Operational Scheduling of Behind-the-Meter Storage Systems Based on Multiple Nonstationary Decomposition and Deep Convolutional Neural Network for Price Forecasting.

In the competitive electricity market, electricity price reflects the relationship between power supply and demand and plays an important role in the strategic behavior of market players. With the development of energy storage systems after watt-hour meter, accurate price prediction becomes more and more crucial in the energy management and control of energy storage systems. Due to the great uncertainty of electricity price, the performance of the general electricity price forecasting models is not satisfactory to be adopted in practice. Therefore, in this paper, we propose a novel electricity price forecasting strategy applied in optimization for the scheduling of battery energy storage systems. At first, multiple nonstationary decompositions are presented to extract the most significant components in price series, which express remarkably discriminative features in price fluctuation for regression prediction. In addition, all extracted components are delivered to a devised deep convolution neural network with multiscale dilated kernels for multistep price forecasting. At last, more advanced price fluctuation detection serves the optimized operation of the battery energy storage system within Ontario grid-connected microgrids. Sufficient ablation studies showed that our proposed price forecasting strategy provides predominant performances compared with the state-of-the-art methods and implies a promising prospect in economic benefits of battery energy storage systems.

Dynamic Regulation of Level Set Parameters Using 3D Convolutional Neural Network for Liver Tumor Segmentation.

Segmentation of liver tumors plays an important role in the choice of therapeutic strategies for liver disease and treatment monitoring. In this paper, we generalize the process of a level set with a novel algorithm of dynamic regulation to energy functional parameters. The presented method is fully automatic once the tumor has been detected. First, a 3D convolutional neural network with dense layers for classification is used to estimate current contour location relative to the tumor boundary. Second, the output 3D CNN probabilities can dynamically regulate parameters of the level set functional over the process of segmentation. Finally, for full automation, appropriate initializations and local window size are generated based on the current contour position probabilities. We demonstrate the proposed method on the dataset of MICCAI 2017 LiTS Challenge and 3DIRCADb that include low contrast and heterogeneous tumors as well as noisy images. To illustrate the strength of our method, we evaluated it against the state-of-the-art methods. Compared with the level set framework with fixed parameters, our method performed better significantly with an average DICE improvement of 0.15. We also analyzed a challenging dataset 3DIRCADb of tumors and obtained a competitive DICE of 0.85 ± 0.06 with the proposed method.

Lung Nodule Image Classification Based on Local Difference Pattern and Combined Classifier.

This paper proposes a novel lung nodule classification method for low-dose CT images. The method includes two stages. First, Local Difference Pattern (LDP) is proposed to encode the feature representation, which is extracted by comparing intensity difference along circular regions centered at the lung nodule. Then, the single-center classifier is trained based on LDP. Due to the diversity of feature distribution for different class, the training images are further clustered into multiple cores and the multicenter classifier is constructed. The two classifiers are combined to make the final decision. Experimental results on public dataset show the superior performance of LDP and the combined classifier.