Custom Industrial Research and Development

SÅGA redox measurement system (project name SOX-GLASS)

 

System for non-contact in situ measurement of the glass melt properties.

A system for non-contact, in-situ measurement of the properties of molten glass, enabling a rapid response to changes in the process and active control of the melting furnace. The purpose of the project is to develop the first long-term stable and reliable system for direct measurement of the oxidative state of molten glass during normal production, allowing manufacturers to achieve savings in the consumption of primary raw materials and energy.

The investment is co-financed by the Republic of Slovenia and the EU through the European Regional Development Fund.
https://www.eu-skladi.si

Implementation of Circular Economy

 

Companies RC eNeM d.o.o. and Culmium d.o.o. can become your partners in achieving circular economy goals.

We offer a comprehensive consulting and implementation service regarding the possibilities of reusing resources. Our approach is based on the DMAIC approach:

• We delve into the production processes within the company.
• We review historical data, including import, cleaning, and merging, examining correlations.
• We conduct an analysis of measurement systems.
• We plan experiments for structured and systematic data collection, enhancing the understanding of processes.
• We analyze data using modern mathematical modeling approaches.
• We use models for predicting changes in the process or assessing the impact of individual process parameters.
• Based on predictions, we implement changes in the process with a defined level of risk.

Case Study: Collaboration with Steklarna Hrastnik

Glass manufacturer Steklarna Hrastnik faced increasing costs for disposing of hazardous waste – dust from the flue gas cleaning system. This dust contained a significant amount of selenium, which plays a role as a decolorizer in glass – it is added to achieve perfectly transparent, so-called extra-flint glass. They were interested in whether they could use the filter dust as a source of selenium, either returning it to the production process or recycling it.
To answer this question, we had to examine the entire composition of the dust – from the main components to trace elements. Additionally, it was crucial to determine whether and how the selenium content in the dust changed concerning the process parameters of the glass melting furnace. We found that the selenium content varied to such an extent that continuous monitoring of selenium concentration in the dust or the use of reliable predictive models for selenium concentration would be necessary to ensure consistent product quality.
For developing a mathematical model, we first analyzed historical furnace operation data. We extracted parameters relevant to selenium vaporization and binding to filter dust and determined the sampling interval for dust. In complex processes with many parameters of different types and distributions, we employed modern mathematical approaches and selected predictive models with greater predictive power on validation data. Since we were limited by the cost of selenium concentration analysis (ICP-EOS method), we looked for a cheaper alternative. It turned out that with proper calibration, the manual XRF method was suitable. Our experts assisted Steklarna Hrastnik employees in establishing a reliable measurement method, making monitoring easier and enabling the collection of more data for modeling.
By combining the use of predictive models and continuous measurements of selenium dioxide concentration in the flue gas cleaning system dust with manual XRF, Steklarna Hrastnik successfully returned 100% of the waste SFP in a pilot attempt, thereby replacing 60% of the primary raw material for glass decolorization.

 

Case Study: Collaboration with Eti Elektroelement

Eti Elektroelement faces the issue of ceramic mass waste during the batch-continuous process of manufacturing high-power network fuses. The final stage of the process involves extruding the mass, but sometimes the mass does not meet the required quality for product manufacturing. Some intermediate production steps are highly energy-intensive. If we could optimize and/or predict the quality of the ceramic mass at the extruder by manipulating process parameters at different sequential stages of the production process (mills, pool, drying-spraying tower), we could reduce waste.
We focused on the EN-6 mass, which represents one of the company’s key products. To connect different process sections and methods of data capture and collection in individual sections, we thoroughly studied the process through process diagrams, control systems, and discussions with process and development technologists. Based on historical data, we sought connections between different sections and causal relationships between phases, especially due to the specificity of the process (a mixture of batch and continuous production).
In mathematical modeling, we encountered many challenges, mainly due to the complexity of the process and the traceability of mass between different sections of the process. Mathematical-statistical approaches such as time interpolation of curves, decision tree modeling, and logistic regression proved to be more effective than classical modeling methods in this case.
Despite a limited set of high-quality data, we managed to determine mass quality with the model down to one class accurately. Since mass quality is a subjective assessment by the operator, such information can be very useful in detecting extremes (e.g., very poor mass) in the early stages of the production process. With this information, the mass can be manipulated appropriately to prevent later waste, and earlier rejection of poor quality mass can save energy required for processing it in later stages.