With the increasing penetration of wind power in renewable energy systems, it is important to improve the accuracy of wind speed prediction. However, wind power generation has great uncertainties which make high-quality interval prediction a challenge. Existing multi-objective optimization interval prediction methods do not consider the robustness of the model. Thus, trained models for wind speed interval prediction may not be optimal for future predictions. In this paper, the prediction interval coverage probability, the prediction interval average width, and the robustness of the model are used as three objective functions for determining the optimal model of short-term wind speed interval prediction using multi-objective optimization. Furthermore, a new Stochastic Sensitivity for Prediction Intervals (SS_PIs) is proposed in this work to measure the stability and robustness of the model for interval prediction. Using wind farm data from countries on two different continents as case studies, experimental results show that the proposed method yields better prediction intervals in terms of all metrics including prediction interval coverage probability (PICP), prediction interval normalized average width (PINAW) and SS_PIs. For example, at the prediction interval nominal confidence (PINC) of 85%, 90% and 95%, the proposed method has the best performance in all metrics of the USA wind farm dataset.
Considering that most of the photovoltaic (PV) data are behind-the-meter (BTM), there is a great challenge to implement effective demand response projects and make a precise customer baseline (CBL) prediction. To solve the problem, this paper proposes a data-driven PV output power estimation approach using only net load data, temperature data, and solar irradiation data. We first obtain the relationship between delta actual load and delta temperature by calculating the delta net load from matching the net load of irradiation for an approximate day with the least squares method. Then we match and make a difference of the net load with similar electricity consumption behavior to establish the relationship between delta PV output power and delta irradiation. Finally, we get the PV output power and implement PV-load decoupling by modifying the relationship between delta PV and delta irradiation. The case studies verify the effectiveness of the approach and it provides an important reference to perform PV-load decoupling and CBL prediction in a residential distribution network with BTM PV systems.
Accurate load forecasting is essential to the operation and planning of power systems and electricity markets. In this paper, an ensemble of radial basis function neural networks (RBFNNs) is proposed which is trained by minimizing the localized generalization error for shortterm and mid-term load forecasting. Exogenous features and features extracted from load series (with long shortterm memory networks and multi-resolution wavelet transform) in various timescales are used to train the ensemble of RBFNNs. Multiple RBFNNs are fused as an ensemble model with high generalization capability using a proposed weighted fusion method based on the localized generalization error model. Experimental results on three practical datasets show that compared with other forecasting methods, the proposed method reduces the mean absolute percentage error (MAPE), mean squared error (MSE), mean absolute error (MAE) by at least 0.12%, 8.46 (MW) 2 , 0.83 MW in mid-term load forecasting (i.e., to predict the daily peak load of next month), respectively, and reduces the MAPE, MSE by at least 0.19%, 2009.69 (MW) 2 and 0.30%, 3697.18 (MW) 2 in half-hour-ahead forecasting and day-ahead forecasting, respectively.
Photovoltaic (PV) generation is increasing in distribution systems following policies and incentives to promote zero-carbon emission societies. Most residential PV systems are installed behind-the-meter (BTM). Due to single meter deployment that measures the net load only, this PV generation is invisible to distribution system operators causing a negative impact on the distribution system planning and local supply and demand balance. This paper proposes a novel data-driven BTM PV generation disaggregation method using only net load and weather data, without relying on other PV proxies and PV panels' physical models. Long Short-Term Memory (LSTM) is employed to build a generation difference fitted model (GDFM) and a consumption difference fitted model (CDFM) derived from weather data. Both difference fitted models are refined by a crossiteration with mutual output. Finally, considering the photoelectric conversion properties, the disaggregated generation results are acquired by the refined GDFM of changing input. The proposed method has been tested with actual smart meter data of Austin, Texas and proves to increase the disaggregated accuracy as compared to current state-of-the-art methods. The proposed method is also applicable to disaggregate BTM PV systems of different manufacturing processes and types.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.