Advanced modeling of HPGR power consumption based on operational parameters by BNN: A “Conscious-Lab” development

• Published in: Powder technology Vol. 381; pp. 280 - 284
• Main Authors: Tohry, A., Yazdani, S., Hadavandi, E., Mahmudzadeh, E., Chelgani, S. Chehreh
• Format: Journal Article
• Language: English
• Published: Lausanne Elsevier B.V 01.03.2021; Elsevier BV
• Subjects: Artificial intelligence; Big data; BNN; Energy consumption; HPGR; Iron ores; Machine learning; Mineral Processing; Mineralteknik; Modelling; Neural networks; Power consumption; Real variables; Energy consumption; HPGR; Big data; BNN; Machine learning
• ISSN: 0032-5910; 1873-328X; 1873-328X
• DOI: 10.1016/j.powtec.2020.12.018

This study, for the first time, is going to introduce the boosted neural network (BNN) as a robust artificial intelligence for filling gaps related to the modeling of energy consumption (power draw) in the industrial scale high-pressure grinding rolls (HPGR). For such a purpose, a new concept called “Conscious Laboratory (CL)” has been developed. CL would be the modeling of variables based on real databases that are collected from the industrial-scale plants. Although using HPGRs have been absorbed attention in many processing plants, a few investigations have been conducted to model the power draw of HPGRs. In this article, BNN was used for modeling relationships between HPGR operational variables, and their representative power draws based on an industrial database. This investigation indicated that the generated CL based on BNN could accurately assess the multivariable relationships between monitoring variables of an HPGR from an iron ore plant. [Display omitted] •Conscious lab (CL) can solve many challenges within industrial powder technology.•Boosted neural network (BNN) can assess linear and nonlinear multivariable relationships.•CL by BNN can quite accurately model power draw of a high-pressure grinding roll.