Exergy-related process monitoring for hot strip mill process based on improved support tensor data description

Abstract

Process monitoring is important for ensuring industrial production safety. If faults are detected in time, maintenance plan will be made to avoid economic losses. Traditional process monitoring methods pay more attention to the utilization of process data and ignore process mechanism. It is necessary to consider the energy flow and the spatial information of different production equipments. As the unity of quality and quantity of energy, exergy contains the performance change information of the process and can be used as another way of achieving the required dimensionality reduction. Moreover, the introduction of spatial information will lead to the increase of data dimension. Support vector data description (SVDD) are oriented to vector data and cannot deal with tensor data directly. To handle above issues, a novel exergy-related process monitoring method based on improved support tensor data description (ISTDD) is proposed in this paper. First, exergy efficiency are calculated and exergy-related process variables are obtained by the minimal redundancy maximal relevance (mRMR). Second, a third-order tensor is constructed with spatial information. Then, the exergy-related monitoring model and its robust version are developed. Finally, case study on a hot strip mill process (HSMP) is given to illustrate the effectiveness of proposed method.