Successful Implementation of Advanced Process Control in Cryogenic Plants.
Tom Lyndrup and Greg Rogers
Enterprise Products LLC
Rafael Aguilar and Duane Morningred
Honeywell Process Solutions
Houston, TX, and Camarillo, CA, USA
Paper to be presented at the 2012 AIChE Spring meeting, Houston, Texas
April 1-5, 2012
Enterprise Products Partners L.P is a natural gas and crude oil pipeline company with headquarters in Houston, Texas. It has 18,368 miles of natural gas pipelines, 13,297 miles of NGL and petrochemical pipelines, 813 miles of Gulf of Mexico crude oil pipelines, 26 Gas Processing plants, 19 NGL and propylene fractionators along with Storage (Salt Dome), Drilling Platforms and NGL Terminals.
The Meeker Gas Plant 1 is a new process unit in Colorado's Piceance Basin, started up in late 2007 with a capacity to process 750 MMcf/d of natural gas and is capable of extracting up to 35,000 b/d of natural gas liquids (NGLs). Meeker, the largest natural gas plant in the continental United States, serves the Piceance Basin and other area deposits comprising the nation's fastest-growing natural gas field, increasing export capacity and providing gas to downstream markets through four major pipelines. Phase II of the complex (completed in the third quarter of 2008), doubled the processing capacity at Meeker to 1.5 Bcf/d of natural gas and 70,000 b/d of NGLs.
Enterprise Products wanted to maximize their Meeker, Colorado Cryogenic I plant's recoveries. This plant is subject to continuous changes in production plans, feed composition and pipeline constraints, which makes control difficult. The goal was to increase the recoveries by at least 1%.
The solution at Meeker Gas Plant utilized Honeywell's Multivariable Predictive Control applications that have steady state models. Step Testing was minimized due to the controller characteristics and the additional tools for online model ID used during the project implementation.
Since the initial commissioning, Enterprise Products has continued to enhance the applications finding new sets of variable limits and automatic update for optimization limits based on process conditions. In addition, the operator workflow has been improved, since the focus is now to understand the unit overall problem and search for a solution, instead of changing limits and setpoints to multiple controllers all the time. APC displays have been effective to show the current process constraints to increase production, as well as improve operational consistency across the multiple operation crews.
This paper summarizes the existing APC and Optimization solution and discusses the details of how the control strategy was recently implemented.