• Case study
• Energy balance
• Mass balance
• Predicitve model approach
• Stock preparation
• Tissue paper

The goal of this research is to streamline the paper manufacturing stock preparation process and cut down on energy costs. In the paper machine system, the refiner unit takes center stage as a critical energy-consuming component. This study uses a detailed mass and energy balance analysis to identify inefficiencies and measure energy inputs, outputs, and losses during the stock preparation stage.
Mass balance analysis not only helps us understand material flows in detail but also acts as a Key Performance Indicator (KPI) to track and enhance process efficiency. Operational teams can utilize this KPI to monitor performance and pinpoint areas for additional enhancement. Data for this study were obtained from a paper mill which is Sofidel Cartiera, Italy, and experimental samples were produced in the plant’s laboratory. The combination of real-world and experimental data provides a robust basis for analysis and model development.
The objective is to develop an initial predictive model that takes into account the complex and non-linear refining process, despite the limited data available. The model aims to identify connections between important operational factors, such as motor load, refining intensity, pulp consistency, and throughput. These connections can then be used to estimate energy usage and suggest process optimizations.
Machine learning plays a pivotal role in this research by enabling the development of predictive models that can analyze large datasets and uncover hidden patterns and relationships. Through the application of sophisticated algorithms, machine learning can process complex, non-linear data to generate accurate predictions and actionable insights. This approach allows for continuous learning and improvement as more data becomes available, ultimately aiding in optimizing the energy efficiency of the refiner unit and other components of the stock preparation process. Furthermore, machine learning can facilitate real-time monitoring and adaptive control, thereby enhancing operational efficiency and reducing energy consumption.
The experimental samples, which served as the dataset for the model, were collected manually, introducing some errors and inconsistencies. Given the complexity and non-linear nature of the refining process, as well as the limited sample size, the model provides initial insights into the factors influencing energy use in the refiner unit. Despite these challenges and the presence of some errors, the predictive model offers a valuable approach for identifying optimal process parameters that can reduce energy consumption. It serves as a foundation for future refinement and validation with larger, more comprehensive datasets, and more advanced machine learning techniques.
The findings from this study suggest several potential strategies for optimizing energy use in stock preparation. These include adjusting process parameters, such as refining intensity and motor load, to more energy-efficient settings, and upgrading equipment to more modern, energy-efficient designs. Implementing these strategies can lead to significant energy cost reductions, thereby enhancing the sustainability and profitability of the paper manufacturing process. In addition, these strategies can contribute to reducing the environmental impact of paper production, aligning with broader sustainability goals and regulations.
This research contributes to the broader field of industrial energy management by presenting an exploratory approach to energy analysis and predictive modeling in stock preparation. The methodologies and insights developed can be further refined and applied to similar industrial processes, promoting greater energy efficiency and cost savings across various sectors. Future work should focus on refining the predictive model by incorporating a larger and more diverse dataset to improve its accuracy. Additionally, experimental validation under various operating conditions should be conducted to further validate and enhance the model's reliability. Continued research should also explore the integration of advanced data collection methods to reduce errors and inconsistencies, thereby providing more robust and actionable insights for optimizing energy use in the stock preparation process. The methodologies and insights developed in this research can serve as a foundation for broader applications in similar industrial processes, promoting greater energy efficiency and cost savings across various sectors.