Quality prediction plays an essential role in industrial process monitoring and optimization. However, the existed methods lack of the ability to extract features in multiphase and nonlinear time series data. To improve the quality prediction accuracy and simultaneously provide some suggestion of the process mechanism, this article proposes a phased dual-attention long short-term memory (PDA-LSTM) neural network, which consists of a phase segmentation module and a prediction module. The whole series data are first divided into phases in the phase segmentation module, and individual models are subsequently constructed for each phase in the prediction module to capture distinct patterns. The LSTM network with input and temporal attention mechanism is utilized to select the most relevant variables and time steps throughout the time sequence. The proposed method is verified with a numerical example and a real injection molding process. Compared with the existed methods, the PDA-LSTM method has the advantages of higher prediction accuracy and the ability to interpret the key input variable and time points of the process data.
To overcome the shortcomings of existing magnetic levitation density measurement methods, such as a narrow measurement range (0.8~2.3 g/cm³) and cumbersome operation when expanding their measurement range, a novel density measurement method for high-density diamagnetic substances is proposed. In this method, the high-density substance is placed on an adjustable inclined plate, and both are positioned in a horizontally placed magnetic levitation device. By adjusting the angle of the inclined plate, the sample can achieve levitation in the device under the combined action of gravity, buoyancy, normal force, and magnetic force. Based on the analysis of the forces on the substance in the paramagnetic solution, the distance between the substance and the center of the left magnet surface as well as the angle of the inclined plate are measured, and a correlation function among the distance, angle, and density is established. Accurate measurement of the density of high-density substances is achieved based on this functional relationship. Meanwhile, this function provides three measurement modes, namely linear, non-linear, and arbitrary-position methods, to cope with different density measurement scenarios, and finally a comprehensive surface graph of measurable density is obtained. Experimental results show that this method can measure the density of small-volume high-density samples (such as copper, tin, zirconia ceramic, aluminum, and glass). It achieves an expanded measurement range and more straightforward operation.