Data-Driven Prediction of Sinter Composition Based on Multi-Source Information and LSTM Network

A Survey of Data-Driven Soft Sensing in Ironmaking System: Research Status and Opportunities

“…In order to combine multiple data, a multisource information fusion scheme was creatively proposed to achieve the FeO content prediction based on infrared thermal image data of sinter cross section and process data. The prediction results show that the extracted multisource features reach a good performance and meet the needs of practical engineering. , …”

“…The prediction results show that the extracted multisource features reach a good performance and meet the needs of practical engineering. 60,61 Apart from chemical compositions, TS, as a critical metallurgical and intrinsic property, is also essential to the wear resistance and anticollision performance of the sintered ore. 62 In general, high strength sinter ore helps to reduce the industry dust output dosage and improve efficiency of the blast furnace, while too low strength can affect the permeability of material surface. 63 Under this background, Ye et al developed a TS estimation method based on LSSVM and local thermal nonequilibrium (LTNE) model, and the proposed scheme was verified in the sinter pot tests.…”

“…According to the engineering practice experience, the desired position of BTP is approximately located in the 58 Integrated model combining multisource information fusion and LSTM network for FeO content prediction. Bai et al 60 Combination model of adaptive particle swarm optimization (APSO) algorithm and least-squares support vector machine (LSSVM) algorithm. Jiang et al 61 Polymorphic measurement method of FeO content of sinter based on heterogeneous features of infrared thermal images.…”

Data‐driven modelling methods in sintering process: Current research status and perspectives

“…Bai et al [29] Integrated model combining multi-source information fusion and LSTM network for FeO content prediction…”

“…For the prediction of sinter ore composition, there has been a lot of representative research work. For example, Bai et al [29] proposed an integrated model combining multi-source information fusion and a long short-term memory (LSTM) network to predict the FeO content in sinter ore. In addition, a combination model of an adaptive particle swarm optimization (APSO) algorithm and a least square support vector machine (LS-SVM) algorithm was proposed to predict the FeO content of sinter ore. [52] Li et al [30] developed a novel dynamic time feature expanding and extracting framework for the FeO content prediction in the sintering process.…”