Evaluation strategy and mass balance for making decision about the amount of aluminum fluoride addition based on superheat degree

Abstract

The purpose of aluminum fluoride (AlF 3) addition is to adjust the superheat degree (SD) in the aluminum reduction process. Determining the appropriate amount of AlF 3 to add has long been a challenging industrial issue as a result of its inherent complexity. Because of the decreasing number of experienced technicians, the manual addition of AlF 3 is usually inexact, which easily leads to an unstable cell condition. In this paper, an evaluation strategy based on the SD for AlF 3 addition is proposed. An extended naïve Bayesian classifier (ENBC) is designed to estimate the states of SD and its trends that represent the current and potential cell condition respectively, and then the process is graded by evaluating the estimated results based on fuzzy theory. The reduction process is divided into a few situations based on the evaluation grades, and mass balance is introduced to determine the amount of AlF 3 addition in each situation. The results of experiments show that the proposed strategy is feasible, and the effectiveness of AlF 3 addition is improved compared to the existing method. Moreover, automatic AlF 3 addition is promising based on the proposed strategy.

Vulnerability assessment for physical protection systems of cave temples: A fuzzy petri net approach

“…Motivated by the above challenges and inspired by recent advances in knowledge representation, reasoning, and fusion techniques, this paper presents an innovative PPS vulnerability assessment method based on an enhanced fuzzy Petri net (FPN), tailored to the specific characteristics of cave temples regarding illegal intrusion [ 4 , 5 ] with the following contributions: A vulnerability assessment indicator system for the PPS of cave temples was established through research on cave temples and the experience of relevant experts. These indicators are more in line with the topography and architectural characteristics of the cave temple and can describe its security system more accurately.…”

A self-supervised temporal temperature prediction method based on dilated contrastive learning

“…With the advent of artificial intelligence and information technologies, the productivity of the modern industrial system increases continuously [9,10]. In aluminum electrolysis production, superheat degree is an important index for the indicator and monitoring of the whole production situation, lengthening the lifespan of the reduction cell, and improving the current efficiency [15,16,17,38]. However, cell temperature detection suffers many limitations such as the hard environment, expensive device cost, and large delay.…”