Search
Use the search bar or filters below to find any TAPPI product or publication.
Filters
Content Type
Publications
Level of Knowledge
Committees
Collections
Journal articles
Data driven modeling to reduce fossil fuel consumption in a lime kiln integrated with biomass gasifier, TAPPI Journal September 2026
ABSTRACT: Biomass gasification, although already known and applied, is currently being consolidated as a sustainable alternative in the pulp industry, contributing to the reduction of fossil carbon dioxide (CO2) emissions and to the utilization of forest residues. In this context, pulp mills are beginning to adopt biomass gasification within their chemical recovery cycles by integrating the technology with lime kilns. However, further process studies are still needed to support and optimize this application. Thus, the present study aimed to analyze lime mud feed temperature by applying an artificial neural network model to a gasification system integrated with a lime kiln. For this evaluation, 10 periods of system stability throughout 2024 were selected under different operational conditions. Data from 28 potential predictive variables were collected, a total of 348 observations. This dataset was then processed using R software, where data treatment and model dimensionality reduction were performed, resulting in 10 predictive variables between gasifier and kiln. Next, the dataset was randomly divided into training (70% of the observations) and testing data. The established neural network model (using “neuralnet” package) was optimized, resulting in a configuration containing one hidden layer with three neurons. This setup enabled optimal estimation of the flue gas kiln outlet temperature, with mean absolute error (MAE) = 3.0°C and root mean square error (RMSE) = 4.1°C, applying the resilient backpropagation algorithm with backtracking — both below the thermocouple’s measurement error (±5.0°C) for the evaluated average temperature range (668.2°C). When the same dataset was modeled using the resilient backpropagation algorithm without backtracking, even better results were achieved: MAE = 2.5°C and RMSE = 3.7°C. Thus, after evaluating these and other configurations, it was concluded that the best model required 10 predictive variables and the backpropagation algorithm without backtracking to determine flue gas kiln outlet temperature. These results provide a better understanding of how gasifier and kiln variables influence the temperature in the lime kiln, which is essential to improve control and optimize the calcination process, avoiding supplementary fossil fuel consumption.
Journal articles
Magazine articles
Application of ATR-IR measurements to predict the deinking efficiency of UV-cured inks, TAPPI Journal January 2022
ABSTRACT: In recent years, ultraviolet (UV)-curable ink has been developed and widely used in various printing applications. However, using UV-printed products (UV prints) in recovered paper recycling causes end-product dirt specks and quality issues. A new method was developed that can distinguish UV prints from other prints by means of attenuated total reflectance infrared (ATR-IR) spectroscopy. Application of this method could allow more efficient use of UV prints as raw materials for paper recycling.First, a mill trial was performed using UV prints alone as raw materials in a deinked pulp (DIP) process. Second, test prints were made with four types of UV inks: a conventional UV ink and three different highly-sensitive UV inks. Each print sample had four levels of four-color ink coverage patterns (100%, 75%, 50%, and 25%). Next, deinkability of all prints was evaluated by laboratory experiments. Finally, each print was measured using the ATR-IR method, and the relationship between the IR spectra and deinkability was investigated. Mill trial results showed that UV prints caused more than 20 times as many dirt specks as those printed with conventional oil-based ink. There were variations in recycling performance among UV prints taken from bales used for the mill trial. Lab tests clearly revealed that not all UV-printed products lead to dirt specks. In order to clarify the factors that affected deinkability of UV prints, the print samples were investigated by lab experiments. Key findings from lab experiments include: • The number of dirt specks larger than 250 µm in diameter increased as the ink coverage increased. • Higher ink coverage area showed stronger intensity of ATR-IR spectral bands associated with inks. These results indicate that deinkability of UV prints could be predicted by analysis of ATR-IR spectra. • Finally, the method was applied for assessment of recovered paper from commercial printing presses. It was confirmed that this method made it possible to distinguish easily deinkable UV prints from other UV prints. Based on these findings, we concluded that the ATR-IR method is applicable for inspection of incoming recovered paper.