A Fuzzy Feed-Forward/Feedback Control System for a Three-Phase Oil Field Centrifuge.
- Ronald E.
- Fred N.
- Paul J.
- Neal
A set of fuzzy controllers was designed and applied to a commercial three-phase oil field centrifuge. This centrifuge is essentially a one of a kind unit. It is used to recover oil from tank bottoms and oil field and/or refinery sludge. It is unique because it can separate oily emulsions into three separate phases, oil, water, and solids, in one operation. The centrifuge is a large but portable device. It is moved form site to site and is used to separate a large variety of waste emulsions. The centrifuge feedstock varies significantly from site to site and often varies significantly during the daily operation. In this application, fuzzy logic was used on a class of problems not easily solved by classical control techniques. The oil field centrifuge is a highly nonlinear system, with a time varying input. We have been unable to develop a physical-mathematical model of the portion of the centrifuge operation that actually separates the oil, water, and solids. For this portion of the operation we developed a fuzzy feedback control system that modeled a skilled operator's knowledge and actions as opposed to the physical model of the centrifuge itself. Because of the variable feed we had to develop a feed-forward controller that would sense and react to feed changes prior to the time that the actual change reached the centrifuge separation unit. This portion of the control system was also a fuzzy controller designed around the knowledge of a skilled operator. In addition to the combined feed-forward and feedback control systems, we developed a soft-sensor that was used to determine the value of variables needed for the feed-forward control system. These variables could not actually be measured but were calculated from the measurement of other variables. The soft-sensor was developed with a combination of a physical model of the feed system and a skilled operator's expert knowledge. Finally the entire control system is tied together with a fuzzy-SPC (Statistical Process Control) filter, used to filter process and instrument noise and a fuzzy conflict resolution code used to keep the feed-forward and feedback control systems working well together.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE
- OSTI ID:
- 976192
- Report Number(s):
- LA-UR-02-3225; TRN: US201009%%592
- Resource Relation:
- Conference: Submitted to: The International Conference on Egineering Design&Automation, Maui, Hawaii, August 4-7, 2002
- Country of Publication:
- United States
- Language:
- English
Similar Records
Comparison of soft computing techniques for a three-phase oil field centrifuge.
Development and Implementation of an Automatic Continuous Online Monitoring and Control Platform for Polymerization Reactions to Sharply Boost Energy and Resource Efficiency in Polymer Manufacturing