Effectiveness of control of the float glass production process

Abstract

The results of simulation modeling of the control of the float glass production process are presented. It is shown that the systems approach to control is effective, making it possible to solve holistically the problem of economical use of raw materials and energy resources while ensuring high glass quality and limiting adverse environmental impacts. The main problem in glass production is to produce high-quality glass while limiting any adverse environmental impacts of the production process. Product quality is formed at all stages of the life cycle; the required quality must be maintained at all stages of the production process. The quality requirements are set by the international standards ISO Ser. 9000. As indicated in the standards ISO Ser. 14,000 attainment of the target environmental indicators can be evaluated by the amount of raw materials used or the energy consumed and the amount of the contaminants NOx , SO 2 , CO, HC, PB and CFC s produced. Any improvements to the production process must be continually focused on control and reducing the variability of product characteristics and the parameters of the production process. These problems can be solved by integrated management systems used a glass plants. In working out corrective actions a decision maker (DM) is guided by the principles of decision making based on analysis of data and information acquired during production. Decision support systems (DSS) are used for real-time data processing. The DM can use the ‘suggestions’ supplied by the DSS or decline them and use a solution based on personal experience in controlling the process. The effectiveness of the corrective actions implemented depends on the likelihood of the ‘suggestions’ being adopted. In this connection there arises the problem of evaluating the effectiveness of the man-machine system as a function of the likelihood of the suggestions generated by the DSS being implemented. We shall make use of the concept of repeating the same test under constant conditions [1]. As elementary outcomes of each individual test we shall distinguish two outcomes: the first event is that the DM takes the ‘suggestion’ generated by the DSS and the second is that the DM declines the DSS ‘suggestion’ and adopts a decision based on personal experience in controlling the process. We shall use a binomial distribution with the probability P of the appearance of the first event to model the suggestions adopted by the DM for correcting the controlled process. We shall now proceed to evaluating the effectiveness of control of the float glass production process at the AGC Borskii Glass Plant, JSC. The operating regimes of the stages of the sheet-glass production lines were picked as the control actions. Since the control actions affect the quality of the glass and the intensity of the environmental impact of the production process simultaneously, this required a systems approach to generating control actions [2]. The problem of controlling the glass production process was formulated in terms of minimizing the losses of energy, raw materials and materials resources and limiting adverse environmental impacts in the production of a planned amount of high-quality glass. The effectiveness of the control problem formulated and of the criterion chosen was evaluated by means of simulation modeling, which was performed using average daily data obtained during the production of polished glass at AGC Borskii Glass Plant, JSC. The sample used for the modeling was representative. The information collected over a period of one year (365 points) makes it possible to fully reflect all aspects of the functioning of two process lines used in the production of sheet glass, take account of the changes in the