Data Collection for RSM
In general, the data feeding into the model are initially organized in the RSM based on their associated plant, workshops, or production sections. They are classified according to various defect sources: Raw Material-Related Defects, Utilities-Related Defects, Batch-Plant Related Defects, Melting-Related Defects, Forming-Related Defects (Tin Bath), Annealing-Related Defects, Cold End-Related Defects, and Handling-Related Defects. This classification allows us to pinpoint the root causes of defects by tracing product issues from their manifestation in the glass back to their origins in the production process. This tracking is supported by four interconnected areas of relation, enabling the model to trace defects to their sources through all stages, from design to operation.

• Design Stage: Providing design criteria, identifying bottlenecks, and outlining limits and capabilities can help refine the model, turning it into a more developed tool.
• Construction Stage: During the assembling, heat-up, and startup stages, real-time information about challenges such as damage to refractory materials or other structural components of the production line is crucial. By incorporating data on potential errors and the expected outcomes of these errors, the model can be further refined. This enhancement allows for deeper analysis and improves its predictive capabilities.
• Commissioning: Commissioning in the context of a glass project involves verifying that all systems and components of a production line operate correctly according to operational requirements. This crucial process includes heating up the furnace, forming systems, and annealing lehr, which significantly impacts the entire campaign life of the production line, including potential defects and incidents during production. Accurately recording this process and incorporating the data into the model is essential for future analysis and optimization of the model.
• Operation Stage: This stage is the primary source of data that feeds into the model, encompassing the following categorized data:
  • Primary Process Parameters: The model can capture primary or main process parameters either through continuous measurements from sensors or from Excel files created in the control rooms of each section. It calculates deviations and detects non-conformities by analyzing these data. The model must consider all relevant parameters for precise analysis. Accurate and comprehensive measurements of these contributing process parameters are also essential for achieving the goals set for the RSM, as it is crucial not to overlook any parameters unless they are obstructed or concealed in the process.
  • Defect Information: Typically, the RSM can identify the type of product defect and its recurring patterns (such as defect density, appearance time intervals, and defect locations on the glass ribbon or article). This information is obtained either from automatic scanners or through data manually entered into the model by the QC team and lab technicians using specific forms generated by the model.
  • Process Incidents: RSM can obtain data on process incidents and their historical development through inputs from production operators, technicians, or inspection companies. These inputs are entered into the model using specific forms that the model generates.
  • Standards and Instructions: Quality standards and operating instructions, including upper and lower limits of process parameters and limits on utilized resources (materials, energy, operators), are crucial for the model. They enable it to identify non-conformities, calculate deviations, and analyze data and potential outcomes. Additionally, these materials are instrumental in training within the model, providing easy access for operators during regular operations or emergency situations.
  • Maintenance of Machinery and Sensors: Proper maintenance of machinery, sensors, and control systems is crucial for ensuring reliable, stable, and consistent production with minimal hardware malfunctions. By incorporating the maintenance schedule, service life of sensors, and repair activity reports into the model, it becomes possible to analyze deviations and identify the root causes of incidents effectively.
  • Production Line Structural Damage or Deficiencies: Chronic structural changes and potential deficiencies across various sections of the production line are critical elements in data analysis. This includes potential damage to batching plant key devices, furnace, tin bath,  and cutting system. These changes need to be incorporated into the model to enhance its effectiveness in analysis.