verlag New York, 2009. 9
10. 17-19 September 2017 - 49th North American Power Symposium
Day Month
1
:
June
: July
:
:
:
:
:
:
: April
30 May
31
00:00
:
23:00
Weather Data
:
:
:
:
: :
:
:
:
:
:
:
: :
: :
:
:
:
:
:
:
PV Power
:
:
(PastObservations)
:
:
:
:
:
:
Forecasts
(Model’s
Outcomes)
at 00:00
Day Month
1
:
June
: July
:
:
:
:
:
:
: April
30 May
31
00:00
:
23:00
Weather Data
:
:
:
:
: :
:
:
:
:
:
:
: :
: :
:
:
:
:
:
:
Models’ Outcomes
:
:
:
:
:
:
: : :
:
:
:
:
:
:
:
:
:
: : :
: : :
Forecasts
PV Power
:
:
(PastObservations)
:
:
:
:
:
:
Combined
Forecasts
(a) (b)
Persistence model and day-ahead Forecasts from MLR, ANN and SVR
Combining by Radom ForestProducing Different Models’ Outcomes
Schematic diagram of producing and ensemble different models’ outcomes
10
Ensemble Forecasts
𝑃𝑒𝑟𝑠𝑖𝑠𝑡𝑒𝑛𝑐𝑒 𝑚𝑜𝑑𝑒𝑙, 𝐹(𝑡 = 𝑃(𝑡 − ℎ𝑜𝑟𝑖𝑧𝑜𝑛
11. 17-19 September 2017 - 49th North American Power Symposium
𝑓𝑅𝐹 =
1
𝐵
𝑏=1
𝐵
𝑇𝑏 𝑥 [Ref. ]
Each tree of the random forest is trained with random samples (random input dataset).
In the forecasting process (test) the trees give different forecasts, which can be then used for
quantifying the uncertainty of the average output and producing the probabilistic forecasts.
The prediction of a given point x of the response
variable is then obtained by averaging the individual
trees outputs:
T. Hastie, R. Tibshirani, J. Friedman, and others, The elements of statistical learning, 2 Edition. Springer-Verlag New
York, 2009.
Note: In fact, tree splitting is not simple as it seems,
since x can comes with dimensions up to 20 dimensions.
Probabilistic Forecasts
Ensemble-based probabilistic forecasts method:
11
12. 17-19 September 2017 - 49th North American Power Symposium
Analog Ensemble (AnEn) method:
S. Alessandrini, L. Delle Monache, S. Sperati, and G. Cervone, “An analog ensemble for short-term probabilistic solar power
forecast,” Appl. Energy, vol. 157, pp. 95–110, 2015.
Probability
Distribution
Observed Solar Power
Given Point ForecastPast Solar Power Forecasts
Probabilistic Forecasts
where 𝐹Given
𝐻𝑟
denotes the given point forecast at an hour Hr, for which the
prediction interval will be estimated, 𝐹Past
𝐻𝑟
the point forecasts at the same hour
of the day.
Notice that all values are normalized in the range [0, 1].
𝐹Given
𝐻𝑟
− 𝐹Past
𝐻𝑟
≤ ε
Schematic diagram of analog ensemble method
ε = 0.1
12
13. 17-19 September 2017 - 49th North American Power Symposium
Persistence probabilistic forecasts method:
Probabilistic Forecasts
Probability
Distribution
Observed Solar Power
Given Point
Forecast
tt-24t-48t-72
Schematic diagram of analog ensemble method
The 10, 20 and 30 recent observed powers are carried out.
It is found that the recent 10 observed solar powers at the given hour with CDF distribution
achieve more accurate persistence probabilistic forecasts.
13
No past forecasts are needed.
14. 17-19 September 2017 - 49th North American Power Symposium
Probabilistic Forecasts
Different distributions of probability
Probability distributions by the cumulative distribution function (CDF)
14
Linear CDF
Max
Min
to derive CDF
Parametric normal-
distributed CDF
Mean
Std. dev.
to derive CDF.
Nonparametric CDF
No mean
neither Std. Dev.
CDF is estimated by
piecewise
nonparametric method
The probabilistic forecasts are estimated by using CDF-1
For example, for a given point forecast at 14:00, June 2nd 2013:
Histogram of the ensemble of RF’s outcomes
15. 17-19 September 2017 - 49th North American Power Symposium
The objective is to determine probabilistic solar forecasts in the form of probabilistic distribution
(in quantiles) in incremental time steps through the forecast horizon.
