Reliable prediction of remaining useful life (RUL) plays an indispensable role in prognostics and health management (PHM) by reason of the increasing safety requirements of industrial equipment. Meanwhile, data-driven methods in RUL prognostics have attracted widespread interest. Deep learning as a promising data-driven method has been developed to predict RUL due to its ability to deal with abundant complex data. In this paper, a novel scheme based on a health indicator (HI) and a hybrid deep neural network (DNN) model is proposed to predict RUL by analyzing equipment degradation. Explicitly, HI obtained by polynomial regression is combined with a convolutional neural network (CNN) and long short-term memory (LSTM) neural network to extract spatial and temporal features for efficacious prognostics. More specifically, valid data selected from the raw sensor data are transformed into a one-dimensional HI at first. Next, both the preselected data and HI are sequentially fed into the CNN layer and LSTM layer in order to extract high-level spatial features and long-term temporal dependency features. Furthermore, a fully connected neural network is employed to achieve a regression model of RUL prognostics. Lastly, validated with the aid of numerical and graphic results by an equipment RUL dataset from the Commercial Modular Aero-Propulsion System Simulation(C-MAPSS), the proposed scheme turns out to be superior to four existing models regarding accuracy and effectiveness.

Precise prediction of short-term electric load demand is the key for developing power market strategies. Due to the dynamic environment of short-term load forecasting, probabilistic forecasting has become the center of attention for its ability of representing uncertainty. In this paper, an integration scheme mainly composed of correlation analysis and improved weighted extreme learning machine is proposed for probabilistic load forecasting. In this scheme, a novel cooperation of wavelet packet transform and correlation analysis is developed to deal with the data noise. Meanwhile, an improved weighted extreme learning machine with a new switch algorithm is provided to effectively obtain stable forecasting results. The probabilistic forecasting task is then accomplished by generating the confidence intervals with the Gaussian process. The proposed integration scheme, tested by actual data from Global Energy Forecasting Competition, is proved to have a better performance in graphic and numerical results than the other available methods.