A framework for enhancing industrial soft sensor learning models

Mattos, João Guilherme; Happ, Patrick Nigri; Fernandes, William; Lopes, Hélio; Barbosa, Simone Diniz Junqueira; Kalinowski, Marcos; Rosa, L. S. da; Novello, Cassia; Ribeiro, Leonardo D.; Ventura, Patricia Rodrigues; Marques, Marcelo Cardoso; Pitta, Renato Neves; Camolesi, Valmir J.; Costa, Luís; Paravidino, Bruno Itagyba; Pereira, Cristiane Salgado

doi:10.1016/j.dche.2023.100112

Cited by 1 publication

(1 citation statement)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The distribution pattern of multivariate time series data such as finance, medical treatment and weather is clear and the fluctuation range is large, which is easy to fill the missing value [10][11][12]. The parameters of industrial systems are relatively stable under normal operating conditions and fluctuate for a long time within a small range, which means that sensor data has a unique distribution [13]. Therefore, when dealing with missing data in industrial systems, it is not only necessary to consider whether the filling values are reasonable, but also whether they conform to the distribution of the original data [14,15].…”

Section: Introductionmentioning

confidence: 99%

Missing data filling in soft sensing using denoising diffusion probability model

Jiang,

Wang,

Shen

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

With the aim of addressing the problem of degradation in soft measurement accuracy due to missing data in industrial processes, a filling method based on the denoising diffusion probability model (DDPM) is proposed here to improve the accuracy of soft measurement modeling. First, missing regions are detected with the help of an improved Isolation Forest algorithm to obtain information such as the locations and numbers of missing data regions. Next, a data generation model is constructed based on DDPM and new samples are obtained. By adjusting the threshold for normal operation of the system and the weight sampler, filler samples that are similar to the distribution of the original data can be filtered from the new samples to form a complete dataset. The feasibility of the proposed missing data filling method is explored through numerical simulations, and its superiority in terms of improving the prediction accuracy of soft measurements is verified in regard to the nickel flash smelting process.

show abstract