In the operation process of wet ball mill, there are often multi-modal and multi-condition problems. In this paper, a multi-view based domain adaptive extreme learning machine (MVDAELM) was used to measure the mill load. Firstly, the correlation relationship between the load parameters and the two views (vibration and acoustic signals of the ball mill) was obtained by Canonical Correlation Analysis (CCA) respectively. Secondly, a small number of labeled data from the target domain were introduced to construct a Domain Adaptation Extreme Learning Machine (DAELM) model under manifold constraints, which solve the mismatch problem caused by the change of working conditions in the multi-condition grinding process. Finally, based on the correlation coefficient obtained before, the two views domain adaptive load parameter soft sensor model was integrated to solve the uncertainty problem in single-modal data modeling. The experimental results show that the proposed method can effectively improve the learning accuracy of the soft sensor model under multi-modal conditions.

Multi-layer Artificial Neural Networks (ANN) has caught widespread attention as a new method for time series forecasting due to the ability of approximating any nonlinear function. In this paper, a new local time series prediction model is established with the nearest neighbor domain theory, in which the hybrid Euclidean distance is used as the similarity measurement between two sets of time series. In order to improve the efficiency, prediction performance, as well as the ability of real-time updating of the model, in this paper, the recombination samples of the model is derived by Deep Extreme Learning Machine (DELM). The experiments show that local prediction model gets accurate results in one-step and multi-step forecasting, and the model has good generalization performance through the test on the five data sets selected from Time Series Database Library (TSDL).