Multirate-Former: An Efficient Transformer-Based Hierarchical Network for Multistep Prediction of Multirate Industrial Processes
- 8 November 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-13
Abstract
Due to the limitations of measurement technology and cost in industrial processes, it is difficult to obtain measured values of variables with different properties, such as flow rate and temperature, under uniform sampling rates. This leads to the ubiquitous multirate sampling characteristics of the collected industrial process data, which brings great challenges to the quality prediction of industrial processes. To address this issue, this paper proposes a novel quality prediction modeling method based on the transformer modal called Multirate-Former for multirate industrial processes. First, the raw data is chunked by arranging data variables to the same sampling rate. Then, the hierarchical coarse-grained and fine-grained complementation of the data is completed by the multi-layer convolution network and transformer. Notably, a novel sampling type coding method is proposed to explore the missing pattern of multirate process data. After the above pre-training, the patched completed dataset and better initial weights are provided for the subsequent fine-tuning process. Finally, the multi-step prediction error of quality variables is exploited to fine-tune the entire network parameters. The proposed method is applied to an industrial debutanizer column and an actual industrial hydrocracking process for multirate multi-step prediction. Experimental results demonstrate that the proposed method outperforms other state-of-the-art methods in dealing with multi-sampling rate types of industrial process data.Keywords
Funding Information
- National Natural Science Foundation of China (92267205, 61988101)
- science and technology innovation Program of Hunan Province (2021RC4054, 2021JJ10065, 2020RC3003)
- CAAI-Huawei MindSpore Open Fund (CAAIXSJLJJ-2022-001A)
This publication has 29 references indexed in Scilit:
- Input selection methods for data-driven Soft sensors design: Application to an industrial processInformation Sciences, 2020
- Torus-Event-Based Fault Diagnosis for Stochastic Multirate Time-Varying Systems With Constrained FaultIEEE Transactions on Cybernetics, 2019
- A deep learning based data driven soft sensor for bioprocessesBiochemical Engineering Journal, 2018
- Modeling Long- and Short-Term Temporal Patterns with Deep Neural Networks2018
- Deep Learning of Semisupervised Process Data With Hierarchical Extreme Learning Machine and Soft Sensor ApplicationIEEE Transactions on Industrial Electronics, 2017
- A Novel Soft Sensor for Real-Time Monitoring of the Die Melt Temperature Profile in Polymer ExtrusionIEEE Transactions on Industrial Electronics, 2014
- FIR model identification of multirate processes with random delays using EM algorithmAIChE Journal, 2013
- Data-Driven Soft Sensor Design with Multiple-Rate Sampled Data: A Comparative StudyIndustrial & Engineering Chemistry Research, 2009
- Data-driven Soft Sensors in the process industryComputers & Chemical Engineering, 2009
- Soft sensors for product quality monitoring in debutanizer distillation columnsControl Engineering Practice, 2005