Abstract: Natural Gas (NG) is a viable energy source that is dependent on pipelines and other infrastructures to reach consumers. To overcome this problem, the gas volume needs to be reduced so that it can be transported through Liquefied Natural Gas (LNG) carriers and pipelines. Currently, the liquefaction process employed is energy intensive as NG needs to be cooled until it liquefies at -162˚C and atmospheric pressure. Hence, the main objective of this research is to reduce the energy consumption of an LNG plant in order to minimize the associated operational costs. The research methodology undertaken was divided into four major tasks: simulation of Propane Pre-cooled Mixed Refrigerant (C3MR) process, mathematical model development, testing of model performance, and validation optimization results. The plant was optimized following the technique of Nonlinear Programming (NLP), where model CONOPT in GAMS was utilized for solving the objective function set. By varying the feed conditions, the maximum profit generated was at $842,467 when the optimum feed temperature was 38°C, pressure at 609 psi, and flowrate at 44220 kgmole/hr. Once, the performance of each variable was tested, the optimum conditions for variables were compared to the initial ones inputted in Aspen HYSYS.
Key words: LNG, NG, CONOPT, GAMS, C3MR Process.
Methodology: The research method is divided into four major tasks: simulation of C3MR process, mathematical model development, testing of model performance, and validation optimization results. In the first stage, a standard case of C3MR LNG liquefaction process will be simulated through Aspen HYSYS and later analyzed in order to duly understand the C3MR process, identify potential optimization variables, and select the best optimization procedure in order to achieve optimal operating conditions. In the second stage, mathematic model will be developed while adhering to the specified constrains. Basically, an object function will be used as a scalar quantitative performance measure, and then a set of equations and inequalities referred to as constraints will be employed to define the performance limits of the C3MR process. Finally, the selected variables will be manipulated in order to satisfy the constraints. In the third stage, the mathematical model will be tested through a suitable GAMS solver. Lastly, in stage four, after the optimization results have been gathered, they will be compared to the initial simulation conditions.
1. Simulation of C3MR Process: Prior to creating the C3MR flow sheet, literature review was conducted to understand the significance of this configuration, analyze the process to determine the data required for certain variables and identify optimization potentials. The flow sheet used followed the structure of that created by Hatcher, Khalilpour, and Abbas (2012), which can be seen below. The simulation properties were obtained from example LNG Plant simulation provided through example cases in Aspen HYSYS. The C3MR configuration was chosen for optimization because it is most widely used efficient MR process that uses Propane to precool NG followed by MR cooling (Hatcher, Khalilpour, and Abbas 2012).
2. Mathematical Model Development: In this stage, the mathematical model was developed based on mass balance and energy equations that represent the main unit operations while adhering to associated thermodynamic property functions. Peng-Robinson equation of state was employed as it is the recommended thermodynamic package for gas processing, refinery, and petro-chemical applications (AspenTech 2011). Sensitivity analysis is conducted to obtain energy equations. In this report several case studies are presented to fully analyze the performance of the LNG plant.
3. Testing of Model Performance: The performance of an LNG plant is identified through the work required per unit mass of gas compressed and liquefied and fraction of the total flow of gas that is liquefied (Dash 2009). In this stage, CONOPT solver in GAMS was used to solve the optimization problem, with the aid of the mathematical model. Basically, the process variables, parameters, scalars, the objective function, equations, equality and inequality constrain were specified. Then, all of the equations have been inputted in GAMS in a particular syntax for it to run. Here, several variables such as product cost and flowrate, raw material cost, feed flowrate, temperature, and pressure, Stream 3 and 5 temperatures were analyzed in order to evaluate their impact on the profit of the plant.
4. Validation of Optimization Results: This stage was directly related to the previous one as it solely depends on the optimization done on GAMS. Once the results were obtained, they were compared to the initial variable conditions that were inputted in Aspen HYSYS. This was basically done through conducting sensitivity analysis again to see the degree of improvement. In the case where the required energy reduction requirement was not met, troubleshooting will be done until the optimization solution has been validated.
Important Remarks: This page only includes Project Abstract and Methodology. The full undergraduate thesis and CUTSE Conference paper are available upon request. [Project Duration: March 2013 – November 2013]
Acknowledgement: This work was fully guided by Dr. Mesfin Getu (project supervisor) with the support from Curtin University, Miri Campus, Sarawak, Malaysia.