The endpoint carbon content is an important target of converters. The precise prediction of carbon content is the key to endpoint control in converter steelmaking. In this study, a real-time dynamic prediction of the carbon content model for the second-blowing stage of the converter steelmaking process was proposed. First, a case-based reasoning (CBR) algorithm was used to retrieve similar historical cases and their corresponding process parameters in the second blowing stage, based on the process parameters of the new case in the main blowing stage. Next, a long short-term memory (LSTM) model was trained by using process parameters of similar cases from the previous moment as the input and the carbon content for the next moment as the output. Finally, the process parameters of the new case were input into the trained LSTM model to produce a real-time dynamic prediction of the carbon content in the second blowing stage. Actual production data were used for the verification, and the results showed that the prediction errors of the proposed model within the ranges of (−0.005, 0.005), (−0.010, 0.010), (−0.015, 0.015) and (−0.020, 0.020) were 25%, 54%, 71%, and 91% respectively, which were higher than the prediction accuracies of the traditional carbon integral model, cubic model, and exponential model.

The stability of mold level fluctuations (MLFs) is crucial for product quality and process efficiency in continuous casting. Abnormal mold level fluctuations, which are typically associated with multiple factors including stopper rod opening, casting speed, and mold width, are known to lead to slab quality defects. In this paper, an Informer-based prediction framework is proposed for the early detection of abnormal MLF. A threshold-based labeling method is developed to quantify the future likelihood and severity of anomalies across different time horizons. Considering the importance of frequency-domain features in mold level prediction, power spectral density (PSD) features are incorporated and smoothed using the exponential moving average (EMA) to enhance predictive performance. Through the integration of temporal and processed spectral features, early indicators of abnormality can be captured, and proactive warnings can be issued. The proposed architecture is validated using approximately 32.5 million data points from a real-world continuous casting process. This approach provides a robust and data-driven solution for predicting and diagnosing abnormal MLF events in continuous casting. Experimental results show that the mean ROC-AUC and PR-AUC reach 0.821 and 0.418, respectively.

During slab continuous casting, mold level fluctuation (MLF) affects the slab quality. MLF is affected by multiple process parameters and multi-scale dynamic characteristics inherent in each parameter. The forecasting of MLF generally adopts time series forecasting models with MAE and MSE used as loss functions. However, MAE and MSE only calculate numerical differences and may cause temporal misalignment between predicted sequences and real sequences. An LSTM-based Seq2Seq multi-step forecasting model for MLF, named LSTM-Seq2Seq-WaveAtt-DILATE, is proposed. Wavelet decomposition is introduced to enhance the mode's ability to capture cross-scale relationships among input parameters. DILATE (DIstortion Loss including shApe and TimE) loss is introduced to explicitly penalize shape distortion and temporal misalignment between predicted and real sequences and ensure accurate shape representation and precise temporal localization. Firstly, input data are decomposed and reconstructed into low-frequency, medium-frequency, and high-frequency components via wavelet decomposition. Each frequency component and the original data are processed using multi-head attention separately. Then, the context vector output by the encoder is passed via Bahdanau attention to the decoder to perform multi-step forecasting. Finally, the model is trained with a combined loss function of MSE and DILATE to enhance forecasting accuracy and address temporal misalignment.The experimental results compared with the LSTM-Seq2Seq benchmark model trained only with MSE loss show that, the proposed method achieves significant performance improvements in MLF forecasting: MAE decreased from 0.517 to 0.150, MSE decreased from 1.426 to 0.118, DTW decreased from 42.11 to 19.00, and TDI decreased from 2.096 to 0.009.