An Approach for Solvent Selection in Extractive Distillation Systems Including Safety Considerations

Nancy Medina-Herrera, Ignacio E. Grossmann, M. Sam Mannan, Arturo Jiménez‐GutiérrezIndustrial 2014 · DOI

TL;DR

This paper introduces a novel methodology for selecting solvents in extractive distillation processes that explicitly incorporates safety considerations from the early design stages. By integrating computer-aided molecular design (CAMD) with a multi-objective optimization framework that balances economic costs with safety risks (quantified by the distance likely to cause death), the approach aims to achieve inherently safer chemical processes.

What Problem Does This Paper Address?

Traditionally, the selection of solvents in chemical processes, particularly for challenging separations like extractive distillation, has primarily focused on performance and economic factors. Safety implications are often addressed as an afterthought, once the process design is largely complete. This reactive approach can lead to suboptimal safety outcomes and potentially costly retrofits. The authors aim to bridge this gap by developing a proactive methodology that embeds safety considerations directly into the solvent selection and process design phases, ensuring that potential hazards are identified and mitigated early on. This is particularly relevant for extractive distillation, where the choice of solvent significantly impacts both separation efficiency and process safety.

Methods

The proposed methodology involves a multi-step approach:

  1. Pre-selection of Solvents: Computer-Aided Molecular Design (CAMD), specifically using the ProCAMD tool within the ICAS software, is employed to screen potential solvents based on desired safety properties (flash point, octanol/water partition coefficient) and performance criteria (boiling point, azeotrope formation). This step identifies candidate solvents that meet predefined safety heuristics.
  2. Consequence Analysis: For the selected candidate solvents, a detailed consequence analysis is performed. This involves modeling potential catastrophic incidents (e.g., BLEVE, UVCE, flash fire, jet fire) and quantifying their impact. The primary safety metric used is the "distance likely to cause death" (DD), derived from established risk analysis guidelines (CCPS).
  3. Multi-Objective Optimization: A multi-objective optimization problem is formulated to find the optimal balance between Total Annual Cost (TAC) and the average distance likely to cause death (DDaver). This optimization is carried out using genetic algorithms, considering design variables such as column stages, feed locations, and solvent flowrates.
  4. Inherently Safer Design: The entire process is geared towards achieving an inherently safer design by applying principles of moderation and minimization, ensuring that the chosen solvent and process configuration minimize risks from the outset.

Datasets

The paper utilizes a database of chemical compounds within the ICAS software for the CAMD pre-selection step. For the consequence analysis, probabilities of catastrophic incidents are taken from the CCPS (Center for Chemical Process Safety) guidelines. No new benchmark datasets are introduced; rather, existing data and models are integrated into the proposed framework.

Key Insights and Ideas

  • Proactive Safety Integration: The core insight is that safety should not be an add-on but an integral part of the early solvent selection and process design, leading to more robust and inherently safer systems.
  • Quantifying Safety with DD: Using the "distance likely to cause death" (DD) as a quantifiable safety metric provides a practical way to incorporate safety into optimization, especially in the early design stages when detailed plant layout is unavailable.
  • Trade-offs are Inevitable: The study highlights that there are inherent trade-offs between economic costs (TAC) and safety (DDaver). The multi-objective optimization framework effectively visualizes these trade-offs through Pareto fronts.
  • Solvent Choice Impacts System Design: The choice of solvent not only affects separation performance but also influences equipment size, operating conditions (e.g., pressure), and consequently, the overall safety profile and economics of the entire distillation sequence.

Key Results

The study applied the methodology to the purification of bioethanol.

  • A pre-selection step identified 282 potential solvents, with 15 being in the ICAS database.
  • Further analysis focused on four solvents: dipropylene glycol, diethylene glycol, ethylene glycol, and dimethylformamide.
  • Dipropylene glycol was identified as a promising inherently safer solvent, offering a favorable compromise between economics and safety, particularly when compared to commonly used entrainers like ethylene glycol. For instance, using dipropylene glycol instead of ethylene glycol showed potential savings of $200,000/yr for similar safety risk levels.
  • The Pareto fronts clearly illustrated the economic and safety trade-offs for each solvent, showing that safer solvents often required higher solvent flowrates and potentially larger equipment, impacting inventory and thus safety principles.
  • Dimethylformamide, while appearing safest in some consequence analyses, raised concerns due to its known toxicity, underscoring the need for comprehensive hazard assessment beyond the scope of the current models.

Limitations and Future Work

  • Limited Solvent Database: The CAMD pre-selection is limited to the solvents available in the ICAS database. The identification of novel, un-synthesized molecules with desirable safety properties is acknowledged as a limitation.
  • Toxicity Modeling: The current consequence analysis did not fully account for toxicity hazards for all solvents, particularly dimethylformamide, due to the lack of specific probit models. The authors suggest developing these models for a more complete safety assessment.
  • Domino Effects: Domino effects (sequential incidents) were not considered in the consequence analysis, which could lead to underestimation of risks in certain scenarios.
  • Future Work: The authors suggest extending this approach to other separation systems that rely on solvents and further developing toxicity models for a more comprehensive safety evaluation. The integration of other inherent safety principles beyond substitution and moderation/minimization could also be explored.

Summary updated

Abstract

The selection of solvents in the chemical industry is typically based on performance and economic considerations. Other relevant aspects, such as the safety implications of a given solvent, are generally left for consideration after the design of the process has been completed. In this work, an approach for solvent selection including safety considerations at the design stage of the process is presented. The safety component is included through a consequence analysis using an average distance for the risk of death as the major parameter. The approach is applied to the design of extractive distillation systems, for which a preselection step for the solvent is used, followed by the formulation of a multiobjective optimization problem in which both economic and safety aspects are taken into account. The approach is applied for the purification of bioethanol, for which solvents that offer the best cost–safety compromise are identified.
BibTeX
@inproceedings{medinaherrera2014approach,
  author = {Nancy Medina-Herrera and Ignacio E. Grossmann and M. Sam Mannan and Arturo Jiménez‐Gutiérrez},
  title = {{An Approach for Solvent Selection in Extractive Distillation Systems Including Safety Considerations}},
  booktitle = {Industrial},
  year = {2014},
  doi = {10.1021/ie501205j},
}
Figure