A mathematical programming model for optimal layout considering quantitative risk analysis
نویسندگان
چکیده
Safety and performance are important factors in the design and operation of chemical plants. This paper describes the formulation of a mixed integer nonlinear programming model for the optimization of plant layout with safety considerations. The model considers a quantitative risk analysis to take safety into account, and a bowtie analysis is used to identify possible catastrophic outcomes. These effects are quantified through consequence analyses and probit models. The model allows the location of facilities at any available point, an advantage over grid-based models. Two case studies are solved to show the applicability of the proposed approach. Introduction Chemical plants must not only be cost effective, but also avoid or minimize the risk of major hazards, which places safety as one of the major components in the operation of chemical plants. History supports this fact. The Texas City refinery explosion in 2005 and the Flixborough disaster in 1974, among others, are examples of lack of safety in chemical plants due to poor layouts and back-up systems. Facility siting and layout is an important item in risk management and safety (Crowl & Louvar, 2002). A good facility siting and a proper layout contribute to an inherently safer plant and better risk management, and may even reduce occupied land and operation costs (Patsiatzis et al., 2004). The Center for Chemical Process Safety (CCPS) has published guidelines for facility siting and layout (AIChE, 2003). The CCPS guidelines, based on industry practice and standards, provide guidance for finding an optimal production site and for proper placing of units within the plant. However, the guidelines do not provide a systematic method for plant layout. Mathematical programming has been applied to model layout problems. Georgadis et al. (1999) have proposed a general mathematical programming approach for plant layout under restrictions of fixed safety distances. Penteado and Ciric (1996) developed an MINLP model for safe process layout considering three possible hazardous incidents in an ethylene oxide plant. Addition of safety devices to decrease consequences in case of an incident was also taken into account. Vazquez-Roman et al. (2010) proposed an MINLP model that considers atmospheric uncertainties under toxic releases using Monte Carlo simulation. Jung et al. (2010a) developed a systematic approach for facility layout considering fire and explosion scenarios using a gridbased MILP model. In a second work, Jung et al. (2010b) reported a MINLP model for facility layout considering toxic releases using CFD software to validate the results. These works have particularly contributed to a better understanding and modeling of the relationship between layout and safety. However, most of them have focused on worst-case scenarios, which give only a partial view of the entire spectrum of risk sources, typically overestimating risk. This work aims at providing a more elaborate analysis of risk sources by considering a complete quantitative risk analysis (QRA). A QRA identifies common scenarios and quantifies their corresponding risk. In this way, a QRA finds possible outcomes, among them the most frequent one, and the scenario with highest consequences. We propose a mathematical model that yields a systematic algorithm for plant layout following CCPS guidelines for facility siting and layout. The proposed model requires a bowtie analysis, which identifies potential catastrophic outcomes given a failure within dangerous process equipment. Once the outcomes are identified, an MINLP model is formulated to find the optimal location for process units and equipment. The objective function considers risk of workers at the units, risk of damage to process equipment, and land and interconnection costs. The objective function is subject to geometry constraints, non-overlapping constraints, scenarios characterization constraints, and consequences quantification constraints considering economic data and wind direction uncertainty. The outline of the paper is as follows. First, we introduce basic concepts and common risk management procedures. Next, we present the problem statement and state some assumptions for the formulation of the MINLP model. The general formulation of the model is explained and relevant constraints such as geometrical relations, disjunctions for nonoverlapping, frequency analysis, consequences analysis, the objective function, and the reformulation of the disjunctions are addressed in detail. Two examples are then used to show the application of the proposed model.
منابع مشابه
A Continuous Plane Model to Machine Layout Problems Considering Pick-Up and Drop-Off Points: An Evolutionary Algorithm
One of the well-known evolutionary algorithms inspired by biological evolution is genetic algorithm (GA) that is employed as a robust and global optimization tool to search for the best or near-optimal solution with the search space. In this paper, this algorithm is used to solve unequalsized machines (or intra-cell) layout problems considering pick-up and drop-off (input/output) points. Such p...
متن کاملSolving a new mathematical model for cellular manufacturing system: A fuzzy goal programming approach
A fuzzy goal programming-based approach is used to solve a proposed multi-objective linear programming model and simultaneously handle two important problems in cellular manufacturing systems, viz. cell formation and layout design. Considerations of intra-cell layout, the intra-cell material handling can be calculated exactly. The advantages of the proposed model are considering machining cos...
متن کاملA mathematical model for the design of distributed layout by considering production planning and system reconfiguration over multiple time periods
In this paper, we develop a new mathematical model that integrates layout configuration and production planning in the design of dynamic distributed layouts. The model incorporates a number of important manufacturing attributes such as demand fluctuation, system reconfiguration, lot splitting, work load balancing, alternative routings, machine capability and tooling requirements. In addition, t...
متن کاملMathematical Programming Models for Solving Unequal-Sized Facilities Layout Problems - a Generic Search Method
This paper present unequal-sized facilities layout solutions generated by a genetic search program named LADEGA (Layout Design using a Genetic Algorithm). The generalized quadratic assignment problem requiring pre-determined distance and material flow matrices as the input data and the continuous plane model employing a dynamic distance measure and a material flow matrix are discussed. Computa...
متن کاملAn optimal model for Project Risk Response Portfolio Selection (P2RPS) (Case study: Research institute of petroleum industry)
In the real world, risk and uncertainty are two natural properties in the implementation of Mega projects. Most projects fail to achieve the pre-determined objectives due to uncertainty. A linear integer programming optimization model was used in this work to solve a problem in order to choose the most appropriate risk responses for the project risks. A mathematical model, in which work structu...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- Computers & Chemical Engineering
دوره 68 شماره
صفحات -
تاریخ انتشار 2014