A computational intelligence scheme for the prediction of the daily peak load

Estimating Energy Forecasting Uncertainty for Reliable AI Autonomous Smart Grid Design

Building safe, reliable, fully automated energy smart grid systems requires a trustworthy electric load forecasting system. Recent work has shown the efficacy of Long Short-Term Memory neural networks in energy load forecasting. However, such predictions do not come with an estimate of uncertainty, which can be dangerous when critical decisions are being made autonomously in energy production and distribution. In this paper, we present methods for evaluating the uncertainty in short-term electrical load predictions for both deep learning and gradient tree boosting. We train Bayesian deep learning and gradient boosting models with real electric load data and show that an uncertainty estimate may be obtained alongside the prediction itself with minimal loss of accuracy. We find that the uncertainty estimates obtained are robust to changes in the input features. This result is an important step in building reliable autonomous smart grids.

Day-Ahead Electric Load Forecasting for the Residential Building with a Small-Size Dataset Based on a Self-Organizing Map and a Stacking Ensemble Learning Method

Electric load forecasting for buildings is important as it assists building managers or system operators to plan energy usage and strategize accordingly. Recent increases in the adoption of advanced metering infrastructure (AMI) have made building electrical consumption data available, and this has increased the feasibility of data-driven load forecasting. Self-organizing map (SOM) has been successfully utilized to cluster a dataset into subsets containing similar data points. These subsets are then used to train the forecasting models to improve forecasting accuracy. However, some buildings may have insufficient data since newly installed monitoring devices such as AMI have no choice but to collect a limited amount of data. Using a clustering technique on small datasets could lead to overfitting when using forecasting models following an SOM network to be trained with clusters. This results in a relatively high generalization error. In this study, we propose to address this problem by employing the stacking ensemble learning method (SELM) that is well-known for its generalization ability. An experimental study was conducted using the electricity consumption data of an actual institutional building and meteorological data. Our proposed model outperformed other baseline models, which means it successfully mitigates the effect of overfitting.

An Intelligent and Dynamic Decision Support System for Nonlinear Environments

Nonlinear time series systems are high dimensional and chaotic in nature. Since, the design of a dynamic and efficient decision making system is a challenging task, a Support Vector Machine (SVM) based model is proposed to predict the future event of a nonlinear time series environment. This model is a non-parametric model that uses the inherent structure of the data for forecasting. The Hybrid Dimensionality Reduction (HDR) and Extended Hybrid Dimensionality Reduction (EHDR) techniques are proposed to represent the time series data and to reduce the dimensionality and control noise besides subsequencing the time series data. The proposed SVM based model using EHDR is compared with the models using Symbolic Aggregate approXimation (SAX), HDR, SVM using Kernel Principal Component Analysis(KPCA) and SVM using varying tube size values for historical data on different financial instruments. The experimental results have proved that the prediction accuracy of the proposed model is better compared with other models taken for the experimentation.