(104ab) Facility Layout Optimization Detailed Research Through MILP and MINLP

Facility layout optimization detailed research through MILP and MINLP

J. Cadena¹, F. Munoz², J. Gómez³, M. Ramos⁴

Chemical Engineering Department, Universidad de los Andes, Bogotá, Colombia

¹je.cadena21@uniandes.edu.co

²fmunoz@uniandes.edu.co

³jorgomez@uniandes.edu.co

⁴ma.ramos43@uniandes.edu.co

Within Process Safety one can find several stages of plant and process design in which decisions can be made in order to have a significant impact on safety performance during start-up, normal operation and turnaround events. One of these stages is the development and choosing of the distribution or layout of a process unit. Within a process unit one may find buildings, equipment and a central dangerous unit area, and all these elements generate a specific cost of build-up and installation with each other, which makes an issue arise: how can the buildings, equipment and dangerous units be located within a specific terrain minimizing risk and costs related to loss and installation? This is exactly what Jung, S. (et al. 2010) developed and presented on their paper in which a facility layout with a central hazardous unit and a defined number of additional buildings was obtained through the assignation of risk scores for each section of the total terrain and the use of MILP. The risk scores were based on VCE and BLEVE events simulated using PHAST modeling program (v.6.53.1). In the present study the same general structure is maintained, but some solving differences and specific variations on risk scores calculation approach. The first difference is the calculation of risk scores based on VCE event using CFD dispersion/explosion simulations, although the actual calculation of risk score from the impacts simulated remains the same. The second difference, and probably the most notable one is the solving strategy used, which consists on the use of both MILP (previous approach) and MINLP. The latter is an effort to use a more accurate calculation of distance between buildings and to the hazardous unit. The title of this study has a strong relation with what is expected, since it is not thought as an objective to necessarily improve previous results, but to research the possibilities of improvement taking a different approach. Finally the risk scores map used in the base work should be used as input for the solving strategy used in this study to compare the global nature of the optimization’s results. As a general conclusion of the work carried out and to be developed, the optimization of the layout and the research of the possible methodologies that allow obtaining a global optimum, means an important move forward towards the optimal usage of separation distances as passive barriers within a process safety scheme of protection.