A Computational Model to Simulate Decoking in Steam-Cracking Furnace Coils: Model Calibration and Validation
AIChE Spring Meeting and Global Congress on Process Safety
2017
2017 Spring Meeting and 13th Global Congress on Process Safety
The 29th Ethylene Producers' Conference
Ethylene Plant Process Control
Monday, March 27, 2017 - 1:15am to 1:38am
Yoshitsugi Kikkawa,
Kikkawa Giken, Yokohama Japan, yoshikikkawa@yahoo.co.jp
David Hill
Manager of Technical Support, Chemstations, Houston,Texas
Abstract
The physical properties of process fluid are essential for the design of the natural gas liquefaction process. The K-values are necessary to determine phase and vapor density; liquid density is also important. Accurate K-values for natural gas can be calculated by equations of state (EOS) such as SRK or Peng-Robinson. Estimation of liquid density is less accurate, particularly at higher pressures where a liquid phase is undesirable for process operations.
Various thermodynamic models are compared for accuracy of K-Value and liquid density. K-Values and densities are compared using SRK, Volume Translated Peng Robinson (VTPR), and Scott Fuller SRK (SFSRK). The SFSRK model demonstrates accurate fit for K-value and liquid density across a wide temperature range.
The VTPR equation is considered successor to PSRK, and can be used to calculate both K-values and liquid densities. Twu alpha parameters are used to tune the model. The VTPR equation demonstrates improved liquid density calculation for LNG at normal operating temperatures, when compared to SRK.
A corrected SRK-type EOS has been developed which improves estimation of liquid density without sacrificing accuracy of calculated K-values. A repulsive force term for SRK is developed, based on knowledge of the hard sphere fluid. The SFSRK model revises the molecular size parameter included in the attractive force. The SFSRK model demonstrates accurate liquid density calculation for LNG at high temperature and pressure.