A Pinball loss function is used to evaluate the accuracy of the probabilistic forecasts. It is a
piecewise linear function which is often used to evaluate the accuracy of quantile forecasts.
J. M. Morales, A. J. Conejo, H. Madsen, P. Pinson, and M. Zugno, Integrating renewables in electricity markets - Operational problems, vol. 205.
Boston, MA: Springer US, 2014.
where Pbq(F, P) is the pinball loss function to the probabilistic forecasts for each hour; 𝐹 is the forecasted value at the
certain q quantile of the probabilistic solar power forecasts, and P is the observed value of the solar power. The quantile q has
discrete values 𝑞 ∈ [0.01, 0.99].For instance, q = 0.9 means that there is a 90% probability that the observed solar power will
be less than the value of the 90th quantile.
Probabilistic Forecasts
Evaluation of probabilistic forecasts:
𝑃𝑏 𝑞(𝐹, 𝑃 =
𝑞 𝐹 − 𝑃 , 𝑖𝑓 𝑃 ≤ 𝐹
1 − 𝑞 𝑃 − 𝐹 , 𝑖𝑓 𝑃 > 𝐹
15
16. 17-19 September 2017 - 49th North American Power Symposium
Graphs of the probabilistic forecasts of the three methods for three days
Results and Evaluation
The lower Pinball (Pb) is, the more
accurate probabilistic forecasts are.
Pinball loss function (Pb):
𝑃𝑏 𝑞(𝐹, 𝑃 =
𝑞 𝐹 − 𝑃 , 𝑖𝑓 𝑃 ≤ 𝐹
1 − 𝑞 𝑃 − 𝐹 , 𝑖𝑓 𝑃 > 𝐹
16
17. 17-19 September 2017 - 49th North American Power Symposium
Monthly Pinball of the probabilistic forecasts of the three methods
Results and Evaluation
Month
Pinball (Pb)
Improvement of
Ensemble Over
Persistence
Analog
Ensemble
Ensemble Persistence
Analog
Ensemble
June 0.0166 0.0097 0.0095 42% 2%
July 0.0176 0.0122 0.0124 29% -2%
August 0.0182 0.0111 0.0117 36% -5%
September 0.0173 0.0116 0.0115 34% 1%
October 0.0149 0.0096 0.0096 36% 0%
November 0.0191 0.0104 0.0106 45% -2%
December 0.0162 0.0090 0.0089 45% 1%
January 0.0179 0.0083 0.0080 55% 4%
February 0.0215 0.0104 0.0098 54% 6%
March 0.0208 0.0128 0.0131 37% -2%
April 0.0194 0.0103 0.0099 49% 3%
May 0.0137 0.0086 0.0080 42% 7%
Average Pb. 0.0178 0.0103 0.0102 42% 1%
Pinball loss function (Pb):
The lower Pinball (Pb) is, the more
accurate probabilistic forecasts are.𝑃𝑏 𝑞(𝐹, 𝑃 =
𝑞 𝐹 − 𝑃 , 𝑖𝑓 𝑃 ≤ 𝐹
1 − 𝑞 𝑃 − 𝐹 , 𝑖𝑓 𝑃 > 𝐹
17
18. 17-19 September 2017 - 49th North American Power Symposium
Results and Evaluation
-8%
-6%
-4%
-2%
0%
2%
4%
6%
8%
Improvement
The Improvements of Ensemble over Analog Ensemble
0.0000
0.0050
0.0100
0.0150
0.0200
0.0250
Pinball
Probabilistic Forecasts
Persistence Analog Ensemble Ensemble
18
19. 17-19 September 2017 - 49th North American Power Symposium
Conclusions
The probabilistic forecasting are quantifying the uncertainty associated with point forecasts.
Combining the forecasts of various models leads to accurate point and probabilistic forecasts.
Throughout the complete year, the ensemble based-probabilistic forecasts are more accurate.
than the analog ensemble and persistence probabilistic forecasts.
The random forest is a powerful ensemble learning method.
The CDF with the assumption of a normal distribution is better than the linear distribution to
produce the probabilistic forecasts.
The nonparametric estimation of CDF without the normality assumption yields a small
improvement (Pb=0.0100 vs. Pb=0.0102 with a normality assumption of CDF).
With additional historical data, the forecasting performance could be improved.
19
20. 17-19 September 2017 - 49th North American Power Symposium
Thanks for Your Listening
Any Question?
http://epic.uncc.edu/
Mohamed Abuella
mabuella@uncc.edu
Energy Production and Infrastructure Center
University of North Carolina at Charlotte