To address the challenge of assessing carbon emissions in the iron ore sintering process, characterized by dynamic complexity, strong temporal correlations, and multi-stage coupling, this study innovatively introduces the Hidden Markov Model (HMM) into the field of carbon emission analysis. We propose a method for process stage identification and carbon emission modeling based on a Gaussian Hidden Markov Model (Gaussian HMM). The model defines the four stages of the sintering process as hidden states and uses six-dimensional flue gas monitoring data (temperature, SO2 concentration, NO concentration, NOx concentration, O2 content, CO concentration) as the observation sequence, with Gaussian distributions describing the emission characteristics of each stage. The research employs the Random Forest algorithm to impute missing values and correct outliers in the raw data, followed by standardization to eliminate scale differences. Model parameters are initialized using Maximum Likelihood Estimation (MLE) and iteratively optimized via the Forward-Backward and Baum-Welch algorithms to enhance the model's fitting capability for complex temporal data. The Viterbi algorithm dynamically decodes the hidden state sequence, enabling an online "predict-until-cooling" monitoring strategy. This strategy accurately determines the optimal cooling timing to balance combustion efficiency with the reaction endpoint. This approach prevents incomplete iron ore combustion and low raw material utilization, while simultaneously reducing emissions of harmful gases and greenhouse gases, thereby achieving the goal of lowering carbon emissions. It provides technical support for the refined management of carbon emissions.
Read full abstract- All Solutions
Editage
One platform for all researcher needs
Paperpal
AI-powered academic writing assistant
R Discovery
Your #1 AI companion for literature search
Mind the Graph
AI tool for graphics, illustrations, and artwork
Journal finder
AI-powered journal recommender
Unlock unlimited use of all AI tools with the Editage Plus membership.
Explore Editage Plus - Support
Overview
571 Articles
Published in last 50 years
Articles published on Utilization Of Raw Materials
Authors
Select Authors
Journals
Select Journals
Duration
Select Duration
545 Search results
Sort by Recency